Ae] L_ Saad af

mal a ee ' Pt wz |

} ) ga peat ae ay.

S24 et & Gee pseee™

-hjs i ) ;

j a erp ASV CU Ee if A MT He taby

‘ t pu pepe Peary ‘

) j ‘ hd ; ere shew #)) y “yes eu

' Bye ‘ mt he Anh eet Ue Ga

ts {+3 yds ey Ly) eee oe “ i

| ' ! Pay ataad at aatoe sea Wihett rip lob ign hi pa ye Fea" meshes foes tei sir? Ks

’ a i i t bi) ty fae Car Heirke telat pelet ate Tsataeett bers ”

j taf ' ; : ud Ba de 44 ps ge bade sat lal @ oe ere Y

, (dd : } ee | } 7 Nake date Vai peo kel Cry Sa is perig ciate vel ne tele be gaye:

5 ; We x4 ‘ SP bka + Get ie iy Bice pe bal eoreeercoae Le hedged

| | eat i Pid Neale i iranian Mittal a rs Replat iets

1 ( ' y , rd j ‘ ‘ a F } 4, Pub i neaei sy Has iy) a4 : Renae Lt dete f, tote

‘ 4 ita ‘ dedi) v

‘ ako ; a tyne vite ie He an .

Ae : 4 ot Bele 4 nae uM

‘ ; An ps 4 Be oh

Pity tg } | me

; i 4 SPte ep Ue i i mon + asaega.ate bpd i de

Vahey bat oro ports tet kt tot sats bet

Ab lite ofa babs isin ie ahd begets Let “} ee

Svitu lan ad eg eo be fepeta re

te he be)

A! re Ste bodet

mths 44 : Mii st

b pep ae

igh dee BN 1

das

: dibsts: ag pote

Let pearts) n ‘st sali

Liye

Huh

Pas

Biren

Mia

Pisce a

Bers

ore ;

pean

ee.

)e'b

eye

eet Wee eres

’

ih i sat

fate wi

reid i ,

gts Pee i

Se +

eid ah Wad a ae Uist

i eieaindoteenseere

eae Spabsekced : bibeee

Pek

tone dy we dee

shad ee

te ree et

Masierche pt 4 ”

RELL Stat V9

ia ae geet

aba de ie es

Fal ares nent wha We i

i. , wi Ay

ett Ay nee P| Me A? eaten sass fat

SerPrht ey i tat eat ecek a bat il eptea Es et vies

f niet Lat aay Ha : ;

inane aise sy

Arg tohe tee A Pert eiea

ie C fiat i ae

4 nate #1 a pelaivisat

see ae Aly

Peat Tae,

$4314 ad Dd

3 Meat Pht $ HAY eed

Vi suatat hy naade inte A

ie Oia ta

ei is Pier ard it Woe i

} }

t ’ he i Te SP f j i

Vistula ida? badd ‘

‘ 4 vei ’ ‘

5 TLR) betel, & ok ' ! ad Pad wean citer

] ‘ rene tive ry } pharm ee eat bo

' ye"y"4 na MM aaa! 4 jad iy ji

ry ) rene , ’ Dud 4A" ieee

Mie A: ’ it) syavsj) Me rt LT oredr

; , ; ; j j oe] Wal vil ey wi

i ' 13 pal reper

; p : ; { b) '

/ oes aul i

\ t beh | ’ dd tae Suess ?

‘ toe “ uit #2 he 4 oe a4 if ees

ii 5 ia ‘ ha? 4 ie) { Ne it K ut tie eet

‘ j Fi t paaibiasae pate ti dist

ye. ; re ae | ; ‘ AeA atbtaneeotitat

i ' ia to . ft +) ae das ae Mor ite rf

" iii } 4 ) Fe ate DB OES Fatiph bev atwted Saneeuee

{ bd jd atat ‘ bie Pa) ee Seatatnert a aane

a) { } : bere "

} } bag Taye er TT HCL) Bi :

vi ' ; fi ee sisted: ees re

j : J y HG tek St Dbl ia taeti that Fel

; } 4 } ‘ 4) ti

aig” j } if - ‘4 : ; HA she os dato te Ns hws vee

5 1 PR SAIROR EPEC SEC tee ry Vee ye vg OG Z rt aN. 7 tab ebyr rh

i? } } j j ‘ gated ‘ § ! aT) rie Oe i; sods koe ‘ serene. id an atte

F é ' s34 44 ype oe oid j Pe dy Pb 8 eb ale as a yw f] uty

eet PRD te Wu Ie bi dad Hi Rr tC ron Dit 1) uate ne ie Ses Rand it

' yi ste th? i i} lh ie yb xmeoee Fb b eh gH Hy rhe ' yr ER at at aHie tit ohh ee

{ 1 ait 4 Huse ay ‘ yi } Be a De Ay ae ies ey can Mra gt: baajhtat Ete

) va) rey a oa, bakes { ay i aed a) 4 ¢a aw ; Seateyet estas at + ve steer reid

ad a) hes ij ' ; my Le pei ‘ eG nee His vel a vipibeyresitenals Made by spl

; ‘ Hit t WG) a ed) mi" eye ‘uty aM 4 soe jeanaya cal sa) pt bruret yeaa Aone vee ae

, I} ‘ ia ryt ashycoald heh ha aE a vial bet et sotige i epee tet

4 ab ' } if} wie We 1 1! Yiaid Abie “4 wre as yedtrtds dace nla ge a naa

whi } ‘ A seuliwe edd ih pe? at ends PGads ea celta

OB tes calli en ii AULA i neath situate ti

’ 4 i i wis a r| ete att LT

A, at v4, ier enh ite aetyioekerot wanenaiones tags

4 1 72 i ' i.) t ' oy ty ied jn atiay, doweniaads

, Hae ay igh eaiyit Roe via) i At ita ta eae Hydeet

j ‘ ‘ ‘ “aoa y ie theo Hed a erred

‘ ns 14h) earns ' td vhsso a} te! myth Pala tebatinns eh Rarity

' aft Pe | wiht j vb Tee 3 pty! Ge “ {hea

‘ ‘ Vw . Pe ou ey nal ‘ narod beat HY aa aaa uaas

ei . j ‘ pie 4 ' ‘ ir od pet ne we ¥ He wht ee ia Hs ret ta

aoe) } i 5 =f bl ' 4 ' My hoe wMin wt La hie apdatah earn ® onieede f eye ay ire nt i

} vy) ii tale } Mr AC CR REDT Ph cd PERLE oh aetce ta ive Keyed ta aie trie (EY ne aye ae Oo (PEE

4 j i 4 ‘ j ‘ PER UNIT ELEC ER CATED tea ’ . pees fo - cae

‘ yt vi ; ‘ f ti ' | ' a POPOM MODE CE toes taenes DR aru er er Uh oh ee

‘ i ! of ‘ MHS vv! i ede a Me macnn oe ir

‘ ; ) ' Lee ivy opel bey a aca. ‘ PA PDO et Leth Lehn kag

‘ j F ee j t ave { ya Vc dota be id

; ‘ 4 ‘ ii ‘ ' | vit ‘ Vp eb rte deded teed a

‘ ‘ epee Bi | ' ‘ vbw AMA sa ES oc one! Daag et?

Men? ele ee a i i ’ pot bli te) rea eed ee eee a teint 4 ps teh hy

ee ae ‘ i ' | ®. +t yer ; : eh} > att Et oe i \* i / q . *

j ‘ ' y F yey ee ad HO HN HUT rr ee Ciao

7 hp ae? a a Mae WW UPA DOE dA Pe Be | PEATE BL YA Ub SE CL DEL Cases Cet f ire hae eg 9 404 BRE aS

" ) eu tr Vy ih) Re Gey ‘ y abelian ey, Plena bad edad: ry

‘ ‘ ' ya beWi tse ve} ‘ vat 5 us

" ae hy Af) MRED RITA OHA THERA oh od. esa rhe hei Brae oy Meer eoeitateaa iae dl bd mY baby py

Wits i" ya vals ; han pa | 3 yew ‘A wade Sat rivee ptitall ‘ Rant eee dae we }

P | i tia eta } yy i shed Wb ote dg ee Ok Dat bbatwr Ape Ae: ede ni ras ae wea! waite 2

ay reld i.) ri ba PeRMTATHUA HEAL Sides, 4 Paigatandes " ne alate

vat ae hile} Ea PAPA AL LC ne nt

, : ‘rh i i \ “4 ‘ Vibe pb eewye beh oo be tha ye pa Oe aye

rani ' 4 ( 4 1) ' eh Vie td We a Ge te tet tate pp

Oe ee he yea vane mre 1 Pie ee paoes pasa} seietehenians iy i tate

‘ re | ' ‘ an ' 4 a) ayet ' ra Pde Eb A yaaa | Phe ae

hd eye Lj i id PPM TA PRT WERE EE Chi BE RT AC I daeh ; i iy): ’ raceme i atad ri

‘ Ne ya N ran pal ; H RCLOSVHERD LG Petar ba eit Be Gat { cape dt a Hig PasMtdit sotaversteus ges de Btwn Bs

' eva Vie ty POR WONG Fureabt Ve yy) VERY pre Tat ee We it ae ieee yee Lie Gaile i PF rentrererecth Mont

7 | Pa ha a Vibe ' bay tho Dyed diode Lehae ida yd Oe eee ye bide de rei eee ceoabe de dei faa enor

vives ' POVETE LT SM LP RE ab ed Py Dye EU hed te tele Lei iet Pee He t) 4 ye Svea ee ee bee Oe

ve ‘ yh yey yyw ivi hie nit y deb tgeQeetbat be E need HUE rere taigen ro

r ’ ' ‘ ert TUE TEC VM aD RAL LH Pa Bee 4 ~

a" , ARM eine By aratatuiae DAL Phra ta beh So adi bedanraae i Veda aaa NRT TA) me cunine

' ' ' 't ’ ebay Vee ' ' vitthe ro wih) He ” wean! 4 lob q

ei vai he LE EAN OM Ye 0d $ ht beat Th : vied rer yy yatta ae J woatenel Le as Na ug a woes Wp baearette Of

i berate bere ‘ eed be a eee CU Reed RUBLE Gielen yr Dede FS the pray Wate MeO N Gs bas FF Ryne ie Wer hin SCA he alee ie

; is 8 Lk pied ‘ bev od ee De Pee a 40) ba dal Tae ha 20g see Cn aia se aft yprase bebe waite on te ee

' vey wil ' ' wi bite be eg be Rt Tighe be fe ed 2 a Me pig ae gaat weve thet tet Ni gues ba wanes eines a“

' yee i ve a red Pe eae pk ae pent } ev Ree pe a) : sy betaiaiebene eheaeiat? itotes nbaeaitbure ie eeachon dhe

‘ i ’ j ‘ ‘ vt ‘ r4 1 re mt yrs vie tr pees te ae o es heleatatt Tate bey epee ee

prrvevewe ‘ tw wi he nd A hitews te ae bbe) jie beeps ba yalbeld : ‘tenet Set ede bene stingdt ede ee

ea wa ' ‘ ‘ jieet ot ‘ 1) Lie hee ee ’ uy uve avporbete ewes CL et

45 ory muy 4 » ' pomen ' J walhote ‘ obs) bntetely teehte yaa

y alk co Ay dee he ik bela t Siete tebe b Watet ty) ian fae yregy an debe

a Ala, boihe iba Rabe he eg ald 2 yan eh ye bm ity he Hyg the b

Hegre heed jeg he we

MOPPT eC Prer rn M ru dl au

Nal ne

Aaa

aRaaAn AR

ie) |

RASS <ARA8 meas?

‘a al VY ey

AANA. al |

vwy A pl t aaalel aanam a

east tettnatannann

i leealete ‘i RAaaa-

} . e a

vane Sea nar malta | yy

2 an aA Amer am r

See annannnens aeesanr

. : NAN Neca! YY

oY FAY NAA saa NAaASARARAAAR AS. teictel

SRaABAAARARann. ARRRAARA ~~ ~ a

Nadel

YN a)

mannan aris:

Aa aaataAn a

‘fr

Aa ioe | ro» Ae a a ae

LPR. NA

—_ 7 ry Oy hp is

z, SIS

io cee

1" —T™| RoONAe "

CL |

w w LAR

~Oneen

vw we veer peetesesee

Sr Se

pais srr aS

2 al >

wey #

e -

e ~.

ee os <5

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLGME XX VITI

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME XOXV LIT

PRINTED FOR THE

TRUSTEES OF THE SOUTH AFRICAN MUSEUM

AND THE

GEOLOGICAL SURVEY OF SOUTH AFRICA

BY NEILL AND CO., LTD., 212 CAUSEWAYSIDE, EDINBURGH.

1929 — 1932.

Le i at oh ge a Phe

AP Wigs as Ave 9

~ : nar ‘ay Np (pep ais ene Hash raat

* e t ry Le ele i ae bd , by r

: mets | :

TRUSTEES OF THE SOUTH AFRICAN MUSEUM.

Sir THomas Murr, C.M.G., M.A., LL.D., D.Sc., F.R.S.

Professor WILLIAM ADAM JOLLY, M.B., Ch.B., D.Sc., F.R.S.S.Afr.

J. G. VAN DER Horst, » Comm.

W. J. THORNE.

Councillor 8. J. HENDRIKSzZ.

SCIENTIFIC STAFF OF THE SOUTH AFRICAN MUSEUM.

Epwin LEoNARD GILL, D.Sc., Director and Keeper-in-Chief.

KEPPEL Harcourt BARNARD, M.A., D.Sc., F.L.S., Assistant Director ; in Charge

of Fishes and Marine Invertebrates.

Miss Star GaRABEDIAN, M.A., Assistant in Charge of the Botanical Department.

REGINALD FREDERICK LAWRENCE, B.A., Ph.D., Assistant in Charge of Reptiles

and Batrachians, Arachnids, and Myriopods.

ALBERT JOHN HeEssz, B.Sec., Ph.D., Assistant in Charge of the Entomological

Department.

LizuwE Drrk Boonstra, M.Sc., D.Sc., Assistant in the Palaeontological

Department.

SipNEY Henry Haveuton, B.A., D. ah ., Honorary Keeper of the ee

Mineralogical, and Dalcontolneical Collections.

LIST OF CONTRIBUTORS.

L. D. BoonstRA.

On the Pareiasaurian Manus

On the Pareiasaurian Pes

On the Hind Limb of the Two Little- ae iParcincanaon Genera :

Anthodon and Pareiasaurus : :

The Osteology and Tas of the Lonomonen Wppasicees B.—Fore

Limb

F. CHAPMAN.

On a Foraminiferal Limestone of Upper Eocene Age from the Alexandria

Formation, South Africa

AY LS puehom:

Some Fossil Plants from the Karroo System of South Africa

Some Fossil Plants from the Gondwana Beds of Uganda

T. W. GEVERS

A Tabulate Coral from the Bokkeveld Beds of South Africa

. H. Haveuton.

On some New Therapsid Genera ;

Notes on some Pareiasaurian Brain-cases .

The Fossil Equidae of South Africa .

S. H. Haucuton and L. D. Boonstra.

An Attempt at a Classification of the Pareiasauria based on Skull

Features .

On the Pareiasaurian Mancables ; :

The Osteology and Myology of the Loconteen Apparatus, A.—Hind

Limb : : 5 : 5 : 3 :

J. V. L. RENNIE.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda)

New Lamellibranchia and Gastropoda from the Upper Cretaceous of

Pondoland (with an Appendix on some Species from the Cretaceous

of Zululand)

L. F. Spats.

On the Cephalopoda of the Uitenhage Beds

PAGE

113

429

437

291

369

395

123

159

131

LIST OF NEW GENERA PROPOSED IN THIS VOLUME.

Brachypareia

Bradysuchus

Choerosaurus

Dolichopareia

Eoarctops

Eretmonia

Hipposaurus

Koalemasaurus

Lycaenodontoides

Nochelesaurus

Notohipparion

Platyoropha

Styracocephalus

Trigonocallista

PAGE

Cotylosauria (Reptilia) HAUGHTON and BoonstRA 84

Cotylosauria (Reptilia) ne as 82

Therocephalia (Reptilia) HAUGHTON 60

Cotylosauria (Reptilia) HAUGHTON and BoonstrRA 83

Gorgonopsia (Reptilia) HAUGHTON 68

Cycadofilices (?) pu Toit 381

Gorgonopsia (Reptilia) HAUGHTON 75

Cotylosauria (Reptilia) HAUGHTON and Boonstra 84

Gorgonopsia (Reptilia) HavGHTon 70

Cotylosauria (Reptilia) HAvGHTON and BoonstRA 83

Perissodactyla (Mammalia) HauGutTon 421

Cotylosauria (Reptilia) HAvGHTON and BoonstRA 84

Styracocephalia (Reptilia) HaucutTon 55

Lamellibranchia (Mollusca) RENNIE 197

DATE OF ISSUE OF THE PARTS.

Part 1. August 1929.

Part 2. February 1930.

Part 3. August 1930.

Part 4. February 1932.

vii

XITI-XV.

XVI-XXXI.

XXXII-XXXVI.

XXXVII.

XXXVIII.

XXXIX.

XL.

XLI.

XLII-XLIV.

LIST OF PLATES.

Albian Fossils from Angola.

Senonian Fossils from Angola.

Manus of Bradysaurus vanderbyli.

Carpus and Digits of Pareiasauria.

Manus of Pareiasaurian.

Proximal Tarsal Bones of Pareiasauria.

Epipodium and Basipodium of Embrithosaurus schwarzi.

Pes of Brachypareia rogersi.

Pes of Bradysaurus vanderbylv.

Cretaceous Cephalopoda.

Cretaceous Lamellibranchia and Gastropoda.

Pareiasaurian Mandibles.

Eocene Foraminifera.

Hind Feet of South African Pareiasaurs.

Stigmaria ficoides.

Karroo Fossil Plants.

Uganda Fossil Plants.

Fore Limbs of Pareiasaurians.

Actaeon

Actaeonella

Alectryonia

Anchistoma

Anisodonta

Anomalina

Anthodon

Arcotia

Astarte

Avellana .

Baiera

Barbatia .

Baroda

Belemnopsis

Bochianites

INDEX OF GENERA.

389

155

153

Brachypareia 84, 99, 116, 272, 330, 472

Bradysaurus 81, 94, 99,115, 276, 312, 455

Bradysuchus

Bulimina .

Bullinella

Cardita

Cardium .

Cercomya.

Cerithium

Chenopus

Chilodonta

Choerosaurus

Cibicides .

Confusiscala

Conus

Cornucarpus

Crassatellites

Cristellaria

Cryptorhytis

Cyclodendron

Cyclothyris

Cymbophora

Cypraea

Cyprimeria

Cyprina

294

233

1 220)

36, 196, 249

. 23

PAGE

Dicroloma 217

Discocyclina 296

Discorbis . 295

Distoloceras 5 : 3 Ld2

Dolichopareia 83, 95, 269, 329, 469

Dosiniopsis : : ss AOD

Embrithosaurus 85, 93, 116, 326, 467

Hoarctops : : =) i068:

Hodesmoceras 141

Equus 410

Eretmonia . ool

Eriphyla . 29, 245

Kurygnathohippus . 419

Kutrephoceras . : 139

Exogyra . 16, 177

Fabularia 293

Fusus 46

Gangamopteris . 372, 397

Gaudryina . 294

Globigerina . 294

_Glossopteris 378, 398

Glycimeris ait ETA.

Goniomya = 238

Gyrodes 40, 212

Hibolites . 156

Hipparion 419

Hipposaurus 75

Hoplitides 151

Inoceramus 13

Johnstonia 390

Index of Genera.

PAGE

Koalemasaurus 84

Lathyrus . . 224

Lima : 22, 183

Liopistha ,, ABO

Lithodomus 22

Lucina ‘ 31

Lycaenodontoides 70

Lysis 211

Macoma . ; . 201

Macrocallista 37, 198, 246

Mactra : 205

Margarites 208

Martesia . : 205

Metacerithium . 42

Modiola 184

Neithea 16, 179, 241

Nemodon é 3 He,

Nerita : : ; 39, 209

Nochelesaurus . 83, 116, 276, 321, 463

Nordenskjoldia . : A 2) LO

Notohipparion . = Saal

Nucula 3 168, 239

Nummulites 295

Operculina 3) 295

Ostrea lays 175)

Palaeomoera 200

Palaeosephaea . ; : . 228

Pareiasaurus . 86, 338, 429, 477

Pareiasuchus 86, 88, 270, 335, 475

Patella : : 200

Pecten seeelaS

Pholadomya 23, 187

Phylloceras . 140

Phyllotheca 374

Pinna 172,

Pirula 3 : 5 ca ee

Platyoropha 84, 116, 324, 466

Pleurodictyum . : : . 128

PAGE

Pleuromya 22, 186

- Plicatula . 20

Polymorphina . 294

Polystomella . ; ; - 295

Propappus . 87, 116, 280, 332, 474

Protocardia : : - | ee

Pterodonta 45

| Pterophyllum 392

Pugnellus 218

Pyropsis . 229

Quingueloculina 293

Ringicula 234

Rogersites 142

Rotalia 295

Schizoneura - tenes

Semifusus 230, 249

Semisolarium =) 20M

2 | Sigmoilina 293

Siphonalia 223

Solecurtus : 202

Sphenophyllum 376

Sphenopteris 377

Spiroloculina - | 20a

Spondylus 215 182

Stigmaria ~ oe

Strombus : 44

Styracocephalus 55

Al;

Tellna 38

Textularia 293

Trajanella : . i200

Trigonarca 9, 169, 240

Trigonia . : ~ -aaKG

Trigonocallista . 197

Trillina 293

Triloculina 293

Tritonidea - 222

Turritella 41, 214

Veniella 26, 191, 242

Venilicardia 2 23) 25

4 wes

ay be

a Sis

= rR

THE 2 GEOLOGICAL SURVEY OF SOUTH . AFRICA

me ontaining a ae

Cretaceous: Fossils from Angola Cfo) haniinn and

=e Gastropoda). By Joun V. L. RENNiE, M.A. (Cape Town),

: “Pee Student, Emmanuel ee Cambridge. (With

as some Ne ew. eee Gok By Sy Ee tenon

Oi aoe Hon. a Pape eign Collections.

sauria a on Skewll Features: a S. H. Havcuton,

_B.A., D.Se., Hon. Curator, and SABE BoonstTRa, Mse.,

Assistant i in Palaeontology. 3

=! Part a -Pareiasaurian Brain- Cise.

Be IBY i

- Part TIL. —On the Pareiasaurian ees By ‘Limowe:

- - D. Boonstra, M.Sc., Assistant in Dalene

mee South African. Museum. ube: Plates and

-Text-figures 27, 28.)

- Part IV.—On the ue saat ian Pes, By Lizuwe D..

- Boonstra, M.Sc. Cee Plates —— and Text-—

figure 29.)

ei Tabulate Coral ao the Boleleaietél Beds of South A frica,

By’ - Le ‘GEVERS, M.A. ,D. Phil. (With Text- foots 30.)

ISSUED AUGUST 1929. PRICE 10s.

PRINTED FOR THE’ % ME bie

S OF THE SOUTH AFRICAN ee

St res

AND THE pees nies

LOGICAL SURVEY OF SOUTH AFRICA | i

ILL AND ©O., LTD, 212 CAUSEWAYSIDE,. EDINBURGH,

By S cae Bn on B.A, D.Se., Hon. Curator of ©

ea Siena toligi¢al Collections. | (With: a oie

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME XXVIII.

1. Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda).—

By Joun V. L. Renniz, M.A. (Cape Town), Research Student,

Emmanuel College, Cambridge.

(With Plates I-V and Two Text-figures.)

SoME few years ago the South African Museum acquired, through the

courtesy of Dr. L. Reinecke, an interesting collection of Cretaceous

fossils from the Loanda province of Angola, collected by the field

officers of the Companhia de Petroleo de Angola. The Cephalopoda

and Echinoidea of this collection have been described by Dr. Haughton,

who was able to show that an Upper Albian horizon exists in the

Loanda embayment containing a fauna essentially similar to that

described by Spath from the province of Benguella, and that the Teba

Formation is undoubtedly Senonian. Previous to the publication of

the above paper the only Senonian forms definitely known from Angola

were three Lamellibranchia from the province of Benguella and a

Didymoceras mentioned by Spath from Barra do Dande.*

The present paper deals with the Lamellibranchia and Gastropoda

of the above collection, and with a small but extremely interesting

* Since this paper was written the writer has seen an important paper by Velez

Mouta and Alexandre Borges, entitled Communication de la Mission Géologique de

? Angola sur le Crétacé du Litoral de V Angola (Districts de Benguela et Mossamedes),

which was read at the fourteenth session of the Congrés Géologique International

(Spain, 1926) and published in Lisbon by the Agéncia Geral das Colonias. The

Cretaceous deposits are described in some detail and extensive fossil lists are given,

including many new records for the area. Roudazrera [= Veniella] drut is figured

for the first time from Angola, from S. Nicolau, and the Pondoland Pseudomelania

sutherlandi (Baily) is listed from the same area. It is doubtful, however, if much

value can be attached to many of the identifications, e.g. Solariwm [ =Semisolarium]

bailyi Gabb, a Pondoland Upper Senonian form, is listed among Albian forms.

VOL. XXVIII, PART 1. 1

2 Annals of the South African Museum.

collection made by Mr. W. J. Reynolds along the coastal strip from

Mossamedes to the Benguella area. Though a large number of the

specimens are very imperfect or in the condition of casts, and many

had, therefore, to be completely ignored, some interesting results

have been obtained, notably the recognition of a Senonian fauna near

Mossamedes and the conclusion that the Senonian of Angola is closely

related to that of Pondoland and Zululand. Some new light is,

therefore, thrown on the palaeo-geography of that period.

The writer wishes to express his thanks to those who have aided the

work from time to time with useful suggestions and criticisms, and

particularly to Dr. 8. H. Haughton, through whose interest the investi-

gation was commenced, to Mr. Henry Woods, M.A., who has kindly

read through the first draft of the paper, and to Mr. L. R. Cox, M.A.,

who has given the writer assistance in the British Museum.

In the following account the numbers in heavy type refer to the

bibliography appended. The South African Museum catalogue num-

bers are also given.

List oF FossiIts.

1. Dombe Grande.

Netthea tricostata (Coq.).

Spondylus angolensis sp. nov.

Crassatellites spp.

Lucina angolensis sp. nov.

Cardium (Laevicardium 2) sp.

Cyprimeria ? sp.

Nerita cf. malhewror Choftat.

Gyrodes cf. gents (Sow.).

2. Catumbella.

Cardium (Trachycardium *) sp.

Chenopus sp.

3. Benguella.

Trigoma cf. ethra Coq.

Neithea tricostata (Coq.).

4. Uchi district.

Trigoma cf. ethra Coq.

Netthea aequicostata (Lam.).

af quinquecostata (Sow.).

a troicostata (Coq.).

Nerita malheiror Chofiat.

Pterodonta cf. inflata d’Orb.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda).

10.

LY.

12.

13.

14.

15.

16.

Chipupo district.

Astarte sp.

Lucina ? sp.

Gyrodes cf. gents (Sow.).

. Camballa area, Cuvo River (various localities).

Trigonoarca ct. diceras (Seg.).

Ostrea vesicularts Lam.

Alectryonia cf. syphax (Coq.).

Exogyra cf. conica (Sow.).

Newthea tricostata (Coq.).

» welwrtsche (Choffat).

. Caputo Hills, Cuvo River.

Metacerithium tremonile (Mich.).

. Cabo Ledo, Quissama.

Inoceramus sp.

. Hills west of well location, Quimbaixie.

Barbatia ? sp.

Astarte sp.

Chenopus (Drepanochilus) reinecker sp. nov.

Mumbondo (various localities).

Chenopus (Drepanochilus) revnecket sp. nov.

Pterodonta afi. elongata d’Orb.

Kast of Dondo—Quixinge Road, Quissama.

Exogyra olisvponensis Sharpe.

Cyprina sp.

Lifune Seep, Lifune River.

Nerita cf. malhevror Chofiat.

Chio (various localities).

Lucina revnecker sp. nov.

Baroda ? spp.

Avellana incrassata (Sowerby).

Muscima.

Baroda ? sp.

Capolo (15 km. S., 75 km. E. of).

Pleuromya ? sp.

Carimba.

Inoceramus sp.

Macrocallista ? sp.

4 Annals of the South African Museum. —

17. Massangano Fort.

Pholadomya aff. tugris Noethng.

Cardium denticulatum Baily.

18. Massangano area, 8S. of river.

Plicatula sp.

Venilicardia cf. nicarser (Coq.).

Cyprimeria % sp.

Nerwia angolensis sp. nov.

Cryptorhytis cf. blerchert (T. and Peron).

19. Near Hombo, Quissama.

Venilecardia cf. nicarser (Coq.).

Protocardia hillana (Sow.).

20. Baba district, Mossamedes Province.

Trigonoarca angolensis sp. nov.

o cf. trichinopolitensis (Forbes).

Nemodon natalensis (Baily).

Venrella drua (Munier-Chalmas).

Cardita barronett Munier-Chalmas.

Cardium (Trachycardium) reynoldsi sp. nov.

Tellina (Palaeomoera *) sp.

Turritella (Haustator ?) cf. acanthophora Miller.

21. Mesados Cavalleros, Mossamedes.

Inma (Mantellum) sp.

Inthodomus sp.

Eriphyla cf. forbesiana Stol.

Various Gastropoda, among which Turritella, Certthvum, Strom-

bus?, Fusus?, Conus?, and Actaeon? occur as casts and im-

perfect forms.

DISCUSSION OF THE FAUNAS.

(1) Albian and Cenomaman of the Benguella Area.—The Albian and

Cenomanian of the Benguella area are well known through the early

researches of Choffat (6, 7) and the more recent descriptions by R.

Bullen Newton (5) and Spath (51) of material collected by Gregory

(24). The age of the so-called “Jnflaticeras-fauna ”’ has been accur-

ately determined by Spath, who states that it probably corresponds

with his Beds XI-XIII at Folkestone, 7.e. near the top of the Upper

Albian, hence need not be further discussed here. No Cenomanian

Ammonoidea are known from Angola, but Chofiat was probably

Cretaceous Fossils from Angola (Lamellubranchia and Gastropoda). 5

right in regarding the beds overlying the Pervinqucerta-bearing * lime-

stone as mainly Cenomanian. The specimens in the present collection

from this area were unfortunately not collected with a due regard to

horizon, and therefore most probably include both Albian and Ceno-

manian forms.

The localities Uchi and Chipupo are new. The occurrence at Uchi

of Trigonva cf. ethra, Nerthea aequicostata, N. tricostata and its varieties,

and Nerita malheirot, all of which occur in or below the Pervinqwmeria

limestone at Dombe Grande or Lobito, suggests that the Albian only is

represented among the forms from this locality. The Chipupo shells

are casts, but include Gyrodes cf. gents, which is an Albian form.

(2) Albsan of the Loanda Embayment.—The presence of the Upper

Albian in the Loanda Cretaceo-Tertiary embayment has been well

established by Haughton (28), who has shown that beds containing

Pervinquierta occur at Benguella Velha, Camballa, and Cabo Ledo.

The Albian age of the Camballa fossils is confirmed by the presence

there of Netthea tricostata, and also of Exogyra cf. conica, Trigonoarca

ef. diceras, and other forms. New localities are Caputo Hills, where the

undoubtedly Albian Metacerithium triomonile occurs ; Chio, from which

comes Avellana incrassata ; and Lifune River. Exogyra olisuponensis

is in North Africa a Cenomanian-Turonian form, and the ages of the

fossils of the remaining localities (9, 10, and 14) are doubtful.

(3) Senonian of the Benguella Area.—In the province of Benguella

the presence of the Senonian was first definitely established by Chofiat

in 1905 ('7), when he described Cardita barroneti Munier-Chalmas and

Roudaireia forbesiana (Stoliczka) from the uppermost horizons of

the Cretaceous of Dombe Grande, though he had recorded from there

in 1888 (6, p. 26) sandstones containing Ostrea baylei Gueranger

associated with impressions of Roudaweia and other forms. C.

barronett is characteristic of the Maestrichtian of Northern Africa,

and R. forbesiana [ Veniella} occurs in the Lower Senonian of India, the

Coniacian and Santonian of Tunis, and has been recorded from the

Manuan Creek area of Zululand, where Spath (48, p. 302) has recog-

nised Coniacian Ammonoidea ; the latter occurs also in the Campanian,

as the South African Museum possesses numerous fine examples from

Umkwelane Hill and Lake Itesa, Zululand, while there is a single cast

undoubtedly representing this species in the Transvaal Museum’s col-

lection of Pondoland shells. The data relating to the ages of the two

horizons recognised by Chofiat are very meagre, but as C. barronets

is recorded in this paper in association with forms related to those of

' * Pervinguieria Bohm 1910 must replace Inflaticeras Stieler 1920. See 52, p. 79.

6 Annals of the South African Museum.

Pondoland, it seems probable that the upper horizon, Choffat’s Gres

a C. barroneti, is equivalent in age to the beds at Baba and to the

Umzamba beds of Pondoland, which are usually regarded as Cam-

panian. No other Senonian forms are known from this province, nor -

are any present in the collection here described.

(4) Senonian of the Loanda Embayment.—The wide extent of the

Senonian in Angola was shown in 1925 (28, p. 265) by Haughton, who

described from Carimba in the Loanda embayment several Ammon-

oidea of undoubtedly Senonian aspect and probably Upper Senonian.

The Lamellibranchia from the Carimba localities give no additional

evidence as to the age of the Carimba fauna.

The Massangano fauna is without doubt Senonian on the evidence

of the forms listed above, and it is therefore interesting to note that

Haughton has described Placenticeras reineckei from the Massangano

area, a species which he has compared with Upper Senonian forms.

Cardvum denticulatum is common in the Campanian of Pondoland.

Pholadomya tigris is a Senonian form, and Nerita angolensis sp. nov.

appears to be more nearly related to N. rugosissima Forbes from the

Ariyalur group of India than to any other species. The Cypromeria is

very like C. analoga (Forbes) from the Trichinopoli group and may

be identical with an undescribed form from the Campanian of Pondo-

land in the collection of the Transvaal Museum.

Of the two species from Hombo, the Venilicardia has been compared

with a Senonian species and occurs also at Massangano, while Pro-

tocardia hillana is a long-ranged and widespread species which is found

also in the Pondoland deposits.

(5) The Senonran of the Mossamedes Littoral.—The collection made

by Mr. W. J. Reynolds in the neighbourhood of Baba, near Moss-

amedes, includes only eight species, but these are very well preserved

and of exceptional interest, as they not only prove the existence of

Campanian deposits far to the south of the hitherto described Senonian

horizons of Angola, but open up interesting questions in connection

with the relationships of the Senonian of Angola to that of Pondoland

and Zululand. Nemodon natalensis is a very distinct species

characteristic of the Pondoland and Umkwelane Hill deposits. T-

gonoarca angolensis sp. nov. has its nearest ally in T. capensis from

Pondoland. Veniella drui has long been known from the Campanian

and Maestrichtian of North Africa and is represented by fine un-

described specimens in the collections of the South African and

Transvaal Museums from Pondoland. Cardita barroneti is from the

Maestrichtian of North Africa, while the Turritella will be compared

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 7

with a Maestrichtian form from Aachen. Trigonoarca trichinopoli-

tensis occurs 1n the Lower Senonian of India, but has not been recorded

from the Senonian of South-Hast Africa. Cardiwm reynolds: sp. nov.

appears to be a very distinct form.

Three of the eight species from Baba are represented in Pondoland

by identical or closely allied forms and two others are Maestrichtian.

The age of the Umzamba beds of Pondoland has been regarded as

Campanian by Woods (60, p. 346), as Upper Santonian by van Hoepen

(50, p. 45), and as Campanian+ Maestrichtian by Spath (50, p. 116).

The Baba fauna can safely be regarded as of the same age as the

Umzamba and Umkwelane Hill faunas, which are probably mainly, if

not entirely, Campanian. Unfortunately there are no Ammonoidea

from Baba to give confirmatory evidence.

The Baba beds appear to be of the same age as Chottat’s Grés a

Cardita barroneti at Dombe Grande, and are perhaps of the same age

as the red calcareous sandstone of 8. Nicolau north of Mossamedes

(from which Gurich has recorded “ Trigonoarca, ressemblant un peu a

Tr. Trichinopolitensis Stol.” and “* Cyprina, voisins de C. forbesi Stol.,”’

but without giving descriptions or figures), which Choffat in 1905

(7, p. 21 and p. 30) suggested was of the same age as his Grés a

Rondaireia Forbest at Dombe Grande.

Apparently of uppermost Cretaceous age is a collection of poorly

preserved specimens, mostly Gastropoda, from Mesados Cavalleros

near Mossamedes. The Gastropoda have a Tertiary aspect, but the

Eriphyla appears to be closely allied to E. forbesiana from the Ariyalur

group and the Lzthodomus has been compared with a Senonian form.

That the Cretaceous does occur near Mossamedes seems to be indicated

by Choffat’s reference to a Cardwum and a Trigoma from “ fazenda

Vidal sur le Monte Cavaleiros ”’ near Mossamedes (7, p. 21), though

Bebiano (2) in his recent valuable map does not show the Cretaceous

much to the south of the 8. Nicolau river.

(6) Palaeo-geographical Considerations.—It has generally been

assumed that in Albian times a Brazilio-Ethiopian continent stretched

across the Atlantic and that direct communication between the

Mediterranean and the Cape did not take place until after the close

of the Cretaceous. The close relations of the Upper Albian fauna of

Angola to that of the Mediterranean have been noted by Spath (51,

pp. 154-158) and Bullen Newton (5, p. 562). Haug (27, p. 1357) has

postulated a temporary connection between the Mediterranean and

the Gulf of Guinea via the Sahara and the Sudan to explain the

affinities of the faunas of those regions, and Gregory (25) has embodied.

8 Annals of the South African Museum.

his views in a map which indicates the range of the Albian sea, which

according to him invaded the Brazilio-Ethiopian continent in the

form of a long arm stretching from the Tethys as far south as Angola.

While the idea of such a Brazilio-Ethiopian continent has received

wide support, it has been discarded in a recent paper by Spath (51a,

p. 196), who points out that the support given by ammonite evidence

to zoological provinces is valueless, as these provinces were based on

comparisons of beds of different ages, and that the Upper Albian

genus Elobiceras, formerly thought to be confined to the Angolan Gulf,

is now known from practically all round Africa. While it must be

admitted that the question of the Angolan Gulf has still to be

definitely settled, it is suggested that too much reliance should not be

placed on forms which are almost universal in their distribution, and

so the land barrier south of Angola during Albian times is provision-

ally accepted.

On the other hand it is certain that permanent communication had

been established between the Mediterranean and the southern seas via

the west coast of Africa in Kocene times, for the marine Tertiary from

the vicinity of Bogenfels on the south-west African coast, originally

described by Béhm and Weissermel (8) as Lower Miocene, is now

known to be of earlier date and related to the Eocene of West Africa

and Egypt (80, vol. 11, chap. xix, and 17, p. 349).

Hitherto the scanty and inconclusive remains of post-Albian

Cretaceous forms in Angola have led investigators to conclude that the

connection between South Africa and Brazil persisted through the

Senonian, in spite of the fact that certain differences between the

otherwise closely allied South African and Southern Indian Senonian

faunas can only be accounted for by direct communication of some

sort between the Natal coast and the seas of Southern Europe. Van

Hoepen (56, pp. 43-45) came to the conclusion that the fauna of

Pondoland showed the same affinity to the Cretaceous of Northern

Africa as to that of India and postulated a direct, uninterrupted coast-

line between Natal and Egypt, and Spath has noted the number of

“ Atlantic ” elements which occur in the Pondoland and Zululand

faunas. The evidence now at hand from Angola, though still lament-

ably scanty, certainly points to migration through Angola in Upper

Senonian times, and therefore is in support of the views put forward

some time ago by Kossmat (81, pp. 39-55) but not supported by

Woods (60, p. 348). It is concluded, therefore, that the great Cam-

panian transgression, which left deposits in Pondoland, Zululand,

Angola, and elsewhere, opened up a sea connecting the Angolan Gulf

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 9

with the southern ocean, and thus permitted intermingling of

“ Atlantic ” and “ Indo-Pacific’ elements in Angola and South-East

Africa. The Campanian probably marked the final severance of

the southern part of the Brazilio-Ethiopian continent. That the

“ Atlantic’ elements in South-Hast Africa could have, in part,

migrated down the east coast is of course still probable.

One is tempted to suggest that the communication may have been

established as early as the Coniacian, following on the Turonian

regression (Turonian deposits probably being absent from Angola),

but evidence for Coniacian forms on the west coast south of the

Cameroons is still lacking.

It is somewhat surprising that Campanian deposits have not been

found between Mossamedes and Pondoland. The occurrence reported

by Merensky (88, p. 18) of Protocardia hillana, Turritella (Zaria)

boner, and T. meadi, three Pondoland forms, on the south-west

African coast south of Bogenfels, would, if substantiated, have a

distinct bearing on the question, but Kaiser (80, vol. ii, p. 1) has been

unable to confirm the report.

DESCRIPTION OF THE SPECIES.

LAMELLIBRANCHIA.

Famity ARCIDAE.

Genus TRIGONOARCA, Conrad.

Trigonoarca angolensis, sp. nov.

(Plate IV, figs. 1-3.)

Material.—Holotype, 6469, an extremely well-preserved right valve.

Paratype, 6466. |

Description.Shell large, thick, trigonal and oblique, slightly

longer than high, well inflated. Anterior margin merging gradually

into the rounded, moderately convex, ventral margin; posterior

margin straight, forming an acute angle with the ventral margin.

Umbonal region much inflated, umbones incurved, with a very well-

marked rounded carina extending to the produced postero-ventral

angle, slightly concave to the postero-dorsal margin ; the part behind

the carina sloping very rapidly, almost vertical. Margins of valves

10 Annals of the South African Museum.

smooth, pallial line entire, with irregular muscle pits on the inner

side. Adductor impressions large, subequal, the posterior one on a

raised platform. Hinge long, arched, and with numerous teeth, the

central short, the lateral longer, with the upper portions bent sharply

outwards at or near the middle. Area large, triangular, with seven

deep ligament grooves separated by raised ridges of the same width

as the grooves ; grooves and ridges sharply bent beneath the umbo.

Outer surface of both specimens poorly preserved, but showing growth

lines of varying strength ; radial ornamentation not seen.

Length 75 mm., height along carina 70 mm., height measured

vertically from the umbo 63 mm., thickness of one valve 30 mm.

Remarks.—This is a very distinct species, most nearly related to

T. capensis (Griesbach) (60, p. 288, pl. xxxiv, figs. 1, 2) from Pondoland.

The latter has a narrower area with fewer, shallower, and more widely

spaced ligament grooves, a relatively longer hinge line, and is less

massive and oblique.

Locality.— Baba district, near Mossamedes.

Trigonoarca cf. diceras (Seguenza).

(Plate I, fig. 9.)

1882. Arca diceras. Seguenza, Atti Accad. Lincei, Roma, ser. 3,

vol. xii, p: 158% ple xaventioomln

1912. Arca (Trigonoarca ?) diceras. Pervinquiére, Et. Pal. Tunis,

Gastr. et Lam. Crét., p. 102, pl. vu, figs. 23, 25, 26 (with

syn.).

1915. Trigonoarca cf. diceras. R. Bullen Newton, Trans. Roy. Soc.

Kdin., vol. li, pt. iii, p. 570.

1918. Arca (Trigonoarca) diceras. Greco, Pal. Italica, vol. xxiv,

p. 29, pl. xix, figs. 14, 15 (with syn.).

Remarks.—A cast from Camballa of a small Trigonoarca resembles

this North African Cenomanian species by reason of its elongate,

almost quadrangular shape, very prominent incurved and distant

umbones placed very anteriorly, high sharp posterior carina extending

to the postero-ventral angle. A similar cast from the ammonite

marls at Lobito has been referred doubtfully to this species by Bullen

Newton.

Locality.—4 km. east of Camballa village, Cuvo River, No. 6828.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 11

Trigonoarca cf. trichinopolitensis (Forbes).

(Plate IV, figs. 8, 9.)

1846. Arca trichinopolitensis. Forbes, Trans. Geol. Soc. Lond.,

vol. vii, p. 150, pl. xv, fig. 16.

1871. Trigonoarca trichinopolitensis. Stoliczka, Cret. Pel. S. India,

p. 353, pl. xviii, figs. 12, 14; pl. xix, figs. 2,3 ; pl. xx, figs. 2,

8, 10.

Materval.—One specimen with valves closed, fairly well preserved ;

No. 6465. :

Description.—Shell moderately large, trapezoidal, convex, longer

than high. Anterior margin evenly rounded and merging into the

rounded ventral margin; posterior region produced, the postero-

dorsal margin almost straight, meeting the ventral margin in a sharp,

produced, postero-ventral angle. Umbones anterior, incurved and

distant, relatively small; a pronounced carina, slightly curved and

concave to the postero-dorsal margin, extending from the umbones to

the postero-ventral angle, dividing the shell into a moderately inflated

area and a narrow, steeply descending postero-dorsal region. Hinge

moderately long, area and hinge not clearly seen.

Ornamentation near the umbo consisting of numerous concentric

ridges crossed by numerous radial ridges, the latter being the stronger ;

the radial ridges interrupted towards the ventral part of the shell by

irregular laminar concentric growths; posterior to the carina the

concentric ornamentation appears to dominate.

Remarks.—Stoliczka has amply figured this species and noted the

variation in shape. The Baba specimen resembles very closely the

more elongated examples from the Trichinopoli group, but an actual

comparison of specimens would be necessary to decide whether it is

identical with Forbes’ species. The ornamentation is peculiar and

appears to be identical with that of the Indian form, in which the

radial ribs likewise tend to become obsolete posteriorly.

T. marest Coquand (11, p. 130, and 89, p. 103) from North Africa has

a considerable resemblance to this species, as has been pointed out by

Pervinquiéere and others ; in this Coniacian form, however, the radial

ornamentation is but poorly developed, whereas in our form the

radial ornamentation predominates on the earlier portions of the test.

Woods has drawn attention to the relationship between the Indian

species and 7’. capensis (Griesbach), but this species and 7. angolensis,

12 - Annals of the South African Museum.

described above, can be readily distinguished by their more quadrate

outline and more prominent umbonal region.

Locality.—Baba district, near Mossamedes.

Genus BarBatia, Gray.

Barbatia %, sp.

(Plate II, fig. 17.)

Remarks.—The collection contains a single left valve of a small

Arcid which may be referred tentatively to Barbatea. The shell is

sub-quadrate, longer than high, moderately convex, with rounded

margins, and is provided with a well-rounded posterior carina; hinge

line long but not the greatest length of the shell; umbo relatively

prominent. The surface is ornamented with numerous straight sharp

radial ribs which are widely spaced and separated by flattened inter-

spaces ; the ribs are closer together behind the carina.

The specimen has the general form of B. meridiana Woods (60, p. 287,

pl. xxxiu, fig. 14), but the hinge line is relatively longer and the orna-

mentation quite different. It is preserved on a limestone slab with

poor specimens of Chenopus reinecker sp. nov. and Astarte sp., frag-

ments of fish remains, and numerous foraminifera, among which Mr.

W. A. Macfadyen, M.A., has identified Bigenerina capreolus d’Orbigny.

Locality.— Hills west of well location, Quimbaixie, No. 6821.

Genus NEmopon, Conrad.

Nemodon natalensis (Baily).

(Plate IV, fig. 4.)

1855. Arca natalensis. Baily, Q.J.G.S., vol. xi, p. 461, pl. xiii,

fig. 2.

1904. Latiarca (?) natalensis. Etheridge, Sec. Rep. Geol. Surv. Nat.

and Zul., p. 77, pl. i, figs. 10-12.

1906. Nemodon natalensis. Woods, Cret. Fauna of Pondoland, Ann.

S. Afr. Mus., p. 289, pl. xxxiv, figs. 3—7.

Remarks.—This fine species is represented in the collection by two

well-preserved left valves which agree in every detail with the

description given by Woods and topotypes in the Sedgwick Museum.

The radial ribs anterior to the carina are more widely spaced and

irregular in size than in Woods, fig. 5a, but more regular than in the

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 13

variety fig. 4, and hence the character of the ornamentation lies

between the two. The type is from the Campanian of Pondoland.

Locality.—Baba district, north of Mossamedes, Nos. 6467, 6477.

Famity PERNIDAE.

Genus INocERAMmuS, J. Sowerby.

Inoceramus spp.

Remarks.—There are some poor examples of Inoceramus in the

collection. In this genus, as in the Ostreidae, the great variability

of the species makes the identification of specimens from distant

localities difficult, and when, as in the present case, only imperfect

material is available, the best that can be done is to indicate the

species-groups to which the forms belong. The evolution of the

English Cretaceous members of the genus has been worked out by

Woods (68).

There are two examples from “ foot of cliff 2 km. E. of Cabo Ledo ”’

(No. 6825), from which locality fragmentary Cephalopoda have been

obtained which, according to Spath (see Haughton, 28, p. 266), are

possibly Turonian. One is a small, moderately inflated form, rather

equilateral, whose periphery is not clearly shown, but which is higher

than long and ornamented with close concentric sulcations and a

suspicion of fine radial striae; it recalls forms belonging to the:

anglicus-crippsi-labiatus group, particularly some young forms of

I. crvppst Mantell. The other specimen from the same locality is

flattened and crushed, but may be the same species. The anglicus-

crippsi-labiatus group ranges from the Albian to the Turonian.

Two specimens from Cabo Ledo (Nos. 6926, 6927), whence Haugh-

ton has described Albian ammonites, are in a fragmentary condition

and probably flattened by pressure, but are not unlike the preceding.

Famity OSTREIDAE.

Genus OstTrEA, Linnaeus.

Ostrea vesicularis, Lamarck.

(Plate I, fig. 6.)

1806. Ostrea vesicularts. Lamarck, Ann. Mus. Hist. Nat., vol. viii,

p. 160, and vol. xiv (1809), p. 375, pl. xxi, fig. 3.

1888. Ostrea szajnochai. Choffat, Mat. Strat. Pal. d’Angola, p. 92,

pln yy; digs 1S.

14 Annals of the South African Museum.

1888. Ostrea vesiculosa. Choffat, ebid., p. 91, pl. v, figs. 15-17.

1912. Ostrea vesicularis. Woods, Cret. Lam. England, vol. u, p. 36,

pl. lv, text-figs. 143-182 (with full discussion and

synonymy).

Remarks.—A single specimen from Camballa belongs to this wide-

spread and variable Upper Cretaceous species, whose variations have

been studied by Woods. It is one of the types showing the grypheate

trend and left posterior sinus; left valve convex with the umbo

prominent and slightly incurved ; posterior part of the shell produced

and inflated, separated from the rest of the shell by a deep sinus ;

right valve deeply concave with a raised fold corresponding to the

sinus in the left valve; surface smooth. The sinus is deeper than in

any figure the writer has seen.

Ostrea szajnochar Choffat from Dombe Grande is a high grypheate

form with small posterior lobe and cannot be distinguished from some

of the specimens figured by Woods.

O. vesiculosa Sowerby has been figured by Choffat from the ammonite

marls at Catumbella, but his specimens are more likely the more

widespread O. vesicularts.

Locality.— Ridge east of Camballa village, Cuvo River, No. 6862.

Genus ALECTRYONIA, Fischer de Waldheim.

Alectryonia cf. syphax (Coquand).

(Plate I, fig. 3.)

1854. Ostrea scyphaz. Coquand, Descr. géol. Prov. Const., p. 143,

pl. iv, figs. 1-4.

1912. Alectryonia syphax. Pervinquiere, Et. Pal. Tunis., Gastr. et

Lam. Crét., p. 203, pl. xiv, figs. 15-18 (with syn.).

Remarks.—A single left valve from Camballa is probably A. syphax

Coquand or perhaps an allied species. Shell very thick, very slightly

inflated, sub-equilateral, with a posterior aliform expansion ; liga-

mental area very large; ornamented with about a dozen strong

radiating folds. It resembles the figures given by Pervinquiére from

the Cenomanian of Tunis. A. dichotoma Bayle is distinguished, accord-

ing to Pervinquiére, by having in general more numerous, finer, and

more freely bifurcating ribs and is a Senonian form.

Locality.—Ridge east of Camballa village, Cuvo River, No. 6799.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 15

Genus ExocyRra, Say.

Exogyra cf. conica (Sowerby).

(Plate I, fig. 18.)

1813. Chama conica. J. Sowerby, Min. Conch, vol. i, p. 69, pl. xxvi,

fig. 3.

1909. Exogyra conica. R. Bullen Newton, Trans. Roy. Soc. 8. Afr.,

vol. i, pt. 1, p. 51, pl. xi, figs. 8-10.

1912. Exogyra conica. Woods, Cret. Lam. England, vol. i, p. 407,

figs. 215-242.

Remarks.—Three exfoliated specimens are perhaps best compared

with EH. conica (Sowerby). The left valve is well angulated, and in

one specimen the umbo is very much incurved and partially free.

The right valve is moderately convex, with flattened border.

Localitves.— Foot of ridge E. of Camballa village, Cuvo River,

No. 6818.

E. of N. end of lagoon, Camballa village, No. 6801.

Exogyra olisuponensis, Sharpe.

(Plate III, fig. 8.)

1850. Exogyra olisiponensis. Sharpe, Q.J.G.S8., vol. vi, p. 185, pl. xix,

figs. 1, 2.

1862. Ostrea overwegr. Coquand (non de Buch), Géol. Pal. 8. Const.,

p- 226, pl. xix, figs. 1-6.

1905. Ostrea (Exogyra) olisiponensis. Choffat, Nouv. Donn. sur la

Zone Litt. d’Angola, p. 44, pl. 1, figs. 4, 5.

1911. Ostrea (Exogyra) olvstponensis. Woods, Pal. Upper Cret.

Nigeria, p. 277, pl. xx, figs. 1-3 (with syn.).

1912. Exogyra olusiponensis. Pervinquiére, Et. Pal. Tunis., Gastr. et

Lam. Crét., p. 174, pl. xiii, figs. 4, 5, 9 (with syn.).

1918. Exogyra olisiponensis. Greco, Pal. Italica, vol. xxiv, p. 5,

pl. xvu, figs. 12-14 (with syn.).

Remarks.—A single left valve from Quissama resembles the figure

of the holotype and later figures very closely, hence no further descrip-

tion is necessary. It is widespread in the Cenomanian of the Medi-

terranean region and occurs also in the Turonian of Portugal and

16 Annals of the South African Museum.

Nigeria. In Angola it has been figured from Dombe Grande in beds

of doubtful age. The type is Upper Turonian.

Locality.—East of Dondo-Quixinge Road, Quissama, No. 6798.

Famity TRIGONIIDAE.

Genus Triconia, Bruguieére.

Trigoma cf. ethra, Coquand.

1912. Trigonia ethra. Pervinquiére, Pal. Tunis., p. 218, pl. xv, figs.

4—7T (with syn.).

1915. Trigona crenulata. Bullen Newton, Trans. Roy. Soc. Edin.,

Vol. li, ptr iit yp. .om lepplan dient A:

1918. Trigonia ethra. Greco, Pal. Italica, vol. xxiv, p. 30, pl. xx,

figs. 1-3 (with syn.).

Remarks. Three poorly preserved specimens probably belong to

this widely spread North African Cenomanian form, rather than to

the allied 7. crenulata Lamarck originally described from France.

The relationships of the two species have been discussed by Greco

and by Pervinquiere ; 7. ethra differs chiefly in having fewer and con-

sequently more widely spaced costae, which are straighter than in

T. crenulata, and in possessing a less inflated shell. In our specimens

the costae are separated by interspaces which are wider than the

costae. The specimens agree very well with the figures given by the

authors quoted. Bullet Newton (5, p. 571, pl. 1, fig. 14) has

recorded as T. crenulata some casts from Catumbella and from Lobito

(where they occur below the ammonite beds) which have the more

widely spaced costae of 7. ethra.

T. subcrenulata d’Orbigny described by White (58, p. 70, pl. v, figs.

2, 3) from Brazil has a considerable resemblance to the Angola forms,

as Bullen Newton has already remarked.

Localities.—Benguella, No. 6420; Uchi district, Nos. 6444, 6453.

Famity PECTINIDARE.

Genus NerrHea, Drouet.

Neithea aequicostata (Lamarck).

1819. Pecten aequcostatus. Lamarck, Anim. sans vert., vol. vi, p. 181.

1912. Pecten (Neithea) aequicostatus. Pervinquiére, Pal. Tunis., Gastr.

et Lam. Crét., p. 135 (with syn.).

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 17

1915. Nethea aequicostata. Bullen Newton, Brach. and Mollusca from

Angola, Trans. Roy. Soc. Edin., vol. li, p. 565, pl. i, figs. 2, 3.

Remarks.—A small right valve from Uchi appears to be identical

with the specimen from the ammonite marls of Lobito described by

Bullen Newton as N. aequicostata. The shell is triangular and only

slightly ovate, narrowing rapidly towards the somewhat pointed umbo,

ornamented with about twenty ribs of more or less equal size; the

ears and transverse striations were not seen. The Angola specimens

differ from the typical N. aequicostata in having a more triangular

shape, narrower umbos, and fewer ribs, and in these respects resemble

N. sergipensis White (58, vol. vii, pl. i, figs. 6, 7, p. 39) from the same

horizon in Brazil, as Bullen Newton has pointed out.

Locality.—Uchi district, No. 6459.

Neithea quinquecostata (J. Sowerby).

(Plate I, fig. 13.)

1814. Pecten quinquecostatus. Sowerby, Min. Conch., vol. i, p. 122,

pl. lvi, figs. 4-8.

1903. Pecten (Neithea) quinquecostatus. Woods, Cret. Lam. England,

vol. 1, p. 202, pl. xxxix, figs. 14-17, and pl. xl, figs. 1-5

(with syn.).

1906. Pecten (Neitthea) quinquecostatus. Woods, Ann. 8. Afr. Mus.,

vol. iv, pl. xxxiv, fig. 14.

1909. Newthea quinquecostata. Bullen Newton, Trans. Roy. Soc. S.

Afr., vol. i, pt.1, p. 58, pl. 1, figs. 1, 2 (with syn.).

Remarks.—A damaged and worn right valve from Uchi certainly

belongs to this widespread species. There are from four to five

subsidiary ribs in each of the interspaces between the main ribs. The

species has not previously been recorded from Angola, but occurs

throughout the Upper Cretaceous of Hurope, North Africa, India, the

Campanian of Pondoland, and the Manuan Creek deposits of Zululand.

It occurs here in a white limestone with Albian forms.

Locality.—Uchi district, No. 6437.

Neithea tricostata (Coquand).

(Plate I, figs. 10-12.)

1862. Janira tricostata. Coquand, Géol. Pal. 8. Const., p. 219,

pl. xi, figs. 3, 4 [non Pecten tricostatus Bayle].

1888. Janira ficalhoi. Chofiat, Mat. Strat. Pal. d’Angola, p. 89,

pl. v, figs. 8-10.

VOL. XXVIII, PART 1. 2

18 Annals of the South African Museum.

1912. Pecten (Neithea) shawi. Pervinquiere, Et. Pal. Tunis., Gastr.

et Lam. Crét., p. 136, pl. ix, figs. 1-6 (with syn.).

1915. Nevthea tricostata. R. Bullen Newton, Trans. Roy. Soc. Edin.,

vol. li, p. 567, pl. 1, figs. 5-7.

1915. Nethea angolensis. R. Bullen Newton, chid., p. 566, pl. 1,

fig. 4.

Materval— Numerous large and small specimens. Also two speci-

mens of the variety with reduced intermediary costae.

Description.Shell triangular, nearly equilateral, considerably

higher than long, ventral margin polygonal, ears small. Right valve

very convex with the umbo greatly incurved; ornamented with

twenty-one ribs, six of which are much stronger than the rest and pro-

ject at the ventral margin ; between each of the main ribs are three

smaller ribs of more or less unequal size, separated by deep concave

interspaces, the central intermediary rib larger than the other two ;

surface with fine concentric striations. Antero- and postero-dorsal

areas sloping downwards or slightly inwards, with a few nearly

obsolete ribs and concentric striations.

Left valve slightly concave with more or less equal ribs, six of the

furrows being a little deeper than the rest.

Remarks.—R. Bullen Newton has described a beautiful and extra-

ordinarily large example of this species from Lobito, but his figure

hardly does justice to the specimen. The species, which is widespread

in the Cenomanian of North Africa, is often confused with N. quadri-

costata (Sowerby), as Newton has observed, but differs in several

important points; the height is relatively greater and in the right

valve the umbo is more strongly incurved ; the ears are much smaller ;

the antero- and postero-dorsal areas slope downwards rather than

outwards; the intermediary ribs are usually markedly unequal.

There are several large examples in the collection, none of which are as

well preserved as that figured by Newton, as well as some small forms

which are more like the examples figured by Pervinquieére.

Variety with Costae Reduced.—N. tricostata is interesting on account

of the frequently exhibited tendency of the intermediary costae to

become reduced. Normally there are three well-developed costae

between each of the six major costae, and of these the middle one is

greater than the other two. Such normal forms are figured, e.g. by

Pervinquiere (figs. 1-3). The lateral intermediary costae in each

group tend to become rudimentary. Pervinquiére figures two such

varieties (figs. 4-6): the variety atropha Peron, in which the lateral

intermediary costae nearest the middle line of the valve become

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 19

reduced, and the variety biatropha Pervinquiere, in which the re-

duction is carried furthest in the lateral intermediary costae most

remote from the middle line. In each case the tendency is to leave

two rather unequal intermediary costae between each of the six major

costae. The degree and manner of the reduction is extremely variable

—Pervinquiere himself figures an intermediary variety—and in the

present state of our knowledge it seems to the writer to be useless to

propose varietal names.

There are in the collection two small right valves from Uchi, in

excellent state of preservation, which show a considerable reduction

of the lateral intermediary costae, and in that respect resemble the

variety atropha. Inshape they are like the typical forms, with narrow,

triangular valves, well inflated and strongly incurved at the umbones,

smooth and very steep areas, and small ears; they show the concentric

striation characteristic of N. tricostata but are perhaps narrower and

more pointed at the umbones than most examples of the species.

Janira ficalhotc Chofiat from the Pholadomya pleuwromyaeformis

horizon at Dombe Grande is apparently an extreme case in which one

of the three intermediary costae in each group has become reduced

to a ridge on the flank of the adjacent major costa. The holotype of

N. angolensis R. B. Newton is a worthless specimen from below the

ammonite marls at Catumbella, in which the reduced costae are

apparently absent owing to the poor state of preservation ; in form it

agrees almost exactly with the Uchi specimens.

Localitues.—Ridge EH. of Camballa village, Cuvo River, Nos. 6819,

6851; Dombe Grande, No. 6404; Benguella, Nos. 6423, 6428 ; Uchi

district, Nos. 6454, 6455.

Variety with reduced costae : Uchi district, Nos. 6436, 6464.

Nerthea welwitschi (Choffat).

1888. Janira welwitscht. Choftat, Mat. Strat. Pal. d’Angola, p. 90,

ple veer.

Remarks.—This species was founded on a single large shell from

Dombe Grande, characterised by about a dozen large costae with one

rather smaller costa between each, and, judging from the figure, of

moderate convexity. The species appears to be more closely allied

to N. quadricostata than to any other species, but the point cannot be

decided in the absence of better material. Some shells from Cam-

balla are doubtfully referred to the species.

Locality. —Ridge E. of lagoon, Camballa village, Nos. 6864-6866.

20 Annals of the South African Museum.

Famity SPONDYLIDAE.

Genus PricatuLa, Lamarck.

Plicatula, sp.

(Plate V, figs. 1-3.)

Remarks.—Five specimens of Plicatula, all with the valves closed,

are present in the collection. Owing to the great variability of the

species and the extraordinary difficulty in defining them, the writer

has been unable to allot definite names to the present specimens.

Unless a very large number of specimens were examined it would be

extremely hazardous to attempt to assign names to the species or

even adequately to describe them, and it would merely add to the

confusion already existing. A large number of species have been

described from North Africa by Coquand, Peron, Pervinquiére, and

others, and the more recent studies of Pervinquiére, who has

described and given excellent figures of numerous examples, show

clearly that the delimitation of the numerous species and varieties is

by no means an easy matter. Under the circumstances the writer

can do no better than record some of the characters exhibited by the

present specimens and indicate some of the previously published

figures which they resemble.

The five Angola specimens diverge considerably from one another,

though they have certain features in common which lead one to

regard them as one and the same species. They are all stout shells,

higher than long, with subequal valves, tapering towards the umbones,

the longest part of the valves nearer the ventral margin than the

umbones; ornamented with numerous very irregular subequal

radiating ribs and prominent irregularly spaced growth laminae.

The chief variation is in the degree of inflation, which is from slightly

convex to moderately convex, and in the number of radiating ribs,

which vary from about twenty to over thirty, and appear to increase

in the later parts of the valves by dichotomy.

These forms bear a considerable resemblance to the Tunisian species

described and figured by Pervinquiére as P. ventilabrum Coquand var.

suffetulensis Pervinquiére (89, p. 154, pl. x, figs. 9, 10) and to the more

finely costate variety of P. ventilabrum (89, pl. x, fig. 11) which are

Senonian forms. These varieties are stout inflated forms which taper

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 21

towards the umbones and bear from twenty to thirty irregular ribs,

interrupted by irregular prominent growth laminae. The var. suffetu-

lensis, however, has fine radiating striae between the costae, which

are not seen in the Angola forms (though they may have originally

been present).

P. ferryt Coquand var. desjardinst Coquand (89, pl. ix, fig. 22), also

Senonian, resembles both the above varieties and our specimens

closely ; it has no finer striae between the costae, a character which is

one of the main differences between P. ventilabrum and P. ferryt,

according to Pervinquiére indicating evolution from different stocks.

Locality.—Massangano area, south of the river, Nos. 6752, 6754.

Genus SPoNDYLUS, Linnaeus.

Spondylus angolensis, sp. nov.

(Plate II, figs. 9-11.)

Description.—Left valve smaller than the right, well inflated, in-

equilateral, quadrangular in outline. Umbonal region prominent,

pointed. Posterior side sloping gently, anterior side with steep slope

from a well-rounded ridge which extends from the umbo to the antero-

ventral angle. Posterior margin moderately convex, merging

gradually into the ventral margin ; anterior margin straight, meeting

the ventral margin at about a right angle. Margin dentate on the

inner side. Ornamented with about thirty stout radiating ribs, sharp

at the summit ; ribs provided with pointed spines at regular intervals,

the spines arranged in concentric rows; ribs separated by deep

grooves of about the same width as the ribs. Right valve greatly

produced, almost cylindrical.

Remarks.—The diagnosis was formed partly from an internal cast

of a specimen in which the valves were united, partly from a mould of

part of the left valve of the same specimen. In the form of the right

valve this species recalls S. coguandianus d’Orbigny (14, vol. ii, p. 663,

pl. cccelii, figs. 9, 10), a Turonian species, but the quadrangular outline

and relatively few ribs easily separate it and point to its being a rather

distinct species.

Locality.—Dombe Grande, No. 6369.

22 Annals of the South African Museum.

Famity LIMIDAK.

Genus Lima, Bruguiére.

Sub-genus Mantellum, Bolton.

Inma (Mantellum), sp.

(Plate tye ies ids)

Remarks.—Among the Mossamedes specimens is the left valve of a

small Mantellum comparable with L. (Mantellum) elongata (Sowerby)

(61, vol. u, p. 34, pl. vi, figs. 5-7) from the Chalk. Shell moderately

convex, oblong, oblique ; antero-dorsal margin long, almost straight,

ventral margin rounded. Ornamentation of about fifteen very strong,

elevated, laterally compressed, sharp-edged ribs, separated by deep

V-shaped furrows.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6485.

Famity MYTILIDAR.

Genus LirHopomus, Cuvier.

Inthodomus, sp.

(Plate III, figs. 5-7.)

Remarks.—Shell elongate, cylindrical, very convex; anterior

truncated ; ventral margin straight ; posterior slightly compressed.

Umbones moderately inflated, highly incurved and almost touching,

terminal. Surface ornamented with concentric growth lines. The

above description applies to a small Lithodomus which may be com-

pared with L. aequalis d’Orbigny (14, vol. iu, p. 295, pl. ccexlvi, figs.

4-6) from the Turonian of France, but our form is more perfectly

cylindrical, the sides being parallel when viewed from above.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6492.

Famity PLEUROMYIDAE.

Genus PLEuRoMyYA, Agassiz.

Pleuromya ?, sp.

(Plate V, fig. 11.)

Remarks.—A single right valve from Capolo is referred doubtfully

to Pleuromya. Shell elongate oval, moderately inflated, inequilateral ,

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 23

thin; umbo prominent, incurved and anteriorly placed; margins

rounded ; anterior part short, posterior long and very slightly com-

pressed. Posterior not gaping, or only a little. The surface is orna-

mented with growth lines and is without punctae.

From the same locality Haughton has recorded Phylloceras surya

(Forbes), which in India occurs in the Valudayur group, and Baculites

sp.

Locality.—15 km. 8., 75 km. EH. of Capolo, No. 6786.

Famity PHOLADOMYIDAE.

Genus PHOLADOMYA, Sowerby.

Pholadomya aff. tegris, Noetling.

(Plate V, figs. 7, 8.)

1902. Pholadomya tigris. Noetling, Fauna Baluchistan, p. 52, pl. xiii,

fig. 6.

Remarks.—A damaged and somewhat worn cast of a specimen with

the valves closed agrees very well with the figures and description of

P. tagris from the Maestrichtian of the Mari Hills. The shell is thin,

well inflated; umbones anteriorly placed, strongly incurved, the

umbonal region well inflated ; ornamentation of growth lines and about

eighteen radiating folds, which are absent from the anterior and

posterior ends of the valves. Noetling states that though the anterior

of his specimen is damaged, it appears to have been short and cut off

straight ; from his figure the shell appears to be truncate anteriorly,

but may have been rounded as in our shell, which agrees in other

particulars. Compare also P. esmarki (Nilsson) (22, vol. 1, p. 272,

pl. clvu, fig. 10, and 34, p. 101, pl. xxxiv, fig. 5) which appears to be

higher and more oblique.

Locality.—Massangano Fort, No. 6767.

Famity CYPRINIDAE.

Genus Cyprina, Lamarck.

Cyprina, sp.

(Text-figs. 1, 2.)

Description.—Shell very large, highly inflated, ponderous, very

inequilateral and oblique. Umbonal regions large, inflated, very

os —s

SSS ~

4 Sees SSS

tin te. =r

oN =

Figs. 1 and 2.—Cyprina, sp. Two views of cast from the Quissama area, No. 6824.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 25

anterior ; umbones relatively small, highly incurved but not touching.

Anterior margin rounded, not produced ; posterior of the shell pro-

duced, postero-dorsal margin convex, postero-ventral extremity

sharply rounded; ventral margin gently convex. Lunular region

large, broad, and depressed. Muscular impressions relatively shallow,

the anterior large and elongated.

Length 10 cm., thickness 8-5 cm., height 8-5 cm.

Remarks.—The description is of a remarkably large internal cast

which seems to be a true Cyprina. The highly inflated and broad

umbonal region, combined with the obliquity and ponderous size,

make this a very distinct species. It occurs with the Cenomanian-

Turonian Ez. olisiponensis. —

Locality.— Kast of the Dondo-Quixinge Road, Quissama, No. 6824.

Genus VENILICARDIA, Stoliczka.

Venilecardra cf. nicarser (Coquand).

(Plate V, figs. 9, 10.)

1862. Cyprina micarser. Coquand, Geol. Pal. Sud. Const., p. 201,

pl. ix, figs. 8-10.

1912. Cyprina (Venilicardia) nicaiser. Pervinquiére, Pal. Tunis.,

Gastr. et Lam. Crét., p. 227, pl. xvi, figs. 3-5.

Materval.—A poorly preserved right valve from Hombo, showing

only roughly the general shape and coarse ornamentation; a cast

of a right valve and a cast of a specimen with the valves united, with

portions of the test adhering, both from Massangano.

Description.—Shell triangular, longer than high, very inequilateral,

moderately inflated ; umbones prominent and placed well forward.

Postero-dorsal margin long, convex ; ventral margin slightly convex ;

the posterior part somewhat truncate ; antero-dorsal margin short,

concave, anterior rounded. Prominent carina extending from the

umbones to the postero-ventral extremity, separating a very steep

postero-dorsal area from the rest of the shell. Ornamentation of

strong concentric ribs, which become finer behind the carina. Muscle

impressions very deep, the anterior deeper.

Remarks.—The specimens bear a certain resemblance to the North

African form in size and proportions, position of the umbo, and

strength of the carina, but the ornamentation appears to be coarser.

C. lageriensis d’Orbigny (14, vol. ii, p. 103, pl. cclxxv) from the

26 Annals of the South African Museum.

Turonian has the umbo placed more anteriorly and the carina less

strong. C. securiformis Sharpe (46, p. 182, pl. xxu, figs. 1-3) is of the

same type but with finer ornamentation.

Localitues—Massangano area, south of the river, No. 6758. Near

Hombo, Quissama, No. 6759.

Genus VENIELLA, Stoliczka.

The group of shells related to Morton’s Venzlia conrad: has given

rise to much confusion, chiefly owing to the misinterpretation by

Stoliczka of Morton’s genus Venilia. The genus Venilia was created

by Morton (85, p. 294, pl. vin, figs. 1, 2, and 35a, p. 67, pl. viu, figs.

1, 2) in 1833 for a shell from New Jersey, V. conradi Morton, which

was characterised by its markedly trigonal shape and incurved

umbones. The original description of the genus is meagre—* an

equivalve bivalve; the hinge with three robust cardinal teeth in

each valve, and an elongated thick lateral on the posterior side,

similar to that of a Unio; anterior muscular impression profound ”—

but the figure given of the right valve shows the relative proportions

and positions of the teeth sufficiently clearly, though that of the left

is too poor to be of much use.

Stoliczka (54, p. 189) in 1871 replaced the name Venzlia (preoccupied,

having been used by Duponchel in 1829 for a group of Lepidoptera

and since used also by Alder and Hancock in 1844) by Veniella, but

at the same time he misunderstood Morton’s genus and wrongly

characterised Veniella as possessing two cardinal teeth and one

posterior lateral in each valve. Stoliczka (54, pp. 197-199, pl. ix and

pl. x, fig. 1) described three species from India with the same remark-

able shape as V. conradi. Two of these he referred to Cyprina

Lamarck, viz. C. forbesiana Stol. and C. cristata Stol., but Meek (82,

p. 150) pointed out some of the characters which separate the former

from Cyprina, and placed both species (the hinge of the latter not

having been seen) in Venzella, quoting Conrad in support of his action.

Stoliczka himself is inconsistent in calling C. forbestana a Cyprina,

for he states that a typical Cyprina is uniformly convex and that the

right posterior cardinal is thick and deeply bifurcate, whereas his

description and figures show clearly that C. forbesiana is a markedly

trigonal, highly carinated shell, with the right posterior cardinal only

feebly sulcated above. The third Indian species was placed in a new

sub-genus Cicatrea, including only C. cordialis Stol., a shell which

resembles the other forms closely in shape, in the arrangement of the

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 27

teeth, and is distinguished, according to Stoliczka, by possessing a

short but deeply bifurcate ligament groove; he placed Crcatrea pro-

visionally under Cyprina.

Further confusion was caused in 1881 when Munier-Chalmas (16,

p. 74) created the genus Roudairera, with a Tunisian species R. drua

as type; he associated Cyprina forbesiana and C. cristata with his type,

but retained Crcatrea as he did not find the bifurcate ligament groove

in R. druc.

Douvillé (15, p. 216) in 1904, in recording Cicatrea cordialis from

Madagascar, was led to regard Crcatrea and Roudaireia as synonyms

and to reject the former as having been badly defined by Stoliczka,

a view which received the support of Pervinquiére (89, pp. 228-229)

in 1912, who, however, restudied the types of R. drui and came to

the conclusion that that species does show the deeply bifurcate liga-

ment groove, and hence he accepted the three Indian forms under

Roudaireca. At the same time Pervinquiére remarks on the lack of

accord between Morton’s original figures, Meek’s figures of the holo-

type of 1876, and Whitfield’s figures of 1885 (59, p. 144, pl. xix, figs.

8-10), and implies that, as none of them quite agree with Roudaireva,

either Venzella is badly defined and must give place to Roudairera

or is a different genus. Trechmann (55, pp. 58-59) has recently

described a Roudairera (R. jamaicensis) and a Veniella from Jamaica

and evidently regards them as distinct.

On the other hand Bullen Newton (4, p. 69) in 1909 regarded Cyprina

forbesiana and C. cristata as belonging to Venvella, with which he was

inclined to make both Crcatrea and Roudaireva synonymous. Wade

(57, p. 77, pl. xxiv, figs. 14-16) has recently given a lengthy descrip-

tion of V. conradz with excellent figures, and it is interesting to note

that he includes R. druc under Veniella.

There can be no doubt that the various species under discussion

belong to one and the same genus, which is characterised by its

markedly trigonal shape and high carina and by its non-bifid right

posterior cardinal, the latter character at once distinguishing it from

both Cyprina Lamarck and Venilicardia Stoliczka. It is not quite

clear what Stoliczka had in mind when he stated that Crcatrea possessed

a short but deeply bifurcate ligament groove, and his figure of a left

valve leaves one in some doubt as to its interpretation ; the writer

has seen an example of V. forbesiana in which the nymph has broken

away from the shell, leaving two parallel grooves, and it is likely that

the diagnosis of Cicatrea is based on a similarly imperfect specimen.

The delicate lamina connecting the top of the left anterior cardinal

28 i Annals of the South African Museum.

with the base of the left median cardinal, which lamina corresponds

to the groove between the right anterior and median cardinals, appears

to have been observed only in V. forbescana. That Veniella should

have precedence over Roudaireia is established ; the figure of the right

valve given by Morton shows clearly enough not only the peculiar

shape and ornamentation of the genus, but also the nature and dis-

position of the teeth, and in particular the non-bifid right posterior

cardinal.

Veniella druc (Munier-Chalmas).

(Plate III, figs. 2, 3.)

1881. Roudatreca druz. Munier-Chalmas, Mission des Chotts Tun.,

p. 16, ploiv; figs. Iisa aie sla

1902. Roudairera drurz. Quaas, Overwegischichten, p. 221, pl. xxiv,

figs. 20-22.

1912. Roudarrera drurz. Pervinquiere, Pal. Tun., Gastr. et Lam. Crét.,

p. 230, pl. xv, figs. 9-13.

1917. Roudawreca auressensis. Fortau, Geol. Surv. Egypt, Pal. Ser.,

_ No. 3, p. 63 (with syn.).

1926. Veniella drurz. Wade, U.S.G.S. Prof. Paper 137, Fauna Ripley

Form. Tenn., p. 77.

Remarks.—The collection includes two damaged specimens which

.can be referred without doubt to this characteristic North African

species. Oneis aright valve whose hinge has been damaged and whose

postero-ventral extension has been removed, the other a left valve

still retaining the postero-ventral portion. They resemble the figures

given by Munier-Chalmas and Pervinquiére very closely, and it is not

necessary to add anything to the description beyond mentioning that

they show very clearly the characteristic ornamentation of broad and

irregular sulcations which become less pronounced with the growth

of the shell, with growth lines behind the carina, and that the carina

is more compressed laterally than in thetype. There is a very obscure

keel running from the umbo between the carina and the dorsal margin

as in the holotype, whereas in Coquand’s (8, p. 207, pl. xii, figs. 10, 11)

Trigonia auressensis, which has been regarded as synonymous by

Peron (38, p. 299) and by Fortau, the keel is very marked ; they can

hardly be the same species.

There are three undescribed right valves in the collection of the

Transvaal Museum from Pondoland which are not distinguishable

specifically from this species, which has hitherto not been described

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 29

from Southern Africa; it is characteristic of the Campanian and

Maestrichtian of North Africa. ©

Locality.—Baba district, north of Mossamedes, Nos. 6470 and 6468.

Famity ASTARTIDAE.

Genus ASTARTE, Sowerby.

Astarte, sp.

(Plate II, fig. 3.)

Remarks.—Three specimens from Quimbaixie are referable to

Astarte but are too poor for further determination. Shells rounded,

somewhat compressed, umbones sub-central and moderate in size,

posterior truncate ; surface with distant concentric ribs with sharp

raised summits.

Locality.—Hills west of well location, Quimbaixie, Nos. 6821, 6822.

Astarte, sp.

(Plate IT, fig. 5.)

Remarks.—A specimen from Chipupo contains numerous moulds

of an Astarte. The following is a description of a plasticine impression.

Shell round, longer than high, moderately inflated ; umbo prominent,

anteriorly directed ; lunule large, cordate, distinctly limited by an

impressed line; surface ornamented with distant, sharp, raised, con-

centric ribs separated by flattened interspaces bearing fine concentric

striations.

Locahity.—Chipupo district, No. 6510.

Genus EripHyua, Gabb.

Eriphyla cf. forbesiana, Stoliczka.

(Plate III, fig. 4.)

1871. E. forbesiana. Stoliczka, Cret. Pal. S. India, p. 181, pl. vi,

figs. 14-16. ;

Description.—A single left valve in good state of preservation but

with the hinge and lunule somewhat obscure. Shell orbicular, very

flat and compressed, the earlier part more inflated ; umbo pointed and

30 Annals of the South African Museum.

anteriorly curved ; lunule must have been very narrow and probably

deep ; escutcheon absent ; ornamentation of numerous sharp, thin,

concentric ribs separated by broad interspaces which are very minutely

striated.

Remarks.—The ornamentation, flat compressed shell, narrow lunule,

and lack of escutcheon agree very well with Stoliczka’s description

of EL. forbesiana from the Arialyur group, but unfortunately his figures

hardly indicate the difference between that species and E. lenticularis

Goldfuss figured from the same horizon.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6488.

Famity CRASSATELLITIDAE.

Genus CRASSATELLITES, Kriiger.

Crassatellites, sp.

(Plate II, fig. 12.)

Description.—Shell triangular, moderately inflated, compressed

ventrally and posteriorly, with a sharp posterior carina, cutting off

a narrow depressed area from the rest of the shell. Anterior well

rounded, the margin passing gradually into the curved ventral

margin. Postero-dorsal margin long, straight, posterior slightly

truncate. Umbo inflated; escutcheon long, narrow; lunule indis-

tinct. Surface with concentric striations.

Remarks.—The shell is too poorly preserved to make exact com-

parison possible, but is of the type of C. regularts d’Orbigny and may

be compared with C. pusilla Coquand (8, p. 198, pl. xi, figs. 12, 13)

from the Cenomanian of Algeria.

Locality—Dombe Grande, No. 6373.

Crassatellites, sp.

(Plate II, fig. 15.)

Remarks.—Similar to the above and preserved in the same type of

matrix, a highly micaceous sandstone, is a left valve apparently

belonging to Crassatellites. It is more elongated and posteriorly com-

pressed and less inflated than the preceding, and the somewhat sharp

carina separates a fairly wide posterior area. Surface with growth

lines.

Locality.—Dombe Grande, No. 6372.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 31

Famity CARDITIDAE.

Genus Carpita, Bruguiere.

Cardita barronett, Munier-Chalmas.

(Plate III, fig. 9.)

1881. C. baronnetti. Munier-Chalmas, Mission des Chotts Tun., p. 70,

pl. xi, figs. 4-8.

1905. C. baronnetr. Choffat, Nouv. Donn. sur la Zone Litt. d’ Angola,

p. 30, pl. 1, fig. 2.

1912. C. barroneti. Pervinquiére, Et. Pal. Tunis., Gastr. et Lam. Crét.,

p. 241, pl. xvin, figs. 1-3 (with syn.).

Remarks.—A single right valve agrees extremely well with the

cotypes figured by Pervinquiére. The shell is circular in outline, with

the prominent umbo curved forward, and the surface is ornamented

with about eighteen stout radial ribs, without tubercles, separated by

narrower grooves; ventral margin slightly crenulate on the inside ;

part of the hinge is visible, showing a stout median cardinal, a long

narrow nymph and ligament groove, and probably a small anterior

cardinal.

From the Maestrichtian of North Africa and recorded also by Chofiat

from Dombe Grande.

Locality.— Baba district, north of Mossamedes, No. 6482.

Famity LUCINIDAE.

Genus Luctna Bruguiere.

Lucina angolensis, sp. nov.

(Plate I, figs. 1, 2.)

Description.—One specimen with the valves closed, tolerably well

preserved. Shell sub-orbicular, longer than high, moderately com-

pressed, equivalve ; hinge line long, gently arched; postero-dorsal

and antero-dorsal margins straight; ventral margin well rounded ;

the posterior probably slightly truncate but only partly present in

the specimen ; the posterior and anterior margins meeting the postero-

dorsal and antero-dorsal margins respectively in rounded angles.

Umbones moderately prominent, almost centrally placed, directed

32 Annals of the South African Museum.

slightly towards the anterior. Lunule small, deep, elongate ; escutch-

eon narrow, very elongate, with. margins raised and sharp. Orna-

mentation of regularly spaced, narrow, sharp, concentric lamellae,

separating very broad flat interspaces with fine concentric striations,

the concentric markings bending rather sharply upwards before reach-

ing the antero- and postero-dorsal margins.

Remarks.—This fine large Lucina resembles L. fallax Forbes from

the Utatur group of India. The examples regarded as typical by

Stoliczka (54, p. 256, pl. xiv, figs. 3-5) differ from our form in several

respects ; they are shorter, more nearly circular shells with the umbo

more anteriorly placed, and the lunule has a more anterior aspect,

whereas our form tends to be quadrate and is a much larger shell ;

the ornamentation appears to be similar.

L. saharica Quaas (41, p. 214, pl. xxiv, figs. 5-7) from North Africa

is more nearly circular in outline and the main lamellae are closer

together. L. subnumismalis d’Orbigny (see Ravn, 42, p. 129, pl. iv,

fig. 211) is similar to L. saharica.

Our shell is most nearly approached by the form described and

figured by Woods (62, p. 279, pl. xx, figs. 8, 9) as Lucona sp. from the

Turonian of Nigeria ; the latter is more quadrate than L. saharica and

in outline and ornamentation resembles our form closely, but differs

in being a more inflated and much smaller shell.

There are a number of internal casts from the Chipupo district

which are probably Lucina, some of which are perhaps this species,

as is also a cast from Dombe Grande.

Localities—Dombe Grande; holotype, Nos. 6334, 6337 ? Chipupo

district, Nos. 6501 ?, 6502 ?, 6518 2, 6519 ?

Lucina reinecker, sp. nov.

(Plate II, figs. 18, 19.)

Description.—Shell sub-orbicular, a little longer than high, moder-

ately compressed. Hinge line long, arched. Postero-dorsal margin

straight, antero-dorsal gently convex, the two diverging from the

umbo at about 120° ; anterior, ventral, and posterior margins forming

a more or less continuous curve, but the posterior rather straighter

than the anterior and giving rise to a slight postero-ventral angle.

Umbo small, moderately prominent, anteriorly directed. Lunule

and escutcheon obscure or absent. Surface with very numerous sharp

concentric ribs of varying strength. Hinge strong, not clearly seen.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 33

Remarks.—This species is easily distinguished from the preceding

by the irregular nature of the concentric ornamentation. Among

Huropean species which are perhaps comparable are L. sanctae-

crucis Pictet et Campiche (40, sér. iv, p. 289, pl. cxxii, fig. 8) from the

Albian, but the latter has a well-defined though narrow escutcheon ;

Lucina sp., figured by Woods (61, vol. ii, p. 152, fig. 3) from the Spilsby

Sandstone, has a more completely rounded outline and apparently

no trace of a postero-ventral angle. The species is founded on a fairly

well-preserved right valve.

Locality.—4 km. Mag. N. of Chio, No. 6796.

Famity CARDIIDAE.

Genus Carpium, Linnaeus.

Sub-genus Trachycardvum, Méorch.

Cardium (Trachycardium) reynoldsi, sp. nov.

(Plate Ill, figs. 10-15.)

Description.—The holotype is a well-preserved right valve and is

accompanied by three other right valves and one left valve. Shell

stout, thick, much inflated, higher than long, inequilateral ; umbones

prominent, curving strongly inward ; anterior margin gently rounded,

merging gradually into the ventral margin; posterior margin straight ;

posterior portion of valve sloping rapidly, the postero-ventral part

produced ; hinge strongly curved with prominent cardinal and lateral

teeth. Adductor impressions not marked. Ornamentation of about

twenty-six very strong, much elevated ribs, which are, on the central

parts of the valve, very high, with spiny summits, and separated by

deep narrow furrows; ribs on the posterior slope lower and more

rounded, the change from the large ribs on the central part of the valve

to the smaller ribs on the posterior being abrupt, giving the shell an

apparent carina; ribs on the posterior part ornamented with large

conical tubercles, the outermost ribs having the largest tubercles ;

ribs on the central part with short spines like the teeth of a saw;

on the anterior part the ribs are lower and with conical tubercles ;

irregular growth lines present. Margins of the valves markedly

dentate, especially on the posterior margin.

One of the specimens, apparently an old shell, has a greatly increased

height without much increase in length, giving the shell a very inflated

and laterally compressed shape.

VOL. XXVIII, PART 1. 3

34 Annals of the South African Museum.

Remarks.—This is a very distinct form belonging to the sub-genus

Trachycardium and differs from forms like C. productum Sowerby and

C. pustulosum Munster (22, pl. cxliv, fig. 6), which it resembles in shape

and size, by having a lesser number of stouter ribs.

Locality.—Baba district, north of Mossamedes: holotype, No.

6480; paratypes, Nos. 6471, 6473, 6474, 6481.

Cardium (Trachycardium *), sp.

(Plate IT, fig. 4.)

Remarks.—The collection includes an incomplete left valve with

damaged exterior surface, best included under Trachycardium on

account of its form. The shell is triangular and very convex and is

ornamented with very numerous ribs and furrows. A rather similar

lot of shells has been recorded by Bullen Newton as Trachycardium cf.

syrvacum (Conrad) (5, p. 573). Compare with C. subproductum Thomas

et Peron, regarded as synonymous with C. productum Sowerby by

Pervinquiére (39, p. 259, pl. xix, figs. 25-27), both Turonian forms.

Locality.—Catumbella district, No. 6415.

Sub-genus Acanthocardium, Gray.

Cardium (Acanthocardium) denticulatum, Baily.

(Plate V, fig. 4.)

1855. Cardium denticulatum. Baily, Q.J.G.S., vol. xi, p. 460,

pl. xi, fig. 4.

Non. 1871. Cardiuwm denticulatum. Griesbach, Q.J.G.S., vol. xxvii,

p: 67, pleut:

1906. Cardium denticulatum. Woods, Ann. S. Afr. Mus., vol. iv,

p. 306, pl. xxxvi, fig. 2, and pl. xxxvu, figs. 1, 2.

Description —An incomplete right valve. Shell stout, oval,

inflated ; umbones median, prominent, and curving strongly inward ;

posterior side of valve sloping rather more rapidly than the anterior ;

margins absent in the specimen and hence the characteristic tooth-like

projections not seen. Ornamentation of about twenty-six strong,

very elevated ribs with spinose summits, separated by deep furrows

which are broader than the ribs; ribs closer together and less spiny

on the posterior part.

Remarks.—This specimen cannot be distinguished from the well-

known form from the Campanian of Pondoland ; it agrees well with

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 35

the description and figures given by Woods, and differs from a specimen

in the Sedgwick Museum only in its slightly smaller size.

Locality.—Massangano Fort, No. 6766.

Sub-genus Laevicardvum, Swainson.

Cardium (Laevicardium *), sp.

(Plate II, figs. 1, 2.)

Description.—Shell trigonal, moderately oblique, inflated, higher

than long. Umbones prominent, pointed, touching, directed some-

what anteriorly, with a rounded ridge extending to the postero-ventral

angle on each valve; behind the ridge is a narrow, steeply sloping

area. Margins gently convex, with rounded postero- and antero-

ventral angles. Surface smooth, with very faint traces of growth

markings.

~ Remarks.—From its shape and lack of ornament this shell would

appear to be a Laevicardium. Coquand (8, pls. x, xi) has figured and

given new specific names to several casts from Algeria; of these

C. paula (ibid., p. 204, pl. x, figs. 5, 6), though a larger shell, appears

to have the same form as our specimen, but the concentric orna-

mentation is more in evidence.

Localityx—Dombe Grande, No. 6330.

Genus ProtocarpiA, Beyrich.

Protocardia hallana (J. Sowerby).

(Plate III, fig. 1.)

1813. Cardium hillanum. J. Sowerby, Min. Conch., vol. i, p. 41,

pl. xiv, upper fig.

1904. Protocardium hillanum var. umkwelanensis. Etheridge, Sec.

Rep. Geol. Surv. Nat. and Zulul., p. 79, pl. 1, fig. 16.

1906. Protocardia hillana. Woods, Cret. Fauna Pondl., Ann. S. Afr.

Mus., vol. iv, p. 307, pl. xxxvii, fig. 6.

1908. Protocardia hillana. Woods, Cret. Lam. Eng., vol. 11, p. 197,

pl. xxxi, fig. 6, and pl. xxxu, figs. 1-6 (with syn.).

1909. Protocardia hillana. Bullen Newton, Trans. Roy. Soc. 8. Afr.,

vol. i, pt. i, p. 76, pl. v, fig. 16 (with syn.).

Remarks.—A poorly preserved left valve of this ubiquitous species

is present in the collection. The concentric ornamentation appears

36 Annals of the South African Museum.

to have been coarse as in the Pondoland examples. The species occurs

in the Cenomanian, Turonian, and Senonian of Hurope, Cenomanian

of North Africa, Senonian of Pondoland, Zululand, and elsewhere.

Locality. Near Hombo, Quissama, No. 6827.

Famity VENERIDAE.

Genus CypRIMERIA, Conrad.

Cyprimerva 2, sp.

(Plate V, fig. 6.)

Description.—Shell rounded, oval, moderately convex, length a little

greater than height, inequilateral. Antero-dorsal margin short, very

slightly convex, passing gradually into the well-rounded anterior

margin; postero-dorsal margin long, gently convex, the posterior

missing inthespecimen. Umbones moderate, close together, anteriorly

curved, in front of which is a somewhat depressed area but no defined

lunule. Ornamentation of fine concentric striae, some stronger than

others.

Remarks.—This shell probably belongs to Cyprimeria and may be

compared with C. analoga (Forbes) (54, p. 178, pl. v, figs. 21-23) from

the Trichinopoli group of India, which it resembles rather closely.

There is an undescribed shell in the Pondoland collection of the

Transvaal Museum with which it is very probably identical, but the

state of preservation of the Angola shell prevents a definite decision.

Locality.—Massangano area, south of the river, No. 6750.

Cyprimeria ?, sp.

(Plate IT, figs. 13, 14.)

_ Remarks.—From Dombe Grande comes a small left valve referred

tentatively to Cyprimeria. It is preserved in the same matrix as

Crassatellites spp. from the same area. The shell is oval, long, and

somewhat compressed, the umbo moderately prominent and in front

of the middle, and the badly worn surface was probably ornamented

only with growth markings. The hinge is fairly strong and appears

to be provided with two diverging cardinals, a long, very thin, posterior

cardinal flanking the nymph, and an anterior lateral of moderate

length. The shell is obscurely carinated near the postero-dorsal

border.

Locality.—Dombe Grande, No. 6374.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 37

Genus Macrocatuista, Meek.

Macrocallista ?, sp.

(Plate V, fig. 5.)

Remarks.—The Senonian Ammonoidea of the Carimba localities

are accompanied by an indeterminable Inoceramus (6835) and

imperfect specimens of a small Venerid. The latter are oval shells,

longer than high and moderately convex, ornamented with fine con-

centric striations. These resemble Meretrix euglypha Woods (60,

p. 305, pl. xxxvi, figs. 7-10) from the Senonian of Pondoland, but

the ornamentation is finer.

Locality. Ridge 1300 m. W., 250 m. S. of well-rig, Carimba Camp,

No. 6838. Possibly also “ Map 65, 8E, Angola,” No. 6877.

Genus Baropa, Stoliczka.

Baroda %, sp.

(Plate II, fig. 20.)

Description.—Shell oblong, much longer than high, moderately

inflated, very inequilateral, expanded posteriorly. Ventral margin

broadly convex ; anterior margin short and well rounded ; postero-

dorsal margin long; posterior produced, with well-rounded margin.

Umbones small, incurved, approximate, placed very anteriorly,

with a small cordate lunular depression in front of them, not limited

by an impressed line. Ligament long and prominent, escutcheon

obscure. Surface ornamented with strong concentric striations.

Remarks.—The specimen possesses closed valves and hence the

determination cannot be accurately made ; the surface is for the most

part considerably damaged. The external features justify comparison

with Stoliczka’s Baroda, founded for two Indian species (54, pp. 167,

168), or possibly with Tapes.

Locality.—34 km. 8., 2 km. N. of Chio, No. 6775a.

Baroda 22, sp.

(Plate IT, fig. 16.)

Description.—Shell oval, oblique, longer than high, moderately

inflated. Umbones small, anteriorly directed, incurved, touching ;

38 Annals of the South African Museum.

umbonal region fairly prominent. Postero-dorsal margin long,

straight. Posterior produced, postero-ventral angle well rounded.

Anterior short, the anterior margin rounded and merging gradually

into the slightly convex ventral margin. Lunule? Escutcheon

absent or indistinct ; ligament moderately long and fairly thick. An

obscure, very much rounded carina extends postero-ventrally from

the umbones. Surface with strong, high, sharply crested and later-

ally compressed, concentric ribs, separated by deep furrows of about

the same width as themselves.

Remarks.—This is a well-preserved shell with the valves closed,

hence in the absence of the hinge the generic position is uncertain.

There is a mould of a right valve of the same species from Muscima.

The ornamentation is very striking; ribs of a similar type are seen

in a new species of Macrocallista from Pondoland to be described

shortly, allied to M. umzambiensis Woods (60, p. 304, pl. xxxvi,

figs. 4-6).

Localities.—34 km. S., 2 km. N. of Chio, No. 67756; Muscima,

No. 6776.

Famity TELLINIDAE.

Genus Trina, Linnaeus.

Sub-genus Palaeomoera, Stoliczka.

Tellina (Palaeomoera *), sp.

(Plate IV, fig. 7.)

Remarks.—Attached to a specimen of Trigonoarca angolensis is a

right valve which from its shape and peculiar ornamentation should

probably be referred to Palaeomoera. Shell oblong, longer than high,

posterior side shorter than the anterior, margins rounded. Umbo

small, directed slightly towards the anterior. Ornamentation on the

posterior of radiating ribs, on the central and anterior of faint spaced

radiating striae, the whole crossed by growth lines giving the posterior

ribs a roughened or serrated appearance. This shell is probably

closely allied to, perhaps identical with, an undescribed Pondoland

example in the collection of the Transvaal Museum.

Locality. Baba district, near Mossamedes, No. 6466.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 39

GASTROPODA.

Famity NERITIDAE.

Genus Nerita, Linnaeus.

Nerita malheiror, Chofiat.

(Plate I, figs. 14, 15.)

1888. Nerita malheiror. Chofiat, Mat. Strat. Pal. d’Angola, p. 82,

ple ive fier 10:

Remarks.—Nerita malheiroi is known from sixteen internal casts

in good preservation from the Pholadomya pleuromyaeformis beds

below the ammonite horizon at Dombe Grande. The short descrip-

tion given by Choffat includes the statement that the species appears

to have been smooth. Our specimen from Uchi has the form and size

of the Dombe Grande species but is in a different state of preservation,

for portions of the shell are present. The thin outer layer of the test

is almost smooth but with faint transverse striae and with the original

colour-markings of the shell still showing as irregular brownish patches ;

the inner portion of the shell has probably been removed, and a cast

of the inner surface of the thin outer layers shows well-marked trans-

verse striae.

There is a small Nerita from Dombe Grande with a similar form to

that of N. malheiror whose worn surface shows traces of growth lines

which may also be this species, and there is a cast of a Nerita in a

bituminous sandstone from Lifune River in the Loanda province

which compares favourably with the type.

Localities.—Uchi district, No. 6450; Dombe Grande, No. 6379 2 ;

Lifune Seep, Lifune River, No. 6867.

Nerita angolensis, sp. nov.

(Plate V, fig. 12.)

Description.—Shell moderately large, sub-globose, somewhat flat-

tened posteriorly, consisting of about three whorls; spire small,

lateral, almost flat; last whorl very large, inflated, flattened

posteriorly, rounded above, not concealing the spire; aperture not

seen; ornamentation of prominent transverse ribs which represent

the expanded edge of the outer lip at successive growth stages, the

40 Annals of the South African Museum.

ribs crenulate, the raised portions tending to follow radial lines,

especially on the upper part of the whorls.

Remarks.—This fine species is founded on a single specimen in which

the aperture and ventral portion were not seen. It seems to be closely

related to but quite distinct from N. rugosissuma Forbes (58, p. 342,

pl. xxv, fig. 6) from the Arialyur group of India, a smaller species

with a slightly higher spire and a rather less flattened posterior, while

the ornamentation consists of more widely spaced ribs.

Locality. Massangano area, south of theriver: holotype, No. 6839.

Famity NATICIDAE.

Genus GYRODES, Conrad.

Gyrodes cf. gents (J. Sowerby).

(Plate II, fig. 7.)

1816. Helix gent. J. Sowerby, Min. Conch., vol. ui, pl. exlv.

1843. Natica gaultina. dOrbigny, Pal. Frang. Terr. Crét., p. 156,

pl. elxxii, figs. 3, 4.

1912. Natica (Gyrodes) gaultina. Pervinquiére, Et. Pal. Tunis., p. 47

(with syn.).

1925. Natica (Gyrodes) genti. Cox, Ann. Transv. Mus., vol. xi, p. 203,

pl. xxxviii, fig. 8 (with syn.).

Remarks.—Four specimens from Dombe Grande in the condition

of casts are very similar to this well-known, variable, and widespread

Albian species; they are large shells having the proportions of

numerous examples from the Gault of Folkestone in the Sedgwick

Museum, with open umbilicus and the posterior portion of the whorls

well depressed near the suture. Faint transverse striations are

retained in one specimen. A discussion of the synonymy and varia-

tions of this species has recently been given by Cox (18, p. 204); the

Angola forms differ from that of Portuguese East Africa described by

Cox in having broad and deep canals bordering the sutures.

Natica feior Choffat (6, p. 81, pl. iv, fig. 9), described from two

specimens from the Pholadomya plewromyaeformis horizon at Dombe

Grande, appears to differ in having a more expanded outer lip and

smaller umbilicus.

Localitves.—Dombe Grande, Nos. 6341, 6349, 6361, 6365; Chipupo

district, No. 6503.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 41

Famity TURRITELLIDAE.

Genus TURRITELLA, Lamarck.

Sub-genus Haustator, Montfort.

Turritella (Haustator %) cf. acanthophora, Miller.

(Plate IV, fig. 13.)

1851. T. acanthophora. Miiller, Mon. Aachener Kreide, ii, p. 32, pl. iv,

fig. 5.

1887. 7. acanthophora. Holzapfel, Moll. Aachener Kreide, Palaeonto-

sraphica, xxxiv, p. 156, pl. xvi, figs. 9, 10, 12.

1912. 7. (Haustator) acanthophora. Cossmann, Pal. Comp., vol. 1x,

p. 117.

Remarks.—A large imperfectly preserved Twurritella has a very

strong resemblance to the figures given by Holzapfel of this species

from the Maestrichtian of Aachen, which has been placed by Cossmann

in the sub-genus Haustator Montfort. The specimen is crushed and

worn and the transverse striation is not shown, but the gently convex

whorls and ornamentation of four widely spaced spiral ribs bearing

coarse tubercles strongly recalls the Aachen species, which is of the

same size and proportions.

Locality.—Baba district, north of Mossamedes, No. 6472.

Turritella, sp.

(Plate IV, fig. 14.)

Remarks.—An imperfect and worn Turritella from Mossamedes is

figured here.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6487.

Famity CERITHIIDAE.

Genus CerirHium, Bruguiére.

Cerithium, sp.

(Plate IV, fig. 6.)

Remarks.—The Mossamedes Gastropoda include one specimen

which may be referred to Cerithiwm, using the term in the widest

sense. The whorls are moderately inflated and divided into two areas

42 Annals of the South African Museum.

by a sharp angulation towards the posterior side, the area behind the

angulation being concave and descending rapidly to the impressed

suture ; spiral striation can just be recognised.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6493.

Famity PROCERITHIIDAE.

Genus METACERITHIUM, Cossmann.

Metacerithium trimonile (Michelin).

(Plate I, figs. 4, 5.)

1838. Cerithium trimonile. Michelin, Mém. Soc. Géol., vol. iu, p. 100,

ple xdietiss oO.

1842. Cerithium trimonile. d’Orbigny, Pal. Frang. Terr. Crét., vol. 11,

PA soos ply cexcxe sal.

1906. Metacerithium trimonile. Cossmann, Paléoconch. Comp., vol. vii,

p. 54, pl. vi, figs. 29-31.

Description.—Shell conical, turriculate, with a spiral angle of about

28°, spire made up of flattened whorls with shallow sutures ; whorls

ornamented with two spiral rows of tubercles, the posterior series

being the smaller and situated close to the suture, the anterior series

larger and forming a ridge situated at some little distance from the

suture; in addition, the whorls are ornamented with faint spiral and

transverse striations; base rather flattened, merging with a well-

rounded angle into the last whorl, ornamented with marked transverse

striations. Aperture not seen.

Remarks.—The species is common in the Albian of Europe and has

been made the genotype of Metacerithium by Cossmann. There

appears to be considerable variation in the arrangement of the

tubercles, for d’Orbigny’s specimens have three rows of tubercles, of

which, however, the anterior is the greatest. The description given

by Cossmann of forms from the same area indicate that there is great

variability in this respect, for the middle row may be feeble or absent.

Four more or less damaged specimens from the Cuvo River area, all

of which retain well-preserved portions of the test, agree with Coss-

mann’s figures in having only two rows of tubercles. Of numerous

specimens in the Sedgwick Museum from the Gault of Folkestone

the majority have the middle row of tubercles developed, but forms

with only two rows are present and these our specimens resemble

closely.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 43

The species from the Arialyur group of India described by Stoliczka

as C. trimonile (58, p. 199, pl. xv, fig. 9, and pl. xix, figs. 2, 3) is probably

another species, as Cossmann has remarked, for the tubercles, arranged

in three spiral rows, almost form transverse ribs.

M. mosense Buvignier (40, sér. ili, p. 293, pl. lxxi, fig. 11) from the

-Cenomanian has three rows of tubercles of which the posterior is

the most prominent, while in M. ornatissimum Deshayes (40, sér. 111,

p. 290, pl. Ixxi, fig. 10) from the Albian the posterior row of tubercles

is stronger than the anterior.

Locality.—Caputo Hills, Cuvo River, No. 6887.

Famity CHENOPIDAE.

Genus CHENnorpus, Philippi.

[=A porrhais, da Costa. ]

Chenopus (Drepanochilus) reineckei, sp. nov.

(Plate II, fig. 8.)

Description.—Shell of moderate size ; spire of five whorls, the earlier

convex and evenly rounded; the last whorl angular owing to the

development of two prominent carinae separating the whorl into a

wide posteriorly sloping portion, a narrow flattened median portion,

and flattened base; sutures deep and sutural angle small; surface

apparently smooth; anteriorly produced into a short canal; outer

lip produced into a wide wing-like expansion with rounded end and

without digitiform extensions ; the posterior portion of the outer lip

retracted, concave to the aperture and sharply upturned at its margin.

Remarks.—The nature of the posterior portion of the outer lip and

the absence of digitation place this species in Meek’s section Drepano-

chilus (82, p. 324), which is placed under Chenopus by Cossmann (12,

vol. vi, p. 75). The absence of ornamentation readily distinguishes

this from other species, except Chenopus coquandi Cossmann (12, vol.

vi, p. 76, footnote) (=C. simplex (Coquand)), which, however, has the

wing differently shaped. The retraction and upturning of the posterior

portion of the outer lip is seen in the Eocene C. decoratus (Locard)

figured by Cossmann (12, vol. vi, pl. vii, figs. 7, 8), but the latter has

tubercles on the carina.

Localities.—16 km. W., 4 km. 8. of Mumbondo, No. 6771; hills

west of well location, Quimbaixie, Nos. 6821, 6823.

ad Annals of the South African Museum.

Chenopus, sp.

(Plate I, fig. 19.)

Remarks.—Two imperfect specimens resemble some species of this

genus. Shell moderately large, last whorl rather larger than half the

height, aperture elongate, outer lip missing; whorls moderately

convex, without carination, sutures well marked, whorls about twice

as broad as high ; ornamentation of numerous flexuous costae which

are slightly concave towards the aperture and crossed by numerous

fine spiral striae ; a few tubercles developed along the posterior portion

of the whorls; anteriorly produced into a canal, basal portion orna-

mented only with spiral striae.

In the ornamentation there is a resemblance to Aporrhais (Drepano-

chilus) calcaratus (Sowerby) (Cossmann, 12, vol. vi, p. 75, pl. iv, fig. 10,

and pl. v, figs. 1, 2, 14), but it is readily distmguished by the absence

of carination. Cox (18, p. 206, pl. xxxvii, figs. 4, 5) has recorded a

smaller but similar species from the Albian of Catuane, Portuguese

Kast Africa, as Chenopus (Arrhoges ?) sp., but the latter has a smaller

spiral angle and more convex whorls.

Locality.—Catumbella, Nos. 6417, 6418.

Famity STROMBIDAKE.

Genus StromsBus, Linnaeus.

Strombus 2, sp.

(Plate IV, fig. 12.)

Remarks.—This small shell has quite the form of some young

examples of Strombus ? incertus (d’Orbigny) (89, p. 27, pl. 11, figs. 19,

20) figured by Pervinquiére from the Cenomanian of Tunis, but differs

in the absence of spiral folds. The surface is delicately sculptured

with growth lines.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6484.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 45

Famity COLUMBELLINIDAE.

Genus Prrroponta, d’Orbigny.

Pterodonta cf. inflata, d’ Orbigny.

(Plate II, fig. 6.)

1842. Pterodonta inflata. dOrbigny, Pal. Franc. Terr. Crét., vol. u,

Pulse pls coxix,

1904. Pterodonta inflata. Cossmann, Paléoconch. Comp., vol. vi,

p. 116, pl. vin, figs. 10-12.

Remarks.—Two large but imperfect specimens from Uchi resemble

very closely the excellent figures given by Cossmann of this species,

the type of which comes from the French Cenomanian. Both

specimens exhibit traces of the varices, but the apertural regions are

damaged and hence the possibility of the forms belonging to Tylostoma

is not excluded; the resemblance to Cossmann’s figures is, however,

extraordinarily close.

Locality.—Uchi district, Nos. 6439, 6442.

Pterodonta afi. elongata, d’Orbigny.

(Plate I, figs. 7, 8.)

1842. Pterodonta elongata. d’Orbigny, Pal. Frang. Terr. Crét., vol. ii,

p- 316, pi cexvin, fig. 2.

1904. Pterodonta elongata. Cossmann, Paléoconch. Comp., vol. vi,

oe Satie

Remarks.—Three well-preserved casts from near Mumbondo in the

Loanda embayment resemble this form in shape and possess a ridge

at each half-volution due to the periodic dilation of the outer lip and

corresponding in position to the internal varices, while the aperture

is notched or sub-canaliculate anteriorly; they thus belong to

d’Orbigny’s genus and not to Tylostoma Sharpe (45) which is holosto-

matous. Cossmann (12, vol. vi, p. 116) has shown that the genera

are distinct. The present specimens have the posterior portion of the

outer lip expanded backwards as in d’Orbigny’s figure, but the anterior

canaliculate parts of the shells are too damaged to show whether the

short canal was twisted or not.

Locality.—9% km. W., 14 km. N. of Mumbondo, No. 6769.

46 Annals of the South African Museum.

Famity FUSIDAE.

Genus CryPTORHYTIS, Meek.

Cryptorhytis cf. bleichert (Thomas and Peron).

(Plate V, fig. 13.)

Remarks.—A rather large cast from Massangano appears to belong

to Meek’s genus Cryptorhytis and resembles C. blechert (Thomas and

Peron) (89, p. 70, pl. v, figs. 12-15) as figured by Pervinquiére and also

C. reynest (Coquand) (8, p. 188, pl. v, fig. 14), both Lower Senonian

forms. The columellar folds are barely suggested by small projec-

tions within the aperture.

Locality.—Massangano area, south of the river, No. 6757.

Genus Fusus, Lamarck.

Fusus %, sp.

(Plate IV, fig. 15.)

Remarks.—The Mossamedes Gastropoda include portions of three

whorls of a fusiform shell which may be placed here tentatively. The

spiral angle 1s very approximately 60°; the whorls are convex,

carinated a little in front of the middle, and ornamented with distant

spiral ribs.

Locality.—Mesados Cavalleros, near Mossamedes, No. 6485.

Famity CONIDAE.

Genus Conus, Linnaeus.

Conus %, sp.

(Plate IV, fig. 10.)

Remarks.—Internal cast, probably referable to Conus. From

Mesados Cavalleros, near Mossamedes, No. 6490.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 47

Famity ACTAHONIDAE.

Genus AcTAEON, Montfort.

Actaeon %, sp.

(Plate IV, fig. 5.)

Remarks.—Among the poor Gastropoda from Mossamedes is one

which can be regarded as an Actaeon or possibly a Tornatellaea. The

spire and part of the last whorl are visible, the whorls of the spire

being rather convex, the spiral angle about 60°. The spiral ribs are

elevated and separated by interspaces wider than themselves.

Locality. Mesados Cavalleros, near Mossamedes, No. 6489.

Famity RINGICULIDAE.

Genus AVELLANA, d’Orbigny.

Avellana incrassata (Sowerby).

(Plates fess 16), 17,)

1817. Auricula incrassata. J. Sowerby, Min. Conch., vol. u1, pl. clxin,

figs. 1-3.

1853. Avellana incrassata. d’Orbigny, Pal. Frang. Terr. Crét., vol. u,

p. 133, pl. elxvii, figs. 13-16.

1909. Avellana cf. wncrassata. R. Bullen Newton, Trans. Roy. Soc.

S. Afr., vol. i, pt. i, p. 30, pl. vin, figs. 6-8.

1925. Avellana incrassata. Cox, Ann. Transv. Mus., vol. xi, p. 207,

pie mexyl, fie. 7.

Remarks.—The collection contains a single specimen which appears

to be identical with the Blackdown species. Though far from being

perfect, the apertural region exhibits traces of the columellar folds.

The axial elements of the sculpture are not evident on the specimen,

but the narrow spiral ribs and furrows agree exactly with specimens

from Blackdown. In South Africa the species has been recognised by

R. B. Newton in the Manuan Creek (Albian) of Zululand and by Cox

from Portuguese East Africa.

Locality.—4 km. Mag. N. of Chio, No. 6797.

48 Annals of the South African Museum.

REFERENCES.

(1) Batty (1855).—“‘ Description of Some Cretaceous Fossils from South Africa,

etc.,” Q.J.G.S., vol. xi, pp. 454-465.

(2) BrBrano (1923).—“* Geologia e Riqueza Mineira de Angola,

Geol. de Portugal, t. xiv, pp. 127-217.

Com. Serv.

(3) Boum unD WEISSERMEL (1913).—‘‘ Uber tertiare Versteinerungen von den

Bogenfelser Diamantfeldern,” Beitr. geol. Erfor. Deut. Schut., Heft v,

pp. 59-111.

(4) Butten NEwron, R. (1909).—“ Cretaceous Gastropoda and Pelecypoda

from Zululand,” Trans. Roy. Soc. 8. Afr., vol. i, pt. i, pp. 1-106.

(5) Butten Newton, R. (1916).—‘ Cretaceous Brachiopoda and Mollusca from

Angola, Portuguese West Africa,” Trans. Roy. Soc. Edin., vol. li, pt. iii,

No. 15, pp. 561-580.

(6) CHorrat, P. (1888).—‘* Matériaux pour ? Etude Strat. et. Pal. de la Province

d’ Angola,” Mém. Soc. Phys. Hist. Nat. de Genéve, t. xxx, No. 2, pp. 1-116.

(7) Cuorrat, P. (1905).—‘‘ Nouvelles Données sur la Zone Littorale d’ Angola,”

Contr. & la Conn. Géol. des Colonies Port. d’Afrique, Comm. du Serv.

Géol. du Portugal, pp. 1-48.

(8) Coguanp, H. (1862).—‘“‘ Géologie et Paléontologie de la Région Sud de la

Province de Constantine,’ Mém. Soc. d’Emul. de la Provence, vol. ii.

(9) Coguanp, H. (1865).—Mon. de l Etage Aptien de l Espagne.

(10) Coguanpb, H. (1869).—Mon. du Genre Ostrea: Terrain Crétacé.

(11) Coguanp, H. (1880).—* Etudes supplémentaires sur la Paléontologie algé-

rienne, etc.,” Bull. de Académie d’ Hippone, No. 15.

(12) Cossmann, M. (1895-1907).—Hssais de Paléoconchologie Comparée.

(13) Cox, L. (1925).—‘“‘ Cretaceous Gastropoda from Portuguese East Africa,’ Ann.

Transv. Mus., vol. xi, pp. 201-216.

(14) D’OrBiany, A. (1842-48).—Paléontologie Frangaise, Terrains Crétacés.

(15) Douviti#, H. (1904).—“* Sur quelques fossiles de Madagascar,” Bull. Soc.

Géol. Frang., sér. iv, vol. iv, pp. 207-217.

(16) Dru er MunreR-CHALMAS (1881).—Mission des Chotts Tunisiens.

(17) pu Tort, A. L. (1926).—Geology of South Africa.

(18) Erueripes, R., Jr. (1904).—“‘ The Umkwelane Hill Deposit,” Second Report

Geol. Surv. Natal and Zululand, pp. 79-93.

(19) Forsss, E. (1846).—‘‘ Rep. on the Fossil Invertebrata from S. India, ete.,”

Trans. Geol. Soc. London, vol. vii, pp. 97-174.

(20) Fortav, R. (1917).—* Cat. des Invert. Fossiles de ’ Egypte, Terr. Crét. 2™°

partie : Lamellibranchia,” Geol. Surv. of Egypt, Pal. Series, No. 3.

(21) Gurnitz, H. B. (1871-75).—‘‘ Das Elbthalgebirge in Sachsen,’ Palaeonto-

graphica, vol. xx.

(22) GoLpFuss (1839).—Petrefacta Germaniae.

(23) Greco (1916-18).—“‘ Fauna Cret. dell’ Egitto,’ Pal. Italica, vol. xxiii,

pp. 93-161; vol. xxiv, pp. 1-58.

(24) Grecory, J. W. (1915).—* Contributions to the Geology of Benguella,”

Trans. Roy. Soc. Edin., vol. li, pt. iii, No. 13, pp. 495-536.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 49

(25) Grecory, J. W. (1922).—‘*‘ Suppl. Note on the Geology of Benguella, etc.,”’

Trans. Roy. Soc. Edin., vol. liii, pt. i, No. 7, pp. 161-163.

(26) GrresBacH (1871).—‘‘ On the Geology of Natal in South Africa,” Q.J.G.S.,

vol. xxvii, pp. 53-72.

(27) Have (1921).—Traité de Géologie.

(28) Haveuton, 8. H. (1925).—‘‘ Notes on Some Cretaceous Fossils from Angola

(Cephalopoda and Echinoidea),” Ann. 8. Afr. Mus., vol. xxii, pt. i,

pp. 263-288. Also translated into French in Com. Serv. Géol. de Portugal

(1924), vol. xv, pp. 79-106.

(29) Hotzapre., E. (1887—89).—‘‘ Die Mollusken der Aachener Kreide,” Palaeon-

tographica, vols. XxXiv, XXXV.

(30) KatsEer, Ericu (1926).—Die Diamantenwiiste Siidwest Afrikas.

(31) Kossmat, F. (1895).—‘‘ On the Cretaceous Rocks of S. India,” Rec. Geol.

Surv. India, vol. xxviii, pp 39-55.

(32) Muzzx, F. B. (1876).—“ A report on the Invert. Cret. and Tert. Fossils of the

Upper Missouri Country,” U.S.G.S. Terr. Rep., vol. ix.

(33) Merensxy, H. (1909).—‘‘ The Diamond Deposits of Liideritzland,”’ Trans.

Geol. Soc. 8. Afr., vol. xii, p. 18.

(34) Mozscu, C. (1874).—‘‘ Mon. der Pholadomyen,” Abhandl. Schw. Pal. Ges.,

vols. i-il.

(35) Morton, 8. G. (1833).—Am. Jour. Sci., Ist ser., vol. xxiii.

(354) Morton, S. G. (1834).—Syn. Org. Rem. Cret. U.S.A.

(36) Mutxumr, J. (1851).—Mon. der Petref. der Aachener Kreideformation.

(37) Noeturne, F. (1902).—** Fauna of Baluchistan, etc.,” Pal. Indica, ser. xvi,

vol. i, pt. 3.

(38) Pron, A. (1890).—Mollusques fossiles de Tunisie.

(39) Pervinquimre, L. (1912).—Etudes de Paléontologie Tunisienne : Gastropodes

et Lamellibranches des Terrains Crétacés.

(40) PictET ET CAMPICHE (1864—71).—* Description des fossiles du Terrain

Crétacé des Environs de Sainte-Croix,”’ Mat. pour la Pal. Suisse, sér. iii-v.

(41) Quaas, A. (1902).—‘‘ Fauna der obersten Kreidebildungen in der libyschen

Wiiste (Overwegischichten, etc.),’’ Palaeontographica, vol. xxx, pt. ii,

pp. 153-334.

(42) Ravn (1902).—Molluskerne i Danmarks Kridtaflejringer : Lamellibranchiater.

(43) Romaness (1916).—“‘ Note on an Algal Limestone from Angola,” Trans. Roy.

Soc. Edin., vol. li, pt. ili, pp. 581-584.

(44) SeauENZA, G. (1882).—“‘ Studi geol. Cret. Italia,” Atti Accad. Lincei, Roma,

ser. iii, vol. xii.

(45) SHarpr, D. (1849).—‘‘ On Tylostoma, etc.,”’ Q.J.G.8., vol. v, pp. 376-380.

(46) SHarps, D. (1850).—‘‘ On the Secondary Rocks of Portugal,” Q.J.G.S.

vol. vi, pp. 135-199.

(47) Sowmrsy, J. (1812-46).—The Mineral Conchology of Great Britain.

(48) Spatu, L. F. (1921).—‘‘ On Cretaceous Cephalopoda from Zululand,” Ann. S.

Afr. Mus., vol. xii, pt. vii, pp. 217-321.

(49) Spatu, L. F. (1921).—‘* On Upper Cretaceous Ammonoidea from Pondoland,”’

Ann. Durban Mus., vol. iii, pt. 2, pp. 39-57.

(50) Sparu, L, F. (1922) —‘“‘ The Senonian Ammonite Fauna of Pondoland,”’

Trans. Roy. Soc. 8. Afr., vol. x, pt. ili, pp. 113-147.

VOL. xxviii, PART 1. 4

50 Annals of the South African Museum.

(51) Sparta, L. F. (1922).—‘‘ Cretaceous Ammonoidea from Angola, etc.,” Trans.

Roy. Soc. Edin., vol. liii, pt. i, pp. 91-160.

(514) Spats, L. F. (1925).—‘‘ On Upper Albian Ammonoidea from P.E. Africa,

etc.,” Ann, Trans. Mus., vol. xi, pt. iii, pp. 179-200.

(52) Spatu, L. F. (1926).—‘‘ On New Ammonites from the English Chalk,” Geol.

Mag., vol. lxiii, pp. 77-83.

(53) SroticozKa, F. (1868).—Cretaceous Fauna of S. India (Gastropoda).

(54) SroriozKa, F. (1871).—Cretaceous Fauna of 8. India (Lamellibranchia).

(55) TRECHMANN (1927).—‘* The Cretaceous Shales of Jamaica,’ Geol. Mag.,

vol. lxiv, pp. 27-42, 49-65.

(56) van Hoerpen, E. C. N. (1921).—‘“‘ Cretaceous Cephalopoda from Pondoland,”

Ann. Transv. Mus., vol. viii, pt. i, pp. 1-48.

(57) WADE (1926).—*‘ The Fauna of the Ripley Formation, Tennessee,” U.8.G.S.

Prof. Paper 137.

(58) WHITE (1888).—** Contributions to the Palaeontology of Brazil,” Archiv.

Mus. Nac. Brazil, vol. viii.

(59) WHITFIELD (1885).—“ Brachiopoda and Lamellibranchiata of the Raritan

Clays, etc.,’’ U.S.G.S. Mon., vol. ix.

(60) Woops (1906).—‘‘ The Cretaceous Fauna of Pondoland,” Ann. S. Afr. Mus.,

vol, iv, pt. vii, pp. 275-350.

(61) Woops (1899-1913).—Mon. Cretaceous Lamellibranchia of England; Mon.

Pal. Soe.

(62) Woops (1911).—‘“‘ Palaeontology of the Upper Cretaceous Deposits of

Northern Nigeria,” Geol. and Geogr. N. Nigeria (Falconer), pp. 273-286.

(63) Woops (1912).—‘‘ Evolution of Inoceramus in the Cretaceous Period,”

Q.J.G.8., vol. xviii, pp. 1-20.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 51

FIG.

. Lucina angolensis sp. nov, Right valve of holotype, from Dombe Grande,

16.

Die

18.

19.

EXPLANATION OF PLATES.

PuateE I.

Albian Fossils from Angola.

The figures are of natural size unless otherwise stated.

No. 6334, p. 31.

. Lucina angolensis sp. nov. Dorsal view of holotype.

. Alectryonia cf. syphax (Coq.). Left valve from Camballa village, Cuvo River,

No. 6799, p. 14.

. Metacerithium trimonile (Mich.). From Caputo Hills, Cuvo River, No. 6887,

p. 42.

. Metacerithium trimonile (Mich.). Detail of ornamentation of part of same

specimen. X 3.

. Ostrea vesicularis Lam. Lower valve from Camballa village, Cuvo River,

No. 6862, p. 13.

. Pterodonta afi. elongata d’Orb. Cast from near Mumbondo, No. 6769

(Albian ?), p. 45,

. Pterodonta aft. elongata d’Orb. Another view of same.

. Trigonoarca cf. diceras (Seg.). Right valve of cast from Camballa village,

Cuvo River, No. 6828, p. 10.

. Neithea tricostata (Coq.). Right valve from Uchi, No. 6455, p. 17.

99 5 35 Variety from Uchi, No. 6464.

9 99 0 Detail of ornamentation at ventral margin of same

specimen as fig. 11, No. 6464. x 3.

. Neithea quinquecostata (Sow.). Right valve from Uchi, No. 6437, p. 17.

. Nerita malheiroi Choftat. From Uchi, No. 6450, p. 39.

aS s 3s Detail of ornamentation on dorsal surface of same

specimen. X 2.

Avellana incrassata (Sow.). Apertural view of specimen from Chio, No. 6797 ;

the columellar folds have been emphasised by the artist to indicate their

position, p. 47.

Avellana incrassata (Sow.). Another view of same.

Exogyra cf. conica (Sow.). Left valve from Camballa village, Cuvo River,

No. 6818, p. 15.

Chenopus sp. From Catumbella, No. 6418, p. 44,

Puate II.

Albian Fossils from Angola.

The figures are of natural size unless otherwise stated.

. Cardium (Laevicardium ?) sp. Left valve of specimen from Dombe Grande,

No. 6330, p. 35.

. Cardium (Laevicardium ?} sp. Anterior view of same specimen.

10.

Annals of the South African Museum.

‘ Astarte sp. Left valve from near Quimbaixie, No. 6821 (Albian ?), p. 29.

. Cardium (Trachycardium ?) sp. Left valve from Catumbella, No. 6415, p. 34.

. Astartesp. Left valve from Chipupo (plasticine impression of mould), No. 6510,

1D AAS)

. Pterodonta ct. inflata @ Orb. Cast from Uchi, No. 6439, p. 45.

. Gyrodes ct. genti (Sow.). Cast from Dombe Grande, No. 6365, p. 40.

. Chenopus (Drepanochilus) reineckei sp. nov. Holotype, from near Mumbondo,

No. 6771 (Albian ?), p. 43.

. Spondylus angolensis sp. nov. Left valve of holotype, from Dombe Grande,

No. 6369, p. 21.

. Spondylus angolensis sp. nov. Anterior view of holotype.

5 5 sp. nov. Mould of surface of left valve of holotype.

. Crassatellites sp. Left valve from Dombe Grande, No. 6373, p. 30.

. Cyprimeria? sp. Left valve from Dombe Grande, No. 6374, p. 36.

vi sp. Hinge of same specimen. x 14.

. Crassatellites sp. Left valve from Dombe Grande, No. 6372, p. 30.

. Baroda ?? sp. Left valve of specimen from near Chio, No. 6775), p. 37.

. Barbatia? sp. Left valve from near Quimbaixie, No. 6821. x 24. (Albian 2),

ps2.

. Lucina reineckei sp. nov. Right valve from near Chio, holotype, No. 6796,

p. 32,

. Lucina reineckei sp. nov. Dorsal view of the holotype.

. Baroda? sp. Right valve of specimen from near Chio, No. 6775a, p. 37.

Puate ITI.

Senonian Fossils from Angola.

The figures are of natural size unless otherwise stated.

. Protocardia hillana (Sow.). Left valve from near Hombo, Quissama, No. 6827,

p. 39.

. Veniella drut (Mun.-Chalmas). Right valve from the Baba district, near

Mossamedes, No. 6470, p. 28.

. Veniella drui (Mun.-Chalmas). Dorsal view of same specimen.

. Eriphyla cf. forbesiana Stol. Left valve from Mesados Cavalleros, near Moss-

amedes, No. 6488, p. 29.

. Lithodomus sp. Dorsal view of specimen from Mesados Cavalleros, near

Mossamedes, No. 6492, p. 22.

. Inthodomus sp. Anterior view of same specimen.

sp. Right valve of same specimen.

. Exogyra olisiponensis Sharpe. Lower valve from east of the Dondo-Quixinge

road, Quissama, No. 6798 (L. Senonian ?), p. 15.

. Cardita barronetti Mun.-Chalmas. Interior of right valve, from the Baba

district, near Mossamedes, No. 6482, p. 31.

Cardium (Trachycardium) reynoldsi sp. nov. Exterior view of the holotype,

a right valve, from the Baba district, near Mossamedes, No. 6480, p. 33.

Cardium (Trachycardium) reynoldsi sp. nov. Hinge of another right valve,

No. 6481.

Cretaceous Fossils from Angola (Lamellibranchia and Gastropoda). 53

FIG.

12. Cardium (Trachycardium) reynoldsi sp. nov. Detail of ornamentation on

posterior slope of same specimen as fig. 11, No. 6481. x 2.

13-15. Cardium (Trachycardium) reynoldsi sp. nov. Three views of an old right

CO -3 & Ct

14,

15.

Roo boo

valve, No. 6471.

PuateE IV.

Senonian Fossils from Angola.

The figures are of natural size unless otherwise stated.

. Trigonoarca angolensis sp. nov. Interior view of right valve from Baba,

No. 6469, p. 9.

. Trigonoarca angolensis sp. nov. Exterior view of same specimen.

5 < sp. nov. Dorsal view of same specimen.

. Nemodon natalensis (Baily). Exterior view of left valve from Baba, No. 6467,

p. 12.

. Actaeon? sp. From Mesados Cavalleros, near Mossamedes, No. 6489, p. 47.

. Cerithium sp. From Mesados Cavalleros, near Mossamedes, No. 6493, p. 41.

. Tellina (Palaeomoera ?) sp. Left valve from Baba, No. 6466, p. 38.

. Trigonoarca cf. trichinopolitensis (Forbes). Left valve of specimen from Baba,

No. 6465, p. 11.

. Trigonoarca cf, trichinopolitensis (Forbes). Detail of ornamentation of umbonal

region of left valve ofsame specimen. X 2.

. Conus? sp. From Mesados Cavalleros, near Mossamedes, No. 6490, p. 46.

. Lima (Mantellum) sp. Left valve from Mesados Cavalleros, near Mossamedes,

No. 6485, p. 22.

. Strombus ? sp. From Mesados Cavalleros, near Mossamedes, No. 6484, p. 44.

. Turritella (Haustator ?) cf. acanthophora Miller. From Baba, No. 6472,

p. 41.

Turritella sp. From Mesados Cavalleros, near Mossamedes, No. 6487, p. 41.

Fusus? sp. From Mesados Cavalleros, near Mossamedes, No. 6485, p. 46.

PLATE V.

Senonian Fossils from Angola.

The figures are of natural size unless otherwise stated.

. Plicatula sp. Specimen from Massangano, No. 6754, p. 20.

» sp. Another specimen, No. 6752.

» Sp. Same specimen, lateral view, No. 6752.

. Cardium (Acanthocardium) denticulatum Baily. Right valve from Massangano

Fort, No. 6766, p. 34.

. Macrocallista ? sp. Left valve of specimen from near Carimba, No. 6838,

pe.37.

. Cyprimeria? sp. Left valve of specimen from Massangano, No. 6750, p. 36.

FIG.

. Pholadomya aff. tigris Noetling. Right valve of specimen from Massangano

. Venilicardia cf. nicaisei (Coq.).

. Pleuromya? sp. Right valve from near Capolo, No. 6786, p. 22.

. Nerita angolensis sp, nov. Holotype, from Massangano, No. 6839, p. 39.

. Cryptorhytis cf. bleicheri (T. and P.). Cast from Massangano, No. 6757, p. 46.

Annals of the South African Museum.

Fort, No. 6767, p. 23.

. Pholadomya aff. tigris Noetling. Dorsal view of same specimen.

. Venilicardia cf. nicaiser (Coq.).

Cast of right valve from Massangano, No.

67586, p. 25.

Left valve of another specimen, No. 6758a.

Plate I.

del. E. T. Talbot.

ALBIAN FOSSILS FROM ANGOLA.

Plate II.

del. E. V. Talbot.

7 ALBIAN FOSSILS FROM ANGOLA.

Plate ITI,

del. E. T. Talbot.

SENONIAN FOSSILS FROM ANGOLA.

—

io)

ape

(2)

jo)

: : :

= Z

: 7 y

s 3

4 >

ae -

+ =}

== 5 2

Ann. S. Afr. Mus., Vol. XXVIII. Plate V.

del. E. T. Talbot.

SENONIAN FOSSILS FROM ANGOLA. |

( 55 )

2. On some New Therapsid Genera. By 8S. H. Haveuton, B.A., D.Sc.,

Hon. Curator, Palaeontological Collections.

(With Text-figures 3-20.)

STYRACOCEPHALIA, Nov.

Styracocephalus platyrhynchus gen. et sp. nov.

WHILST on a recent collecting expedition for the South African

Museum, Mr. L. D. Boonstra obtained from the Tapinocephalus beds

on the farm Boesmans Rivier, Beaufort West Division, a somewhat

crushed skull of a Dinocephalian which is manifestly different from

anything that has hitherto been described.

The most immediately striking features of the specimen are the

large backwardly projecting tabular horns, the massiveness of the

bones in the postorbital region, the small temporal opening, the

swollen cheek-like quadratojugal, and the shallowness of the snout.

Examination of the somewhat crushed palate shows that here, too,

are certain unusually interesting features; and, in spite of the

impossibility of tracing some of the suture-lines owing to the peculiar

mode of preservation of the skull, the type proves of sufficient import-

ance to merit description.

_In general appearance the skull is somewhat reminiscent of the

peculiar form called Burnetia nurabilis by Broom, and the resemblance

is heightened by the spongy nature of the bones of the region behind

the snout. Burnetia, however, comes from the base of the Middle

Beaufort Beds, some thousands of feet higher in the Karroo sequence

than this new form; and such resemblances as exist between the two

are therefore of very great interest.

As preserved, the skull has a maximum length of about 400 mm.,

and was probably about 420 mm. long when complete, and its

maximum width is about 290 mm. This latter is exaggerated by dorso-

ventral crushing, which has flattened out the large “cheeks.” The

minimum interorbital width is 97 mm., that between the temporal

openings about 160 mm.

56 Annals of the South African Museum.

The snout is fairly long, narrow, and shallow, having altogether an

appearance of feebleness when compared with the rest of the skull.

In dorsal view the main features of the skull are the very large and

thickened postorbital bones, the ‘“‘ roofed’ appearance of the skull,

and the lateral backwardly projecting horns. The postorbital com-

pletely shuts out the orbit in dorsal view, so that the orbit looks

outwards and slightly forwards. The temporal opening is very small

and looks wholly outwards.

The exact limits of the postorbital cannot be traced, but it is the

most massive bone in the skull. From the swollen supra- and post-

orbital crests it slopes gradually downwards and backwards, articulat-

Fic. 3.—Side view of type skull of Styracocephalus platyrhynchus

gen. et sp. nov.

ing in a long straight suture with the parietal and tabular, and being

apparently excluded from most of the upper border of the temporal

opening by an anterior projection of the squamosal.

The pineal foramen is small and stands in the middle of a broad, low

swelling on the parietal surface. The tabular horns are very prominent.

The limits of the tabular, squamosal, and quadratojugal are not

visible ; the last bone forms a swollen rounded angle to the cheek.

The flattening of the skull has caused the occipital plate to lie in

almost the same plane as the palate. The basioccipital condyle was

small and rounded and the foramen magnum was small. Above the

foramen is a large median boss on the supraoccipital, from which boss

superiorly runs a median ridge which reaches the occipital crest, and

two oblique ridges which almost reach the crest. These probably

mark the limits of the interparietal. Laterally the supraoccipital boss

is continued as a strong, rounded ridge of bone which runs slightly

upwards and forms the upper border of an elongate triangular de-

pression, whose lower border is formed by the paroccipital process and

On some New Therapsid Genera. 57

whose outer border is probably mainly squamosal. It is at the outer

-end of this depression that the small post-temporal fossa was prob-

ably located.

The details of the ventral surface of the basicranium are obscure

but the general features are peculiar. In front of the occipital

condyle there is a small median boss which lies posterior to the basi-

Fie. 4.—Dorsal view of type skull of Styracocephalus platyrhynchus

gen. et sp. nov.

sphenoidal vault. This latter is fanked by two bony swellings which

seem to form the anterior borders of the fenestrae ovales and are

therefore the basisphenoidal tubera. Instead, however, of the anterior

ridges from the tubera coalescing to form a median pterygoid ridge,

the anterior borders of the lateral swellings terminate suddenly in

faces that look forwards and outwards, and pass abruptly into the

almost flat pterygoidal plates whose outer portions run back towards

58 Annals of the South African Museum.

the quadrates. The lateral borders of the pterygoids approximate

anteriorly and then suddenly turn outwards to form the fairly strong

lateral pterygoid processes. In advance of these the chief features of

the palate are the absence of a suborbital vacuity and the presence

of series of strong teeth on the pterygoids and palatines on ridges

which form the borders to a median elongate oval shallow vault.

The stapes is a short bone with a swollen distal end.

The dentition of the upper jaw is not completely preserved. There

was certainly a large canine (which is now displaced), oval in cross-

section and not serrate, and a series of ten or eleven molars whose

cross-sections are elongate-oval and which decrease slightly in size

from the front backwards.

It can be seen from the brief description given that the skull agrees

in many features with that of Burnetia. Of the palate of that form

Broom says, “ So far as can be seen, it resembles that of the Gorgon-

opsians more than that of other Therapsids. As in the Gorgonopsian

palate, there are a pair of dentigerous areas partly formed by the

posterior and main ends of the palatines and partly by the pterygoids.

The pterygoids have large transverse processes like those of the

Therocephalians. There appears to be, as in the Gorgonopsians, no

suborbital opening in the palate ; if there is one it must be very small.

The posterior part of the pterygoid is very short.”

This description applies equally well to the palate of Styracocephalus,

which shows, in addition, a vaulted area between the dentigerous

ridges of the palate, a feature also displayed by the Gorgonopsia.

The structure of the back part of the palatal view, however, is not

so strongly reminiscent of the Gorgonopsia. The flat triangular plate

which forms the posterior part of the pterygoids is like that seen in the

Tapinocephalia (Mormosaurus, etc.) ; but the peculiar basisphenoidal

tubera and presamed massive basipterygoid processes are very different

from the conditions seen in either the Tapinocephalia or Titanosuchia,

or in the Therapsids generally. It is a feature far more nearly

paralleled in North American forms assigned to the Cotylosauria,

such as Labidosaurus or Captorhinus.

It will thus be seen that Styracocephalus does not readily fit into

any of the recognised sub-orders. Except for its basipterygoid

processes it is Therapsid in structure. Broom has considered the

Therapsid stem as having been derived from a Cotylosaurian close

to Captorhinus ; and this form gives further evidence as to the truth

of that conclusion, retaining as it does the early basipterygoid feature

alongside Therapsid advances. When considered as a Therapsid, the

On some New Therapsid Genera. 59

skull obviously contains features which are characteristic partly of

Therocephalia, partly of Gorgonopsia, and partly of the Dinocephalia.

The Therocephalian and Gorgonopsian features are seen mainly in the

palate ; the Dinocephalian features occur in the posterior palate and

basicranium, in the thickening of the bones of the postorbital and

Fig. 5.—Palatal view of type skull of Styracocephalus platyrhynchus

gen. et sp. nov.

supraorbital regions, and in the low, weak snout. In addition there

are the interesting specialisations shown by the tabular horns and the

swollen quadratojugals. .

In describing the type of Burnetia mirabilis, Broom was constrained

to found for its reception a new sub-order of Theriodontia which he

called the Burnetiamorpha, distinguished from the other sub-orders

by “ the remarkable specialisations ’ which the only known genus has

undergone. Some of these specialisations are also seen in the type

60 Annals of the South African Museum.

of Styracocephalus ; but this genus from a much lower horizon also

shows the retention of a primitive Cotylosaurid character, a slight

advance on the completely roofed skulls in the possession of a very

small temporal opening, and features which ally it to the Dinocephalia

on the one hand and to the Gorgonopsia on the other. A genetic

relationship between the Dinocephalia and Gorgonopsia has been

argued by Broom and others from a study of the external characters

of the skull, and by the present author from examination of the

details of the brain-case ; and in Styracocephalus we seem to have the

specialised representative of a group which combines characters of

both sub-orders and retains primitive features of that branch of the

Cotylosauria from which all the Therapsids presumably sprung. It

becomes necessary, therefore, to raise the group to which it belongs

to the rank of a sub-order which can be called the Styracocephalia, of

which this skull represents the only known genus. Discussion of the

relations between this sub-order and the much later Burnetiamorpha

must remain in abeyance until better preserved skulls of both are

obtained.

THEROCEPHALIA.

Choerosaurus dejagert gen. et sp. nov.

This new genus and species is found on a very interesting and almost

perfect little skull and lower jaw found by me some years ago in the

shale bank just above the house on Kuils Poort, Beaufort West

(S. Afr. Mus. Cat. No. 8797). It is obviously a member of the family

Scaloposauridae, but differs from any hitherto-described genus in

that family by its peculiar specialisations.

The skull is slightly distorted laterally, and the lower jaw is dis-

placed. The tip of the snout is missing and some of the bone weathered

away from the top and sides of the skull. The quadrate and articular

of the left side are crushed out of their true positions, the occipital

plate is not quite complete, and the palate is distorted. Nevertheless,

it is possible to give a sufficiently complete account of the chief

features to enable the identity of the form to be recognised.

The whole skull is long and narrow, and the orbits lie entirely in

the posterior half of the skull. The most striking features, however,

are the enormously swollen rounded boss at the front of each maxilla,

and the smaller but still very pronounced swelling on the dentary at

the posterior end of its lower border.

The dentition is of the type common in the Scaloposauridae where

more than one canineistherule. In the upper jaw the dental formula

On some New Therapsid Genera. 61

is 15 ¢c.3 m.12. All the teeth are simple poimted cones without

serrations. The 5 incisors occupy a length of jaw of 9°5 mm.; behind

the 5th isa diastema of 455mm. Then follow 3 canines all protruding

from the base of the boss on the maxilla; of these the anterior two

are slightly larger than the incisors, but the third is a much larger

Fic. 6.—Lateral view of type skull of Choerosaurus dejageri

gen. et sp. nov.

true canine tooth. The three teeth occupy 65 mm. The first molar

occurs directly behind the large canine, and the molar series occupies

20mm. The molars decrease slightly in size posteriorly.

The chief measurements of the skull are as follows :—

Greatest length . , ; =» QS) roryan,

Greatest width . 3 ! ‘ Spice eo ae

Snout to front of orbit . Babs. anne

Snout to pineal foramen : BOs.

Interorbital width i Be eats,

Intertemporal width : : ‘ Pl igees.

It is not possible to delimit most of the bones of the skull. The

nasal is long and comparatively narrow. The frontal forms almost

the whole of the upper border of the orbit, and there is a median ridge

where it meets its neighbour. The parietal was probably short. The

pineal foramen is small and lies just in front of the occipital crest, as

in the Gorgonopsia.

The jugal is a long slender rod of bone which has no upwardly

ascending postorbital process. It is probable that the postorbital

arch was not entire but was of the type seen in Scaloposaurus and in

Bauria. The only evidence of the postorbital arch that remains is

a tiny fragment attached to the matrix which fills the orbit and

temporal opening.

62 Annals of the South African Museum.

The squamosal is a somewhat reduced bone which overlaps the jugal

in front and passes inwards between the parietal and tabular. On the

occipital plate there is a small interparietal and a larger supraoccipital

which forms the upper margin of the post-temporal opening.

The lower jaw is slender with a Therocephalian type of mentum.

The full dentition is not visible, but there is a large canine correspond-

ing with that in the upper jaw. The

jaw is mainly composed of dentary,

which has a long, strongly developed

coronoid process. The outer face of

the dentary is furnished at the hinder

lower corner with a strong boss, on to

the lower hinder side of which the

angular overlaps. Above the boss

the bone is furnished with a longitu-

dinal groove. The inner face of the

dentary is grooved for the reception of

the splenial.

The angular shows, externally, a

large oval surface which is corrugated

and has a notch in its lower border.

The coronoid process of the dentary

is crushed down on to it so that no

surangular shows.

The articular is crushed out of posi-

tion and now les partly within and

partly below the angular. It has a

saddle-shaped articular surface which

/ is still in partial contact with the

displaced quadrate.

The chief features of the posterior

Fic. 7.—Dorsal view of type skull half of the imperfectly displayed pal-

S ae dejagert gen. et ate are the shortness of the pterygoid-

vomerine bar, the small suborbital

vacuity, and the large pyriform interpterygoid vacuity. In these

features the genus agrees with Akidnognathus and other Scalopo-

saurids.

Pelvis.—Associated with the type skull of Choerosaurus dejagert was

found an almost complete pelvis with the articulated proximal part

of a left femur. At first I thought that the pelvis was too large to

be part of the same skeleton as the skull; but, as in the Gorgonopsian

On some New Therapsid Genera. 63

Lycaenops, the total length from front of pubis to back of ischium is

greater than the maximum length of the skull, and as this new pelvis

is actually shorter than the skull of Choerosaurus I am inclined to

Fic. 8.—Lateral view of right side of pelvis of Choerosaurus dejagert. [The

ilium is slightly crushed towards the mid-line of the body from above.]

associate the skull and pelvis as the remains of one individual. It

must be recognised, however, that the association is not absolute, and

in future it might be necessary to dissociate the pelvis from the new

genus. In any case, its fea-

tures are sufficiently new and

noteworthy to meritdescription.

The pelvis has its two sides,

and the bones of its two sides,

in articulation.

Itium.—The most immedi-

ately striking feature of the

ilium is its large pre-acetabular

process, which differs entirely

from the large plate-like an-

terior portion of the ilium of

Dicynodon, Lystrosaurus, Thro-

naxodon, or Diademodon. A

strong, rounded ridge runs from

Fic. 9.—True lateral view of right

ilium of Choerosaurus dejagert.

the strong supra-acetabular process in a slightly curving direction

towards the upper anterior corner of the ilium. Below this ridge the

64 Annals of the South African Museum.

pre-acetabular surface is a thin plate of bone, concave externally,

thickening superiorly, and passing at almost a right angle at the ridge

to the main external surface of the ilium. If the latter were approxi-

mately vertical in life (on both sides it now faces upwards and out-

wards) then the concave portion of the pre-acetabular process must

have looked outwards, downwards, and forwards. The anterior end

of the ilium is truncate. The post-acetabular process is considerably

shorter than the anterior wing and, although slightly incomplete, was

undoubtedly bluntly rounded. The upper edge is almost straight.

Overhanging the deeply cupped acetabulum—of which the ilium forms

more than one-third—is a strong supra-acetabular boss, in front of

which the acetabular border is notched.

The ilium thus shows a well-marked advance in the direction of

what, from the evidence afforded by monotremes and marsupials,

must be regarded as the ancestral mammalian condition. There is

a distinct approximation to the trihedral type seen in Didelphys, with

a well-defined surface for the origin of the iliac muscle; the area for

insertion of the gluteal muscles is reduced when compared with that

of Cynognathus or Dicynodon ; and the post-acetabular portion of the

iliac plate is also reduced and nearer to the acetabular border.

Pubo-Ischium.—The pubis and ischium of both sides are quite

distinct. The pubic and ischial tuberosities are missing, but the

dimensions of the bones can be fairly accurately estimated.

At the acetabulum the pubis is massive and there is a pronounced

swelling at the anterior junction with the ilium for the origin of the

psoas muscle. Below the acetabulum the pubis rapidly becomes a

thin plate with a slightly swollen rounded outer edge. The direction

of this plate is forwards and downwards, and it gradually curves more

downwards until, at the ventral edge, it stands almost vertical. The

ventral anterior edge is thin, but swells laterally to the pubic tuberosity.

The whole bone is much shorter than wide. It meets its neighbour

in a symphysis which is not complete anteriorly ; and a large part of

its hinder border is excavate, forming the anterior border of the large

mammal-like obturator foramen.

The other border of this large foramen is formed by the ischium,

which is much longer than the pubis and meets its neighbour in a long

symphysis. Except near the acetabulum the bone is a thin plate

which narrows posteriorly. The pelvic opening was wide and shallow.

The large obturator fenestra is more mammal-like than in any

previously described Therapsid. There is no separate pubic foramen

such as is found in lizards. There is, however, evidence—in a small

On some New Therapsid Genera. 65

notch in one side of the symphysial line—that a small pubo-ischiadic

vacuity was also present. There can be little doubt that the large

fenestra here is homologous with the fenestra for the obturator externus

in mammals and not, as Williston suggested, due to an enlargement of

an opening between the four bones of the two sides of the plate.

Fie. 10.—Ventral view of pelvis of Choerosaurus dejageri.

Acetabulum.—The acetabulum, into the formation of which all three

pelvic bones enter, is deeply cup-shaped and not perforate ; and it is

overhung by a very strong supra-acetabular crest formed by the ilium.

At the puboischiac suture the border of the acetabulum is slightly

depressed. There is a large notch behind, and a smaller notch in

front of, the supra-acetabular boss. The posterior notch extends down

to the iloischiac suture.

The dimensions of the pelvis are as follows :—

Antero-posterior length of iliac crest . . prob. 36 mm.

Max. height of ium ; ; : : Eee

Width of acetabular portion of ilium . . a ish 8

Length of pubis in mid-line ? : Pe prop. la: ;;

Length of ischium in mid-line . of) prob: 32° ;,

Length of obturator foramen. : OR:

Width ,, He een : 20:

VOL. XXVIII, PART 1. 5

66 Annals of the South African Museum.

Femur.—Lying in the left acetabulum is the proximal half of the

left femur. The femur was a slender bone, whose proximal end is not

expanded as it is in Diademodon. ‘The head is flexed inwardly on the

shaft and has a very pronounced swelling on the ventral side, greater

than that seen in Moschops or in the femur figured by Gregory as

Aelurosaurus (?). Between the outer end of the proximal surface

and the external trochanteric margin the bone is thin with a sharply

rounded edge—the collum femoris. The bone as a whole is com-

pressed antero-posteriorly and widened laterally. The trochanter

Fie. 11.—Proximal end of left femur of Choerosaurus dejageri.

a, Dorsal view ; 6, ventral view ; c, proximal view.

major is a not very prominent short ridge on the lateral side of the

dorsal (anterior) face of the bone below the collum femoris. In the

trochanteric region the bone is not expanded as it is in Diademodon.

Between the trochanter major and the medial dorsal ridge, which runs

down from the head, the dorsal surface of the bone is shallowly

excavate.

The trochanter minor is much stronger than the trochanter major,

and lies almost at the external edge of the ventral (posterior) surface

of the bone. At its proximal end it runs inwards towards the head,

and proximal to it there is a shallow fossa. Anterior to its upper end

there is a narrow, short, and shallow intertrochanteric fossa.

The most striking things about the arrangement of the trochanters

are the lateral position of the trochanter minor, as compared with other

allied forms, and the reduction of the trochanter major. As a con-

sequence of the former, the area of insertion on the bone of the pubo-

ischio-femoralis internus muscle (=iliac+-psoas+ pectineus) is greatly

On some New Therapsid Genera. 67

increased ; this is correlated with the wide area of origin of the iliac

muscle on the illum, and indicates a muscle of great strength. It is

obvious, too, from the position of the trochanter minor that the dorsal

side of the femur must have faced outwards as well as forwards.

Discussion.—Hitherto, far less attention has been paid by students

of the Therapsida to details of the postcranial skeleton than to details

of the skull and lower jaw ; and discussion as to the relations between

the Therapsida and the Mammalia have been based largely upon the

latter. The discovery of the form under consideration seems to call

for more stress being laid upon the evolution of the limbs and girdles

than has hitherto been the case.

The pelvic girdle in the Dinocephala has been well described for

Moschops by Gregory. Broom has given us details of the same region

in Dicynodon and in Diademodon, and Watson in Lystrosaurus ; and I

am indebted to Dr. Broom for a drawing of the pelvis of the Gorgon-

opsian Lycaenops, which has not yet been published, and for a de-

scription of a Therocephalian pelvis. Figures, too, have been given of

the pelvis in the Dromasauria by Broom.

Comparison of the pelvis of Choerosaurus with those of the other

Therapsids known shows that the former is further advanced along

the mammalian line than any of the other types, even including that

of Diademodon, which is representative of a group that has the most

mammal-like skulls of all the Therapsids. Choerosaurus occurs in the

middle of the Hndothiodon zone of the Lower Beaufort Beds; and

although its skull is specialised it lacks a complete postorbital arch,

and could well be considered as a specialisation from an earlier form

which had the features necessary for a pre-mammalian skull. The

pelvis of the Upper Beaufort Cynognathus and Diademodon is less

mammalian than that of Choerosaurus ; and it would seem, in con-

sequence, as if the line of Therapsids which ultimately gave rise to

the ancestral mammals was already differentiated in Lower Beaufort

times, and may have evolved from some early Therapsid stock con-

temporaneously with the lines that gave rise to the various known

orders of mammal-like reptiles. Broom considers that this pre-

mammalian stock may be the Anningiamorpha, which is known from

a single skull, Anningia megalops, from low down in the Tapinocephalus

zone of the Lower Beaufort Beds.

The postcranial skeleton in the members of the Scaloposauridae

Akidnognathus, Ictidognathus, Ictidodon, Icticephalus, Simorhinella,

and Scaloposaurus is unknown, and it is therefore impossible to say

definitely that Choerosaurus belongs to the same family. In view of

68 Annals of the South African Museum.

its skull characters and dentition, however, I group it as a specialised

genus of the Scaloposauridae until such time as definite evidence forces

it to be placed in a new family.

GORGONOPSIA.

Hoarctops vanderbyli, gen. et sp. nov.

Three Gorgonopsians only have hitherto been described from the

Tapinocephalus zone, viz.—EHriphostoma microdon, Scylacognathus

parvus, and Galesuchus gracilis. Of these the first two are known from

Fic. 12.—Side view of type skull of Loarctops vanderbyli

gen. et sp. nov.

rather incomplete skulls and have never been adequately figured ;

and Galesuchus lacks the snout.

There is in the South African Museum collection a rather weathered

skull (S. Afr. Mus. Cat. No. 5598) from the well-known locality,

Abrahams Kraal, Prince Albert District, which seems to differ

sufficiently from the above-named forms to be considered as a separate

type. It is of the general dimensions of Scylacognathus parvus, and

has a short snout as in that genus ; but the dental formula is different,

as are certain other features. From Eriphostoma it differs in not

having a laterally compressed snout and in its dentition.

The chief dimensions of the skull in question are as follows :—

Greatest length : . prob. 129 mm.

Greatest width : ~ probs 94k

Snout to front of orbit. : Noe S

Snout to pineal foramen . : OM

Interorbital width . 36

Intertemporal width : 39

On some New Therapsid Genera. 69

The tip of the snout is weathered away ; but the snout was obviously

low and broad, and the preorbital region—as in Arctops—was quad-

rangular in cross-section. The nostril was wider than high and looked

mainly forwards. The weathered snout shows that the median in-

terchoanal bar of the palate is definitely made up of two prevomers.

The number of incisors in the upper jaw 1s not determinable ; but

there are definitely only three incisors in the lower jaw, and it is highly

probable that, as in Eriphostoma, Oey.

there were only four incisors in |

the upper. The roots of the last

two incisors in the right premaxilla

are preserved; and the incisor

series occupied not more than 10-5

mm. Behind the last incisor there

is a wide diastema of about 138

mm., corresponding to the position

of the lower canine. The upper

canine has an antero-posterior

diameter of 6 mm. Behind it is

a diastema of 8 mm., and then

follow 3 small backwardly directed

pointed molars, occupying another

8mm. In the lower jaw 2 molars

only can be seen, so that the den-

tal formula is i.4 c.t m.3. The

teeth are apparently all simple

and unserrated. The width of a

frernout atthe scammes is about 626 !s_Doral view of rype skull of

Eoarctops vanderbyli gen. et sp. nov.

43 mm.

The top of the skull is flat and the pineal foramen is in front of the

middle of the temporal opening. Nevertheless, it is almost at the

back of the top of the median bar, as almost immediately behind the

occipital plate begins. This slopes backwards slightly to the foramen

magnum and is concave from side to side.

There is a preparietal wholly in advance of the pineal foramen,

and the postfrontals are of the usual Gorgonopsian type. The large

frontals form an appreciable part of the supraorbital border. The

prefrontal is large and wholly on top of the skull.

The occipital plate does not show sutures clearly. A large median

interparietal lies wholly on the back of the skull. Below it is a smaller

supraoccipital. The tabular, if present, must have been small.

70 Annals of the South African Museum.

The articular surface of the quadrate lies considerably below the

level of the occipital condyle and below the level of the cheek-teeth.

Weathering of the back of the skull shows that the main portion of

the quadrate was a vertically placed plate lying pressed against the

front of the squamosal. Its upper edge is on the level of the middle

of the foramen magnum.

The palate is not well displayed ; but it is obvious that the den-

tigerous border of the maxilla lies considerably below the ventral

surface of the prevomers, so that there was probably at least an

incipient secondary palate. As in Arctops, there is a distinct step in

the outline of the upper jaw in front of the canine. Between the

palatal portions of the pterygoids there is a deep V-shaped vault ;

and the ridges bounding this are dentigerous. The basisphenoidal

(vomerine) tubera lie well below the level of the occipital condyle.

The lateral process of the pterygoid is strong and deep. In front of

it there seems to be evidence of a small suborbital vacuity ; but this

may be due to the displacement of the ectopterygoid.

Lycaenodontoides bathyrhinus gen. et sp. nov.

This new genus and species is founded on an almost complete skull

and lower jaw and partial skeleton found at Oudeberg, Graaff Reinet,

and presented to the South African Museum by the Very Rev. J. H.

Whaits (S. Afr. Mus. Cat. No. 3329).

The skull lacks the portion behind the pineal foramen, and in general

appearance agrees with Aelurosaurus felinus. It is, however, larger,

has a different dental formula, and the details of its structure

differ.

The dental formula is 1.5 c.1 m.6, as in Lycaenodon longiceps. The

five upper incisors are long and pointed, the third being the largest,

and the fifth considerably smaller than the others; the incisors

occupy 26 mm. Following a diastema of 8 mm. comes the large

canine, serrated posteriorly in at least the upper part of its crown and

having an antero-posterior diameter of about 10 mm. The diastema

between it and the first molar is 9 mm. long. The molars are

moderate-sized, simple, pointed, backwardly curved, and serrated on

the posterior border in the upper half of the crown. The whole dental

series occupies 79 mm.

Behind the canine the edge of the maxilla bends down, and is convex

to a greater degree than in Aelurosaurus felinus ; but there is no step

in the dentigerous border in front of the canine.

On some New Therapsid Genera. cht

The nostril is almost terminal, with no internasal septum. The

septomaxilla has a turbinal process as in other Gorgonopsia.

< Rupees 7 ‘4

- | Vi Lijjjys

Fie was AA ; oe ee.

Fic. 14.—Lateral view of type skull of Lycaenodontoides bathyrhinus

gen. et sp. nov.

The chief measurements are as follows :—

Snout to front of orbit : ey <i ey GOP maT

Height of orbit : Seat De NGA

Height of skull at 3rd ein ' By OORIE.:

Height of skull at canine. ‘ ieee.

Snout to front of pineal foramen . eb ie a.

Interorbital width . RS yaa

The top of the skull is pitted and rugose, and passes over by a rounded

surface gradually into the less rugose cheeks.

The parietal region was probably as wide as the interorbital ;_ but,

unlike most Gorgonopsia, the preparietal is very small. In this it

agrees with Lycaenodon longiceps and with Lycaenoides angusticeps,

and approximates to Sycosaurus, where the preparietal is altogether

absent.

The orbit is round, and the jugal forms a considerable portion of its

anterior lower border. The lachrymal is fairly small. The prefrontal

is large.

The lower jaw is furnished with a strong mentum. The symphysial

72 Annals of the South African Museum.

part of the jaw makes an angle with the lower border of the ramus

similar to that in Aelurosaurus tenuirostris. The splenial forms part

of the symphysis. The angular extends far forward as a slender

wedge between the dentary and splenial.

Shoulder Girdle—Among the portions of the postcranial skeleton

preserved is a fairly complete shoulder girdle. The chief features

displayed are the large size

pS of the clavicles and inter-

we Wi SSSQ .

HA . clavicle.

fy AV \V"\\\\ .

/j . \\\\ \ we

j . \\\\ © \\

The scapula was prob-

ably about 100 mm. long.

Its proximal end is broad,

with a broad, thin, anterior,

supracoracoid plate. Above

this the bone narrows con-

siderably, and then swells

into a thin plate whose dis-

tal width is about 38-40

mm. There is no acromion

process. The scapula forms

half of the glenoid cavity,

which faces largely down-

wards, and there is a strong,

ridged, supraglenoid border.

The precoracoid and cor-

acoid are largely covered in

the specimen in ventral

view by the clavicle and

interclavicle, and undoubt-

edly in life the precoracoid

was largely underlain by the

mesial expansion of the

clavicle. The precoracoid

forms part of the border of the glenoid cavity. This participation

of the precoracoid in the glenoid is an unusual feature in the Therap-

sids, as far as the structure of the shoulder girdle is known within

the group.

The clavicle is an unusually large bone with a strongly expanded

mesial end. In anterior view it is a bent blade some 90 mm. in length,

which is wider distally than proximally. In ventral view it is seen

that the mesial portion of the bone is widely expanded towards the

MXK

x »: S S > ‘

CR Wass

Fic. 15.—Dorsal view of type skull of Lycae-

nodontoides bathyrhinus gen. et sp. nov.

On some New Therapsid Genera. 73

posterior end of the interclavicle, forming a triangular plate which

meets its neighbour in the middle line below the interclavicle, extend-

ing as far back as the suture between the precoracoid and coracoid.

Its anterior edge in ventral view is straight, its posterior edge slightly

concave. Its length in

the mid-line of the body

is 37 mm.

The interclavicle is a

large, elongate, median

plate with, probably, lat-

eral wings lying on the

clavicles in its anterior

half. Ventrally it has an

elongate boss in the mid-

line just behind the clav-

icles. The length of the

bone is about 90 mm.

Hand.—Among the

skeletal elements pre-

served there is also the

larger portion of a right

manus with the bones in

articulation with one an-

other. It differs con-

siderably from the manus

of Aelurognathus tagri-

ceps. The third and fourth

digits are complete, and

each definitely contains

only three phalanges, so

that the digital formula Fic. 16.—Shoulder girdle of type of Lycae-

nodontoides bathyrhinus.

was in all probability 2, 3,

3, 3, 3. |

Of the carpus we have the radiale and two closely articulated

centralia. The former has a trefoil outline seen dorsally, but is

transversely elongate in palmar view; while the centralia—closely

pressed to the radiale—are much larger in palmar than in dorsal view.

The inner centralia is larger than the other. There are five carpalia.

Carpalia 1 and 2 articulate with the Ist centrale; carpal 3 has its

proximal border between the two centralia and touching the radiale ;

carpal 4 is the largest of the series, and must have articulated with

a, Anterior view; 6, ventral view.

74 Annals of the South African Museum.

the now-missing ulnare besides touching the 2nd centrale; whilst

carpal 5 is a very small bone. There was possibly a rudimentary

prepollex between the radiale and the first carpal.

Of the metacarpals the IVth is the longest ; but it is more slender

than the IIIrd, which has a much broader shaft. In the digits the

Fic. 17.—Manus of type of Lycaenodontoides bathyrhinus.

a, Palmar view; 0, dorsal view.

claws are long and slender; and on the palmar side of the claw of

digits 3 and 4 there is a knob-like boss near the proximal end. The

first and second phalanges are comparatively short and broad. The

measurements of the metacarpals and digits are as follows :—

Length, | MEE | a

Metacarpal I ; 10-0 mm. 9-5 mm. 33

Metacarpal II. Oe Oe 12-5 mm.

Digit 1 iLO ws 13-05) 12-07

Metacarpal III_ . INT Se) os TSO 15-Oo ae

Digit 1 10:53 14:0) 5. ses

Digit 2 ISO) ee Oa: 1 Ope

Digit 3 : 26-00 5, 1l=Oy < »

Metacarpal IV. PAU) S as Igle5, ae 43-3) ae

Digit 1 9:00". SON ‘TO 0s

Digit 2 Asai sO =Qvide 5 EO» £2 10-ba

Digit 3 yas ae NOOR 43 ae

Metacarpal V LicO<.,, OOS TAO

Digit 1 : ae LO-Oraes :

On some New Therapsid Genera. 75

Hipposaurus boonstrar gen. et sp. nov.

This new and very striking form was found on a recent collecting

expedition by Mr. L. D. Boonstra at Klein Koedoeskop, in the Division

of Beaufort West, at a fairly high horizon in the Tapinocephalus zone

(S. Afr. Mus. Cat. No. 8950). The type consists of an almost complete

skeleton, but hitherto only the skull and lower jaw have been partially

cleared from the highly intractable matrix. It is proposed to give

a description of the whole skeleton at a future date.

Fie. 18.—Lateral view of type skull of Hipposaurus boonstrai

gen. et sp. nov.

The skull is of somewhat remarkable shape. The snout is long,

narrow, and deep, the depth increasing regularly from the anterior

end to the plane of the antorbital border. The sides of the snout are

nearly vertical, and they pass by a fairly sharp curve into the narrow,

slightly convex, upper surface. Immediately in front of the orbit the

sides are widely excavate into a basin-like depression whose contain-

ing wall is very thin. This depression was probably overhung by a

preorbital ridge.

At about the plane of the middle of the orbits the top surface of the

skull is bent, so that the top of the skull makes an obtuse angle with

the top of the snout. There are well-marked supraorbital crests,

between which and the central ridge the frontals are broadly concave.

In their posterior part this concavity is continued as a well-defined

irregular groove which extends back to the level of the pineal foramen.

76 Annals of the South African Museum.

The pineal foramen is surrounded by a rounded boss of bone and is not

far in front of the occipital crest.. The intertemporal region is much

broader than the interfrontal. The postorbital arch is massive, and

the temporal fossa small and looking outward and slightly backward.

The limits of the bones of the top and sides of the skull cannot all be

defined, but the following facts are visible. The preparietal is a small

narrow bone lying wholly

in front of the pineal

boss. It is flanked by

the parietal, which is a

triangular - shaped bone

with the middle of the

occipital crest as its apex

and having its posterior

corner acutely wedged

between the tabular and

the postorbital. The

latter forms the anterior,

upper, and part of the

posterior borders of the

temporal fossa. The

postfrontal and prefron-

tal are both large bones.

The frontal sends for-

ward a wedge between

the nasals, separating

them for some distance.

The occipital plate

stands at right angles to

Fic. 19.—Dorsal view of type skull of Hippo- aoe ee ec

saurus boonstrai gen. et sp. nov. and is strongly concave

with a weak, median,

vertical ridge. The parietals form a slightly overhanging crest, below

which a large median interparietal extends down the plate nearly

to the top of the foramen magnum. The sutures between supra-

occipital and paroccipital are not apparent; but the supraoccipital

is certainly shallow in the middle line and expands considerably later-

ally. The tabular is a high, narrow plate of bone which faces in-

wards and backwards, lies behind the parietal and squamosal, flanks

the interparietal and supraoccipital, and passes down almost to touch

the quadrate.

On some New Therapsid Genera. i

The foramen magnum is higher than wide. The exoccipitals

apparently meet above it and form its lateral borders. The basi-

occipital forms the rather small rounded condyle, which lies well

above the level of the articular surface of the quadrate.

In hinder view the quadrate mass is displayed as a high, broad plate

of bone, with a rounded bulge in the middle of the lower half of its

posterior surface. I am not able to distinguish a separate quadrato-

jugal in the mass.

The basicranial region has been partly cleared, but the palate is as

yet undeveloped. The basicranial tubera are deep, and between them

Fie. 20.—Occipital view of type skull of Hipposaurus boonstrar

gen. et sp. nov.

the palatal surface has a short, deep, rounded excavation. The

vomerine-pterygoid bar is comparatively short. The lateral pterygoid

processes are strong and deep, their lower corners showing below the

lower border of the lower jaw when the jaws are shut. Passing from

the foramen ovalis to the inner angle of the quadrate is a rod-like

stapes, whose outer end rests just above the posterior end of the

quadrato-pterygoid bar. The stapes points mainly downwards and

slightly backwards and outwards.

The dental formula of the upper jaw is 1.5 c.l m3. The fifth

incisor is smaller than the others ; all are simple, pointed teeth, finely

serrated on the posterior border. The five incisors occupy 22 mm.,

and are followed by a diastema of 6mm. The canine is large, with an

78 Annals of the South African Museum.

antero-posterior diameter of 12 mm. at the maxillary edge. The first

molar is separated from it by a diastema of 11 mm., and the three

small molars occupy 16 mm.

The chief measurements of the skull are as follows :—

Maximum length

Maximum breadth

Snout to front of orbit

Snout to front of antorbital dopteauine

Snout to bend in mid-line of skull

Length of top of skull

Interorbital width .

Intertemporal width

Height from bottom of dudtieate to sop of cea :

174 mm.

LOU

LOO

3. Pareiasaurian Studies.

Part I.—An Attempt at a Classification of the Parevasauria based on

Skull Features—By 8S. H. Havenuton, B.A., D.Sc., Hon.

Curator, and L. D. Boonstra, M.Sc., Assistant in Palaeontology.

Tue identification of Pareiasaurian material from the Karroo Beds

of South Africa offers certain peculiar difficulties, and hitherto very

few attempts have been made to classify the various examples which

have been collected, and only one attempt has been made to divide

the known species into clearly defined genera. This was done by

Watson, who based his classification on the material in the British

Museum collection and on one skeleton in the South African Museum,

and endeavoured to separate the following Pareiasaurian genera :

Pareivasaurus, Owen; Propappus, Seeley ; Anthodon, Owen; Brady-

saurus, Watson; and Embrithosaurus, Watson. Watson’s classifica-

tion was not accepted by Broom, who has subsequently described new

forms under the generic name Pareiasaurus on the grounds of estab-

lished usage, whilst recognising a necessity for further subdivision,

and suggesting that Watson’s generic names should be relegated to

the status of sub-genera. :

One of the difficulties which faces the student of Pareiasaurian

systematics is the variability shown by individual skulls, a variability

which, as Broom has pointed out, might almost induce the worker

to make each skull the type of a separate species. This variability

is, in large part, due to post-mortem deformation, and the writers

of this paper have attempted to use as a basis of classification features

which are the least likely to be affected by such changes as lateral

or dorso-ventral compression. The number of specimens we have

been enabled to examine is far greater than any other worker has

studied, and we feel that we are justified in putting forward this

scheme of classification as a basis for future study and criticism.

The classification is based largely on the features of the skull and

lower jaw. Where possible, it has been checked by reference to the

characters of the postcranial skeleton ; and the result must be taken

as a basis for a much fuller description of all the material we have

80 Annals of the South African Museum.

studied. That material consists mainly of the large collection of

Pareiasauria which the South African Museum has amassed by

collection and by donation ; but the senior author has also had the

opportunity of examining critically the historic specimens in the

British Museum and the type of P. acutirostris in the Albany Museum.

We are also indebted to Dr. R. Broom for some sketches and notes

upon the somewhat incompletely described species whartsz and

strubent, whose types are now in the American Museum of Natural

History.

Teeth.—As is well known, the teeth of Parelasauria are single-rooted,

and have crowns which are broader than thick and furnished with

a series of marginal cusps. Examination of the dental series of any

single skull will show that, although the anterior teeth may differ

from the posterior teeth in actual and relative dimensions, the teeth

throughout the series usually agree with one another in the nature

and number of the cusps.* <A study of the cusping of a series of skulls

shows, however, that an obvious distinction exists between forms

like serridens, serrarius, peringueyt, and omocratus on the one hand

(with numerous cusps arranged round the edge in a regular segment

of a circle), and bainz and bombidens on the other (with a few cusps),

with schwarz: occupying a somewhat intermediate position.

The actual number of teeth present seems to be, within certain

limits, a function of the age of the skull; and any scheme of classifica-

tion which takes cognisance of the number of teeth can only be based,

therefore, upon the space occupied by a complete series, especially in

view of the fact that the dimensions of the anterior teeth usually

differ appreciably from those of the posterior ones.

Top of Skull_—Post-mortem compression may have the effect of

masking the true form of the snout of a skull or alter the relation

of the descending cheek to the rest of the skull; but, beyond a twist-

ing, whose presence can easily be detected and whose amount can be

estimated, it can have little effect upon the broad, flat plate of bones

which forms the dorsal interorbital and postorbital surface. We have,

therefore, taken into consideration the shape of this surface, defining

that shape in terms of the ratio between the interorbital width and

that between the centres of the prominent tabular bosses, designating

the latter for convenience the tabular width. This has supplied a

ready means of distinguishing three groups among the skulls having

teeth with few cusps—viz. those whose interorbital width is greater

than, approximately equal to, or less than the tabular width.

* An exception to this occurs in skulls placed in the new genus Nochelesaurus.

Pareiasaurian Studies.—Pavri TI. 81

The Cheek.—Kxamination of a well-preserved skull such as that of

Pareiasuchus peringueyt shows that, although the two cheeks are

generally alike, there are minute differences between the sculpture

and arrangement of the bosses on the two sides. The hinder border

of the cheek is furnished with a series of bosses which, in different

species, differ in prominence. In some species, too, the cheek is

obviously long ; in others, comparatively short. This last difference

we have used in our classification, estimating the “length” by the

ratio between the distance between the top of the tabular boss and

the lowest corner of the cheek, and that between the hinder border

of the orbit and the middle of the cheek.

Basis of Classification.—The large and earliest Pareiasaurs have

teeth with elongate crowns carrying a few marginal cusps; the

Pareiasauria from the upper half of the Lower Beaufort Beds have

broad, shortened teeth with somewhat numerous cusps. We have,

therefore, made our primary subdivision on the nature of the teeth,

recognising at the same time that forms in one of the new subdivisions

may be actually closely allied to forms in another. The other factors

we have used in defining genera and species are—(a) the shape of the

top of the skull, (b) the shape of the snout, (c) the depth and nature

of the cheek, and (d) the shape of the lower jaw. The classification

thus obtained is admittedly an artificial one; but we believe that by

its use as a basis for the identification of species it will be possible

ultimately to arrive at some phylogenetic classification within the

group, linking together successive species in the time-scale.

DIVISION A.

Forms having teeth with few cusps (less than 9) arranged irregularly

round the crown.

SUBDIVISION Aa.

Forms with interorbital width appreciably less than the “ tabular

width.”

1. Genus BrapysaurRus, Watson.

Snout broad and rounded. Cheek shallow. Six cusps on each

tooth.

Bradysaurus baini Seeley.

Posterior border of cheek fairly smooth.

VOL. XXVIII, PART 1. 6

82 Annals of the South African Museum.

To this species we assign the following specimens :—

a. The type in the British Museum, from De Bad, Prince Albert Division.

6. Skull and skeleton 4347 in S. African Museum, from Abrahams Kraal, Prince

Albert Division.

c. Skull and skeieton 5127 in S. African Museum, from Leeuw Rivier, Beaufort

West Division.

d. Skeleton 3533 in S. African Museum, from Hottentots River, Beaufort West

Division.

e. Partial skull and skeleton 4999 in S. African Museum, from Grootfontein, Prince

Albert Division.

f. Skull in Titbingen Geol.-palaontolog. Institut, from Abrahams Kraal, Prince

Albert Division.

Bradysaurus seeleyt sp. nov.

Posterior border of cheek with well-defined bosses.

This species is based upon the two specimens in the British Museum

which were described by Seeley as Parezasaurus bombidens. The one

of these which may be looked upon as the type of the new species is

No. 49426, and came from Palmietfontein; the other (R1970) is

known as the Tamboer specimen.

B. bombidens was founded by Owen on a very fragmentary maxilla

and lower jaw—a very unsatisfactory type which cannot accurately

be placed in our classification; the teeth of seeleyz, however, seem

to be relatively broader and flatter than in the type of bombidens.

To this species, too, we assign the 8. African Museum skeleton

No. 5624, from Groot Kruidfontein, Prince Albert Division.

Bradysaurus vanderbyli sp. nov.

Posterior border of cheek very smooth.

The type of this new species is a very large skull and lower jaw—

No. 3718 in the 8. African Museum collection—from Abrahams Kraal,

Prince Albert Division. To the species we also assign No. 6242

(S. Afr. Mus.), from Zwarts Siding, Prince Albert Division, and

No. 8941 (8S. Afr. Mus.), from Mynhardts Kraal, Beaufort West

Division. . |

2. Genus BRADYSUCHUS, nov.

Snout pointed. Cheek deep. Cusps unknown.

Bradysuchus whaitsi (Broom).

Posterior border of cheek with bosses. Large tabular bosses.

Median nasal boss. Two bosses on lower border of angular.

Parevasaurian Studies.—Part I. 83

We have not seen the type and only known specimen, which is

a skull and lower jaw from Fraserburg Road, now in the American

Museum of Natural History, but Dr. Broom has favoured us with

a sketch of the top of the skull.

SUBDIVISION Ab.

Forms with interorbital width approximately equal to the “ tabular

width.”

3. Genus NOCHELESAURUS, nov.

Snout rather pointed. Cheek deep. Teeth massive. Number of

cusps 6 to 8.

Nochelesaurus strubent (Broom).

Posterior border of cheek without bosses. Single median nasal

boss. Deep, short lower jaw with large angular boss.

The type is a lower jaw from Abrahams Kraal, now in the American

Museum of Natural History. We have assigned to the species, by

comparison of the lower jaws, a complete skull and lower jaw from

Blaauwkranz, Prince Albert Division (No. 5019 in the S. African

Museum), and another specimen (No. 5590, 8. African Museum) from

the type locality.

Nochelesaurus alexanderz sp. nov.

Snout less pointed than in strubent. Posterior border of cheek with

strong bosses. Skull heavily ornamented.

Type.—No. 6239 in 8. African Museum, from Boesmanskop, Beau-

fort West Division. This species is named in honour of Dr. A. L.

du Toit.

4. Genus DoLICHOPAREIA, nov.

Snout very pointed. Cheek deep.

Dolichoparera angusta sp. nov.

Posterior border of cheek with strong bosses. Skull heavily

ornamented. ‘Two nasal bosses close together with possibly a median

boss. Deep jaw with large angular boss.

The type is a skull and lower jaw from Boesmanskop, Beaufort

West Division (No. 6238 in S. African Museum). No. 3717 (8. African

84 Annals of the South African Museum.

Museum), from Leeuw Rivier, Beaufort West Division, is also assigned

to this species.

SUBDIVISION Ac.

Forms with interorbital width appreciably greater than “ tabular

width.”

5. Genus KOALEMASAURUS, nov.

Orbit in anterior half of skull.

Koalemasaurus acuttrostris (Broom).

Cheek broad and low. Posterior border of cheek with low bosses.

Angular boss low.

Type.—Skull and lower jaw from Hottentots River, Beaufort West

Division, in the Albany Museum, Grahamstown.

6. Genus BRACHYPAREIA, nov.

Orbit not in anterior half of skull. Cheek broad and low.

Brachypareva rogersi (Broom).

Cheek‘rather smooth. Snout rounded.

The type consists of part of a postcranial skeleton without skull,

from Hoedemakers Kraal, Prince Albert Division. The South African

Museum contains an almost complete skeleton (No. 5012), from

Abrahams Kraal, Prince Albert Division, in which the limb-bones

and vertebrae agree very closely with those of the type. This is

therefore taken as a neotype, and the skull-characters described

from it.

Brachypareia watsoni, sp. nov.

Cheek with fairly marked bosses and rugose surface. Snout

pointed.

Type.—No. 6240 (S. African Museum), from Abrahams Kraal,

Prince Albert Division.

7. Genus PLATYOROPHA, nov.

Orbit not in anterior half of skull. Cheek narrow and deep.

Parevasaurian Studies.—Part I. 85

Platyoropha broom, sp. nov.

Posterior border of cheek thin, with bosses in lower half only.

Snout very broad and rounded.

Type.—Skull No. 5002 (S. African Museum), from Vogelfontein,

Prince Albert Division.

DIVISION B.

Forms having teeth with 9 cusps arranged irregularly around border

of crown.

SUBDIVISION Ba.

Forms with interorbital width approximately equal to the “ tabular

width.”

8. Genus EMBRITHOSAURUS, Watson.

Snout rather pointed. Cheek deep. Teeth with 3 anterior and

3 posterior cusps.

Embrithosaurus schwarzi Watson.

Posterior border of cheek with pronounced bosses. Two nasal

bosses.

Type.—Complete skeleton in 8S. African Museum, from Hoogeveld,

Lot A, Prince Albert Division.

DIVISION C.

Forms having teeth with 9 or more cusps arranged regularly around

edge of crown.

SUBDIVISION Ca.

Top of skull narrow.

9. Genus ANTHODON, Owen.

Interorbital width small. Skull small. Cheek deep.

Anthodon serrarius Owen.

Surface of skull smooth.

The type specimen is an imperfect skull and anterior vertebrae

(Brit. Mus., No. 47337), from Stylkranz, Graaff Reinet Division.

86 Annals of the South African Museum.

The 8. African Museum possesses a weathered skull (No. 4020), from

Dalham, Graaff Reinet Division, from which full details of the palate

and basicranium can be obtained. The postcranial skeleton is hitherto

unknown.

SUBDIVISION Cb.

Top of skull broad.

10. Genus PAREIASAURUS, Owen.

Skull pointed. Cheeks very deep. Ihum strongly inclined forwards.

Ischium short.

Parevasaurus serridens Owen.

Crowns of teeth quadrangular in lateral view and with pronounced

lingual cingulum.

The type consists of the cast of a skull, a fragmentary lower jaw,

pelvis, scapula, dorsal vertebrae, and scutes (Brit. Mus. No. R4063),

from Blinkwater, Fort Beaufort Division. No other specimen is

known, but it is highly probable that Paresasaurus russouwi Seeley

is synonymous with serridens. The type of russowwz is a right upper

jaw and dentary with angular boss, from Klipfontein, Fraserburg

Division (Brit. Mus. No. R1996). In its dimensions and the nature

of its teeth it agrees closely with serridens.

11. Genus ParElasucHuS Broom and Haughton.

Snout rounded. Cheek deep and rugose. Ilium as in Parevasaurus.

Ischium long.

Parevasuchus perinqueyt Broom and Haughton.

Cheek very rugose with massive boss at genal angle. Massive

shallow angular boss.

The type is a complete skeleton in the S. African Museum, from

Dunedin, Beaufort West Division.

Parevasuchus nasicornis sp. nov.

Cheek less rugose than in peringueyi, with genal angle pointing

slightly backwards. Nasal bosses very prominent ; other bosses on

top of skull reduced. Palate shorter than in peringueyi. Horn-like

angular boss.

Pareiasaurian Studies.—Part I. 87

The type is a crushed but almost complete skeleton in the S. African

Museum (No. 3016), from Graaff Reinet.

SUBDIVISION Cec.

Skull unknown.

12. Genus PRropappus, Seeley.

Tlium not much inclined forwards. Ischium short.

Propappus omocratus Seeley.

The type is a partial skeleton, without skull, in the British Museum

(R4064), from Hast Brak River, Fort Beaufort Division. There is in

the 8. African Museum the specimen described under this name by

Broom, from Welgevonden, Graaff Reinet Division. In the dimension

and shape of the limb-bones and vertebrae the two specimens agree

closely enough to suggest that Broom was correct in his assignation

of the Welgevonden specimen. If so, then omocratus is peculiar in

that it possesses two horn-like knobs on the angular, of which the

posterior is the larger, thereby agreeing with Bradysuchus whaitst only

among known forms.

Propappus parvus Haughton.

This species is based upon a pelvis, some vertebrae, and some

dermal ossicles, and is mainly characterised by the short upstanding

ilium and the very large ischial tuberosity.

88 Annals of the South African Museum.

Part Il.—WNotes on some Parevasaurian Brain-Cases—By 8. H.

Havucuton, B.A., D.Sc., Hon. Curator of Palaeontological

Collections.

(With Text-figures 21-26.)

THE fullest account of the Pareiasaurian brain-case, as far as the

South African members are concerned, has been given by Watson, who

described the condition in a skull from Hottentots River which must

probably be assigned to the genus Bradysaurus. In working through

the collection in the South African Museum the present author has

been able to study several well-preserved specimens, most of which

have been almost wholly cleared of their very hard matrix by Mr. L. D.

Boonstra. As some of them show features different from those dis-

played by Watson’s specimens, the following notes will probably be

found to contain some points of interest.

Parevasuchus sp. indet.

The first skull to be examined was the weathered top and basi-

cranium of an unidentified species of Pareiaschus from Dunedin,

Beaufort West (8. Afr. Mus. No. 2367). In addition to displaying the

brain-case and quadrate regions, this skull shows the sutures between

the bones of the postorbital region, seen from within.

Basioccipital.—The condylar portion of the bone is weathered away.

The bone forms the floor of the foramen magnum. Laterally it supports

the exoccipital ; between the two no sutures can be definitely traced.

Exoccipital.—The exoccipital forms the lateral wall of the foramen

magnum. It consists of a fairly thick upper plate which contracts

ventrally to form a pillar, oval in section, resting on the basioccipital.

The dorsal plate is overlapped laterally by the paroccipital and meets

the supraoccipital medially. The pillar forms the hinder inner border

of the large jugular canal. Just anterior to the pillar the upper

surface of the basioccipital is pierced by a foramen, which is the opening

to a short canal whose outer end opens into the inner wall of the jugular

canal. This must be the canal for the passage of the XIIth nerve.

On the inner wall of the brain-case there is a deep groove between the

exoccipital and supraoccipital.

The outer wall of the large jugular canal is formed by a downwardly

directed pillar of the paroccipital. The inner end of this bone rests

on the exoccipital. The outer process is massive and forms the floor

Pareiasaurian Studies.—Part IT. 89

of the post-temporal vacuity. The anterior face is overlapped by the

pro-otic ; and there is a short articulation with the supraoccipital.

Below the pro-otic the front face of the paroccipital is notched, and

forms the posterior border of the fenestra ovalis, which is smaller than

the foramen jugulare.

The large foramen jugulare lies below the level of the foramen

magnum. The floor of the canal leading from it runs forwards and

inwards and then rises suddenly to the general level of the floor of the

Fic. 21.—Occipital view of skull of Pareiasuchus sp.

(S.A.M. Cat. No. 2367).

brain. The inner opening for the XIIth nerve is on the general level

of the floor of the brain and is continued forwards as a slight channel

on the surface of the basioccipital. Between the inner ends of the

jugular canals the floor of the brain carries a short, broad, shallow,

median ridge. In advance of this the floor is slightly concave, with

two narrow longitudinal ridges separating a median from two lateral

grooves. The ear cavity consists of a ventral pit bounded anteriorly

by the paroccipital pillar, lying slightly more lateral than the foramen

jugulare, and an anterior excavation in the hinder face of the pro-otic.

The notch in the anterior border of the paroccipital is presumably the

exit of the VIIIth nerve. The vestibule for the ear is apparently only

90 Annals of the South African Museum.

partly separated from the brain by bone, and not wholly as in Watson’s

specimen of Bradysaurus.

The course of the VIIth nerve is not quite certain, but it probably

passed through the pro-otic and made its exit slightly in advance of

the fenestra ovalis. The front of the pro-otic is smooth, and has a

well-defined notch at the bottom for the exit of the Vth and VIth

nerves. The inner face of the pro-otic is deeply grooved by a hori-

zontal channel which passes forwards and outwards.

ee aw, Le,

Fig. 22.—Palatal view of skull of Pareiasuchus

(S.A.M. Cat. No. 2367).

The ventral end of the pro-otic rests on the basisphenoid, and

continues inwards to form part of the floor of the brain-case. Between

this articulation and the basipterygoid process the side wall of the

basisphenoid is deeply hollowed out, and on the floor of the pit so

formed there is a large, oval, external carotid foramen. This com-

municates by means of a forwardly and inwardly directed canal with

the single, median, internal carotid foramen which lies in a groove on

the floor of the pituitary fossa. This latter is overhung by the bony

floor of the brain-case, the dorsum sellae, and is widely open on either

side. The groove containing the internal carotid foramen is flanked

Parevasaurian Studies.—Part IT. 9]

by two prominent bony ridges. Anterior to the pituitary fossa there

is a single median wall of bone which splits superiorly, and is carried

forwards and upwards to form the bony wall of the anterior part of

the brain. Just in advance of the pineal foramen the brain is roofed

by the parietals ; in front of the parieto-frontal suture the “ sphen-

ethmoid ” contracts and becomes continuous over the constricted part

of the brain. In front of this constriction the bone forms the floor

and side-walls only, leaving the frontals to form the roof of the brain.

The external opening for the IInd-IVth nerves lies fairly low down,

just in advance of the level of the pineal foramen.

There is no contact between the pro-otic and the skull-roof. The

pro-otic has an ascending process which receives the outer spur of the

on ey. "

a ~ SI j (Re

5 ; \ \ | bier on qe J Ss ©) Per dag,

PS. Pit Foss. Ex.Ca.Fo.

Fic. 23.—Lateral view of basicranium and brain-case of Pareiasuchus sp.

(S.A.M. Cat. No. 2367).

massive supraoccipital. __In the specimen under consideration, owing

to slight crushing, this process appears to be a separate bone ; but its

true nature can be seen by an examination of the type of Parevasuchus

peringueyt, where the ascending process is continuous with the main

body of the pro-otic. Between it and the supraoccipital there is a

distinct longitudinal canal whose floor is formed by the inwardly

directed flange of the pro-otic.

The relations between the quadrate and pterygoid seem to differ

somewhat from those of the Pareiasaurian figured by Broom as

Pareiasaurus bombidens. The anterior process of the quadrate

becomes a thin plate of bone passing in front of the posterior ramus of

the pterygoid, which thins rapidly posteriorly and passes upwards

nearly to the roof of the skull medially to the ascending plate of the

quadrate. In longitudinal section this quadrate ramus of the pterygoid

is somewhat triangular, and its upper surface is hollowed out into a

92 Annals of the South African Museum.

broad groove which is deepest at the end nearest the basipterygoid

process. This broad groove, which does not pass on to the quadrate,

was presumably for the reception of the anterior face of a massive

stapes. A similar groove has been figured by Watson in the Car-

boniferous Embolomerous Stegocephalian Lozomma allmanr; but

there the groove passed over on to the quadrate. At its outer end the

groove in Pareiasuchus has a vertical anterior wall; at the inner,

deeper end it has in addition a horizontal floor. Its outer end definitely

faces towards the auditory region.

The late Professor Sushkin figured (Pal. Zeitschr., vin, 1927,

p. 295, fig. 25) a similar recess on the posterior face of the “ post-

quadrate plate of the pterygoid’ in Pareiasaurus Karpinsku, which,

apart from its forward-sloping quadrate, exhibits many points of

similarity to Pareiasuchus. The recess in both forms does not extend

on tothe quadrate. Sushkin states that in the Pareiasauria from North

Dwina, “‘ wherever the posterior face of the quadrate is well preserved

there exists a tubercle lying close to the suture of the pterygoid, at

one-third of the height of the quadrate from its condylar end; in

larger and more aged specimens this tubercle is more prominent. I

think that it presents the point of attachment of the distal end of the

stapes.” This tubercle lies lateral to the recess on the pterygoid.

Watson has suggested that the stapes of the Pareiasaurians was a

slender rod ; but the presence of this large recess and consideration

of the bone in Captorhinus agutz, as figured by Sushkin (fig. 18), suggest

that it was a rather massive bone, swollen in the middle and at its

proximal end.

Anthodon serrarius.

The skull in the South African Museum collection assigned to this

species shows some interesting features of the quadrate region. It

lacks the whole of the posterior part of the brain-case, but differs from

Parevasuchus in some striking characteristics of the remaining part.

Between the basisphenoidal tubera, and to a greater extent between —

the basipterygoid processes, the basisphenoid is deeply vaulted; and

the parasphenoidal rostrum hes high above the pterygoid plate, so

that there is a deep, large, interpterygoid vacuity which is continued

forwards on the palate as an elongate triangular slit.

The basisphenoid-pterygoid suture is very different from that of

Pareiasuchus. In the latter it is practically longitudinal ; in Anthodon

it is almost wholly transverse in ventral view, and vertical. The

posterior ramus of the pterygoid passes directly outwards from the

iS)

Parevasaurian Studies.—Part II. 9

outer corner of the suture; it lies behind the quadrate, which sends

forward a long vertical plate to lie in front of the pterygoid. The

quadrate ramus of the pterygoid

is also a high, vertical plate,

whose posterior surface is sud-

denly bent backwards ventrally

so as to form a _ pronounced

groove whose deepest part is

medial and just in advance of

the basisphenoid-pterygoid sut-

ure. As in Pareiasuchus, this

groove does not extend on to the

quadrate.

The lateral wall of the basi-

sphenoid is rather high and con- Fic. 24.—Basisphenoid, pterygoid, and

. . quadrate of a skull of Anthodon

cave, and lies outside the ante- es ee

rior fragment of pro-otic that is

preserved. At the bottom of the concavity the bone is pierced by

the external carotid foramen.

Anteriorly the upper part of the pro-otic is a thin plate of bone,

concave medially, its ventral portion being bent inwards so as, pre-

sumably, to meet its neighbour in the middle line on the floor of the

brain-case.

Embrithosaurus schwarzt.

The basicranial region in the type skull is rather crushed and the

limits of the various bones not easy to determine ; nevertheless it

shows some interesting features.

The condyle differs from that of other described Pareiasaurs in the

more saucer-like character of its articular surface, the prominent

rounding of the outer part of this surface which is seen in other forms

being here reduced to an insignificant rim or edge. The distance

between this edge and the basisphenoidal tubera is much shorter here

than in Watson’s specimen of “‘ Pareiasaurus ” from Hottentots River.

The pro-otic and paroccipital are indistinguishably fused. The

fenestra ovalis is large, and its hinder border is deeply notched

superiorly for the passage of the VIIIth nerve. The whole brain-case

is wide and low. Owing to the crushed condition of the specimen it is

uncertain whether there is a separate external foramen for nerve XII ;

but there are indications on the right side that this was separate from

the foramen jugulare.

94 Annals of the South African Museum.

Bradysaurus bain.

The description of the condition in this species is based upon an

examination of a finely preserved skull from Abrahams Kraal, now

in the Geol.-palaontolog. Institut at Tibingen; for permission to

study this I am greatly indebted to Drs. F. von Huene and H. Hennig.

fi) OY IN

fil BN ANZ |

Fic. 25.—Occipital view of skull of Bradysaurus baini in Coll.,

Geol. Institut, Tubingen.

The condyle differs from that of Hmbrithosaurus, the articular

surface being gently rounded and having no distinct marginal rim.

As in Watson’s specimen, the opening for the XIIth nerve is a small

foramen lying behind the much larger opening for nerves IX—XI, and

separated from it by a slight ridge of bone. The foramen ovalis is

much smaller relatively than in Paresasuchus and is nearer to the

foramen jugulare.

There is a very pronounced depression in the posterior face of the

quadrate ramus of the pterygoid, near the outer end of which occurs

Pareiasaurian Studies.—Part II. 95

the pterygoid-quadrate suture. Near this suture, and on the quadrate,

is the small boss of bone described by Sushkin as a probable point of

attachment of the stapes. There isa large quadrate foramen, elongate

in a vertical direction, seen on both the posterior and anterior surfaces

of the bone near the quadratojugal suture. In posterior view the

upper part of the quadrate articulates with the squamosal along a

vertical ridge, on the mesial side of which the plane surface of the

quadrate looks inwards and backwards, and on the lateral side of which

the squamosal surface is transverse.

The paroccipital and pro-otic are indistinguishably fused. The

paroccipital process is very massive. On the outer posterior surface

of the exoccipitals, on a level with the top of the foramen magnum,

there is a shallow depression for articulation with the pro-atlas. The

exoccipitals certainly meet above the foramen magnum.

The palate differs from that of Anthodon in that the medial portions

of the pterygoids meet in a long suture to form a shallow vaulted

triangular plate, broadening posteriorly, which covers the whole of the

parasphenoid in ventral view.

Dolichopareia angusta.

The basicranium of the type (8. Afr. Mus. No. 6238) has been

cleared of matrix. The brain-case and basicranium are of the usual

Pareiasaurian type, but certain differences of detail are visible when

comparison is made with Pareiasuchus.

The articular surface of the condyle is deeply and evenly concave,

and its rim is broadly rounded. The upper lateral corners are pro-

duced in square-cornered flat plates. The basisphenoid is longer than

in Pareiasuchus, so that the inner end of the quadrate ramus of the

pterygoid les some distance in advance of the plane of the lateral

opening of the pituitary fossa. In Parevasuchus it les below and

behind the same opening. Further, in Dolichopareia the floor of the

pituitary fossa is above the level of the foramen ovalis ; in Paretasuchus

it is shghtly below. The paroccipital process has its lower border

passing strongly downwards, so that the large jugular canal looks more

downwards than backwards.

The floor of the brain-case, between the inner openings of the jugular

canals, has a strong median ridge and boss, on either side of which the

floor of the canals slopes down steeply. These canals are separated

from the vestibules by knobs of bone, which do not rise up to meet

the side-wall of the brain-case. The ear-cavity is formed by a lower

96 Annals of the South African Museum.

excavation, low down in the lateral floor of the brain-case and an

upper vault in the lateral wall, here formed of pro-otic.

The foramen magnum is slightly higher than wide; but in front

of the exoccipital pillars the brain-case widens rapidly. The upper

PR. fe At ie

< a i @ "

ROSS ]

—s 4 ie geN a

\+ s RA\

Cs LURAY

—

Fic. 26.—Lateral view of basicranium and brain-case of type skull

of Dolichopareia angusta.

edge of the pro-otic is a sharp ridge which dips slightly from the

dorsum sellae to a point above the anterior edge of the fenestra ovalis,

and then rises rather more abruptly as it swings in towards the mid-

line of the skull.

Pareiasaurian Studies.—Part ITT. 97

Part Il].—On the Parevasaurian Manus.— By Lizuwe OD.

Boonstri, M.Sc., Assistant in Palaeontology, South African

Museum.

(With Plates VI-VIII and Text-figures 27, 28.)

HISTORICAL.

In 1892 Seeley (1), in describing Bradysaurus baini, mentioned a

number of points in the structure of the manus. These conclusions

were based on an incomplete right manus, the elements of which were

partially displaced. In giving a brief summary of Seeley’s views I

base my criticisms on a study of a plaster cast made in the British

Museum under Seeley’s direction.

Seeley was of the opinion that there were two rows of bones in the

carpus. The proximal row is represented by a large ovate bone

lying immediately under the radius. This bone he rightly considered

to be the principal carpal; but erroneously regarded it as a composite

bone formed by the fusion of radiale and intermedium. He mentioned

a second ossification to the right of the main one, and expressed the

opinion that there probably were further ossifications of the proximal

tow. This second bone noted by Seeley is not one of the proximal

carpals but a centrale, as will be shown more fully in the sequel.

In the distal row he noted two bones articulating with the main

carpal, and he thought that they are probably centralia. If this were

the case there should be still another row of carpals—the distalia.

This means that there are three and not two rows of carpal elements.

In reality the two bones are distalia with which the metacarpals of

the tirst and second digits, respectively, articulate.

The metacarpals he described as short and flattened.

Seeley was very uncertain about the digits, which he rightly thought

were five in number. He had no certain evidence of more than three

phalanges in any digit, but considered it possible that one digit may

have had four. He erroneously stated that the digits with the larger

claws are situated on the ulnar side. In the cast the larger and

stronger digits are on the radial or preaxial side. Is this discrepancy

due to a slip of the pen in the description ?

In 1894 Seeley (2), from a further study of the specimen, stated

that the phalangeal formula appears to be 3, 4, 4, 3, 3.

In 1903 Broom (3) described Embrithosaurus schwarzi (or, as he

VOL. XXVIII, PART l. 7

98 Annals of the South African Museum.

ceé

then called it, Pareiasaurus serridens) and said of the manus, “a

considerable number of isolated phalanges have been found, but

without any indications as to which digits they belong to.”” Accom-

panying this paper is a restoration of the whole animal. The manus

is represented as having the phalangeal formula, 2, 3, 4, 5, 4.

In 1904 Boulenger (4) arbitrarily stated that the phalangeal

formula of the British Museum specimen of Bradysaurus baini is

2, 3, 3, 4, 3, and that he could not understand how Broom could

reconstruct a Pareiasaurian with the formula 2, 3, 4,5, 4. As will be

shown in the sequel, Boulenger was probably right, but as he gave no

reasons for this guess, I quite agree with Broom (5) where he says,

“Tf Seeley was unable to tell even to which foot the different toes

belonged . . . it is difficult to see how Boulenger obtained his result.”

In 1908 Broom (5), in answer to Boulenger, referred to his specimen

of Embrithosaurus schwarz, saying that two digits had their phalanges

in contact. These he inferred to be the second and third digits.

These have three and four phalanges respectively. This specimen is in

the South African Museum Catalogue, No. 8034. I have developed these

two digits and find them both to be composed of a metacarpal and

three phalanges (Pl. VII, fig. 2). I have shown these to Dr. Broom,

and he agrees that, now they are developed, it is clear that both have

three phalanges. Which digits they are, remains, of course, uncertain.

In 1913 Broom (6), in a paper on “The Manus and Pes of

Pareiasaurus,”’ based on a specimen which he thinks in all probability

to be Bradysaurus baint, gives the following analysis of the manus: the

proximal row consists of three elements—a large radiale, a narrow inter-

medium, and a small, nearly circular ulnare ; the middle row is repre-

sented by a small rounded second centrale, and in his sketch he

reconstructs a small first centrale ; four distalia were doubtfully present,

but Broom sketches in the fifth, which he considers lost in preservation ;

the first and second metacarpals and the first phalanx of the first

digit are mentioned ; a pisiform is sketched as articulating with the

postaxial surfaces of ulna and ulnare. This manus had its elements

disarranged during excavation, and the contacts were lost. In a

verbal communication Broom states that he is only sure of his diagnosis

of radiale, ulnare, intermedium, and second centrale.

From this review we learn that the Pareiasaurian manus is very

inadequately known. In fact, we are only moderately confident about

the presence of radiale, ulnare, intermedium, and second centrale.

Pareiasaurian Studies.—Part ITT.

MATERIAL.

During October 1928 I collected a perfect left manus, and in

subsequently going through the collection in the South African Museum

I found a considerable amount of supplementary material—enough

to justify an inquiry being instituted into the structure of the

Pareiasaurian manus.

The material on which this paper is based can be tabulated as

follows :—

S.A.M. ;

: Geographical Geological

oe Animal. Locality. Horizon. Remarks.

3533 | Bradysaurus baint | Hottentots River, | T'apinocephalus | Imperfect carpus,

Beaufort West. Zone, high. metacarpals and

phalanges incom-

plete.

4350 | Bradysaurus sp. juv.| Leeuw River, | Tapinocephalus | Radiale, Ist, 2nd,

Beaufort West. Zone, high. and 3rd digits in

articulation.

5012 | Brachypareiarogersi| Abrahams Kraal, | Tapinocephalus|Complete right

Prince Albert. Zone, low. manus.

8941 | Bradysaurus van- | Mynhardts Kraal,| T’'apinocephalus | Complete left manus.

derbyli. Beaufort West. Zone, high. |Incomplete right

manus.

5127 | Bradysaurus baint | Leeuw River, | Tapinocephalus <Proximal carpalia,

Beaufort West. Zone, high. | centrale, 3rd and

4th distalia.

5624 | Bradysaurus seeleyi | Groot Kruidfon- | T'apinocephalus | Two radialia.

tein. Zone, low.

8934 | Bradysaurus sp. Leeuwfontein, Tapinocephalus | Right carpus and dis-

Beaufort West. Zone, high. placed phalanges.

3016 | Pareiasuchus nasi- | Graaff Reinet. Endothiodon Proximal carpalia

cornis. Zone. and centrale.

8034 | Embrithosaurus Hoogeveld A, | T'apinocephalus | Displaced phalanges

schwarz. Beaufort West. Zone, mid. and metacarpals.

DESCRIPTION.

My conclusions are based chiefly on 8941—a perfect manus—and to

a lesser degree on 5012—a complete manus, which has, however, the

carpals slightly disarranged owing to the fact that the hand was

fossilised in such a position that the digits are bent towards the ventral

surface of the fore-arm. The hand is, as it were, standing on its

knuckles.

100 Annals of the South African Museum.

(2) BASIPODIUM.

The carpus is complete and has the various elements in undisturbed

articulation in 8941. In 3533, 8934, 5012, 5127, and 3016 the

proximal row is intact, and, with the exception of 5012, the centrale,

the third, and the fourth distale.

The proximal row of the basipodium consists of three bones—the

radiale, intermedium, and ulnare. Broom (6), in his sketch of 3533,

inserts a small pisiforme articulating with ulna and ulnare. My study

of this same specimen, corroborated by a study of another specimen

of Bradysaurus baini, 5127, does not support this supposition. It

does not appear as if the strong articulation between ulna and ulnare

needs the further support of a pisiforme (PI. VII, fig.1). In 8941 no

pisiforme was noted during the excavation, although particular care

was taken to certify this fact, because the distal end of the ulna was

decomposed, and I realised that it would crumble away on exhumation

and make determination at a later date impossible. In 3016, an

armoured “ Pareiasuchus,’ the articulation of ulna with ulnare is

not very firm, and there is a space on the postaxial side of the junction

of these two bones. In this case the possibility of the presence of a

pisiforme should not be excluded. In 5012 the disarrangement of

the carpals already referred to makes it difficult to ascertain whether

a pisiforme could have been present. In any case, no pisiforme is

preserved.

Radiale.-—The radiale is the main carpal bone. Its proximal

surface has a large elliptical surface for articulation with the radius,

which bears the greater part of the weight resting on the fore limbs.

On its dorsal surface there is a strong transverse ridge at right angles

to the long axis of the limb. This ridge is very clearly marked in

8941 and 5127. On its palmar surface there is a similar but weaker

ridge and situated slightly more distally. These ridges limit the

amount of movement of the radius-radiale joint in the vertical plane.

Distally the radiale has two articulatory surfaces for the first and

second distalia. This fact 1s very clearly shown in 8941 and also in

the cast of the British Museum specimen of Bradysaurus bavnt.

On its medial or postaxial side there is a facet for the articulation of

the centrale. In the two specimens of Bradysaurus baini the inter-

medium stretches beyond the distal ends of the radius and ulna, and

articulates with the proximal postaxial corner of the radiale.

The size of the radiale varies considerably in the different members

of the Pareiasauridae.

Fie. 27.—Dorsal view of reconstruction of left manus of Bradysaurus vanderbyli

(S.A. Mus. Cat. No. 8941). x 4.

R=radius. 1-4=distalia.

U=ulna. M=metacarpals.

r=radiale. (1)-(3) =phalanges.

u=ulnare. J-~V =Number of digit,

i=intermedium.

=centrale.

Fic. 28.—Dorsal view of reconstruction of right manus of Brachypareia

rogersi (S.A. Mus. Cat. No. 5012). X 4.

(Lettering as for fig. 27.)

102 Annals of the South African Museum.

Measurements.

Brachypareia| Bradysaurus | Bradysaurus | Pareiasuchus

rogersi. baini. vanderbyli. | nasicornis.

mm. mm. mm. mm.

(a) Proximo-distal 62 62-68 64 58

(6) Pre-postaxial . 73 110-125 85 70

Intermedium.—The intermedium is a, medium-sized bone laterally

compressed but elongated proximo-distally and dorso-ventrally. It

is wedged in between the radius, ulna, ulnare, radiale, and centrale.

Its palmar surface is much greater than its dorsal in Bradysaurus

baint (5127). It appears to extend farther distally in the two

specimens of Bradysaurus baini and also in Pareiasuchus nasicornis

(3016) than in Bradysaurus vanderbylv.

Measurements.

Bradysaurus Bradysaurus Pareiasuchus

baint. vanderbyli. nasicornis.

mm. mm. mm.

(a) Proximo-distal . 70-74 AT 50

(6) Pre-postaxial : 34-40 LT. 16

Ulnare.—The ulnare is a moderately large, nearly circular bone.

Preaxially the intermedium is strongly articulated to it. On its

preaxial side, z.e. the side articulating with the intermedium and

centrale, there is in 8941 (Pl. VI), a very distinct notch. This notch is

not at all clearly shown in all the other specimens. These, however,

have a space between the three bones—ulnare, intermedium, and

centrale—very similar to that drawn by Williston (7) for Ophiacodon.

This space and the notch in 8941 is the perforating foramen noticed

in many reptiles. Distally the ulnare articulates with the fourth

distal carpal. In his sketch of 3533 Broom (8) inserts a fifth distale

articulating with the ulnare. In 5127, also a specimen of Bradysaurus

baint (Pl. VII, fig. 2), there is no possibility of a fifth distale articulating

with the ulnare in addition to the fourth, since the latter occupies the

Parevasaurian Studies.—Part IIT. 103

whole distal surface of the ulnare. In 8941, however, there is, as

preserved, a space (Pl. VI) between the disto-postaxial surface of the

ulnare and the metacarpal of the fifth digit. I am quite sure that

there was no ossified bone in this space, neither did the matrix show

any indication of a chondrified element. In any case, there is no

reason why the fifth digit should not articulate directly with the

ulnare as, for instance, in Limnoscelis.

Measurements.

Brachypareia| Bradysaurus | Bradysaurus | Pareiasuchus

rogerst. baini. vanderbyli. | nasicornis.

mm. mm. mm. mm.

(a) Proximo-distal 39 64-70 64 45

(>) Pre-postaxial . 30 52-82 60 33

The Median Row.—The median row of carpal bones is represented

by a single bone—the second centrale. It has already been remarked

that Broom (6) sketched two centralia for 3533, but, as he himself

acknowledges verbally, he is only sure of the proximal row—the

contacts of the other elements having been lost in excavation.

In Bradysaurus baini the centrale is wedged in between radiale,

ulnare, intermedium, and the third and fourth distalia. In these

two specimens the palmar surface of the centrale is much greater than

on the dorsal aspect.

In 3016 the centrale is proportionally a large bone, whereas it is

quite moderate in size in 5012. This is rather a great variation in

animals of about the same size.

In Bradysaurus vanderbyli (8941) the centrale is a laterally com-

pressed bone very similar in shape to the intermedium. Its relation

to the other bones is similar to the condition in Bradysaurus baina. It

differs, however, from the latter in that there is not much difference

in the size of the dorsal and palmar surfaces.

[ TABLE.

104 Annals of the South African Museum.

Measurements.

Brachypareia| Bradysaurus | Bradysaurus | Pareiasuchus

rogerst. baini. vanderbyli. | nasicornis.

mn. mm. mm. mm.

(a) Proximo-distal 32 47-50 38 34

(0) Pre-postaxial 26 37-43 25 48

The Distal Row.—I have already referred to the uncertainty

attached to Broom’s (6) identification of Bradysaurus baini. All I

can say of a study of the type specimen (3533) is that there were at

least four well-developed distalia present. Which distalia are present

cannot be determined. In any case there is no degenerated distale

as is the case in 8941.

In 8941 four distalia are present (Pl. VI and text-fig. 27). There is

no doubt about the determination. The first and second distalia are

in articulation with the metacarpals of the first and second digits

respectively. Proximally they articulate with the radiale. They

are well-developed ellipsoid bones slightly compressed proximo-

distally. The third distale is a small spherical rudimentary bone

articulating with the preaxial half of the proximal articulatory

surface of the metacarpal of the third digit. Proximally it articulates

with the second centrale. It appears to be in process of reduction.

The fourth distale is a strong bone roughly quadrilateral in shape.

Proximally it articulates with the second centrale and the ulnare. It

articulates distally with the metacarpals of the third and fourth digits.

The fourth distale is generally the strongest of the distalia, and it is

quite evident that it is here assuming, in conjunction with its original

function, the role of the disappearing third distale. Attention has

been drawn to the fact that there is, as preserved, a space in 8941

between the metacarpal of the fifth digit and the ulnare, and reasons

have already been given why the presence of a fifth distale should be

considered improbable. In the specimen of Bradysaurus baini (5127)

the fourth distale is a very strong bone occupying the whole of the

distal surface of the ulnare. There can be no question of the presence

of a fifth distale. In this specimen there is also a small degenerate

third distale. As there is no degenerate distale in 3533, Broom’s

reconstruction becomes still more doubtful.

In both Bradysaurus baina and vanderbyli the elimination of the

Pareiasaurian Studies.—Part IIT. 105

fifth distale must be accepted. This tendency towards elimination is

shown by the reduction of the third distale. In Brachypareia rogersi

(5012) there are only four distalia. They are, however, disarranged,

and it is only by analogy that we may consider them to be the first

four.

Measurements.

Bradysaurus baint. Bradysaurus vanderbyli.

mm. | mm. mm. | mm ; mm. |; mm. | mm. | mm.

(a) Proximo-distal .| 2 2 208 |) A328 P21 | 32

(6) Pre-postaxial .| 2 Q B05) 2 S60 63416). \).35

(0) METAPODIUM.

There are five metacarpal bones—one for each digit. These bones

are short proximo-distally and flattened dorso-ventrally. Their

articulatory surfaces are expanded while the shafts are constricted.

Owing to the dorso-ventral flattening, the pre-postaxial constriction

is very marked. With the exception of the first metacarpal in 8941,

the metacarpals agree more or less in shape. They decrease in

massiveness in pre-postaxial direction ; but the fourth and fifth are

longer bones. The fourth is absolutely, while the fifth is proportion-

ally, longer than the first three metacarpals.

First Metacarpal.—In specimen 8941 the first metacarpal is a broad

five-sided bone. On its proximal postaxial surface there is a concave

articulatory surface for the first distale. This articulation is a ball-

and-socket one allowing movement in all planes. On the proximal

preaxial surface there is a convex articulatory surface. Receiving

this there is a groove on the distal preaxial surface of the radiale.

When this joint functions it allows of lateral movement on the part

of the first digit, which is, in general, more mobile than the other digits.

Whence Williston’s (7) generalisation. The first and fifth metacarpal

are more freely mobile on the wrist than the other three. Distally

this metacarpal has a bipartite articulation with the first phalanx.

The preaxial portion of the articulation is the stronger.

Second, third, fourth, and fifth Metacarpals.—Proximally the second

metacarpal articulates wholly with the second distale.

106 Annals of the South African Museum.

The third metacarpal has two articulatory surfaces on its proximal

end. The main articulation is with the rudimentary third distale.

In addition it has a small articulatory surface on its postaxial margin

with the fourth distale. |

The fourth metacarpal has a longer and more slender shaft. Proxi-

mally it also has two articulatory surfaces. The chief one lies

preaxially and articulates with the strong fourth distale, which lies

between the third and fourth metacarpal. On its postaxial margin ~

there lies a small facet.

The fifth metacarpal is of the same type as the fourth, but is smaller

and still more slender. It has a facet on its proximal preaxial margin.

This facet and the postaxial facet of the fourth metacarpal may

together have articulated with a fifth distale. If this were the case,

however, there should be a single articulatory surface on the distal

end of the ulnare. There is certainly no well-marked single facet on

the distal end of the ulnare, but there is an indication of two.

Although I do not absolutely exclude the possibility of a fifth distale,

I prefer to interpret the condition in 8941 otherwise, viz. that the

preaxial facet of the fourth metacarpal, as well as the whole facet of

the fifth, articulated directly with the ulnare. A similar condition is

found in Limnoscelis (8) and Captorhinus (9), but although the fifth

digit would in this case be very mobile, it certainly was not so mobile

as the first.

Measurements in Muillimetres.

Bradysaurus vanderbyli. Brachypareia rogerst.

(a) Proximo-distal | 27-41} 36 | 40 | 46 | 40 | 2 | 30 | 32 | 32 | 23

(6) Pre-postaxial .| 52 | 45 | 47 | 42/30] ? | 32) 31 | 30] 21

In the cast of Bradysaurus baini the difference in the size of the

elements of the five digits is not very great. In the specimen of

Bradysaurus vanderbyli (8941), and in that of Brachypareia (5012),

there is a marked decrease of size in pre-postaxial direction. In fact,

the fifth digit is so small that I am of the opinion that it is fast be-

coming rudimentary in these two forms. It is quite possible that

Pareiasaurian Studies.—Part ITI. 107

B. bain would still have a well-developed fifth digit. It is interesting

to note that with decrease of size there is a corresponding diminution

in the size of the fifth digit. We eagerly await the discovery of the

manus of the small and light Anthodon.

For a general account of the phalanges and terminal claw Seeley’s (1)

work should be consulted.

First Digit.—This digit is composed of a single phalanx and a

terminal claw.

The phalanx is large, short, wide, and dorso-ventrally flattened.

Its proximal end is divided into two separate articulatory surfaces

corresponding to the bipartite nature of the distal end of the

metacarpal.

The terminal claw is the largest of all the digits. For description

and illustration see Seeley (1).

Measurements.

Bradysaurus vanderbyli. | Brachypareia rogers.

Phalanx. Claw. Phalanx. Claw.

mm. mm. mm. | mm.

(a) Proximo-distal 29 70 17 60-70

(0) Pre-postaxial . 55 39 139 40-44

Second Digit.—There are three elements in this digit.

The phalanges have single articulatory surfaces. As in all the

phalanges these two are more constricted medially on their palmar

sides than dorsally.

The claw is only slightly smaller than that of the first digit.

Measurements.

Bradysaurus vanderbyli. Brachypareia rogers.

Ist 2nd Ist 2nd

Phalanx.) Phalanx.| Claw. | Phalanx.| Phalanx.| Claw.

mm, mm. mm. mm. mm. mm.

(a) Proximo-distal 20 30 65 15-17) T2214") 63-718

(6) Pre-postaxial . 45 43 46 40-34 | 38-34 | 39-40

| i

i !

itl

108 Annals of the South African Museum.

Third Digit.—In the third digit there are also three elements.

The first and second phalanges are similar to those of the second

digit, but the claw is markedly smaller than that of the previous

digits.

Measurements.

Bradysaurus vanderbyli. Brachypareia rogersi.

Ist 2nd 1st 2nd

Phalanx.| Phalanx.| Claw. | Phalanx.) Phalanx.| Claw.

mm, mm. mm, mm. mm. mm.

(a) Proximo-distal 23 25 54 | 17-15 | 17-13 | 55-68

(6) Pre-postaxial . 43 27 40 | 36-31 | 35-33 | 36-36

Fourth Digit.—In 8941, 5012, and in 4350 this digit is composed

of three bones.

The terminal claw is again slightly smaller than the previous ones.

Measurements.

Bradysaurus vanderbyli. Brachypareia rogersi.

Ist 2nd Ist 2nd

Phalanx.| Phalanx.|} Claw. | Phalanx.| Phalanx.| Claw.

mm. mm. mm. mm. mm. mm.

(a) Proximo-distal 20 22 s4 2-15 | @-14 | 2-49

(6) Pre-postaxial .| 37 32 34 23 | %?18 | ?31

Fifth Digit.—What the fifth digit of B. baini looked like we do not

know. According to Seeley (1) it was well developed and therefore

functional. It is quite possible that in the heavier Bradysaurus a

wider spread of foot was necessary to support the massive weight—a

weight the heavier because of its bad balance ; whereas in the lighter

Bradysaurus vanderbylt and Brachypareia rogersi the fifth digit would

be of less moment. Is there a still further decrease in the small

Anthodon ?

Pareiasaurian Studies.—Part ITT. 109

In addition to the difference in size, difference in habitat could, of

course, also play a réle. Of this, however, we know practically

nothing.

In the specimen of Bradysaurus vanderbyli (8941) there are two

elements. ‘The first phalanx is slender and light, but is proportionally

longer than that of any other digit. The terminal claw is much more

flattened than that of the other digits. It is also less pointed.

In the specimen of Brachypareia rogersi (5012) there is only one

element—the terminal claw articulating directly with the metacarpal.

It would appear that the condition represented in these two speci-

mens are two steps in a series of reductions.

Measurements.

Bradysaurus vanderbyli. Brachypareia rogersi.

Ist Phalanx. Claw. Ist Phalanx. Claw.

mm, mm. mm.

(a) Proximo-distal 20 33 absent 16

(b) Pre-postaxial . 30 25) % 19

COMPARISONS.

Embryological research (11) shows that the tetrapod limb probably

passed through a stage with seven digits. Palaeontologically the

view (12) has been expressed that Hryops indicates a condition where

more than five digits were developed. If we take Hryops as showing

a primitive condition, the primitive carpus had many elements, viz.

carpal formula 4, 1, 3, 5.

In another rhachitomous temnospondyl—Trematops—the carpal

formula is 4, 1, 1 (2), 5, and according to Williston (13) the phalangeal

formula is 2, 3, 3 (4), 4, 2 (3).

Nearly all the other Amphibia, both extant and extinct, have four

digits in the manus and are not considered primitive.

It seems certain that the five-toed reptiles were not evolved from

four-toed Rhachitomi, but that both originated from some five-toed

Embolomeri as yet unknown.

110 Annals of the South African Museum.

Among that heterogeneous assemblage, grouped together as Cotylo-

sauria, there is great divergence in the manus.

The primitive reptilian manus is taken to have the carpal formula

4,1, 1, 5 and the phalangeal formula 2, 3, 4, 5,3 (4). This conclusion

is based on ontogenetical and palaeontological research.

Conodectes (=Seymouria) stands on the border line between

Amphibia and Reptilia, but unfortunately the manus is unknown.

Williston (14) reconstructed the manus with the carpal formula

4,1, 1, 5 and phalangeal formula 2, 3, 4,5, 3. He is sure that this

must be correct for Seymourza, because he has examined the condition

in Limnoscelis !

According to Amalitzky, some of the North Russian fossils belong

to the family Seymouridae. In Koilassia he gives the uncertain

phalangeal formula ?, 1, 2, 2 (3), 1 (2) for the hind limb. He adds

that there are not more than three elements in both the third and

fourth digit.

Among the Captorhinidae Williston (9) states that Parzotichus

(=Captorhinus ?) has the carpal formula 3, 1, 1, 4 and the phalangeal

formula probably 2, 3, 3, 3 (4), 2. In the carpus it is the first distale

which is taken to be absent. According to Case (15), Labidosaurus

has the carpal formula 3,1,1,4. The first and second distalia are taken

as fused. Notwithstanding the fact that Williston (16) reconstructed

the phalangeal formula as 2, 3, 4, 5, 3, 1t remains unknown.

Among the Diadectidae Diasparactus (17) has the carpal formula

3 (4), 0, 1, 3 (4). The fifth distale is taken to be absent. The

proximal row consists of large well-developed bones, whereas the

distalia are very small and were probably largely cartilaginous in life.

The phalangeal formula is considered to be 2, 3, 4,5, 33; but they are

not very well preserved in the specimen described.

Among the Limnoscelidae, Williston (8) states that the carpal

formula is 4, 0, 1, 2. The two ossified distalia are the third and

fourth. The first and second are considered to have been cartilaginous

as well as the second centrale, while the fifth distale is absent alto-

gether. It is practically certain that the phalangeal formula is the

primitive reptilian one, viz. 2, 3, 4, 5, 3.

Among the Triassic Procolophonidae, von Huene (18) states that

they all have the carpal formula 3, 1,0-1 ; 4-5. The Procolophonidae

is the only Cotylosaurian family in which the second centrale is absent.

The same author states that Telerpeton and Kozloskiosaurus have the

primitive reptilian phalangeal formula—2, 3, 4, 5, 3. This is also

Boulenger’s (4) opinion on Telerpeton and Broom’s (19) on Procolophon.

Pareiasaurian Studies.—Part ITT. Ta

Among the Pareiasauridae, Bradysaurus vanderbylt and Brachy-

pareia rogerst have the carpal formula 3, 0,1, 4. The third distal is

degenerate and the fifth missing. In Bradysaurus baini the carpal

formula is probably 3, 0, 1, 4, or 5. The phalangeal formula is

2, 3, 3, 3, 2 in Bradysaurus vanderbyli and 2, 3, 3, 3, 1 in Brachypareia

rogerst. This formula is uncertain in Bradysaurus bainr. It certainly

is not 2, 3, 4, 5, 3. I consider the view held by Boulenger (4) and

others, that the formula is 2, 3, 3, 4, 3, more probable, and I even

consider it very probable that the fourth digit also had only three

phalanges.

In this brief comparative review it is clear that the Cotylosaurian

manus is very imperfectly known, and one wonders how many of the

forms really possess the primitive phalangeal formula that is usually

assigned to them. In any case, until we are better informed, we may

state that of the Cotylosauria the Pareiasauridae is the only family

members of which have been shown to have digits composed of less

elements than the digits of the currently accepted primitive manus.

ADDENDUM.

Since the above was written a manus from Abrahams Kraal has

been developed (S. Afr. Mus. Cat. No. 5015). The postaxial side

has been damaged by weathering—so much so that the fifth digit is

missing, only a half of the ulnare is preserved, and the intermedium

is also slightly weathered. The second and third distal carpals are

not preserved. The fourth distal carpal is as well developed as in the

South African Museum Catalogue, No. 8941, and supports the third

and fourth metacarpals. The phalanges of the fourth digit have

suffered some injury during life. Osteological recovery has caused

the first two phalanges to coalesce. It is, however, clear that this

digit was composed of three elements.

For the rest, the specimen agrees with the conclusions based on the

previously considered material.

In the plate (Pl. VIII) a photograph is given of the palmar surface.

The specimen is probably one of Bradysaurus vanderbylt.

112 Annals of the South African Museum.

HXPLANATION OF PLATES.

Puhate VI.

Dorsal view of left manus of S.A. Mus. Cat. No. 8941, Bradysaurus vanderbyli.

x §. The ventral surface was first developed out of the block of matrix.

This side was then embedded in plaster of Paris. The dorsal side was then

developed. In the photograph the bones are therefore shown in the exact

position in which they were found.

Pratr VIE.

Via. 1. Ventral view of right carpus of Bradysaurus baini, S.A. Mus. Cat. No. 5127.

X zs.

», 2. (a) Palmar view of digit of Hmbrithosaurus schwarzi. xX 3.

(6) Dorsal view ot another digit of Embrithosaurus schwarzi. S.A. Mus,

Cat. No. 8034. x: 3.

Puate VIII.

Palmar view of the incomplete manus of S.A. Mus. Cat. No. 5015, probably

Bradysaurus vanderbyli. X 4.

WORKS CITED.

(1) SpeLeyY.—Phil. Trans. Roy. Soc., vol. 183 B, 1892.

(2) SrELEyY.—Phil. Trans. Roy. Soc., vol. 185 B, 1894, p. 1024.

(3) Broom.—Ann. S. Afr. Mus., vol. iv, pt. ii, 1903.

) BouLtEencER.—Proc. Zool. Soc. London, 1904.

(10) Wit1iiston.—Biol. Bull., vol. xvii, 1909.

(11) Srerner.—Anat. Anzeiger, Nr. 22, Band 53, 1921.

(12) Grecory, Minss, and NospLze.—Bull. Amer. Mus. Nat. Hist., vol. xlviii, 1923.

(13) Wituiston.—Journ. Geol., vol. xviii, Chicago, 1909.

(14) Wiziiston.—Journ. Geol., vol. xix, Chicago, 1911.

(15) Case.—Revision of Cotylosauria of North America, 1911.

(16) Wiiuiston.—Journ. Geol., vol. xvi, 1908.

(17) Casze.—Permo-carboniferous Vertebrates from: New Mexico, 1913,

(18) Hunne.—Centrbl. fur Mineral. etc., 1911.

(19) Broom.—Ree. Alb. Mus., vol. i, 1903.

Ann. S. Afr. Mus., Vol. XXVIII. Plate VI,

MANUS OF BRADYSAURUS VANDERBYLI.

Neill & Co., Lid.

Ann. 8. Afr. Mus., Vol. XXVIII.

CARPUS

ies 2:

AND DIGITS OF PAREIASAUBIA.

Bite Vil:

Neili & Co., Lid.

Plate VIII.

Ann. S. Afr. Mus., Vol. XXVIII.

MANUS OF PAREIASAURIAN.

& Co., Lid.

evil

Pareiasaurian Studies.—Part IV. 113

Part IV.—On the Parevasaurian Pes.—By LikUWE D. Boonstra,

M.Sc., Assistant in Palaeontology.

(With Plates IX-XII and Text-figure 29.)

HISTORICAL.

THE proximal row of tarsal elements Seeley (1) found to consist of

one bone—a composite structure formed by the fusion of the fibulare

and intermedium. The presence of a nutritive foramen indicates the

line of fusion. Although not preserved in his specimen he considered

that a tibiale was present in life.

Broom (2), in 1903, was of the opinion that the proximal element

of Embrithosaurus schwarz. probably consisted of the united tibiale,

intermedium, and fibulare.

In 1908 Broom (3) stated that the proximal tarsal element of

Propappus omocratus differed in contour from that of Parevasaurus,

and “it probably represents the tibiale, fibulare, and intermedium

fused together.”’

Of Brachypareia rogerst Broom (4) said, “* The proximal tarsal bone,

which certainly represents the conjoined tibiale and fibulare with

possibly the intermedium, is of the well-known Pareiasaurian type.”

In 1913 Broom (5) considered the proximal element to represent

at least the tibiale and fibulare. Comparing it to the condition in

Procolophon, he thought that the intermedium must also be included.

In the distal row of tarsals Seeley (1) recognised four or five elements.

Broom (3) wrote that in Propappus omocratus the distalia are small

and cartilaginous for the most part. Two are, however, ossified, one

being the first distal tarsale.

In 1913 Broom stated that in the pes of Bradysaurus bain four

distal tarsalia are developed. The fifth distale he considered to have

been unossified.

In 1892 Seeley (1) said, “ There appears to have been five digits in

the hind foot . . . the phalanges, including the claw, appear to have

been three in number.” The foot was broad and short. The bones

of the pes were not found in articulation, and considerable doubt

must be attached to Seeley’s identification. In 1903 Broom (2) recon-

structed Embrithosaurus schwarzi with the phalangeal formula of the

hind foot as 2, 3, 3, 4, 3. Our exact knowledge was, however, not

increased thereby. Boulenger (6), in 1904, stated that the phalangeal

VOL. XXVIII, PART 1. 8

114 Annals of the South African Museum.

formula of Seeley’s specimen of Bradysaurus baini is 2, 3, 3, 4, 3.

Boulenger, as most subsequent authors, laboured under the false

impression that the phalangeal formulae of hind- and fore-feet were the

same. Attention has repeatedly been drawn to the digging function

of the fore-limbs. This difference in function surely ought to induce

us to consider it probable that there exists a difference in structure.

Of Propappus omocratus (3) Broom wrote, ““ Now it can with much

| certainty be stated that two digits in the pes had more than three

| phalanges. These are the third and the fourth.” The fifth digit

| could not be determined and was probably feebly developed. Broom

summed up by assigning to both Propappus and “ Pareiasaurus ”’ the

phalangeal formula 2, 3, 4, 5, 3. The material on which Broom

based these considerations is supposed to be in the South African

Museum. I have not been able to find it.

In his paper, “ On the Manus and Pes of Pareiasaurus,’ Broom (5)

stated that the formula is 2, 3, 3, 4,3. The difference in the phalan-

geal formula of the closely allied Propappus and Pareiasaurus was

ascribed to difference in body-weight. Unfortunately, the material

now preserved in the South African Museum is not sufficient to enable

me to confirm Broom’s determinations; of “‘ Paresasaurus”’ there

are not sufficient phalanges to make even three digits.

8941 Bradysaurus

Mynhardts Kraal,

Tapinocephalus

MATERIAL.

The material on which this paper is based can be tabulated as

follows :—

= Geographical Geological

Cat. No. Name Locality. Hemesnt Remarks.

Complete left pes in

| vanderbyla Beaufort West Zone—high articulation.

5012 Brachypareia Abrahams Kraal, | T'apinocephalus | Incomplete right pes, 3rd

| rogerst Prince Albert Zone—low and 4th digits certain.

8034 | Hmbrithosaurus | Hoogeveld, Lot A, | Tapinocephalus | Right and left proximal

schwarzt Beaufort West Zone—middle tarsals in articulation.

5 Paz Bradysaurus Leeuw River, Beau- | T'apinocephalus | Right proximal tarsal in

bani fort West Zone—high articulation.

5015 Bradysaurus Abrahams’ Kraal, | Tapinocephalus | Proximal tarsal, distalia,

vanderbyla Prince Albert Zone—low and metatarsals.

5624 | Bradysaurus Groot Kruidfontein, | Tapinocephalus | Right and left proximal

| seeleyt Prince Albert Zone tarsals in articulation.

8953. | Brachypareia Hoedemakers Kraal, | Tapinocephalus | Left proximal element

rogerst Beaufort West Zone in articulation.

5590 Nochelesaurus | Abrahams Kraal, | Tapinocephalus | Three distal tarsalia in

strubent Prince Albert Zone—low position,

Pareiasaurian Studies.—Part IV. 115

In addition to these there are in the collection about a dozen

disarticulated proximal tarsal elements, a number of disarticulated

distal tarsals, metatarsals, phalanges, and terminal claws.

DESCRIPTION.

(a) BastPopium.

Proximal Row.—In the proximal row of the Pareiasaurian tarsus

there is only one element. That this bone is of a composite structure

has long been known. From the

historical review it has, however,

become clear that diverse views have

been held as to which of the three

proximal bones are included.

In a consideration of the reptilian

tarsus two points must be taken into

consideration.

(a) A passage for nutritive vessels

passes in between the elements of the

proximal row. It is to-day a text-

book verity that this passage always

passes between fibulare and inter-

medium.

(b) In a recent paper Broom has

brought forward a convincing array

of facts which tend to show that the

radiale is the most inconstant of the

three bones of the proximal row.

The single element of the Pareia-

Fig. 29.—The reconstructed left pes

saurian tarsus has a penetrating fora-

men. It must therefore be composed

of at least the fibulare and inter-

medium. The tibiale may also be in-

cluded, but in view of the fact that the

tibiale has been shown to be generally

inconstant this is not very likely.

In the study of eight specimens in

of Bradysaurus vanderbyli (S.A.

Mus. Cat. No. 8941). It has

been deemed advisable to give

a reconstruction as in the origi-

nal (cf. Plate XII) the proximal

part of the tarsus is somewhat

crushed. Three distal tarsals

are sketched. In Brachypareia

rogersi (S.A. Mus. Cat. No. 5012)

there were probably four. 3.

which the tarsal element is in articulation it has become apparent

that preaxially there is a considerable overhang of the tibia. To

support this overhang I think that a small separate cartilaginous

tibiale was present in life. This view was also held by Seeley.

—— oo —_ eee

116 Annals of the South African Museum.

The general form of this tarsal element has been well described by

Seeley, who, however, did not inform us completely of the nature

of the palmar aspect. The foramen opens about in the centre of the

bone. From the foramen a groove stretches distally and slightly

preaxially. It reaches the distal surface at the preaxial margin of

the distal articulatory surface. A shallower groove stretches distally

and postaxially from the foramen, and opens on the distal surface

on the postaxial margin of the articulatory surface. These grooves

housed the vessels which penetrated the bone at the foramen. In

previous descriptions little was said of the distal surface of the proximal

tarsal. From the material at my disposal it is certain that the distal

surface carried one large elliptic articulatory surface. With what did

this articulate ? The distal tarsals are small pebble-like bones, and

there are no centralia. The only possible explanation is that some

of the distals (presumably the third and fourth) were encased in a

large cartilaginous pad, and that this pad articulated with this large

distal articulatory surface.

Although the proximal tarsal elements of the different Pareiasauridae

agree in general features, it becomes evident that there are a number

of types when minor differences are taken into account.

(a) The proximal tarsal elements of the three species of Bradysaurus,

the two species of Nochelesaurus and of Embrithosaurus compose one

group. In these forms the bone is large and massive. The ridge

separating the tibial and fibular articulations is strong, but it does

not reach the distal margin of the bone. The distal articulatory

surface is large, and is carried on distal prolongation of the bone.

This projection is composed chiefly of the fibular part.

(6) Approaching closely to the first group is the element of Platy-

oropha brooms (S. Afr. Mus. Cat. No. 5002). This bone, although

massive, is smaller, higher, and of less lateral extent. The median

dorsal ridge extends practically to the distal margin. The distal

articulation is not carried on a distal prolongation of the bone.

element is small. The smallness is mainly due to the decrease in size

of that part formed by the intermedium. The grooves on the palmar

aspect are very marked. In this group may also be included the

proximal tarsal of Brachypareva rogersi (S. Afr. Mus. Cat. Nos. 8953

and 5012). In the whole group the median ridge does not extend to

the distal margin. The distal articulatory surface is situated on a

slight prolongation of the bone. This is most marked in 8953, least

in 5012; 1058 occupies an intermediate position.

Parevasaurian Studies.—Part IV. 117

(d) In two unidentified specimens the tarsal bones are aberrant.

In South African Museum Catalogue No. 4348 the bone is very broad

but short and low. The weak median ridge extends to the distal

margin. The distal articulatory surface is not carried on a prolonga-

tion of the bone, and is furthermore not clearly demarcated from the

rest of the distal surface. In South African Museum Catalogue

No. 4346 the bone agrees with 4348 in being broad, short, and low.

The median ridge is, however, strong, and the distal articulatory

surface, although not situated on a distal prolongation of the bone,

has its limits defined.

Measurements.

a. b. C. d.

Brady-

aad Platy- Baad

eroup. oropha. Pro- rachy- 4348 | 4346.

pappus. pareia.

mm. mm, mm, mm. mm, mm.

Proximo-distal 88-110 85 82 60-74 97 71

Pre-postaxial . | 137-166 | 1138 aid 98-118 | 180 | 138

Dorso-palmar. 57-86 82 59 50-55 Uy 458

Central Row.—No bones in the central row of the tarsus have been

found. The presence of an intertarsal jot would cause a reduction

of central elements.

Distal Row.—The distal tarsalia appear to ossify late in life. In

our specimen of Bradysaurus vanderbyli (S. Afr. Mus. Cat. No. 8941),

although the rest of the pes is well ossified, there is no trace of ossified

distalia. In the incomplete pes of Bradysaurus % baini (S. Afr. Mus.

Cat. No. 3533), on which Broom based his conclusions, two large

distal elements are at present preserved in the collection. Broom

figured four. In one of our specimens of Brachypareza rogersi (S. Afr.

Mus. Cat. No. 5012), the fourth and fifth metatarsals articulate with

a medium-sized distale. In South African Museum Catalogue No.

5590, probably a specimen of WNochelesaurus strubent, Haughton

sketched three distalia in his field-book. These he considered to be

the second, third, and fourth.

This is all we know of the distal row of tarsals. In view of the

fact that the fourth and fifth digits of 5012 articulate with the same

distale, I think we are justified in assuming that there are four distal

tarsalia in the Pareiasaurian pes.

118 Annals of the South African Museum.

(6) METAPODIUM.

In Bradysaurus vanderbyli (8941) there are four digits, and con-

sequently four metatarsals. There is no doubt that this animal

only had four toes. The whole pes was excavated in a block of

matrix. On either side of the preserved digits there was matrix but

no indication of another digit. In the specimen of Brachypareia

rogers (5012) a previous developer has disarticulated the digits, and

I am only sure of the three last digits. As the second digit from the

postaxial side has four phalanges it is certainly the fourth digit. In

this animal there were five toes.

In both 8941 and 5012 the first metatarsal is an irregular pentagonal

bone, much broader than long. Its proximal surface is bipartite. The

postaxial half only articulates with the fourth distale.

The second metatarsal has the shape usual to normal metatarsals.

It is a short stout bone.

The third metatarsal approaches still more closely to the usual

form of a metatarsal. It is longer than the second and has more

of a shaft.

The fourth metatarsal is the longest of all, with a well-constructed

shaft and expanded ends.

The fifth metatarsal of Brachypareia rogersi (S. Afr. Mus. Cat.

No. 5012) is a small and slender bone. It is evident that the fifth

digit was feebly developed. Thisisalso Broom’s opinion. As already

stated, there is no fifth metatarsal in 8941—Bradysaurus vanderbylt.

Measurements in Millimetres.

8941. 5012.

Proximo-distal . | 40 | 40 | 49 | 57 | 33 | 30 | 36 | 45 | 30

Pre-postaxial . . | 60 | 51 | 45 | 44 | 40 | 30 | 29 | 37 | 22

Dorso-palmar . . | 27 | 29 | 34 | 40 |13 | 20 | 21 | 220eis

It has already been mentioned that 8941 has only four digits.

Unfortunately the fifth digit of 5012 is only represented by the

metatarsal.

Pareiasaurian Studies.—Part IV.

119

First Digit.—In both 8941 and 5012 this digit is composed of a

broad short phalanx and a long terminal claw.

Proximo-distal

Pre-postaxial

Dorso-palmar

Measurements.

Ist Phalanx.

8941. 5012. 8941.

mm. mm, mm.

24 20 42,

AT 26 46

23 15 30

Claw.

5012

mm.

Second Digit.—The second digit is composed of three bones less

massive than those of the first.

relatively longer.

Proximo-distal

Pre-postaxial .

Dorso-palmar .

Third Digit.—The third digit differs little from the preceding one,

with the exception of the first phalanx.

Measurements.

Ist Phalanx. 2nd Phalanx.

8941. | 5012. | 8941. | 5012.

about twice as long as that of the preceding digit.

is not so evident in 5012.

Proximo-distal

Pre-postaxial .

Dorso-palmar .

Measurements.

Ist Phalanx. 2nd Phalanx.

8941. | 5012. | 8941. | 5012.

The terminal claw is less broad and

Claw.

8941. | 5012.

mm. mm

45 45

45 on

24 20

This phalanx is in 8941

This difference

Claw.

8941. 5012.

mm. mm

Ab 46

42 31

25 21

120 Annals of the South African Museum.

Fourth Digit—The fourth digit has four elements. The terminal

three differ little from the three elements of the second and third digits.

The first phalanx is, however, much longer than the first phalanx of

the third digit. It is a strong bone with a distinct shaft and with

expanded ends. In fact, it looks very much like a metatarsal.

Measurements.

Ist Phalanx. | 2nd Phalanx.| 3rd Phalanx. Claw.

8941. | 5012. | 8941. | 5012. | 8941. | 5012. | 8941. | 5012.

mm. mm. mm. mm. mm. mm. mm. mm.

Proximo-distal . | 30 | 20 | 24 | 14 | 20 | 13 | 40 | 41

Pre-postaxial . | 43 | 31 ) 37 | 22 |-'35> | 249)" SOMimeg

Dorso-palmar'. . | 32 1.19) | 24 |) 16° |) 20 |) bp) 2s a eee

Fifth Digit.—In 5012 there is a fifth metatarsal preserved but no

phalanges. Judging from the metatarsal the phalanges must have

been small and slender. The terminal claw was probably feebly

developed.

COMPARISON.

In a comparison of the pes with the manus (7) the specialisation

of the proximal row of tarsal elements becomes very evident. The

manus has the primitive three elements preserved, whereas in the

pes only two are preserved, and that in the form of a highly specialised

composite “ heel-bone.”’

In 8941, where we have both the left manus and pes, a difference

in the rate of ossification is apparent. Whereas the distal carpalia

are well ossified, there is no trace of ossified distal tarsalia.

With the development of an intertarsal joint the pes has lost the

elements in the central row of the tarsus. The carpus has no such

joint, and retains a centrale.

A striking difference is the fact that the pes has four digits in 8941,

whereas the manus has the more primitive number, five.

The first metacarpal and metatarsal are differentiated from the

other metacarpals and metatarsals in a similar way. Both are broad

irregular bones.

The first difference that strikes us in the acropodium is the different

number of digits in pes and manus in Bradysaurus vanderbyli. The

Pareiasaurian Studies.—Part IV. 121

fourth digit of the pes differs from the corresponding digit of the

manus in the possession of four phalanges.

The terminal claws of the manus are much more developed than

those of the pes. The Pareiasaurians used their fore-limbs for digging

as wellas for locomotion. The hind-limbs had only the latter function.

The cause for the difference in the development of the claws is to be

sought in this difference of function.

Broom (8) has recently reviewed our knowledge of the reptilian

tarsus, but a few comparative remarks on the various Cotylosaurian

families will not be out of place here. The proximal row of

tarsals consists of tibiale and fibulare in Seymouria (9); of fibulare

and a fused intermedio-tibiale in Limnoscelis (10); of tibiale and

fibulare in Labidosaurus (11); of fibulare and a conjoined intermedio-

tibiale in Captorhinus (12); and of a tibiale and a fibulare in Dva-

sparactus (13). Inno Cotylosaurian are three separate bones preserved.

The reduction is brought about either by loss or fusion. In the

Pareiasauridae this tendency towards reduction of separate elements

has reached its height in the production of a single highly specialised

composite bone.

The central row of the tarsus is represented in Seymourza by one

centrale ; in Labidosaurus and Captorhinus there is either one or two ;

in Diasparactus no centralia are preserved. In this respect also the

Pareiasauridae are further removed from the primitive condition.

The primitive phalangeal formula is taken to be 2, 3, 4,5, 4. In

the Cotylosauria the phalangeal formula of the hind foot is known

with certainty only in Limnoscelis. Bradysaurus vanderbylr, with

only four digits and the reduced phalangeal formula 2, 3, 3, 4, 0,

is without doubt more highly specialised than any other known

Cotylosaur. Brachypareva rogersi, with its five digits, approaches the

condition of the other Cotylosaurs, but even it is highly specialised.

The evidence obtained from our study of the Pareiasaurian pes

justifies our stating that as far as the hind limb is concerned the

Pareiasauridae are the most highly specialised of the Cotylosaurs.

= =

— a

es oS aS See

122 Annals of the South African Museum.

EXPLANATION OF PLATES.

PLaTE IX.

Proximal tarsal elements of various Pareiasaurians. x#.

(a) S.A. Mus. Cat. No. 5127, Bradysaurus baini—dorsal view, right tarsal.

(6) S.A. Mus. Cat. No. 4348, undetermined—dorsal view, right tarsal.

(d) S.A. Mus. Cat. No. 5002, Platyoropha broomi—dorsal view, left tarsal.

(ec) S.A. Mus. Cat. No. 4346, undetermined—dorsal view, right tarsal. _

(f) S.A. Mus. Cat. No. 8953, Brachypareia rogersi—dorsal view, left tarsal.

PLATE X.,

Dorsal view of right proximal tarsal element of Embrithosaurus schwarzi (S.A. Mus.

Cat. No. 8034), in articulation with the epipodium. x2. a=fibula; b=tibia;

¢=proximal tarsal element.

Puate XI.

Dorsal view of right pes of Brachypareia rogersi (S.A. Mus. Cat. No. 5012). The

epipodium, proximal tarsal element, and the last three digits are as im situ ;

the first two were disarranged, and their position is only to be taken as

approximate. X4';.

PLate XII.

Dorsal view of left pes of Bradysaurus vanderbyli (S.A. Mus. Cat. No. 8941). All

the elements are in the exact relations in which they were fossilised. x2.

WORKS CITED.

(1) StrLey.—Phil. Trans. Roy. Soc., vol. 1838, 1892.

(2) Broom.—Ann. S. Afr. Mus., vol. iv, pt. ii, 1903.

(3) Broom.—Ann. 8. Afr. Mus., vol. iv, pt. viii, 1908.

(4) Broom.—Ann. S. Afr. Mus., vol. vii, pt. v, 1912.

(5) Broom.—Ann. S. Afr. Mus., vol. vii, pt. vi, 1913.

(6) BouLENcER.—P.Z.S., 1904.

(7) BoonstRa.—Preceding paper.

(8) Broom.—P.Z.S., 1921.

(9) WirListon.—Journ. Geol., vol. xix, 1911.

(10) Wixtiston.—Ameda, Permian Vertebrates, 1911.

(11) WiLuiston.—Journ. Geol., 1908.

(12) Wittiston.—New or Little Known Permian Vertebrates, 1909.

(13) WiLLIston and Casz.—Permo-carboniferous Vertebrates from New Mexico,

1913.

Ann. S. Afr. Mus., Vol. XXVIII. Plate IX.

a b

PROXIMAL TARSAL BONES OF PAREIASAURIA.

Neill & Co., Ltd.

Plate X.

Ann. S. Afr. Mus., Vol. XXVIII.

PODIUM OF EMBRITHOSAURUS SCHWARZI.

EPIPODIUM AND BA

Neill & Co., Lid.

Ann. S, Afr, Mus., Vol. XXVIII.

PES OF

BRACHYPAREIA ROGERSI.

Plate XI.

Neill & Co., Ltd.

Ann. S. Afr. Mus., Vol. XXVIII.

PES OF BRADYSAURUS

VANDERBYLI.

Plate XII.

Netll & Co., Ltd.

(-128 )

4. A Tabulate Coral from the Bokkeveld Beds of South Africa.

| By T. W. Gevers, M.A., D.Phil.

(With Text-figure 30.)

In view of the rarity of fossil corals in the Bokkeveld beds of South

Africa, special interest attaches to the tabulate coral recently found

by the writer at Gamka Poort, particularly on account of the wide

distribution of this genus in the Lower Devonian of Europe and

America. No tabulate coral has as yet been recorded from the

Devonian of South Africa.

Pleurodictyum bokkeveldense sp. nov.

Locality and Stratigraphic Horizon.—The specimen to be described

was found at Gamka Poort in the Zwarteberg mountains of the Cape.

Its stratigraphic position is near the bottom of the series—that is to

say, within the lowest of the five shale horizons of the Bokkeveld,

below the first sandstones.

Mode of Occurrence.—Only one specimen was found, and this occurs

in the form of a cast in a tough micaceous and arenaceous mudstone.

Description.—Compound, corallum discoidal and rather flat. Outer

edge not completely preserved, but outline appears to be slightly

oval, almost circular. Upper surface of corallum rather convex.

Walls of corallites mostly thick, varying in thickness from -2 to -9 mm.,

with fairly numerous, irregular pores. Septa rudimentary, indicated

by nodose and denticulated radial ridges of varying strength. They

are weakly defined within the lower interior portion of cells by radial

lines of low, inconspicuous, granular spikes, but more strongly de-

veloped in upper portion of cells around their edges, where they are

represented by fairly consistent and regular nodose ridges.

Since the specimen is preserved as a cast, the visceral cavity of

each corallite being filled with mud, no tabulae can be discerned.

Several of the marginal corallite-casts, however, exhibit a distinct

groove running around their outer marginal side half-way up their

124 Annals of the South African Museum.

preserved heights. Possibly these features represent rudimentary

tabulae, or tabulae in incipient stages of growth.

The individual corallites diverge from the centre of the base and

are oriented more or less at right angles to the curved, convex surface

gree

Fia. 30.—Pleurodictyum bokkeveldense. x 24.

of the corallum. All of them taper towards the base, thus having

a distinct funnel-shaped appearance. This feature is particularly

strongly developed in the smaller and narrower, probably younger,

cells that are irregularly situated between wider and deeper ones.

While generally polygonal, their outline varies considerably. None

of them are hexagonal in form. The larger and deeper ones, while

still polygonal, are almost rounded, circular, and oval in outline.

Rectangular shapes are not infrequent in the case of the smaller,

narrower cells.

A Tabulate Coral from the Bokkeveld Beds of South Africa. 125

The specimen shows 34 cells in all and, although not completely

preserved along part of the margin, the total number of cells probably

was not more than about 40 to 43. The great variety in outline of

the individual cells has already been mentioned, but a still more

characteristic feature is their great divergency in size. While the

largest attain a width of 5 to 6 mm., casts of the smallest are repre-

sented merely by a long or short tapering spike of 2 to 3 mm. diameter

at the base of the cast (surface margin of corallite). Owing to the

great divergency in shape and size of the individual cells, a true radial

symmetry of the corallum is not at once apparent. If the missing

marginal portion of the corallum, however, is reconstructed, it is seen

that there is a large central cell round which the others are concen-

trically grouped. The great divergency of the cells, however, par-

ticularly in size, to a large extent obscures this radial symmetry.

Sometimes a number of large cells adjoin, and groups of small ones

also occur. The latter, however, do not occur as clusters, but where

present in greater numbers exhibit a linear arrangement in the inter-

vening space between larger ones. These features distinguish the

form under discussion from other known species of the same genus.

Dimensions.—Unfortunately the specimen is not completely

preserved in some of the marginal portions of the corallum, but the

latter possesses a longest diameter of 2-6 cm. and a shortest diameter

of 2-4 cm. approximately.

Affinities.—This form is very different from Pl. lenticulare Hall.v

var. lawrenttum commonly found in the Helderbergian (Lower

Devonian) of New Scotland, North America. The latter is characterised

by its very large and few cells, the walls of which are strongly marked

by nodose and broken septa. There is a central cell, hexagonal in

form, bounded by six other similar cells. The lenticular corallum

frequently grows to a large size, showing three cycles of hexagonal

cells around the central cell.

From this species Pl. bokkeveldense differs mainly through the

marked irregularity in the size, shape, and arrangement of the indi-

vidual cells.

A somewhat closer relationship is shown to Pl. styloporum, found \

in the Middle Devonian of North America, the corallum of which,

however, is not nearly as discoidal, but attains a considerably greater

height.

A still closer relationship is shown to the well-known PI. problemati-

cum Goldf. which is very common in the spirifer-bearing greywackes

and sandstones of the Lower Devonian of the Rhineland. This form

an Ab 126 Annals of the South African Museum.

' possesses a flat, discoidal corallum of circular or oval outline and is

also invariably found as a cast. The corallum of Pl. bokkeveldense,

however, appears to be slightly more convex and exhibits a far greater

irregularity in the arrangement, size, and shape of the individual

cells.

Pl. bokkeveldense shows the closest affinity to Pl. amazonicum Katzer

ete |

a) | (Geologie des unteren Amazonasgebietes, p. 268, pl. ix, fig. 1 a-d,

| 1903), which is common in the Lower Devonian Conularia beds of

| ‘Bolivia. (Kozlowski: “‘ Faune Dévonienne de Bolivie,’ Annales de

| Paléontologie, 1923, vol. xu, p. 97, pl. x, fig. 1.) It differs, however,

| from this form by the very much greater number of its corallites and

the far greater irregularity displayed in their dimensions, shape, and

| distribution.

\ | | It is interesting to note that all these closely allied forms occur in

| the Lower Devonian of such widely separated localities and that,

< -— “78 a 2 —

eS SS eS

moreover, their modes of occurrence and the rocks in which they

occur are singularly alike.

Biological Notes—It has already been mentioned that a radial

symmetry such as that displayed by Pl. lenticulare is somewhat

obscured in the case of Pl. bokkeveldense. It is interesting to note in

this connection that also the perfect radial symmetry of the former

type is only apparent and that in reality a bilateral symmetry forms

the ground plan of these tabulate corals.* The outlines of the cells

| | are forced on them by purely mechanical means and are conditioned

| by the space available. The mutual influence of the individual cells

|| on further growth causes subsequent cells to become progressively

| more and more polygonal in outline, at first pentagonal and finally

In the case of Pl. bokkeveldense there is a prominent large cell

situated very near the centre of the corallum. Its shape, however,

while definitely polygonal, is irregular, and it is not followed by a

regular cycle of more or less uniformly shaped cells. On the contrary,

| | it is surrounded by a number of small and large cells in irregular, not

| strictly cyclic distribution. Succeeding this irregular cycle, a more

| regular and strongly marked cycle follows consisting of five large

cells, not strictly uniform in size, separated by five considerably

smaller cells of various sizes and stages of growth. That marginal

part of the corallum, in which a sixth larger cell with its interstitial

| | hexagonal.

Connecticut Acad. Arts and Sciences, vol. viii, 1891, p. 201; ‘“‘ Symmetrical Cell

* Beecher, Ch. E., “‘ The Development of a Palaeozoic Poriferous Coral,’’ Transac.

Development in the Favositidae,” zhid., p. 215.

A Tabulate Coral from the Bokkeveld Beds of South Africa. 127

small cell belonging to this cycle without a doubt was situated, is not

preserved. Then follows a marginal cycle, most regular and distinct

of all, of cells not as large as the inner cells, but showing a greater

uniformity in size and outline. While the majority of the larger

interior cells, though still recognisably polygonal, are almost rounded,

even circular in outline, the cells of the marginal cycle have a pro-

nounced polygonal outline, the latter varying with the amount of

space available and the contours of the cells of the preceding cycle.

It is to be noted that the outer contour of these marginal cells are

almost invariably rounded (convex) and parallel to the almost circular

outline of the whole corallum. The same feature is noticeable in the

case of the preceding cycle of cells. Sharp angles generally only

occur on sides facing the interior of the corallum.

Although the outer marginal cycle of cells appears so very much

more regular, small, incipient interstitial cells on the very margin of

the corallum may still be made out, thus showing that growth and

enlargement were still actively going on when the colony of polyps

perished, probably as a result of being detached from its foundation

or buried by mud.

It is also to be noted that the interior cells, in accordance with the

convexity of the surface of the corallum, are deepest and become

progressively shallower towards the margin. This applies also to the

smaller, 2.e. younger, interstitial cells.

It is apparent from the above description that also in the case of

the corallum of Pleurodictyum bokkeveldense growth evidently began

from a central cell around which new cells gradually developed, and

that in its primitive state its original ground plan, though not nearly

as regular, resembled that of Pl. lenteculare. At an early stage of its

growth at least six cells appear to have developed around the central

cell, their different proportions being determined by the irregularity

in outline of the central cell. Originally in juxtaposition and in close

contact with themselves and the central cell, as growth went on

they gradually began to diverge from the latter and one another like

the sprouting stems of some varieties of grass. Their walls and that

of the central cell appear to have been irregular in thickness. By

nature already rather thick, more calcite was secreted by the various

polyps in those parts where the divergence and interstice between

neighbouring walls was greatest in order to effect a strengthening

and prevent the disruption of the gradually growing corallum. Where

the divergence and interstice between neighbouring walls became too

great to be filled up by secretion and strengthening of opposite walls,

128 Annals of the South African Museum.

and where sufficient space became available, a new polyp took up its ©

| abode. The space thus set free within the corallum was utilised for

| the formation of a new cell, its outline being determined by the angles

made by the adjoining walls and the angle of tapering of the new

visceral cavity by the angle of divergence of the older cells from the

central cell and one another, 7.e. purely mechanically by the space

available. Where this gradual divergence proceeded more slowly

and the potential interstice could still conveniently be filled up by

further secretion, 7.e. a thickening of the adjoining and opposite

walls, the formation of a new cell was correspondingly delayed. This

| process seems to have been continuous with the gradual addition of

new cells, z.e. the growth of the whole corallum. Where the diver-

gence of the older cells from one another and particularly from the

central cell was sufficiently great, several new cells were formed in

close proximity within the additional space thus available.

That there was a limit to this gradual enlargement of the corallum,

and that it proceeded more or less uniformly around the whole of its

margin and along certain definite lines, is shown by the fact that in

each species of Plewrodictyum the same definite circular or oval

outline is always maintained. Asa matter of fact the whole compound

corallum appears to behave like a single individual.

Reference must also be made to the fact that in the case of this

specimen of Pleurodictyum bokkeveldense there is no trace of the worm-

like appendage so frequently found within the centre of the corallum

of other species of the genus Pleurodictyum. Although the nature

of this appendage by some is still debated, there seems no doubt that

it actually represents the cast of a worm either living as a parasite

or in symbiosis with the coral, particularly in view of the fact that in

Kurope and America by no means all specimens exhibit this pheno-

menon. In addition, its frequency varies with the species found. Thus

in North America it is most common in the Middle Devonian form

Pleurodictyum styloporum, while in the other species it occurs with

lesser and varying frequency. In Hurope Plewrodictyum problemati-

cum most frequently exhibits it. Further, since forms in various

stages of growth, small and large, show the appendage to be of corre-

sponding dimensions, it is certain that the worm grew with the coral.

| The fact that the worm has been overgrown by the corallum, and

| individual cells appear to have been affected by it both as regards

| outline and size, seems to indicate that it must have taken up its

eee

abode within the corallum at a fairly early stage of development of

I} | the latter.

— -

:: — } —— ——--:: —_ c y =. —_ ——

: : >a =¥ ; SSS = == = as

= = 5 , .—o = Fé ra > =e z Se E — 7 = = = r- Pa i — —— =

a A == Pe. (a See > E 2 = te e-> — Be = === PID) 7 : $33

= SS — —oe- eet ye = = = FPR SS = SSS = > Le =

SSS SSS SS SS = == ane Ss aS a SF> SSSt SS = Saas =

e aN

A Tabulate Coral from the Bokkeveld Beds of South Africa. 129

Dacqué * favours the view that this curious phenomenon represents

a case of symbiosis and that the worm was not a parasite. He argues

that since the secretion of calcite is possible at any place within a

coral exteriorly (theca) and interiorly (septa, dissepiments, and

tabulae), the coral could easily have encased the worm by secreting

calcite around it, analogous to the formation of a pearl within an

oyster, if the worm had been an unwelcome intruder. Since, however,

the coral made no attempt to do this, and the corallum, moreover,

shows no sign of injury to its growth, he concludes that the worm

lived in symbiosis with the coral and was not a parasite.

It is interesting to speculate whether any other specimens of the

genus Pleurodictyum that may be found in the Bokkeveld beds in the

future will exhibit this remarkable and interesting phenomenon.

* Dacqué, E., Vergleichende biologische Formenkunde der fossilen niederen

Tiere, vol. ii, p. 473, Berlin, 1921.

VOL. XXVIII, PART l. 9

. On We A teotpoas boy he Uae Hee By a Bie

i, eb a an G. 5. ne Plates XIN-XV and As Text- “

oe a Coates of See Cee an ppendie on aa a)

oe Species from. the. Cretaceous of. Zuliland). Lo By. JOHN

fe Nee RENNIE, M. AS Webb. Research Scholar, ‘Department: : oh

eae of Geology, ‘University of | ee Town, ey Plates ee

RE and 2 Text- REEL 2

Cee S Eh 7 ao

3 oe ‘DSc, Hon Canine -Palaeonto- :

es eS and~ Lie De -Boonstra, M.Sc.,

ant in Palaeontology. oe Plates: XXXII-

Tand Text- Sige ae

ai ie

oe eae ‘South A ii ‘By Shae

ap Ae 8. eae - ete. aa) op eee :

Pid pa, s

NAR ak

peittee ? i

ae ye EM

oa - PRINTED FOR THE :

‘or THE SOUTH AFRICAN MUSEUM

A AND. THE |

1. SURVEY OF SOUTH AFRICA

(3%)

5. On the Cephalopoda of the Uitenhage Beds.

By L. F. Spatu, D.Sc., F.G:S.

(With Plates XIII-XV and 1 Text-figure.)

CONTENTS.

PAGE PAGE

I. INTRODUCTION . ; ai hie! B. Order Ammonoidea—conid.

IT. REMARKS ON THE UITENHAGE Genus Rogersites,

FAUNA . : 2 gy Spath . : . 142

III. Synopsis or CEPHALOPODA Family Neocomitidae,

AND DESCRIPTIONS OF NEW Spath. : y st

SPECIES . : : = leg Genus Hoplitides, v.

A. Order Nautiloidea So Koenen emend.

Family Nautilidae, Sayn . . barge Cog

Owen emend. Spath 139 Genus Distoloceras,

Genus Hutrephoceras, Hyatt. : Ste?

Hyatt. : . 139 Incertae Sedis_ . . 153

B. Order Ammonoidea . 140 Genus SBochianites,

Family Phylloceratidae, Lory . ; . 153

Zittel emend.. . 140 C. Order Belemnoidea, . 155

Genus Phylloceras, Family Belemnopsidae,

Suess . ‘ . 140 Naef emend. . . 155

Family Desmocera- Genus Belemnopsis,

tidae . : . 141 Bayle . . 155

Genus Hodesmoceras, Genus Aibolites

Spath . : 2) 14k (Montfort), Mayer-

Family Olcostephani- Hymar : > 156

dae, Spath . . 142

I. INTRODUCTION.

The Invertebrata of the Uitenhage fauna have received such ad-

mirable and exhaustive treatment at the hand of Dr. F. L. Kitchin *

that a fresh discussion of some of the Cephalopods may seem un-

necessary. I have, however, lately had the privilege of examining

a very fine collection of Uitenhage Ammonoidea, made by Dr. S. H.

Haughton in the course of his survey work. This collection com-

prises not only some new forms, notably an example of a genus

(Hodesmoceras) not hitherto recorded from these beds (or indeed from

* “The Invertebrate Fauna and Palaeontological Relations of the Uitenhage

Series,’ Annals 8S. Afr. Mus., vol. vii, part 2, No. 3, 1908.

VOL. XXVIII, PART 2. 10

een

132 Annals of the South African Museum.

Africa), but specimens of previously known forms that on account

of more favourable preservation supplement our knowledge. Addi-

tional illustrations of these species or their suture-lines may thus

prove generally acceptable. Moreover, a number of specimens in

the British Museum, not previously recorded, on closer study also

proved worthy of description or illustration. In the present paper

the number of Cephalopoda known from the Uitenhage Beds is thus

brought up to twenty; eight of these were not represented in the

collections examined by Dr. Kitchin.

My best thanks are due to Dr. 8. H. Haughton for submitting this

interesting fauna to me, and to Dr. W. D. Lang, F.R.S., the Keeper

of the Geological Department of the British Museum (Natural History),

for facilitating my work by allowing me, as always, to make the

fullest possible use of the rich collections in his charge.

Il. REMARKS ON THE UITENHAGE FAUNA.

The most striking features of the Cephalopod fauna of the Uitenhage

Beds are, first, the abundance of gigantic Olcostephanids, apparently

unknown (in a similar size) from anywhere else, also of Bochianites

africanus, and, second, the apparent scarcity of the other forms

recorded, as of Belemnites and Nautili. This may be due partly to

accidents of collection, but it is a characteristic that cannot fail to

impress any observer who compares the Uitenhage fauna with a cor-

responding assemblage from another part of the world. Thus at

Speeton,* in bed D,, there occur Rogersites (though smallin comparison

with those of the Uitenhage Beds) associated with the same type of

Bochianites—that is to say, there occur the two dominant Uitenhage

genera, in the company of similar Hoplitides and Distoloceras. One

of the last was, in fact, attached to an Uitenhage species (Dzstoloceras

spinosissimum, Hausmann). There can be no doubt that the Uiten-

hage fauna is of Upper Valanginian age and corresponds to my

Hoplitidan age or to what in the south of France has been called

the “zone of Saynoceras verrucosum and Kalranella roubaudiana.”

Unfortunately no finer zoning of the beds in the Uitenhage Series is

as yet possible. 3

In the Speeton succession Belemnites form the most important

element of the cephalopod fauna; and as Danford ft has shown,

* Spath, ‘‘ The Ammonites of the Speeton Clay and the Subdivisions of the

Neocomian,”’ Geol. Mag., vol. lxi, 1924, table to p. 80.

+ ‘Notes on the Belemnites of the Speeton Clays,’ Trans. Hull Geol. Soc.,

vol. v, part 1, 1906, p. 6.

On the Cephalopoda of the Uitenhage Beds. 133

“they swarm at the very compact base of D,, Belemnites lateralis

and B. russiensis being perhaps most abundant, while the Compound

Nodular Band (D,) seems mainly tenanted by B. explanatoides.”” In

Northern Germany, also, according to Stolley,* Acroteuthis, almost

the only group of Belemnites represented in the Lower Neocomian,

is of the greatest importance, whilst Hzbolites, so abundant in the

Mediterranean Valanginian, is practically absent. The fresh-water

beds of the south of England and the Boulonnais, of course, yield

no Cephalopods; but when marine beds appear again, at the south-

eastern border of the Paris Basin (Aube, Yonne), only doubtful

Rhynchoteuthis are recorded. In the Southern Mediterranean facies,

however, in addition to the Hibolites just mentioned and the character-

istic flattened Belemnites of the genus Duvalia, there is a rich ammonite

fauna, including not only numerous Neocomitids, but abundant

Phylloceratids and Lytoceratids. The last are entirely unknown

from the Uitenhage Beds, and forms of the other two ammon-

ite families are extremely rare in comparison with the dominant

Olcostephanids and Bochianites.

In North Africa we meet with the same Mediterranean facies ;

and in Tunisia, at least, the pyritised ammonites of the Valanginian

marls (in localities ike Hammam Lif) are the same as those of South-

eastern France. But according to Haug 7} the abundance of Duvalia

in the Lowest Cretaceous of Madagascar makes it necessary to attach

this to the Mediterranean Province and indicates in a certain manner

the existence oi direct communications with the Tethys.

The Neocomian fauna recorded by Wray t from Fernao Vellozo

in Mozambique (Portuguese Hast Africa) was regarded as presenting

a facies resembling that of the Uitenhage Series. Newton § referred

to the fossils, presented by Mr. Wray to the British Museum, as

awaiting systematic description ; and as his two ammonites are before

me, I may briefly discuss. them in this connection. One was quite

correctly compared to Rogersites schenki (Oppel), and does not differ,

except in its small size, from such Uitenhage Rogersites (admittedly

of very narrow specific interpretation) as those listed below under

R. sphaeroidalis (e.g. No. 194). The second example, described by

* “* Belemniten der Unteren Kreide Norddeutschlands,” IV. Jahresb. Nieder-

sachs. Geol. Ver. Hannover (Geol. Abt.), 1911, p. 177.

+ Traité de Géologie, vol. ii, fasc. 2, 1907, p. 1232.

t ‘‘ Observations sur la géologie du district de Mozambique,’ Comm. Serv. Géol.

Portugal, vol. xi, 1915, pp. 69-84.

§ Appendix to Teale, ‘‘ Geology of Portuguese East Africa, etc.,” Trans. Geol.

Soc. S. Afr., vol. xxvi (1923), 1924, p. 157.

134 Annals of the South African Museum.

Mr. Wray as a small Phylloceras, “‘ not sufficiently well preserved to

allow of specific determination,’ I believe to be a Neolissoceras,

probably N. grasianum (d’Orbigny) itself, which species, according

to Kilian,* ranges from the boissierd zone or the Infravalanginian up

to the Hauterivian. It also is a Mediterranean element.

But any attempt to consider the Uitenhage fauna more closely

comparable to such an assemblage (with Polyptychitids) as that

found at Speeton, than to Mediterranean faunas seems to be negatived

by the comparatively rich Lytoceras fauna known from Mahiba Hill,

west of Port Amelia, Portuguese East Africa. When the late Mr. R. B.

Newton fj first recorded this fauna, including Belemnites, he stated

that I had considered the Lytoceras fragments insufficient for strati-

graphical purposes, but quite well supporting a lower Cretaceous

horizon. Unfortunately Mr. Newton at that time did not realise

the value of some other fragments in the same collection, for he only

showed me some of the more favourably preserved examples of

Lytoceras of that assemblage. Even these are not specifically deter-

minable, but the portion of the periphery of a Neocomitid (Lytvcoceras

of the type of L. regalis [Bean] or Neocomites neocomiensis [d’ Orbigny]

as figured by Sayn ft), and the impression of a fragment of the Uiten-

hage Bochianites africanus are decisive and unmistakable. These

are accompanied by similarly characteristic forms of the Belemnite

genus Duvalia. From Madagascar the latter is recorded together

with the Infravalanginian genus Protacanthodiscus (group of Hoplites

andreaet, Kilian), doubtful Neocomitids, and the later Rogersites

madagascariensis (Lemoine §), which, as Kilian || has pointed out,

belongs to the group of FR. atherstoni, and cannot be compared to the

boreal genus Swmbirskites. There is thus no doubt that the Valan-

ginian ammonites of both Madagascar and Portuguese Hast Africa

confirm Haug’s view, already quoted, and it appears probable that

the curious resemblance between Rogersites and the Polyptychitids,

hitherto known almost entirely from the boreal province, is a case

of homoeomorphy, although both sprang from Spvticeras.

In view of the occurrence of Valanginian deposits and genera with

Mediterranean affinities further south, it would not be surprising to

* In Frech, Lethaea Geognostica, II, Mesozoicum, 3. Kreide, I, 1 (1910), p. 174.

{ Loc. cit., Trans. Geol. Soc. S. Afr., vol. xxvi, 1924, p. 156.

t ‘“‘ Ammonites pyriteuses des marnes valanginiennes du S. EH. de la France,”

Mém. Soc. Géol. France, vol. xv, fasc. 2, Mem. No. 23?, 1907, pl. vii, figs. 5b, 60.

§ Etudes Géologiques dans le Nord de Madagascar, Paris, 1906, p. 182, pl. i,

fies:

|| Loc. cit., Lethaea, 1910, p. 215.

On the Cephalopoda of the Uitenhage Beds. 135

find them also in Tanganyika Territory. No ammonites of undoubted

pre-Hauterivian and post-Portlandian age, however, seem as yet to

have been discovered in this part of Africa. Krenkel * already thought

it probable that there was a considerable gap between the Upper

Jurassic and the Lower Cretaceous; but he wrongly assumed some

Hauterivian forms he described to be comparable to Valanginian

Uitenhage species. His “ Holcostephanus’”’ dacquei belongs to the

sulcosus group (=Subastrerta, Spath) which, at Speeton, is even later

than the true Olcostephanus of the basal Hauterivian. Also Krenkel’s

“new variety’ of Phylloceras rogersi, Kitchin, since renamed by

Zwierzycki, is as unsuitable for exact dating as are obviously the

lamellibranchs. Among the Tanganyika forms later described by

Zwierzycki f as coming from the Trigonia schwarz beds, there is,

indeed, a form that may be a Valanginites, namely, “ Holcostephanus”’

crassus Zwierzycki; there are also some very doubtful fragments

attributed to “ Hoplites”’ ci. neocomiensis (d’Orbigny), and a

“ Bochianites”’ that on account of its more elaborate suture-line

does not compare well with the Valanginian species known. But in

the new collections from the Trigonia schwarz beds of Tendaguru,

now before me, there is not a single form that can be attributed to a

pre-Hauterivian species; and the “ Astverca”’ forms from Mikadi,

especially, including all the species described by Zwierzycki, are true

Holcostephanus and show no close similarity to the presumably earlier

Uitenhage Rogersites here discussed. On the other hand they are

associated not only with Hauterivian Holcodiscids and Crioceratids,

and with Barremian Heteroceras, Lytoceratids, etc., but even with

Lower Aptian Ancyloceras and Procheloniceras, in the same facies (from

Niongala). There is apparently a conformable succession from the

Hauterivian to the Aptian; but there are no undoubted Valanginian

ammonites, and perhaps no marine equivalent of the Uitenhage beds.

Whether the Upper Saurian Beds of Tendaguru with traces of

a Wealden flora, but now correlated with the Purbeck,t are of a

* “Untere Kreide von Deutsch-Ostafrika,’ Beitr. Pal. Geol. Osterr.-Ung.,

vol. xxiii, 1910, p. 230.

+ “‘Cephalopoden Fauna der Tendaguru Schichten in Deutsch-Ostafrika,”’

Wiss. Ergeb. Tendaguru Exped., 1909-12, pt. 3, Archiv f. Biontologie, vol. iii,

Heft 4, 1914, p. 83.

t See Dietrich, ‘‘ Das Alter der Trigonien Schichten am Tendaguru,”’ Centralblatt

fir Mineralogie, etc., B, 1927, p. 63. Since the above was written, Dr. F. L.

Kitchin’s important paper, “‘On the Age of the Upper and Middle Deinosaur

Deposits at Tendaguru, Tanganyika Territory,” has appeared (Geol. Mag., vol.

Ixvi, No. 779, May 1929, pp. 193-220).

136 Annals of the South African Museum.

corresponding age is doubtful. It seems improbable that they com-

pletely bridge the wide gap between the basal Hauterivian above

and the Trigonia smeer beds with Portlandian ammonites below.

But it may be advisable to discuss the relations of these smeez beds

with the so-called Umia group of Kachh, since this is also often

compared with the Uitenhage Series, and since we cannot trace the

Neocomian farther northward, nothing pre-Barremian being known

from either Jebel Moghara, east of Suez,* or from Somaliland.; On

a previous occasion,t however, [ recorded certain doubtful (either

entirely new or else poorly preserved) ammonites from Somaliland

that might have belonged to the privasensis zone of the uppermost

Jurassic (Tithonian) or the bozssiera zone of the Lowest Cretaceous

(Infra-Valanginian). They were somewhat reminiscent of Medi-

terranean types, like Pomel’s forms from Lamoriciére in Algeria and

the Infra-Valanginian of Tunisia, and resembled the fauna of the

Jowest Cretaceous of the Argentine, since described; and although °

I stated that the existence, in that part of Somaliland, of beds of so

late an age was not yet proved, yet I thought it would tend to confirm

a temporary marine transgression across Northern Africa during

Tithonian and Lower Cretaceous times. It may be recalled in this

connection that the Upper Kimmeridgian ammonites described, in-

cluding Virgatitids, were also entirely new, and that nothing like

them was known from the whole of Africa or the Indo-Madagascan

Province.

Now, in Kachh we have the Katrol Beds of Gudjinsir in the north-

west, which were considered by Waagen to represent the lowest

Katrol, apparently resting directly on the Dhosa Oolite. In reality

these beds are much later than the Katrol Beds of the Katrol Range

and the south of Kachh generally. They yield an abundance of

Haploceras elimatum (Oppel) in addition to Hildoglochiceras, a Streb-

lites, Ptychophylloceras angelim, P. gemminum (Oppel), etc., and

Perisphinctids of the sparsiplicatus group to be described in the next

part of my Revision of the Jurassic Cephalopod Fauna of Kachh. This

is not unlike the fauna of the Trigonia smeer beds of Tanganyika, per-

haps also of the glauconitic sandstones of Antsalova in Madagascar ;

* H. Douvillé, “‘ Terrains Secondaires dans le Massif du Moghara. Paléont.,”

Mém. Acad. Sci., vol. liv (2), 1916, pp. 89 ff.

+ Mayer-Eymar, ‘‘Uber Neocom-Versteinerungen aus dem Somaliland,”

Vierteljahrschr. Naturf. Ges. Ziirich, vol. xxxviii, 1893, pp. 1-17.

t “‘ Ammonites and Aptychi,” Part VII of Monograph on Collection of Fossils

and Rocks from Somaliland, Hunterian Museum, Glasgow, 1925, pp. 111-164.

= Tes OS <3 Fs Se

ei SO ae |

On the Cephalopoda of the Uitenhage Beds. 137

and it is now taken to be of Portlandian age rather than Upper

Kimmeridgian. The Zama shales of Wynne* are possibly still

higher. On the other side of the Jumara Dome (north-west) where

there is an unexamined series of rocks between the Dhosa Oolite

scarp and the lowest ammonite-bearing bed of the Umia group, the

latter has yielded an abundance of often gigantic Virgatosphinctes fF

of the denseplicatus-frequens group; but there is a second ridge of

similar oolitic rock with still later Tithonian ammonites behind it,

yielding new species comparable to some described from the Proniceras

beds of the Crimea and Mexico, and to forms of Himalayitidae of the

Spiti Shales. This higher fauna was entirely unknown to Waagen,

who, on the other hand, wrongly listed Katrol forms from the Umia

Beds.

Separated from these Lower Umia scarps with undoubted Tithonian

ammonites by further ridges and a thickness of barren sandstones

of probably several hundred feet, there follows a Trigonia Bed; and

above a further unknown thickness of sandstones of the Upper Umia

group, le the remnants of marine Aptian beds at Ukra Hill. The

matrix (red and yellow limestones and ironstones) is identical with

that of many of the Jurassic forms. According to Mr. Raj Nath,

who has lately investigated the Kachh deposits and kindly allowed

me to figure some of his ammonites in the forthcoming memoir, the

presence of a fault on one side and the igneous mass of Ukra Hill

on the other, prevent the establishment of a definite succession up

to the Aptian. Throughout the great thickness of these sandstones,

however, as Wynne showed long ago, plant remains are common

and there is no trace of any ammonites of, apparently, the privasensis

zone of the uppermost Tithonian, certainly not of the Infra- Valanginian

and the whole of the Lower, Middle, and Upper Neocomian (Valan-

ginian, Hauterivian, and Barremian). Trigonia in Kachh, also, have

been found in the Katrol, Lower and Upper Umia groups and well

above the horizon at which most of the plant fossils have been obtained.

Whatever the relations of the lamellibranchs (of unknown ranges)

of the Upper Umia group to those of the Uitenhage Series may be,

the striking dissimilarity in the ammonite faunas of the two areas is

now explained by their great difference in age.

In the fauna of the Spiti Shales, Valanginian elements, notably

* See Spath, “‘ Revision of the Jurassic Cephalopod Fauna of Kachh (Cutch),”

Pal. Indica (N.S.), vol. ix, Mem. No. 2, fasc. 2, p. 159.

+ One of these giants encloses T'rigonia retrorsa, Kitchin, and numerous Acantho-

thyris multistriata, Kitchin.

i ——

‘—

1? weg

a SS eS en Se = as 2 ress baa R . Ss

138 Annals of the South African Museum.

Rogersites again occur, and I have lately been able to examine a lower

Neocomian fauna from Southern Persia with the characteristic

ammonites of the Valanginian Marls of the south of France. But

the Olcostephanus known from the Salt Range (Kalabagh and Chichali .

Pass) and from the Samana area (in the similarly glauconitic

Belemnite Beds of Thal) * are Lower Hauterivian forms, and are

as closely allied to those of e.g. the Crimea (where Upper Valan-

ginian Rogersites also occur) as to those of the Trigonia schwarz: Beds

of Tendaguru.

Turning now to South America, we find Valanginian faunas with

Rogersites and Bochianites resembling the Uitenhage forms well

developed in Colombia, and I have on a previous occasion correlated

these “‘ Valanginites Beds” of Colombia with deposits of Hoplitidan

age from localities as far apart as Speeton and Spiti. In Mexico,

Bose ¢ has lately compared with the Uitenhage species, R. atherstona

and R. baini, some forms of his Astzerza Beds which probably include,

and are certainly underlain by, Valanginian beds. We could com-

pare the Uitenhage ammonites with these forms at least as well as

with the Valanginian species known from farther south, e.g. “ Astierza ”’

laticosta Stehn t from the Argentine Andes which on account of its

association with forms of the “ zone of Spiticeras damesi (Steuer),”’

comparable to the new Somaliland forms above discussed, is probably

of earlier Valanginian date, although somewhat intermediate between

Valanginites and Rogersites and resembling the Hast African V. ?

crassus (Zwierzycki). It is clearly premature to generalise from the

scanty data of the distribution of comparable Valanginian ammonites

at present available, and I may repeat what I wrote in connection

with the discussion of an Albian fauna from Nigeria: § “ As in the

case of the Hotriassic fauna of Hast Greenland lately recorded, the

distribution of the African Cretaceous Ammonites might well be used

as evidence in favour of the permanence of the Continents, 7.e. in the

present case, of the huge ancient land-mass south of the Sahara, since

Pre-Cambrian times, with only its fringe occasionally submerged.”’

* See Spath, “‘ The Fossil Fauna of the Samana Range and Neighbouring Areas,”

Part V. The Cretaceous Cephalopods, etc., Mem. Geol. Surv. India., Pal. Indica,

N.S. (in press), pl. viii, figs. 1-3.

+ ‘‘ Alounas Faunas Cretacicas, etec.,”” Bol. Inst. Geol. Mexico, No. 42, 1923,

p. 76, pl. ii, figs. 3-6, p. 77, pl. iii, figs. 1, 2.

t “‘La Fauna Neocomiana de la Cordillera Argentina, etc.,” Actas Acad. Nac.

Cienc. Rep. Argentina, vol. ix, p. 62, pl. ii, figs. 8, 8a.

§ “The Albian Ammonoidea of Nigeria,” Appendix to The Nigerian Coalfield,

Section II, Bull. No. 12, Geol. Surv. Nigeria, 1928, p. 54.

On the Cephalopoda of the Uitenhage Beds. 139

Ill. Synopsis oF CEPHALOPODA AND DESCRIPTIONS

or New SPECIES.

Ciass Cephalopoda.

A.—OrvER NAUTILOIDEA.

Famity NAUTILIDAE, Owen emend. Spath.

- Genus HuTREPHOCERAS, Hyatt, 1894.

1. Eutrephoceras witenhagense, sp. nov.

(Text-figs. la, b.)

1856. Nautilus undeterm. Sharpe. ‘“‘ Secondary Fossils from South

Africa,’ Trans. Geol. Soc.

(II), vol. vii, p. 201.

1908. a sp. Kitchin. “ Invertebrate Fauna of the Uitenhage

Series,” loc. cit., p. 225.

Type.—The specimen recorded by Sharpe from the Sundays River

(B.M., No. 11034, Geol. Soc. Coll.).

~ Diagnosis.—Coiling occlusal, with umbilicus nearly closed. Whorl-

section rounded, slightly compressed at first, later flaring, with greatest

thickness at inner third and no edge to high umbilical wall. Suture-

line with slight umbilical saddle and shallow lateral lobe, straight

across venter. Annular lobe strongly developed. Test entirely

smooth, thick.

Measurements of Type :—

Diameter in mm. (about). > 135

Height of outer whorl, in per cent. of diameter (about) 55

Thickness _,, r. fe 74.

Urabilicus; : , i

Remarks.—The holotype of this species, merely recorded by Sharpe

as being 54 inches in diameter, is not in a perfect state of preservation,

although it shows all the characteristic features. The outer whorl is

septate to the end, but this is largely corroded, as is one side of the

earlier half of the outer whorl. The test is, however, preserved in

patches, and the inner whorls, on being broken out, allowed of a fairly

satisfactory reconstruction of this species. The excentric position of

the almost dorsal siphuncle is probably not a specific character.

140 Annals of the South African Museum.

The form is close to Nautilus boissieri, Pictet,* with a similarly

straight suture-line, but a thinner whorl-section. There is nothing

TEXT-FIG. 1.—Hutrephoceras uitenhagense, sp. nov. (a) Outline sectional view and

(6) tracings of four septal edges. (Reduced to 2 linear.) Sundays River

(B.M., No. 11034, Geol. Soc. Coll.).

like the present species among numerous Nawtilz in the British Museum

from the Hauterivian to Aptian deposits of the Tendaguru District,

Tanganyika Territory.

B.—OrpER AMMONOIDEA.

Famity PHYLLOCERATIDAH, Zittel emend.

Genus PHYLLOCERAS, Suess, 1865.

2. Phylloceras rogersi, Kitchin.

1908. Phylloceras rogersi, Kitchin. ‘Invertebrate Fauna of the

Uitenhage Series,” loc. cit.,

p. 179, pl. viii, figs. 19, 19a—e.

—-— -- - _ 7 — ~ 4 ~ -

it = = fa Sa .

ot — es p 1 me = a Jd =:

i — ie 2 Sito Soe Sebi tag ae 2 fie int aS ——

a > SS E tS ne ae 8 =

ea a a ET _ 2 — - - Sa - es ~ —— ore -

* Mélanges Paléontologiques, vol. ii, 1866, ‘‘ Faune a Terebratula diphyoides de

Berrias,”’ p. 58, pl. viii, fig. 4.

—

On the Cephalopoda of the Uitenhage Beds. 141

Non 1910. Phylloceras rogersi, Kitchin. var. nov. Krenkel, ““Untere

Kreide von Deutsch-Ost-

afrika,’ Beitr. Pal. Geol.

Osterr.-Ung., vol. xxiii, Heft

A e225. pla rxxdt,| ie. 79

(=P. krenkelt, Zwierzycki,

1914, loc. cit., p. 84).

The holotype described by Kitchin still remains the only available

specimen. Its suture-line and inflated whorl-shape suggest reference

to Ptychophylloceras, i.e. the semisulcatus group, but there are neither

ventral ridges nor umbilical sulci and the lineate ornamentation seems

to prove that the species must be included in Phylloceras s.s.

Uhlig * took this form to represent the “ northern element ”’ in the

Uitenhage fauna, together with Belemnopsis africanus (Tate), which

he considered to show that Himalayan Belemnites spread as far as the

extreme south of Africa. The rarity of these two species makes them

of little value for far-reaching generalisations.

Famity DESMOCERATIDAE, Zittel.

Genus EKopEesmoceras, Spath, 1923.

3. Hodesmoceras haughtoni, sp. nov.

(Pl. XIII, figs. 2a-e.)

Type.—A specimen from “Shore of pan, Zoutpan, Uitenhage ”’

(S.A.M., No. 227, S.H.) here figured.

Diagnosis.—Coiling platygyral (with wide and flat, compressed

whorls), subleptogyral (thickness under 33 per cent. of the diameter),

subangustumbilicate (umbilicus rather narrow), with narrowly arched

venter and steep but rounded umbilical wall. Test with traces of

very faint sigmoidal striation. Suture-line simple, with high external

lobe and trifid first lateral lobe (see Pl. XIII, fig. 2e).

Measurements of holotype :—

Diameterinmm. . : ' 5 ial

Height of outer whorl (in per cent. of diameter). OU

Thickness 5 ‘ Hs at . 8

Width of umbilicus a if fe : 20

* “Marine Reiche, etc.,’’ Mitt. Geol. Ges. Wien, vol. iv, 1911, p. 408.

142 Annals of the South African Museum.

Remarks.—The unique example of this species, unfortunately, is

incomplete, terminating in a. septal surface, so that it probably

represents merely the inner whorls of a larger form like Hodesmoceras

celestint (Pictet and Campiche).* This has the same type of suture-

line which does not differ from that of many Jurassic Haploceratids,

at a corresponding size. The suture-line here figured (enlarged x 9)

was taken from the last quarter of the outer whorl of the nucleus

represented in fig. 2a (Pl. XIII), itself enlarged three diameters ; but

at the end of the outer whorl there are still only three lateral and two

auxiliary lobes. The outer whorl was figured separately, since its

dorsal aspect (fig. 2d) 1s instructive. The inner whorls figured in

figs. 2a, b were taken out of this outer whorl, but the intervening

portions (not figured) are partly corroded.

Neolissoceras, which also occurs in beds of the age of the Uitenhage

Series (and is, indeed, far commoner in Europe), shows the charac-

teristic whorl-shape and peculiar umbilical wall already at small

diameters. To judge by a number of specimens in the Lamplugh Col-

lection (B.M., Nos. C32366-70, from Autan, Drome) which show the

suture-line perfectly, this is more Haploceratid and less Desmoceratid

than that of the present species, and the unsymmetrically divided

first lateral saddle especially is more like that of the Upper Jurassic

Haploceras.+

Famity OLCOSTEPHANIDAH, Spath, 1924.

Genus RoceErsITES, Spath, 1924.

4. Rogersites atherstoni (Sharpe).

1856. Ammonites atherstoni, Sharpe. Trans. Geol. Soc. (I1), vol. vu,

p: 196, pl. xxi, tes) daa:

1908. Holcostephanus atherstont (Sharpe). Kitchin, “ Inverteb.

| Fauna, Uitenhage Series,”

fon Se

1909. Astieria atherstona (Sharpe). Wegner, “ Uebersicht Astrerca

Formen,’ Neues Jahrb. ie

Mins eten(h)s pacts

1909. Holcostephanus atherstoni (Sharpe). Hatch and Corstorphine,

Geol. of S. Air) “pas0s:

text-fig. 76a.

= “Verrain Crétacé de Ste. Crorx, I,” Pal. Suisse, Il, 2, 1860; pp. 276, 3am

plo xxxix, figs. 1; 2.

t See Spath, “ Revision of the Jurassic Cephalopoda of Kachh,” Pal. Indica

(N.S.), vol. ix, Mem. No. 2, fasc. 4, pl. Ixxxi, fig. 6a.

On the Cephalopoda of the Uitenhage Beds. 143

1924. Rogersites atherstoni (Sharpe). Spath, Geol. Mag. (Speeton

Clay), p. 87.

Kitchin’s exhaustive discussion of this species has not prevented

authors from continuing to use Sharpe’s name for other forms from

various parts of the world. This is due partly to the reduced original

figure which does not clearly convey the fact that the holotype repre-

sents merely the inner whorls of a gigantic form. On the other hand,

examples like those figured by Burckhardt * or by Béset from

Mexico as Astreria cfr. atherstona and A. ex. aff. atherstoni represent

the outer whorls of Olcostephanus of the astierianus-filosus group such

as are common in the south of France (e.g. B.M., No. C31110, from

Moustiers Ste. Marie, Basses-Alpes).

The great authority on Lower Cretaceous Ammonites, the late

Prof. W. Kilian,t was more fortunate in his identifications, but he

protested against Kitchin’s numerous species, and the exclusion, from

R. atherston, of various European forms. Kilian considered this

species to occur “ without any doubt” in the Upper Valanginian of

the Jura region, but on the same and the following pages (213 and 214)

he characteristically called the forms from the Jura and the south of

France first “mere varieties’’ of Sharpe’s species and then Holco-

stephanus (Astieria) atherstoni and Holc. (Ast.) bainz “ mere varieties ”’

of Astierca forms common in France. On the inspection of figures,

like those of Baumberger’s § Astierva cf. atherstont or A. imbricata,

A. actinota and A. leptoplana (Baumberger), which appear to represent

the inner whorls of large Rogersites, it certainly seems probable that

Sharpe’s species also occurs in the Valanginian of Hurope; but they

must be kept distinct from the true Olcostephanus (“‘ Astieria’’) of the

Lower Hauterivian, and the correct identification of immature nuclei,

in any case, is generally impossible. Thus itis very uncertain whether

Kilian’s || Holcostephanus atherstona (Sharpe), var. nov., which was

later named by Wegner § var. densicostata, is really a young Rogersites ;

* “ Faune Jurassique de Mazapil,” Bol. Inst. Geol. Mexico, No. 23, 1906, p. 185,

pl. xi, figs. 2, 3.

+ “‘ Algunas Faunas Cretacicas, etc.,”’ Bol. Inst. Geol. Mexico, No. 42, 1923,

Me 1, pl..ii, fies. 1, 2.

{ In Frech, Lethaea Geognostica, II, 3, 1910 (fasc. 2), p. 213.

§ ‘‘ Fauna d. Unteren Kreide im Westschweizerischen Jura,’’ Abhandl. Schweiz.

Pal. Ges., vol. xxxiv, 1908, pl. xxv, fig. 4; text-fig. 123, p. 15; pl. xxvi, figs. 1-4

only.

| ‘‘ Sur quelques fossiles remarquables de lHauterivien, etc.,”’ Bull. Soc. Géol.

France (4), II, 1902, p. 865, pl. lvii, figs. la, b.

q Loc. cit., “‘ Astieria Formen,”’ 1909, p. 82.

144 Annals of the South African Museum.

and even some undoubted immature specimens of Rogersites in

Dr. Haughton’s collection cannot be satisfactorily distinguished from

similar constricted young of the later Olcostephanus, or even the

(usually more coronate) Subastieria, although they remain almost

smooth to a diameter at which the last is already highly tuberculate

and costate. In Olcostephanus subfilosus * also the innermost whorls

are less cadicone and more prominently tuberculate and ribbed.

A large but fragmentary specimen (No. 320, 8.H.) in the new

collection shows that at a diameter of 230-240 mm. the ribbing has

become slightly more distant and the umbilical tubercles very high

but more widely spaced. Two other examples (Nos. 313 and 318,

S.H.) are unfortunately less favourably preserved, whilst two more

large examples (Nos. 315 and 319, S.H.) 7 are intermediate between

R. atherstoni and the form described below as R. sphaeroidalis. They

retain the open umbilicus of the former species (31 per cent. at 225 mm,

diameter) and show twenty strong spines round the umbilicus (as

compared with only about thirteen or fourteen in the more involute

R. sphaerordalis), but these transitional forms have more globose

earlier whorls.

5. Rogersites sphaeroidalis, nom. nov.

(PI) CHI hiest5)s PIP exeVesironle)

1908. Holcostephanus cf. atherstoni (Sharpe). Kitchin, “ Inverteb.

Fauna, Uitenhage Series,”

loc. cut., p. 193.

1909 ? be baint (Sharpe). Hatch and Corstorphine,

Geol. of ‘SS. Aft jpaeccor

text-fig. 73 (left 2).

1910 ey (Astieria) cf. atherstont (Sharpe). Kitchin,

Kilian, loc. cit. (Lethaea,

pt. 11), p. 214.

In his description of a form, close to #. atherstoni, but more inflated,

Dr. Kitchin discussed a specimen in the British Museum (No. 46534)

which he considered to show close agreement. This example again,

in its very depressed outer whorl, resembles the specimen (No. 195,

S.H.) t here figured, which may be taken as the holotype of the

present species. The inner whorls (Pl. XV, fig. 1) are not strikingly

* Spath, loc. cit., Geol. Mag., 1924, p. 76, =Olcostephanus astierianus, Pavlow,

non d’Orbigny, Argiles de Speeton, Moscou, 1892, p. 136, pl. xvii (x), fig. 15.

+ These two and the three preceding specimens are from “ just behind the house,

Cuyler Manor, Uitenhage.”’

{ From Cuyler Manor, Uitenhage.

On the Cephalopoda of the Uitenhage Beds. 145

different from the true R. atherstoni, but the whorl-section soon be-

comes very depressed. The portion of the outer whorl figured in

Pl. XIII, fig. 5 has a thickness or breadth of 115 mm. as compared with

a whorl-height of only 33 mm. (in the siphonal plane), and is thus even

more depressed than the outline-section 4b of Pl. XV (R. kitchini sp.

nov.). Later the whorl-thickness decreases again, and in a specimen

(No. 317) of 250 mm. diameter, the shape does not differ greatly from

that of the transitional forms discussed above, but the width of the

umbilicus is only 24 per cent. of the diameter, and there are only

thirteen or fourteen large and distant spines round the umbilicus. The

ribbing is distinctly coarser than in R. atherstoni, also in the example

(B.M., No. 46534) recorded by Dr. Kitchin, of about the same size

as Sharpe’s type of R. atherstonc; but in another very large specimen

(No. 316, S.H.), indistinguishable from the present species, it remains

close, as in ft. atherstoni. This, therefore, may also be regarded as a

passage-form between the two species, and there is another smaller

example (No. 524, 8.H.) that also seems to have rather closer costation

towards the end, whilst in four still smaller specimens (Nos. 194, 312,

314, 322,* S.H.) only the inflated whorl-shape is relied on for reference

to R. sphaeroidalis. A typical body-chamber portion in the Geol.

Soc. Coll. (B.M., No. 34198, labelled Amm. atherstoni, but un-

localised) agrees with the example (No. 46534) discussed by Dr.

Kitchin, but some crushed Rogersites from the A. G. Bain collection

(B.M., Nos. C32206-7, labelled Amm. atherstoni, Sundays River) may

well belong to the true R. atherstont.

6. Rogersites wilmanae (Kitchin).

1908. Holcostephanus wilmanae, Kitchin. “ Inverteb. Fauna, Uiten-

| hage Series,” loc. cit., p. 195,

pleixe tess Pla:

1909. Astierva psilostoma (Neumayr and Uhlig), var. wilmanae,

Kitchin. Weener, loc. cit.,

“ Astrerra Formen,” p. 86.

1910. Holcostephanus (Astierta) wilmanae, Kitchin. Kilian, loc. cit.

(Lethaea, fasc. 2), p. 214.

This form is undoubtedly very close to Neumayr and Uhlig’s

R. psilostoma, with the same type of peristome, and to the forms

separated by Wegner as var. pictetr and var. koenent. The species

* Specimens 312, 314, 322, as well as 316 and 317 above mentioned, are from

‘just behind Cuyler Manor, Uitenhage’’; 194 is also from ‘‘ Cuyler Manor” ;

524 from “ Cliffs on Zoetgeneugd, Sundays River.”

yn ae

146 Annals of the South African Museum.

is again listed independently only because the holotype is partly

crushed and its inner whorls are as yet unknown, and it is not certain

that they are comparable to those of a passage form between R. psilo-

stoma or R. wilmanae and R. bainz here figured (Pl. XIII, fig. 3; Pl. XIV,

fig.4; Pl. XV, fig. 2).* It will be noticed that in this the constrictions

are not noticeable; on the body-chamber, which occupies nearly

half of the outer whorl, the ribbing becomes coarser (as in R. baind),

but it might, perhaps, also be included in a more comprehensive and

almost universal species R. psilostoma. The suture-line of this

transitional example is well shown (see Pl. XIV, fig. 4), but while there

is general agreement with the diagrammatic suture-line figured by

Pictet ¢ it is doubtful whether it is identical with the suture-line of

the true R. psilostoma (Neumayr and Uhlig).

A specimen in the Geol. Soc. Collection (B.M., No. C32205, from

the Sundays River, labelled ““ Amm. atherstont”’) is not deformed

by crushing, like the holotype, and whilst showing the closest agree-

ment in external characters with Neumayr and Uhlig’s{ type of

R. psilostoma, shows an even more inflated whorl-section, though the

periphery does not become so broad as that of R. convolutus (v.

Koenen).§

7. Rogersites baini (Sharpe).

1856. Ammonites baini, Sharpe. Trans. Geol. Soc. (II), vol. vu,

p. 197, pl. xxi, ies 2a

1908. Holcostephanus baini (Sharpe). Kitchin, “ Inverteb. Fauna,

Uitenhage Series,” p. 187.

1909. Astierca baina (Sharpe). Wegner, loc. cit. (Astieria Formen),

Ooze

1909. Holcostephanus baini (Sharpe). Hatch and Corstorphine,

Geol. of S. Atrica, )paya0s,

text-fig. 760.

HOO: A (Astverva) bavna (Sharpe). Kilan, loc. cit.

(Lethaea, fasc. 2), p. 214.

This well-characterised species is represented by a typical fragment

(No. 525, 8.H., from the “Clifis on Zoetgeneugd, Sundays River)

* B.M., No. C32204, ex Geol. Soc. Coll., from Sundays River, labelled “‘ Amm.

baini.”’

{ “ Terr. Crét. Ste. Croix,” Pal. Suisse, IT. 2, 1860, pl. xlii, fig. 5.

t ‘“‘Ammonitiden a. d. Hilsbildungen Norddeutschlands,” Palaeontogr., vol.

xxxii, 1881, p. 149, pl. xxxii, figs. 2, 2a.

§ ‘“‘ Ammonitiden d. Norddeutschen N eocom,’’ Abhandl. K. Preuss. Geol. Land.

Anst., N.F., Heft. xxiv, 1902, pp. 146, 412, pl. xxxix, figs. 4a, b.

On the Cephalopoda of the Uitenhage Beds. 147

agreeing with the outer whorl of the holotype (B.M., No. 10976A,

Geol. Soc. Collection). A second example (No. 583, 8.H., from

Welbedachtsfontein) shows greater resemblance to the Sundays River

example in the British Museum (No. 52052) referred to by Dr. Kitchin

(under Holcostephanus cf. baint, p. 200), but this is not separable

from the present species and differs mainly in unimportant details

of ribbing and whorl-shape.

8. Rogersites crassicostatus, nom. nov.

1908. Holcostephanus cf. bain (Sharpe). Kitchin, ‘“ Inverteb.

Fauna, Uitenhage Series,”’ loc.

Clix, Pa d0e) Die ix, teed ple x,

ieee

This form does not differ greatly from the last, but considering that

Rk. baini is a small species and that the holotype of R. crassecostatus,

v.e. the septate specimen figured by Kitchin, at 92 mm. diameter

does not include the body-chamber, the differences seem at least as

important, for systematic purposes, as those between R. rogersi and

R. schenki (Oppel).* The holotype has been described by Kitchin

in great detail, and since there are no additional specimens I can

add but little to this description. The coarser ribbing of the inner

whorls of R. crassicostatus seems to me an important distinguishing

character, in addition to the difference in the course of the rib-curve

or radial line.

9. Rogersites rogerst (Kitchin).

1908. Holcostephanus rogerst, Kitchin. ‘“ Inverteb. Fauna, Uiten-

hage Series,’ loc. cit.,

oe CONS Tole Wx ale he

Tolle, 2%) Lule

1909. Astierta rogers (Kitchin). Wegner, loc. cit. (Astverva Formen),

p- 88.

1909. Holcostephanus rogersi, Kitchin. Hatch and Corstorphine,

Geol. of S. Africa, p. 303,

text-fig. 76c.

LO: - (Astierta) rogersi (Kitchin). Kilian, loc. cit.

(Lethaea), p. 214.

* “‘ Ostindische Fossilreste, etc.,’’ Palaeont. Mitteil. Mus. K. Bayer. Staates,

vol. i, 1863, p. 286, pl. Ixxxi, figs. 4a—-c: Uhlig, “ Fauna of the Spiti Shales,” Pal.

Indica, Ser. XV, vol. iv, pt. 1, 1903, p. 130, pl. xviii, figs. 2a-e.

VOL. XXVIII, PART 2. id

148 Annals of the South African Museum.

A number of immature specimens * in the new collection may be

referred to this species, but they are too small to be distinguished

from the young of such closely allied species as R. schenki (Oppel).

Another such immature specimen} is intermediate between the

present species and R. bain, and two very small examplesit are

specifically indeterminable.

10. Rogersites modderensis (Kitchin).

1908. Holcostephanus modderensis, Kitchin. “ Inverteb. Fauna,

Uitenhage Series,”’

loc. cit., p. 202) pl:

x, figs. 3, 3a.

1909. Astieria modderensis (Kitchin). Wegner, loc. cit. (Astierva

Formen), p. 89.

1910. Holcostephanus (Asteria) modderensis, Kitchin. Kilian, loc.

cut. (Lethaea), p.

214.

1924. Rogersites modderensis (Kitchin). Spath, loc. cit. (Geol. Mag.),

. p- 86.

What is probably a large example of this species (No. 693, S.H.)

was collected by Dr. Haughton in the cliffs on Zoetgeneugd, Sundays

River. Itis about 165 mm. in diameter and retains the small umbilical

tubercles, but the whorl-section becomes less depressed with increase

in size.

11. Rogersites kitchini, sp. nov.

(Pl. XV, figs. 4a, b.)

Type.—The Zwartkops specimen (B.M., No. C761) figured in PI.

XV, figs. 4a, b (reduced to half the natural size).

Diagnosis.—Coiling rather close (subangustumbilicate), perpachy-

gyral (whorls extremely thick). Shape cadicone, with very high and

steep umbilical wall and very broadly arched venter. Umbilical

edge very pronounced, and provided with prominent spines (18 on

outer whorl, successively fewer towards interior). Ribs across venter

blunt and comparatively distant. Suture-line unknown.

* Nos. 310, 311, S.H., from “‘ just behind the house, Cuyler Manor, Uitenhage ”’ ;

423, from Picnic Bush, Zwartkops River; also one specimen from one mile north

of Coega Station.

+ No. 582, S.H., from Welbedachtsfontein, Uitenhage.

t No. 410, S.H., from Picnic Bush, Zwartkops River ; and 5075 from Sundays ,

River.

On the Cephalopoda of the Urtenhage Beds. 149

Measurements of holotype :—

Diameter in mm. A eae : . about 300

Height of whorl (at 220 mm. diameter) », 40 per cent.

Thickness of whorl ( ye ean) a 80 af

Umbilicus ( es Sea. | - 30 <

Remarks.—This gigantic species may be considered to be a morpho-

logical transition between the genus Polyptychites and the typical

Rogersites. The peripheral aspect is reminiscent of such large forms

of Polyptychites as that figured by Neumayr and Uhlig* as

Olcostephanus marginatus (non Phillips), Roemer, but the umbilical

tuberculation is that of a Rogersites. The genus Valanginites Sayn fT

also produced somewhat similar forms, e.g. V. perinflatus, Matheron t

and V. stephanophorus Matheron,§ but they are finely ribbed and thus

represent a Polyptychitoid offshoot quite independent of Rogersites.

A large but fragmentary specimen (No. 5070) from the Sundays

River, sent by Dr. Haughton, of about 175 mm. (and the general

appearance of what a gigantic R. schenky may be supposed to be like)

seems closer to the present species than to any other described form

of Rogersites, but the umbilical tubercles are less prominent.

12. Rogersites otoitordes, sp. nov.

(Pl. XIV, figs. la, b.)

Type.—The Zwartkops (Railway Cutting) specimen (No. 876, S.H.)

figured in Pl. XIV, figs. la, 0.

Diagnosis.—Coiling rather close (subangustumbilicate), pachygyral

(whorls thicker than high), with whorl-section not so depressed as in

holotype of R. atherstont and umbilical tubercles more prominent.

Ribbing comparatively coarse, as in fh. modderensis, but finer on the

inner whorls. Peristome flared, projecting laterally and ventrally

but with inward-bent “ears.” Suture-line unknown.

Measurements of holotype :—

Diameter in mm. oe : ; 3 ; ; . 100

Height of last whorl (in percent.ofdiameter) . . 40

Thickness of last whorl ( ‘ aa ye » BC

Umbilicus ( ss % ec 0)

* Loc. cit.. Ammonitiden Hilsbildungen, 1881, pl. xxix.

+ See Kilian, loc. cit. (Lethaea, IT, 2, 1910), p. 196; Genolectotype, A. perinflatus,

Matheron (see Spath, Geol. Mag., 1924, p. 86).

t ‘‘ Recherches Paléontol. dans le Midi de la France,” 1878, pl. B20, figs. 7a, 6.

§ Ibid., figs. 4a, b (left by Sayn in ‘‘ Astierta’’ according to Kilian).

150 Annals of the South African Museum.

Remarks.—This species is probably closest to R. rogersi, which

differs in its less coronate whorl-shape with less prominent tubercles

and slightly closer costation. The peculiar flared peristome, with

its wavy outline, is unfortunately not perfectly preserved, so that

the outline drawing of fig. 1b is somewhat diagrammatic. It gives a

better idea, however, of the shape of the Penuloine than would photo-

graphic top and front views.

R. psilostoma (Neumayr and Uhlig) has a similar peristome, but is

also less coronate. R. schenki (Oppel) is undoubtedly closer, but it

displays only the tubercles of the inner whorls, not part of the ribs

as well; itis also more depressed and does not show the characteristic

change from a comparatively close costation of the inner whorls to

the coarse ribbing of the outer whorl. The latter is reminiscent of

the peculiar ornamentation of Polyptychites, and R. otowtoides is

probably a passage-form between this boreal stock and Rogersites

of the atherstona group. The somewhat similar R. boussingaulti

(d’Orbigny) * also belongs to the latter group.

A fragmentary example (No. 581, 8.H., from Welbedachtsfontein,

Uitenhage), doubtfully attached to the present species, may perhaps

represent the inner whorls of a passage-form to R. psilostoma.

13. Rogersites witenhagensis (Kitchin).

1908. Holcostephanus witenhagensis, Kitchin. “Inverteb. Fauna,

Uitenhage Series,”

loc. cit., payZ0e:

ple Satz

1909. Asteria uttenhagensis (Kitchin). Wegner, loc. cit. (Astierra

Formen), p. 89.

1909. Holcostephanus witenhagensis, Kitchin. Hatch and Cor-

storphine, Geol. of

S. Africa, p. 295,

text-fig. 73 (right).

1910. a (Asteria) uitenhagensis Kitchin. Kilian, loc.

cit. (Lethaea, 11, 2),

p. 214.

There is a crushed fragment (No. 692, S.H.) from “ clifis on Zoetge-

neugd, Sundays River,” of a large example of this species, recognisable

by the striate whorl-side, well visible in the umbilicus and the unusual

width of the latter. Kilian thought this species close to the inflated

* < Rossiles de Colombie, etc.,’’ Paris, 1842, p. 32, pl. i, figs. 1, 2.

On the Cephalopoda of the Uitenhage Beds. 151

variety of “ Astierra”’ sayni, Kilian,* but apart from the inflation of

their inner whorls, the passage forms between R. atherstoni and R.

sphaeroidalis above discussed are really much closer to the present

species than is the French form.

Fam. NEOCOMITIDAH, Spath, 1924.

Gen. HopiitTipEs, v. Koenen, 1902 emend., Sayn.

14. Hoplitides subanceps (Tate).

(Pl. XIII, figs. 4a—c.)

1867. Ammonites subanceps, Tate. South African Fossils, Quart.

Journ. Geol. Soc., vol. xxiii,

p. 150, pl. vu, figs. 3a, b.

1882. a. ee Be Holub and Neumayyr, Fossilien

der Uitenhage Formation,

loc. cit., p. 213.

1896. Reineckera ne (Tate). Newton, Journ. Conch., vol.

Wall, INO. Bs jose.

1908. Solgeria « Kitchin, Inverteb. Fauna,

Uitenhage Series, loc. citt.,

| jos Zs

1909. Ammonites 2 - Hatch and Corstorphine,

Geol. of S. Africa, p. 303,

text-fig. 76d.

This form is very close to species of Hoplitides common in the

Valanginian of the south of France, e.g. H. deperets and H. provincialis,

Sayn.t The suture-line of the former species was apparently drawn

very diagrammatically, but it has the same oblique, bipartite lateral

lobe as that of the present species. Unfortunately Tate’s type,

septate to near the end, if not entirely, is still the only example

known.

* “* Sur le Néocomien des environs de Moustier,”’ Bull. Soc. Géol. France (3),

vol. xxiii, 1896, p. 976 (=Ammn. asticrianus, pars, d’Orbigny, Pal. Frangaise, Terr.

Crét., I, 1840, pl. xxviii, fig. 4 only).

+ ‘‘ Ammonites pyriteuses des marnes valanginiennes du S.E. de la France,”

Mém. Soc. Géol. France, vol. xv, fase. 2 (No. 23), 1907, pp. 58, 59, pl. viii, figs. 7

and 10, text-fig. 26 on p. 59. '

————

——

== Sees eee

152 Annals of the South African Museum.

Genus DistoLoceras, Hyatt, 1900.

15. Distoloceras spinosissimum (Hausmann).

(PPO XSI fees)

1837. Ammonites spinosissimus, Hausmann. “ Beitrage zu der

Geognost. Consti-

tut. v. §.Admike

Gottinger Gelehrte

| Anzeiger, p. 1458.

1882. Crioceras spinosissimum (Hausmann). Holub and Neumayr,

“Fossilien der’

Uitenhage Forma-

tion,’ Denkschr.

K. Akad. Wiss.

Wien, vol. xliv,

p. 273, pl. ii

1896. a 28 (Hausmann MS.). Holub and Neumayr,

Newton, Journ.

Conch., vol. vii,

No. 9; poe:

1908. pS he (Hausmann). _Neumayr, Kitchin,

‘“‘ Inverteb. Fauna,

Uitenhage Series,”

loc. ctt., p. 225.

1924. Distoloceras cf. spinosissimum (Hausmann). Spath, loc. cit.

(Geol. Mag.), p. 75.

This species was not represented either in the collections examined

by Dr. Kitchin or among the new material forwarded by Dr. Haughton.

But there are two examples in the British Museum, one collected by

Capt. Rocke (Geol. Soc. Coll., No. 11085), which is now figured ; the

other (No. C10819, Miss Vaughan Williams, 1906) consisting of a

similar whorl fragment with the impression of the previous whorl

(not in contact) preserved in the matrix.

The innermost whorls are unfortunately as yet unknown, and in

the more complete example figured by Holub and Neumayr the

apparent uniformity of the ribbing of the earliest portion preserved

(perhaps due to corrosion) may suggest wrong comparisons. The

complex suture-line of the specimen here figured seems to agree with

that of other Distoloceras.

On the Cephalopoda of the Uitenhage Beds. 153

The writer agrees with Uhlig * in considering that Hyatt’s + genus

Distoloceras (for Ammonites hystrix [Bean MS.], Phillips, in Neumayr

and Uhlig) has nothing to do with the Lytoceratid genus Pictetia,

Uhlig; and no palaeontologist would now use the family Ancylocera-

tidae in Hyatt’s interpretation. This is no reason, however, why the

name Distoloceras should not be used for the hystrix-curvinodus group

and more or less uncoiled allies, leading to what I separated as

J uddiceras.t

16. Dvastoloceras sp. ind.

1908. Acanthodiscus sp. Kitchin. “Inverteb. Fauna, Uitenhage

Series,” loc. cit., p. 207.

Dr. Kitchin compared a doubtful fragment to Acanthodiscus

hystricoides (Uhlig),§ but this does not belong to the group of A.

radvatus (Bruguiére) to which I have previously || restricted the

genus Acanthodiscus. It is possible that the fragment is closer to

the species above discussed than Dr. Kitchin thought, the poor

preservation of the earlier whorls of Holub and Neumayr’s specimen

showing what is probably a spurious uniformity of costation. Some-

thing similar, however, is found in Dvrstoloceras roemeri (Neumayr and

Uhlig),4] Koenen, of a higher bed (radzatus zone of the basal Haute-

Tivian), a species which, according to Kilian,** is one of the few German

criocones that also occurs in the Neocomian of the south of France.

INCERTAE SEDIS.

Genus BocutanitEs, Lory, 1898.

17. Bochianites africanus (Tate).

(Pl. XIV, figs. 2a—c, 3a-c; Pl. XV, figs. 3a-c.)

1867. Hamites africanus, Tate. “Secondary Fossils from South

Africa,’ Quart. Journ. Geol.

* “ Fauna of the Spiti Shales,” Pal. Indica, Ser. XV, vol. iv, fasc. 2 (1910), p. 168.

+ In Zittel’s ‘ Text-book of Palaeontology,” first English edition of Eastman

(1900), p. 588. { Loc. cit., Geol. Mag., 1924, p. 84.

§ ““Cephaloden Fauna der Teschener und Grodischter Schichten,’’ Denkschr.

K. Akad. Wiss. Wien, vol. lxxii, 1902, p. 39, pl. i, figs. 8a, b.

|| Loc. cit., Geol. Mag., 1924, p. 87.

q Loc. cit. (Ammonitiden Norddeutsch. Neocom.), 1902, p. 294, pl. xvi, figs. 5a—c.

** Loc. cit. (Lethaea, 1910), p. 271.

154 Annals of the South African Museum.

Soc., vol. xxii, p. 150} spliiae

figs. 5a-d.

1908. Hamites africanus, Tate. Kitchin, Inverteb. Fauna, Uiten-

hage Series, loc. cit., p. 225.

There are now over twenty fragments of this species before me *

which may be considered to be, next to the gigantic Rogersites, the

most characteristic element of the Uitenhage fauna, and, as already

mentioned, has been found also in Portuguese East Africa. The

suture-line here figured (Pl. XV, fig. 3d) is composite, being taken

partly from a large fragment comparable to that figured in Pl. XIV,

figs. 2a—c, partly from one of Tate’s syntypes (Pl. XIV, figs. 3a-c).

This suture-line is essentially the same as that of B. neocomiensis

(d’Orbigny).t The small example illustrated in Pl. XV, figs. 3a-c, is

the original of Tate’s fig. 5c, but his fig. 56 represents merely a

fragment of the body-chamber of a larger individual. In the circum-

stances it seems advisable to consider his most complete example 5a,

here refigured, to be the lectotype.

The Speeton forms (from bed D,) which I ¢ listed as B. neocomiensis

(d’Orbigny) have a more circular whorl-section, but Karsten’s §

Baculites granatensis may well be specifically identical with the present

species. B. maldonadi (Karsten) || has a more elliptical and less

triangular whorl-section, but also the same type of ribbing.

Bochianites gerardi (Stoliczka) § with a similar suture-line has a

more circular section, and B. oosteri, Sarasin and Schondelmayer **

differs not only in the absence of strong costation, but in its far less

simplified suture-lines.

Bochianites was formerly believed by the writer to have originated

* Represented in Dr. Haughton’s collection from Cliff on Kuduskloof (Nos. 255

and. 360); Kloof on Colchester (Nos. 661-663 and two unnumbered lots) and Cliff

on Zoetgeneugd (No. 268a-—c ?). The examples in the British Museum are from the

Sundays River mouth (Prince Alfred’s Rest), cliff above Tunbridges, and

M‘Loughlin’s Bluff (Geol. Soc. Coll., ex Dr. Rubidge, G. W. Stowe, Major Rocke

Colls.). t Pal. Frangaise, Terr. Crét., vol. i, 1842, p. 560, pl. cxxxviii, fig. 4.

t Loc. cit., Geol. Mag., 1924, pp. 75 and 86, B.M., Nos. C32375-7.

§ “ Geognostische Verhaltnisse des westlichen Columbien,” Amtl. Ber. 32. Vers.

Deutsch. Naturf. und Arzte, Wien, 1856 (1858), p. 105, pl. ii, fig. 1.

|| Poed., pl. u, fig. 2.

J “‘ Geological Sections across the Himalayan Mountains, etc.” Mem. Geol.

Surv. India, vol. v, pt. i, 1865, p. 110, pl. x, fig. 3.

** Etude monographique des Ammonites du Crétacique inférieur de Chatel-St.

Denis, pt. 2, Mém. Soc. Pal. Suisse, vol. xxiv, 1902, p. 179, pl. xxiv, figs. 3 and 4,

and text-fig. 6, p. 180.

On the Cephalopoda of the Uitenhage Beds 155

from a Neocomitid stock, e.g. Distoloceras (via Juddiceras=group of

Crioceras curvicosta, v. Koenen), but the possible connection with the

Tithonian genus Protancyloceras, Spath (= group of Ancyloceras

gumbeli, Oppel, and A. gracile, Oppel, in Zittel), has yet to be investi-

gated. Its systematic position is thus uncertain.

18. Bochianites glaber, Kitchin.

1908. Bochianites glaber, Kitchin. “ Inverteb. Fauna, Uitenhage

Series,” loc. c2t., p. 181, pl.

vill, figs. 20, 21.

1909. oe Pe S Hatch and Corstorphine, Geol.

of 8. Africa, p. 303, text-

fig. 76f.

No additional examples of this species have been collected.

Bochiamites undulatus, v. Koenen,* from the Lower Aptian, with which

Dr. Kitchin had compared his species, I have on a previous occasion Tf

stated to be a homoeomorphous development of a different stock.

C.—OrpvER BELEMNOIDEA.

Fam. BELEMNOPSIDAH, Naef emend.

Gen. BELEMNopsIS, Bayle, 1878.

19. Belemnopsis africanus (Tate).

1867. Belemnites africanus Tate. “Secondary Fossils from South

Africa,’ Quart. Journ. Geol.

Soc., vol. xxii, p. 151, pl. vu,

figs. 2a, b.

1909. Bs ne BF Boehm, Centralbl. f. Miner.,

etc., p. 564.

1909. b se a Hatch and Corstorphine, Geol.

of 8. Africa, p. 303, text-fig.

76e.

1911. 5 6 a Uhlig, “Marine Reiche,” loc.

ctt., p. 408.

1927. Belemnopsis _,, oS Spath, Kachh Revision, i, loc.

ou. (Pal. Indica), p. It.

* Loc. cit., Ammonitiden Norddeutsch. Neocom., 1902, p. 393, pl. xxxv, fig. 13.

+ ‘‘ Notes on Ammonites,” I, Geol. Mag., 1919, p. 30.

156 Annals of the South African “Museum.

This form was not represented in the collections studied by Kitchin.

In addition to the holotype (B.M., No. 26890) refigured by Boehm

there is now a second small fragment (S. Afr. Mus., No. 665, 8.H.,

from Kloof on Colchester, Sundays River). Representing merely the

alveolar end of a guard, it could not be distinguished from a cor-

responding portion of the Upper Jurassic B. gerardi (Oppel), and its

allies, discussed in 1927. It may be recalled here that Neumayr

had considered this species to be a representative of the “ absoluti,”

and that Haug * thought it to belong with certainty to the genus

Cylindroteuthis.

Gen. Hisouites (Montfort), Mayer-Hymar, 1883.

20. Hubolites, sp. ind.

1908. Belemnites sp. Kitchin. “‘ Inverteb. Fauna, Uitenhage Series,”

loc. cit., p. 210. |

The fact that the two fragments recorded by Dr. Kitchin were

considered to represent subfusiform species, makes it probable that

they are to be referred to the genus Hibolites so abundant in the

Valanginian of the Mediterranean facies. In the North German and

Speeton successions Hzbolites are becoming of importance only in

the Middle Neocomian.

* Traité de Géologie, vol. ii, fase. 2, 1907, p. 1230.

On the Cephalopoda of the Uitenhage Beds. 157

EXPLANATION OF PLATES.

Puate XIII.

. Distoloceras spinosissimum (Hausmann). Septate whorl-fragment (unlocal-

ised), from the Geol. Soc. (ex Capt. Rocke) Coll. (B.M., No. C32194.)

P. 152.

. Hodesmoceras haughtoni, sp. nov. (a, 6) inner whorls, enlarged ( <3); (c,d) part

of outer whorl (enlarged X 2), and (e) suture-line (diagrammatic and enlarged

X 9) of holotype, from ‘‘ Shore of pan, Zoutpan, Uitenhage.’ (No. 227, 8.H.)

P. 141.

. Rogersites aff. wilmanae (Kitchin). Peripheral view of the transitional example

figured in Pl. XIV, fig. 4, and Pl. XV, fig. 2. Sundays River. (B.M., No.

C32204.) P. 146.

. Hoplitides subanceps (Tate). Enlarged side and peripheral views of holotype

( X2) and its suture-line (diagrammatic and enlarged <5). Sundays River.

(B.M., No. 10996, Geol. Soc. Coll.) 12 Jee

. Rogersites sphaeroidalis, sp. nov. Peripheral view of part of the outer whorl

of the example figured in Pl. XV, fig. 1, showing suture-lines. Cuyler Manor,

Uitenhage. (No. 195, S.H.) P. 144.

PLATE XIV.

. Rogersites otoitoides, sp. nov. Side view (a), with whorl-section and peristome

(5). Railway cutting, Zwartkops. (No. 876, S.H.) P. 149.

. Bochianites africanus (Tate). (a and c) lateral, (6) ventral views, (d) outline

whorl-section of a specimen from Kloof on Colchester. (V=venter; D=

dorsum.) P. 153.

. Bochianites africanus (Tate). Two lateral (a, b) and dorsal (c) views of lecto-

type (6 and ¢ are inverted). Prince Alfred’s Rest, Sundays River mouth.

(B.M., No. C25228.) P. 153.

. Rogersites aff. wilmanae (Kitchin). Part of side view (enlarged x2) of Sundays

River example figured in Pl. XIII, fig. 3 and Pl. XV, fig. 2. (B.M., No.

C32204). P. 146.

PuatTE XV.

. Rogersites sphaeroidalis, sp. nov. Side view of inner whorls of the specimen,

of which part of the outer whorl is figured in Pl. XIII, fig. 5. Cuyler Manor,

Uitenhage. (No. 195, S.H.) P. 144.

. Rogersites aff. wilmanae (Kitchin). Side view of. the Sundays River example

figured in Pl. XIII, fig. 3 and Pl. XIV, fig. 4. (B.M., No. C32204.) PP. 146.

. Bochianites africanus (Tate). Two lateral (a, c) and ventral (6) views of one

of the syntypes (Tate’s fig. 5c) and suture-line (d), composite and enlarged

x 2-5. Prince Alfred’s Rest, Sundays River mouth. (B.M., No. C25229.)

P. 153.

. Rogersites kitchini, sp. nov. Side view (reduced to half and not quite central)

and restored outline whorl-section of earlier portion (at about x) of holo-

type. From Zwartkops. (B.M., No. C761.) P. 148.

* + "

” oe %, ~ t Fant

. : * .

7 * .

4 a

' y

te XITT.

Pla

Neill & Co., Ltd.

L. F. SPAtTH.

Ann. 8. Afr. Mus., Vol. XXVIII.

L. T. SPATH.

Plate XIV.

Neill «& Co., Lid.

Ss F =

; =< hy

i i uM e

F ¥ . i ‘

‘ - oN ,

: . i ;

re: rf * ea

i Fry Reuss pa a) <

it ; : P= i <

, arte + - « ie

~ de we ~ oar ” “

z . « v af a

ie i x on = i - <

PP » oF pany LY 7” . 4)

i ’ a .

= z

: ‘

"I 4 f

. we i ‘} ,

3 Fe 4

b an ‘

’ 4 ;

;

{

aie yw o.

Pat + + ,

a ’

i )

, 4 »

“

‘ ¥

’

we “ ‘ ns

f ' ‘

c i

’ ; :

49 ,

” . fh é

eet y } i

xh, ' aa ,

Plate XY.

Ann. 8. Afr. Mus., Vol. XXVIII.

Newl & Co., Lid.

L. F. SPATH.

( 159 )

6. New Lamellibranchia and Gastropoda from the Upper Cretaceous of

Pondoland (with an Appendix on some Species from the Cretaceous

of Zululand).—By Joun V. L. Renniz, M.A., Webb Research

Scholar, Department of Geology, University of Cape Town.

(With Plates XVI-XXXTI and 3 Text-figures.)

CONTENTS.

PAGE PAGE

1. INTRODUCTION . 161 | 3, Sprciric DxEscriptions—

Historical : = 161 LAMMELLIBRANCHIA—Continued.

Acknowledgments . 162 Pleuromya africana (R.

2. Discussion . 163 Etheridge, jr.) . 186

Stratigraphy : = 163 Pholadomya umzambiensis

Palaeontological Zoning . 163 sp.nov. . 187

Age and Distribution LalG5 e ef. elliptica

3. SpEciFIc DESCRIPTIONS . . 168 Minster . 188

LAMELLIBRANCHIA . 168 Goniomya umzambiensis sp.

Nucula kaffraria sp. nov. . 168 nov. . : ; . 188

Trigonoarca elongata sp. nov. 169 Cercomya arcuata (Forbes) . 189

* Be capensis (Gries- *Tiopistha (Psilomya) corru-

bach) . = LO gata Woods . 190

*Nordenskjéldia natalensis Veniella drut (Munier-

(Baily) . ‘ = 170 Chalmas) aol

Glycimeris cf. subplanata = forbesiana (Sto-

(Stoliczka) ee i! liczka) . aelo2

Pinna vanhoepeni sp. nov. . 172 Astarte 2? amapondensis sp.

» cf. decussata Goldfuss 174 NOV. =< : : 93

Ostrea sp. . . 175 Crassatellites haughtoni sp.

Alectryonia cf. arcotensis HOVs < ‘ . 193

(Stoliczka) 5 iG Anisodonta 2? umzambiensis

Exogyra decussata Goldfuss . 177 sp. nov. é . 194

Pecten (Camptonectes) kaff- Dosiniopsis geverst sp. nov.. 195

raria sp. nov. er alyas) Cyprimeria? natalensis

*Neithea quinquecostata Sp. nov. : : . 196

(Sowerby) : . 179 Cyprimeria ? (Cyclorisma ?)

Spondylus cf. calcaratus ef. analoga (Forbes) . 196

Forbes . ‘ . 182 Trigonocallista spathi sp.

Lima (Acesta) obliquistriata nov. . : . 198

Forbes . : . 183 *Macrocallista (Callistina)

Modiola typica (Forbes) . 184 euglypha (Woods) . 198

* These species have already been specifically determined from the Umzamba Beds.

160

PAGE

3. Sprciric DEScRIPTIONS—

LAMELLIBRANCHIA—continued.

Palaeomoera umzambiensis

Sp. nov. . 200

a sp. : . 200

Hs haughtoni sp.

nov. . > 201

Macoma papyracea sp. nov. 201.

Solecurtus (Azor) woodsi sp.

NOV. 3 3 . 202

Cymbophora rogersi sp. nov. 203

Mactra ? kaffraria sp.nov. . 205

Martesia? sp. . 4 . 205

GASTROPODA . 206

Patella kaffraria sp. nov. . 206

*Semisolarium bailyi (Gabb). 207

Chilodonta (Agathodonta)

africana sp.nov. . . 207

* Margarites (Solariella) radia-

tula (Forbes) . : . 208

*Nerita umzambiensis Woods 209

* ,, kaffraria Woods ~. 209

Trajanella dutoiti sp. nov. . 210

Lysis capensis sp. nov. Ss ozallil

Gyrodes tenellus Stoliczka . 212

*Turritella (Haustator) meadi

Baily . 214

Arcotia vanhoepent sp. nov. 215

*Confusiscala ornata (Baily) . 216

_ Dicroloma (Perissoptera)

bailyi (R. Etheridge, jr.). 217

Dicroloma (Perissoptera) spp. 218

Pugnellus contortus (Forbes) 218

. 220

Pirula (Protopirula) capensis

sp. nov. : . 222

(Cantharulus)

kaffraria sp. nov. . 222

umzambiensis

. 223

ERED (Mazzalina) geversi

‘sp. nov. . 224

225

Cypraea chubbi sp. nov.

Tritonidea

Siphonalia ?

sp. nov.

*Oryptorhytis rigida (Baily) .

6 pseudorigida sp.

MOVE 4s 5 274i

“sy ? sp. . 228

Annals of the South African Museum.

PAGE

3. Spreciric DEscRIPTIONS—

GasTROPODA—continued.

Paleopsephaea scalaris sp.

nov. . 228

En kaffraria, sp.

nov. . 229

Pyropsis geversi sp.nov. . 229

Semifusus (Mayeria) acuti-

carinatus sp. Nov. . 230

Semifusus (Mayeria) umzam-

biensis sp. nov. 5 . 231

Semifusus (Mayeria) kaff-

TATIUS SP. NOV. : . 231

Actaeonella (Trochactaeon)

woodsi sp. nov. . 232

Bullinella sp. . : . 233

Ringicula woodsi sp. nov. . 234

Anchistoma umzambiense sp.

nov. é F : . 235

4. APPENDIX (CRETACEOUS OF

ZULULAND) : . 236

DESCRIPTION OF THE SPECIES. 239

LAMELLIBRANCHIA . . 239

Nucula zululandensis sp.

nov. . : . 239

Trigonoarca capensis

(Griesbach) . 240

Neithea quadricostata

(Sowerby) . ; . 241

Cardium (A.) denticulatum

Baily. . 242

Veniella etheridger Newton 242

» forbesiana (Sto-

liczka) . 244

Eriphyla lenticularis

(Goldfuss) . : 245

Macrocallista (C.) anon

sont (Newton) . 246

Macrocallista (C.) zulu-

landensis sp. nov. . 247

Cyprimeria? sp... . 249

GASTROPODA . 249

Semifusus (M.) naicleners

Sp. Nov. . 249

BRACHIOPODA . 250

Cyclothyris spp. . 250

5. REFERENCES . . : . 257

* ‘These species have already been specifically determined from the Umzamba Beds.

New Lamellibranchia and Gastropoda of Pondoland. 161

1. INTRODUCTION.

Historical.—No additions have been made to our knowledge of the

Lamellibranchia and Gastropoda of the Upper Cretaceous deposits

of the coast of Pondoland since the publication in 1906 of Woods’

important monograph on the Cretaceous Fauna of Pondoland. It is

the purpose of this paper to describe the Lamellibranchia and Gastro-

poda contained in three collections made subsequently to the publi-

cation of the above monograph. The Cephalopoda of these collections

have already been described in important papers by E. C. N. van

Hoepen and L. F. Spath, and have yielded important results.

The highly fossiliferous beds near the mouth of the Umzamba were

discovered more than a century ago by H. F. Fynn, and were brought

to the notice of the scientific world by Garden in 1855 (28a).* The

collection made by the latter was described in the same year by

Baily (2), who was able to compare the fauna with those of Southern

India. A number of new species and the stratigraphy of the beds

were described in 1871 by Griesbach (25), who, on the basis of what

must now be regarded as insufficient collecting, divided the sequence

of beds into five main zones, which he believed to represent a wide

range of time in the Upper Cretaceous.

The locality was not again visited until 1901, when Rogers and

Schwarz (then members of the Cape of Good Hope Geological Com-

mission) recovered a considerable number of fossils and made obser-

vations on the stratigraphy which were included in the annual report

for that year (52). The collecting made it clear that the beds represent

only a short time-range in the Upper Cretaceous, since all the common

species were found to range through the deposit. The new material

was the occasion for a paper by Chapman in 1904 (8a) on the Fora-

minifera and Ostracoda, and in 1906 the fauna was monographed by

Woods (77), who brought the total number of molluscan species up to

about eighty, and stated that the beds were Campanian and probably

referable to one zone.

The stratigraphy of the Umzamba Beds was described by du Toit

in 1913 (14) and again in 1920 (15), and also by Plows in 1921 (49).

Three extensive collections of material from the same locality have

been made in recent years. That made by H. C. N. van Hoepen is

in the Transvaal Museum, Pretoria, and the Cephalopoda contained

in it, together with some purchased specimens, were described by

* The numbers in heavy type refer to the list of references at the end of the

paper.

162 Annals of the South African Museum.

van Hoepen in 1920 (64) and 1921 (65); a pre-Campanian age was

claimed for the fauna.

A smaller but very interesting collection was made by W. J. Plows

and E. ©. Chubb for the Durban Museum, the Cephalopoda of which

were described by Spath in 1921 (57). An important work on the

Cretaceous Cephalopoda of Zululand, in which frequent references

are made to the Umzamba fauna, was published by the same author

in 1921 (58), and was followed by a review of the Ammonoidea of

Pondoland, including the description of new species contained in a

collection presented to the British Museum by the Natal Government

in 1894 (59). Spath’s researches led him to believe that both

Campanian and Maestrichtian Ammonoidea occur in the Pondoland

fauna and that the Umzamba Beds could and should be zoned (59,

pa LG):

To investigate the possibility of distinguishing palaeontological

zones in the beds was the object of a visit paid to the deposits by

Gevers in 1923, on behalf of the South African Museum, Cape Town.

The results of very careful collecting, bed by bed, are set forth in an

accompanying table.

Acknowledgments.—The material described in this paper comprises

the following: (a) a large collection obtained by Dr. EH. C. N. van

Hoepen for the Transvaal Museum, Pretoria; (6b) a collection made

for the Durban Museum by the Curator, Mr. E. C. Chubb, and by

Mr. W. J. Plows; (c) a large assemblage of species collected by

Dr. T. W. Gevers for the South African Museum, Cape Town.

The writer wishes to express his indebtedness to Dr. 8S. H. Haughton,

Honorary Keeper of the Palaeontological Collections in the South

African and Transvaal Museums, for encouragement in the under-

taking of this work, for supplying him with a list of the Ammonoidea

contained in the Gevers Collection, and for help in various ways.

The writer is particularly grateful to Dr. L. Gill, Director of the

South African Museum, for arranging for the publication of this

paper. To Mr. EH. C. Chubb and Mr. C. J. Swierstra, Curators of the

Durban and Transvaal Museums respectively, thanks are due for the

use of the material in those institutions. The writer’s thanks are due

to Dr. T. W. Gevers for permission to publish in tabulated form the

results of his careful collecting, The bulk of the work was carried

out in the Sedgwick Museum, Cambridge, and the writer wishes to

express his thanks to Mr. Henry Woods, M.A., F.R.S., for help in

various Ways.

A special note of appreciation must be expressed to Miss E. T.

New Lamellibranchia and Gastropoda of Pondoland. 163

Talbot, of Cambridge, for her enthusiasm and untiring patience in

the preparation of the plates for this and a previous paper.

The catalogue numbers of the specimens are given in the case of

those belonging to the collections of the South African and Transvaal

Museums. The Lamellibranchia, Gastropoda, and Kchinoidea of the

Gevers Collection are distributed as follows (see Table A, p. 251) :—

S.A. Mus., 8365-8389, bed 10. S.A. Mus., 8619-8623, bed 23.

8390-8453) 4. 18a. : 8624-8627 ,, 12.

Rm 6154-8488. TY . 8628-8643 ,, Ba.

Peo 848028510 6 Ola. 2 8675-8686 ,, 5A.

» 8520-8562, 4a. f 8687-8697 ,, 18.

Bee el de ee 8698-8705 ,, QI.

» 8572-8605 ,, 3. - 8706-8723 ,, 16.

8606-8616 4. 8724-8734 tre

. 8617-8618 ,, 15.

2. Discussion.

Stratigraphy.—The deposits have been described in considerable

detail by Rogers and Schwarz (52), du Toit (14, 15), and Plows (49).

It suffices here to recall that the strata consist of thin alternating

bands of sandy, calcareous clays, sometimes pebbly, and hard, shelly

limestones, which occur intermittently along a stretch of coast about

25 miles long from near the mouth of the Umtentu in Pondoland to

the mouth of the Umpenyati in Alfred County, Natal. The beds

have a slight dip seawards (about 2° to 4°), and rest unconformably

upon a planed surface of Table Mountain Sandstone. The beds are

only exposed at intervals along the coast, as ledges which only

occasionally appear above mean sea-level, and the beds are covered |

inland by blown sand. Though typical Umzamba species have been

found at the northernmost outcrop at the Umpenyati, and also in

the Umgazana outlier south of Port St. Johns, nearly all the material

hitherto described has been obtained from one locality, the so-called

Fossil Head immediately north of the Umzamba mouth. Here the

Umzamba Beds rise into a cliff over 30 feet high, above which are

bush-covered slopes. There is reason to believe that all the fossils

hitherto collected from this locality were obtained from beds ranging

from between low-tide mark to about 20 feet above mean sea-level.

Palaeontological Zoning.—Rogers and Schwarz (52, p. 43) were of

the opinion that the greater number of the species existed throughout

VOL. XXVIII, PART 2. 12

164 Annals of the South African Museum.

the whole of the time occupied by the deposition of the rocks seen at

the Fossil Head section, but that in the coarser deposits, which were

formed in shallower or more disturbed water than the finer, only

the stronger shells were preserved. The more robust species, like

Glycimeris africana, Crassatellites africanus, and Nordenskjoldia

natalensis, were found by them in both coarse and fine-grained beds,

in the former much abraded ; the more delicate shells were obtained

from the finer-grained beds alone.

All subsequent collecting has tended to strengthen the conviction

that one palaeontological zone only is represented, though the distri-

bution of the species within the beds has not hitherto been presented

in tabulated form.

The results of very careful collecting by Gevers are set forth in

Table A. The first section was made at the best exposed portion of

the cliff, the lthological bands being numbered from Bed 1 (the

lowest) to Bed 23 (the highest bed accessible from the base of the

cliff). By lowering himself over the top of the cliff at this spot,

Gevers was able to recover a considerable number of fossils from a

bed situated considerably higher than Bed 23, the intervening beds

being inaccessible ; this top bed (Bed T) is about 10 feet above Bed 23.

The second section was made some yards to the north of the first,

the bands being numbered from Bed 1a (the lowest) to Bed 214 (the

highest bed accessible from the base of the cliff). Owing to the

discontinuous nature of the lithological bands, it was not possible to

correlate the two sections with certainty.

It is evident from Table A that all the common Pondoland Lamelli-

branchia and Gastropoda have a great vertical range in the deposits.

Bed T, which is considerably higher than Bed 23 (and Bed 214),

contains essentially the same fauna as, for example, Bed 44 and Bed 3

near the base of the deposits. It should be pointed out that, as far

as can be ascertained, no fossils have hitherto been collected from a

horizon higher than Bed 23.

The evidence for the range of the Ammonoidea in the deposits is

unfortunately rather meagre. In the first place, the common Pseudo-

schloenbachia griesbacht was collected by Gevers from beds ranging

from Bed 3 to Bed 23, and the allied P. umbulazi, as well as Hauer-

ceras gardent, have almost as wide arange. The only species obtained

from the uppermost bed (Bed T) was Hoploscaphites sp., but the same

species was obtained from near the base in Bed 44. Pseudophyllites

indra was found at the base of the section by Rogers and Schwarz

(teste Woods, 77, p. 347). Spath (59, p. 116) drew attention to the

New Lamellibranchia and Gastropoda of Pondoland. 165

fact that Plows (49, p. 63) has recorded Mortoniceras soutoni from a

bed well above that containing P. indra; but M. soutoni was recorded

from the basement bed by Rogers and Schwarz (52, p. 41) and was

found there also by Gevers. Gaudryceras varicostatum was obtained

by Gevers from Bed 164; the “ Ammonites kayev,” recorded by

Griesbach, is probably identical and was obtained from his zone “ d,”’

which is considerably lower in the section.

There is as yet no evidence for assuming that more than one zone is

present in the Umzamba Beds, and it seems highly probable that with

further collecting the ranges of many of the species will be greatly

extended. It should be borne in mind that each visit to the locality

has resulted in the recovery of new species, and that a large proportion

of the species are still represented by one or two specimens only.

Moreover, it is quite evident from the various accounts that have

been given of the stratigraphy, that the lithological bands are lenti-

cular in form.and have little lateral extension. The alternation of

thin bands of different kinds of sediment suggests that the deposits

were formed under conditions of rapid current-change, and it may

well be that the whole of the Umzamba Beds (as now exposed) was

laid down in a relatively short space of time.

Age and Distribution of the Fauna.—The geographical distribution

of the Lamellibranchia and Gastropoda is set forth in Table B (p. 254),

which is in part an extension of that given by Woods (77, pp. 349, 350).

Spath (58, p. 269) has pointed out the very great differences between

the Cephalopoda of the Umzamba Beds and those of Umkwelane Hill

in Zululand, differences which appear to be due to a difference of

facies between the two deposits. As is pointed out in the appendix

to this paper, these differences are not so evident among the Lamelli-

branchia and Gastropoda, for a very large proportion of the species

from Umkwelane Hill are identical with Pondoland species. In spite

of obvious differences, due to the absence of certain common Pondo-

land species in the Zululand deposits, the faunas resemble each other

rather closely.

Common Pondoland species have been met with at Incomanini on

the Komati River in the neighbourhood of Delagoa Bay (du Toit, 16,

p- 327), but the fossils from this locality are not well known.* The

fossils from the region between the Zambesi and Sabi Rivers in the

central part of Portuguese East Africa, recovered by Teale and

described by R. B. Newton (44), appear to represent a different facies.T

* This fauna is not included in Table B.

+ “Central Mocambique ” in Table B.

166 Annals of the South African Museum.

Haughton (26) and the writer (51) have recently described a small

Bie, collection of Upper Senonian species from the coast of Angola. The

1s | h ' | fauna is of very great interest, since several species are identical with

iy Pondoland types, notably Nordenskjoldia natalensis and Cardium

denticulatum, and the writer has concluded (51, pp. 6-9) that the

i continents bordering the Southern Atlantic had become separated

\ ! by Campian times, allowing of free communication between Pondoland

and Angola round the south of the African continent. The occurrence

! of “ Atlantic > Ammonoidea in South-East Africa, and particularly at

| Umkwelane Hill, is thus accounted for.

| | Comparison has frequently been made between the Pondoland

fauna and the faunas of the Ariyalur and Valudayur Beds in Southern

| India. It suffices here to repeat Woods’ assertion (77, p. 344) that

} the resemblance between the faunas is due in the main to a similar

| assemblage of genera and the occurrence of a number of allied species.

| i 7 Although the Lamellibranchia and Gastropoda of the Umzamba Beds.

now number over 100 species, only 14 of these occur also in Southern

oe India, and some of the latter are widespread species (e.g. Nezthea

| quinquecostata, Eriphyla lenticularis). On the other hand, a very

i) | large proportion of the species are peculiar to the Umzamba Beds

‘ 1 or to South-East Africa, so much so that it seems legitimate to regard

aie | South-East Africa (Pondoland, Zululand, and possibly Portuguese

von i" | Hast Africa) as constituting a distinct Upper Cretaceous faunal

f ae province in the Indo-Pacific region.

The Upper Cretaceous deposits of Madagascar contain a fauna poor

| in Ammonoidea which resembles that of Southern India rather than

Ae that of Pondoland. The occurrence of Potamides, Cerithium, and other

| genera indicates an estuarine facies (see Cottreau, Ga).

The exact age of the Umzamba Beds relative to the Furonese

| succession is still rather doubtful. Spath (59, p. 116) accepted a

, Campanian plus Maestrichtian age for the Umzamba Beds, and

i || believed that “the many large Mortoniceras, characteristic of South

| i | Africa, are pre-Maestrichtian ” (58, p. 265). It has been pointed out

al above that there is no evidence for believing that more than one zone

is represented in Pondoland; in addition, the presence of the wide-

hi | spread Pseudophyllites indra admits of a very close correlation of the

. Umzamba Beds with deposits elsewhere in the Indo-Pacific region.

| The position has recently been summarised by Marshall (35, p. 199)

in the following terms: “In the general account of his ‘ Geology,”

however, Haug classes all those strata that contain the important

: fossils Gaudryceras kaye and Pseudophyllites indra in the Maestrichtian.

New Lamellibranchia and Gastropoda of Pondoland. 167

formation. These strata are in general those that have been classed

as Senonian or Santonian or Maestrichtian by various authors. This

point is at least apparent. All those strata that contain the two

fossils named above are in all probability of the same age; and, since

those authorities that are familiar with the species consider that they

are closely related to Gaudryceras planorbiforme and Gaudryceras

collott, it becomes equally certain that this formation is very high

in the Cretaceous succession, perhaps even in the Maestrichtian.

Included in this are the Ariyalur, Valudayur, Nanaimo, Lower Chico,

Quiriquina, Seymour Island, New Zealand, and Pondoland Upper

Cretaceous formations.”’ While the Maestrichtian age has been

accepted by Haug, van Hoepen (65, p. 45) is “inclined to regard the

Pondoland Beds as of Upper Santonian age,” and Marshall (85, p. 203)

prefers ‘‘ to take the Upper Santonian or the Lower Campanian as the

more probable European equivalent” of the New Zealand deposits.

On the whole, an Upper Campanian age for the Umzamba Beds appears

to be most acceptable.

Spath (58, pp. 264-272) has drawn attention to the admixture of

“ Atlantic’ and “‘ Pacific’? Ammonoidea in South-East Africa; and

Marshall (35, p. 201) has concluded from the absence of Mortoniceras

and other “ Atlantic ’ types in the faunas of Seymour Island (Antarc-

tica) and New Zealand, that “‘ South Africa derived its Indo-Pacific

elements directly from India along the coast of the western Indian

Ocean, and that there was no association with the countries at the

south of the Pacific Ocean.” This belief is greatly strengthened by

the occurrence of Upper Senonian species in Angola with affinities

with both North Africa and Pondoland, since it demonstrates that

the South Atlantic had come into being by Campanian times and

that there could not have been direct coastal communication between

Pondoland and Seymour Island. Moreover, Pondoland and Seymour

Island have no Lamellibranchia and Gastropoda in common. At the

same time the coastal communication between Pondoland and

Southern India may not have been as direct as Marshall seems to

imply.

The molluscan fauna of the Umzamba Beds of Pondoland now

totals over 160 species; and since new species have been found in

every collection that has so far been made, it is probable that the list

will be extended in the future. It is to be remarked that in spite of

the new collections, no Brachiopoda are yet known from the beds.

= SS

168 Annals of the South African Museum.

3. SPECIFIC DESCRIPTIONS.

LAMELLIBRANCHIA.

FAMILY NUCULIDAE.

Genus Nucuua, Lamarck.

Nucula kaffraria, sp. nov.

(Pl. XVI, figs. 2-5.)

Material.—Three well-preserved valves showing little of the interior,

two right and one left. Holotype, right valve, T. Mus., No. 1647;

paratypes, T. Mus., Nos. 1645, 1646.

Description.—Shell triangular, convex, slightly compressed anteri-

orly. Antero-dorsal border long, almost straight, very slightly

convex, postero-dorsal shorter, concave; ventral margin convex,

meeting the postero-dorsal in a sharp angle; anterior extremity a

rounded angle. Umbones prominent, curved slightly inwards and

posteriorly, placed near the posterior end. Lunule * posterior, deep,

broad, cordate, flattened, ornamented ; escutcheon anterior, elongate,

not distinct from the rest of the shell. Margins of valves coarsely

crenulated within.

Ornamentation reticulate, of regular concentric ribs with interspaces

broader than themselves, and radiating ribs. Towards the anterior

the concentric ornamentation tends to predominate, and the radial

ribs are separated by interspaces wider than themselves; towards

the posterior and up to the bounding ridge of the lunule the radial

ribs become stronger and closer together; on the lower part of the

lunule the concentric ornamentation predominates, but towards the

umbo there are a few strong, rounded, radial ribs.

No. 1645. No. 1646. No. 1647.

Length ; 5 lS) sang, 15 mm. 17 mm.

Hicightye si cen oa lene 4)

Remarks.—In form and ornamentation this beautiful little shell is

not unlike the Gault N. pectinata Sowerby (76, vol. i, p. 16), though

the latter is readily distinguished by its wider and deeper lunule, its

greater length, and by the fact that the concentric ornamentation

* For the use of the terms “lunule”’ and ‘‘escutcheon’’ in Nuculidae, see

Woods, Mon. Cret. Lamell. of England, vol. i, p. 12.

New Lamellibranchia and Gastropoda of Pondoland. 169

is definitely subordinate to the radial. The Pondoland species appears

to be more nearly related to NV. pulvillus Miller (28, vol. xxxv, p. 201,

pl. xxi, figs. 7, 8), but the antero-dorsal margin is straighter and the

apical angle smaller than in the Aachen form.

Famity ARCIDAE.

Genus TRIGONOARCA, Conrad.

Trigonoarca elongata, sp. nov.

(ETE XOVa siirese iG sais)

Material Holotype, T. Mus., No. 1635. Specimen with both

valves preserved in part only, showing the area of the left valve,

surface abraded.

Description.—Shell trapezoidal, moderately convex ; considerably

longer than high. Anterior margin convex, curving regularly and

merging into the almost straight ventral margin. Umbones small,

incurved, well in front of the middle; carina extending from the

umbones to the very produced postero-ventral angle, rounded, the

part posterior to it sloping rapidly to the posterior margin. Hinge-

area moderately long, narrow, with few ligament grooves. Orna-

mentation of numerous radial ribs separated by very narrow grooves,

crossed by numerous growth lines, the latter tending to be stronger

ventrally and behind the carina.

Length (approx.), 55 mm.; height, 30 mm.

Remarks.—This species is considerably longer than any hitherto

recorded from Southern Africa or from the deposits of Southern India,

as well as being somewhat compressed in comparison with other

members of the genus. Some specimens of TZ. trichinopolitensis

(Forbes) (62, p. 353, pl. xix, fig. 2) approach it in length, but the

Indian species seems always to be more convex than the present form,

and to possess rather larger and more prominent umbones. 7.

archiaciana (d’Orbigny) (18, p. 235, pl. cecxxii) is very similar in pro-

portions, but possesses a well-marked second carina posterior to the

major one; this is a Turonian form.

170 Annals of the South African Museum.

Trigonoarca capensis (Griesbach).

(PLL cies 1)

1871. Arca capensis, Griesbach: Q.J.G.S., vol. xxvii, p. 66, pl. ii,

fig. 10. }

1906. Trogonoarca capensis, Woods: Ann. 8. Afr. Mus., vol. iv,

p. 288, pl. xxxiv, figssieee

Remarks.—The species has been well described by Woods, but the

interior has not been figured. The Transvaal Museum collection

contains an exceptionally fine specimen, No. 216, which shows the

interior. The anterior adductor impression is large, subcircular, and

placed close to the anterior end of the hinge. The posterior adductor

impression is large and placed on a raised platform. Pallial line entire,

with elongated muscle pits on the inner side. The interior should be

compared with that of 7. angolensis Rennie from Angola, described

in a recent paper in these Annals (51).

Genus NoRDENSKJOLDIA, Wilckens.

Nordenskjéldia natalensis (Baily).

1855. Arca natalensis, Baily: Q.J.G.S., vol. xi, p. 461, pl. xin,

MEN, 2,

1904. Latiarca (%) natalensis, Etheridge: Second Rep. Geol. Surv.

Natal and Zululand, p. 77, pl. 1,

figs. 10-12.

1906. Nemodon natalensis, Woods: Ann. S. Afr. Mus., vol. iv,

pe 209, ple xxv. tesa le

1929. Nemodon natalensis, Rennie: Ann. 8. Afr. Mus., vol. xxviii,

p. £2) qollaing dies

Remarks.—The writer is indebted to Dr. 8. H. Haughton for calling

-his attention to the resemblance between this well-known Pondoland

species and Wilckens’ genus Nordenskjéldia (73, pp. 26-30), which

was formed to receive the two Indian species Arca disparilis d’ Orbigny

—the genotype—and Arca japetica Forbes (62, pp. 350-352), and a

species from Snow Hill and Seymour Island, Antarctica, N. norden-

skjoldi Wilckens (73, p. 26, pl. u, figs. 8-11). Nordenskjéldia is

probably allied to Nemodon Conrad, and is separated from it chiefly

by reason of its stouter, more inflated, less elongated valves, much

larger area, usually coarser radial ornamentation and marginal crenu-

New Lamellibranchia and Gastropoda of Pondoland. iit

lations. The hinges of the genera are very similar, but the lateral

teeth of Nemodon are longer and thread-like. For good figures of the

genotype of Nemodon, N. eufaulensis (Gabb), and other species, see

Wade (67, pp. 42, 43, pls. viii, 1x).

The genus Nordenskjéldia is now known from the Ariyalur and

Trichinopoli Groups of Southern India, South-East Africa, Southern

Angola, and Antarctica. Nemodon appears to be confined to the

eastern part of North America.

The Pondoland species was apparently overlooked by Wilckens

when he described the genus.

Genus GLYCIMERIS, da Costa.

(= Pectunculus, Lamarck.)

Glycimeris cf. subplanata (Stoliczka).

(Pl. XVI, figs. 8-11.)

Material.—Two specimens, 8.A. Mus., Nos. 8415, 8515.

Description.—Shell rather small, sub-orbicular, compressed, flattened

towards the margins, longer than high, the greatest length at or a

little above the middle. Umbones prominent. Margins rounded,

the hinge margin tending to be straight for a short distance with feeble

angulations at either end. Ligamental area very small, narrow, and

rapidly descending. Surface with broad, flattened or slightly rounded,

radiating ribs, separated by very narrow grooves, and with very faint

traces of concentric ornament.

Remarks.—The description given above applies to the larger

specimen, No. 8415, which differs from the common Pondoland species,

G. africana (Griesbach) (77, p. 291, pl. xxxiv, figs. 8-12), in several

respects. In the latter the shell is in all full-grown examples higher

than long and the greatest length is below the middle, the anterior

and posterior margins tend to slope towards the umbones, and the shell

is stouter and more inflated, while the ribs tend to be narrow and on

the posterior side unequal in size. The general aspect of the former

suggests a comparison with Stoliczka’s Axinaea subplanata (62, p. 347,

pl. xvu, figs. 28-30; pl. xlix, fig. 10), and also with the figures given

by that author of Forbes’ A. subauriculata (62, p. 349, pl. xvii, figs.

31, 32), but not with Forbes’ originals (20, p. 150, pl. xvii, fig. 13) ;

but the figures given in the Pal. Indica are unfortunately lacking in

clearness, and the details of the ornamentation can only be guessed

at from the somewhat meagre description. G. subplanata seems to be

172 Annals of the South African Museum.

most closely allied to the present species by reason of its rather com-

pressed valves, its length, and the narrowness of the liigamental area,

but the identification can be made certain only by an actual com-

parison of material from the two areas.

The shells described by Etheridge as A. subauriculata (Forbes) ?

from the Umsinene River deposit, Zululand, are similar in outline

but apparently more inflated (18, p. 75, pl. 1, figs. 11, 12).

The ornamentation of G. africana appears to vary considerably, a

phenomenon which may be due in part to variation in the state of

preservation, and Woods has stated that young examples are more

orbicular, as long as high, and with the greatest length about the

middle; hence the separation of young examples of the two Pondo-

land species would be a difficult matter. The smaller of the two

specimens, No. 8515, though it agrees with the larger in outline and

poor convexity, and should be referred with it to G. subplanata, shows

a distinct resemblance to Woods’ figure of a small example of

G. africana (op. cit., fig. 12), not only in shape but in the nature of

the radial ornamentation.

qf Famity PINNIDAE.

rH .

ee FI Genus Pinna, Linnaeus.

Pinna vanhoepent, sp. nov.

(Pl. XVIII, fig. 9; text-fig. 1.)

Se See

Material.—The holotype, T. Mus., No. 1630, is a fine specimen

with the valves closed, the anterior and posterior regions missing.

Two fragments in the collection of the Durban Museum have also

been examined.

\a wa Description.—Shell large, straight, well inflated, elongated. Margins

vo Ke nearly straight, the dorsal very slightly concave. Section quad-

il rangular. Valves not fissured along the middle line, but each divided

he near the middle into two parts which meet at an angle. Dorsal part

sik | of each valve flattened and ornamented towards the umbones with

| about 7 rounded ribs, which are increased by intercalation to about

14, the secondary ribs remaining weaker than the primary. Ventral

| part of each valve slightly convex, becoming more flattened posteriorly,

| its dorsal portion ornamented with about 4 primary ribs, its median

] portion with about 4 weak secondary ribs. Ribs on all parts of

ee ed

-— = >

wo 8 ig

| the valve separated by broad, slightly concave interspaces ; ribs and

New Lamellibranchia and Gastropoda of Pondoland. 173

interspaces crossed by fine, rather irregular, concentric ridges, which

are closer together posteriorly, and on crossing the median secondary

TEXxtT-FIG. 1.—Pinna vanhoepeni, sp. nov. Cross-section to

show degree of inflation.

ribs of the ventral part of the valve, curve towards the umbones and

are continued to the ventral margin, which they meet at an acute

angle. )

Remarks.—This species resembles P. laticostata Stoliczka (62, p. 385,

pl. xxv, figs. 2,3; pl. xxvi, fig. 4) from the Ariyalur Group of India

in proportions and absence of the median fissure, and in the nature

and disposition of the concentric ridges, but is at once separated from

it by the intercalated secondary ribs, a feature which is seen also in

the dorsal portions of the valves of P. arata from the Trichinopoli

Group.

i 174 Annals of the South African Museum.

) : i ! Pinna cf. decussata, Goldtfuss.

Hii | | (Pl. XIX fien9); text-figs: 2)°3.)

ME iba) | j :

Ni iT Material—One more or less complete specimen, considerably

Ve | damaged posteriorly, with the surface of the left valve in a moderately

Bi |

\| i |

| f |

tii |

H Val

Vai

| fi

2 ft | 2

| TExtT-FIG. 2.—Pinna cf. decussata, TExtT-FIG. 3.—Pinna cf. decussata,

ii || Goldfuss. Cross-section to show Goldfuss. Right valve. x #.

Ti degree of inflation.

good state of preservation, S.A. Mus., No. 8692, and a fragment of

| the dorsal part of another individual, S.A. Mus., No. 8443.

i Descrvption.—Shell large, straight, elongated, not well inflated ;

| , margins nearly straight, the ventral rather convex; shell flattened

He \ i posteriorly, apical angle small (35°). Valves not medially carinated

>» ia '

‘ ser ae . }

a = Bile —- a — or

=== === Se SS ee er a ;

-— - a = = sage 7

+ SS

New Lamellibranchia and Gastropoda of Pondoland. 175

but gently rounded, and fissured for their greater length. Dorsal

part of the valve flattened, ornamented with about 7 ribs, the dis-

tance between which increases very gradually in passing towards the

posterior. Dorsal portion of the ventral part of the valve ornamented

with about 6 ribs, closer together than in the dorsal part of the

valve, and diminishing in strength ventrally. Ventral portion of the

ventral part of the valve ornamented with strong, curved growth lines

and folds, which meet the last rib at an acute angle and do not appear

to be continued across the ribbed portion of the valve—but this last

character may be due to a certain degree of exfoliation. Shell thin.

Remarks.—This shell belongs to the group of P. decussata Goldfuss

(24, vol. 1, p. 166, pl. cxxviii, figs. 1, 2; and 76, vol. ui, p. 99, pl. xiii,

figs. 4-6; pl. xiv, fig. 1) and P. cretacea (Schlotheim) ; the latter is a

more elongated species with the ventral lines and folds less distinctly

curved, as Woods has remarked (76, p. 101). The present specimen

is more widely triangular, 2.e. has a larger apical angle, than the

forms figured by Goldfuss, but agrees with it in the character of the

ribs and degree of inflation ; the concentric ornamentation is, however,

not evident on the dorsal part of the Pondoland shell.

Stoliczka (62, p. 384) and Wilckens (72, p. 124) have indicated the

relationship of P. arata Forbes (62, p. 384, pl. xxiv, fig. 5; pl. xxv,

fig. 1; pl. xxvi, fig. 5) to P. decussata, the Indian species differing

only in the presence of secondary ribs on the dorsal portions of the

valves. Wilckens (72, p. 123, pl. v, fig. 2) has described a closely

related species, P. morenoi, from the Upper Cretaceous of South

Patagonia, which appears to be very close to the present species, but

has a smaller apical angle.

Famity OSTREIDAE.

Genus Ostrea, Linnaeus.

Ostrea sp.

(Pl. XVII, figs. 1, 2.)

Material.—One large left valve, T. Mus., No. 1571. Probably also

T. Mus., No. 1564, and two specimens in the collection of the South

African Museum, Nos. 8555 and 8557.

Deseription.—Shell large, subcircular in outline, rather longer than

high, compressed, more convex towards the centre, with a flattened

bounding region. Exterior with strong growth markings at irregular

176 Annals of the South African Museum.

intervals, and where well preserved with faint growth striae, and with

obscure, discontinuous, radial plications giving the surface a roughened

appearance; umbo pointed, small. Interior with a central, moder-

ately concave region, higher than long, slightly produced postero-

ventrally, separated from the border by an obscure linear groove ;

this is bounded by a broad, flattened shelf, which commences on

either side of the area as large, lateral, wing-like expansions, particu-

larly wide on the posterior side, and is continued in rather narrower

form round the ventral margin; near the umbo the shelf is covered

with small, irregular ridges or crenulations. Area small, triangular,

and ligament pit triangular, shallow. Muscular impression subcentral

(relatively to the entire shell), but somewhat posterior (relatively to

the central concave portion), subcircular.

The remaining specimens are of younger forms, and indicate that

the juvenile shell was circular, smooth in outline, and without the

flattened border.

Remarks.—This oyster resembles in form O. nummus Coquand

(5, p. 136, pl. xliv, figs. 10-12) from the Cenomanian of France, a

species in which the flattened bounding region is relatively narrower,

the outline more nearly circular, and the surface smoother than in

the present form. The Santonian, O. licheniformis, figured by Coquand

(5, p. 91, pl. xxxvu, figs. 17-19) from France is not unlike the present —

form in some respects; but there again the bounding region is rela-

tively narrower, especially in the upper part; this is regarded by

Woods (76, vol. 11, p. 381) as a form of O. semzplana J. de C. Sowerby,

in which the radial folds are usually more pronounced.

Genus ALECTRYONIA, Fischer de Waldheim.

Alectryonia cf. arcotensis (Stoliczka).

(Pl. XVII, figs. 5-7.)

Material.—Two left valves, 8.A. Mus., Nos. 8437, 8439.

Remarks.—These shells appear to be close to the forms described

from the Ariyalur Group by Stoliczka as Ostrea (Alectryonia) arco-

tensis (62, p. 471, pl. xlii, figs. 3-7). The shells are small, subovate,

slightly bent, rather convex; the surface is ornamented with several

stout, subequal, rounded, radial ribs, which are separated by deep,

rounded furrows; the margin is dentate; growth lines also occur ;

the adductor scar is large, bean-shaped, and placed towards the

New Lamellibranchia and Gastropoda of Pondoland. 177

posterior border. The left valves figured by Stoliczka are provided

with fewer and stouter ribs, and one of the Pondoland examples is

more convex than any of the Indian forms, hence the hesitation in

identifying the forms from the two regions.

A rather similar but larger specimen has been compared with the

Indian forms by Wilckens (72, p. 125, pl. v, fig. 4); it is from the

Upper Cretaceous of South Patagonia.

Genus Exoeyra, Say.

Exogyra decussata, Goldfuss.

(Pl. XVII, figs. 3, 4.)

1834. Exogyra decussata, Goldfuss: Petref. Germ., p. 35, pl. Ixxxvi,

moot (s

1869. Ostrea decussata, Coquand: Mon. Genre Ostrea, p. 30, pl. vu,

(with syn.).

Material.—Two left valves, S.A. Mus., Nos. 8562, 8578.

Description.—Left valve convex, elongate-oval in outline, inequi-

lateral, margins rounded. Umbo curved inwards and backwards,

distorted by a circular area of attachment. Strong carina extending

in a curve from the umbo to the postero-ventral extremity, dividing

the shell into a more or less flattened posterior region and a slightly

rounded anterior. Surface with coarse growth markings, and near

the carina with irregular radial folds or ridges.

Remarks.—The writer does not hesitate to place these specimens

in the species created by Goldfuss, a species which is widespread in

the Upper Senonian of Europe and has also been found in N. Africa.

The high carina and radial ribs are characteristic. The species does

not appear to have been found in India, though a similar species,

E. ostracina (Lamarck), is common in the Ariyalur Group (62, p. 459,

pl. xxxv, figs. 6-12; pl. xxxvi, figs. 1-4).

| 178 Annals of the South African Museum. .

| FAMILY PECTIN IDAE.

| Genus PrctEen, Miiller.

Sub-genus CAMPTONECTES, Meek.

(Pl. XVI, figs. 12-15.)

| Pecten (Camptonectes) kaffraria, sp. nov.

1906. P. (Camptonectes) sp.,. Woods: Ann. 8. Afr. Mus., vol. iv,

| | p. 297, pl. xxxv, figs: 125g:

Ain) | | Materval—Six specimens, T. Mus., Nos. 1598-1603. Types,

pai | Be i Nos. 1601 (left valve) and 1602 (right valve).

! Description (after Woods).—Right valve ovate, slightly inequi-

lateral, with the dorsal third pointed, and the margin of the remain-

| ing part forming a regular curve. Antero-dorsal margin rather

i) longer than the postero-dorsal.

\ | Ornamentation consists of numerous fine radial furrows. Near the

Hl median part of the valve the furrows are nearly straight and often

iil discontinuous ; they are separated by broad, flat interspaces on which

fine concentric lines may occur. On the anterior and posterior parts

of the valve the furrows become more curved, deeper, and more con-

tinuous, and the interspaces become narrower, more raised, and rib-

like. The furrows do not show punctae.

Anterior ear large, with its anterior margin nearly at right angles

to the hinge line, ornamented with fine ridges parallel with the

anterior margin ; byssal sinus of moderate size. Posterior ear smaller

than the anterior, with its outer angle obtuse, ornamented with fine

radial ribs and growth-ridges.

Left valve with ornamentation similar to, but rather coarser than,

that on the right valve. Anterior ear large, with its outer angle nearly

rectangular. Posterior ear smaller than the anterior, with its outer

angle obtuse.

Length, 17 mm.; height, 19 mm. |

There is nothing to add to this description. Most of the specimens

are considerably larger than that described by Woods. The measure-

ments of No. 1601, with the ventral margin restored, are: length,

34 mm.; height, 35 mm. No. 1600 would, when complete, have

approximately the following dimensions: length, 54 mm.; height,

57 mm.

Remarks.—Woods states that the specimens examined by him

New Lamellibranchia and Gastropoda of Pondoland. 179

belong to the same type as P. virgatus Nilsson, but appear to differ in

having finer ornamentation, and resemble most nearly the example

figured by Romer from Texas as P. virgatus. With more material

from Pondoland at hand, it is possible to state that the ornamentation

is very much finer than in the Swedish species, and that it is extremely

unlikely that the present examples belong to the same species as the

small type figured by Nilsson (45, p. 22, pl. ix, fig. 15) and Hisinger

(27, p. 52, pl. xvu, fig. 3). In the case of the shell figured by Romer

from Texas (53, p. 66, pl. vii, fig. 5), the ornamentation is certainly

coarser than in any of the present specimens, and, moreover, there

appears to be a more pronounced system of fine concentric lines.

Under the circumstances it 1s desirable to have a new specific name

for the Pondoland forms.

The absence of punctae in the furrows, combined with the exceed-

ingly fine ornamentation, separates the shells at once from P. striato-

punctatus Romer and P. curvatus Geinitz (76, vol. i, pp. 157-162).

Genus NeirHea, Drouet.

Neithea quinquecostata (Sowerby).

(Pl. XVITI, figs. 1-5.)

1814. Pecten quinquecostata, J. Sowerby: Min. Conch., vol.i, p. 122,

pl. lvi, figs. 4-8.

1903. ,, (Neithea) quinquecostatus, Woods: Cret. Lam. Eng-

land, vol. i, p. 202, pl. xxxix,

figs. 14-17; pl. xl, figs. 1-5

(with full synonymy).

21904. Neithea sp., R. Etheridge, jr.: Second Rep. Geol. Surv. Natal

and Zululand, p. 73, pl. ii,

fig. 13.

1906. Pecten (Neithea) quinquecostatus, Woods: Ann. S. Afr. Mus.,

WOls) ii jos “Nepal Wec-oraya

fig. 14.

1907. Neithea quinquecostata, R. Etheridge, jr.: Third Rep. Geol.

Surv. Natal and Zululand,

pp. (0; pla iv, fies. 8,9:

1909. , 4 R. Bullen Newton: Trans. Roy. Soc.

Sree Volo i (pts tp. 58,

pl. ii, figs. 1, 2.

Remarks.—This ubiquitous species has been recorded from the

VOL. XXVIII, PART 2. 13

et ——_ ~ -- = — —

ae is Fg

PS - — a =

a - 3 E> p py a J J a: : — = ; ——

— — — i x Pd ‘ : : } =

3 ~< Sia Fs 3S m 3 n = ate _ = = a ee z za Ss

} aa Ga = J ——— — ——— = aes - 2 : > ~ - - -

; = - ——= ~~ 23 = y. 3B ¥ 7 7 - ee ———. —— — — : = ~~ anne

st = ¥ ~ = — : — - = —_ ‘ BP Pe > ren —— ——— a :

Ft a a ee eer = eee = —_ bs _ 3 : = =a S + — a ~-

—_—_—_—_ _—__ Se Se . F as . —— z Se —e ee = “Sai =

ea ieaiiiat =— ‘ x = Se —— y ze eee a wis = SS

rr =

180 Annals of the South African Museum.

Umzamba Beds in Pondoland, and in Zululand from the Manuan

Creek area, the Umsinene deposit, and Umkwelane Hill; the South

African Museum possesses a damaged right valve from Angola which

has recently been described in these Annals by the writer (51).

Woods figured only the exterior of a right valve from Pondoland.

The collections made by Dr. Gevers includes several well-preserved

specimens of both valves, including a small right valve (No. 8420),

in which the concentric striations are preserved with exceptional

clearness. The writer has also had the opportunity of examining a

larger right valve, in which the rather remarkable structure of the

hinge-plate is seen to perfection; this beautiful specimen is in the

collection of the Durban Museum.

Hinge-plate in Netthea.—The following is a description of the hinge-

plate of the Durban specimen. Ears subequal, triangular; dorsal

edges of the ears straight, horizontal, and somewhat thickened ;

lateral edges straight, vertical, thin, and sharp. Hinge-line provided

with a row of well-developed transverse denticles, which are in appear-

ance remarkably like the teeth of a taxodent shell. Ears slightly

concave and smooth below the hinge-line ; the posterior ear provided

with two ridges on the ventral part, placed obliquely, and subparallel

to one another, the lower one being the distal end of the auricular

crura. Resilium pit deep, triangular, with the umbo arched over its

dorsally pointed end. On each side of the pit is a very stout, elevated,

laterally compressed, elongated, tooth-like process; the ventral end

of each tooth is thickened and projects below the margin of the hinge-

plate ; each tooth thins out and becomes less elevated dorsally, but-

the ends are curved round immediately beneath the umbo into two

small, projecting, tooth-like tubercles.

A specimen of a left valve in the collection of the South African

Museum (No. 8380) shows two small plate-like outgrowths which

project upwards from the hinge-line on either side of a median notch ;

immediately below each plate is an oblique slit. When the valves

are united the notch evidently receives the tip of the highly incurved

right umbo, and the plate-like outgrowths fit into the sockets above

the large tooth-like processes of the right valve, the latter fitting into

the slits below the plates on the left valve.

The interior of the right valve of Nezthea has seldom been figured.

The only figures seen by the writer in which the interior is clearly

shown are included among the unusually fine figures given recently

by Wade (67, p. 64, pl. xxi, figs. 1-5) of specimens from the Upper

Cretaceous of Tennessee. These are quite typical examples of

New Lamellibranchia and Gastropoda of Pondoland. 181

N. quadricostata (Sowerby), but are erroneously referred to N. quin-

quecostata (Sow.), which is placed under Pecten without reference to

the Neithea group. The figures demonstrate fairly clearly in both

valves the structures described above for the allied N. quinquecostata,

though the resilium pit of the right valve is hidden by the arching

of the umbo.

Dall (11, p. 690) makes the following observations on the hinge-plate

of Pecten: “ The ends of the resilium are received by subtriangular

or oval pits in the umbonal region. These pits may be shallow or

deep; their basal margin sometimes projects slightly into the cavity

of the valves ; their apex is always nearly coincident with the umbonal

point of the valve. In a few species, in the right valve, the lateral

margins of the pit are raised into tooth-like processes, which fit into

corresponding depressions in the opposite valve (e.g. P. swiftw Bernh.),

but these are not homologous with the so-called teeth of Plicatula and

Spondylus. Outside of these, radiating fan-like from the apex of the

valve, are frequently found one to three pairs of more or less prominent

laminae, which I call the cardinal crura, and further away and below,

on the ridges which mark the lower boundary of the ears, will some-

times be found another pair, only distally conspicuous, which I have

named the auricular crura. The cardinal crura are most conspicuous

in heavy shells, especially such as Pecten proper and Lyropecten, and

serve to adjust the closing of the valves, as does the hinge armature

of the Teleodonts. In a few species the crura are sufficiently prom-

inent to actually interlock with the valves half open ; in many others

hardly any trace of them is visible.”

Excellent figures of various species of Pecten showing the cardinal

crura (‘cardinal ribs’’) and auricular crura have been given by

Verrill (68, p. 51, pl. xix, figs. 6-9).

Concerning Spondylus, Dall (11, p. 494) makes the following state-

ment: “ Itis through a study of the Pectens and very young Spondyla

that I have been able to satisfy myself that the so-called teeth of

Spondylus (and of some Pectiniform relatives) are not homologous

with either Prionodont or Teledont teeth, but are a modification of

certain ridges which reinforce the auricles in many Pectens and which

I named, in 1886, the auricular crura. I have been able to trace this

modification of the crura to a point where it absolutely harmonises

with the incipient teeth of Spondylus at an age when the original

arca-like teeth of Pecten and Spondylus still occupy the cardinal

margin.”

In the opinion of the writer, the large tooth-like processes which

182 Annals of the South African Museum.

flank the resilium pit in both N. quinquecostata and N. quadricostata

are modifications of the dorsal ends of the auricular crura. They are

therefore homologous with the crural teeth of Spondylus and Plicatula.

If the tooth-like flanking processes quoted by Dall in the case of

P. swiftic are of the same nature as those seen in Neithea, then the

writer is unable to understand the distinction between them and the

crural teeth of Spondylus. Dall states that P. swifti is close to

Pecten (Chlamys) jparmeleei Dall from the American Pliocene (11,

p. 708), so it is evidently rather different from Nevthea.

The genus Neithea differs in several points from the other Pectens,

notably in the development of crural teeth and the retention of well-

developed denticles on the hinge-line. The auricular crura of the

true Pectens cannot be called tooth-like, except perhaps in the case

of P. swift, and well-developed denticles on the hinge-line occur

only in juvenile forms, being obsolete or quite absent in the adult.

The strong arching of the nght valve, the ornamentation, and the

remarkably efficient articulation, make this a very distinct genus.

It is concluded that Nezthea should be placed somewhat apart from

Pecten and its numerous subdivisions (Chlamys, Camptonectes, etc.).

The striking development of the crural teeth shows a remarkable

parallel with the Spondylidae. It may be remarked, however, that

the type of Nezthea is N. aequicostata (Lamarck), the hinge of which

does not appear to have been seen. This view is a departure from the

opinions of former workers; Fischer (19, p. 946), Verrill (68, p. 60),

and Dall (81, p. 457) have regarded Nezthea as a section of Pecten

(sensu stricto), the type of which is P. maximus Linnaeus.

Famity SPONDYLIDAE.

Genus Sponpytus, Linneus.

Spondylus cf. calcaratus, Forbes.

(Pl. XX, figs. 5-7.)

Cf. 1846. S. calcaratus, Forbes: Trans. Geol. Soc. Lond., vol. vu,

p. 155, pl. xvii, fig. 2.

Cf. 1846. S. subsquamosus, Forbes: Ibid., p. 154, pl. xvi, fig. 1.

Cf. 1871. S. calcaratus, Stoliezka: Cret. Pel. 8. India, p. 448, pl.

0.0.04 HI Gee (5 8)5, 10)

Cf. 1906. z Boule and Thevenin: Ann. de Pal., vol. i,

p- 49.

New Lamellibranchia and Gastropoda of Pondoland. 183

Material.—Three large upper (left) valves in moderately good

condition and showing the hinge-line, but somewhat worn externally,

and a single right valve affixed to one of these; T. Mus., Nos. 1567-69,

also, doubtfully, T. Mus., No. 1566.

Description.—Left valve large, thick, moderately inflated, ovate,

higher than long, very slightly inequilateral. More or less pointed

in the umbonal region, ventral and lateral margins well rounded.

Surface irregular, undulating. Ears small, worn in the specimens,

the posterior slightly larger, separated by a deep fold and rounded

ridge from the rest of the shell, with fine radial ornamentation.

Smooth or irregularly toothed within the central margin. Orna-

mentation of very numerous radiating ribs, rounded or laterally

compressed, separated by deep grooves which may be narrower than

or as wide as the ribs, both ribs and grooves being crossed by numerous

growth striae which give to the ribs a faintly rugose appearance.

Hinge stout, area triangular, with a deep median ligament groove.

Right valve attached by most of its surface, free parts with poorly

developed radial ribs and strong growth striae. Inner surface with

radial grooves corresponding to the external ribs, and with toothed

ventral margin. Umbonal part much produced and strongly curved

anteriorly, with large pointed area.

Remarks.—S. calcaratus seems to be a very variable species. The

specimens differ considerably in the degree of convexity of the left

valve and spacing of the radial ribs, and none of them show the presence

of a few stronger ribs and the scale-like tubercles seen in most (but

not all) of the Indian forms described by Forbes and Stoliczka from

the Trichinopoli Group. Nevertheless the Pondoland shells should

be referred at least tentatively to the Indian species, with which they

agree in proportions, nature of the ribbing, and the curious twisting

of the umbonal regions of the fixed valve. The species seems to be

very close to the European 8S. truncatus Goldfuss (18, p. 666, pl.

eccclix).

Famity LIMIDAE.

Genus Lima, Bruguiére.

Sub-genus acesTA, H. and A. Adams.

Lima (Acesta) obliquistriata, Forbes.

(Pl. XIX, figs. 7, 8.)

1846. Lima obliqui-striata, Forbes: Trans. Geol. Soc. Lond., vol.

vii, p. 154, pl. xviii, fig. 13.

iid

184 Annals of the South African Museum.

1871. Radula (Acesta) obliqui-striata, Stoliczka: Cret. Pel. 8. India,

. p. 421, pl. xxx, figs. 2-5, 13.

Material—An incomplete left valve, T. Mus., No. 1631. The

Durban Museum possesses two specimens, somewhat worn.

Description.—Shell compressed, considerably higher than long,

moderately oblique. Anterior margin straight, ventral well rounded ;

postero-dorsal and postero-ventral margins moderately convex, and

meeting in a rounded angle above the middle. Umbones pointed,

small, placed close to the anterior margin. Apical angle about 70°.

Anterior ear small, narrow, separated from the shell by a deep de-

pression ; posterior larger, oblique, well marked off from the rest of

the shell. Anterior area very narrow, depressed, elongate, limited

by a sharp edge from the rest of the shell.

Ornamentation of numerous rounded ribs, somewhat flattened,

separated by grooves narrower than themselves; the ribs increase

in size from the umbo ventrally without increase in number. Surface

with concentric growth lines, which cross ribs and grooves alike, on

the posterior side somewhat obliquely; and with more prominent

growth ridges at irregular intervals.

Length, 32 mm.; height, 50 mm.

Remarks.—This species seems to be identical with that from the

Ariyalur Group described by Forbes and Stoliczka, with which it

agrees in shape and ornamentation. Stoliczka states that the ribs

when well preserved are ornamented with numerous spinulose scales ;

in the Pondoland specimens the best-preserved ribs have a somewhat

scaly appearance due to the crossing of the growth lines. It resembles

most L. rauliniana and L. ornata of d’Orbigny, as remarked by

Stoliczka, but is distinguished from the former by having fewer and

more closely set ribs, and from the latter by its more closely set ribs

and finer scales.

Famity MYTILIDAE.

Genus Moprota, Lamarck.

Modiola typica (Forbes).

(Pl. XTX, figs. 1-6.)

1846. Mytilus (Modiolus) typicus, Forbes: Trans. Geol. Soc. Lond.,

vol. vil, p- 152; plexxive

fig. 4.

New Lamellibranchia and Gastropoda of Pondoland. 185

1866. Modiola typica, Zittel: Biv. der Gosaugeb., p. 78, pl. xi,

fig. 5.

Hew e Stoliczka: Cret. Pel. S. India, p. 377, pl.

xxl, figs. 12-15.

Material.—Two fine specimens, T. Mus., Nos. 1595, 1596; three

small forms, apparently young specimens of the same species, T. Mus.,

Nos. 1597, 1598. Also South African Museum, Nos. 8535, 8619,

8675.

Description.—Shell elongate-oval, curved; anterior and posterior

rounded ; ventral margin gently concave. Shell well inflated, the

part of greatest convexity represented by a rounded ridge extending

from the umbones to the postero-ventral extremity, curved, flanked

on the anterior side by a slight depression. Dorsal margin straight.

Umbones obtuse, touching, anterior. Anterior margin produced

slightly beyond the umbones. Ornamentation of sharp, somewhat

irregular, concentric ribs and faint concentric striae, the ribs being

stronger on and posterior to the ridge of greatest convexity. A bundle

of faint radial striae extends from the umbones postero-ventrally along

the anterior flank of the ridge.

Remarks.—The above description is of the large specimens. There

are in addition three small specimens which appear to be young forms

of the same species. These show some variability in shape and

strength of the concentric ornamentation and do not appear to possess

the fine radial striae towards the middle seen in the typical specimens,

but do not possess characters sufficiently marked to lead one to regard

them as a different species. They are shorter and less inflated towards

the umbones than WM. kaffraria Woods (77, p. 294, pl. xxxv, fig. 5)

from the same deposits.

M. reversa Sowerby (76, vol. i, p. 94, pl. xv, figs. 15-18; pl. xvi,

figs. 1-3) is closely allied, but the ridge of greatest convexity appears

to be more marked and less gently rounded than in M. typica, while

its radiating striae are confined to the shallow sulcus anterior to the

ridge, whereas in M. typica they are situated on the anterior slope of

the ridge.

;

186 Annals of the South African Museum. |

Famity PLEUROMYIDAE.

Genus PLeuromya, Agassiz.

Pleuromya africana (R. Etheridge, jr.).

(Pl. XX, figs. 8-10.)

1907. Myopsis (%) africana, R. Etheridge, jr.: Third Rep. Geol.

Surv. Natal and

Zululand, p. 81,

pl. 1, figs. 9,0:

1909. Pleuromya africana, R. Bullen Newton: Trans. Roy. Soe. S.

Afr., vol. 1; ptaieape

84, pl. vin, figs. 1-3.

Material—Two well-preserved specimens with the valves closed,

T. Mus., Nos. 1657, 1658.

Description.—Shell elongate-oblong, inequilateral, convex. Umbones

prominent and incurved. Anterior margin rounded, ventral margin

convex, the posterior region produced and compressed with rounded

margin. Greatest convexity in front of the umbones; an obscure

carina extends from the umbones to the antero-ventral margin,

cutting off a rapidly sloping antero-dorsal part. Shell thin, with

small anterior and posterior gapes. One specimen has a large part

of the shell removed and shows the large pallial sinus, very broadly

rounded, and the relatively small, rounded, posterior adductor im-

pression. The dorsal margins posterior to the umbones differ in the

two valves, the right overlapping the left, the left being modified for

that purpose by being flattened or bevelled along the margin.

Ornamentation of concentric sulcations and rounded ridges, the

grooves being rather narrower than the ridges, and of irregularly

distributed growth lines, the concentric ornamentation tending to

be finer near the umbones. The entire surface of the shell is covered

with minute granulations, scattered for the most part in haphazard

manner, but with a tendency to arrangement in radial lines. At the

posterior end there are in some specimens minute, irregular, radial

riblets or wrinklings. |

The measurements of the more complete specimen, which agree

with those given by Bullen Newton, are as follows :—

Length . . 45 mm.

Height. : : a et abeee

Thickness . , : Pa AA ae

New Lamellibranchia and Gastropoda of Pondoland. 187

Remarks.—This species was referred by Bullen Newton to Pleuromya,

and compared with Pholadomya neoconnensis Leymerie and Myopsis

untoides Agassiz, both of which are regarded as synonymous with

Panopea gurgitis (Brogniart) by Woods (76, vol. ii, p. 222). The

resemblance of the present species to some forms of P. gurgitis is

based on superficial characters, shape, concentric corrugations, and

on the presence of the minute granulations, and not on the characters

of the hinge, which has not been seen in the present species. The

overlapping of the dorsal margin of the left by that of the right valve

at once separates P. africana from P. gurgitis, and strongly supports

Newton’s placing of the species in Plewromya. The type is from the

Umsinene deposit (Zululand), and the species occurs also at Umkwelane

Hill and in the Manuan Creek area.

A similar species is Panopea ? (Pleuromya %) clausa Wilckens from

Antarctica (73, p. 68, pl. ii, fig. 10).

Famity PHOLADOMYIDAE.

Genus PHoLtapomya, G. B. Sowerby.

Pholadomya umzambiensis, sp. nov.

(RIS SOXe tras 3 4.)

Material.—The species is founded on a single specimen with the

valves closed ; though somewhat weathered and in the form of a cast,

the form and ornamentation are indicated sufficiently well to justify

its description as a new species. S.A. Mus., No. 8475.

Description.—Shell small, elongate, medially inflated, posteriorly

expanded and compressed. Umbones small, moderately prominent,

incurved, continuous, placed well in front of the middle. Dorsal

margin long, straight; anterior margin well rounded and merging

gradually into the moderately convex ventral margin; posterior

apparently truncate, but may have been evenly rounded. Shell pro-

bably very thin. Ornamented with numerous (about 35), relatively

strong, radiating ribs, which are towards the anterior separated by

rounded interspaces wider than themselves, and towards the posterior

more closely crowded; the ribs cover the entire shell, except for a

very small antero-dorsal area and a narrow postero-dorsal area.

Remarks.—The species differs from other Upper Cretaceous species

by reason of its small size and numerous ribs spread over most of the

shell.

188 Annals of the South African Museum.

Pholadomya cf. elliptica, Miinster.

(Pi. XEXS nes We d22)

Cf. 1839. P. elliptica, Miinster: Goldfuss’ Petr. Germ., vol. u1, p. 273,

plechan, fig: 1.

Cf. 1843. P. royana, d’Orbigny: Pal. Frang., Terr. Crét., vol. in,

p- 360, pl. ceclxvi.

Cf. 1874. P. elliptica, Moesch: Mon. der Phol., Mém. Soc. Pal.

Suisse, p. 104, pl. xxxiv, figs.

| 3, 4.

Ck 1912: > Pervinquiére: Pal. Tunis., Gastr. et Lam.

Crét., p. 289 (with syn.).

Material.—A single fairly well-preserved specimen with the valves

closed, T. Mus., No. 1586.

Description.—Shell thin, moderately inflated, elongated. Um-

bones inflated, moderately prominent, placed towards the anterior.

Anterior slightly produced with rounded margin ; posterior produced.

Ornamented with eight radiating, widely separated ribs, which are

absent from the anterior and posterior ends of the shell; and with

strong concentric growth rings.

Remarks.—This seems to be very close to the Kuropean species ;

the resemblance to the fig. 3 of Moesch and the fig. 3 of d’Orbigny

is particularly evident—these being varieties with comparatively

few ribs. P. elliptica is widespread in the Santonian and Campanian

of Northern Africa, France, and elsewhere in Europe.

Genus Gontiomya, Agassiz.

Goniomya umzambiensis, sp. Nov.

(Pl. XX, figs. 1-4.)

1906. Goniomya sp., Woods: Ann. 8. Afr. Mus., vol. iv, p. 310,

pl) seein ie 9:

Material.—Three specimens, one retaining portions of the test.

Holotype, T. Mus., No. 1588; paratypes, T. Mus., Nos. 1587, 1589.

Description.—Shell small, elongate-oblong, moderately convex,

inequilateral. Anterior moderately produced, the margin rounded,

the antero-dorsal margin making an angle of about 45° with the

ventral. Ventral margin straight. Posterior produced, truncate

and gaping; posterior margin oblique, forming an acute angle with

New Lamellibranchia and Gastropoda of Pondoland. 189

the ventral margin. Umbones moderately prominent with an obscure

carina extending postero-ventrally, separating a compressed ribless

area from the rest of the shell. Lunule and escutcheon elongate and

very narrow. Ornamentation of prominent ribs and faint concentric

striations ; anterior to the umbones is a ribbed area on which the ribs

extend obliquely backwards; posterior to the umbones are ribs

extending obliquely backwards and more or less parallel to the

posterior margin; the anterior and posterior ribs are connected by

less well-developed horizontal ribs, which are very slightly concave

to the ventral margin and meet the anterior and posterior ribs at

sharp angles; where the horizontal median ribs are very poorly

developed, the shell appears to be ribless towards the centre.

Remarks.—This species resembles in form and ornamentation

several species from rather lower horizons in Europe. G. archiaci

(Pictet and Renevier) from the Lower Greensand 1s relatively shorter,

and the posterior ribs are nearly vertical and curve round to meet

the horizontal median ribs (76, vol. ii, p. 254, pl. xli, figs. 4,5). In

G. rauliniana (d’Orbigny) (18, p. 353, pl. ccelxiti, figs. 4, 5), from the

Albian of France, the posterior ribs are placed obliquely forwards,

while G. caudata Agassiz=Ph. agassizi d’Orbigny (18, p. 352, pl.

ecclxiil, figs. 1, 2) is not so markedly truncate posteriorly and has the

posterior ribs curved.

The genus does not appear to have been recorded from the Cretaceous

of Southern India.

Famity ANATINIDAE.

Genus Cercomya, Agassiz.

Cercomya arcuata (Forbes).

(PIS ES fie.) 6:)

1846. Anatina arcuata, Forbes: Trans. Geol. Soc. Lond., vol. vi,

p. 143, pl. xvi, fig. 5.

1871. Anatina (Cercomya) arcuata, Stoliczka: Cret. Pel. 8. India,

De Opole tity shies. A.

1906. x a , Boule et Thevenin: Ann. de Pal.,

VOL pss: pl. 1, figs 7.

1923. Cercomya cf. arcuata, Bullen Newton: Trans. Geol. Soc. 8.

Afr., vol. xxvi, pp. 148-9.

Material.—A single right valve, in the condition of a cast but with

portions of the shell adhering, T. Mus., No. 1655.

iit

190 Annals of the South African Museum.

Description.—Shell elongate, moderately compressed, inequilateral,

highest in front of the umbones, tapering posteriorly. Anterior com-

pressed, the margin broadly rounded. Posterior produced, narrow,

with two carinae extending to the extremity from the incon-

spicuous umbo. Dorsal margin moderately concave, the ventral

almost straight. Ornamented with distant, rounded, concentric

folds and faint concentric striae, the folds not continued across the

posterior carinae ; and with radial rows of minute tubercles.

Remarks.—The resemblance to the figure given by Stoliczka of a

specimen from the Valudayur group is very great, and there can hardly

be any doubt as to the identity of the two forms. The ornamentation

of radial rows of minute tubercles, which seems to be characteristic

of the genus, was not noticed by Stoliczka. The species has been

recorded by Boule and Thevenin from Madagascar, and by R. Bullen

Newton from Portuguese Hast Africa.

Famity POROMYIDAE.

Genus LiopistHa, Meek.

Section pstLoMyA, Meek.

Inopistha (Psilomya) corrugata, Woods.

(Pl. X XI, figs. 1-4.)

1906. L. (Psilomya) corrugata, Woods: Ann. S. Afr. Mus., vol. iv,

p. 309, pl. xxxvulmigs:

10-12.

Material.—Species founded by Woods on three internal casts in

the Museum of the Geological Society of London. The numerous

specimens seen by the writer show that the types are juvenile

forms and consequently that the description must be considerably

augmented.

Transvaal Museum, Nos. 1640-1644, 1648, 1653, 1691.

South African Museum, Nos. 8536, 8537, 8599, 8680, 8694, 8710.

Among the Transvaal Museum specimens are four adult forms with

the shell beautifully preserved and several juvenile casts.

Description.—Shell oval, slightly inequilateral, inflated, rather

longer than high; margins rounded anteriorly and posteriorly,

ventral margin gently curved. Umbones prominent, curved inward

and slightly forward. Shell surface nearly smooth in the main, with

New Lamellibranchia and Gastropoda of Pondoland. 191

faint growth lines, but ornamented towards the umbones with rounded

concentric ribs or corrugations; these corrugations disappear with

the growth of the shell (7.e. are only formed in the younger stages)

and are more marked in the casts than in those specimens which

retain the test ; there is a certain amount of variability in the strength

of these corrugations, and some examples seem to lose them at an

earlier stage than others. The surface of the shell is covered with

radial rows of minute pits, these being most marked towards the

periphery, and finest and closest together on the dorsal region posterior

to the umbones, and often completely absent from the central parts

of the shell. There seems to be a very shallow groove extending

posteriorly from the umbo, parallel to the postero-dorsal margin.

Remarks.—This seems to be a very distinct species. The dorsal

corrugations are analogous to those of the otherwise very different

English species, L. gigantea (Sowerby), but are more strongly developed.

The Indian species of the same age (Ariyalur Group), P. globulosa

Forbes, does not possess the corrugations at all, and is more truncate

posteriorly.

Famity CYPRINIDAHE.

Genus VENIELLA, Stoliczka.

Veniella druz (Munier-Chalmas).

(Pl. XXI, figs. 5-9.)

1881. Roudaireia drui, Munier-Chalmas: Mission des Chotts, p. 76,

pleiven ties. (7, and

pole Ws) diets al;

1912. Roudawreia drur, Pervinquiére: Pal. Tunis, Gastr. et Lam.

Cret pe 230, ph xv,

figs. 9-13.

1917. Roudairera auressensis, Fortau: Geol. Surv. Egypt, Pal. Ser.,

| No. 3, p. 63 (with syn.).

1926. Veniella drui, Wade: U.S.G.S. Prof. Paper 137, Fauna

: Ripley Form. Tenn.,

jo. Ue

Material.—Three right valves in the Transvaal Museum Collection,

Nos. 1632-1634, one showing the complete hinge. There are two

right valves in the collection of the South African Museum, Nos. 8538

and 8474, and two in the collection of the Durban Museum.

Description.—Shell triangular, oblique, higher than long. Umbones

192 Annals of the South African Museum.

prominent, incurved and anteriorly directed. Posterior margin

slightly convex or sometimes nearly straight, meeting the ventral

margin in an acute angle; ventral margin straight or nearly so;

anterior margin well rounded. Lunule flattened, bounded by an

impressed line. A remarkably high carina extends from the umbones

to the produced postero-ventral angle, dividing the surface of each

valve into a rounded, convex anterior and a flattened, rapidly sloping

posterior. An obscure keel runs for a short distance from the umbones

on the posterior region, about midway between the carina and the

postero-dorsal margin.

Hinge stout ; in the right valve with three cardinals ; the posterior

long, oblique, not bifid, laterally compressed; the median small,

tubercle-like, placed obliquely above the anterior; anterior stouter

than the median and separated from it by a deep groove. Nymph

stout, elongated. Posterior lateral long, with a long groove above

it for the reception of the corresponding tooth of the left valve.

Surface of the valve with irregular concentric sulcations and growth

lines, finer behind the carina.

Remarks.—These Pondoland forms agree in all particulars with

the Tunisian examples so well described by Munier-Chalmas. The

writer has described and figured an example from the Baba district

of Angola, near Mossamedes, in a recent paper in these Annals (51).

Cyprina cristata Stoliczka (62, p. 198, p. 1x, fig. 1), from the Ariyalur

Group, known only from the single specimen figured by Stoliczka, may

well be a young example of this or an allied species.

aie ae

Veniella forbesiana (Stoliczka).

(PEO XOX ies 10%)

1871. Cyprina forbesiana, Stoliczka: Cret. Pel. S. India, p. 197,

pl. ix, figs. 2-8.

1905. Roudaieria forbesiana, Choftat: Nouvelles données, Angola,

p. 42, pl. 1, fig. 3.

VR ag Sy a aS 7 Bl

SS Sa 8 Sane See

SY be i #3 Pervinquiére: Pal. Tunis, Gastr. et

Lam. Crét., p. 232, pl. xv,

figs. 14, 15. .

1909. Veniella forbesiana, R. Bullen Newton: Trans. Roy. Soc. 8.

Afr., vol. 1; p: 6i,7 pega

figs. 1-4.

Remarks.—There is a cast of a small example of this species in the

collection, T. Mus., No. 1590. It agrees exactly with some beautifully

New Lamellibranchia and Gastropoda of Pondoland. £95

preserved small shells from Zululand, in the collection of the South

African Museum, described in the Appendix to this paper.

Famity ASTARTIDAE.

Genus ASTARTE, Sowerby.

Astarte ? amapondensis, sp. nov.

(PI XVI tigs. 7, 8:)

Material.—One specimen with the valves closed, very well preserved,

S.A. Mus., No. 8639.

Description.—Small, rounded, compressed, about as long as high.

Umbo small, moderately prominent, pointed, curved slightly forwards,

and placed a little in front of the middle. From the umbo a straight

furrow extends to a sinuosity in the middle of the posterior margin,

cutting off a very much compressed postero-dorsal area. A similar

but somewhat curved furrow extends from the umbo to an angulation

of the anterior margin, cutting off a small laterally compressed area.

Surface of each valve with six well-marked, flattened, concentric

lamallae, the interspaces between which are covered with minute con-

centric striations.

Remarks.—The species apparently belongs to Astarte, but is quite

castinct from A. griesbachi from the same deposit (77, p. 300, pl. xxxv,

figs. 18, 19).

Famity CRASSATELLITIDAE.

Genus CRASSATELLITES, Kriiger.

Crassatellites haughtoni, sp. nov.

(Pl. XXT, figs. 12, 13.)

Material.—The holotype is a well-preserved left valve, S.A. Mus.,

No. 8429.

Description.—Shell subquadrangular, moderately convex, longer

than high. Umbonal region moderately inflated ; the umbo promi-

nent, curved forward, placed in front of the middle. From the umbo

a well-defined, rounded carina extends backwards to the postero-

ventral margin, dividing a broad, flattened, sloping posterior area

from the rest of the valve. Lunular margin concave ; anterior margin

well rounded and merging gradually into the gently convex ventral

194 Annals of the South African Museum.

meeting the posterior margin in a very obtuse angle; posterior

ad margin gently convex, meeting the ventral margin in a rounded angle.

| I 4 Lunule moderately large, elongate, depressed, bounded by a raised

, ; ridge. Escutcheon long, very narrow, steeply descending, sharply

abt limited dorsally. Hinge stout; the two cardinals relatively short,

a diverging, the posterior one obliquely placed, both strong and ele-

\ vated ; cartilage pit moderately large, deep, very oblique. Margins

\| of valve coarsely crenulate within. Surface ornamented with strong

A concentric ribs and furrows, ribs and furrows of about the same width ;

“at behind the carina the ribs become somewhat laminar and fine con-

i, margin ; postero-dorsal margin very slightly convex, almost straight,

centric ribs appear between them; the ornamentation is somewhat

i irregular.

| Remarks.—This fine species 1s very different from C. africanus

Woods (77, p. 303, pl. xxxv, fig. 21; pl. xxxvi, figs. 1-3) from the

4 same deposits, both in form and disposition of the teeth and cartilage

pit, and is not likely to be confused with it.

any It more nearly resembles C. macrodonta (Sow.) (55, p. 417, pl. xxxvi,

\eil fig. 8), originally figured from Gosau by J. Sowerby, and since well

figured and described by Zittel (80, p. 150, pl. vii, figs. 2, 3), but the

Pe | | umbones of the Gosau species are more massive, higher, and with

: | | greater anterior curvature, the escutcheon is wider, the teeth are

| longer and more laterally compressed, and the lunule is usually more

pe { sunken. The same remarks apply to the Ariyalur forms described

by Stoliczka (62, p. 295, pl. v, figs. 12-14) and referred to the Gosau

| species.

Famity SPORTELLIDAKE.

Genus AnisopontTa, Deshayes.

Anisodonta *? umzambiensis, sp. nov.

(Pls XOX tie. id)

Material.—A right valve, in excellent preservation, but not showing

the hinge, T. Mus., No. 1654.

Description.—Shell elongate, inflated. Umbonal region swollen,

placed in front of the middle; umbo small, incurved, near the hinge

margin. A strong carina extends obliquely backwards from the umbo

to the postero-ventral angle, separating a wide, flattened, or slightly

concave dorsal area. Postero-dorsal margin long, straight ; posterior

margin shorter, straight, meeting the postero-dorsal margin in a sharp

=S—

See

———

pel

ol 78

Pay

i i

a |

“a

{

New Lamellibranchia and Gastropoda of Pondoland. 195

angle of about 125° and the ventral margin in an angle of a little less

than 90°. Ventral margin long, very gently convex ; anterior short,

its margin well rounded and merging gradually into that of the

ventral side. Surface with concentric striae and occasional stronger

growth lines, which bend sharply at the carina, become more pro-

nounced on the dorsal area, and meet the postero-dorsal margin

somewhat obliquely.

Remarks.—The single specimen described above seems to be unique

among Cretaceous Lamellibranchia, and its position cannot be defi-

nitely ascertained without a view of the hinge. It is referred very

doubtfully to Anisodonta Deshayes, whose type comes from the

Eocene of the Paris Basin, but the Pondoland shell is considerably

longer and more markedly carinated. In shape it approaches some

members of the genus Unio, but the prismatic layer is well developed.

Famity VENERIDAE.

Genus Dosintopsis, Conrad.

Dosinropsis geversi, sp. Nov.

(Pl. XXI, figs. 14-16.)

Material.—The holotype is a well-preserved left valve exhibiting

the hinge and interior, 8.A. Mus., No. 8518. Several imperfectly

preserved specimens probably belong to the same species, S.A. Mus,,

Nos. 8370, 8374, 8565, 8567, 8617, 8678.

Description.—Shell moderately thick, rounded, oval, moderately

convex ; longer than high, inequilateral. Antero-dorsal margin con-

cave or nearly straight; anterior margin rounded, passing gradually

into the curved ventral margin; postero-dorsal margin long, con-

vex; posterior margin rounded. Umbones small, pointed, anteriorly

eurved. Lunule elongated, flattened, bounded by an impressed line.

Escutcheon narrow. Pallial sinus moderately large, rounded.

Surface ornamented with fine concentric striae and growth lines.

Hinge of the left valve with three cardinals; the median and

anterior diverging from beneath the umbo; the median robust; the

posterior oblique, slender, joined to the flattened nymph ; the anterior

lateral elongate, parallel to the lunular margin; the posterior lateral

hardly evident.

Length, 30 mm. ; height, 26 mm.

Remarks.—This species appears to be extremely close to D. sub-

VOL. XXVIII, PART 2, 14

. a -

eS ee ee

196 Annals of the South African Museum.

rotunda (Sowerby) (76, vol. u, p. 181, pl. xxvii, figs. 1-6) from the

Upper Greensand of Blackdown; the points of difference are the

more nearly circular outline of the Blackdown form and the slightly

less curved and more centrally placed umbones ; the Pondoland form

seems to be a thicker shell and is distinctly more inflated ; the hinge

and pallial sinus of the left valves agree in almost every detail.

——F ———

—s

— J , a

= ————= =

| |

| Genus CypRIMERIa, Conrad.

wef | Cyprumeria ? natalensis, sp. nov.

rf | | | (Pl. XXI, figs. 19-21.)

1 Hi .

1 Hi

| Material.—One specimen with the valves closed, well preserved,

| | in the collection of the Durban Museum.

Description.—Shell small, oval, longer than high, only moderately

VAM convex. Umbones relatively prominent, anteriorly curved, touching,

| placed in front of the middle. Margins rounded, the anterior and

nai | posterior margins convex. Noescutcheon. Lunule moderately large,

ih | elongately cordate, raised in the middle, limited by an impressed line.

Surface polished, with traces of concentric sculpture.

Remarks.—This beautiful little shell is quite different from hitherto

described Pondoland and Zululand Venerids, and has the form of

Cyprimeria. The generic position is uncertain in the absence of hinge

characters.

Sub-genus CycLorisMA, Dall.

5 > ee Re:

| | Cyprimeria * (Cyclorisma *) Be Gower (Forbes).

| | | eae (Pl. XXI, figs. 17, 18.)

el -Remarks.—A single specimen (T. Mus., No. 1579) appears from its

uaa | shape and proportions, and poor development of the lunule, to belong

: to Cyclorisma Dall; it consists of the two valves, with finely preserved

i | exterior, but shows nothing of the interior. The shell is rounded,

| | oval, almost orbicular, moderately convex; length rather greater

| than height; moderately inequilateral. The antero-dorsal margin

| is short, slightly convex, passing gradually into the well-rounded

| ventral margin; postero-dorsal margin long, almost straight, but very

| slightly convex ; posterior margin somewhat truncate. The umbones

| are small, close together, anteriorly curved, and beneath them is a

small, depressed area corresponding to the lunule. The surface is

ornamented with fine concentric lines, some stronger than others.

New Lamellibranchia and Gastropoda of Pondoland. 197

Length, 32 mm.; height, 29 mm.

The species may be compared with C. analoga (Forbes) (62, p. 178,

pl. v, figs. 21-23) from the Trichinopoli Group, with which it agrees

fairly closely ; the Indian form is, however, relatively longer. The

writer has recently described a very similar form from the Senonian

of Angola (51).

Genus TRIGONOCALLISTA, nov.

This genus is proposed for two striking shells from the Senonian

of Pondoland, Meretrix umzambiensis Woods (77, p. 304, pl. xxxvi,

figs. 4-6), which may be taken as the type, and a new species, 7. spathi.

The genus differs from Macrocallista chiefly in the combination of a

markedly trigonal shape and large escutcheon with the presence of

an entire right posterior cardinal.

The group may be thus defined: shell stout, triangular, more or

less truncate posteriorly, umbonal region very prominent, umbones

with strong anterior curvature, very inequilateral. Lunule large,

deep, cordate, well limited. Escutcheon large, wide, depressed, with

a strong bounding ridge. Hinge very strong, teeth in the left valve

consisting of three cardinals, the posterior joined to the nymph but

well developed and elongate, the two others diverging from beneath

the umbo; the anterior lateral remarkably long and stout, the posterior

lateral long and thin. In the right valve there are two cardinals

below the umbo, a long, thick, superficially grooved posterior cardinal

and long pits to receive the left laterals. Ornamentation concentric.

Pallial sinus rounded, horizontal, not large.

Trigonocallista is remarkably stout and trigonal, and easily recog-

nised by its very strongly curved umbones and well-marked escutcheon.

The hinge is similar to that of Jukes-Browne’s Callistina (29, p. 156),

of which M. plana (Sowerby) (76, vol. ii, p. 192, pl. xxx, figs. 1-6) is

the type, but differs from it in two important points, (1) the left

anterior lateral is much longer and stouter and (2) the right posterior

cardinal is not bifid but merely grooved along its summit. Cytherea

polymorpha Zittel (80, pt. i, p. 126, pl. i, fig. 6), which was associated

with M. plana by Jukes-Browne, is rather more triangular than the

type of Callistina, but otherwise agrees perfectly with it. Cytherea

renauxiana d’Orbigny (18, vol. ii, p. 447, pl. cceclxxxvi, figs. 1-3) is

very similar in shape to the two Pondoland species and possesses a

similar escutcheon, but the hinge has not been figured.

198 | Annals of the South African Museum.

Trigonocallista spathi, sp. nov.

(Pl. XXII, figs. 1-9.)

Material—Founded on several well-preserved specimens in the’

collection of the Durban Museum. Also Transvaal Museum, No. 1580;

South African Museum, No. 8707.

Description.—Shell stout, triangular, very inequilateral. Umbonal

region very prominent, umbones with strong anterior curvature.

Lunular margin deeply concave; anterior margin well rounded,

passing gradually into the strongly curved ventral margin; postero-

dorsal margin very long, slightly convex, almost straight ; posterior

margin truncate, meeting the postero-dorsal and ventral margins in

obtuse angles. Lunule large, cordate, much depressed, bounded by a

strong groove, outside of which is a raised ridge. Escutcheon large,

wide, depressed, somewhat flattened, bounded by astrong ridge. Hinge

of left valve strong; posterior cardinal strong, elevated, joined to the

nymph, the median and anterior cardinals diverging from beneath

the umbo, the anterior lateral stout, straight, and very long. In the

right valve the median and anterior cardinals diverge from beneath

the umbo ; posterior cardinal large, slightly curved, elevated, narrow-

ing upwards and with a slight groove; large deep pits for the left

laterals. Pallial sinus moderately deep, rounded at the apex.

Ornamentation consisting of regular concentric ridges and grooves,

the ridges being elevated and sharp, and giving place to growth lines

on escutcheon and lunule.

4 i

aa |

}

a ‘|

‘

| '

~ fi 1 ]

1 |

5 a

| Soe

i, q |

i} 1

| a]

ts Bis

’ |

| i ha

- *, a |

mt “4| i |

, a) a t}

im at |

| a

nl Oi

| f

xh i | ||

. 1] |

yet be

(a). (0). (c). No. 1580.

Length . 56mm. 43 mm. 62 mm. 60 mm.

Height. Stra, OMe as 54, Doi

Remarks.—This fine species is very close to WM. wumzambiensis Woods

in the structure of the hinge, but is easily distinguished by the greater

posterior truncation in the larger examples and the rather different,

coarser ornamentation. |

Genus Macroca.uista, Meek.

Sub-genus CALLISTINA, Jukes-Browne.

Macrocallista (Callistina) euglypha (Woods).

(Pl. XXII, figs. 10-12.)

1906. Meretrix euglypha, Woods: Ann. 8. Afr. Mus., vol. iv, p. 305,

pl. xxxvi, figs. 7-10.

New Lamellibranchia and Gastropoda of Pondoland. 199

Material. Woods did not see the hinge of this species. The

collection made by Gevers includes numerous specimens, two of which,

S.A. Mus., Nos. 8588 and 8589, have the hinge quite undamaged.

Description (partly after Woods).—Shell oval, of moderate or

small convexity. Margin near the lunule concave. Anterior margin

well rounded and forming a continuous curve with the ventral margin.

Posterior margin rounded, but often slightly truncate. Postero-

dorsal margin with a gentle curvature. Umbones of moderate size,

curving slightly. Lunule small, elongate, bounded by a groove.

Ornamentation consists of numerous strong, sharp, regularly

concentric ribs, separated by deep and narrow grooves. The ribs

are usually continued on to the lunule.

Hinge in the right valve with three cardinals; the anterior and

median small, laterally compressed, plate-like, diverging from beneath

the umbo; posterior cardinal oblique, long, divided, the posterior

part longer than the anterior part. In front of the cardinals in the

right valve there is a small elongate pit. Left valve with stout

anterior and median cardinals diverging from beneath the umbo,

the anterior nearly vertical ; posterior cardinal long, slender, oblique,

situated on the flank of the nymph; anterior lateral elongate, ridge-

like.

Remarks.—The hinge of this species, which was described by Woods

as a Meretrix, agrees exactly with that of Sowerby’s Venus planus

(76, vol. 11, p. 192, pl. xxx, figs. 1-6), which was made the type of

Callistina by Jukes-Browne (29, p. 156).

Woods has discussed the relations between WM. euglypha and Cytherea

hornesi Zittel from Gosau, Cytherea fabulina Stoliczka from Southern

India, and other species.

The writer has had the opportunity of comparing perfect specimens

of M. euglypha and Meretrix andersoni Bullen Newton. The latter

is from the south side of the Manuan Creek Valley, and is described

below among the Zululand forms as Macrocallista (Callistina) anderson.

The two species possess almost identical hinges, and are without doubt

very closely allied. The Zululand species is, however, distinctly more

inflated and is usually longer than M. euglypha, which possesses a much

smaller lunule.

Se ee Sl EE EN ET ae as ee

= = a ~ ow = a

> en £ 3 \ e

=~ 5 P 3 3 pre “= =

= ees 3 7 eee we ‘ \ = Se >

BS a ee Pe SET et - as — eo

-- : + = {sf ee 4 = - ——

a4 all -- ge Sy oe — E. | — *.

+—_§-—f-

200 : Annals of the South African Museum.

Famity TELLINIDAE.

Genus PALAEOMOERA, Stoliczka.

Palacomoera umzambiensis, Sp. NOV.

(Pl. XXIII, figs. 1-3.) —

Material.—A single left valve, very well preserved, T. Mus., No.

1639. |

Description.—Shell elongate-oval, compressed, inequilateral, length

about 13 times the height. Anterior part longer than the posterior,

umbo small with a very slight anterior curvature. Ventral margin

slightly convex, meeting the posterior margin in a very obtuse angle ;

posterior and anterior margins rounded. A very obscure, rounded

carina appears to extend from the umbo to the postero-ventral angle,

where the margin is slightly raised. Hinge with a single anterior

tooth, long and obliquely placed.

Ornamentation, on the posterior part, of numerous rounded radiating

ribs separated by grooves which are wider than the ribs; on passing

the obscure carina the ribs come closer together, and towards the

middle of the shell the intervening grooves pass into spaced, radiating

striae; the latter cease abruptly beneath the umbo, but mbs and

grooves reappear antero-dorsally. The whole shell is covered with

growth striae of varying strength, which give to the ribs a somewhat

rugose appearance. | |

Remarks.—The form of the shell and the single oblique anterior

tooth place this species in Stoliczka’s Palaeomoera. It is very similar

to Palaeomoera inaequalis (Sow.) (76, vol. ii, p. 173, pl. xxvui, figs. 2-8)

from the Upper Greensand, but in the latter the radial ornamentation

is confined to the posterior portion of the shell. T. strigata Goldfuss

(24, vol. 11, p. 234, pl. exlvii, fig. 18), the type of Palaeomoera, has fine

radial ornamentation over the entire surface, while the Indian species

P. inconspicua (Sow.) (62, p. 129, pl. iv, figs. 6-8) is entirely without

radial ornamentation.

Palaeomoera sp.

(Pl. XXIII, figs. 4, 5.)

Material.—One right valve, T. Mus., No. 1638.

Description.—Like the preceding in form, but differing in the details

of the ornamentation. On the posterior part are numerous strong

ribs and grooves of equal width; by flattening and broadening of the

New Lamellibranchia and Gastropoda of Pondoland. 201

ribs this area soon passes into an area ornamented with spaced, sharp,

radiating grooves, which become more widely spaced anteriorly. The

whole shell is covered with sharp, regular, concentric lines.

Remarks.—The hinge was not clearly seen, but appears to be that

of a Palaeomoera. The shell resembles the preceding rather closely

and only further collecting will show whether it is a different species.

Palaeomoera haughtont, sp. nov.

(Pl. XXIII, figs. 6-8.)

Materval.—One right and two left valves, not showing the hinges ;

holotype a left valve, T. Mus., No. 1637; paratypes, T. Mus., Nos.

1649, 1650. The writer has also seen a specimen with the valves

connected, in the collection of the Durban Museum.

Description.—Shell oval, elongate, considerably longer than high,

moderately compressed, inequilateral. Umbo small, moderately

prominent, behind the middle. Anterior long, the margin well rounded

and merging quite gradually into the convex ventral margin. Posterior

shorter, truncate; a rounded carina extending from the umbones

to the postero-ventral angle of about 120°; posterior margin short,

straight, and nearly vertical, meeting the short, slightly convex

postero-dorsal margin in a rounded angle. Surface almost smooth,

with faint growth striae, and at fairly regular intervals stronger growth

rings, and with numerous faint, spaced, discontinuous radial lines.

Remarks.—This species is quite distinct from the preceding and

from any species from Southern India, and from its form would appear

to be a true Palaeomoera.

Genus Macoma, Leach.

Macoma papyracea, sp. nov.

(PE OXeXU Ties 1 12.)

Material.—One specimen with the valves closed, very well pre-

served, T. Mus., No. 1636.

Description.—Shell oval; longer than high, moderately inflated,

very thin. Umbones small, fairly prominent, pointed, touching,

behind the middle and posteriorly directed. Anterior longer than the

posterior, broad, rounded, the margin passing quite gradually into

the convex ventral margin. Posterior shorter, narrowed, compressed,

‘

‘i

Ti why

ae he

ais

} +

) 4

| |

bed

. 1}

;

| \

ity

Paks tet

é v

2 ae a ae

Se eS S

Se TS

202 Annals of the South African Museum.

slightly truncate at the tip, flexured. Surface with very numerous

growth lines, the strength varying.

Remarks.—This interesting shell is probably a Macoma, but may

possibly be a true Tellina. In form it approaches M. lata (Gmelin),

regarded by Dall as identical with the genotype, and M. praetenuis

(Leathes MSS.) (Wood), both Red Crag species (75, pp. 228-231,

pl. xxi, figs. 6, 7), sufficiently well to justify the inclusion of the species

in Macoma. As the crucial point is the absence of lateral teeth in

Macoma and their presence in Tellina, the point cannot be decided

without more complete material. The only Cretaceous form that

appears to be comparable is the shell described by Whiteaves as

Cuspidaria suciensis (69, p. 376, pl. xlvi, fig. 2), a smaller shell with a

relatively narrower and more produced posterior.

Famity SOLENIDAE.

Genus SoLtecurtus, Blainville.

Sub-genus Azor, Gray.

Solecurtus (Azor) woodsi, sp. nov.

(Pl. XXIII, figs. 18-20.)

1906. Solecurtus? (Azor ?) sp., Woods: Ann 8. Afr. Mus., vol. iv,

p. 308, pl. xxxvu, fig. 7.

Material—The writer has seen numerous specimens—Transvaal

Museum, Nos. 1582-1585; South African Museum, Nos. 8580, 8582.

Also some examples in the collection of the Durban Museum. The

diagnosis has been made from several specimens ; holotype Tvl. Mus.,

No. 1584, a large right valve which, however, does not exhibit the

hinge; Tvl. Mus., No. 1585 is the interior view of a left valve ; while

S.A. Mus., No. 8580 is a left valve showing the hinge.

Description.—Shell elongate, inequilateral, slightly curved and

moderately convex, with a distinct carina extending from the umbo

to the angular postero-ventral extremity. Umbo small, placed a

short distance in front of the middle. Dorsal margin gently convex ;

ventral margin concave; anterior margin rounded; _ posterior

truncate, the posterior and ventral margins meeting at an angle of

about 80°. Lunule small, not distinct. In the holotype a ridge

extends posteriorly from the umbo near the dorsal margin, becoming

obsolete some distance before reaching the posterior margin.

New Lamellibranchia and Gastropoda of Pondoland. 203

Anterior adductor scar smaller than the posterior, oval, irregular in

outline ; posterior scar larger, rounded. Pallial sinus deep, the end

broadly rounded. Hinge of the left valve with a single oblique tooth

beneath the umbo.

Surface ornamented with strong, irregular growth lines, meeting

the dorsal margin anterior to the umbones at an acute angle; these

bend sharply upwards on crossing the carina and increase in strength,

meeting the postero-dorsal margin obliquely.

Remarks.—Woods recorded two specimens, both very imperfect,

as Solecurtus ? (Azor ?). The species is very similar to d’Orbigny’s

Solen guerangert (18, vol. iti, p. 321, pl. cccli, figs. 1, 2), but is less

elongated and posteriorly produced, and the umbo is more nearly

central, while the postero-dorsal angle is more rounded. Stoliczka’s

Tagellus albertinus (62, p. 102, pl. iv, fig. 3) has a considerable super-

ficial resemblance to S. woodsi, but that author states that there are

two small teeth in the left valve, whereas S. woodsi possesses only one.

Famity MACTRIDAE.

Genus CymMBopHora, Gabb.

Cymbophora rogersi, sp. nov.

(PIX XT figs: 15-17.)

Material.—The holotype is a well-preserved left valve, T. Mus.,

No. 1651. There are two specimens in the collection of the South

African Museum, Nos. 8372 and 8459, both incomplete.

Description.Shell above the average in size, triangular, longer

than high, inequilateral, moderately inflated. Umbo small, high,

pointed, anteriorly curved, almost central. Antero-dorsal and

postero-dorsal margins diverging from the summit at about 117°,

the former long and straight, the latter gently convex. Ventral

margin forming a broad curve, posterior straight and truncate,

anterior very short and truncate. Lunule and escutcheon very

narrow and indistinct. A weak, rounded carina extends obliquely

backwards from the umbo to the postero-ventral angle and immedi-

ately in front of it is a shallow, rounded sulcus which widens gradually

in the ventral direction and results in a concavity. or sinuosity in

the ventral margin in front of the angle. A faint sulcus can be made

out extending obliquely forwards from the umbo to the very short,

truncate, slightly concave anterior margin.

t Ee See

204. Annals of the South African Museum.

Ornamentation of numerous strong, concentric lamellae, which

become progressively stronger ventrally ; these lamellae are weakened

on the posterior carina.

Hinge strong, thick, with a large, deep, oblique, spoon-shaped

chondrophore with sharply raised bounding ridges on either side ;

cardinal tooth relatively small, like an inverted V in shape, and

distinct from the raised anterior margin of the chondrophore. Anterior

lateral tooth very strong and elevated; the posterior very strong,

longer but less elevated than the anterior ; laterals not cross-ribbed.

Adductor impressions large, subequal, sunken. Pallial sinus short,

rounded. .

Measurements of the holotype: length, 46 mm. ; height, 38 mm. ;

thickness, 14 mm.

Remarks.—The genus Cymbophora was proposed by Gabb (28,

p. 180) for Mactra ashburnert Gabb from the Cretaceous of California,

distinguished by possessing a spoon-shaped chondrophore with raised

margins, a slender cardinal distinctly separated from the chondrophore

margin in the left valve, and large lateral teeth. According to Dall

(11, p. 879), who studied the typical species, the attached ends of the

resilium are convex and the margins of the pit therefore elevated ;

and the ligament is fixed on the convex margin of the pit or on the

side of the ventral lamina, or partly on both. The absence of a

special septum separating the ligament and the resilium caused Dall

to place the group as a sub-genus of Spisula, though Meek (87, p. 204)

evidently considered that the raised posterior margin of the chondro-

phore is homologous with the septum of Mactra, for he placed

Cymbophora under Mactra. Anderson (1, p. 74) has referred Gabb’s

species to Mactra without reference to Cymbophora. On the other

hand, Wade (67, p. 95) has recently retained Cymbophora as a genus.

In the opinion of the writer the absence of a special septum

separating the ligament and resilium is a point of resemblance between

Cymbophora and Spisula. At the same time the ligament and resilium

are very definitely separated by the raised posterior margin of the

chondrophore, which leaves an elongated slit behind it for the reception

of the ligament. The nature of the chondrophore and its relation to

the ligament-slit and the cardinal tooth, as well as the unusually large

laterals, warrant the retention of the genus. Dall has already re-

marked (loc. cit.) that the Cymbophora type of hinge occurs in most

Mesozoic species of Mactridae.

C. gabbiana (Anderson) (1, p. 74, pl. vii, fig. 156), from the Lower

Chico of California, resembles the Pondoland species closely in form

New Lamellibranchia and Gastropoda of Pondoland. 205

and is strongly grooved concentrically, but appears to be without a

posterior sulcus.

C. alta (Meek) (87, p. 210, pl. xxxvu, fig. 2) is relatively higher

and more pointed, and with finer concentric sculpture. C. gracilis

(Meek), figured-by Wade (67, p. 95, pl. xxxi, fig. 8), is not unlike the

Pondoland species; it is considered by Wade to bear a striking

resemblance to Mactra (?) zulu Etheridge (17, p. 82, pl. ii, figs. 17-19)

from Umkwelane Hill, but the latter is certainly not identical with

C. rogerst. |

Genus Mactra, Linnaeus.

Mactra? kaffrarva, sp. nov.

(PIXE fies. 95 110:)

Materval.—A well-preserved left valve, T. Mus., No. 1652.

Description.—Shell triangular, compressed, longer than high,

inequilateral. Umbo small, umbonal region prominent. Ventral

margin gently convex ; postero- and antero-dorsal margins diverging

at about 100°; both straight, and meeting the ventral margin in

rounded angles. From the umbo sharp carinae run to the anterior

and posterior angles, cutting off on their dorsal sides narrow, steep,

almost vertical areas, the anterior being wider than the posterior.

Surface almost smooth, with very faint growth and radial lines.

Remarks.—This is a very distinct little shell, referred tentatively

to Mactra on account of itsform. =

Famity PHOLADIDAE.

Genus Martestia, Leach.

Martesia ? sp. )

(Pl. XXIII, figs. 13, 14.)

Materral.—An incomplete cast with the valves closed, T. Mus.,

No. 1592:

Description.—Shell moderate in size, elongate-ovate. A strong

furrow descends obliquely backwards from the umbones to the ventral

margin, dividing each valve sharply into an inflated anterior and a

compressed, tapering posterior. Anterior well inflated, rounded, the

antero-ventral margin-rounded and.oblique, anteriorly sub-truncate ;

ornamentation of strong, regular, oblique ribs separated by wider

Bi,

HM hi |

oa!

l Bal

i f |

- ~

;

‘

8 Hh

- |

oe |

> '

Ob)

by i tit

206 Annals of the South African Museum.

interspaces and traces of radial ribs. Umbones moderately inflated,

not laterally compressed, longer than the anterior, tapering when

viewed from the side, the ventral margin straight and horizontal, the

dorsal oblique ; ornamentation like that on the anterior part. Acces-

sory plates not seen; anterior with small or no gape.

Remarks.—This interesting shell is not sufficiently well preserved

for the generic position to be definitely ascertained, but it has the form

of a Martesia. Pholas cithara Morton (70, p. 187, pl. xxv, figs. 14-16),

regarded by Gabb as a Martesia, is readily distinguished by having

the posterior more quickly tapering and laterally compressed, a less

deep dividing furrow, more strongly developed radial ornamentation,

more prominent umbones and higher anterior region. M. cuneata

Meek (37, p. 259, pl. xxx, fig. 8) is very much smaller and with the

concentric ribs well developed, but is less oblique in the anterior

region.

Pholas reticulata Miller (28, p. 140, pl. vii, figs. 1, 2) rega ded by

Stoliczka (62, p. 23) as probably a Martesia, from the Senomian of

Aachen, resembles the present species closely in shape, but the radials |

are more strongly developed, and the present species is more oblique

anteriorly.

GASTROPODA

Famity PATELLIDAE.

Genus PATELLA, Linnaeus.

Patella kaffraria, sp. nov.

(Pl. XXIV, figs. 1-4.)

Materval— Two specimens; holotype, S.A. Mus., No. 8477,

paratype, S.A. Mus., No. 8572.

Description.—Shell moderately convex, with the apex obtusely

pointed, not incurved, and placed well in front of the middle; the

sides are straight, or only very slightly convex. Aperture oval,

considerably longer than wide, with wavy margin. Surface with

stout, rather irregular, radial ribs, and narrow furrows; on the

posterior side the ribs are of two sizes, the larger and smaller alter-

nating; on the anterior side there are more numerous, finer ribs;

the ribs are crossed by irregular growth markings.

Remarks.—lIt need hardly be stated that the genus Patella is here

used in the widest possible sense. The species is apparently distinct

from any previously described from the Cretaceous.

New Lamellibranchia and Gastropoda of Pondoland. 207

Famity KUOMPHALIDAE.

Genus SEMISOLARIUM, Cossmann.

Semrsolarium bailyr (Gabb).

1855. Solarium pulchellum, Baily: Q.J.G.S., vol. xi, p. 457, pl. xii,

| fig. 3 (non d’Orbigny, 1850 ;

Prodr. de Pal., 1, p. 104).

RST 1: 5, wiebeli, Griesbach: Q.J.G.S., vol. xxvu, p. 65,

plan, fie; (6:

1861. ee (Architectonica) bailyr, Gabb: Proc. Amer. Phil. Soc.,

vol. vi, p. 95.

1906. i bailyi, Woods: Ann. S. Afr. Mus., vol. iv, p. 315,

pl. xxxviul, figs. 4, 5.

1915. Semesolarvwm barlyz, Cossmann: Paléoconch. Comp., vol. x,

p- 156.

Remarks.—Cossmann has included this species and Architectonica ©

kossmati R. B. Newton (48, p. 23, pl. vii, figs. 11, 12) in his genus

Semisolarium in the family Euomphalidae. Solariwm sp. ind.,

R. Etheridge, jr., from Umkwelane Hill, Zululand (17, p. 88, pl. u,

figs. 35-37), may be identical with S. badly, but the specimens seen

by the writer from that locality are too poor for comparison.

Famity TROCHIDAE.

Genus CHILODONTA, Etallon.

Sub-genus AGATHODONTA, Cossmann.

Chilodonta (Agathodonta) africana, sp. nov.

(Pl. XXIV, figs. 8-10.)

Materval.—One specimen, 8.A. Mus., No. 8630, consisting of two

whorls, the aperture and ornamentation fairly well preserved.

Description.—Shell small, trochiform, a little elongated, with a

spiral angle of about 45°. Whorls very convex, the height of the

whorls less than half their width; sutures deep, subcanaliculate.

Aperture almost circular, entire, comparatively large; outer lip

moderately thick; inner lip somewhat thickened; columella pro-

vided with a large, median, amorphous tooth, in front of which is a

stout, conical tubercle. The plane of the aperture is inclined to the

sutures at an angle of about 50°. Surface ornamented with numerous

208 ‘Annals of the South African Museum.

stout spiral ribs, separated by grooves which are a little narrower

than the ribs; in places the ribs appear to be somewhat nodular,

the nodules being arranged in transverse lines, and the penultimate

whorl bears several obscure transverse folds.

Remarks.—The sub-genus Agathodonta was created by Cossmann to

include three Cretaceous species which differ from Chilodonta in

several respects. In Agathodonta the shell is more elongated; the

transverse striae are not continuous from suture to suture, but are

present only on the spiral ribs, where they form tubercles; the

aperture has only a few tubercles, and the columella is provided with

two teeth instead of one. The type of Agathodonta is Trochus dentiger

d’Orbigny (12, p. 185, pl. clxxvu, figs. 9-12), figured also by Cossmann

(6, vol. xi, p. 200, pl. vu, figs. 8-11), from the Neocomian of France.

C. africana differs in no essential features from C. dentigera; the

tubercles on the riblets are less in evidence in the Pondoland species,

and the obscure transverse folds seem to be peculiar to the species.

The whorls of (. guyotiana (Pictet et Roux) and C. tollotiana (Pictet

et Roux) (48, vol. i, pp. 202, 203, pl. xix, figs. 8, 9) are much less

convex.

Genus MarGarirTes (Leach MSS.) Gray.

Sub-genus SOLARIELLA, Wood.

Margarites (Solariella) radiatula (Forbes).

(PAS Dexa ie 1s)

1846. Trochus radiatulus, Forbes: Trans. Geol. Soc. Lond., ser. 2,

vol. vii, p. 120, pl. xii, fig. 11.

1868. Solariella radiatula, Stoliczka: Cret. Gastr. 8S. India, p. 375,

pl.. xxiv, figs. 17-192 eee

xxvii, figs. 8, 9.

1906. Margarita radiatula, Woods: Ann. 8. Afr. Mus., vol. iv, p.

310; pl. xxxvu, fies IS

1918. Eumargarita (Solariella) radiatula, Cossmann: Paléoconch.

? .Comp., vol. xi, p. 260.

Remarks.—The examples from Pondoland seen by Woods did not

exhibit the fine spiral striae characteristic of young specimens from

the Ariyalur Group. The present collection contains numerous

specimens, most of which, though apparently perfectly preserved,

are quite smooth or covered only with faint growth lines; two,

however, S.A. Mus., Nos. 8575, 8631, show spiral striae, which are

New Lamellibranchia and Gastropoda of Pondoland. 209

stronger on the posterior half of the last whorl than on the base of

the shell; in front of the sutures the spiral striae are crossed by fine

transverse ribbings.

The generic name Margarites (Leach MSS.) Gray, 1847, must be

retained for this group, according to the rules of nomenclature.

Margarita Leach, 1819, must be discarded as the name was first used

by that author in 1814 for a Lamellibranch ; Humargarita Fischer,

1885, which was retained by Cossmann in 1918, must give place to

Margarites owing to priority. The name Margarites has been used

subsequently by Mojsisovics in 1889 for an Ammonite.

Famity NERITIDAE.

- Genus Nerita, Linnaeus.

Nerita umzambiensis, Woods.

(BI ROXd: tigs5 165 We)

1906. Nerita umzambiensis, Woods: Ann. 8. Afr. Mus., vol. iv,

ps old als xacxvil, ies. 14s

1155}. |

Remarks.—The species was founded by Woods on three specimens.

The Transvaal Museum collection contains one specimen which is

considerably larger than either of the examples figured by Woods.

In form and ornamentation it agrees exactly with the types, so that

there is nothing to add to the original description of the species. T.

Mus., No. 301.

Length, 35 mm.; breadth, 43 mm.

Nerita kaffraria, Woods.

(GEE DOG E s3ta, 5),

1906. Nerita kaffraria, Woods: Ann. S. Afr. Mus., vol. iv, p. 311,

Dl SRV. 16.

Remarks.—The species was founded by Woods on a single specimen,

in which traces of the original colour-banding still exist. In the

collection cf the Transvaal Museum there is a single specimen referable

to this species in which the colour bands are preserved with remarkable

clearness. The coloration consists of alternate, broad, sharply limited,

spiral bands of dark brown and dirty white ; the two brown bands are

situated respectively on the anterior side near the margin of the

210 Annals of the South African Museum.

aperture and immediately behind the part of greatest convexity.

T. Mus., No. 1604.

Fawtty PYRAMIDELLIDAE.

Genus TRAJANELLA, Popovici-Hatzeg.

Trajanella dutoitr, sp. nov.

(Pl. XXIV, figs. 16-18.)

1855. Turritella Renauxiana, Baily: Q.J.G.S., vol. xi, p. 458.

1867. Eurchrysalis gigantea, Stoliczka: Cret. Gastr. S. India,

p- 290 (reference only).

1870. 6 3 Griesbach: Q.J.G.S., vol. xxvii, p. 65.

1906. Pseudomelania (Oonia) sp., Woods: Ann. 8. Afr. Mus., vol.

iv, p. 313, pl. xxxviu, fig. 1.

1909. Trajanella sp., Cossmann: Paléoconch. Comp., vol. viii,

p- 109.

Material.—Holotype and paratype in the Durban Museum

Collection.

Description.—Shell large, thick, elongately oval. Spire relatively

short, composed of at least five whorls; spiral angle about 50°;

whorls of the spire flat or very slightly convex; sutures not con-

spicuous. Aperture relatively small, lunate, narrowing considerably

in the posterior direction, rounded anteriorly ; aperture somewhat

oblique and anteriorly truncated, so that the interior is seen when

the shell is viewed from the front. Outer lip thin; inner lip with a

thin deposit of callus. Surface smooth, with moderately sinuous

growth striae.

Remarks.—The collection of Pondoland fossils made by Captain

Garden and described by Baily in 1855 included an imperfect specimen,

which Baily erroneously referred to Turritella Renauxiana d’Orbigny.

The specimen was seen by Stoliczka in the Museum of the Geological

Society of London and identified with the Trichinopoli-Ariyalur

species, in which he was followed by Griesbach. The specimen

was not, however, figured until 1906, when Woods referred it to the

group Oonia Gemmellaro. Cossmann regarded it as a new species of

Trajanella. The Pondoland species should be referred to Trajanella

on account of its peculiar eulimoid shape and truncated aperture ;

Oonia, which according to Cossmann does not occur above the

Barremian, is more conical in outline, and the whorls of the spire are

more distinct, while the aperture is straight.

New Lamelhibranchia and Gastropoda of Pondoland. 211

T. dutoiti is very closely related to the Indian species, 7. gigantea

(Stoliczka) (61, p. 289, pl. xxi, figs. 3-5), but differs in that the spire

is less elongated, the aperture relatively smaller (if the restoration

given by Stoliczka is correct) and the sutures less distinct. In the

Indian species the inner lip appears to be provided with a thicker

deposit of callus.

Sturm has figured a similar form from the Chlomek Beds in Saxony

(63, p. 67, pl. v, fig. 1), but Cossmann questions its identification with

the Indian species.

Famity FOSSARIDAE.

Genus Lysis, Gabb.

Lysis capensis, sp. nov.

(Pl. XXIV, figs. 19-21.)

Material.—Holotype, T. Mus., No. 1615; paratype, T. Mus., No.

1614. There is also a cast in the collection of the South African

Museum, collected by Dr. Gevers, No. 8534.

Generic Diagnosis (after Gabb, Pal. Calif., vol. i, p. 138).—** General

form like Stomatia. Shell sub-spiral, very oblique ; spire moderately

prominent; whorls costate. Aperture narrow, oblique; outer lip

simple ; columellar lip straight and rather heavily incrusted. Um-

bilicus broad, but entirely closed by a concave expansion of the

incrusting layer of the inner lip.”

Description.—Shell moderately thin, semi-ovate, oblique, below

the average in size. Spire small, pointed, obliquely placed. Whorls

three, rapidly increasing in size, sutures impressed. Last whorl

large, forming most of the shell, rounded posteriorly, flattened

anteriorly, with six stout, sharp, spiral ribs, widely spaced on the

posterior and closer together on the anterior, with broad, flattened

interspaces ; third rib from the posterior more elevated than the

others, marking the angulation of the whorl; interspaces crossed by

numerous oblique, irregular, growth lines, which are extended across

the ribs and in places tend to give them a serrate or roughened appear-

ance. Aperture oblique. Outer lip not seen. Inner lip smooth,

entire, very slightly concave; umbilical region represented by a

broad, concave, smooth area of callus, sharply carinated along its

outer edge.

Remarks.—Gabb founded the genus Lysis for a new and interesting

little species from the Upper Cretaceous of California, L. duplicosta

VOL. XXVIII, PART II. 15

22S

==" = SS SSeS

=... = =

SS ==

Ps, SES

- =) > a > >

+~ > ~., ae — & a = 5. iil a ii Se

2) a eee

rar gg rae

==.

: , Ps Reo . (SS SMS PI —— ee =

= : = - BESS: - = ee Se SS eS ee eee ee

. : = -— — :

Ss eS =e = = = = = Se == = = = =

Soe a ee ee San =e = — = : = = = seed oe

212 Annals of the South African Museum.

Gabb, but gave no indication of what he considered to be its affinities,

other than stating that the general form was like Stomatia, which is

almost entirely a recent group. Stoliczka (61, p. 158) remarks that

it ““. . . has the general form of Fossar, and partly that of Natzca

and Stomatia,”’ and “it is, properly speaking, very much allied to

N. carinata Sow. sp., which is very probably a Fossar.” Fischer

placed it in the Naticidae (19, p. 768), while in Zittel (81, p. 558) it is

included under Thaisidae. The writer has here placed it in the

Fossaridae on account of its form and ornamentation, and considers

it to be closely allied to Phasianema 8. Wood (6, vol. x, p. 88), from

which it differs principally in the complete closing of the umbilicus ;

in other respects the two groups seem to be very similar, and it may

well be that Lyszs should be regarded merely as a sub-genus of Fossarus.

The only other figured Cretaceous species which appears to be at all

like Lysis is the Natica ? carinata Sowerby (56, p. 343, pl. xviii, fig. 8)

quoted by Stoliczka, a form which is at once distinguished by its open

umbilicus and stouter shell, but in general resembles the Fossaridae

rather closely.

It is of extraordinary interest to be able to record a species of this

genus in the South African Senonian deposits. L. capensis resembles

the genotype sufficiently closely in general features (few whorls,

obliquity, ribbing, nature of the umbilicus) as to leave no doubt as

to the relations of the two species. It is distinguished from the

Californian species by its rather more inflated and less oblique body-

whorl, the angulation which persists in each of the specimens examined,

and a slightly larger spire.

Famity NATICIDAE.

Genus GyRoDEs, Conrad.

Gyrodes tenellus, Stoliczka.

(Pl. X XV, figs. 1-7.)

1868. Gyrodes tenellus, Stoliczka: Cret. Gastr. §. India, p. 306,

pl. xxii, fig. 14. |

21904. os sp., R. Etheridge, jr.: Second Rep. Geol. Surv. Natal

and Zululand, p. 88, pl. ui,

fig./o2¢ plan, fies is

1906. >. sp. Woods: Ann. S. Afr. Mus., vol. iv, p. 317) pr

KV eS. Ost: |

New Lamelhibranchia and Gastropoda of Pondoland. 213

Remarks.—The collection made by Gevers includes numerous,

mostly small, well-preserved examples of Gyrodes. Though these

show a certain amount of variation in form, the writer believes them

to belong to one and the same species, and can find no point of differ-

ence between them and G. tenellus, described by Stoliczka from the

Trichinopoli and Ariyalur Groups. The variation lies in the width

of the whorls and height of the spire—a common feature of Naticoid

shells, and in the depth of the depression of the posterior part of each

whorl.

Woods has figured two low-spired examples from Pondoland,

which resemble the figures given by Stoliczka very closely, as Woods

has observed. Some of the Pondoland shells at first sight appear to

be rather different by reason of their wider whorls and consequently

more elevated spires, but are connected with the low-spired forms

through intermediary varieties. The forms from Umkwelane Hill

figured by Etheridge most probably are but high-spired examples

of G. tenellus, though they are too poor to make the identification

absolutely certain.

In all varieties the aperture is large, the outer lip thin and the inner

lip without a deposit of callus. The umbilicus is very large and deep,

and is bordered by a moderately sharp spiral ridge, on the outside of

which is a second, more rounded ridge. The surface is covered with

numerous fine, oblique, growth lines; in one or two examples an

exceedingly fine spiral striation is discernible. The posterior portion

of each whorl is in some specimens only flattened, in others the

posterior shelf is concave, and the edge of the whorl rises to the

suture ; there does not appear to be any relation between the nature

of this part of the shell and the height of the spire.

The high-spired varieties approach G. pansus Stoliczka (61, p. 305,

pl. xxu, figs. 9-13) in form, but the crenulations characteristic of this

species do not occur in any of the Pondoland examples, and the species

attains to greater dimensions.

G. manuanensis R. B. Newton (48, p. 18, pl. vii, figs. 9, 10) from

Zululand, probably Senonian, is much larger and shows traces of

marginal crenulations.

<—

, {

“|

th i |

Hie

wae

10)

in| |

a i

5 Bt

| dell Hy

1 ei

. 7

1 Va in

i! y

ab |

18 t

}, |

ee es

;

| ,

| "

{ |

me be)

}

| |

a ein

| a |

a a

‘' {| i i

: | el

: oP Ty

: | }

i} yi |

ip Me) Tt)

y a

ip 1

| |

| P '

’ th 2 ’

" fl .

i hy j

We rm

: il hi Ps

. Cae

] ing

| q

iF i |

ee) miei |

| eo se

" it '

7 a

aH f

1) ¥

dea) ial {

| 1H Oo

pee { |

Y |

5 cH }

i it |

1 |

We S| fae

7 eG "hohe

a al ih

+ 1]

z i

’ |

\ 1H

) a hie

; On hei ba

He! inl |

ad ,' ||

u ya

oad] H

| ete) ah hal

y |

a : Mill

{ a |

mK Sah

| Ps}

i | |

q ii!

“he Ue

‘?

eatin |

4 aA a ba

’ I

whe

el ;

yah 4a) ;

cat bh /

} il

B.| & |

ify ;

hy al |

ai i i

OS, |

) a

rit ! /

at {! i

Pay pe! tal)

ny | af

eerie

tbe |

Be, | i

ei tt

ty) ia i

i heme

oe 4d aaa

ca) nt bl

Tele

|e I

ea; }

) ’

214 Annals of the South African Museum.

Famity TURRITELLIDAE.

Genus TURRITELLA, Lamarck.

Sub-genus HAUSTATOR, Montfort.

Turritella (Haustator) meadt, Baily.

(Pl. XXIV, figs. 12-15.)

1855. Turritella meadw, Baily: Q.J.G.S., vol. xi, p. 458, pl. xii,

fig. 6.

Material.—The collection contains numerous worn specimens and

casts of small Turritellids, which may be assigned to 7. meadi; S.A.

Mus., Nos. 8400, 8501, 8573 are fairly well preserved.

Description.—Shell small, turriteliform, of about 12 whorls, with

a spiral angle varying from 15° to 20°. Whorls flattened or very

gently convex towards the sutures; sutures well marked, varying

in depth; the anterior side of each whorl sometimes obscurely

carinated, with a very narrow area in front of it descending steeply

to the suture. Surface ornamented with numerous spiral riblets,

every second or third being slightly stronger than the others; and

covered with faint, simuous growth markings, which are a little

posterior to the middle of each whorl, deeply concave towards the

aperture.

Remarks.—The specimens seen by Baily are mostly casts, and the

figure given by him is a restoration based on a fragment. The

specimens in the present collection show great variation in the pro-

portions of the whorls; the spiral angle, which in Baily’s figure is

about 18°, varies from 15° to 20°, and the forms with the larger angle

tend to have deeper sutures than those with small angles. The

ornamentation, though variable and somewhat irregular in each

individual, is of the same type throughout.

The species is close to 7. imbricataria Lamarck, the type of

Haustator Montfort (6, vol. ix, p. 114, pl. vin, figs. 10, 11). The

three Senonian species of Haustator from Southern India, 7. dispassa

Stoliczka, T. pondicherriensis Forbes, and T. gemina Stol. (61, pp. 217,

218, pl. xvi, xix), are larger forms with different ornamentation and

not likely to be confused with 7. meadi. .

New Lamellibranchia and Gastropoda of Pondoland. 215

Genus ArcotTtiA, Stoliczka.

Arcotia vanhoepent, sp. nov.

(Pl. XXV, figs. 8-10.)

Material.—The holotype is a very well preserved specimen in the

collection of the Durban Museum; also S.A. Mus., No. 8393.

Description.—Shell small, elongately conical. Spire elongated,

composed of 5 moderately convex whorls, which are carinated in

front of the middle, the posterior part of the whorls somewhat

flattened and tapering posteriorly. Sutures well marked. Last

whorl large, well rounded, carinated, moderately flattened at the

base. Aperture of the holotype damaged on the exterior side, outer

lip missing; aperture more or less rounded, subangular on the

anterior side; plane of the aperture parallel to the axis of the shell ;

outer lip probably thin; inner lip smooth. Umbilicus open, narrow,

tube-like. Protoconch smooth, dextral, paucigyrous.

Ornamentation of relatively stout spiral ribs and transverse

tiblets. In front of the middle the carina 1s accentuated by a strong,

rounded, spiral rib. Posterior to the carinal rib, on the flattened part

of the whorls, are five weaker spiral ribs separated by concave inter-

spaces wider than themselves. Anterior to the carinal rib on each

whorl of the spire is a single spiral rib, in front of which is a steep

descent to the suture. On the rounded part of the last whorl there

are three moderately prominent spiral ribs separated by wide inter-

spaces, in each of which lies a very weak spiral rib. On the base

there are about seven moderately strong ribs, between each pair of

which is a weak rib. The spiral ribs and their interspaces are crossed

by transverse growth lines of considerable strength—they may almost

be termed riblets—which run almost straight from suture to suture,

but are very slightly concave towards the aperture. The intersection

of spiral ribs and transverse riblets results in a series of knots or

tubercles along the former.

Remarks.—Stoliczka founded Arcotia (61, p. 212) for a single

species from the Trichinopoli Group, A. indica Stol. (ebid., p. 215,

pl. xvi, fig. 12; pl. xix, fig. 6), the genus being thus defined: Are.

testa turrita, elongata, crassvuscula ; anfractibus spiraliter striatis ;

strus wmerementi rectis, non sinuatis ; columella excavata ; apertura

angulate-rotundata, antice subeffusa. The open umbilicus and straight

growth lines separated the genus from Twrritella. Cossmann (6,

vol. ix, p. 126) has regarded Arcotia as a synonym of Mesaha Gray,

216 — Annals of the South African Museum.

and stated that Stoliczka’s figures show that A. indica has the same

form, ornamentation, and aperture as the type of that genus. With

this view the writer cannot agree; the description and figures given

by Stoliczka indicate perfectly clearly the open umbilicus and straight

growth lines which characterise Arcotia ; on the other hand, Mesala |

invariably has the umbilicus closed and the growth lines and outer 4

lip possess the sinuous character seen in the majority of Turritellids.

Arcotia is therefore retained as a genus, and its close relation to

Turritella and to Mesalia is regarded as questionable.

A. vanhoepeni differs from the genotype in several respects, and is

not likely to be confused with it; the latter is a larger form with

the whorls more definitely flattened above the carinal rib, and has ‘

fewer spiral ribs, while the transverse lines are weaker and do not cut ,

the summits of the spiral ribs into tubercles. Twrritella (Zaria)

ventricosa Forbes (61, p. 227, pl. xvu, fig. 15; pl. xix, figs. 22, 23),

which was doubtfully referred to Mesalia by Cossmann, is certainly ,

not an Arcotia, and was probably rightly placed in Zaria by Stoliczka.

Arcotia margaritata Frech (22, p. 180, pl. xvi, fig. 13) from the ‘

L. Senonian of Central Europe has weaker spiral ribs and smaller .

spiral angle. :

Famity SCALIDAE.

Genus ConFrusiscaLa, de Boury.

Confusiscala ornata (Baily).

1855. Scalaria ornata, Baily: Q.J.G.S., vol. xi, p. 459, pl. xii, fig. 2.

1906. Scala ornata, Woods: Ann. S. Afr. Mus., vol. iv, pt. 7, p. 314,

pl. xxxviu, figs. 2, 3.

Remarks.—Cossmann (6, vol. ix, p. 75) places this species in the

genus Confusiscala de Boury, of which the type is S. dupiniana :

d’Orbigny, and to which S. decorata (Romer) also belongs. The x

species was compared with these European forms by Woods, but

the aperture has not yet been figured.

New Lamelhibranchia and Gastropoda of Pondoland. 217

Famity CHENOPIDAH.

Genus DicroLoma, Gabb.

Sub-genus PERISSOPTERA, Tate.

Dicroloma (Perissoptera) balyz (R. Etheridge, jr.).

(Pl. XXV, figs. 11-15.)

1904. Alaria? bailyr, Etheridge: Sec. Rep. Geol. Surv. Natal and

Zululand, p. 83, pl. iu, figs. 4-8.

1906. Dicroloma (Perissoptera) sp., Woods: Ann. 8. Afr. Mus., vol.

iv, p. 319, pl. xxxviul, fig. 14.

Material_—_Two specimens, both incomplete, one with the wing

perfectly preserved, T. Mus., Nos. 1612, 1613. A perfect specimen

in the collection of the Durban Museum, but with the surface somewhat

worn. Diagnosis partly after Woods and Etheridge.

Description.Shell elongate. Spire formed of 8 slightly convex,

non-carinate whorls; apical angle 36°; ornamented with strong,

oblique, slightly curved, transverse ribs, about 14 per whorl, separated

by broad, rounded interspaces; ribs and interspaces crossed by

numerous minute spiral ribs, which are coarser and more widely

separated just in front of the sutures than elsewhere, and are inter-

rupted by faint growth lines. Last whorl rounded, ornamented with

minute spiral ribs and faint growth lines; the transverse ribs are

represented by a few swellings on the part of greatest convexity,

giving the appearance of an indistinct ridge posterior to the middle

of the whorl. Anterior canal not seen. Outer lip thin, expanded

into a large wing which is truncated a little obliquely and is more or

less quadrangular in outline and without digitations, gently convex

on its outer edge, concave above and below, rather sinuous towards

the suture ; wing ornamented with irregular growth lines and minute

spiral ribs; wing reaching to about the middle of the penultimate

whorl.

Remarks.—The writer does not hesitate to identify the Pondoland

shells with the species figured by Etheridge from Umkwelane Hill.

On comparison with the present specimens, however, it would seem

that the strength of the spiral ribs towards the posterior side of each

whorl has been slightly exaggerated in Etheridge’s figures. The

shape of the wing and its attitude (the angle between wing and spire

is rather greater than 90°), together with the relatively small apical

218 Annals of the South African Museum.

angle and relatively few transverse ribs, separate this shell from other

Senonian species. D. schlotheima (Romer) from Aachen (28, vol. vi,

pl. vu, fig. 13) seems to be very close to it, but has more numerous

transverse ribs, and the wing 1s expanded to a certain extent posteriorly.

D. papilionacea (Goldfuss) (24, vol. iti, p. 18, pl. clxx, fig. 8) is a very

different species ; judging from the figure given by Goldfuss, a larger

shell with more numerous transverse ribs which are continued over

the body whorl; subsequent authors, however, have assigned to

D. papilionacea forms in which the transverse ribs are more or less

obsolete on the body whorl, as in our species, and less numerous on

the spire; the forms described by Stoliczka from Southern India

(61, p. 31, pl. ui, figs. 9, 10) do not appear to be identical with the

type of Goldfuss, and have the body whorl free from transverse ribs,

but the examples figured are considerably larger than D. baily.

Dicroloma (Perissoptera) sp.

(Pl. XXV, figs. 16-18.)

Remarks.—One specimen in the collection, S.A. Mus., No. 8699,

differs from D. bailyi in that the strong transverse folds are not

developed. The surface is covered with strong transverse lines and

riblets of varying strength, and with faint spiral lines which are

strongest just in front of each suture. It may be but a variety of

D. bailyr.

Dicroloma (Perissoptera) sip.

(RE SOXVe nes 195220:)

Remarks.—8.A. Mus., No. 8494 is a small Perissoptera with strong

transverse ribs as in D. bailyr. It differs from that species, however,

in that the entire shell is covered with fairly strong spiral ribs, which

are stronger and less numerous than the corresponding spiral lines of

D. bailyi.

Famity STROMBIDAE.

Genus PuUGNELLUS, Conrad.

Pugnellus contortus (Forbes).

(Pl. XXV, figs. 24-26.)

1846. Strombus contortus (G. B. Sowerby MSS.), Forbes: Trans.

Geol. Soc. Lond., vol. vu,

p. 129, pit xv eo:

New Lamellibranchia and Gastropoda of Pondoland. 219

1871. Pugnellus contortus, Stoliczka: Cret. Gastr. S. India, p. 19,

pl i, figs. 1-5.

Material.—A single specimen with the anterior missing but with

the outer lip intact, T. Mus., No. 1611.

Description.—Shell very thick and solid, elongate in the direction

of the axis, rounded, consisting of about 6 whorls, with a spiral angle

of about 60°. Spire much shorter than the last whorl, of 5 whorls,

almost entirely covered by callus, the sutures just visible dorsally.

Last whorl rounded ; with about 7 stout transverse ribs in the dorsal

half, placed obliquely, very slightly curved, thickened into tubercles

along the region of greatest convexity, thus giving to the whorls a

slight angularity; these ribs die away anterior to the tubercles ;

surface otherwise smooth. Aperture long, moderately wide. Outer

lip thick, expanded into a large wing, with a very stout lobe which

has a pointed, projecting process on the posterior end and a sharp

ventral crest; on the inside of the lobe is a flattened area with a

stout median ridge extending from an anterior to a posterior notch.

Inner lip much thickened with callus, which extends over all the

ventral part of the last whorl and over most of the spire right up to

the apex.

Remarks.—This fine species is according to Stoliczka very char-

acteristic of the Trichinopoli Group. The Pondoland specimen agrees

very closely with the figures of a well-preserved specimen given by

that author (figs. 1, la—-d). P. uncatus (Forbes) (61, p. 22, pl. ii,

figs. 9-13) is readily distinguished by the more delicate and more

humerous curved ribs and more tapering anterior.

P. crassicostatus Noetling from the Maestrichtian of Baluchistan

(46, p. 61, pl. xvi, figs. 3, 4) is founded on casts and appears to be

closely allied to P. uncatus. This species has been figured by Boule

and Thevenin from Madagascar (8, p. 46, pl. 1, fig. 3), but the identi-

fication is based on imperfect material.

P. auriculatus Woods (77, p. 319, pl. xxxviii, fig. 15) from Pondo-

land has more angular whorls and different ornamentation.

220 Annals of the South African Museum.

Famity CYPRAEIDAE.

Genus Cypragza, Lamarck.

Cypraea chubbi, sp. nov.

(Pl. XXVI, figs. 1-6; pl. XXVII, figs. 1-4.)

Materval.—The holotype is a very large example in the collection

of the Durban Museum. Apparently belonging to the same species

are two smaller examples in the same collection, and a very small

shell, T. Mus., No. 1624.

Description.—Shell large, robust, ovoid, longer than high, narrower

in front than behind, with flattened base. Highly involute, the

spire visible but small. Base smooth, periphery of the base coarsely

crenulate ; dorsal surface covered with numerous low tubercles,

which are arranged in oblique rows. Aperture moderately wide,

elongate, gently curved, narrow behind, widening gradually in front.

Posterior notch deep, rounded in posterior view, with thick, slightly

extended lips which form a very short canal. Anterior notch narrow,

twisted in the direction of the termination of the apex, probably

without extended lips. Columella smooth, without folds, and bearing

a slight depression. Outer lip straighter than the inner, with a

crenulated margin (about 18 “teeth’’), thick; inner lip less well

defined, inflated, obscurely crenulated at the anterior end.

Remarks.—The two remaining specimens in the Durban Museum

Collection are smaller, and although they resemble the holotype in

shape and in the nature of the aperture, they differ in certain ways

which are probably due to the fact that they have not reached the

adult state. The specimens are somewhat worn, but nevertheless it

would seem that the dorsal surfaces possessed traces of the tubercles

that characterise the adult. In one specimen the base is only slightly

flattened, and without the coarsely crenulated periphery which in

the adult greatly accentuates the flattening of the base; at the same

time the posterior notch is without the incipient canal seen in the

adult. In the other specimen the periphery of the base is partially

produced and very slightly crenulate, and the posterior notch is a

little sub-canaliculate; the specimen is slightly smaller than the

other, but appears to be intermediate in character between it and the

holotype.

T. Mus., No. 1624, is very much smaller than the above specimens,

but is of exactly the same proportions, is quite smooth, and with

<< Ss a

——— ee ea ee

a eo)

SS Ve oe ae

New Lamellibranchia and Gastropoda of Pondoland. 221

simple anterior and posterior notches. The spire is barely visible.

The aperture is like that of the larger examples. It is very probably

a young form of the above.

It is very interesting to find specimens of this genus, so rare in

Cretaceous rocks, among the Pondoland forms. C. chubbi is unique

among Cretaceous members of the genus in size and ornamentation.

The two smaller Durban Museum specimens are very like the forms

for which Jousseaume created the section Bernayia (6, vol. v, p. 156) ;

they agree with Bernayia in proportions, smoothness of shell, and in

the structure of the columella and aperture. Like Bernayia the

posterior notch is provided with slightly projecting lips, and the

columella is smooth, without folds, and with a slight depression.

The ornamentation, crenulate periphery, and sub-canaliculate posterior

notch of the holotype at once differentiates the species from Bernayia,

to which, however, it is probably allied.

The smallest specimen resembles some of the forms described by

Forbes and Stoliczka from Southern India. C. cunliffi Forbes

(61, p. 450, pl. xxviii, fig. 29) has a smaller posterior notch and the

surface is finely reticulated. C. (Epona) globulina Stol. (61, p. 451,

pl. xxviii, fig. 30), like the preceding from the Ariyalur Group, is

more inflated and apparently possesses small projecting lips both

anteriorly and posteriorly. The only European form at all like it

seems to be the Danian C. bullaria (von Schlotheim) (50, p. 22, pl. u,

figs. 4, 5), but the latter is relatively shorter and with a more curved

aperture.

Famity DOLIIDAE.

Genus Prruta, Lamarck.

Sub-genus PROTOPIRULA, Nov.

This sub-genus is proposed for the Pondoland species described

below, P. capensis sp. nov., characterised by the absence of the

typical ornamentation of Pirula s.s. and by the poor development of

the anterior canal.

According to Fischer (19, p. 662) and Cossmann (6, vol. v, p. 140)

the genus is characterised by having a thin, piriform shell, with the

surface reticulated or nodular, the aperture large and elongated, the

canal long and arched, the lip smooth and thin, the columella smooth

and sinuous. The genus appears in the earliest Tertiary, and increases

until the present day, forming a small but natural group, all the

members of which are characterised by the marked reticulated

222 Annals of the South African Museum.

surface and long canal, though some forms are more or less nodular

or sub-carinated. Protopirula is proposed for some well-preserved

specimens, which differ from the genotype in that they possess no

ornamentation other than faint growth lines and in not possessing

a produced anterior canal. It may be regarded as ancestral to

Pirula s.s., just as the Upper Senonian Protodolium Wilckens, P.

speightt (Trechmann), of New Zealand (74, p. 18, pl. iv, figs. 3-5), is

ancestral to Doliuwm, and differs from the latter in the absence of the

canal. Both the genera of Dolidae are thus seen to have Upper

Cretaceous forerunners in which the canal is absent or poorly

developed.

Porula (Protopirula) capensis, sp. nov.

(Pl. XXV, figs. 21-23.)

Materval.—Two well-preserved shells. Holotype, T. Mus., No.

1626 ; paratype, T. Mus., No. 1625. Also three imperfect specimens

in the collection of the Durban Museum.

Description.—Shell moderate in size, highly involute, thin. Spire

very small, consisting of about two whorls, pointed. Last whorl

forming most of the shell, oval, elongated and convex, tapering very

gradually towards the anterior. Aperture large, outer lip thin,

smooth, and convex; inner lip smooth; anterior not produced into

a canal but somewhat truncate. Ornamentation solely of faint

growth striations, gently convex towards the aperture.

Famity BUCCINIDAE.

Genus TRITONIDEA, Swainson.

Sub-genus CANTHARULUS, Meek.

Tritonidea (Cantharulus) kaffraria, sp. nov.

(Pl. X XV, figs. 27, 28.)

Materval.—The holotype is an almost perfect specimen in the

collection of the Durban Museum.

Description.—Shell ovate, fusiform. Spire composed of 5 turreted,

carinated whorls; spire less than the body-whorl in height; the

carina on each whorl is situated a little in front of the middle and

divides the whorl into a flattened, sloping, posterior part and a

narrower, flattened, anterior part. Sutures well marked. Body-whorl

New Lamellibranchia and Gastropoda of Pondoland. 223

large, inflated, rounded, only obscurely carinated. Aperture large,

ovate; outer lip thin; columella twisted, with a swelling below;

inner lip large, smooth, with a moderately thick callus.

Surface coarsely and somewhat irregularly ornamented. Whorls

of the spire with about 15 transverse ribs with furrows between

of about the same width; these ribs are obliquely placed, swell

into tubercles along the carina and weaken towards the posterior

suture; they are crossed by 6 or 7 narrow, widely separated spiral ribs.

On the body-whorl the transverse ribs are somewhat irregular and

nodular where crossed by the spiral ribs; finer spiral riblets are

intercalated between the principal spiral ribs. On all the whorls

there are numerous fine, rather sinuous, transverse riblets, or promi-

nent growth lines. ;

Remarks.—Meek formed the section Cantharulus for the Missouri

species Fusus vaughani Meek (87, p. 379, pl. 32, fig. 5, text-fig. 48 ;

6, vol. iv, p. 173, pl. vu, fig. 1), which was placed under Tritonidea

by Cossmann. JT. kaffraria differs from the type mainly in having

a wider aperture, but agrees with it in the structure of the columella,

in general shape and in the nature of the ornamentation. Cossmann

refers to the section, but with some doubt, 7. pondicherriensis (Forbes)

(61, p. 127, pl. xi, figs. 10-12) from the Trichinopoli Group, and T.

gosauicum (Zekeli) (79, p. 82, pl. xv, fig. 1). The latter is much

smaller than the Pondoland form, has fewer transverse ribs, and is

not carinated. The Indian species resembles T. kaffraria rather

closely, and has the same ovate aperture, but the whorls of the spire

are not carinated, and the species differs in the details of the orna-

mentation.

Famity CHRYSODOMIDAE.

Genus SrpHonaLia, Adams.

Siphonalia * umzambiensis, sp. nov.

(RE XOX) tie, 5.)

Material._—Holotype S.A. Mus., No. 8416; paratype, S.A. Mus.,

No. 8509.

Description.—Shell fusiform. Spire turreted, composed of 4

convex non-carinated whorls, less than the body-whorl in height ;

spiral angle about 55°. Last whorl large, rounded. Aperture

ovate. Outer lip not seen; inner lip smooth and without colu-

mellar folds.

224 Annals of the South African Museum.

Ornamentation of strong transverse folds separated by rounded

furrows of about the same width, folds and furrows crossing the whorls

from suture to suture; about 12 folds to the volution. These are

crossed by numerous narrow, spiral ribs, which are separated by

interspaces wider than themselves; on the posterior part of the last

whorl additional weaker spiral ribs appear, one between each pair of

principal spiral ribs. There are in addition fine growth and spiral

lines forming a network.

Remarks.—This species appears to have the form and ornament of

Siphonaria, but the absence of the anterior canal in both specimens

introduces some uncertainty.

The Trichinopoli shells figured by Stoliczka (61, p. 124, pl. mi,

figs. 8, 9) as Tritonidea requieniana (d’Orbigny) may be identical with

this species, or are at any rate closely allied. Stoliczka thought that

the Indian forms belonged to d’Orbigny’s Fusus requieniana (12, p. 342,

pl. cexxy, fig. 3), with which F. buchi Miller (40, p. 35, pl. v, fig. 15)

was identified. Holzapfel has since studied F. buchi and placed it in

Chrysodomus (28, p. 102, pl. x, figs 9-12), a genus which is, however,

without strong axial ornament; at the same time the identity of the

species with that of d’Orbigny was denied.

S. umzambiensis has the axial folds more strongly developed

than S. buchi, and the radial ribs are fewer and more prominent.

D’Orbigny’s species is much larger, and appears to be characterised

by relatively few and hence widely separated transverse folds, and

the anterior canal is straight.

Famity FUSIDAE.

Genus LatHyrus, Montfort, em. Cossmann.

Sub-genus MAzZZALINA, Conrad.

Lathyrus (Mazzalina) geverst, sp. nov.

(Pl. X XVII, figs. 6-8.)

Material.—One well-preserved specimen, 8.A. Mus., No. 8486.

Description.—Shell small, ovate. Spire rather short, of about

4 moderately convex whorls; spiral angle about 70°; sutures well

marked. Last whorl much larger than the spire, rounded, attenuate

anteriorly. Aperture elongately oval, produced into a short, slightly

bent, anterior canal. Inner lip curved, smooth; columella provided

New Lamellibranchia and Gastropoda of Pondoland. 225

towards the anterior with three relatively stout, slightly oblique folds,

of which the anterior is placed a little distance away from the other

two, which are close together.

Surface ornamented with low, rounded, spiral riblets, and shallow

furrows of about the same width as the riblets. The riblets are most

numerous towards the anterior of the last whorl and are obsolete on

the part of greatest convexity ; a few riblets are rather prominent just

behind the part of greatest convexity, both on the last whorl and on

the whorls of the spire; there is one rather deeply impressed furrow

close to the suture on the posterior side of each whorl. The shell is

covered with transverse growth striae; the penultimate whorl bears

obscure, transverse folds, which are also traceable on part of the last

whorl on the part of greatest convexity.

Remarks.—The character shown by this shell is quite in agreement

with the description of Mazzalina given by Cossmann, and with his

figure of the species L. stamanea (Tate) (6, vol. iv, p. 51, pl. ii, fig. 17).

The only Cretaceous species indicated by this author is Fasciolaria

buccinoides Meek (87, p. 358, pl. xxxi, fig. 8), from the Fox Hills

Group of Missouri, a larger and more evenly ornamented species ; it

is very like L. geversi in shape, and sometimes shows similar weak

transverse folds on the convex part of the whorls.

The single specimen from the Ariyalur Group, described by Stoliczka

as Neptunea rhomboidalis (Zekeli) (61, p. 120, pl. x, fig. 21), appears to

be rather like the Pondoland shell, but the anterior canal is more

twisted ; columellar folds are not indicated in the figure, but if they

are present, then the shell is certainly not identical with the Gosau

forms described by Zekeli (79, p. 80, pl. xiv, fig. 9), and referred to

Sycum (Palaeatractus) by Cossmann (6, vol. iv, p. 83). It is possible

that Stoliczka’s Volutonutra canaliculata from the Trichinopoli Group

(61, p. 100, pl. 1x, figs. 12, 13) belongs to Mazzalina, but the species was

founded on imperfect material.

Genus CryPToRHYTIS, Meek.

Cryptorhytis rigida (Baily).

(Pl. XXVIT, figs. 9-12.)

1855. Voluta rigida. Baily: Q.J.G.8., vol. x1, p. 459,

pl. xu, fig. 14.

Non. 1868. Fasciolarva rigida. Stoliczka: Cret. Gastr. 8. India,

pmlOo woljex, wes. LONG.

ae 7S

= =

= —- “g 2 = —_-—— ——s _ = eo

i ~~ “ —F = . = r= ° =

=k bas inne ee , r : oe b . 7 ~ : ~ = -

* — ef _ = ae Ss : =<“, = = :

: z = : aa LEO 7 ~ See aia x- <

4 *- 7 7 - > ee a, —— = ae > - ee ea: we z

i Fa - = = - — - =e =

=: = = a ~ < < = aS se

226 Annals of the South African Museum.

Non. 1906. Cryptorhytis rigida. Woods: Ann. 8. Afr. Mus., vol.

iv, p. 321, pl. xxxix, fig. 2ipiad

fig. 1.

Material.—Three specimens in the collection of the Durban Museum

have been seen by the writer. Two of these are large, well preserved,

but incomplete anteriorly; the third is about half grown, and is

beautifully preserved.

Description.—Shell fusiform, scalariform. Spire shorter than the

last whorl, consisting of five convex whorls with deep sutures.

Posterior part of whorls with a concave depression or sulcus, having

a tumid margin or ridge near the suture. Spiral angle, 55°-60°.

Ornamentation consists of numerous, broad, rounded, transverse

folds, which are usually indistinct on the posterior sulcus, but reappear

on the tumid posterior margins of the whorls, and are continued

anteriorly over the greater part of the last whorl. The transverse

folds are from 14-19 per volution, and are separated by deep, con-

cave grooves of about the same width as themselves. Numerous

moderately strong spiral ribs are present, and are separated by

interspaces wider than themselves; the spiral ribs become stronger

on the anterior part of the last whorl. Growth lines are present.

Aperture oval-elongate, angular behind. Outer lip not clearly seen.

Inner lip thin, concave. Columella without folds. Anterior canal

moderately long, shghtly bent.

Length (of young example), 14°5 mm.; breadth, 8 mm.

Remarks.—Baily’s type has been lost. The following is the original

description :—

“VY. testa oblonga; spira conica ; anfractibus superne angulatis,

longitudinaliter costatis ; costis im angulis turgidis, et prope suturam

subobsoletis, spwraliter sulcatis ; apertura angulata, elongata.

‘“ Leneth of broken specimen, 1;’5 inch; probable entire length,

13% inch.

‘“ Breadth, 4$ths of an inch.

“Shell somewhat cone-shaped, with about 18 longitudinal ribs,

coarsely striated spirally. The summit of the whorls is formed at the

suture into a rim with slight continuations of the longitudinal costae

and striae of growth. The plications of the columella are not exposed.”

The description is accompanied by a single figure showing the dorsal

view of a nearly perfect specimen.

The specimens which were well described and figured by Woods as

C. rigida belong to a different but closely allied species, to which the

name C. pseudorigida sp. nov. is given below. The writer has had the

New Lamellibranchia and Gastropoda of Pondoland. 227

opportunity of comparing three examples of C. rigida with two of

C. pseudorigida, all being well preserved. The species differ chiefly in

the number and disposition of the transverse ribs. In the former the

transverse ribs are 14-18 in number, and they are continued anteriorly

down the slope of the base of the body-whorl, becoming extinct before

the canal is reached ; posteriorly the ribs are thickened very slightly

just in front of the sulcus, where they disappear, but their positions are

indicated behind the sulcus by fairly well-defined swellings on the

tumid posterior edges of the whorls; the grooves between the transverse

ribs are concave and straight; the spiral ribs are moderately strong

and numerous. In C. pseudorigida the transverse ribs are 10-12 in

number, and they are not continued much in front of the point of

greatest convexity of the body-whorl ; posteriorly the ribs are broadest

in front of the sulcus; the transverse ribs are absent from the sulcus

and from the tumid posterior edges of the whorls ; the grooves betweep

the transverse ribs are shallow and ill-defined; the spiral ribs are

much stronger than in C. rigida and are not quite as numerous. The

differences given above are clearly brought out in the figures given by

the writer and those given by Woods, and there can be little doubt that

Baily’s figure belongs to the species here regarded as C. rigida.

Examples from the Trichinopoli Group (61, p. 109, pl. x, figs. 10-16)

were identified by Stoliczka with the Pondoland species, but the Indian

form is larger and more elongated, with a smaller spiral angle, and

with three columellar folds, besides differing in the details of the

ornamentation.

Baily compared his specimen with Voluta cincta Sowerby, which

Stoliczka regarded as probably identical with Hemifusus cinctus Stol.

(61, p. 114, pl. x, figs. 17, 18), which is rather similar to the present

species but has a shorter spire, broader body-whorl, and more numerous

transverse ribs. Cossmann (6, vol. iv, p. 72) has cited this species as

probably belonging to Gabb’s section Perissolax, but it is more like

Cryptorhytis in proportions.

Cryptorhytis pseudorigida, sp. nov.

1906. Cryptorhytis rigida. Woods: Ann. S. Afr. Mus., vol. iv,

pr ook, plo socxix fie. 2; pl. xls. 1.

Non. 1855. Voluta rigida. Baily: Q.J.G.S., vol. xi, p. 459, pl. Xl,

fig. 14.

Remarks.—The differences between this species and C. rigida have

been given above. The writer selects as holotype the specimen from

VOL. XXVIII, PART 2. 16

228. Annals of the South African Museum.

which were taken Woods’ figs. 2a, 2b, and 2c, pl. xxx1x (in the collection

of the South African Museum). The specimens used by the writer in

arriving at this determination are 8.A. Mus., Nos. 8596, 8717.

Cryptorhytis, ? sp.

(Pl. XX VII, figs. 13, 14.)

Remarks.—The collection includes two examples, 8.A. Mus., Nos.

8523, 8640, of a small fusiform shell which is probably a Cryptorhytis.

As the aperture and anterior canal are missing in both specimens,

the generic position cannot be definitely ascertained. Externally it

resembles the Cryptorhytis sp. figured by Woods (77, p. 322, pl. xl,

fig. 2) from the same deposits, but is much smaller and the spiral ribs

anterior to the angulation of the last whorl are decidedly weaker and

narrower. C. pseudorigida is more elongate and has more prominent

transverse folds.

Genus PALEOPSEPHAEA, Wade.

Paleopsephaea scalaris, sp. nov.

(Pl. X XVII, figs. 23, 24.)

Material.—A single complete specimen, in places with the surface

badly preserved, T. Mus., No. 1606.

Description.—Shell fusiform, much elongated. Spire of about

6 whorls, longer than the last whorl; whorls convex, sutures well

marked. Last whorl large, gently convex, tapering anteriorly.

Spiral angle 35°. Aperture elongated, narrow, angular above, pro-

duced anteriorly into a long, slightly curved canal. Outer lip thin,

slightly convex ; columella with folds, one oblique fold being visible.

Whorls with stout transverse folds, about 10 per whorl, stretching

from suture to suture, with elevated crests, accentuating the convexity

of the whorls, the folds on successive whorls contiguous; folds

separated by broad, rounded interspaces, almost smooth, but with

very faint indications of growth and spiral striae.

Remarks.—This species appears to be closely allied to P. mutabrlis

Wade (67, p. 123, pl. xl, figs. 4, 5, 8) from the Upper Cretaceous of the

Ripley Formation of Tennessee, which is the type of Paleopsephaea

Wade, placed by that author in the Volutidae, here regarded as allied

to Cryptorhytis. It differs from the type in having less swollen, sharper,

transverse ridges and fainter spiral striation. The genus does not

. =

SS ae 35.

wa =

SS SS SES TSS

New Lamellibranchia and Gastropoda of Pondoland. 229

appear to be known from Southern India, and the European species

placed in the genus by Wade, Fusus subsemiplicatus d’Orbigny, Voluta

roemert Geinitz and Mitra nana Miiller (for references, see Holzapfel,

28, pp. 95-100, and Wade, 67, p. 123) are more remotely related to the

type than is the Pondoland form.

Paleopsephaea kaffrarva, sp. nov.

(PEOCXVIL, figs. 155 16.)

Material.—Two specimens, including the holotype, in the collection

of the Durban Museum; also S.A. Mus., Nos. 8479, 8505, 8632.

Description.—Shell fusiform, elongated. Spire of about 6 whorls,

longer than the last whorl; whorls convex, sutures distinct. Spiral

angle about 35°. Aperture elongated, narrow, angular above,

produced anteriorly into a long canal. Outer lip thin; inner lip

slightly concave; columella without folds. Whorls with stout,

rounded, transverse folds, about 10 per whorl, stretching from suture

to suture; and with numerous strong, narrow, spiral ribs, which are

separated by interspaces wider than themselves.

Remarks.—This species resembles the preceding in form, apical

angle and absence of the posterior sulcus, but differs in being without

columellar folds and in having strong spiral ribs. It is not unlike

P. pergracilis Wade (67, p. 124, pl. xl, figs. 3, 7), but has more trans-

verse folds and no columellar folds.

Famity TURBINELLIDAE.

Genus Pyropsis, Conrad.

Pyropsis geversi, sp. nov.

(Pl. XXVII, figs. 17-19.)

Material.—The holotype was collected by Gevers, and is now in

the collection of the South African Museum, No. 8551. There is

also a specimen in the collection of the Transvaal Museum, No. 1616,

and one in the Durban Museum collection.

Description.—Shell relatively small, thick, globose, consisting of

about four whorls. Spire very short, whorls concave, sutures well

marked. Last whorl very large, forming the bulk of the shell, with

a Carina posterior to the middle line ; anterior to the carina the whorl

is convex, ornamented with three or four spiral ribs, which together

230 Annals of the South African Museum.

with the carina bear stout tubercles at regular intervals where they

are crossed by growth ridges; posterior to the carina the whorl is

concave, rising towards the suture into a prominent tubercled ridge ;

surface with numerous growth lines and at regular intervals strong

growth ridges. Aperture circular; outer lip thick, bevelled, smooth

within: inner lip large, smooth. Anteriorly apparently expanded

into a long canal, the base only preserved in the present specimens ;

last whorl constricted where it merges into the canal.

Remarks.—This fine species appears to be very distinct, the more

noteworthy features being the strong tuberculation, the circular

outline of the aperture, which is only slightly angular at the carina

and at the anterior and posterior end, and the anterior constriction

of the last whorl. It is thus very different from the large P. africana,

described by Woods (77, p. 322, pl. xxxviu, fig. 17; pl. xxxix, fig. 1)

from the same beds. The species recalls the smoother Tudicla eximia

Stoliczka (61, p. 151, pl. xu, figs. 5-8) from the Ariyalur Group. It

resembles to a certain extent P. gracilis Wilckens (72, p. 23, pl. i,

fig. 16), but this Patagonian form is more oblique and with a more

prominent spire ; the species has also been figured by Wilckens from

Antarctica (78, p. 93, pl. iv, fig. 7).

Genus SEMIFUSUS, Swainson.

Subgenus MAYERIA, Bellardi.

Semifusus (Mayeria) acuticarinatus, sp. nov.

(Pl. XXVII, figs. 25, 26.)

1906. Semifusus ? (Mayeria?) sp., Woods: Ann. 8. Afr. Mus.,

vol. iv, p. 324, pl. xl, fig. 5.

Material.—One specimen, almost complete, T. Mus., No. 1620.

Description.—Shell fusiform, elongated, consisting of 6 whorls.

Spire about as long as the last whorl (excluding the extended anterior

canal) ; whorls angular, with a remarkably high, laterally compressed,

sharp carina, which is placed anteriorly to the middle of the whorls,

dividing the whorls into a wide, concave area on the posterior side

and a narrower concave area on the anterior side; a second much

smaller carina is placed on the anterior side of the whorls close to the

shallow but well-marked suture. Anterior much produced into a long,

straight, anterior canal. Aperture elongated, angular above, narrow,

outer lip smooth. Inner lip without callus or columella folds. Orna-

New Lamellibranchia and Gastropoda of Pondoland. 231

mentation of numerous small spiral ribs, varying in strength, separated

by spaces wider than themselves ; ribs crossed by faint growth striae

which are concave towards the aperture.

Remarks.—The writer has compared this species with specimens of

the Oligocene S. (Mayeria) errans (Soland.) (6, vol. iv, p. 94, pl. iv,

fig. 13) and does not hesitate to place it in the same group, for it

agrees very closely in proportions, shape of the whorls and aperture,

and in the nature of the ornamentation. Its nearest ally seems to be

the Lagena secans of Stoliczka (61, p. 138, pl. xi, figs. 19, 20), founded

on imperfect material, but the Ariyalur species differs in that the

second carina, which on the last whorl is rather strongly developed,

does not appear on the whorls of the spire as it is embraced by suc-

ceeding whorls; this species is regarded by Cossmann as probably

belonging to Mayeria (loc. cit., p. 94).

Semifusus (Mayeria) umzambiensis, sp. nov.

(Pl. XXVIII, figs. 1, 2.)

Materral.—A single complete specimen, T. Mus., No. 1617.

Description.—Shell fusiform, moderately elongated, consisting of

about 6 whorls. Spire shorter than the last whorl; whorls angular,

with a conspicuous carina, placed near the suture on the anterior

side of the whorls, dividing the whorls into a broad, flattened, or

slightly concave posterior area and a narrow, steep, anterior area.

Carina on the last whorl with stout conical tubercles, about 11 to

the half-volution ; tubercles slightly elongated in the direction of the

growth striae. Anterior tapering, produced into a long anterior

canal. Aperture oval and angular above, narrowed and elongated

anteriorly. Without callus or columellar folds. Ornamentation an-

terior to the tubercled carina of the last whorl of numerous, widely

separated, sharp ribs; surface with well-marked sinuous growth

striae, concave towards the aperture.

Remarks.—This species agrees closely with S. (Mayeria) errans

(cited above) and must undoubtedly be placed in the same group.

There do not seem to be any Cretaceous species with which it can be

compared. |

Semifusus (Mayeria) kaffrarvus, sp. nov.

(Pl. XXVIII, figs. 3-6:)

1906. Semifusus? (Mayeria?) sp., Woods: Ann. S. Afr. Mus.,

VoOliiv.p. o20, pl. xl, fot.

232 ; Annals of the South African Museum.

Material.—Holotype, T. Mus., No. 1618; paratype, T. Mus., No.

| 1619. There is also one specimen in the Durban Museum collection.

my (| Description.—Shell fusiform, of about 5 whorls. Spire shorter than

i | the last whorl, whorls angular and sharply carinate, the carina situated

close to the anterior side, dividing the whorls into a flattened or

slightly concave posterior area and a narrow, steep, anterior area.

Carina on last whorl, with a few broad swellings or tubercles. Aper-

ture sub-circular, outer lip thickened; inner lip smooth, with a

deposit of callus ; anterior expanded in a short, slightly twisted canal.

} Ornamented with small, spiral, widely separated ribs anterior to the

||| carina of the last whorl, fine spiral lines, and growth striae; behind

the carina on the last whorl there are only sinuous growth striae and

fine spiral lines, the latter sometimes present near the carina only.

Remarks.—This species is closely allied to the last, with which it

agrees very closely in the shape of the whorls and ornamentation.

Lit It is readily distinguished by its relatively shorter canal and by the

i shape of the aperture. In S. wmzambiensis the outer lip is markedly

angular and the aperture elongate, whereas in S. kaffrarius the carina

disappears towards the aperture, so that there is no trace of angularity

in the ventral view, the aperture being almost circular. This feature

is seen even in immature forms (T. Mus., No. 1619).

Famity ACTAEKONIDAE.

Genus ACTAEONELLA, d’Orbigny.

—<, SS > ae Sie

1é - — er 2 Pa pas < = 2 ee

ae eS Se JF FS syer=

Ra A * : = at

22S eS SSS SS =—

aes ——

Sub-genus TROCHACTAEON, Meek.

cara

Actaeonella (Trochactaeon) woodsi, sp. nov.

(Pl. XXVIII, figs. 17-21.)

1906. A. (Trochactaeon) sp., Woods: Ann. 8. Afr. Mus., vol. iv,

Po20s, pla sday hee te

Material.—Three specimens, T. Mus., Nos. 1621-1623; holotype,

T. Mus., No. 1621. Also five examples in the collection of the Durban

Museum. .

Description.—Shell elongately oval, stout. Spire of about 6

| whorls, moderately convex, the shape obscured by a row of very

| large, stout tubercles; spiral angle of 70°-80°. Last whorl longer

| than the spire, cylindrical in form, tapering gradually towards the

anterior; base tapering more rapidly, posterior part with a sloping

ledge adjacent to the well-marked but shallow suture, the upper part

S= ES

e.. i

— =

a a

SS=

= ==

ee ==

—

= —-

_———

New Lamellibranchia and Gastropoda of Pondoland. 233

of the whorl ornamented with 10-12 very stout tubercles. Surface

smooth, except for faint growth lines on the last whorl, slightly convex

towards the aperture. Columella short, with three strong folds, of

which the posterior is much stronger than the other two, and the

anterior is weakest. Outer lip thin. 7

Remarks.—This fine species is very closely allied to A. beyricha

Drescher (see references in Woods), but should be regarded as a distinct

species. The European species has in general a shorter spire and a

more rounded and less cylindrical shape, and fewer tubercles along

the posterior side of the whorls; in these features, however, it is

very variable. The figures given by Drescher also show that the

three columellar folds are equally strong, whereas, in all the Pondoland

specimens examined, the posterior fold is much more strongly developed

than the other two. A. beyrichi is a Lower Senonian form.

A. cretacea (Miiller) (28, p. 82, pl. vii, figs. 11, 14-16), from the

Aachen Greensand, appears to be very similar but is too poor for

exact comparison.

A. cylindracea (Stoliczka) (61,'p. 419, pl. xiv, figs. 10-14), from the

Utatur and Trichinopoli Groups, is without tubercles, and the whorls

taper both anteriorly and posteriorly.

A. salomoms Fraas (10, p. 383, pl. xxxv; figs. 3-6) is devoid of

tubercles and has a more tapering spire.

Two specimens in the Durban Museum collections are almost

completely devoid of tubercles, but traces can be made out, their

apparent absence being due at least in part to the state of preserva-

tion ; the tubercles may not have been strongly developed.

Famity BULLIDAE.

Genus BULLINELLA, Newton.

(=Cyxicuna, Loven.) ©

Bullinella sp.

(Pl. X XVII, figs. 20-22.)

Material.—One imperfect specimen, S.A. Mus., No. 8527.

Remarks.—Shell small, cylindrical; aperture rounded below and

angular above ; surface with spiral lines separated by broad, smooth

interspaces. The specimen is too imperfect for exact comparisons,

the anterior and posterior being considerably damaged. It is com-

parable with B. inermis (Stoliczka) (61, p. 431, pl. xxvii, fig. 20) from

234 Annals of the South African Museum.

the Ariyalur Group, and B. thomsoniana (Wilckens) (74, p. 24, pl. v,

figs. 6, 7) from the Upper Cretaceous of New Zealand. Two species

of Cylichna have been described by Etheridge from Umkwelane Hill,

C. fusuliniformis (17, p. 87, pl. ui, figs. 33, 34) and C. griesbachi (abid.,

p. 86, pl. in, figs. 14, 15).

Famity RINGICULIDAE.

Genus Rineicuua, Deshayes.

Ringicula woodsi, sp. nov.

(Pl. XXVIII, figs. 7-11.)

Material.—The holotype, 8.A. Mus., No. 8528, is well preserved,

but the aperture is damaged and partly concealed, and the callus

on the inner lip is only partly visible. A second example, S.A. Mus.,

No. 8493, is less well preserved and traces only of the ornamentation

appear on the last whorl, but portions of the aperture are visible,

showing callus, columellar folds and part of the outer lip. Also S.A.

Mus., No. 4873 from the railway cutting, Umfolisi, Zululand, and

No. 5114 from Umkwelane Hill.

Description.—Shell small, moderately elongate. Spire of 4 slightly

convex whorls, sutures well marked, spiral angle about 55°. Last 5

whorl larger than the spire, convex, rounded. Surface polished, i

provided with numerous, narrow, non-punctate grooves; grooves

consisting of two fine impressed lines placed close together and

sometimes rather wavy ; grooves most numerous towards the anterior

of the last whorl, where they are regularly spaced and separated by

wide, flattened interspaces; towards the posterior of the last whorl

and on the spire the grooves are more widely spaced and situate at

irregular intervals. Aperture moderately large, more or less oval,

outer lip thick. Inner lip provided with a thick deposit of callus ;

and towards the anterior end with one stout columellar fold, in front

of which there appears to be an indication of a second, probably

smaller, fold. } |

Remarks.—The two specimens here described agree in general

features with the genotype, R. ringens (Lamarck) (6, vol. 1, p. 118,

pl. iu, figs. 12-14), and with the Aachen species figured by Holzapfel,

R. hagenowi (Miller) (28, vol. xxxiv, p. 86, pl. vii, figs. 1-7). The 4

species appears to be more nearly allied to R. acuta Forbes (61, p. 424,

pl. xxvu, fig. 3), from the Trichinopoli Group, but the callus of the

S D = 7 ,

jr ie aa

a See peerie ee ee _

——

- . : '

Spel

ahs

=— S-

New Lamellibranchia and Gastropoda of Pondoland. 235

inner lip is thicker and the spiral ornamentation more strongly

developed in R. woodsi.

Famity HELICIDAE.

Genus ANcHIStomaA, Klein.

Anchistoma umzambiense, sp. nov.

(PIX XVIIT, mes, 12=16.)

Material.—One specimen, partly a cast, with the test adhering to

the spire, and the aperture incomplete, T. Mus., No. 1628.

Description.—Shell planorboid. Spire flattened, of about 5 convex

rounded whorls with strongly impressed sutures ; whorls of the spire

ornamented with numerous very regular transverse riblets and grooves,

which are convex towards the aperture. The whorls increase in

width very gradually and regularly. Last whorl convex, with its

greatest width above the middle. Umbilicus wide and deep, the bases

of the whorls very convex. Aperture more or less oval, narrowed

above and flattened below. Outer lip contracted or compressed at

about the middle, but teeth or folds corresponding to the contraction

are missing. Inner lip with only one horizontal fold, situated at about

the middle; the fold being elevated and plate-like, projecting to the

centre of the aperture, concave to the upper suture, and descending a

little towards the aperture. The flattened base of the outer lip bears

a stout, upwardly projecting fold, almost triangular in section.

Remarks.—In view of the extreme rarity of the Helicidae in Creta-

ceous deposits, it is of great interest to be able to describe a species

referable to the family among the Pondoland forms. Though the

single specimen is far from being perfect and the presumably reflexed

portions of the aperture are missing, sufficient data are obtainable to

justify a new specific name. The generic position is uncertain and

the name Anchistoma is here used in a wide sense.

The only species hitherto recorded from Cretaceous horizons in any

way related to A. wmzambiense, are the four species described by

Stoliczka (61, pp. 6-12, pl. i) from the Upper Cretaceous of Southern

India—where they occur, as does the present species, among marine

forms. Three of these were referred to Anchistoma Klein and one to

Macrocyclis Beck. The specimens were carefully re-examined by Nevill

(42, p. 128) in connection with his investigations into the modern

terrestrial Mollusca of the region, and the determinations made by

Stoliczka of the first three species received his support, with the

——

236 Annals of the South African Museum.

addition that he was of the opinion that they belonged to Helder’s

section, Gonostoma, and were not related to either Coridlla Adams or

Plectopylis Benson of the modern Indian fauna. The fourth species

he considered to belong to Patula Helder. A. umzambiense does not

appear to be very close to A. cretacewm Stoliczka, the only species

in which the apertural armature was seen, for the latter is provided

with more numerous and differently arranged folds, but the outer

lip is similarly contracted or compressed near the middle.

4,.— APPENDIX.

CRETACEOUS OF ZULULAND.

The Cretaceous fauna of Zululand was described by R. Etheridge,

jr. (17, 18), G. C. Crick (8), and R. Bullen Newton (43) in the years

1904-9, when it was shown that various Upper Cretaceous horizons

were represented there. The similarity of the matrix of different

horizons unfortunately led to the mixing of fossils from different beds,

and the locality names were used in a very wide sense. At the same

time a large proportion of the material is in a poor state of preservation

—in striking contrast to the material from Pondoland—and thus many

of the species erected by Etheridge and Newton are of very doubtful

value and uncertain horizon. The area is at present being investigated

in the field by Dr. E. C. N. van Hoepen, whose results are awaited with

interest; a preliminary account of the Manuan Creek deposits has

already been published (66).

The South African Museum has since acquired large collections of

Zululand fossils made by A. L. du Toit, W. J. Wybergh, J. 8. Hedges,

and others, the Cephalopoda of which have been the subject of an

important work by Spath (58). While the Ammonoidea have yielded

results of great importance, the Lamellibranchia and Gastropoda are

for the most part in such a poor state of preservation and fragmentary

condition, that to attempt to describe them would merely add to an

already overburdened literature. Some of the forms, however, merit

description either because they are unusually good examples of known

forms or as new species.

The following species are described below, all species which have

previously been described from the area :—

Albian. Neithea quadricostata (Sowerby).

Veniella etheridger R. B. Newton.

Macrocallista (Callistina) andersoni (R. B. Newton).

New Lamellibranchia and Gastropoda of Pondoland. 237

Senonian. Trigonoarca capensis (Griesbach).

Cardium (Acanthocardium) denticulatum Baily.

Venella forbesiana (Stoliczka).

Eriphyla lenticularis (Goldfuss).

Tn addition, the following new species are described (together with

three forms belonging to the Brachiopod genus Cyclothyris—the

phylum being new to the area) :—

Senonian. Nucula zululandensis sp. nov.

Macrocallista (Callistina) zululandensis sp. nov.

Cyprimeria? sp.

Semifusus (Mayerra) natalensis sp. nov.

The Senonan. Spath has commented upon the dissimilarity

between the Cephalopod fauna of Pondoland and that of the approxi-

mately equivalent horizons at Umkwelane Hill and the Umfolosi

valley. Thus out of 29 species of Cephalopoda from Umkwelane Hill

and neighbouring localities, only one ammonite (Pseudoschloenbachia

umbulazt) is identical with a Pondoland species, together with some

uncoiled and straight forms (Bostrychoceras *, Diplomoceras ?, Baculites)

and a Nautilus (58, p. 269). The difference becomes all the more

striking when it is seen that the Upper Senonian of the Manuan Creek

area, not far to the north, is more like that of Pondoland than Umkwe-

lane Hill (2bid., p. 302). The discrepancy between the Umkwelane

Hill—Umfolosi and Pondoland faunas, however, is not so noticeable

when the other elements of the fauna are taken into account, though

many of the commoner Lamellibranchia and Gastropoda of the

Pondoland deposits have not yet been found at Umkwelane Hill

(e.g. Trigonia shepstoner, T. elegans, Pecten (Chlamys) amapondensis,

Crassatellites africanus, Trigonocallista wmzambiensis, Cardvum gries-

bacht, Pseudomelania sutherland:). The writer has been able to

compare numerous forms from both areas; the following Pondoland

species are represented from Zululand * in the collection of the S.A.

Museum, or have already been recorded by Etheridge and Newton :—

(a) Umkwelane Hill, Umfolosi valley.

Trigonoarca capensis (Griesbach).

Nordenskjoldia natalensis (Baily).

Inoceramus expansus Baily.

Neithea quinquecostata (Sowerby).

* Spath has given a map showing the fossiliferous localities in the Umfolosi

valley, Ann. 8. Afr. Mus., vol. xii (1921), p. 222.

i} ii)

HY! {)

238 Annals of the South African Museum.

Pleuromya africana (R. Etheridge, jr.).

Inopistha (Psilomya) corrugata Woods.

Veniella forbestana (Stoliczka).

Eriphyla lenticularis (Goldfuss).

Cardium (Acanthocardium) denticulatum Baily.

Protocardia hillana (Sowerby).

Gyrodes tenellus Stoliczka.

Turritella (Zaria) boner Baily.

us (Haustator) meadi Baily.

-Confusiscala ornata (Baily).

Dicroloma (Perissoptera) barlyz (R. Etheridge, jr.).

Rostellites capensis Woods.

Ringicula woodsi sp. nov.

(b) Lake Itesa (or Eteza), Umfolosi valley.

Nordenskjoldia natalensis (Baily).

Glycumeris africana (Griesbach).

Veniella forbesiana (Stoliczka).

Cardium (Acanthocardium) denticulatum Baily.

Turritella (Zaria) bonet Baily.

Dicroloma (Perissoptera) bailyi (R. Etheridge, jr.).

Trigonoarca capensis (Griesbach).

Inoceramus expansus Baily.

Neithea quinquecostata (Sowerby).

Pleuromya africana (R. Etheridge, jr.).

Gyrodes tenellus Stoliczka.

Turritella (Zaria) boner Baily.

Dicroloma (Perissoptera) bailyi (R. Etheridge, jr.).

Rostellites capensis Woods.

Ringicula woodsi sp. nov.

(d) Umfolosi valley, east of railway. |

Inoceramus expansus Baily.

Pleuromya africana (R. Etheridge, jr.).

(e) Manuan Creek valley, south side.

Nordenskjéldia natalensis (Baily).

Neithea quinquecostata (Sowerby).

Pleuromya africana (R. Etheridge, jr.).

Vemella forbesiana (Stoliczka).

Protocardia hillana (Sowerby).

The above lists indicate a close similarity between the Lamellibranch

New Lamellibranchia and Gastropoda of Pondoland. 239

and Gastropod assemblages of the various Zululand Upper Senonian

localities, and it would seem legitimate to conclude that the absence

of certain common Pondoland forms is not entirely due to the vagaries

of collecting, but is due to a difference of facies between the two sets

of deposits—an explanation which has already been advanced by

Spath (2bed., p. 270). According to du Toit (16, p. 327), Trigonia

shepstonet and Trigonocallista wmzambiensis, neither of which occur

in Zululand, are found with other typical Pondoland forms at In-

comanini on the Komati River in the southern part of Portuguese

Kast Africa; the Transvaal Museum possesses specimens of Norden-

skjoldia natalensis, Glycimeris africana, Trigonocallista umzambiensis

and Macrocallista euglypha from the same locality. The Upper

Senonian is represented in the northern part of Zululand, for the South

African Museum possesses the following specimens :—

(f) West of the Pongola River, Ingwavuma district.

Nordenskjoldia natalensis (Baily).

Pleuromya africana (R. Etheridge, jr.).

The Albian. The Albian fauna of the Manuan Creek area has

been dealt with principally by R. B. Newton and by Spath, in the

papers quoted above. Spath (60) has recently described Upper

Albian Ammonoidea from Catuane, Portuguese Hast Africa. An

important link between the two areas is indicated by the occurrence

in northern Zululand of the following Manuan Creek species :—

(g) Pongola River (Ingwavuma road drift).

. Trigona crickt R. B. Newton.

Neithea quadricostata (Sowerby).

Veniella etheridger R. B. Newton.

DESCRIPTION OF THE SPECIES.

LAMELLIBRANCHIA.

FAMILY NUCULIDAE.

Genus Nucuta, Lamarck.

Nucula zululandensis, sp. nov.

(EX XS tres.) 0. 7.)

Material.—One specimen with the valves closed, the shell surface

partly removed.

240 Annals of the South African Museum.

Description.—Shell triangular, rather elongated, moderately convex.

Antero-dorsal margin long, gently convex; postero-dorsal margin

short, straight. Ventral margin convex, long, meeting the antero-

and postero-dorsal margins at sharp angles. Umbones small, curved

inwards, and placed near the posterior end. Lunule posterior, narrow

and elongated, ill-defined. Ornamentation where seen of extremely

fine concentric striations. |

Length, 15°5 mm.; height, 9 mm.; thickness of two valves, 6 mm.

Remarks.—The species is unlike any of those described by Stoliczka

from Southern India, and is readily distinguished from the Pondoland

form N. kaffraria described above by its elongate shape, absence of

ornament, and obscure lunule. JN. planata Deshayes (76, vol. i, pl. u,

figs. 11-15) is similar, but the anterior part is shorter and more

rounded.

Locality.— Railway cutting, Umfolosi, S.A. Mus., No. 4869,

collected by W. J. Wybergh.

Horizon.—Upper Senonian.

Famity ARCIDAKE.

Genus TRIGONOARCA, Conrad.

Trigonoarca capensis (Griesbach).

1871. Arca capensis, Griesbach: Q.J.G.S., vol. xxvii, p. 66, pl. i,

figs. 10) 10a) b

1904. Trigonoarca umzambaniensis, R. Etheridge, jr.: Sec. Rep.

Geol. Surv. Natal and Zulu-

land, p. 76, pl. 1, figs. 1-5:

1906. Trigonoarca capensis, Woods: Ann. 8. Afr. Mus., vol. iv,

p. 288, pl. xxxiv, tiesaaleeze

Remarks.—Etheridge assigned some specimens of Trigonoarca from

Umkwelane Hill to Baily’s T. wmzambaniensis (2, p. 460, pl. xin,

fig. 1), and attempted to demonstrate points of difference between

that species and 7. capensis (Griesbach). It is highly probable that

Baily’s specimen belonged to T. capensis, which is one of the common-

est species in the Umzamba beds; as the figure given by Baily is

poor and the type is missing, the writer follows Woods in disregarding

the earlier name. It follows that the Zululand forms must either be

assigned to J’. capensis or receive a new name. The writer has

examined some half-dozen examples from Umkwelane Hill and numer-

ous specimens from Pondoland, and concludes that the former belong

New Lamellibranchia and Gastropoda of Pondoland. 241

to the same species. The exterior surfaces of the Umkwelane Hill

specimens are on the whole better preserved than the examples from

Pondoland, but the ornamentation is of the same nature. The

steepness of the posterior slope varies considerably among the forms

from both areas; the Zululand specimens usually have a steep

posterior slope, while those from Pondoland are inclined to be more

rounded ; this character is not, however, a constant feature, for the

Pondoland forms sometimes have the slope as steep as in the

examples from Zululand ; the objection raised by Etheridge, therefore,

disappears. ;

Localities—Umkwelane Hill, S.A. Mus., Nos. 5099, 5404, 5452,

5471, 5483, collected by A. L. du Toit and J. 8. Hedges; Railway

cutting, Umfolosi, No. 4830, collected by W. J. Wybergh.

Horizon.—Upper Senonian.

Famity PECTINIDAE.

Genus NeitHEa, Drouet.

Neithea quadricostata (J. Sowerby).

(Pl. XXX, figs. 10-12.)

1814. Pecten quadricostata, J. Sowerby: Min. Conch., vol. i,

! | Doel wolesliva ios leer

1903. Pecten (Neithea) quadricostatus, Woods: Cret. Lam. Eng-

land, vol. i, p. 210, pl. xl,

figs. 6, 7.

1909. Neithea quadricostata, R. Bullen Newton: Trans. Roy. Soc.

Se LR OI My Oy TORE

pl. u, figs. 18-21.

Remarks.—This widespread species has been figured by R. B.

Newton from the Manuan Creek area. The present collection includes

several exceptionally well-preserved specimens, and as the figures

given by Newton are rather poor, one of these is figured here.

Localities.—Manuan Creek, 8.A. Mus., No. 2590; Pongola River,

S.A. Mus., No. 3971; Ingwavuma Road Drift, Pongola River, S.A.

Mus., No. 5529.

Horizon.—Albian.

242 Annals of the South African Museum.

Famity CARDIIDAE.

Genus Carpium, Linnaeus.

Sub-genus ACANTHOCARDIUM, Gray.

Cardium (Acanthocardium) denticulatum, Baily.

1855. Cardium denticulatim, Baily: Q.J.G.8., vol. xi, p. 460, pl.

: xi, fig. 4.

1904. Cardium bullen-newtom, R. Etheridge, jr.: Sec. Rep. Geol.

Surv. Natal and Zululand, p.

78, pl. ii, figs. 14-16. :

1906. Cardium denticulatum, Woods: Ann. 8. Afr. Mus., vol. iv,

p. 306, pl. xxxvi, fis) tees

XRKVG, fesse 2:

Remarks.—C. bullen-newtont was created by Etheridge for some

specimens from Umkwelane Hill, which were said to differ from the

Pondoland species in that the umbonal region is more tumid, with the

result that the anterior and posterior slopes are more pronounced.

The writer has seen numerous specimens from Umkwelane Hill and

Lake Itesa, as well as from Pondoland, and considers them all to

belong to Baily’s species. The tumidity of the umbonal region is

subject to a certain amount of variation in both the Zululand and

Pondoland forms, and the numerous examples seen by the writer all

agree in having sharp, denticulated costae and toothed margins.

Localities —Umkwelane Hill, 8.A. Mus., Nos. 5106, 5113; Umkwel-

ane Hill, S.A. Mus., No. 5500, collected by A. L. du Toit; Lake Itesa,

Umfolosi, S.A. Mus., No. 4893, collected by W. J. Wybergh.

Horivzon.—Upper Senonian. .

Famity CYPRINIDAE.

Genus VENIELLA, Stoliczka.

Venella etheridge, R. Bullen Newton.

(Pl. XXXT, figs. 1-3.)

1909. Veniella etheridger, R. B. Newton: Trans. Roy. Soc. 8. Afr.,

p. 69; pl. vii, digs, 556:

Material.—The collection includes several fine specimens, including

two specimens showing the hinge of the left valve. The holotype

New Lamellibranchia and Gastropoda of Pondoland. 243

described by Newton is an imperfect specimen, showing part of the

left hinge only.

Description.—Shell large, remarkably thick, strongly inflated,

trigonal. Umbones prominent, strongly curved inwards and anter-

iorly, well inflated ; posterior carina extending from the umbones to

the produced postero-ventral angle, strong, not developed into a

raised ridge by lateral compression. Posterior region very steep,

almost vertical. Antero-dorsal margin concave; ventral margin

long, broadly rounded, meeting the antero-dorsal margin in an angle ;

postero-dorsal margin long, very slightly convex ; the ventral margin

at an angle of about 70° with the postero-dorsal margin. Lunule

large, wide, deep, bounded by a high, rounded ridge but not defined

by an impressed line. Escutcheon defined towards the umbo by a

ridge, the surface sloping rapidly away from the ridge, which dies

away posteriorly.

Left valve with a long, high, stout, posterior cardinal, situated

against the flattened nymph; median cardinal stout, high, with a

feeble sulcation ; anterior cardinal consisting of a stout tubercle and

an anteriorly extending ridge; the anterior and median cardinals

separated by a relatively narrow socket; postero-lateral tooth ill-

defined, consisting of a raised and thickened portion of the postero-

dorsal margin, which fits into a deep groove on the opposite valve.

Ornamentation of numerous concentric striae and growth ridges

and occasional elevated laminae; the posterior area with numerous

oblique ridges which bend sharply at the carina; lunule with numerous

closely set ridges.

Anterior adductor impression deep, enclosed in the produced

anterior part of the shell.

Remarks.—The large and massive species is readily separated from

other members of the genus, though it resembles large specimens of

V. forbesiana, from which it differs in having a more massive hinge

area, a less elevated and more obtuse carina, and a less well-developed

left posterior lateral tooth.

Localities—Manuan Creek, S:A. Mus., Nos. 2723a-f, collected by

the resident magistrate of Ubombo; Middle branch of the Manuan

Creek, S.A. Mus., Nos. 4928, 4975, 4986, collected by W. J. Wybergh ;

Pongola River, 8.A. Mus., No. 3972; Ingwavuma Road Drift, Pongola

River, S.A. Mus., No. 5534.

Horivzon.—Albian.

VOL. XXVIII, PART 2. 17

244 Annals of the South African Museum.

Vemella forbesvana (Stoliczka).

(Pl. XXIX, figs. 1-5.) |

1871. Cyprina forbesiana, Stoliczka: Cret. Pel. 8. India, p. 197,

pl. ix, figs. 2-8.

1905. Roudawrera forbestana, Chofiat: Nouvelles données, Angola,

p. 42, pl. 1, fig. 3.

1909. Venella forbesiana, R. B. Newton: Trans. Roy. Soc. 8. Afr.,

vol. 1, pt.1, ps Oiagple

vil, figs. 1-4.

1912. Roudaireia forbesiana, Pervinquiére: Pal. Tunis., Gastr. et

Lam. .Cret.,. pam2aes

pl. xv, figs: 145i:

I) Material Numerous specimens, several small examples showing

at . the hinge.

iia Description.—Shell elongately-ovate, trapezoidal, moderately to

Pe ‘ + ~S> — = —.. — - es

Ei : “s ; 2 oe = et ~ _- — y - - + =. = _ =— - ~

— - t~ ah <2 L, _ ee. am se srs eer = _ — > — = =

SS SSS = a ae

LN} | well inflated. Umbones prominent, strongly curved inwards and

i i 4 forwards. Posterior carina strongly developed, sharp, running from

i Ni | the umbones to the produced postero-ventral extremity ; posterior

ail | slope moderately steep. Posterior margin well rounded; postero-

| | ventral margin rounded and slightly pointed ; antero-dorsal margin

' TU | short, concave; anterior produced and sharply rounded; ventral

Ld | margin gently rounded, often sinuous in front of the postero-ventral

ee | angle. Escutcheon bounded by an obscure keel running from the

| Lib | umbones posteriorly, about midway between the carina and the

i) postero-dorsal margin, dying away ventrally. Depressed area beneath

the umbones, but no defined lunule. Adductor impressions well

marked, the anterior excavated.

ty || Hinge of right valve: nymph narrow, with deep grooves above

Al \ | and below; posterior cardinal long, oblique, laterally compressed,

ae By stout, with a feeble groove along its summit; median cardinal reduced,

{| plate-like, situated obliquely above the anterior and separated from

i] it by a deep groove ; anterior cardinal stouter ; in front of the anterior

ii cardinal a small tubercle is developed which fits into a small pit inferior

if to the left anterior cardinal; posterior lateral elongated, with a deep,

i long pit to receive the left lateral, with fine transverse striations.

lj | Hinge of the left valve: posterior cardinal long, high, curved,

ly | ; oblique, close to the nymph, which is flattened and separated from

hy | it by a deep groove ; median cardinal trigonal, stout, elevated, with

il | a feeble median sulcation; anterior cardinal stout, trigonal, elongated

oF SS

2 2 FRE

————=_

——— z

a —~

mime:

————=3

—

= BE wees:

1 = = SSS

—

~ as

== =

New Lamellibranchia and Gastropoda of Pondoland. 245

anteriorly, the top connected with the base of the median cardinal

by a thin lamina; posterior lateral stout, high, transversely striated.

Sculpture on the anterior of equidistant concentric elevated laminae

separated by broad, finely striated, shallow sulcations, sometimes more

irregular ; posterior to the carina there are striations only.

Remarks.—The species is widespread, as it is known not only from

the Trichinopoli Group but also from the Mediterranean region and

from Angola. A small example is figured in this paper among the

Pondoland forms. It is easily distinguished from R. druc by its less

markedly triangular shape, more round outline, and less steep posterior

slope. The curious lamina connecting the left median and anterior

cardinals has not been seen in the other species.

Localities.—Recorded from the Manuan Creek area by R. B.

Newton ; Umkwelane Hill, 8.A. Mus., Nos. 5103, 5105, collected by

J. 8. Hedges; Umkwelane Hill, S.A. Mus., No. 5471, collected by

A. L.du Toit ; Lake Itesa, Umfolosi, S.A. Mus., Nos. 4887, 4896, 4899,

collected by W. J. Wybergh.

Horizon.—Upper Senonian.

Famity ASTARTIDAE.

Genus EripHyLa, Gabb.

Eriphyla lenticularis (Goldfuss).

(Pl. XXIX, figs. 8-10.)

-1840. Lucina lenticularis, Goldfuss: Petref. Germ., vol. 11, p. 228,

plaexixa, me. 1G:

1870. Eriphyla lenticularis, Stoliczka: Cret. Lam. S. India, p. 181,

pl. vi, figs. 7-13.

1904. Hriphyla lenticularis, R. Htheridge, jr.: Sec. Rep. Geol. Surv.

Natal and Zulu-

leincl, jos (Os jal ae

figs. 20, 21.

1906. Astarte (Eriphyla) lenticularis, Woods: Ann. S. Afr. Mus.,

vol. iv, p. 301, pl.

xxxv, fig. 20 (with

full syn.).

Remarks.—The collection contains two small specimens from Lake

Itesa, which closely resembie some larger forms from Pondoland. As

they are very well preserved and show the hinge, figures of one of them

246 Annals of the South African Museum.

are given here. Woods figured only the exterior of a right valve in

his Pondoland paper, while the figures given by Etheridge of examples

AO | from Umkwelane Hill are rather poor.

lan ill i i Locality.—Lake Itesa, Umfolosi, 8.A. Mus., No. 4884, collected by

# | W. J. Wybergh.

| Horizon.—Upper Senonian.

Famity VENERIDAE.

Genus Macrocatuista, Meek.

Sub-genus CALLISTINA, Jukes-Browne.

Macrocallista (Callistena) anderson (R. Bullen Newton). |

(Pl. XXX, figs. 13-18.)

1909. Meretrix anderson, R. B. Newton: ‘Trans. Roy. Soc. S. Air.,

vol. 1, pt) 1, paMioamele

vi, figs. 7-9.

Material. Numerous specimens, showing hinges of both valves.

al || Description.—Shell oval, moderately convex, considerably inequi-

lateral. Antero-dorsal margin slightly concave, short; anterior

margin rounded and passing gradually into the ovally curved exten-

sive ventral margin; postero-dorsal margin long, slightly convex,

oblique, produced ; postero-ventral end narrow, subangulate. Um-

bones contiguous, small, moderately curved anteriorly. Lunule

moderately depressed, cordiform, bounded by a groove. Escutcheon

t} not defined. Pallial sinus deep, slightly ascending, rounded at —

| | | | the end.

Hinge in the right valve with three cardinals; the anterior and

median laterally compressed, plate-like, diverging from beneath the

umbo; posterior cardinal oblique, long, divided, the posterior part

longer than the anterior part. In front of the cardinals in the right

valve there is a moderately deep, elongate pit with very slightly

raised upper and lower margins, the lower margin being rather more

raised than the upper. Left valve with stout anterior and median

cardinals diverging from beneath the umbo, the anterior nearly

vertical ; posterior cardinal long, slender, oblique; anterior lateral

elongate, ridge-like.

Ornamentation of numerous close, equidistant, concentric ridges

separated by deep grooves ; continued on to the lunule.

Remarks.—R. B. Newton, who did not see the hinge of this species,

hy |

Pall

il |

New Lamellibranchia and Gastropoda of Pondoland. 247

described it as a Meretrix, and stated that it appears to be intermediary

between J. de C. Sowerby’s Venus faba and V. ovalis, both from the

Upper Greensand of Blackdown. The former of these, however, is

now referred to Cyprimeria, and the latter to Clementia (76, vol. 1,

pp- 187 and 191) on account of the disposition of the teeth ; while the

teeth of M. andersoni indicate clearly that it belongs to the same

group as Sowerby’s Venus planus (bid., vol. 11, p. 192, pl. xxx, figs. 1-6),

made the genotype of Callistina by Jukes-Browne (29, p. 156).

There is a resemblance to Stoliczka’s Cytherea (Callista) fabulina

(62, p. 174, pl. xvi, figs. 31-33) from the Trichinopoli and Utatur

Groups, especially in its strong and regular ornamentation, but the

Indian shell is shorter and more oval in outline than the present

species. |

Newton notes a strong resemblance to Venus pacifica Moricke (89,

p- 103, pl. vii, fig. 7), which has been made a synonym of Venus

auca d’Orbigny and placed under Cytherea by Wilckens (71, p. 243,

pl. xix, figs. 12-14). The right posterior cardinal in the excellent

figures given by Wilckens appears to be entire and there is no mention

of its being bifid in the text, hence that species cannot be included

under Callistina. The Quiriquina species is larger than M. andersona

and possesses a much deeper and more rounded pallial sinus.

Pervinquiére (47, p. 268, pl. xx, fig. 17) has recently assigned to this

species three specimens from the Senonian of Tunis, but his identifica-

tion rests only on external characters and must be regarded as doubt-

ful. The Tunisian form seems to be a little less extended posteriorly

and anteriorly and has a rather less prominent umbo.

M. (Callistina) euglypha (Woods), described earlier in this paper

among the Pondoland forms, is very closely allied, but is shorter,

more compressed, and with a much smaller lunule.

- Localities.—Tributaries of the Manuan Creek (R. B. Newton) ;

south side of the Manuan Creek Valley, 8.A. Mus., Nos. 4926, 4936,

4938, collected by W. J. Wybergh.

Horizon.—Du Toit (16, p. 318) places the species among the Albian

forms, but it may well be of later age.

Macrocallista (Callistina) zululandensis, sp. nov.

(Pl. XXIX, figs. 11-16.)

Material.—Three left valves, two of which show the hinge. The

holotype is fairly well preserved but the posterior is missing; S.A.

Mus., No. 4889.

ee rr

—,

—

» Ja

ee oe >

~ St PE aE : ae 7

SSeS SS. ae ee —

SSS eee OS eS Se

_—- — _ — = -

= > 2

eS a 24 Fa a eS

——. — Z

= — ee eee

248 Annals of the South African Museum.

Description.—Shell oval, inequilateral, slightly convex. Antero-

dorsal margin concave; anterior margin rounded, passing gradually

into the uniformly convex ventral margin. Umbones moderately

small, pointed, with considerable anterior curvature. Lunule elongate-

ovate, slightly concave but raised towards the centre, limited by a

distinct groove. Escutcheon not defined.

Hinge of the left valve consisting of three cardinal teeth; the

anterior and median cardinals diverging from beneath the umbo, the

anterior narrow and nearly vertical, the median thicker, very slightly

curved and oblique, both tapering dorsally ; posterior cardinal elon-

gate, very thin, tending to be obsolete, situated on the flank of the

moderately stout nymph. Antero-lateral strong, very long, parallel

to the lunular margin, thicker and higher at its posterior end.

Surface ornamented with numerous concentric ridges separated by

furrows of rather less width than the ridges, not continued on to the

lunule.

Remarks.—This species differs from M. plana (Sowerby) (76,

vol. ii, p. 192, pl. xxx, figs. 1-6), the type of Callistina, in the length

of the anterior lateral tooth and coarse ornamentation, features

which give it a strong resemblance to Callistina wilckenst Woods

(78, p. 31, pl. xv, fig. 8; pl. xvi, figs. 10, 11; pl. xvu, figs) -a)yeand

C. thomsont Woods (2bid., p. 32, pl. xvi, figs. 4-7; 36, p. 595, figs.

55,56), which have been elevated to a new sub-genus T7kia by Marwick

(36, p. 595), with the latter as type. M. zululandensis has an even

longer antero-lateral tooth than C. wilckensi ; it is less inflated, the

umbo is more pointed, and the lunule appears to be more elongate.

M. sculpturata (Stoliczka) (62, p. 173, pl. vu, 7-9) appears to be

relatively longer and with more widely separated ribs than either of

these.

The lunule of Etheridge’s Cytherea (?) kaffraria (17, p. 81, pl. u,

figs. 20-22), also from Umkwelane Hill, is stated to be inconspicuous

and hardly defined from the anterior ends.

Localities —Lake Itesa, Umfolosi, S.A. Mus., No. 4889, collected

by W. J. Wybergh; Umkwelane Hill, 8.A. Mus., Nos. 5107, 5495,

collected by J. 8. Hedges and A. L. du Toit.

Horizon.—Upper Senonian.

New Lamellibranchia and Gastropoda of Pondoland. 249

Genus CypRIMERIA, Conrad.

Cyprimeria ? sp.

(Pl. X XTX, figs. 17-19.)

Remarks.—S.A. Mus., No. 4885, collected by W. J. Wybergh at

Lake Itesa, Umfolosi, appears to be closely allied to Cyprimeria ?

natalensis sp. nov. described above from Pondoland, but the absence

of a view of the hinge prevents a definite generic determination. It

differs from the Pondoland species in being relatively higher, the

umbones being more pointed and with greater anterior curvature.

In addition, the surface is covered with distinct concentric striations.

The lunule is slightly larger in the Zululand form, being moder-

ately large, elongately cordate, raised in the middle, limited by an

impressed line.

Horizon.—Upper Senonian.

GASTROPODA.

Famity TURBINELLIDAE.

Genus SEMIFUSUS, Swainson.

Sub-genus MAYERIA, Bellardi.

Semifusus (Mayeria) natalensis, sp. nov.

(RE PXOXOe figs) 125)

Material.—Holotype, S.A. Mus., No. 5407, surface very well pre-

served, complete except for the aperture.

Description.—Shell fusiform, moderately elongated. Spire com-

posed of about 6 sharply carinated whorls. Spire longer than the

last whorl. Carina situated near the centre of the whorls, but a

little to the front, dividing the whorls into an upper and a lower

flattened area. Sutures well marked. Last whorl large, rounded,

anteriorly produced into a long canal.

Surface ornamented with numerous fine spiral ribs and with trans-

verse lines. Carina of the whorls of the spire with tubercles at regular

intervals.

Remarks.—This species is closely allied to the two Pondoland

species described above, S. umzambiensis and S. kaffraria, but differs

from them in having the carina placed near the centre of the whorls

and in the ornamentation of numerous fine spiral ribs which occur

on all parts of the whorls. In the two Pondoland species the spiral

250 Annals of the South African Museum.

ribs are fewer in number, stronger, and absent or considerably reduced

on the posterior side of each whorl.

Locality.—Umkwelane Hill, collected by A. L. du Toit.

Horizon.—Upper Senonian.

BRACHIOPODA.

Famity RHYNCHONELLIDAE.

Genus CycLotuyris, M‘Coy.

Cyclothyris spp.

(Pl. XXX, figs. 3-9.)

Remarks.—One of the noteworthy features of the fauna of the

Upper Cretaceous deposits of Southern Africa is the great scarcity of

the Brachiopoda; the only mention that has so far been made of

the occurrence of the phylum is the remark by van Hoepen that he

had obtained a Terebratulid from the Manuan Creek area (66, p. 220).

This is somewhat surprising, in view of the fact that Stoliczka de-

scribed as many as 21 species, belonging to 4 families, from the equi-

valent horizons of Southern India. It is of great interest, therefore,

to be able to record three Rhynchonellids in the Zululand deposits,

belonging apparently to three species, though it is to be regretted

that their state of preservation does not permit of close comparison

with other forms.

Species (a).—S.A. Mus., No. 4878, collected by W. J. Wybergh from

Upper Senonian beds in a railway cutting near the Umfolosi, is a

moderately inflated sub-triangular shell; the lower valve is slightly

more inflated than the upper and has a small, sharply pointed, slightly

curved apex, small foramen and small deltidial plates; the median

sinus is very slight and each valve is ornamented with about 22 sub-

equal ribs. The apical angle is 107°. Figures 5, 6.

Species (b).—8.A. Mus., No. 4879, collected by W. J. Wybergh at

the same spot as the above, is very similar, but the median sinus is

more evident and the apical angle is only about 85°. The valves are

ornamented with about 22 ribs, 8 of which are situated opposite

the sinus. Figures 7-9.

Species (c).—S.A. Mus., No. 4942, collected by W. J. Wybergh on

the south side of the Manuan Creek Valley, presumably from Senonian

deposits, is similar to the last, but has only 16 ribs. The apical region

is missing. Figures 3, 4.

251

New Lamellibranchia and Gastropoda of Pondoland.

x SI 2X mS |) OX

x x

x x x

ES NN oS

Ss || os x x

SN OK x

x x x x xX

x | 205

x XS || 2X

x x Ze Nees ees TS | 2S el) eS

x x XS |) 2X x PMNS Pa

x x x || oS ZS | eS

x K

x x

x x ux Wes |i 2s x PAW Neal |e: Steels

x x x x VW oS | eX x x as | 2x

* PSM eS Il 2S I oS x

NAG ESI MOAI | NAIC INL PAE SNA be te CAG IG CHE AIL OEE SE clea ay

X XXX XK

PON

“nm

Alreg, wmnynynoyuap (‘ysunoy) unipany

oruueyy wwojzybnoy Ly

" SpooM snunoisfn sajzrajpssn49

(SSNJP[O)) Seumpnoyus, ophydrpy

oIuuOY sisuapuodnun ,ajunjis y

(Seu[ey)-Tolun) wmup vpjavua A

SspooM vppbns100 (vhiwmopis) nyisidovy

: HUY Ssiswaiquozun DiwopnjioYy gy

: : . (soqioj7) vo1dhy njovpo pr

(Aqramog) nynjpsovanbumb oe N

Spoo0 MA sisuadno ut

h DEGecr 1 susuapuodnurn (shy) wajoag

: Ayreg sunbaya 4

yorqsorry vauojsdays vwuobr4 7,

: : * ssnjploy vywssnoop vihbox sy

a eat (euro) sisuajooip "Jo piUoh.joa) py

: : oruuery "ds vas”

; : * Ajveg snsundaa snwupsa00uy

: : SSMJP[OH Yyossndap "Jo uur

"(ey ZO1[049) Dyoundgns “yo es

: * (qouqsomy)) vunorisfo sisal)

: * (Apveg.) sasuajpypu vipjolyswap.o Ay

* (YoRqsoLLg) sisuadvs no1mouobrs, J,

‘ : * SpooM DUDIPIIAW DYDQIWOG

“VIHON VUE TTA

"U0IJOOg puodveG “UOTPOOS PIL

"SIOAOK) “M ‘, “Iq 4q opeu uUOTZd9][09 oY UO poseq ‘spog vquIeZUTy OY} UTYYIM sotoods oY} Jo UOTZNQII4SIp oy} SUIMOYS

‘V WIdvy,

Annals of the South African Museum.

252

x Pen lien

Sil fae

x x x x

x x SNM es

x x i Nee | oS ees

x x | X

x x Se cS eX

Xx USNR Nes

x x Ze oS | eS

x x x Xx

x x x

Xx x | X

x x x x

NANG I Val MMA VAST VIE) AS | VAAN Uy

"U0T}OG Puodeg

Me x

UN SN AOS x

2X || 2S | Xk x

KK x

x x x

x x

YS M28

x x

x ll OX x

"UOTYOOY 4SALT

“panurjuoo— "VY ATEHVY,

KKK Oa

xX xX

: (‘al ‘ospt

aOU TE | YT) why (wrapdossrsag’) pUuwojosouUT

‘ds (sapyssod py =) sndouaeyp

é ; * oruuoyy zuadaoyuna Dyooipy

: Alege, wpnow (40jnjsnv 77) oe

: : A[reg, wauog (ning) 0p)0;144N J,

: : * BY ZOIOI snjjawaz sapouhp

HEEL DIDLISYINUW (DYDUVT) 10140 AT

E 5 eluuoy sisuadno sash

(A[veg) wpunprayyns nvwojawmopnas qT

: ; SPOOM SIsUaIquiDZUN )L.10 AT

(soqaoyq) AT AO 1124 Hmj08)) SapMOB.LD

* guy]

pupor4fo (nywopoywby ) pyUopo pry)

yt (qq'exy) ahjing wmnrumjosimagy

; ' gtuuey vind ffoy 07)01%

“VGOdOULSV‘)

* otuuay wswabos nsoydoquhy

ocruinne aie 00%) (40zP) snzunaajoy

- "“(spo0 4)

Dudiina (nUYysyvO) 1811000100 PT

: oluueyy wyyods s

(Spoo AA) sisuarqunzun njpsypnoouobr4, 7,

p oluuory isuaaab sisdovuasoce

SpooMm wyonqsarwb (‘yyupopy) wnipsng

“panuyjuoo—VIHONVUAITIA WW V'T

253

x X

Poe x

New Lamellibranchia and Gastropoda of Pondoland.

A[leg iwsaquof saysouua yy

SspooM Sninucqun sninpisspy

xX X

“VECIONTHO}W

* (soq10,7 ) COT sp1av0wmojdige

x . . . . ‘ygedg

(SIN YoU) wnunoiifp *49 spsaoo0osvuaydgy

" -yyedg (“SIA

YOU) — 4yongsarsb a

"yyvedg

twavudnofopnasd “

° (Ajreg) wnjnqun vriyonquaojyosopnas J

Kale : teers) VWUOINOS SDLAIVUO}LO JA

¢ - ds sayvydnoso,do py

(uadooyy uea) snunoify "ye sapnspupo jy

: * (Apleg) wuapinb spsaoisann eT

(uodooy] uvA) xaydunis "jo wrsanpyay

: uodooy urea saysuadns sajvuobv.a,],

* uoda0xyT uea wnwjsooipa spsaohapnopy

‘VadOdOTVHdH)

KAK OK x

x x

x SX

x x

x xX

xXx

: * spoom vydunsed nyohidrsny

x : : * eluuey wspoom nnoiburay

: : ; * otuuey ‘ds nyjauyjng

x : * SpooM SIVUOIpisaUL DILDI20UDO

: : (yoeqsoty) vunif[py 4 Sst fT

* Spoo(A sisuadno sazrjjajsoay

: eluuey wsuavab SS

x : : * spoom vuvoifo sisdoshg

x : * otuuey niwouf[oy nanydasdoang

* oruuoy “ds 4 es

x : ‘ ouuoy vprbisopnasd suhysojdhip

x " ommludey wuaaab (vuynzev jy) snahyoT

: erUuleyy sysuoequmeun d Dypuoydes

b6F8 “W'V'S “ds

x 6698 "W'V'S “ds (nsajdossisag ) nuojo1017

XXX

X X

xX X

254 Annals of the South African Museum.

TABLE B.

Showing the distribution of the Pondoland Upper Cretaceous Lamellibranchia and

Gastropoda. Identical species are indicated by x. The occurrence of an allied

species is shown by the letter a.

Southern India. ca oars Southern Hemisphere.

Umzamba Beds, Pondoland.

Utatur Group

Trichinopoli Group.

Ariyalur Group.

Valudayur Beds.

Trigonoarca Beds.

Upper Senonian.

Lower Senonian.

Pre-Senonian.

Zululand.

Central Mogambique.

Madagascar.

Angola.

South Patagonia.

Antarctica

LAMELLIBRANCHIA.

Nucula kaffraria Rennie

Nuculana sp. Woods

Arca sp. Woods 3

Barbatia meridiana W oods

as sp. Woods .

Trigonoarca capensis (Griesbach) ; a x a {

elongata Rennie :

Nordenskjoldia natalensis (Baily) s nee Nee x x a

Glycimeris africana (Griesbach) . : x

Bh ef. subplanata (Stoliczka) . x

i Pinna vanhoepeni Rennie .

| Pinna cf. decussata Goldfuss

| Inoceramus expansus Baily : ; a x a

th Ostrea spp. :

) Alectryonia cf. aaenieneis (Stoliczka) ; x a

Hi} | Exogyra decussata Goldfuss A : a

s sp. Woods . ; : : !

Trigonia shepstonet Griesbach :

» elegans Baily :

Pecten (Chamys\ amapondensis ‘Gries-

| bach :

. a capensis Woods

F i o (Aequipecten) kossmatt Woods .

ae aE Tae on

BR TSS

ES EE SSE

a FE ERs — ee

| » (Camptonectes) kaffraria Rennie a

Neithea quinquecostata (Sowerby) Seal ea | OK | KM a ea

Spondylus cf. calcaratus Forbes . x ‘|

Tima (Acesta) obliquistriata Forbes. Xx |

Mytilus sp. Woods : : : ;

Modiola kaffraria Woods . :

| “s typica (Forbes) . Xx x :

#) || Pleuromya africana (R. Etheridge, are ye als x ?a i

Pholadoma ya umzambiensis Rennie , | ;

;

MS ef. elliptica Minster

Goniomya umzambiensis Rennie

Cercomya arcuata (Forbes) . Xx x os

! i | | Liopistha (Psilomya) corrugata W oods. ‘

i} | | Veniella drut (Munier-Chalmas) . : 2a fry

forbesiana (Stoliczka) . : x a x xX | a3) Sex |

New Lamellibranchia and Gastropoda of Pondoland.

TaBLE B.—continued.

Southern India.

ere lee |e

Umzamba Beds, Pondoland. Se We ea

Sissi s

S/2/5|2

ae ee fe

LAMELLIBRANCHIA—continued.

Astarte griesbachti Woods . ‘ é ?a

» 2? amapondensis Rennie A

Eriphyla lenticularis (Goldfuss) . : x

Crassatellites africanus Woods

ss haughtont Rennie . : a

Anisodonta ?? umzambiensis Rennie

Cardium (Acanth.) denticulatum Baily. a

ae griesbacht Woods

Protocardia hillana (Sowerby), var. . a

Dosiniopsis geversi Rennie

Cyprimeria ? natalensis Rennie . ‘

Bs (Cyclorisma ?) cf. analoga

(Forbes) . x

Trigonocallisia umzambiensis (Woods)

spathie Rennie .

Macrocallista (Callistiia) euglypha

(Woods)

Palaeomoera umzambiensis Rennie : Pane:

ie haughtont Rennie

ny sp. Rennie

Macoma papyracea Rennie

Solecurtus (Azor) woodsi Rennie

Cymbophora rogerst Rennie

-Mactra ? kaffravia Rennie .

Martesia ? sp. Rennie

Teredo sp. Woods

GASTROPODA.

Patella kaffraria Rennie

Semisolarium bailyi (Gabb) ‘ ;

Chilodonta (Agathodonta) africana

Rennie

Margarites

(Forbes)

Nerita wmzambiensis Woods

» kaffraria Woods ;

Trajanella dutoiti Rennie . : a\a

Pseudomelamia sutherlandi (Baily)

Lysis capensis Rennie : a al

Natica (Lunatia) multistriata Baily

Gyrodes tenellus Stoliczka . d : Kaox

(Solariella) radiatula

COR Wy OX loos

Trigonoarca Beds.

Europe and

255

Southern Hemisphere.

Antarctica.

N. Africa.

2 (ETe S s

om | om | a | S& S

ae) se alae ease

Satelite ecelercsed| eh Se lites

a || ei So) Scala tenia

= 1 © | So) oy | Ss

SS ge | ee eer es

| ea CAE A esas lear io la

SS OE TST ce reg an face

a S js =

Blo} A <= 5

=a a oe

Sale|

x | Ss |) 2k | dK

x x

SNAG AANA G

aur J

256 Annals of the South African Museum.

TaBLE B—continued.

Southern India. | ge pene Southern Hemisphere. |

Umzamba Beds, Pondoland.

Utatur Group

Trichinopoli Group.

Ariyalur Group.

Valudayur Beds.

Trigonoarca Beds.

Upper Senonian.

Lower Senonian.

Pre-Senonian.

Zululand.

Central Mocambique.

Madagascar.

Angola.

South Patagonia.

Antarctica.

GASTROPODA—continued.

Turritella (Zaria) bonei Baily . 5 x |x

(Haustator) meadi Baily

Arcotia vanhoepent Rennie : : a

Confusiscala ornata (Baily) - - a

Chenopus [ Aporrhais] sp. Woods

Dicroloma (Perissoptera) bailyi (R.

Etheridge, jr.) . ; x

Dicroloma (Perissoptera) spp. Rennie ‘

Pugnellus auriculatus Woods

us contortus (Forbes) : : x a

ss sp. Woods

Cypraea chubbi Rennie :

Pirula (Protopirula) capensis Rennie :

Tritonidea (Cantharulus) kaffraria

x X

Rennie ‘ ; a

Siphonalia ? umzambiensis Rennie , a

Lathyrus (Mazzalina) geversi Rennie . a

Cryptorhytis rigida (Baily). ; - a

sf pseudorigida Rennie

As sp. Woods :

- ? sp. Rennie

Paleopsephaea scalaris Rennie

kaffrarvia Rennie .

Pyropsis africana Woods .

ze geverst Rennie

Pirifusus bailyi Woods - :

Semifusus (Mayeria) acuticarinatus

Rennie Z é : : : a

Semifusus (Mayeria) wmzambiensis

Rennie :

Semifusus (Mayeria) kaffrarius Rennie.

Volutilithes sp. Woods :

Rostellites capensis Woods

Cancellaria meridionalis Woods .

sp. Woods : ‘ g

Turris 2 kaffraria (Griesbach)

Actaeon sp. Woods : F t

Actaeonella (Trochactaeon) woodsi

Rennie P 5 : ¢ a

Bullinella sp. Rennie : : : a a

Ringicula woodsi Rennie . - : a

Eriptycha perampla Woods

Anchistoma umzambiense Rennie

New Lamellibranchia and Gastropoda of Pondoland. 257

5. REFERENCES.

Useful lists of references are given by Woods (76), Pervinquiére (47), and

Marshall (35).

(1) AnpERSon, F. M. (1902).—“ Cret. Deposits of the Pacific Coast,’’ Proc. Calif.

Acad. Science, ser. 3, vol. ii, No. 1.

(2) Batty, W. H. (1855).—“ Descriptions of some Cretaceous Fossils from South

Africa,” Q.J.G.8., vol. xi, pp. 454465.

(3) BouLz et THEVENIN (1906).—“ Fossiles de la Céte orientale de Madagascar,”

Ann. de Pal., vol. i, pp. 43-59.

(3a) CHapman, F. (1904).—* Foraminifera and Ostracoda from the Cretaceous

of Kast Pondoland, etc.,” Ann. 8S. Afr. Mus., vol. iv, pp. 221-237, pl. xxix.

(4) CHorrat, P. (1905).—‘‘ Nouvelles Données sur la Zone Littorale d’Angola,”’

Contr. 4 la Conn. Géol. des Colonies Port. d’Afrique, No. ii,”’ Comm. du

Serv. Geol. du Portugal, pp. 1-48.

(5) Coguanp, H. (1869).—Mon. du Genre Ostrea: Terrain Crétacé.

(6) Cossmann, M. (1895-1921).—Essais de Paléoconchologie Comparée.

(6a) CottREAv, J. (1922).—* Fossiles crétacés de la Céte orientale (de Mada-

gascar),” Ann. de Pal., t. xi, pp. 111-192, pls. ix—xix.

(7) Cox, L. R. (1925).—“ Cretaceous Gastropoda from Portuguese East Africa,”

Ann. Transv. Mus., vol. xi, pp. 201-216.

(8) Crick, G. C. (1907).—* Cretaceous Fossils of Natal, Part 3: The Cephalo-

poda,” Third Rep. Geol. Surv. Natal and Zululand, pp. 161-250.

(9) Crick, G. C. (1907).—‘“* The Cretaceous Rocks of Natal and Zululand and their

Cephalopod Fauna,” Geol. Mag., vol. iv, pp. 339-447.

(10) Dacgu&, E. (1903).—‘‘ Mittheilungen iiber den Kreidecomplex von Abu

Roash bei Kairo,” Palaeontographica, vol. xxx, 2, pp. 227-292.

(1r) Daun, W. H. (1890-1900).—“ Tertiary Fauna of Florida,” Trans. Wagner

Free Inst. Sci., vol. iii.

(12) D’OrRBIaNny, A. (1842-3).—Paléontologie Frangaise, Terrains Crétacés, vol. ii,

‘* Gasteropodes.”’

(13) D’ Orzsiany, A. (1843-7).—Ibid., vol. iii, ‘‘ Lamellibranches.”’

(14) pv Tort, A. L. (1913).—* The Geology of Pondoland,” Ann. Rep. Geol. Surv.

U.S. Afr. for 1912, pp. 153-180.

(15) pu Tort, A. L. (1920).—“ The Geology of Pondoland and Portions of Alfred

and Umzimkulu Counties, Natal: An Expl. of Sheet 28 (Pondoland),”’

Geol. Surv. U.S. Afr., 45 pp.

(16) pu Tort, A. L. (1926).—The Geology of South Africa.

(17) ErueripGe, R., jr. (1904).—‘* Cretaceous Fossils of Natal, Part 1: The

Umkwelane Hill Deposit,” Second Rep. Geol. Surv. Natal and Zululand,

pp. 79-93.

(18) ErHeripcs, R., jr. (1907).—“* Cretaceous Fossils of Natal, Part 2: The

Umsinene River Deposit,’ Third Rep. Geol. Surv. Natal and Zululand,

pp. 67-90.

(19) Fiscumr, P. (1887).—Manuel de Conchyliologie.

(20) Forszs, E. (1846).—“‘ Report on the Fossil Invertebrata from 8. India, etc.,”

Trans. Geol. Soc. Lond., vol. vii, pp. 97-174.

258 Annals of the South African Museum.

(21) Fortav, R. (1917).—‘‘ Cat. des Invert. Fossiles de Egypte, Terr. Crét., 2me

partie : Lamellibranchia,” Geol. Surv. Egypt, Pal. Series, No. 3.

(22) Frecu, F. (1887).—“ Die Versteinerungen der Untersenon Thonlager zwischen

Suderode und Quedlinburg,” Zeitsch. d. D. Geol. Ges., Jahrg. 1887, pp.

141-202.

(23) Gass, W. M. (1864).—“‘Cretaceous Fossils,’ Geol. Surv. Calif.; Palaeontology,

vol. i, sect. iv, pp. 57-217.

(23a) GARDEN, R. J. (1855).—** Notice of some Cretaceous Rocks near Natal,

South Africa,” Q.J.G.S. London, vol. xi, pp. 453-454.

(24) Goupruss, A. (1827-44).—Petrefacta Germaniae.

(25) GRIESBACH (1871).—** On the Geology of Natal, in South Africa,” Q.J.G.S.,

vol. xxvii, pp. 53-72.

(26) Haveuton, S. H. (1925).—‘‘ Notes on some Cretaceous Fossils from Angola

(Cephalopoda and Kchinoidea),”’ Ann. 8. Afr. Mus., vol. xxii, pp. ies

(27) Histneer, W. (1837).—Lathaea Suecica.

(28) HouzapreE., E. (1887—89).—‘* Die Mollusken der Aachener Kreide,”’ Palaeonto-

graphica, vols. xxxiv, XXXv.

(29) JUKES-Browng, A. J. (1908).—‘* On the Genera of Veneridae represented in

the Cretaceous and Older Tertiary Deposits,’ Proc. Malac. Soe., vol. viii,

pp. 148-177.

(30) Kinran, W. (1902).—‘‘ Uber Aptien in Siidafrika,’ Centrall. f. Min., etc., 3,

p. 465.

(31) Kossmat, F. (1895).—‘‘ The Cretaceous Rocks of S. India,’’ Records Geol.

Surv. India, vol. xxviii, pp. 39-55.

(32) Kossmat, F. (1895-6).—‘* Untersuchungen iiber die siidindische Kreide-

formation,” Beitr. z. Pal. u. Geol. Oesterr.-Ung., Bde. 9, 11.

(33) Kossmat, F. (1897).—‘‘ The Cretaceous Deposits of Pondicherri,’ Records

Geol. Surv. India, vol. xxx, pp. 51-110.

(34) KrenkEL, E. (1910).—‘‘ Die Aptfossilien der Delagoa-Bai (Siidostafrica),”’

N. Jahrb. f. Min., 1, pp. 142-168.

(35) MarsHatt, P. (1926).—‘** The Upper Cretaceous Ammonites of New Zealand,”

Trans. N.Z. Inst., vol. lvi, pp. 129-210.

(36) Marwick, J. (1927).—“‘ The Veneridae of New Zealand,” Trans. N.Z. Inst.,

vol. lvii, pp. 567-635.

(37) Muex, F. B. (1876).—“‘ A Report on the Invert. Cret. and Tert. Fossils of the

Upper Missouri Country,” U.S.G.8. Terr. Rep., vol. ix.

(38) Morscu, C. (1874).—‘‘Mon. der Pholadomyen,” Abhandl. Schw. Pal.

Gesellsch., vols. i, ii.

(39) Méricke (1895-6).—(In Sternmann, G.: “ Beitr. z. Geol. u. Pal. v. Siid-

amerika,’ Neues Jahrb. f. Min., beil-bd. x, pp. 1-118.)

(40) Muxuer, J. (1847-59).—Mon. der Petrefacten der Aachener Kreideformation.

(41) Muniter-CHatmas (1881).—(In Rovpatre: Mission dans les Chotts Tunisiens. )

(42) Nevitt, G. (1881).—‘‘ New . . . Mollusca of the Indo-Malayan Fauna,”

Journ. Asiatic Soc. Bengal, vol. i, pt. ii, No. 3.

(43) Newton, R. BuLLEeN (1909).—“* Cretaceous Gastropoda and Pelecypoda from

Zululand,” Trans. Roy. Soc. 8. Afr., vol. i, pt. i, pp. 1-106.

(44) Newton, R. BuLLEN (1924).—‘‘ Appendix to Paper by E. O. Teale on the

Geology of Port. E. Africa between the Zambesi and Sabi Rivers,” Trans.

Geol. Soc. 8. Africa, vol. xxvi, pp. 141-159.

New Lamellibranchia and Gastropoda of Pondoland. 259

(45) Niusson, 8. (1827).—Petrifacta Suecana Formationis Cretaceae.

(46) Nortiine, F. (1902).—“‘ Fauna of Baluchistan, etc.,” Pal. Indica, ser. 16,

WO, pe. U1.

(47) PeRvinqutére, L. (1912).—Etudes de Paléontologie Tunisienne : Gastropodes

et Lamellibranches des Terrains Crétacés.

(48) Proret eT Roux (1847-53).—Description des Mollusques Fossils qui se

trouvent dans les Grés Verts des Environs de Genéve.

(49) Piows, W. J. (1921).—** The Cretaceous Rocks of Pondoland,” Ann. Durban

Mus., vol. iii, pp. 58-66. |

(50) Ravn, J. P. J. (1902).—Molluskerne i Danmarks Kridtaflejringer ; Lamelli-

branchiater.

(51) Renntz, J. V. L. (1929).—“‘ Cretaceous Fossils from Angola (Lamellibranchia

and Gastropoda), Ann. 8. Afr. Mus., vol. xxviii, pp. 1-54.

(52) Rogers, A. W., and Scuwarz, HE. H. L. (1902).—‘‘ Geological Survey of the

Rocks in the Southern Parts of the Transkei and Pondoland, including a

Description of the Cretaceous Rocks of E. Pondoland,”’ Ann. Rep. Geol.

Comm. Cape of Good Hope, for 1901, pp. 38-46.

(53) Romer, F. (1852).-—-Die Kreidebildungen von Texas.

(54) SowERBy, J., and SowErsy, J. pE ©. (1812-46).—The Mineral Conchology

of Great Britain.

(55) SowERBy, J. DE C. (1835).—(In SEpG@wick and Murcuison: Trans. Geol.

Soc. Lond., ser. 2, vol. iii, pp. 417-420.)

(56) SowERBy, J. DE C. (1836).—(In Firron: Trans. Geol. Soc. Lond., ser. 2,

vol. iv, pp. 335-348.)

(57) Sparx, L. F. (1921).—‘* On Upper Cretaceous Ammonoidea from Pondoland,”’

Ann. Durban Mus., vol. iii, pp. 39-57.

(58) Spatu, L. F. (1921).—‘* On Cretaceous Cephalopoda from Zululand,” Ann.

S. Afr. Mus., vol. xii, pp. 217-321.

(59) Sparu, L. F. (1922).—‘‘ On the Senonian Ammonite Fauna of Pondoland,”’

Trans. Roy. Soc. S. Afr., vol. x, pp. 113-147.

(60) Spatu, L. F. (1925).—“‘ On Upper Albian Ammonoidea from Port. E. Africa,

etc.,’ Ann. Transv. Mus., vol. xi, pp. 179-200.

(61) Sroniczka, F. (1868).—‘‘ Cretaceous Fauna of S. India: Gastropoda,”

Pal. Indica.

(62) SrouiczKa, F. (1871).—‘‘ Cretaceous Fauna of S. India: Lamellibranchia,”

Pal. Indica.

(63) Sturm, F. (1901).—‘‘ Der Sandstein von Kieslingswalde in der Grafschaft

Glatz und seine Fauna,’’ Jahrb. d. k. Preuss. Geol. Landesanstalt, vol. xxi,

f. 1900, 2, ii, pp. 39-98, pls. 11—xi.

(64) vAN Hoepen, E. C. N. (1920).—“‘ Descriptions of some Cretaceous Ammonites

from Pondoland,” Ann. Transv. Mus., vol. vii, pp. 142-147.

(65) vAN Horpgy, E. C. N. (1921).—‘‘ Cretaceous Cephalopods from Pondoland,”’

Ann. Transv. Mus., vol. viii, pp. 1-48.

(66) van Hoepen, E. C. N. (1926).—‘‘ Oor die Krytafsettinge van Soeloeland,”’

S. Afr. Journ. Sci., vol. xxiii, pp. 216-222.

(67) Wapz, B. (1926).—‘‘The Fauna of the Ripley Formation, Tennessee,”

U.S.G.8. Prof. Paper, No. 137.

VOL. XXVIII, PART 2. 18

260 Annals of the South African Museum.

(68) VeRRILL, A. E. (1899-1900).—‘‘ A Study of the Family Pectinidae, with a

Revision of the Genera and Subgenera,”’ Trans. Connecticut Acad., vol. x,

pp. 41-96. ;

(69) WuitHaves, J. F. (1903).—“* On some Additional Fossils from the Vancouver

Cretaceous,” Geol. Surv. Canada, Mesozoic Fossils, vol. i, pt. v.

(70) WuiITrIELD, R. P. (1885).—‘‘ Brachiopoda and Lamellibranchiata of the

Raritan Clays,’ U.S.G.S. Mon., vol. ix.

(71) WitcKEens, O. (1904).—‘‘ Revision der Fauna der Quiriquina-schichten,”

N. Jahrb. f. Min., Beil.-Bd. xviii, pp. 181-284.

(72) WitcKENS, O. (1907).—Die Lamellibranchiaten, Gastropoden, etc., der

oberen Kreide Stidpatagoniens,”’ Ber. Naturforsch. Gesellsch. Freiburg,

Bd. xv, pp. 97-166.

(73) WitcKEns, O. (1910).—‘ Die Anneliden, Bivalven und Gastropoden’ der

Antarktischen Kreideformation,”’ Wiss. Ergeb. der Schwed. Sidpolar-

exped., 1901-1903, Bd. iii, lief xii. |

(74) WitcKens, O. (1922).—‘‘ The Upper Cretaceous Gastropods of New Zealand,”

N.Z. Geol. Surv. Pal. Bull., No. 9.

(75) Woop, 8. V. (1850).—‘‘ Mon. Crag Mollusca, Part II: Bivalves,” Mon. Pal.

Soe.

(76) Woops, H. (1899-1913).—‘‘ Mon. Cretaceous Lamellibranchia of England,”

Mon. Pal. Soc.

(77) Woops, H. (1906).—‘‘ The Cretaceous Fauna of Pondoland,”’ Ann. 8. Afr.

Mus., vol. iv, pt. vii, pp. 275-350.

(78) Woops, H. (1917).—‘“‘ The Cretaceous Faunas of the North-eastern Part of —

the South Island of New Zealand,” N.Z. Geol. Surv. Pal. Bull., No. 4. 4

(79) ZEKELI, F’. (1852).—‘‘ Die Gasteropoden der Gosaugebilde,” Abhandl. d. k. k.

Geol. Reichsanstalt, Wien.

(80) ZitreL, K. (1864-6).—‘“‘ Die Bivalven der Gosaugebilde,” Denkschr. d. k.

Akad. d. Wiss., Wien, Math.-nat. Classe, vols. xxiv, xxv. P

(81) ZitreL, K. (1913).—Text-Book of Palaeontology, English edition, ed. C. R. |

Eastman.

EXPLANATION OF PLATES.

The figures are of natural size unless otherwise stated.

PratE XVI.

PAGE

1. Trigonoarca capensis (Griesbach) . : : ; : : an 70s

Interior view of left valve ; T.M. 216.

2-5. Nucula kaffraria sp. nov. : :

2. Holotype, right valve ; T.M. 1647.

Ds Be part near anterior X 3.

4, oe dorsal view.

oa a posterior view.

168 —

As - 6, 7. T'rigonoarca elongata sp. nov. : : ; ; : « GS

5 a ye 6. Holotype, dorsal view ; T.M. 1635.

4 fe + right valve.

8-1l. Glycimeris ct. subplanta (Stoliczka) , a : bat) aes 2 eel aie

8. S.A.M. 8415.

9. a Solos.

10. = ,, natural size.

ll. A =» Gorsal vaew <2.

12-15. Pecten (Camptonectes) kaffraria sp. nov. . : : ‘ < ~) is

12. Type of left valve ; T.M. 1601.

13. Left valve ; T.M. 1600.

14. Type of right valve ; T.M. 1602.

15. $3 S ns part of surface x 5.

Plate XVI.

TF 4p

Shak Chie tes

E. T. Talbot, dei, New & Co., Lid.

i NY dat

Be raat

; if Wf i i

pan

1 a

aint) ;

Ne an

PuLatTe XVII.

TES, = PAGE

2s Ostreaisps at , : : , : : , : : Pe br (5

1. Left valve, exterior view ; T.M. 1571.

DARN Bes »» interior view.

3, 4. EHxogyra decussata Goldfuss . t ; : : ; Rene

3. Left valve, posterior view ; S.A.M. 8562.

bee » exterior view.

5-7. Alectryonia cf. arcotensis (Stoliczka) : ; , ; : ened iS

5. Left valve; S.A.M. 8439.

Gia ss interior view ; S.A.M. 8437.

me: a3 exterior View.

Plate XVII.

E. T. Talbot, del. Neill & Co., Ltd.

WEI ney Pag Hi i

Ne: ah

Neh Se

eit

wae iM i

re kit a |

nine aM ' i" fia

iM),

7 i 1 Hy

f th

ee Y ote

& Pp ean

LAL

ia et nh

‘ j

Minit

“hss

‘

Puate XVIII.

RGSS aun PAGE

1-5. Neithea quinquecostata (Sowerby) . : : : - ats

1. Left valve, exterior view ; S.A. M. 8380.

Hae ke » interior view.

3. Right valve, exterior view ; Durb. Mus.

Ao tiptar: » hinge seen from below x 2.

Out a5 >, interior view.

6. Cercomya arcuata (Forbes) . : P ‘ : a . 189

Cast of left valve ; T.M. 1655.

7, 8. Astarte ? amapondensis sp. nov. . : : = 5 5 = 7198

7. Holotype, right valve ; S.A.M. 8639.

8. - dorsal view.

9. Pinna vanhoepeni sp. nov. . : : 5 ° : : Ld?

Right valve of holotype ; T.M. 1630.

Plate XVIII.

Neill & Co., Lid.

& "6 ALaiye Se

‘ve 4

5 nat gnats

Dy eee

te tegho7 bhp

hhhil 1 Shige ely fia

ca Pi recs

Ann. 8. Afr. Mus., Vol. XXVIII.

£. T, Talbot, del.

‘\

[

ied

: itl tn

PLATE XIX.

FIGS. : PAGE

1-6. Modiola typica (Forbes) F : : ' é : : . 184

1. Left valve ; T.M. 1596.

DN a dorsal view.

3. Juvenile example, left valve ; T.M. 1597.

4, 5 35 dorsal view.

5. Juvenile example, left valve ; T.M. 1598.

6. a an dorsal view.

7, 8. Lima (Acesta) obliquistriata Forbes é : ; : : . 183

7. Left valve ; Durb. Mus.

8. Right valve ; Durb. Mus.

9. Pinna cf. decussata Goldfuss p F A : 4 ; ae eee

Left valve ; S.A.M. 8692.

£. T. Talbot, del.

Plate XIX.

Neill & Co., Ltd.

RON

fa’ ni

ri r Ads

PLATE XX.

FIGS.

1-4. Goniomya umzambiensis sp. nov. . ; :

1. Holotype, right valve x 2; T.M. 1588.

Par is natural size.

tes s dorsal view.

4, Paratype, right valve ; T.M. 1589.

5-7. Spondylus ct. calearatus Forbes

5. Right valve ; T.M. 1568.

6. Left valve, interior view ; T.M. 1569.

Doe: ws exterior view.

8-10. Pleuromya africana (R. Etheridge, jr.)

; 8. Right valve ; T.M. 1657.

Oiye® x, dorsal view.

102 Gas as part of surface X 3.

11,12. Pholadomya ct. elliptica Minster

11. View of left valve ; T.M. 1586.

12. of “a ve dorsal view.

13, 14. Pholadomya wmzambiensis sp. nov.

13. Holotype, right valve ; S.A.M. 8475.

14. 3 dorsal view.

PAGE

188

182

186

188

187

Plate XX.

Ann. S. Afr. Mus., Vol. XXVIII.

Neill & Co., Lid.

£. T., Talbot, del.

FIGS.

1-4.

5-9.

10.

11.

12, 13.

14-16.

17,18. Cyprimeria ? (Cyclorisma 2) cf. analoga (Forbes)

19-21.

PuatE XXI.

Liopistha (Psilomya) corrugata Woods

1. Right valve ; T.M. 1640.

DR »» part of surface X< 4.

ies » IT.M. 1648.

Atri bes », dorsal view.

Veniella drut (Munier-Chalmas)

5. Right valve ; T.M. 1632.

6.. 5 T. M. 1634.

7. Another right valve ; T.M. 1633.

8. ah A dorsal view.

9. a ic interior view.

Veniella forbesiana (Stoliczka)

Small right valve ; T.M. 1590.

Anisodonta 2? umzambiensis sp. nov.

Right valve ; T.M. 1654.

Crassatellites haughtoni sp. nov.

12. Holotype, exterior view ; S.A.M. 8499.

13. s interior view.

Dosiniopsis geversi sp. nov.

14. Holotype, left valve, interior view ; S, A. M. 8: 518.

15. on exterior View.

16. ae hinge X 2.

17. Left valve ; T.M. 1579.

Tits tess nA dorsal view.

Cyprimeria ? natalensis sp. nov.

19. Holotype, right valve ; Durb. Mus.

20. Me anterior view.

21. a dorsal view.

PAGE

190

191

192

194

193

195

196

XXVIII. Plate X XT.

16 18 20 21

E. P. Talbot, del. Neill & Co., Ltd.

Th ;

iets Ut

ih i

he has et

4), Get ee

Puate XXII.

FIGS. PAGE

1-9. Trigonocallista spathi sp. nov. : : : 3 : : - 198

1. Type of left valve, dorsal view ; Durb. Mus.

De a i exterior View.

3. gh af interior view.

4. Smaller left valve, dorsal view ; Durb. Mus.

a As exterior view.

ES 5 hinge.

. Type of right valve, interior view ; Durb. Mus.

> 9 exterior view.

35 - dorsal view.

i ee

10-12. Macrocallista (Callistina) euglypha (Woods) . i : : ~) gS

10. Hinge of right valve x 3; S.A.M. 8588.

11. Left valve; S.A.M. 8589.

12 oS sn hinge X 3.

Plate XXII.

Neill & Co., Lid.

. 8S. Afr. Mus., Vol. XX VITT.

£. T. Talbot, del.

ly te

Mh My Me

ta ;

lg WE ite

i“ ie

i Way

FIGS.

9, 10.

Whe te

13, 14.

18-20.

PLATE ee

Palaeomoera u.nzambiensis sp. nov.

1. Holotype, left valve, exterior view ; T. M. 1639.

Ds My hinge X 2.

3. » part of surface xX 3.

. Palaeomoera sp.

4. Right valve, axterGr: view ; T. M. 1638.

De ls » part of surface x 3.

. Palaeomoera haughtoni sp. nov.

6. Holotype, left valve ; T.M. 1637.

he right aie ; T.M. 1649.

8. of dorsal view.

Mactra? kaffraria sp. nov. j :

9. Left valve, holotype ; T.M. 1652.

LOS », dorsal view.

Macoma papyracea sp. nov. . ; :

11. Holotype, dorsal view ; T.M. 1636.

es - view of left valve.

Martesia ? sp.

13. Dorsal view ae specimen ; ‘T. M. 1592.

14. oe 5 = left valve.

. Cymbophora rogersi sp. nov. .

15. Holotype, left valve, exterior view ; T. M. 1651.

16. EAs interior view.

ie a dorsal view.

Solecurtus (Azor) woodsi sp. nov. -

18. Holotype, right valve ; T.M. 1584.

19. Paratype, hinge of left valve ; S.A.M. 8580.

20. bis interior of left valve ; T.M. 1585.

PAGE

200

201

205

201

203

202

Ann. S. Afr. Mus., Vol. XXVIII.

< 44- ANS

vt SAR ‘

£. T. Talbot, del.

Plate XXIII.

Neill & Co., Lid

BA Wise oh

Leys aN '¢

hia

Peet,

FIGS.

1-4.

11.

12-15.

16-18.

19-21.

Puate XXIV.

Patella kaffraria sp. nov.

1. Holotype, dorsal view ; S.A.M, 8477.

Ds MS lateral view.

3. Paratype, lateral view ; S.A.M. 8572.

4. a dorsal view.

. Nerita kaffraria Woods

Specimen, showing original colour bandee fi M. 1604.

. Nerita umzambiensis Woods .

6: ee view of unusually large specimen ; T. M. 301.

. Dorsal view of same.

. Chilodonta (Agathodonta) africana sp. nov.

8. Holotype, natural size ; S.A.M. 8630.

9. < dorsal view xX 2.

10. an apertural view x 2.

Margarites (Solariella) radiatula (Forbes)

Specimen, showing spiral striae ; S.A.M. 8631.

Turritella (Haustator) meadi Baily

12. S.A.M. 8400

EO Wea et )D)

14. 5. Sa73:

Twos. one of the whorls x 6.

Trajanella dutoiti sp. nov.

16. Holotype, apertural view ; 5, Diab: Mins:

ize BA basal view.

18. ed dorsal view.

Lysis capensis sp. nov. ;

19. Holotype, dorsal view ; T. M. 1615.

20. Es apertural view.

Palle “H posterior view.

PAGE

206

209

207

208

214

210

211

1. Talbot, del.

Plate XXIV.

Neill & Co., Ltd.

11-15.

16-18.

19, 20.

21-28.

24-26.

27, 28.

PLATE XXV.

. Gyrodes tenellus Stoliczka

1-3. Low-spired form; S.A. M. 8730.

4, 5. High-spired form ; S.A.M. 8714.

6; 7% Fr x S.A.M. 8478.

. Arcotia vanhoepent sp. nov.

8. Holotype, dorsal view x 3 ; Tach Mas.

9. an natural size.

10. - basal view xX 3.

Dicroloma (Perissoptera) bailyi (R. Etheridge, jr.)

11. Dorsal view of complete specimen ; Durb. Mus.

2 ys »» apertural view of complete specimen.

13. Apertural view ; T.M. 1612.

14. portion of surface enlarged X 3.

15. wine TM, 16138.

Dicroloma (Perissoptera) sp.

16. Dorsal view ; S.A.M. 8699.

| hy eae a portion of surface enlarged x 3.

1 io Figuah o ss apertural view.

Dicroloma (Perissoptera) sp.

19. Dorsal view ; S.A. M. 8494.

20. os », portion of surface eniateed DK ess

Pirula (Protopirula) capensis sp. nov.

21. Holotype, posterior view ; T.M. 1626.

22. x. apertural view.

23. 3 dorsal view.

Pugnellus contortus (Forbes) .

24. Apertural view ; T.M. 1611.

25. 3 dorsal view.

26. Posterior view.

Tritonidea (Cantharulus) kaffraria sp. nov.

27. Holotype, apertural view ; Durb. Mus.

28. _ dorsal view.

PAGE

212

215

217

218

222

218

Plate X XV.

Neili & Co., Lid.

Med

Pag it)

ina)

PLATE XXVI.

FIGS. PAGE

220

1-6. Cypraea chubbi sp. nov. : : :

1. Holotype, basal view ; Durb. Mus.

Oe a posterior view.

3. Bs dorsal view.

?4. Smaller specimen, posterior view ; Durb. Mus.

25. a 3 basal view.

26. BA a dorsal view.

Ann. S. Afr. Mus., Vol. XX VITI. Plate XXVI.

ore

£. T. Talbot, del. Neill & Co., Lid.

eee

Mek

4 ,

’ Ci:

Pratt X XVII.

il FIGS. PAGE

| ) 1-4. ? Cypraea chubbi sp. nov. . ‘ : : : ; » 220

| 1. Posterior view ; Durb. Mus.

+ 2. Small example, posterior view ; T.M. 1624.

A | Sn os Bs dorsal view.

, | | Aas. 33 basal view.

| 5. Siphonalia ? umzambiensis sp. nov. : 223

i Holotype, dorsal view, the last whorl broken away ; 8. A M. 8416.

ii I 6-8. Lathyrus (Mazzalina) geversisp. nov. . ‘ : ; >, 224

ii 6. Holotype, dorsal view x 2; S.A.M. 8486.

ey) | re us apertural view.

\ i” 8. is natural size.

| i 9-12. Cryptorhytis rigida (Baily) . ‘ A . 225

| 9. Perfect specimen, apertural view xX 2; Dark, Mus.

! | 10. S ae natural size.

. | Blea ness = dorsal view.

12. Larger specimen ; Durb. Mus.

i 13,14. Cryptorhytis ? sp. ; 5 : 5 : : . 228

! 13. Apertural view ; S. A. M. 8640.

iT 14. “0 , dorsal view.

if 15,16. Paleopsephaea kaffraria sp. nov. . : : : ; BP 45)

15. Holotype, dorsal view ; Durb. Mus.

my) | 16. % apertural view.

th q C na no 17-19. Pyropsis geversi sp. nov. : : : ; : : . 229

3 | 17. Holotype, apertural view ; . SAM. 8551.

l 1, 18. op dorsal view.

i | 19. 5 posterior view.

a |

| }

1 | 20-22. Bullinella sp. ine. Se

} | 20. Specimen, apertueal view X 3; 's, AM. 8597.

M Ze As dorsal view x 3.

I 2s x9 natural size.

ath 23,24. Paleopsephaea scalaris sp.nov. . : : . . 228

mh || 23. Holotype, apertural view ; T. M. 1606.

\ . 24. 3 dorsal view.

i) | 25, 26. Semifusus (Mayeria) acuticarinatus sp. nov. . : : : . 230

lt H ! 25. Holotype, apertural view ; T.M. 1620.

| | I 26. 2 dorsal view.

ee — ———e — ——— i ——————— ee —

Plate X XVII.

Neill & Co., Lid.

£. T. Talbot, del.

a SS ee

- --—— = = : SS

—— ee _ —— ———— ——

ee ———— = SS ==

Se a ee oe SS ee eee

— a ———$——_—$—$$5——_—- — — -—__<—__ — Foe

ee = — ~ = = —

wm aes =——~— =

Lory

it Ale it a)

FIGS.

. Semifusus (Mayeria) umzambiensis sp. nov.

3-6.

12-16.

17-21.

Puate XXVIII.

1. Holotype, apertural view ; T.M. 1617.

os es dorsal view.

Semifusus (Mayeria) kaffraria sp. nov. . :

3. Holotype, apertural view ; T.M. 1618.

4, 33 dorsal view.

5. Paratype, dorsal view ; T.M. 1619.

6. 55 apertural view.

. Ringicula woodsi sp. nov.

7. Holotype, apertural view x 3; 'S. A. M. 8528.

8. ze dorsal view.

9. af natural size.

10. Paratype, dorsal view x 3; S.A.M. 8493.

Ad: Be apertural view.

Anchistoma umzambiense sp. nov. .« ‘ :

12. Holotype, posterior view ; T.M. 1628.

US sf basal view.

14, Ey apertural view.

15: a aperture X 3.

16. $5 part of spire x 6.

Actaeonella (Trochactaeon) woodsi sp. nov.

17. Nearly smooth example, dorsal view ; Dab: Maa,

18. ts 5 Be apertural view.

19. Holotype, apertural view ; T.M. 1621.

20. SS dorsal view.

21. Large example, apertural view ; T.M. 1623.

PAGE

231

235

232

Plate XXVIII.

Neill & Co., Lid.

PLATE XXIX.

FIGS. PAGE

1-5. Veniella forbesiana (Stoliczka) : : : : . 244

Upper Senonian of Lake Itesa, iitblone

1. Right valve; S.A.M. 4887a.

De es A interior view X 2.

Su bss =p dorsal view.

4. Left valve ; S.A.M. 4896.

pia tae nd hinge xX 2; S.A.M. 48878

6, 7. Nucula zululandensis sp. nov. : : : : . 239

Upper Senonian, railway cutting, ‘Umrolont

6. Holotype, left valve ; S.A.M. 4869.

is a, dorsal view.

8-10. Hriphyla lenticularis (Goldfuss) . ; : : . 245

Upper Senonian of Lake Itesa, Urntolosis

8. Left valve, exterior view ; S.A.M. 4884.

Ora >, interior view.

1Otaes,. 5» _ hinge <2:

11-16. Macrocallista (Callistina) zululandensis sp. nov. : A : . 247

Upper Senonian of Lake Itesa, Umfolosi.

11. Holotype, left valve, hinge ; S.A.M. 4889.

1. A exterior view.

He35 mr! dorsal view.

Upper Senonian of Umkwelane Hill, Umfolosi.

14. Paratype, left valve, dorsal view ; S.A.M. 5495.

15. 35 exterior view.

16. = hinge.

17-19. Cyprimeria? sp.. ; : : . 249

Upper Senonian of Take ieee Unrolord

17. Right valve, exterior view ; S.A.M. 4885.

18. A »» dorsal view.

19. ss anterior view.

Plate X XIX.

Neill & Co., Lid.

Hil FIGS.

fe 1,2

‘il

' | ‘6

ay il 24

- |

i it

a 5, 6.

p |

| 7-9

[

' 10-12.

: i 13-18.

PLATE XXX.

. Semifusus (Mayeria) natalensis sp. nov.

Upper Senonian of Umkwelane Hill, Wintotesa

1. Holotype; S.A.M. 5407.

oe : part of surface of last whorl.

. Cyclothyris sp.

Senonian, south site of Manan Grek valle

3. Dorsal valve ; S.A.M. 4942.

Ae ales, oA anterior view.

Cyclothyris Sp.

Senonian, railway cutee Dmtolod

5. Dorsal valve ; S.A.M. 4878.

6. Hs a apical view.

. Cyclothyris sp.

eee railway outnne: lumtoleet

. Dorsal valve ; S.A.M. 4879.

‘ 35 Bs anterior view.

9. 45 Le apical view.

Neithea quadricostata (Sowerby)

Albian of the Manuan Creek.

10. Anterior view ; 8.A.M. 2590c.

11. Right valve of same specimen.

12. Left = a mS

Macrocallista (Callistina) andersoni (Newton) .

Albian ?, south side of the Manuan Creek Valley.

13. Left valve, dorsal view ; S.A.M. 4936.

1a es > exterior view.

I ce apes Bre imaveges NE Bh.

16. Right valve, interior view ; S.A.M. 4938.

le As » exterior view.

18. =p ao ingen <a

PAGE

249

250

241

246

Plate XXX.

Neili & Co., Lid.

E. T. Talbot, del.

PuaTeE XXXII.

FIGS. PAGE

1-3. Veniella etheridgea R. B. Newton . : : : : : . 242

Albian of the Manuan Creek.

1. Left valve, interior view ; S.A.M. 2723.

Di aes » exterior view.

Se 5 3 dorsal view.

Ann.

S. Afr. Mus., Vol. XXVIII.

Plate XX XT.

Neill & Co., Lid.

pees = s ~~ =< a

Se — —-

( 261 )

7. Pareiasaurian Studies.

Part V.—On the Parevasaurian Mandible.—By S. H. Havueuron,

B.A., D.Sc., Hon. Curator, Palaeontological Collections, and

L. D. Boonstra, M.Sc., Assistant in Palaeontology.

(With Plates XXXII-XXXVI and Text-figures 4-16.)

HISTORICAL.

In the type specimen of Pareiasaurus serridens the lower jaw shows

considerable post-mortem deformation. Owen (1) was therefore

unable to describe the form very well, but he gave a section of the

ramus which is substantially correct.

Owen (1) made rather a serious error in his description of Brady-

saurus bombidens (= Pareiasaurus bombidens), in that he mistook the

ventral for the dorsal surface of the skull. What he described as the

mandible is in reality the maxillary and pre-maxillary.

The lower jaw of Anthodon serrartus was very inadequately de-

scribed by Owen (1). The fact, however, that he noted that the teeth

differed from those of the two other Pareiasaurians in having the more

numerous cusps regularly arranged is of importance.

In his account of Bradysaurus seeleys (described as Parevasaurus

bombidens) Seeley (2) gave the first good description of the general

form of the Pareiasaurian mandible. His account of the detailed

arrangement of the various bones 1s, however, incorrect.

Seeley’s (3) description of the mandible of Bradysaurus bani ( =

Parevasaurus bani) does not add much to his previous work on that

of Bradysaurus seeleyi. He noticed, however, that the arch appears

wider and more horseshoe-shaped when viewed from its ventral side

than it does from the alveolar aspect.

The jaws of Pareiasaurus roussowt were fragmentary, and the only

point of importance noticed by Seeley (3) was that the cusps of the

teeth were numerous and regular as in Anthodon.

Further additions to our knowledge of the Pareiasaurian mandible

were made by Broom (4) in his description of Propappus omocratus.

262 Annals of the South African Museum.

The presence of two angular bosses was recorded, but it was erron-

eously stated that there was no coronoid preserved. In the sequel a

figure (fig. 14) is given from which it will be clearly seen that the

anterior portion of the coronoid is preserved.

Watson (5) in his study, ““On some Reptilian Lower Jaws,” made

the first attempt at a detailed elucidation of the relations of the

various mandibular bones. Although his conclusions were incorrect

in a number of points, he was prepared to commit himself by giving

figures, which previous authors had not attempted.

In 1913 Broom (6) studied Koalemosaurus acutirostris (= Pareia-

saurus acutirostris), and noted the presence of both coronoid and the

_weak coronoid process. He was uncertain of the presence of the

prearticular.

In the same year Broom and Haughton (7) gave an inadequate

account of the mandible of Parevasuchus peringueyr. No figures were

given. Since this jaw has been properly freed of matrix it has become

evident that Broom and Haughton had no justification for expressing

uncertainty about the presence of the coronoid and the exact nature

of the prearticular and surangular.

In the following year Broom (8) recorded the presence of two

angular bosses in Bradysuchus whats: (= Parevasaurus whaitst).

It was not until 1924 that the structure of the Parelasaurian man-

dible was correctly elucidated. In that year Broom (9) gave three

figures which may be taken as applying in general features to all

Parelasaurians.

This paper is based on a large number of specimens of nearly all the

species of Parevasauridae. Our first object has been to determine

whether the general structure is the same for all species, and whether

this confirms Broom’s account. Secondly, an attempt has been made

to record the smaller variations with the object of determining their

value as specific characters. Lastly, some space has been devoted to

comparative considerations.

MATERIAL.

For purposes of reference it may be useful if the material on which

this paper is based is tabulated as follows :—

Cat. No.

4347

5624

3718

8941

5019

5590

6239

8944

3717

6238

5012

6240

Name.

Bradysaurus

baint

Bradysaurus

seeleyt

Bradysaurus

vanderbyla

Bradysaurus

vanderbyla

Nochelesaurus

strubenr

Nochelesaurus

strubent

Nochelesaurus

alexander

Nochelesaurus

alexanderi

Dolichopareia

angusta

Dolichopareia

angusta

Brachypareia

rogerst

Brachypareia

watsont

Embrithosaurus

schwarzr

Anthodon

Serrarvus

Pareiasuchus

peringueyt

Pareiasuchus

NASUCOPNIS

Propappus

omocratus

Bradysaurus

seeleyt

Indet.

Brachypareia

watsont

Dolichopareia

angusta

Bradysaurus

bane

Indet.

Nochelesaurus

strubent

Bradysaurus sp.

Brachypareia

watson

Indet.

Brachypareia

watson

Indet.

Geographical

Locality.

Abrahams Kraal,

Prince Albert

Groot Kruidfontein,

Prince Albert

Abrahams Kraal,

Prince Albert

Mynhardt’s Kraal,

Beaufort West

Blaauwkranz, Prince

Albert

Abrahams Kraal,

Prince Albert

Boesman’s Kop,

Beaufort West

Klein Koedoeskop,

Beaufort West

Leeuw River, Beau-

fort West

Boesman’s Kop,

Beaufort West

Abrahams Kraal,

Prince Albert

Abrahams Kraal,

Prince Albert

Hoogeveld, Lot A,

Prince Albert

Dalham, Graaff

Reinet

Dunedin, Beaufort

West

Commonage, Graaff

Reinet

Commonage, Graaff

Reinet

Vogelfontein, Prince

Albert

Melkboschwater,

Prince Albert

Veldman’s_ Rivier,

Prince Albert

Hottentot’s River,

Beaufort West

Abrahams Kraal,

Prince Albert

Grootfontein, Beau-

fort West

Abrahams Kraal,

Prince Albert

Vogelfontein, Prince

Albert

Fraserburg Road,

Prince Albert

Sewefontein, Prince

Albert

Mynhardt’s Kraal,

Beaufort West

Vogelfontein, Prince

Albert

Vogelfontein, Prince

Albert

Riet Kuil, Beaufort

West

263

Geological

Horizon.

Tapinocephalus

Zone—low

Tapinocephalus

Zone—low

Tapinocephalus

Zone—low

Tapinocephalus

Zone—high

Tapinocephalus

Zone—low

Tapinocephalus

Zone—low

Tapinocephalus

Zone—high

Tapinocephalus

Zone—high

Tapinocephalus

Zone—high

Tapinocephalus

Zone—high

Tapinocephalus

Zone—low

Tapinocephalus

Zone—low

Tapinocephalus

Zone—middle

Cistecephalus

Zone—low

Border of Hndo-

thiodon and

Cistecephalus

Zones

Endothiodon

Zone

Endothiodon

Zone

Tapinocephalus

Zone—umiddle

Tapinocephalus

Zone—low

Tapinocephalus

Zone—low

Tapinocephalus

Zone—high

Tapinocephalus

Zone—low

Tapinocephalus

Zone—high

Tapinocephalus

Zone—low

Tapinocephalus

Zone—middle

Tapinocephalus

Zone—low

Tapinocephalus

Zone—middle

Tapinocephalus

Zone—high

Tapinocephalus

Zone—middle

Tapinocephalus

Zone—middle

Tapinocephalus

Zone—high

Remarks.

Complete arch.

Articular end of right

ramus missing.

Complete arch.

Articular end of right

ramus,fragmentof left.

Mid. part right ramus.

Complete arch.

Articular end of left

ramus missing.

Complete arch.

Ant. 4 of right ramus,

fragment of left.

Complete arch.

Left ramus and anterior

part of right.

Right ramus.

Anterior part of arch.

Left ramus.

Right ramus and pos-

terior part of left.

Fragments of both rami.

Anterior part of arch.

Anterior part of left

ramus.

Left ramus.

Posterior part of right

ramus.

264 Annals of the South African Museum.

Dr. Broom has been so kind as to send us a number of jaws in his

collection for examination. This material consists of ? of the left

ramus, the posterior part of the right ramus of Bradysaurus seeleyr; a

complete arch of Brachypareia watsoni; and the anterior portion of

the arch of Brachypareia watsoni. —

GENERAL STRUCTURE.

The mandibular arch lies wholly within the maxillary arch, so that

the teeth it carries pass on the inside of the maxillary teeth. In

consequence the flattened surfaces of the serrate crowns of the man-

dibular teeth are directed externally, whereas those of the upper arch

lie internally. The two flattened surfaces of the crowns are thus

opposed to each other, but on account of there being a considerable

interval between them they do not come in contact. Thus the teeth

neither cut nor ground, but served simply as rugosities to prevent the

soft vegetation from slipping out after a mouthful had been taken.

In shape the mandibular arch is more or less U-shaped with the rami

parallel, converging or diverging posteriorly. In the narrow-snouted

forms (e.g. Dolichopareia angusta) the pointed nature of the snout

hardly affects the breadth of the mandibular arch. In the broad-

snouted forms (Nochelesaurus strubent particularly) the rami are short,

and this ratio between length and breadth produces an appearance

of a broader arch. The greater length of the rami in the narrow-

snouted forms produces the contrary effect.

From evidence partly derived from the nature of the teeth it has

been maintained that the Pareiasauridae fed on very soft cryptogamous

plants. In view of this fact, it is remarkable that the lower jaw should

be so massive as it undoubtedly is. This massiveness is, however, in

keeping with the general robustness of the skeleton. These animals

had in fact run to bone. The American Cotylosauria (10) are in general

also characterised by a massive mandible; in fact, that of Labido-

saurus hamatus is proportionally much more massive than that of any

Pareiasaur. Amongst the different species of the Pareiasauridae a

difference in the massiveness is apparent. Nochelesaurus strubent has

a very deep mandible, whereas in forms like Brachypareia rogers: it is

quite shallow. This difference in depth is most apparent at the

symphysis menti. .

Concomitant with the general variation in body-size amongst the

various species of Pareiasaurs, there is a considerable degree of

variation in the absolute dimensions of the mandible ; e.g. the length

Pareiasaurian Studies.—Part V. 265

is 175 mm. in Anthodon serrarvus and 365 mm. in Bradysaurus vander-

byli. Relatively, however, there is no very great difference in length.

The most obviously striking feature in the Pareiasaurian lower jaw

is the presence of a protuberance on the mid-ventral surface. This

protuberance varies in shape and size in the various species from a

slight pyramidal swelling to a large process in some cases resembling a

bovine teat, in others a door-knob, whilst in others it is horn-like. In

two Pareiasaurs two such processes are developed. The larger is

situated posteriorly and is homologous with the single process in the

other species. Immediately in front of the larger process there is in

Bradysuchus whaitsi and Propappus omocratus a similar but much

smaller process. In the discussion on the insertion of the depressor

muscles it will be seen that the larger processes served as points of

insertion. The smaller anterior processes could not have this function.

It may just be possible that they are pathological. The presence of

these bosses on the angular is unique among reptiles. An analogous

(not homologous) development is, however, present amongst the

Elotheres—a group of giant pigs. By the development of this localised

surface for muscular insertion a reduction in massiveness of the jaw

becomes possible, which is not the case in a form like Labidosaurus

hamatus where the insertion is diffuse and spread over the whole

ventral angular surface. The other end of the attachment (origin) of

the depressor muscles is on the posterior border of the cheek. It may,

therefore, be possible to bring the great development of the cheek into

correlation with the massiveness of the mandible.

With the exception of Nochelesaurus strubent, where the ratio is

equality, that part of the jaw lying anteriorly to the angular boss is

longer than that lying posteriorly. In all the species the posterior

portion of the jaw is much less massive than the anterior. The

articulation is in fact weak. The forward position of the depressor

muscles has resulted in a reduction of the rdle played by the posterior

articulatory portion of the jaw.

There are five fenestrae leading into the interior of the mandible

which is (with the exception of the posterior end formed by the

articular) hollow. These fenestrae are, however, relatively small and

do not materially decrease the mass of the mandible. The largest is

the supra-meckelian fossa situated on the dorsal surface immediately

posterior to the dentigerous border. It is bounded by the coronoid,

surangular, articular, and prearticular. The inner mandibular surface

is perforated by two foramina. More or less in the same plane as the

angular boss is a small foramen bounded by the angular, splenial, and

—

266 Annals of the South African Museum.

prearticular. Lateral to the median thickened symphysis there is a

narrow slit-like perforation which separates the dentary from the

splenial. On the outer surface a small foramen perforates the sur-

angular. On the outer surface of the dentary there is a quite small

foramen. Through the supra-meckelian fossa and the fossa lateral to

the symphysis, muscles as well as blood and nerve vessels passed.

The other three foramina were probably solely for the reception of

nutritive vessels and nerves.

As in all aplacental vertebrates, the Parevasauridae have a composite

mandible. It contains seven out of the ten elements which have been

determined in the vertebrate lower jaw. These seven elements are

best treated separately. They are—Articular, Angular, Surangular,

Prearticular, Splenial, Coronoid, and Dentary.

Articular.—The hinge between the lower jaw and the skull hes

between the articular and the immovable quadrate. The articular is

a moderately developed solid bone—in contradistinction to the other

bones forming the lower jaw, which are either splint-like or curved

bones enclosing the intra-mandibular cavity. The articulatory

surface is bipartite, consisting of two cotyli separated by a median

ridge running in an antero-posterior direction. The outer cotylus is

the smaller, and it is only slightly concave. The larger inner cotylus

is deeply concave. Posterior to the cotyli the articular has only a

small extension which forms a weak postarticular process. In fact

this posterior extension is in reality a wall or rampart limiting the

extent to which the jaw can be opened, and it certainly does not serve

as the main point of insertion of the depressor muscle of the mandible.

Anteriorly the articular extends for a short distance as a plug or

wedge into the intra-mandibular cavity, where it is firmly clasped

by all the bones lying immediately anterior to it. Ventrally it is

supported by the angular, which extends a considerable distance up

against the posterior ventral surface. Externally the surangular

flanking the articular stretches right up to the outer border of the

external cotylus. Internally the articular is clasped by the pre-

articular. This flange has, however, not such a great posterior

extension as the surangular has. The articular forms the posterior

border of the supra-meckelian fossa.

This is the only substitution bone in the lower jaw. All the others

are investing or dermal bones.

Angular.—The angular is a large bone of great antero-posterior

extent. The posterior extent has been noted -in describing the

articular. Anteriorly the angular is practically confined to the ventral

Pareiasaurian Studies.—Part V. 267

surface. Superficially it tapers to a point and disappears underneath

the splenial and dentary. From cross-sections it has, however,

become evident that underneath the covering bones the angular

stretches far forward. Externally the posterior edge articulates with

and slightly overlaps the surangular, but its anterior border is

overlapped by the dentary which forms a long squamous suture.

Internally the prearticular overlaps the angular along the suture

running from the infero-posterior meckelian foramen to nearly as far

as the weak postarticular process. The inferior border of the above-

mentioned foramen is for its greater part formed by the angular.

The massiveness of the angular is due to the presence of the solid

boss or bosses. The posterior boss formed the main surface of insertion

of the depressor muscles. Amongst the various species the number,

size, and shape of the angular boss or bosses differ considerably. In

the general account of the jaw as a whole these differences have been

touched upon, and in the systematic part of this paper it will be

determined whether the boss has any value as a specific character.

Surangular.—The surangular is a thin bone lying on the outer

surface of the jaw forming about one-fifth of that surface. At its

supero-posterior border the bone bifurcates. The inner limb is closely

applied to the articular. Posteriorly the bone supports the articular.

Anteriorly it articulates with the coronoid and dentary and is over-

lapped by the latter.

At its inferior border the surangular is overlapped by the angular.

Superiorly the surangular forms the posterior half of the external

border of the supra-meckelian fossa. At about the centre of the

bone there is a small, nearly circular nutritive foramen.

Prearticular.—The prearticular is also a thin bone. It is of con-

siderable size, forming in the more primitive species roughly a half of

the inner surface of the jaw. In the species from higher geological

horizons the prearticular is smaller. In function the prearticular can

be compared to the surangular. Both bones flank and support the

articular—the former on its inner and the latter on its outer surface.

Stretching forward the prearticular is in its turn clasped by the

two limbs of the posteriorly bifurcated splenial, and is furthermore

partly covered by the coronoid so that its full extent is not seen

in superficial view. Cross-sections, however, show that the bone

stretches as a thin plate right up to the superior border of the jaw.

Posteriorly to where it is covered by the coronoid the prearticular

forms the internal margin of the supra-meckelian fossa. Just below

the fossa a ledge is developed on the prearticular. This ledge serves

——_

Sa SSS SSS Se eS ae 4

268 Annals of the South African Museum.

as a surface of insertion of the inner slip of the muscle closing the jaw.

This ledge is more strongly developed in the long-jawed forms.

Coronoid.—The coronoid is a small splint-like bone lying chiefly on

the prearticular plate.

Posteriorly it forms the anterior border of the supra-meckelian fossa

and then passing to the outer surface of the jaw, where it is partly

covered by the extreme posterior point of the dentary, it expands

into a moderate coronoid process. Here its posterior border articu-

lates with the surangular so that it forms the anterior portion of the

external border of the supra-meckelian fossa as well as the whole

anterior margin. Anteriorly it is clasped in a groove in the posterior

margin of the upper limb of the splenial.

The surface of the coronoid process in particular shows that it

formed a surface for muscular insertion.

Splenial.—The splenial is strongly developed and forms the greater

part of the infero-anterior third of the lingual surface of the jaw. At

a plane a little anterior to the plane of the angular boss the posterior

end of the splenial bifurcates. The lower splint-like limb stretches

posteriorly as far as the plane of the angular boss (the posterior larger

boss in Propappus omocratus, a form with two bosses). Here it forms

the whole anterior margin of the infero-posterior meckelian foramen.

The inferior margin of this limb overlaps the angular and its superior

margin the prearticular. The upper splint-like limb does not stretch

so far posteriorly as the lower, and its posterior margin is thickened

and contains a groove wherein the anterior edge of the coronoid fits.

In the more advanced forms (Pareiasuchus peringueyr), where the

prearticular has a reduced superficial surface, the bifurcation of the

splenial takes place more posteriorly.

Anteriorly the splenial forms the lower portion of the symphysis.

On the intero-ventral surface of the symphysis the splenial forms a

well-developed posteriorly directed tuberosity, which served as a point

of insertion of the muscles supporting the floor of the mouth-cavity.

Internally the posterior edge of the upper limb gradually reaches up

to the level of the alveolar border, where it covers the dentary. It

then descends rapidly to form the lower border of the slit-like anterior

foramen into the meckelian cavity which here separates the dentary

and splenial. At the symphysis dentary and splenial meet again.

The ventral surface of the anterior third of the jaw is formed by the

splenial. This surface is more or less flattened, and the suture between

the splenial and dentary runs along the outer edge of this ventrally

flattened surface, so that externally the splenial is hardly visible

Pareiasaurian Studies.—Part V. 269

except at the median plane where it rises to a more or less extent on

the outer symphysal surface.

The degree to which the splenial is externally visible varies amongst

the different species, being more strongly developed at the symphysis

in the later forms (e.g. Parevasuchus peringuey?).

Dentary.—The dentary is a large strong bone forming the greater

part, including the whole anterior portion, of the external surface

of the jaw.

Posteriorly it overlies the angular and surangular in a broad

squamous suture; its upper posterior point also partly overlies the

coronoid as that bone crosses over to the outer surface of the jaw. It

has already been noted that internally the dentary is practically

covered by the coronoid and splenial. Only the upper portion near

the symphysis is uncovered. The two dentaries (as are also the two

splenials) are ankylosed and form a very firm and strong symphysis.

The number, nature, and arrangement of the teeth vary amongst

the different species. These differences will be considered in the

systematic account.

CROSS-SECTIONS (Figs. 4 and 5).

As much of the detailed arrangement of the elements forming the

lower jaw has been studied from cross-sections it appears advisable

to give a selection of such drawings. These drawings are all from

irregular fractures across the jaw (r.=right; J.=—left; post.=

posterior surface of fracture ; ant.=anterior surface of fracture).

In all the illustrations of the lingual surface of the mandible a

cross-section through the symphysis is given. If these are referred

to it will become evident that the symphysis is strong; that in the

older forms the splenial takes a lesser part than the dentary in the

formation of the symphysis, whereas in the younger forms the splenial

part of the symphysis is strong.

I. 1. post. Fracture at plane of third tooth, Dolichopareia

angusta. At this plane the jaw is composed of only two

thick and strong bones. The external surface is mainly

formed by the dentary, which also here forms the upper

part of the lingual surface. The splenial mainly forms

the ventral surface and the lower portion of the lingual

surface. It is to be noted that the splenial is already

commencing to ascend towards the lingual alveolar

border in the form of a thin splint.

8 6

| Btls

—— oe

[ i 270 Annals of the South African Museum.

Il. 1. ant. Fracture at plane of fifth tooth, Parevasuchus

| | peringueyt. Note that the splenial has advanced

Vi higher up the lingual surface. The meckelian cavity

is small.

III. 7. post. The same fracture as the above. A feature to be

noted here is the extent to which the splenial forms the

i internal surface. From the symphysis the splenial

pik rises rapidly towards the inner alveolar border.

‘a IV. v. ant. Fracture at plane of sixth tooth, Pareiasuchus

| iy peringueyt. The high upward extension of the splint-

| li like portion of the splenial is to be noted. Note the

] i great forward extension of the angular in the interior

| i of the jaw after it is no longer visible on the surface.

——_— = =,

V. l. ant. Fracture at plane of eleventh tooth, Dolichopareia

ail angusta. On the upper lingual surface the anterior end

of the coronoid is seen. The forward extension of the

coronoid is therefore considerable.

VI. 1. post. Same fracture as preceding one. From this section

| it is clear that the prearticular is well developed in this

T form and that it stretches far forward. The coronoid

: is also a well-developed bone. Only the lower limb of

i ht the bifurcated splenial is seen.

VII. l. ant. Fracture at plane of twelfth tooth, Pareiasuchus

a peringueyr. It is of interest to compare this section

Di with the two preceding ones. Although this section is

1) more posterior the prearticular does not show on the

lingual surface. The prearticular is at this plane a well-

) i developed splint bone but is covered by the splenial

i and coronoid.

| VIIt. 1. post. Same fracture as preceding one. Here—only

slightly more posteriorly than in the preceding drawing

—the prearticular has a small lingual surface.

IX. vr. ant. Fracture at plane just behind last tooth, Dolicho-

pareva angusta. The coronoid is seen as it is passing to

the external surface of the jaw, where it is partly

covered by the dentary. The lower limb of the bifur-

cated splenial is seen. The upper limb does not stretch

sofar backward. Externally the surangular is beginning

to show on the surface. Internally it is already well

developed, but is covered by the posterior end of the

dentary.

“we

BoA

’

uN ARS >

vow kw

> ert

+77

.—Cross-sections.

Fic.

272

III

XII.

XIII.

XIV.

XV.

XxGVile

Annals of the South African Museum.

. 7. ant. Fracture just behind the angular boss, Pareza-

suchus peringueyt. The section passes through the

posterior inferior meckelian foramen. The prearticular

lies wholly on the surface. The dentary is cut through

just in front of its posterior termination. The anterior

extension of the articular becomes evident, as does also

its plug-like nature.

r. post. Same fracture as IX. In this section the coronoid

has passed over to the outside. Here it forms the

external border of the supra-meckelian fossa whose

inside margin is formed by the prearticular.

r. ant. Fracture just in front of articular surfaces, Anthodon

serrarvus. The posterior part of the external border

of the supra-meckelian fossa is here formed by the

surangular. The coronoid terminates just in front of

this plane. Another point to notice is how the articular

is being clasped by the angular, prearticular, and

surangular.

r. post. Same fracture as preceding one. Note the

bifurcated superior margin of the surangular. The

prearticular is seen immediately anterior to the plane

where it tapers to a point.

l. post. Fracture through plane of middle of supra-meckelian

fossa, Brachypareia watsonr. Note the small foramen

through the surangular. The section of the prearticular

is noteworthy. Its posterior border shows the thicken-

ing to form the ledge for the insertion of the inner slit

of the adductor muscles.

r. post. Fracture through plane of anterior edge of articular

surface, Parevasuchus peringueyr. Note the bifurcated

nature of the surangular. The groove between the two

limbs forms a posterior continuation of the supra-

meckelian fossa. Notice also how the articular now

plugs the meckelian cavity and how securely it is

clasped by the bones lying anterior to it.

l. ant. Fracture through median plane of articulatory

surface. The angular is still supporting the articular.

The shallowness of the external cotylus and the con-

cavity of the inner one is well shown.

Fig.

S. Ang.

5.—Cross-sections.

274 Annals of the South African Museum.

SYSTEMATIC ACCOUNT.

Although the series of jaws in the South African Museum is a very

large one, considerable difficulty has been encountered in determining

the characters on which a scheme of classification based on the jaws

could be founded. It has been found that species with dissimilar

skulls have very similar mandibles, and again that species of the

same genus have very dissimilar mandibles and approaching more

closely species of other genera. ;

Another difficulty encountered was the fact that unfortunately the

crowns of the mandibular teeth are never very well shown. If the

jaws were open or altogether disarticulated prior to fossilization the

teeth are either very much weathered or are more or less destroyed in

freeing them from the intractable matrix. In those specimens again

where the jaws are closed it is very difficult to expose the mandibular

teeth. The teeth of the upper jaw are, however, similar in structure

to the mandibular teeth, and as these are generally preserved quite

well they may be examined and the conclusions derived from them

may be taken to represent the nature of the mandibular teeth.

Examining the cusps it is found that there are three main types of

Pareiasaurian teeth, viz. :

A. Teeth with less than nine cusps.

B. Teeth with nine cusps.

C. Teeth with more than nine cusps.

On this distinction the Parecasauridae have been classified into three

main groups.

Although the number of teeth varies according to the age of the

animal, there do nevertheless appear to be three groups with a more or

less defined number of teeth, viz. :

a. Forms with few teeth, 11-14. To this group belong all the

forms with numerous cusps except Parevasaurus serridens.

b. Forms with many teeth, 19-20. This group contains two forms

—Dolichopareia angusta ; Bradysaurus seeleyi (and Brady-

saurus bombidens ?). |

c. Forms with an intermediate number of teeth, 15-16. The rest

of the Pareiasauridae belong to this group.

There is yet a further difference in the nature of the dentition.

(a) In some forms the crowns of the teeth overlap each other

antero-posteriorly to a considerable extent, as in Nochele-

Pareiasaurian Studies.—Part V. 275

saurus strubeni, Bradysaurus seeleys and Embrithosaurus

schwarzt. One specimen of Bradysaurus baini, 8.A.M. Cat.,

No. 4347, also has the crowns considerably overlapping.

(b) In other forms there is only a slight overlap. This condition

is seen in Dolichoparera angusta, Brachyparera rogersi and

watsoni, and in all the other specimens of Bradysaurus

bain.

all the forms with more than nine cusps. It includes,

however, also Nochelesaurus alexanderi and Bradysaurus

vanderbylr.

These and other points to be enumerated, although they do not

permit of a good grouping, will nevertheless assist in classifying most

lower jaws.

GROUP A.

(1) Forms with broad massive rami of moderate length. In these

forms also the infero-ventral symphysal protuberance is

well developed; the symphysis is high and strong; the

prearticular is large; the splenial although strong does not

reach high up on to the external symphysal surface; the

arch is U-shaped with the posterior part of the rami diverging.

a. Forms with two angular bosses.

Bradysuchus whaitse.

6. Forms with one strong knob-like boss.

(1) Forms with many teeth, 19-20, and with crowns

considerably overlapping.

Bradysaurus seeleyr (Pl. XXXII, fig. 1).

In this species the anterior part of the arch also

appears more pointed.

(2) Forms with an intermediate number of teeth,

15-16, and with crowns slightly overlapping.

Bradysaurus bana (fig. 6).

(3) Forms with an intermediate number of teeth,

15-16, and with practically no overlap of the

crowns.

Nochelesaurus alexanderz (Pl. XXXII, fig. 2).

Bradysaurus vanderbylv.

VOL. XXVIII, PART 2. ng

“ll na

i ‘ain

| | |

f |

: | |

A —<— a

| |

ei) il

| ii |

Via

|! H

Fic. 6.—Bradysaurus baini. * 4. Ventral view of the mandibular arch. Com-

posite figure of 8.A.M. Cat., Nos. 4347 and 4999.

Ang.=angular. Pr. Art. =prearticular.

Spl. =splenial. Sur. Ang.=surangular.

Fic. 7.—Brachypareia rogersi. X 4. Outer view of right ramus. S.A.M. Cat.,

No. 5012.

Cor.=coronoid. Dent.=dentary. Art.=articular.

Fic. 8.—WNochelesaurus strubeni. x 4. Inner view of left ramus. Composite

figure of S.A.M. Cat., Nos. 5590 and 5019.

276

Parerasaurian Studies.—Part V. Qin

(2) Forms with broad light rami of moderate length. These forms

are further characterised by a moderately strong symphysis ;

a moderate or weak infero-ventral symphysal protuberance,

weak or moderate knob-like angular boss; 15-16 teeth

slightly overlapping.

a. Forms with a weak angular boss and weak splenial

portion of symphysis; the arch is U-shaped with the

posterior part of the rami strongly diverging.

Brachypareva rogersi (fig. 7).

Koalemasaurus acutirostris.

b. Forms with a moderate angular boss and moderate

splenial portion of symphysis; the arch is U-shaped

but with the posterior part of the rami converging.

Brachypareva watsom (Pl. XX XITI).

(3) Forms with broad massive rami of great length. This group is

further characterised by a moderate symphysis; moderate

infero-ventral symphysal protuberance ; well-developed pre-

articular with a very strong ledge; very well differentiated

surface of insertion of the anterior pterygoid muscle; the

arch is U-shaped but with posterior part of rami converging ;

the angular boss 1s teat-like ; 19-20 teeth slightly overlapping.

Dolichopareva angusta (Pl. XXXIV, figs. 1 and 2).

(4) Forms with broad massive short rami; symphysis strong and

high; height over angular boss is enormous; the teat-like

boss is very large; the part of the ramus anterior to the

angular boss is approximately equal to the part posterior,

whereas in all the other species the former is greater than the

latter; the arch is U-shaped with the posterior half of the

rami diverging.

Nochelesaurus strubent (fig. 8).

It must be noted how dissimilar the jaws of the two species

of Nochelesaurus are.

GROUP B.

The arch is not very broad; the rami are of moderate length ;

moderate teat-like angular boss; the prearticular is quite large; the

splenial part of the symphysis is moderately strong; the infero-

| | 278 Annals of the South African Museum.

' ventral symphysal protuberance is well developed; 15-16 teeth

| | considerably overlapping.

il Embrithosaurus schwarz (Pl. XX XV).

ib |

) hit

| GROUP C.

1 | i (1) Forms with a broad, medium-sized arch; prearticular small ;

' I splenial part of symphysis strong with splenial considerably

ay ul exposed on the external symphysal surface ; 14 non-overlap-

| | ping teeth; arch U-shaped with posterior portion of rami

Hi diverging. |

a a. Angular boss knob-like ; symphysis moderately high.

He Parevasuchus peringuey? (figs. 9 and 10; Pl. XXXVI, fig. 2).

3 Spl.

Inf Post Meckelian For.

| Fic. 9.—Paretasuchus peringueyi. x 4. Inner view of left ramus. S.A.M. Cat.,

i No. 2337. Drawing by Miss M. van der Merwe.

Cor. Pric

| Fie. 10.—Pareiasuchus peringueyi. Xx 4. Outer view of leftramus. S.A.M. Cat.,

i) No. 2337.

| 6. Angular boss horn-like and directed inwards; low

| symphysis.

| if Parevasuchus nasicornis (Pl. XXXVI).

(2) Forms with a broad arch of small dimensions; prearticular of

medium size; splenial part of symphysis strong with splenial |

Pareiasaurian Studies.—Part V. 279

Sup Post. Meckelian For.

Post Inf Meckelian For.

Fig. 11.—Anthodon serrarius. 3. Inner view of right ramus. S.A.M. Cat.,

No. 4020. Drawing by Miss M. van der Merwe.

Cor. Dent.

Fie. 12.—Anthodon serrarius. Xx 4. Outer view of right ramus. S.A.M. Cat.,

No. 4020.

Fig. 13.—Anthodon serrarius. xX 4. Ventral view of mandibular arch. Posterior

part of left ramus restored. S.A.M. Cat., No. 4020. Drawing by

Miss M. van der Merwe.

280 Annals of the South African Museum.

considerably exposed on the outer symphysal surface; 11

non-overlapping teeth; angular boss a weak pyramidal

swelling ; arch U-shaped with posterior ends of rami parallel.

Anthodon serrarvus (figs. 11, 12, 13).

(3) Forms with two angular bosses; rami of moderate size; pre-

articular of medium size; symphysis of moderate strength,

splenial with a not very great share in symphysis; 14 non-

overlapping teeth.

Propappus omocratus (fig. 14).

Fic. 14.—Propappus omocratus. x +. Inner view of right ramus. S.A.M.

Cat., No. 1058.

(4) Forms with rather pointed, massive arch; symphysis high and

strong ; 15-16 non-overlapping teeth. :

Pareiasaurus serridens.

The jaws of Bradysaurus bombidens are imperfectly known.

The jaws of Platyoropha broom: and Propappus parvus are

altogether unknown.

Paretasaurian Studies.—Part V. 281

Measurements.

Height Height Width Width

Animal. Length. R aE eet over wee

ee neuer | Articular. Sang la

physis. Boss. Boss.

Bradysaurus

bain a Vier 95 LTO 320 325

Bradysaurus

seeley2 340 65 170 205 240

Bradysaurus

vanderbylr 360 92 130 267 265

Bradysuchus

whartsi

Nochelesaurus

strubent 300 103 185 257 248

Nochelesaurus

alexander 345 98 7S 350 320

Dolichopareva

angusta O47 82 162 300 315

Koalemosaurus

acutirostris

Brachypareva

rogerst 288 70 95 (?) 280 255

Brachypareva

watsons 326 78 134 260 270

Embrithosaurus

schwarzr 335 85 145 270 250

Anthodon

serrarvus 175 35D (5 165 165

Parevasuchus

peringueyr 265 60 — 110 290 260

Parevasuchus

NASICOYNUS 225 50 90 265 245

Parevasuchus

serridens 285 15 128 240 200

Propappus

omocratus as 55 125

MUSCULATURE.

The adductor or temporal group of muscles close the jaw. In

the Pareiasauridae, with their massive mandibles, these muscles

must have been large and strong with firm points of attachment.

These muscles are inserted in the supra-meckelian fossa and on the

ULE eee

~ ee) ee ee

282 Annals of the South African Museum.

surrounding bones. The supra-meckelian fossa is not as large as in

other Cotylosaurs. The presence of a moderately developed coronoid

process and the presence of a well-differentiated ledge on the pre-

articular, however, provide very good surfaces for muscular insertion.

It appears to be a general rule among Cotylosaurs that a large well-

developed supra-meckelian fossa is accompanied by a weak coronoid

process and wice versa. In the durophagous Diadectes with its very

heavy jaw both are well developed.

Adductor Group (fig. 15).

Fig. 15.—Pareiasuchus peringueyi. Xx +. Inner view of left ramus, showing

probable muscular connections. S.A.M. Cat., No. 2337.

D.M. =depressor mandibulae.

P.A. =pterygoideus anterior.

C.M.P. =capiti-mandibularis profundus.

C.M.M.=capiti-mandibularis medius.

C.M.S. =capiti-mandibularis superficialis.

1. Capiti-mandibularis superficialis. This slip of the adductor

muscle lies, as its name implies, most external. Its chief

surface of insertion was probably on the internal surface of

the coronoid process, while some of its fibres probably also

entered the supra-meckelian fossa.

2. Capiti-mandibularis medius. This median slip was probably

wholly inserted on the internal surfaces of the bones sur-

rounding the supra-meckelian fossa.

3. Capiti-mandibularis profundus. This most internal slip had

its chief point of insertion on the ledge developed on the

prearticular. Some fibres probably also entered the supra-

meckelian fossa.

The origin of this tripartite muscle was undoubtedly from the inner

surface of some of the bones forming the massive stegocrotaphic

Pareiasaurian Studies.—Part V. 283

“cheek”? or temporal region. The area of origin was probably

confined to the anterior portion of the squamosal, the lower portion

of the postorbital and the jugal. The muscle passed through the

fenestra basitemporalis.

Another muscle belonging to the adductor group was the ptery-

goideus anterior. This muscle probably had its origin from the lateral

arm of the triradiate pterygoid and possibly also from the posterior

edge of the ectopterygoid. Its surface of insertion is very well marked

in a number of jaws, where there is a well-marked groove on the

internal ventral border of the jaw immediately behind the angular

boss. The insertion is therefore partly on the prearticular and partly

on the angular.

Depressor Group. |

The jaw was opened by the contraction of the depressor mandibulae

(digastric). The origin of this muscle must have been from the more

or less rugose posterior border of the squamosal. The area of insertion

could not have extended higher up than the otic notch, because in that

case it would have interfered with the auditory apparatus.

As the retro-articular process of the jaw is weak, only a small

portion of the fibres of the depressor muscle were probably attached

to it. In Hryops this process is also weak, and Adams (11) is of the

opinion that the major portion of the depressor muscle was attached

to the ventral surface of the articular. In the Parevasauridae a more

localised area of insertion has been developed. There does not seem

to be any doubt but that the angular boss had as its main, if not sole,

function the reception of the muscles opening the jaw. The presence

of a subsidiary angular boss in Bradysuchus whaits: and Propappus

omocratus is functionally not easily accounted for. It does not seem

possible that any of the fibres of the depressor mandibulae were attached

to it. It may be possible that this growth is a sign of super-activity

of an area where osteogenesis must undoubtedly be vigorous in order

to produce the normal angular boss.

In connection with the mechanism of the mandible there are some

interesting points correlated with the difference in form of the jaws

of the different species.

In the long-jawed forms (e.g. Dolichopareia) gravity would assist

materially in the opening of the jaw, so that it is not surprising to note

that neither the retro-articular process nor the angular boss is appreci-

ably larger than in the medium- and short-jawed forms. In the

long-jawed forms a greater force is necessary to close the jaw, and we

note that better provision is made for the insertion of the adductor

284 Annals of the South African Museum.

muscles. In Dolichoparera angusta, for instance, the prearticular

ledge is strong, the coronoid process is robust, and the depression

and ridge on the inner ventral surface (for the insertion of the

anterior pterygoid) are better differentiated than in any other jaw

examined.

In the short-jawed forms (e.g. Nochelesaurus strubenr) gravity

would play a lesser role. The depressor muscle would therefore have

to be strong. This supposition is verified by examining the large,

strongly developed angular boss.

In connection with the origin of the depressor muscles an interesting

deduction is possible. In the imperfectly adapted Pareiasaurs the

muscular attachments would not have been perfectly developed. The

Pareiasaurs from the lower geological horizon must be considered as

approaching the primitive condition most closely. In these forms the

posterior border of the squamosal is fairly smooth. Consequently

the depressor muscles were not strongly attached at their origin. In

later forms the attachment would have improved. That this was the

case 1s shown by the fact that the later forms have a more rugose

posterior border of the cheek. Of the South African forms, Pareza-

suchus perrngueyr from the Cistecephalus zone has developed the most

highly rugose cheek.

This rugosity of the cheek is seen to a more pronounced degree in the

Pareiasaurs from North Russia which, on palaeontological grounds,

are considered to be of Cistecephalus zone age; and the most bizarre

development within the family is displayed in the Scottish Triassic

genus Elginia.

It will be very interesting to examine the jaws of the Russian forms

in comparison, and we eagerly await the description of these forms

which is to appear shortly.

COMPARATIVE AND EVOLUTIONARY (fig. 16).

Recently it has become more and more apparent that the ancestors

of the Amphibia must be sought amongst the primitive Devonian

representatives of a family of Crossopteryian fishes—the Osteolepidae.

The structure of the Osteolepid lower jaw has been worked out by

Watson (12) in the case of Megalichthys. The jaw is composed of

nine or possibly ten separate bones, viz.: articular, angular, surangular,

prearticular, two splenials, dentary,and two or possibly three coronoids.

The following points are of interest :—

1. The prearticular is a large bone. Posteriorly it supports the

TIA Pr Art. — ; eo)

Fic. 16.—Series of figures to show some advances in the structure of the lower jaw.

I. Megalichthys sp. x 4. After Watson.

II. Orthosaurus pachycephalus. x 4. After Watson.

III. Conodectes (=Seymouria baylorensis). x 1. After Watson.

IV. Bradysaurus bombidens?. x 4. After Broom.

Art. =articular. Dent. =dentary.

Ang. =angular. Cor. =coronoid.

Spl. =splenial. Sur. Ang.=surangular.

Pt. Spl.=postsplenial. Pr. Art. =prearticular.

285

286 Annals of the South African Museum.

articular and then stretches right up to the anterior end of the jaw,

where it meets its fellow and forms the lower half of the symphysis.

2. The splenials are practically confined to the external surface of

the jaw.

3. The angular is restricted to the external surface, and is not

pierced by foramina. .

4. The two (or possibly three) coronoids are dentigerous and stretch

far forward.

5. There is a moderate retro-articular process.

Recent research has, furthermore, tended to show that the Carbon-

iferous embolomerous Amphibia must be looked upon as the group

which contained the ancestors of the Reptiles.

The best known lower jaw of the Embolomerz is that of Orthosaurus

pachycephalus (12).

The following points in its structure are of interest :—

(1) The prearticular is still a large bone as in Megalichthys. It does

not stretch so far forward and does not enter the symphysis.

(2) The splenials show an advance on those of Megalichthys in that

they form part of the inner surface of the jaw. In fact the splenial,

although it hardly enters the symphysis, occupies much of the anterior

portion of the jaw, which in Megalichthys is formed by the prearticular.

(3) The angular is more visible in lingual view than is the case in

Megalichthys and is perforated by foramina.

(4) There are three coronoids which in most Temnospondyls are

dentigerous.

(5) The retro-articular process is weak.

It has been generally accepted that Conodectes (=Seymourta) is the

most primitive of Reptiles (13). It stands, in fact, on the border-line

between Amphibia and Reptiles. It is therefore not surprising to

note that its lower jaw contains ten elements as in some of the

ancestral Hmbolomeri and possibly also the Osteolepidae.

For our comparison the following points are interesting :—

(1) The prearticular is still further reduced in size.

(2) The splenials occupy a great part of the anterior lingual surface.

The splenial just commenced to take a share in the formation of the

symphysis in Orthosaurus. In Conodectes it forms a large part of the

symphysis.

(3) The angular occupies a greater part of the lingual surface than it

does in Megalichthys and Orthosaurus. It is perforated by a single

foramen.

(4) There are three dentigerous coronoids. The third coronoid has

Parevasaurian Studies.—Part V. 287

its posterior part slightly elevated to form the rudiment of a coronoid

process.

(5) The postarticular process is weak.

All the other Cotylosaurs show advances on the primitive reptilian

condition as shown by Conodectes.

The postsplenial is absent in all other Cotylosaurs except Pantylus

(14).

The first and second coronoids are lost in all higher reptiles, but may

still be present in Pantylus. The dentigerous nature has also a

tendency to disappear.

The other bones are constant in Cotylosaurs.

It has already been noted that in this series of ascending forms the

splenial commenced as a bone of the external surface, and that it

subsequently gradually migrated to the inner surface. This tendency

has continued within the group of Cotylosaurs.

The splenial rises nearly up to the lingual alveolar border in

Captorhinus, Diadectes, Labidosaurus, Procolophon (15) and all the

Parevasauridae.

In Limnoscelis it rises right up to the dentigerous border.

This upward lingual tendency of the splenial is accompanied by a

bifurcation of the posterior end of the bone. The initial stage of this

bifurcation is seen in Captorhinus, Diadectes, and Pantylus. It is more

marked in Labidosaurus, and reaches its greatest development in

LInmnoscelis and in the Pareiasauridae.

Primitively (in Megalichthys) the splenial had no share in the sym-

physis. The stages of the acquirement of such a share are seen

in Orthosaurus and in Conodectes until it became a character of the

Cotylosaur mandible. After having acquired this share in the

formation of the symphysis the splenial may, within the group of

Cotylosaurs, continue along this line of advance. In fact, in the

Pareiasauridae the splenial forms a very strong ventral half of the

symphysis. So that although the initial tendency of the splenial was

to migrate to the lingual surface of the jaw, it again begins to extend

on the outer surface. Its external exposure is, however, confined to

the symphysal region. Within the group of South African Pareiasaurs

a continuation of this development can be noticed. In the younger

forms Pareiasuchus and Anthodon for instance, the splenial occupies

more of the external symphysal surface than it does in Bradysaurus

bani, and this form in its turn shows more of the splenial than does

Brachypareia rogersi and B. watsoni, Dolichopareia and Nochelesaurus.

The absence of a coronoid process must be looked upon as a primitive

288 Annals of the South African Museum.

character. It is not present in the Osteolepidae and Embolomeri, and

is only slightly developed in Conodectes. In some of the heavy-jawed

Cotylosauria the coronoid process is also weak, viz. in Labidosaurus,

Pantylus, Captorhinus, LInmnoscelis. In these forms the lack of

development of the coronoid process is accompanied by the presence

of a large supra-meckelian fossa. In these forms the adductor

muscles were inserted mainly on the inner surface of the bones bound-

ing the fossa. Consequently the coronoid process did not function as

a point for muscular insertion and did not therefore become robust.

With the development of a strong coronoid process these muscles had

a good point of insertion. The supra-meckelian cavity no longer

needed to be large. The development of the ledge on the prearticular

would have a similar effect. Forms with a small supra-meckelian

cavity are :—Procolophon and all the Pareiasauridae. The duro-

phagous Diadectes (16) had both a large supra-meckelian fossa and a

robust coronoid process.

The prearticular decreased in size through the series—Osteolepidae,

Embolomert, and Cotylosauria. Within the group of South African

Parevasauridae a continuation of this reduction is noticeable. The

Cistecephalus zone form, Parevasuchus peringueyt, has a prearticular

with an exposed surface much smaller than have the forms from the

Tapinocephalus zone, e.g. Bradysaurus bain.

ACKNOWLEDGMENTS.

Our thanks are due to Dr. R. Broom for allowing us to examine

some of his material ; and to Miss M. van der Merwe for the drawings

of Parevasuchus and Anthodon.

Pareiasaurian Studies.—Part V. 289

WORKS CITED.

(1) OwENn.—Cat. Foss. Rept. 8. Afr., 1876.

(2) SpELEY.—Phil. Trans. Roy. Soc., B, vol. 179, 1888.

(3) SppLEy—Phil. Trans. Roy. Soc., B, vol. 183, 1892.

(4) Broom.—Ann. 8. Afr. Mus., vol. iv, pt. vii, 1908.

(5) Watson.—Ann. Mag. Nat. Hist., ser. 8, vol. x, 1912.

(6) Broom.—Rec. Albany Mus., vol. ii, 1913.

(7) Broom.—Ann. 8. Afr. Mus., vol. xii, pt. i, 1913.

(8) Broom.—Am. Mus. Journ., xiv, 1914.

(9) Broom.—Proe. Zool. Soc., vol. i, 1924.

(10) Casze.—Revision of the Cotylosauria of North America, 1911.

(10) Witiiston.—The Osteology of the Reptiles, 1919.

(11) Apams.—Ann. New York Acad. Sci., vol. xxviii, 1919.

(12) Watson.—Phil. Trans. Roy. Soc., B, vol. 214, 1926.

(13) Watson.—Proce. Zool. Soce., vol. 11, 1918.

(14) Menzt.—Journ. Geol., vol. xx, 1912.

(15) Watson.—Proce. Zool. Soc., vol. ii, 1914.

(15) Huene.—Centrbl. Min., 1920.

(16) Casz.—Journ. Geol., vol. xv, 1907,

(16) Casz.—Journ. Geol., vol. xiii, 1905.

EXPLANATION OF PLATES.

PLATE XXXII.

Fig. 1. Bradysaurus seeleyi. Xx 7. Ventral view of Arch. S.A.M. Cat., No. 5624.

» 2. Nochelesaurus alexanderi. X 35. Ventral view of mandibular arch.

S.A.M. Cat., No. 8944.

Puate XXXII.

Fig. 1. Brachypareia watsoni. X 4. Dorsal view of mandibular arch. 8.A.M.

Cat., No. 6240.

» 2. Pareiasuchus peringueyt. X 4. Ventral view of mandibular arch. Mirror

image. §S.A.M. Cat., No. 2337.

PLATE XXXIV.

Fig. 1. Dolichopareia angusta. x %. Inner view of right ramus. S.A.M. Cat.,

No. 62388.

», 2. Dolichopareia angusta. Outer view of the preceding jaw.

», 93 Dolichopareia angusta. X< 4. Outer view of left ramus. S.A.M. Cat.,

No. 3717.

PLateE XXXV.

Fig. 1. Embrithosaurus schwarz. X< 4. Inner view of right ramus. S.A.M.

Cat., No. 8034.

» 2. Embrithosaurus schwarzi. Outer view of the preceding jaw.

PLATE XXXVI.

Pareiasuchus nasicornis. X 7. Ventral view of mandibular arch. S.A.M. Cat ,

No. 3016.

Plate XXXII.

| en. Senne Vus.,, Vol. XXVIII.

;

Fic

Fig

cs)

nn. §. Afr. Mus., Vol. XXVIII.

| Plate XXXII.

}

Fic. L.

Sur. Ang.

Ang. Prot.

Teil & Co., Lid.

Me Mus. Vol. XXVIIL Plate XXXIV.

Wine, Il,

iG, B.

Neill & Co., Lid.

Gr. Mos., Vol. XXVIII. Plate 5

Fie. 1.

Nie, Be

Neill & Co., Ltd.

nn. S. Afr. Mus., Vol. XXVIII. Plate XXXVI.

Nel & Co., Lid.

( 291 )

8. On a Foraminiferal Limestone of Upper Eocene Age from the

Alexandria Formation, South Africa.—By FrepERick CHAPMAN,

A.L.S., F.G.8., F.R.MLS., etc. (Palaeontologist to the Australian

Commonwealth).

(With Plate XX XVII.)

IntTRopDucTOoRY NOTE.

At the beginning of July 1925 Dr. Sidney H. Haughton, B.A., F.G.S.,

kindly sent me a series of four rock slices of limestone from the

Alexandria Formation of the Bathurst Division of the Cape, together

with rock specimens from four localities in the same area.

A further consignment of six micro-slides were sent on December

14th, 1925, which had been especially prepared at my request so as

to show more satisfactorily the equatorial chambers of the large

discoidal foraminifera, which in the first set of slices were seen to be

packed together in the plane at right angles to the section.

The rocks under examination show beautifully preserved and

abundant tests of Discocyclina pratii and D. varians, together with

other and smaller foraminifera that are seen scattered throughout

the matrix.

DESCRIPTION OF THE Rock MATERIAL.

Locality.—Birbury, Bathurst, South Africa.

Matrix of rock consists of a mosaic of calcitic grains, often finely

granulated or full of inclusions. Glauconite grains form a small

proportion of the ground mass, whilst the embryonic chambers as

well as those of the equatorial series of the Discocycline foraminifera

are also frequently filled with the same apple-green mineral.

The organic remains seen in the rock specimens form about 50 per

cent. of the whole, and represent Foraminifera (Dzscocyclina, etc.),

Polyzoa (apparently Cyclostomes), Pelecypod shell-fragments, sections

of Kchinoid plates and spines, and numerous glauconite grains repre-

senting the infillings of Ostracods and Foraminifera.

The glauconite varies in colour from pale to olive green. It has

invaded the zodidal cells of polyzoa as well as the minute chamberlets

of the foraminifera.

The smaller foraminifera sometimes have the chambers empty or

nearly so, but occasionally are more or less filled with minute calcitic

VOL. XXVIII, PART 2. 20

! i 292 Annals of the South African Museum.

: i}

| crystals, and these have continued growing beyond the surface of

4 the test, like a fringe around it.

iin

Tah

j hi CONTENTS OF THE Rock SECTIONS.

ll: Slide 4350. : Slide 4459.

iii Quinqueloculina subrotunda (Mon- | Triloculina trigonula (Lam.).

| ila tagu). Cf. 7 rilina sp.

ie Sigmoilina sp. Sprroloculina sp.

\ hij, Gaudryina cf. bradyi Cushman. Gaudryina sp.

ith Polymorphina cf. gibba d’ Orb. Anomalina ammonordes (Reuss).

ab Globigerina conglobata Brady. Nummulites sp.

1h Discorbis globularis (d’Orb.). Discocyclina pratt. (Mich.).

/ i) | Anomalina ammonoides (Reuss). = varvans (Kauf.).

lk 2 Operculina sp.

i Discocyclina pratti (Michelin). Slide 2260

Globigerina conglobata Brady.

Seon: Rotalia calear a’ Orb.

| Polyst . ¢Cratt

Textularia agglutinans d’Orb. ae of. or

| Globigerina conglobata Brady. Se

Discocylina pratti (Mich.).

Discocyclina pratti (Mich.).

Slide 4461.

ei Slide 4352. Textularia gramen dOrb.

j bh Anomalina ammonoides (Reuss). Gaudryina sp.

aS Discocyclina pratti (Mich.). Anomalina ammonoides (Reuss).

hy Cibicides lobatulus (W. and J.).

Slide 4353. Discocyclina pratti (Mich.).

i i! | Sa ay: be varians (Kauf.).

Wit Sigmoilina sp.

: , | Globigerina bulloides d’Orb. Slide 4462.

of Cibicides lobatulus (Walker and

Fabularia sp.

Textularia gramen da’ Orb.

Anomalina ammonordes (Reuss).

Discocyclina pratt. (Mich.).

oat Jacob).

ie Discocyclina pratti (Mich.).

Slide 4458.

) | Bulimina pupoides d’ Orb. Slide 4463.

i Cristellaria crepidula (Fichtel and | Gaudryina sp.

: fl | | Moll). Polymorphina lactea (W. and J.).

ai Globigerina conglobata Brady. Anomalina ammonoides (Reuss).

Discocyclina prattr (Mich.). Discocyclina pratti (Mich.).

varians (Kauf.). a varians (Kauf.).

On a Foraminiferal Limestone of Upper Eocene Age. 293

NoTES ON THE SPECIES RECOGNISED IN THE ROCK-SLICES.

(Classification according to H. B. Brady.)

Fam. MILIOLIDAE.

Sptroloculina sp.

The specimen here recognised in the thin section (4459) is cut

exactly in the median plane. By the long elliptical form and thickened

shell-wall it appears to belong to S. limbata Brady.

Triloculina trigonula (Lam.).

One typical example of this, or a related form, occurs in slide 4459.

Quinqueloculina subrotunda (Montagu).

A small but fairly typical specimen of this species is found in

slide 4350. Although the test is normally of the brownish translucent

character under a low power, it is seen, under a higher magnification,

_ to have a finely granular structure, though not tubulose.

Sigmoilina sp.

Specimens of this genus occur in slides 4350 and 4353. They do

not afford any particular characters for specific determination.

Cf. Trallona sp.

Some sections of a milioline test are found in slide 4459, which in

some respects resemble that of the genus Trillona, having a triloculina

form with cavernous shell-wall.

Fabularia sp.

A remarkably fine section of the test of a Fabularia is seen in

slide 4462. It may possibly be identical with F. discolithes. This

example shows a distinct megaspheric proloculum and a perforated

external shell-wall. 7

Fam. TEXTULARIIDAE.

Textularia agglutinans d’Orb.

A very fine example, cut parallel with the perpendicular axis

occurs on slide 4351, and there are other fragmentary examples in

the same section.

294 Annals of the South African Museum.

Textularia gramen Wd Orb.

Several examples, more or less incomplete, occur in slides 4461

and 4462. The species is here distinguishable from 7. agglutinans by

the coarser test and more rapidly widening chambers.

Gaudryina ef. bradyr Cushman.

A species of Gaudryina occurs here, in slides 4350, 4459, 4461,

and 4463. The sections are not complete enough to determine the

species, but they resemble G. bradyi rather than G. pupoides d’Orb.

Bulimina pupordes d’ Orb.

A single example of this species, with a megaspheric proloculum,

occurs on slide 4458.

Fam. LAGENIDAE.

Cristellarva crepidula (F. and M.).

A single example occurs on slide 4458, which has a rather thick-

walled test, although the general form is typical of the species.

Polymorphina lactea (W. and J.).

A broadly oval test occurs in slide 4463.

Polymorphina cf. gibba d’Orb.

The example recognised in slide 4350 is a large specimen of sub-

globular form and with the sutures almost flush ; the walls are fairly

thick. It is a large and well-grown form of P. gibba.

Fam. GLOBIGERINIDAE.

Globigerina bulloides d’Orb.

A fairly large specimen of this form occurs in slide 4353. The

test is somewhat thicker than is usual in this species.

Globigerina conglobata Brady.

Several examples of this species occur in slides 4350, 4351, 4458,

and 4460. They closely resemble living examples in the thickness

and spinose surface of the test.

On a Foraniniferal Limestone of Upper Eocene Age. 295

Fam. ROTALIIDAE.

Discorbis globularis (d’Orb.).

Several specimens of this species are seen in slide 4350.

Anomalina ammonoides (Reuss).

This species is by far the commonest rotaline form in these rock

sections. They are seen in every position, and the tests of some are

filled with glauconite. The slides in which this form occurs are

4350, 4352, 4459, 4461, 4462, and 4463.

Cibicides lobatulus (W. and J.).

Several typical examples are seen in slides 4353 and 4461. This

_ species is distinguished from that of Anomalina in having fewer

chambers in the whorl and in the superior face of the test being plane.

Rotalia calcar d’Orb.

This comparatively shallow water form is represented by several

examples in slide 4460.

Fam. NUMMULITIDAE.

Polystomella cf. craticulata (F. and M.).

Two specimens occur in slide 4460, one in median section.

2 Operculina sp.

A single transverse section of an operculine form occurs in slide 4350.

It may, however, be equally referable to a nummulite of the J.

planulata type, though of fewer whorls.

Nummulites sp.

Two sections occur in slide 4459, reminding one of N. planulata.

One section is at right angles to the other. They may belong either

to NV. planulata or N. variolaria.

296 Annals of the South African Museum.

Fam. ORBITOIDIDAE.

Discocyclina prattr (Mich.).

This is a most striking and abundant species in the present series.

It occurs in each of the rock slices, and the structure of the test is

most beautifully preserved.

It is seen that the glauconite has infilled the empty chambers of

the tests in many cases, and the invasion of the glauconite seems to

have started by the infilling of the comparatively large initial chambers,

including the proloculum.

Discocyclina pratit is a well-recognised Bartonian species (Upper

Kocene in Kurope).

Discocyclona varians (Kauf.).

This is distinguished from D. prattt in having the central part of

the disc thickened. The sides of the vertical sections of the test are

seen to slope rapidly away to the periphery. This species occurs

in slides 4458, 4459, 4461, 4462, and 4463. It is not nearly so common

as D. prattv. The occurrence of this species seems here to indicate

a slightly lower horizon than Upper Hocene (Upper Lutetian).

EXPLANATION OF PLATE.

Fic. 1.—Discocyclina pratti (Michelin). Vertical section of test. Slide 4458.

x 16.

», 2.—D. pratti (Mich.). Portion of test, showing the chamberlets of the equa-

torial zone. Slide 4460. x 16.

Ann. S. Afr. Mus., Vol. XXVIII. Plate XX XVII.

Discocyclina pratti (Mich.). Upper Eocene, Birbury, Bathurst,

South Africa. x 16.

F.C. photo. Neill & Co., Ltd.

=n Mes

“TRUSTEES OF oy S SOUTH AFRICAN MUSEUM = ee

“NEILL. AND ae uD, 212 CAUSEWAYSIDE, EDINBURGH. ee.

( 297 )

9. Pareiasaurian Studies.

Part VI.—The Osteology and Myology of the Locomotor Apparatus.

A.—Hind Limb.—By 8S. H. Havueutron, B.A., D.Sc., Hon.

Curator, Palaeontological Collections, and L. D. Boonstra,

M.Sc., Assistant in Palaeontology.

(With Plate XX XVIII and Text-figures 1-60.)

HISTORICAL.

In 1856 Owen (23) described and figured (pl. xxxiu, figs, 4-7,

and pl. xxxiv, fig. 1) parts of the pelvis and sacrum of Propappus

omocratus as that of Dicynodon tigriceps.

In 1862 the inner view of the left innominate of Propappus omocratus

was again figured (24) (pl. xxiv, fig. 4) as that of Dicynodon trgriceps.

Also figured (pl. xxiii, figs. 1, 2, and pl. xxiv, fig. 3) as Dicynodon

tagriceps were parts of the pelvis and sacrum of the type specimen

of Parevasaurus serridens.

In 1876 Owen (25) again figured (pl. xxxvi, figs. 2-4, and pl.

Xxxvi, fig. 2) parts of the pelvis and sacrum of Propappus omo-

cratus as Dicynodon trgriceps. The left innominate of Propappus

omocratus was also figured (pl. xxvii, figs. 1, 2) as that of Dicynodon

leoniceps. Parts of the pelvis and the sacrum of the type specimen

of Parevasaurus serridens were here again figured (pl. xxxvi, fig. 1,

and pl. xxxvil, fig. 1) as that of Dicynodon tigriceps.

In 1888 Seeley (27) gave (text-fig., p. 107) an outline drawing of the

pelvis of Parevasaurus serridens as that of Dicynodon tigriceps, and

also (text-fig., p. 103) figured the pelvis of Propappus omocratus as

that of Dicynodon leoniceps. In the same paper Seeley also described

and figured (pl. xu, figs. 1, 2, and pl. xix, figs. 1, 2) parts of the

sacrum and pelvis of Bradysaurus seeleyi as that of Parevasaurus

bombidens (Palmiet Fontein specimen). In the same year he (5)

described as a humerus the femur of Propappus omocratus.

In 1890 Lydekker figured (21) the pelvis (text-fig. 26, p. 120)

previously described as that of Dzcynodon tigriceps and also as

D. leoniceps, and correctly referred it to Propappus omocratus. He,

however, accepted Seeley’s identification of the femur as a humerus

of Propappus omocratus.

VOL. XXVIII, PART 3. 21

298 Annals of the South African Museum.

In 1892 Seeley (29) figured (pl. xvi, pl. xxi, pl. xxu, figs. 1, 2,

and text-figs. 8, 9, 12, and 13) and described the sacrum and pelvis, a

crushed femur, tibia, fibula, and incomplete foot of Bradysaurus baina

as Pareiasaurus baini. Seeley now figured (text-figs. 8, 9) the femur

of Propappus omocratus, and from a comparison with the material

of Bradysaurus baini correctly identified it as a femur, but proposed

the name Pareiasaurus (Propappus) minor. He ignored Lydekker’s

identification of the pelvis of Propappus omocratus.

In 1903 Broom (5) described and figured (pl. xv, figs. 12-22,

pl. xvi) the pelvic girdle, femur, tibia, fibula, and part of the foot

of Embrithosaurus schwarzi in error as Pareiasaurus serridens.

In 1908 Broom (6) gave a short account of the pelvis of a specimen

referred to Propappus omocratus. Of this specimen he also described

and figured (pl. xlv, figs. 2-5) the femur, tibia, and part of the foot.

In 1912 Broom (7) figured (pl. xx, figs. 14-21) and described the

pelvis, femur, tibia, fibula, and the proximal tarsal element of Pro-

pappus rogerst. He also described and figured (pl. xxi, figs. 26, 27)

the pose of the limb.

In the following year Broom (8) described and figured (text-fig.,

p. 355) an incomplete foot of a specimen referred to Bradysaurus baw.

In the same year Broom and Haughton (10) described and gave

photographs (pl. iv, figs. 1-4, and pl. v) of the pelvis, sacrum, and

femur of Pareiasuchus peringueyt, and Haughton (16) described and

photographed the pelvis of Propappus parvus.

In his classification of the Pareiasaurian genera Watson (31) used

some pelvic features for his definitions. He also drew attention to

the parts of the skeletons of the types of Propappus omocratus and

Parevasaurus serridens, which had previously been associated with

Dicynodon.

In 1922 Romer (26) described and figured (pl. xlvi) the probable

muscle attachments to the femur of Propappus.

In 1929 Hartmann- Weinberg (15) described and figured (Abb. 14-20)

the tarsus of a number of Russian and South African Pareiasaurs.

In the same year Boonstra (8) described and figured (pls. ix, x,

x1, and xu, and text-fig. 29) the foot of a number of South African

species.

GENERAL OSTEOLOGY.

Pelvis (text-figs.1—4, 7-11, 15-17, 21, 25, 26, 30, 32, 33, 38, 39, 43, 47—

50, 52, 53).—The Pareiasaurian pelvis consists of at least three pairs of

bones. In all the species these bones are strong ; in some, however,

Pareiasaurian Studies.—Part VI. 299

they are, in addition, very massive, whereas in others they must, in

comparison, be called “ light.”” These three bones—the ilium, ischium,

and pubis—are very firmly united to each other in ankylosis to form

an osinnominatum. The two innominate bones are hardly less firmly

united to each other than are their constituent elements. The sym-

physis, which is thickened in all the forms and ventrally forms a strong

carina, 1s very markedly thickened in the forms with narrow pelvis

(text-fig. 1). Dorsally the girdle is completed by a number of sacral

ribs, which attach the girdle to the vertebral

column. To withstand the strain of this attach-

ment a closer union between the vertebrae con-

cerned is brought about, and it is found that

in some forms two, and in others three or four,

vertebrae are ankylosed to form a sacrum. The

sacral ribs are not all equally developed. In ge ae

some species the strongest rib is the first, whilst “pS

in others it is the second. This would indicate T®xt-Fic. 1.—Longitudi-

; ; nal sections through the

that different vertebrae are concerned, as 1t symphysis. x 4h.

must be taken that the strongest ribis attached (a) Bradysaurus baini

to the primitive single sacral vertebra. In those (2933). _(b) Himbritho-

ie saurus schwarz (8034).

forms where the first sacral rib is the strongest,

its vertebra is taken to represent the first sacral ; where the second is

the strongest, the first rib belongs to a lumbo-sacral vertebra ; where

the fourth sacral rib is attached to a vertebra not ankylosed to its pre-

decessor, this is taken to represent a caudo-sacral vertebra. The third

rib is always attached to a vertebra ankylosed to its predecessor, so

that it is taken to be a true sacral vertebra which has gone through

a caudo-sacral stage. In all the Pareiasaurs the main sacral attach-

ment of the ilium lies, asin Mammals, anterior to the plate of the pelvis,

and is, therefore, pre-acetabular; in the Mammals the sacral ribs

abut on the postero-internal surface of the ilium, in the Pareiasaurs

on the anterior and medial portions of the internal surface of the ilum.

In shape the pelves vary. In some forms the internal outline is

in anterior view broadly U-shaped, with a broad, flattened, dorsal

surface to the floor, and the sides meeting either at a very or moderately

obtuse angle; in others it is V-shaped with a narrow floor, with its

sides meeting at a fairly acute angle. The narrow pelves are much

higher than most of the broad ones, because in them the medial

portions of the plate form part of the sides. In absolute height some

of the broad pelves nearly approach the dimensions of the narrow ones.

This is, however, due to the general greatness, but proportionally

300 Annals of the South African Museum.

they are low. Furthermore, although some forms have upstanding

and others inclined iliac blades, this hardly affects the total pelvic

height.

Ilium.—The ilia vary greatly in size and shape. Some of these

varieties were used by Watson (31) in defining the Pareiasaurian

genera known to him. With reference to the vertical plane passing

through the centre of the acetabulum, the ilium is inclined at a smaller

or greater angle. In the former case we shall speak of an upstanding

or vertical ilium, and in the latter of a forwardly directed iium. The

dorso-ventral measurement of the iliac blade may be small or great,

and we thus distinguish low or high blades; similarly, some blades

are long and others short. In those forms with forwardly directed

blades it is found that the iliac shaft is either long or short, whereas

in those with upstanding ilia the shaft is invariably long. This shaft

may be more or less constricted—a character concomitant with either

upstanding or inclined ilia. Posteriorly the iha may have weakly

or strongly developed processes—a feature not connected with the

forward inclination or upstanding nature of the ilium. The antero-

ventral border of the iliac blade varies greatly ; so does the external

surface. The external surface may be flat or concave in antero-

posterior direction, and the antero-ventral edge may be turned out-

wards to some degree, or it may not be everted at all. This eversion

of the edge must not be confused with another variation of the ihum ;

in postero-anterior direction the whole blade of the ilium may be

diverging with reference to the medial vertical plane—a feature which

reaches its height in Mammals. This divergence seems to be restricted

to the forms with forwardly

directed iia. Many of these

forms have broad pelves, and

the width between the two

ilia therefore becomes very

great. Above the acetabu-

lum the iium forms a well-

developed ridge or buttress.

The internal surface of the

iliac blade is divided into two

TEext-F1G. 2.—Internal view of pelvis, showing distinct articulatory surfaces

ph attachments of he sacral nbs, EM by a dorso-ventraly directed

ridge (text-fig. 2). Anterior

to this ridge a slightly excavated surface articulates with either

the first large sacral rib or in other forms with the first and

Pareiasaurian Studies.—Part VI. 301

second ribs, 2.e. a lumbo-sacral and the first true sacral rib. Pos-

teriorly to the ridge a shallow excavated surface articulates with the

third and fourth ribs or with the second, third, and fourth. The

anterior articulation is a very firm one, and may extend down the

shaft of the iliium to opposite the external supra-acetabular buttress.

In dorsal view the upper edge of the iliac blade is triangular in

shape—the internal ridge forming the apex (text-fig. 3). In those

forms with an everted anterior edge the apical angle is more acute

— Ss ay

TExtT-FIG. 3.—Dorsal view of the sacrum and the pelvis of

Bradysaurus baini. X yy. (9104.)

than in the forms with no eversion, and the anterior part of the

internal surface is directed very much forwards (text-fig. 39).

The anterior surface of the ilium below the shaft bulges out into

a well-developed swelling, which is continued by the pubis. This

swelling is narrow in the forms whose pelves are generally narrow.

The pubo-ilac suture crosses where this swelling is most prominent,

viz. at about the level of the centre of the acetabulum.

Posteriorly as well as medially there is a similar bulging. Pos-

teriorly the ilio-ischiadic suture also lies across the most prominent

part of the swelling; this is, however, slightly higher than the level

of the centre of the acetabulum. From these two points—the most

anterior and the most posterior of the iliac suture—the suture con-

verges ventrally in rounded curves to a point slightly below the

level of the centre of the acetabulum, but slightly anterior to it. The

ium thus forms about five-eighths of the acetabulum.

Ischium.—The posterior iliac bulge is continued by the ischium in

ventro-posterior direction to slightly below the plane of the centre

of the acetabulum. In dorsal view the dorso-lateral border of the

302 Annals of the South African Museum.

ischium then curves downward, and, extending posteriorly as a

‘* shaft,”’ terminates in a swollen ischial tuberosity. The ischial plate,

which apparently forms over two-thirds of the total pubo-ischiadic

plate, extends medially from the ilio-ischiadic suture and the dorso-

posterior shaft-like border, and lies either more or less horizontally in

the species with broad pelves or more or less obliquely in the narrow

forms. Meeting with its fellow it forms a strong thickened symphysis,

which in ventral view forms a strong carina. The ischial symphysis

is thickened where it apparently meets the pubis, and posteriorly it

is more or less notched. In antero-posterior direction the dorsal

surface 1s convex, whereas the ventral keel is concave. This dorsal

convexity stretches backwards for about half the ischial length; the

symphysis then thins out and dorsally becomes concave, and ventrally

the keel becomes convex. Between the ischial tuberosity and the

thickened posterior symphyseal edge the ischial border is rounded but

fairly thin. In lateral direction the dorsal surface of the ischium is

concave in both the broad and narrow forms, with, however, the

medial portion of the surface slightly flattened in the broad forms.

In both the broad and narrow forms the ventral surface—from the

inferior edge of the acetabulum and the posterior-lateral ischial

“ shaft’ to the symphyseal keel—is concave. This forms a good

surface of attachment for the adductor muscles. The posterior

border of the acetabulum bears a buttress similar to the supra-

acetabular buttress on the ilium. The length from this buttress to

the ischial tuberosity differs in the various species, and the ischial

“shafts” are parallel or diverging. The extent to which the pos-

terior ischial border is excavated also varies. In all, however, the

tuberosities are the posterior points.

The ischium forms about a quarter of the acetabulum.

Pubis.—It has not been possible to determine the pubo-ischiadic

suture in all the species; this is due to the firm union of the two

bones. The suture is apparently a straight one passing just behind

the pubic foramen. Probably only one-eighth of the acetabulum is

formed by the pubis. From the ilio-pubic suture on the anterior

border the anterior pubic border is rounded and directed downwards,

terminating at its most lateral point—the pubic tubercle, which is

not prominent, and situated at the level of the ventral border of the

acetabulum. From here the broadly rounded pubic border converges

to meet its fellow at the symphysis. This roundness is due to the

fact that the pubic edge is turned downwards in the broad forms

and downwards and outwards in the narrow ones, so that in posterior

ETB REI AEDT LE TEE REIS RS OE OS

Pareiasaurian Studies.—Part VI. 303

view there is a distinct lip or rampart which lies considerably below

or lateral to the surface of the pubo-ischiadic plate. The edge of

this rampart is roughened, and it would appear that in life cartilage

was attached to it, forming a kind of cartilaginous prepubis. The

dorsal opening of the pubic foramen is dorso-ventrally oblique, and

pierces the upper posterior corner of the pubis between the pubo-iliac

and pubo-ischiadic sutures. Ventrally it opens just behind the lip

or rampart of the pubis. This is in general the shape of the anterior

pubic border. There are, however, considerable differences amongst

the various Pareiasaurs. In some of the broad forms the whole

anterior border is turned downwards, forming in anterior view a

continuous straight rounded edge. In other broad forms this down-

wardly turned lip or rampart

does not continue right up to the \ Sis Sera \\ if

symphysis, but is separated from I \

a distinct symphyseal tongue by a Hyp \ ,

notch; whilst in an isolated form AN LZ

y : a —

the medial edge of the lip forms Wei = =

WI//7

is Bin

a medially directed process, which Sy es HH ‘: | ay

is apparently formed by a distinct Ss tal it Na | LG

ossicle separated from the pubis

by a suture. This appears to be TEextT-Fic. 4.—Ventral view of the pelvis

Saba ; to show the prepubic processes. In-

a distinct prepubis (fig. 4). Inthe ent. x 3. (9116.)

narrow forms the lip is separated

from the very strong symphyseal tongue, which in one form is large

with an antero-ventrally directed dorsal face.

Lime (figs. 5, 6).

Taken as a whole the Pareiasaur limb is short, massive, and clumsy.

The pelvis hangs between the two limbs, which are permanently bent

at the knee, and are directed outwards with more or less obliquely

outwardly directed propodials.

The limb is composed of five segments. The first segment—the

propodial—consists of one element—the femur. Its proximal end is

longer than high, and consequently its adjustment in the more or

less rounded acetabulum is not finely adapted. On the flattened or

slightly concave or convex end there must, in life, have been a strong

development of cartilage, which made the articulation mechanically

more perfect. The distal end of the femur bears three articula-

tory facets. Two face distally and ventrally, and are separated

304 Annals of the South African Museum.

by a groove and articulated with the proximal end of the tibia.

As is the case with all the articulations, this one is not well-

adapted. In life thin cartilaginous pads must have been present to

make the movements more smooth and defined. The amount of

movement at the knee must have been small. The angle through

which the tibia moved from complete extension to maximum flexion

could not have been greater than 45°. On the postaxial surface a

dorso-ventrally directed groove functioned as facet on which the

fibula moved. The joint was not firm, and the plane of action pre-

cludes all possibility of the femur being partly supported by the

fibula. All the vertical thrust was transmitted to the femur through

and at right angles to the facets for the tibia.

The second segment—the epipodial—consists of two unequal

elements—the tibia and fibula. In the standing position both these

elements were practically vertical and parallel to each other. Both

have well-constricted shafts, with the greatest curvatures facing each

other. The spatium interosseum was therefore great. The proximal

surface of the tibia is terminal, and adapted to fit the two parts

of the femur. Between the surfaces articulating with the two

facets on the femur the tibial surface is raised. This eminence

fits into the intercondylar sulcus, to which it was attached by a

ligament.

The distal facet of the tibia is imperfectly divided into two. The

whole surface is oblique—the outer edge (dorsal) being more distal

than the inner (ventral). The postaxial part of this surface fits

very insecurely on the dorso-proximal surface of the preaxial half

of the composite proximal tarsal element. As the vertical thrust

transmitted through the tibia passed through this ill-adapted articula-

tion a strong development of cartilage must be postulated. The

preaxial part of the tibial facet is concave, separated from the post-

axial portion—albeit indistinctly. This surface lies preaxial to the

tarsal element, so that there is here an overhang of the tibia. One

of us has previously expressed the opinion that a cartilaginous element,

probably representing the tibiale, must have been present in life.

A cartilaginous tibiale and a great development of cartilage between

the tibia and the proximal tarsal element are essential to make it

mechanically possible for the weight of the body to be transmitted

vertically on to the basipodium. Considering the articulatory surface

of the tibia and the proximal basipodial element by themselves, a

fair amount of movement intero-externally (dorso-ventrally) appears

possible. That this possibility is not realised will become apparent

Pareiasaurian Studies.—Part VI. 305

when the articulation between fibula and proximal tarsal element is

taken into account.

The proximal facet of the fibula is semi-terminal, facing more or

less preaxially. In shape it is roughly rectangular, with the longest

sides directed ventro-dorsally, so that it is at right angles to the distal

facet. Articulating with the side of the femur no direct thrust could

be withstood by the fibula.

The distal facet of the fibula is large and oval, with its long axis

pre-postaxially. The articulation between the fibula and the proximal

basipodial element is the most well-developed joint in the whole limb.

Its surface is slightly concave, and when some cartilage was present

the joint must have been a very fine one, and would allow very

little play between fibula and foot. As the rotatory movements of

the foot are brought about by forces acting through the fibula, this

bone is placed in an advantageous position to fulfil its chief function.

Although the tibia and fibula move more or less in unison, their

functions are totally different. The tibia is a supporting element,

whereas the fibula regulates the rotatory movements of the foot.

The third segment of the limb—the basipodium (Pl. XX XVIII)

—is represented in the Pareiasaurs by two sub-segments. The

proximal sub-segment is represented by a single long element. As

was shown elsewhere, it is composite. In addition to this ossified

element, it seems necessary to postulate the presence of a cartilaginous

element. From the above it has already become evident that the

rectangular, proximal, basipodial element is closely united to the fibula

and insecurely fitted on to the tibia, and thus not well-adapted to

receive the weight of the body. Its distal surface bears no distinct

facets for the more distal sub-segment, which is composed of four

pebble-like bones. These are the distal tarsals—there are no free

centralia in the Pareiasaur foot. Between the proximal tarsal element

and those distal tarsals a great amount of movement was possible. This

constituted an intertarsal joint. The ventral surface of the proximal

element appears to be adapted for fitting into a cartilaginous “ heel.”

As the bone itself is directed obliquely outward, a cartilaginous heel

must be postulated, otherwise it would be impossible for the foot to

be firmly put on the ground and bear the weight of the body.

The fourth and fifth segments—the meta- and acropodials—are

composed of short, squat elements, which had very little play amongst

themselves. The toes are directed obliquely outward in the standing

position. Relative movements between the toes were reduced to a

minimum. The meta- and acropodial probably acted as one mass.

;

306 Annals of the South African Museum.

This mass with moderately developed facets for the distal tarsals had

quite a considerable possibility of dorso-ventral movement. At the

intertarsal joint, therefore, the last two segments of the limb moved

in relation to the third segment.

In summary it may be stated that at the knee and “ ankle” joint

there was little movement, as was also the case between the toes; at

the hip and intertarsal joint there was more freedom of movement.

The hip joint not being a well-developed ball-and-socket joint, had

considerably less play than is the case in Mammals, and also than in

the Therapsids, where the head of the femur, although not situated on

a neck, is by the curvature of the bone somewhat preaxially directed.

Femur (figs. 12, 18, 22, 27, 31, 34, 40, 44, 45, 51).—The femora of

the different genera and species agree closely in nearly all the essential

points of their structure. The only outstanding point of dissimilarity,

apart from a variation in size, lies in the preaxial curvature, which

is correlated with a more preaxially directed proximal articulation.

This feature is concomitant with a more upright position of the limbs,

which is considered to be an advance on an oblique or horizontal

disposition of the propodials. It is therefore of interest to note that

the forms from the higher geological horizons show this feature very

clearly. A decrease in size accompanies this line of advance. The

more subordinate points of similarity and dissimilarity will be treated

systematically after the general considerations of the femora as form-

ing a homogeneous group.

The Pareiasaurian femur is a short and broad bone with greatly

expanded ends. In the larger forms it is very large and massive,

whereas in the smaller species it is of medium size; but even here it

must be considered to be squat, even though, as in Propappus, it is,

in comparison to the larger femora, long and more slender. Even

here the shaft forms considerably less than half of the total length.

The bone is to some extent twisted on its shaft, so that (in the right

femur) the distal endis rotated through about 10° in clockwise direction.

In Propappus and Brachypareia the rotation is even greater—about

15°. The proximal articulatory surface is long and egg-shaped, with

the narrower point directed postaxially. It is flat, slightly convex or

concave. Postaxially it either tapers off imperceptibly into the

posterior border of the external trochanter or is separated from this

by a very slight notch. The external trochanter is directed postero-

ventrally, and passes either gradually into a flange of bone, which

nearly always reaches the postaxial surface of the fibular condyle, or

is more abruptly separated off from this flange. The trochanter is

Pareiasaurian Studies.—Part VI. 307

situated further distally in some forms than in others. It is the

remnant of the postaxial limb of the primitive Y system of adductor

ridges. Preaxially the articulatory surface ends abruptly. The

proximal preaxial surface is broadly rounded. At about the upper

third of the bone the internal trochanter arises. This cannot be

considered to be a distal continuation of the ventro-preaxial edge

of the bone as the external trochanter is of the postaxial border.

The internal trochanter arises proximally slightly ventral to the

ventro-preaxial edge. It then extends distally as either a low, strong,

and fairly straight ridge, or as a prominent, not very massive, sig-

moidally curved ridge. The internal trochanter of the low, strong, and

straight type decreases gradually as it extends distally, when it is

directed medially and fades out as it approaches the popliteal depres-

sion; the internal trochanter of the prominent, medium, and curved

type ends distally fairly abruptly, but a more or less distinct ridge

continues distally. In some it fades away in medial-distal direction ;

in others it continues right up to the facet of the radial condyle, and

in this case it forms the ventral preaxial border. These different

distal continuations of the internal trochanter are the remnants of

the stem of the primitive Y. In no femur is there on this ridge any

indication of a tuberosity for the reception of the coccygeo-femoralis,

z.e. of a fourth trochanter.

Ventrally between the internal and external trochanter the femur

is excavated—deeply in those forms with a fairly distinct external

trochanter, and either deeply or fairly shallowly in the others. In

this hollow—the intertrochanteric fossa—the pubo-ischio-femoralis

externus inserted. Distally the excavation shallows and passes with-

out a distinct border into a slight pre-postaxially concave surface,

which deepens as it reaches the tibial facets to form a depression—

the popliteal fossa. This depression is distally separated from the

intercondylar sulcus by a distinct border.

Dorsally the ilio-femoralis flange is for its proximal third preaxially

bounded by a ridge, which in some forms ends distally very abruptly.

Preaxially to this ridge there is, on the proximal third of the surface,

a shallow groove for the reception of the pubo-ischio-femoralis internus.

Distally this groove fades away into the dorsal and preaxial surface

of the shaft. In longitudinal direction the dorsal surface has a

moderate or strong curvature. In the latter case, especially in

Brachypareia, this feature is reminiscent of the femur of the tortoises,

where the proximal articulatory surface lies more or less on the

dorsal surface.

308 Annals of the South African Museum.

Distally and ventrally the two condyles are separated by a deep

notch—the intercondylar sulcus. On the dorsal surface, just proximal

to the sulcus, there is a short, narrow, but deep groove, where the

femoro-tibialis was most firmly attached. The fibular condyle extends

further distally than does the tibial. This is due to the fact that the

former serves as surface of muscular origin, whereas the latter has

no such function. The fibular condyle is directed postaxially and

dorsally, whereas the tibial condyle is directed preaxially and ventrally.

The distal corner is rugose, and serves for the origin of the peronaeus

longus and the extensor communis digitorum. The postaxial surface

is excavated to form a groove for the reception of the proximal end

of the fibula. In Brachypareva the proximal border of this groove

is formed by a distinct tubercle, which limits the proximal movement

of the fibula.

Semi-ventrally the facets of the two condyles both receive the tibia.

Both are more or less kidney-shaped. The preaxial is the longer, and

its distal edge is also the edge of the tibial condyle; whereas distally

and postaxially of the postaxial facet lies the articulatory surface

for the fibula, and postaxially to this lies the already mentioned

surface of origin of the muscles for the fibula and the toes.

Tibia (text-figs. 13, 19, 23, 28, 29, 35, 37, 41, 46).—-In all the species

the tibia is a short massive bone with a well-constricted shaft and

dilated ends. The postaxial curvature is the greater. The spatium

interosseum is therefore great. The proximal end is greatly expanded,

and bears a bipartite articulatory surface. The two facets are directed

obliquely downward—the one postaxially and the other preaxially—

and meet at an angle at about the median line. Here a slight ridge

is developed, which near the external dorsal border forms a tubercle

to which the knee-ligament is attached. The preaxial facet is both

longer and broader than the postaxial one.

On the dorsal surface, beginning at the most dorsal point of the

preaxial facet, a strong cnemial crest runs distally for nearly half the

length of the bone. The strength of the cnemial crest is accentuated

by the development of a deep groove which postaxially runs parallel

to the crest. In some tibiae there is also a slight groove preaxial to

the cnemial crest. This is believed to be a feature chiefly due to

crushing. A little more than half-way down the bone an eminence is

developed on the postaxial surface of the shaft. This is quite weak

in most tibiae, but in that of Propappus omocratus it is a very strong

tubercle. Internally (ventrally) a diagonal ridge running from the

proximal postaxial corner to the distal preaxial corner separates two

Pareiasaurian Studies.—Part VI. 309

moderately concave surfaces for muscular attachment. The distal

articulatory surface is roughly oval with a slight ventral notch,

thereby producing an appearance of being kidney-shaped. The long

axis lies pre-postaxially. Whereas the proximal surface is at right

angles to the long axis of the bone, the distal surface is oblique—the

dorsal surface of the bone being longer than the ventral. The articu-

latory surface is imperfectly divided into two facets. The preaxial

facet is flattened or weakly convex, whereas the postaxial facet is

weakly concave. The latter facet articulates with the proximal

tarsal bone, while the former must in life have rested upon a carti-

laginous “‘ tibiale.”’

The differences in the tibiae of the various species are small and of

little value as- specific characters, and will be discussed in the sys-

tematic account.

Fibula (text-figs. 14, 20, 24, 36, 42).—The fibula is, in comparison

to the tibia, a light and slender bone. This appearance of slenderness

is increased by the fact that it is always longer than the tibia—in

some species only a few millimetres, whilst in others as much as

50 millimetres. It has a long, straight or curved shaft connecting the

two expanded ends. The distal end is expanded pre-postaxially and

is greater than the proximal expansion. The expansion is greatest

preaxially, with the result that the preaxial curvature is greater than

the postaxial. This produces a wide spatium interosseum. Dorso-

ventrally the distal end is also somewhat thickened. On the preaxial

ventral surface the distal end bears a ledge for muscular attachment.

The distal articulatory surface is very well developed. Only the

preaxial two-thirds of the distal end articulates with the composite

proximal tarsal bone. On the postaxial third ligaments running to

the more distal segments of the limb were probably attached.

The proximal end, expanded dorso-ventrally, is preaxially inclined,

and lies in a plane which makes an angle of 45° to the plane of

the distal expansion. The dorso-postaxial surface of the head is

rough and serves for muscular attachment. From here along the

postaxial border are two areas for muscular insertion, which are

separated at the proximal third of the bone by a weak ridge or

tubercle. The articular facet is not terminal. As it fits on to the

side of the femur it faces half proximally and half ventro-preaxially.

Just distal to the facet is a hollow bounded on the preaxial border by

a swelling, which serves for muscular attachment.

Foot (text-figs. 5, 6, 59, 60, Pl. XXX VIII).—Since Boonstra’s

account of the Pareiasaurian foot appeared (Ann. 8. Afr. Mus.,

310 Annals of the South African Museum.

vol. xxviii, pt. 1) we have seen Hartmann-Weinberg’s paper on

the carpus and tarsus, and in the interval one of us has had the

opportunity of collecting eighteen feet from the Tapinocephalus Zone.

A study of this additional material has shown that some corrections of

the views based on the former scanty material are necessary. It has

become apparent that the mounters of the skeletons of Hmbritho-

saurus (Bradysaurus) baini and seeleyi in the South African Museum

Text-Fic. 5.—True lateral TExtT-FIG. 6.—True anterior view

view of the pelvis and of the right half of the pelvis

hind limb in the stand- and limb—with the limb in

ing position. Brachy- the forward position at the

pareia rogerst. X zy. commencement of the stride.

(8953.) Brachypareia rogersi. X zy.

(8953.)

articulated the left tarsal element to the right foot, and vice versa.

In Boonstra’s description of this bone this error can be corrected by

substituting pre- for postaxial and proximal for distal, and vice versa.

Proximally the postaxial two-thirds of the surface is formed by a

well-defined, elliptical, articulatory surface. With this the fibula

articulated. Preaxially the articulation for the tibia lies more dorsal

and is not very well defined. The dorsal opening of the nutritive

foramen lies near the dorsal edge, as in the Russian forms.

We still hold the view that this proximal tarsal element is composite

and composed chiefly of the fibulare and intermedium. Through the

courtesy of Professor Young, of the University of Cape Town, sections

Pareiasaurian Studies.—Part VI. 311

have been made of this bone. There is no definite arrangement of

spongiosa and compacta to show which bones are incorporated.

Distally there is an indistinct indication of the incorporation of a

centrale. These results, although negative, make it impossible for

us unreservedly to accept Hartmann-Weinberg’s views. It is still

maintained that the specimens examined by us have the proximal

row of the basipodium composed of one large element, representing

the fused fibulare and intermedium, with possibly also a centrale

incorporated and a cartilaginous tibiale.

No free centrals are present. Although the elements of the distal

row appear to ossify late in life and are not well preserved, there are,

when the foot is fully ossified, four distal tarsals. The metatarsals

of the fourth and fifth toes articulate with a common distale; the

other metatarsals have a distale each.

The metapodium consist of five metatarsals. The acropodium con-

sists of five digits. In view of the fact that all of the eighteen feet

recently collected and examined have five digits, it appears most

reasonable to assume that in the specimen of Bradysaurus vanderbylt,

previously described, the fifth digit was lost before fossilisation, and

that in life five digits were present. The phalangeal formula is

23343.

MATERIAL.

Name. Pelvis. Femur. Tibia. Fibula. Foot.

Bradysaurus bainr complete left, right | left, right left, right | proximal

elements

>» x “5 part of right} right part of right| parts

a a ila missing rhe Hs ae a

Ss a4 incomplete | right right

> 99 part left bie ae

a5 oe as ae left part left complete

ao os complete left, right | left, right left, right | left, right

> pad wie incomplete ae left left, right 3

sf seeleyt . complete left, right | left, right left, right | proximal

elements

9 3p complete left, right | left, right left, right | left, right

99 0 incomplete | part left a oF ws

> aly Wks bes left, right | left, right left, right | left.

35 vanderbyli incomplete | parts of parts of parts of wie

left, right left, right | left, right

>» bs complete left, right | left, right part of part of

left, right | left, right

> > complete left, part part left part left part left

right |

* Skull.

312

MATERIAL—continued.

Annals of the South African

Museum.

No. Name. Pelvis. Femur. Tibia. Fibula. Foot.

5590 | Nochelesaurus strubent complete left left part left part left

6239* i alexanderi . oe left right Mss ae "4

6238*| Dolichopareia angusta Ae ae right part right a

5002*| Platyoropha broom: . incomplete | left, right Si 5 ‘

6242 | Hmbrithosaurus schware incomplete | right left, right left oe :

8034* Fy es complete left, right | left, right left, right | parts left, |

right | —

9148 Be Me part left, left, right left, right | left, right |

right x

8953 | Brachypareta rogersi incomplete | left, right | left, right left parts

4350* 3 bn incomplete | part right | right right right

5012* x BS parts left, right | right right incomplete

right

2337*| Pareiasuchus peringueyr complete left — parts

3016* 8 NASLCOTNIS complete left crushed right als rs.

4996 5 cf. peringueyi | incomplete | part left, Be right

right

4344 aS cf. nasicornis | incomplete | left, right | part right

1058 | Propappus omocratus parts left, right | left, right

1058a* 5 - ae left st

2351 : parvus incomplete

* Skull.

SYSTEMATIC.

Genus BrapysauRus Watson.

Pelvis.—The pelvis is large and its constituent bones massive. It

is high or moderately high and very broad. The broadness is to some

extent masked by the height. In anterior view the inner border is

broadly U-shaped. With the sacral ribs in position its opening is

irregularly oval, with the lateral diameter slightly greater.

Sacrum.—Four vertebrae are concerned in its attachment.

three, or all four centra are coalesced. The sacral ribs are massive.

In one species the first rib and in the other two species the second is

the strongest. In the first case the corresponding vertebra is coalesced

to its fellows; in the latter it is not. The two last ribs, although

lighter than their predecessors, are nevertheless strong.

Ilium.—The ilium is forwardly directed, with its blade strongly

or moderately diverging. The anterior edge is hardly or slightly

everted. The posterior iliac process is large and projects slightly or

moderately. The iliac blade is long or very long and fairly low. The

outer surface is slightly or moderately concave. ‘The iliac shaft is

short and moderately constricted. The anterior swelling of the iliac

border is strong and broad. The posterior swelling is strong and broad.

Two,

Pareiasaurian Studies.—Part VI, 313

Ischium.—The ischial “ shaft” is long or very long and practi-

cally horizontal, and runs parallel to its fellow with a ridge just

posterior to the ischial buttress in one species. The posterior border

continues the strong broad swelling initiated by the ilium. At the

pubo-ischial suture the symphysis is moderately thick. The dorsal

surface of the ischial plate is slightly concave in antero-posterior

direction. The medial portion of the ischial plate is practically hori-

zontal, thereby producing the great pelvic breadth. The posterior

edge is straight and in the same plane as the posterior end of the

symphysis, or concave with the ischial tuberosity more posterior than

the symphysis. |

Pubis.—The anterior border of the pubis continues the prominent

broad swelling of the ium. From the ilio-pubic suture to the not.

prominent pubic tubercle the surface is strongly convex and directed

medially. The step from the pubic tubercle to the pubic plate is

rounded and in medio-posterior direction fades away. A distinction

between this step and the anterior edge of the pubic plate is thus

indicated, although it is by no means well-marked. Below this step

the pubis bends downwards to form a strong rampart. This down-

ward bending includes or does not include the symphysis, which is

only slightly prolonged to form a weak or moderate tongue or no

tongue at all. The notch is wholly absent or moderate. These

features produce a very deep, rounded, anterior pubic border. The

acetabulum is large, oval or irregularly rounded, moderately deep, and

directed outwardly and slightly downwardly.

Femur.—The femora of the three species are very similar. They

are of fairly large to very large size and of massive build. The width

across the greater trochanter and also over the distal epicondyles is

large. The posterior edge of the flange for the insertion of the ilio-

femoralis is very or moderately broadly rounded. The greater tro-

chanter is situated fairly or very much distally, and it is proximally

more or less distinctly demarcated from the edge descending from

the articulatory surface. Distally the trochanter is not very clearly

distinct from the ilio-femoralis flange. The internal trochanter is

strong, straight, and prominent or not very prominent. The “fourth

trochanteroid”’ ridge is strong or very strong and fades away medially.

The shaft is short and only moderately constricted. The dorsal

curvature is moderate or fairly strong, but the proximal articulation

is not in the least directed preaxially, and the tibial facets are directed

downwards. The intertrochanteric fossa is of medium depth.

Tibia.—The tibiae of the genus are fairly or very long, massive or

VOL. XXVIII, PART 3. 22

314 Annals of the South African Museum.

very massive. The proximal articulatory surface is very long and

very broad. The postaxial facet is smaller or very much smaller

than the preaxial facet. The distal facet is long and broad or narrow.

The eminence of the postaxial border of the shaft is weak.

Fibula.—In comparison with the other genera the fibula of Brady-

saurus is long. It is also much or a little longer than the tibia. It

is straight and very strong, with a prominent ridge or longitudinal

tubercle on the postaxial border of the shaft. The expansion of the

ends is great.

Foot.—The foot has the first four distals ossified in one species.

The first toe is either much broader than any of the others or only

slightly broader than the second, third, and fourth. The fourth digit

is very long; the fifth very slender.

Bradysaurus baini (Seeley).

Pelvis (text-figs. 3, 7-11).—The anterior iliac edge is hardly everted.

In the type the posterior iliac process is not very prominent ; in two

TEXtT-FIG. 7.—Lateral view of the left side of the pelvis of Bradysaurus

baini, from the cast of the type. x 4.

other specimens (3533, 5127) referred to the species it is quite pro-

minent. The iliac blade is very long and the outer surface slightly

concave. The ischial “shaft” is very long with a ridge near the

post-acetabular ischial buttress. In the type and in 3533 the posterior

border is straight, whereas in 5127 it is concave. There is no sym-

physeal pubic tongue in the type and 3533, so that the whole of the

pubic border is bent down; in 5127 there is a slight tongue, and in

4999 a medial notch with two processes lateral to it. The two middle

sacral vertebrae are coalesced. The second sacral rib is the strongest.

Trxt-Fic. 8.—Anterior view of the pelvis and sacrum of Bradysaurus baint.

X gy. (3533.)

Trxt-F1g. 9.—Lateral view of the left side of the pelvis of Bradysaurus baini.

X gy. (3533.)

TrExt-Fic. 10.—Anterior view of the pelvis and sacrum of Bradysaurus baini.

xX gy. (5127.)

315

Text-Fic. 11.—Ventral view of the pelvis of Bradysaurus baini.

X gy. (4999.)

\ ;

TEext-F1G 12.—The femur of Bradysaurus baint. xX 4. (5127.)

i(a) Ventral. (6) Dorsal. (c) Anterior. (d) Proximal.

TExtT-FIG. 13.—The tibia of Bradysaurus baint. X }. (5127.)

(a) Dorsal. (6) Ventral. (c) Posterior. (d) Distal. (e) Proximal.

316

Pareiasaurian Studies.—Part VI. 317

Femur (text-fig. 12)—The femur is very large. The flange for the

ilio-femoralis is broadly rounded. The external trochanter is situated

very much distally and distinct from the proximal edge. The internal

trochanter is not very prominent. The rem-

nant of the stem of the Y is a very strong

ridge, continued from the anterior branch to

the preaxial facet. The dorsal curvature is

fairly strong.

Tibia (text-fig. 13).—The tibia is fairly long

and very massive. Proximally the postaxial

facet is smaller than the preaxial one. The

distal end is long and narrow.

Fibula (text-fig. 14).—The fibula is a little

longer than the tibia. There is a strong pyyrmo.14.—Thefibula

ridge on the postaxial border of the shaft. of Bradysaurus bait.

Foot.—The dorsal ridge on the proximal ~ aoe!)

tarsal element is very strong. In 9136 the (2) Bea

first and fourth distals are present; they are

pebble-like bones. The elements of the first digit are not very broad.

Bradysaurus seeleyi Haughton and Boonstra.

Pelvis (text-figs. 15-17).—The anterior iliac edge is hardly everted.

The posterior iliac process is moderately projecting. The iliac blade

Trxt-Fic. 15.—Lateral view of the left side of the pelvis of Bradysaurus seeleyt.

X yy. (9169.)

is very long and its outer surface is slightly concave. The ischial

“shaft”? is long. The posterior border is slightly concave. A

rounded step forms the dorso-anterior pubic border. The down-

wardly directed rampart does not reach the symphysis. There is a

FFF =

- — 2“

318 Annals of the South African Museum.

moderate pubic tongue with a moderate notch lateral to it. The

last three sacral vertebrae are coalesced. The second rib is the

strongest.

TExtT-Fric. 16.—Posterior view of the pelvis TExt-Fic. 17.—Lateral view of the right

and sacrum of Bradysaurus seeleyi. X 5. side of the pelvis of Bradysaurus seeleyi.

(5624.) X go. (5624.)

Femur (text-fig. 18).—The femur is very large; the flange for the

ilio-femorahs is moderately rounded. The external trochanter is not

— f 2,

Aes

“ ote earese

= er eg Sh =

EE EES

igen ey ==

4 iter} s—- re SS

Lt AG oe pe ea

rs le ah Se es

SESS ee RS

—_—————— eee ee eT

= SSS SSS

——<— ———— Sos

sec

er > ew

—— ‘x

TExtT-FIG. 18.—Femur of Bradysaurus seeleyi. xX 4. (5624.)

(a) Dorsal. (6) Anterior. (c) Ventral. (d) Proximal.

situated very much distally ; it is not very distinct from the proximal

edge, and still less distinct from the edge of the ilio-femoralis flange.

Pareiasaurian Studies.—Part VI. 319

The internal trochanter is not very strong. The stem of the primitive

Y is strong, running from the internal trochanter to the tibial condyle.

The dorsal curvature is moderate.

Tibia (text-fig. 19).—The tibia is very long and massive ; proximally

TExt-FIG. 19.—Tibia of Bradysaurus seeleyi. X }. (5624.)

(a) Dorsal. (6) Ventral. (c) Posterior. (d) Proximal. (e) Distal.

the postaxial facet is much smaller than the preaxial one; the distal

end is long and broad.

Fibula (text-fig. 20).—The fibula is much

larger than the tibia. On the postaxial border

there is a prominent longitudinal tubercle.

Foot (Pl. XXXVIII, a, 6). — The dorsal

ridge of the proximal tarsal element is strong.

In 9168 all the four distals are ossified ; they

are not pebble-like but irregularly rectangular.

The elements of the first digit, especially the

terminal claw, are very broad.

TExt-Fic. 20.— Fibula

of Bradysaurus seeleyv.

x 4. (5624.)

(a) Ventral. (6) Dorsal.

Bradysaurus vanderbyli Haughton and

Boonstra.

Pelvis (text-fig. 21).—The anterior iliac edge

is slightly everted. There is hardly any posterior process. The

iliac blade is long and its outer surface moderately concave—with

the concavity extending down the shaft. The ischial “shaft” is

long. The posterior ischial border is straight. The whole anterior

pubic border is turned downwards to form a rampart. There is no

tongue and no notch lateral to the symphysis. Four sacral vertebrae

are coalesced. The first rib is the strongest.

Femur (text-fig. 22).—In 8941 the femur is fairly large, whereas in

320 Annals of the South African Museum.

3718 it is large and massive. In 9169 it is fairly large and massive.

The flange for the ilio-femoralis is proximally rounded, but distally

hardly obscures the shaft. The external trochanter is situated much

TEXtT-FIG. 21.—Lateral view of the TExtT-FIG. 22.—Femur of Bradysaurus van-

right side of the pelvis of Brady- derbyli. X< 4. (9169.)

saurusvanderbyli. X 5. (8941.) (a) Ventral. (6b) Dorsal. (c) Anterior.

proximally and is not distinct from the postaxial border. The internal

trochanter is strong and fairly prominent. The stem of the Y is

moderate, is directed medially, and fades out before reaching the

tibial facets. The dorsal curvature is very moderate.

Tibia (text-fig. 23).—The tibia is fairly long and of medium stout-

ness ; proximally the postaxial facet is very much smaller than the

preaxial ; the distal end is long and broad.

Fibula (text-fig. 24).—The fibula is a little longer than the tibia.

On the postaxial border there is a medium longitudinal tubercle.

TExtT-FIG. 23.—Tibia of Bradysaurus TExt-Fic. 24.—Fibula of Bradysaurus

vanderbyli. XxX 4. (9169.) vanderbyli. xX 4. (9169.)

(a) Dorsal. (6) Ventral. (a) Dorsal. (b) Ventral. (c) Anterior.

Foot.—The dorsal ridge on the proximal element is moderate. The

first, second, and fourth distals are ossified as pebble-like bones.

The elements of the first digit are moderately broad.

Parecasaurian Studies.—Part VI. Sil

Genus NocHELESAURUS Haughton and Boonstra.

Pelvis (text-figs. 25-26).—The pelvis is large and its constituent

bones massive. It is very high and narrow. The height somewhat

accentuates the narrowness. In anterior view the inner border is

V-shaped. With the sacrum and ribs in position the opening is

roughly pear-shaped.

Sacrum.—Three vertebrae are concerned in the attachment of the

pelvis to the vertebral column. The two anterior ones are coalesced.

It is just possible that the vertebra anterior to those has a rib assisting

in the attachment of the sacrum. The sacral ribs are massive. The

first sacral rib is the strongest. The second and third, though lghter

than the first, are, nevertheless, strong.

Itium.—The ilium is vertically directed, with its blade parallel to

the long axis of the body. The eversion of the anterior edge of the

blade is moderate. The posterior process is large and strong, and

projects moderately. The bladeis short and high. The outer surface

is moderately concave. The iliac shaft is of medium length and

moderately constricted. The swelling of the anterior iliac border is

strong and narrow. The posterior swelling moderate and narrow.

Ischiwm.—The ischial “ shaft” is short and inclined slightly down-

ward and outward, thereby making the pelvic cavity appear broader

in posterior view. At the pubo-ischial suture the symphysis is very

thick. The dorsal surface of the plate is very slightly concave in

antero-posterior direction. The medial portion of the plate meets its

fellow at an angle. The posterior brim medially converges in anterior

direction, so that the tuberosities are much more posterior than the

symphysis. The posterior border continues the iliac swelling, which

is moderate and narrow.

Pubis.—The anterior border continues the strong narrow swelling

of the ilium above it. The pubic tubercle is not very distinct from

the pubo-iliac swelling. The pubis extends ventrally and medially

in a rounded convex curve, which fades away on reaching the anterior

border. laterally and ventrally from this curve a shallow groove

separates the antero-dorsal surface of the rampart from this curve.

The rampart is directed downwards, forwards, and outwards. Medially

it 1s separated from the symphyseal pubic tongue by a notch. The

notch is fairly deep, and the tongue strong and directed downwards

and forwards.

The acetabulum is large, rounded, of moderate depth, and directed

outwards and only slightly downwards.

322 Annals of the South African Museum.

Femur.—There are quite a number of differences between the femora

of the two species. In both the femora are large and massive. The

width across the greater trochanter is large or medium, and the width

over the distal epicondyles is medium. The flange for the insertion

of the ilio-femoralis is very broadly rounded or practically straight.

The greater trochanter is situated much proximally or fairly far

distally, and is proximally distinct or fairly indistinct from the

proximal postaxial edge. Distally the greater trochanter is continuous

with or only fairly demarcated from the ilio-femoralis flange. The

internal trochanter is strong, straight, not or fairly prominent. The

stem of the primitive Y is fairly strong, but dies away in the direction

of the preaxial facet. The dorsal curvature is fairly weak, so that

the tibial facets look much downwards.

Tibia.—The tibiae of both species are long and massive; the

proximal end is long and broad; the postaxial facet is smaller than

the preaxial one; the distal end is long and narrow; the eminence

on the postaxial border of the shaft is very weak or moderately

developed.

Fibula.—Imperfectly known.

Foot.—There is a moderate dorsal ridge on the proximal element ;

the second, third, and fourth pebble-like distals are present ; the rest

of the foot is unknown.

Nochelesaurus strubeni (Broom).

Pelvis (text-figs. 25, 26).—As for the genus.

Femur (text-fig. 27).—The width across the external trochanter is

very large. The ilio-femoralis flange sweeping distally includes the

external trochanter in a broad curve carried right up to the fibular

condyle. The indistinctly demarcated external trochanter is situated

much proximally. The internal trochanter is fairly prominent.

The intertrochanteric fossa is deep.

Tibia (text-fig. 28).—The tibia is slightly longer and has a more con-

stricted shaft than that of alexanderi; the postaxial tubercle is weak.

Fibula.—The fibula is imperfectly known.

Foot.—As for the genus.

Nochelesaurus alexandert Haughton and Boonstra.

Pelvis.—The pelvis is unknown.

Femur.—The width across the external trochanter is medium.

Proximally the external trochanter is distinctly demarcated ; distally

Trxt-Fic. 25.—Lateral view of the left TExtT-FIG. 26.—Anterior view ofj,the

side of the pelvis of Nochelesaurus pelvis and sacrum of WNochele-

strubem. X 75. (5590.) saurus strubent. X pe. (5590.)

TEXtT-FIG. 27.—Femur of Nochelesaurus strubent. X 4. (5590.)

(a) Ventral. (6) Anterior. (c) Dorsal. (d) Proximal.

Trxt-Fricg. 28.—Tibia of Nochelesaurus strubeni. xX 3. (5590.)

(a) Dorsal. (b) Ventral. (c) Posterior. (d) Proximal. (e) Distal.

323

324 Annals of the South African Museum.

it is continuous with the ilio-femoralis flange, which has a straight

edge ; it is situated fairly far distally. The internal trochanter is not

prominent butis strong. The intertrochanteric fossa is fairly shallow.

TEext-Fic. 29.—Tibia of Nochelesaurus alexandert. x 4. (6239.)

(a) Dorsal. (6) Ventral. (c) Posterior. (d) Proximal. (e) Distal.

Tibia (text-fig. 29).—The tibia is slightly shorter and has a less

constricted shaft than that of strubeni; the postaxial border has a

strong tubercle.

Fibula.—The fibula and foot are unknown.

Genus PLatyoropHa Haughton and Boonstra.

Platyoropha broom: Haughton and Boonstra.

Pelvis (text-fig. 30).—The pelvis is large and its constituent bones

massive. It is broad and high. In anterior view the inner edge is

U-shaped. With the sacral ribs

in position it is oval.

Sacrum.—Four vertebrae are

concerned in its attachment, of

which the three posterior ones

have their centra coalesced. The

sacral ribsarestrong—the second

being the strongest. The first 1s

only slightly weaker than the

second. The third and fourth

are moderate.

Ilium.—The ilium is greatly

half of the pelvis of Platyoropha broomi. . Pes

x dy (5002.) forwardly directed, with its blade

strongly diverging anteriorly.

The antero-ventral edge is only slightly everted, but the whole blade

is concave antero-posteriorly. The posterior iliac process is large and

Pareiasaurian Studies.—Part VI. 325

projecting posteriorly. The blade is long and low, and only slightly

concave dorso-ventrally. The shaft is of moderate length and

moderately constricted. The anterior swelling of the iliac border is

strong and broad, as is also the posterior swelling.

Ischiwm.—The ischial “ shaft’ is long and downwardly and out-

wardly directed, thereby making the pelvic cavity appear broader in

posterior view. ‘The posterior border continues the broad swelling

initiated by the ilium. At the pubo-ischial suture the symphysis is

thickest. The dorsal surface of the plate 1s slightly concave in antero-

posterior direction. The medial portion of the plate is practically

A ite >

A eo

TExt-F1G. 31.—Femur of Platyoropha broomi. x }. (5002.)

(a) Dorsal. (b) Anterior. (c) Ventral. (d) Proximal.

flat. The posterior brim is straight, with the ischial tuberosity only

slightly more posterior than the symphysis.

Pubis.—The antero-lateral border of the pubis continues the swell-

ing initiated by the ilium above it. The pubic tubercle is not clearly

differentiated, and from the ilio-pubic swelling the pubis extends in a

rounded convex curve, first downwards and then medially towards

the symphysis. There is therefore no ledge as in the other forms.

The symphyseal portion is not preserved, but it appears that the

ventral bulwark of the downwardly bent pubis extends to the sym-

physis. In that case it is probable that no pubic tongue and no

notch was developed.

The acetabulum is large, rounded, of moderate depth, and directed

outwards and slightly downwards.

Femur (text-fig. 31).—Both femora appear to be somewhat crushed

326 Annals of the South African Museum.

preaxially. Making allowances for this, it becomes apparent that the

femur is large and massive. The width across the trochanter and

over the distal epicondyles is not very large. The flange for the ilio-

femoralis is not wide and its edge is straight. The greater trochanter

is situated fairly far distally and is distinctly demarcated both proxi-

mally and distally. The internal trochanter is strong, prominent,

and straight, but directed somewhat preaxially. The “ fourth

trochanteroid ”’ ridge forms the ventro-preaxial edge right up to the

condyle, but it is not very strong. The fact that it lies on the edge

makes it appear stronger. The dorsal curvature is fairly strong and

the tibial facets are therefore directed more distally. The proximal

articulatory surface is somewhat preaxially directed, even when

allowances are made for distortion. The intertrochanteric fossa is

fairly deep.

Tibia.—The tibia, fibula, and foot are unknown.

Genus EMBRITHOSAURUS Watson.

Embrithosaurus schwarzt Watson.

Pelvis (text-figs. 32, 33).—The pelvis is large and the constituent

bones massive. It is very high and narrow, and in anterior view

V-shaped. With the sacral ribs in position the opening is pear-shaped.

Four vertebrae are concerned in its attachment, of which the two

inner ones are coalesced. There is some doubt as to whether the

lumbo-sacral had a rib abutting on to theilium. The first undoubted

sacral vertebra has a very strong sacral rib, which extends downwards

along the ilium to just above the level of the supra-acetabular iliac

ledge. The second true sacral rib (the third counting that of the

lumbo-sacral vertebra) is much weaker than the first. The rib from

the caudo-sacral vertebra is weak.

Ilum.—The ilium is upstanding, with its blade diverging anteriorly,

thereby producing a considerable iliac width notwithstanding its

narrow pelvis. There is practically no eversion of the anterior edge.

The posterior process of the iliac blade is large, but not projecting

greatly. The blade is of medium length and height. The external

surface is practically flat. The iliac shaft is long and fairly constricted.

The anterior swelling of the ilac border is strong but narrow. The

posterior swelling fairly narrow and not very prominent.

Ischium.—The ischial “ shaft” is short and diverging downwards

posteriorly, with a slight indication of a ridge on the upper surface of

Pareiasaurian Studies.—Part VI. 327

the tuberosity. The latter character producing a great ischial width

makes the pelvis appear broader viewed posteriorly than it does from

an anterior aspect. Where the pubo-ischial suture probably lies, the

ischial symphysis is very thick. This great thickness decreases

rapidly posteriorly, producing a thin posterior rim and a downwardly

directed, flat, dorsal, ischial surface. The medial portion of the ischium

is oblique, thereby producing the height and narrowness of the pelvis.

The posterior border runs obliquely forward to the symphysis, 2.e.

TExtT-FIG. 32.—Anterior view of the pelvis Trxt-ric. 33.—Lateral view of the

and sacrum of Hmbrithosaurus schwarz. right half of the pelvis of

X qh. (8034.) Embrithosaurus schwarz. X zy.

(8034.)

the posterior edge of the symphysis lies anteriorly to the edge of the

tuberosity.

Pubis.—The antero-lateral border of the pubis does not continue

the swelling, as shown by the ium just above it. From the ilio-pubic

suture to the pubic tubercle this surface is slightly concave. From

the pubic tubercle a step runs obliquely towards the symphyseal

tongue, but although the pubic tongue extends far forward there is no

notch. Below this step the pubis extends vertically downwards, so

that the downwardly directed pubic bulwark is not roundedly con-

tinuous with the upper true pubic border. As this. bulwark does not

extend to the symphysis there is the appearance of a notch lateral to

the symphyseal pubic tongue. This is, however, solely due to the fact

that the bulwark lies posterior to the step which runs from the pubic

328 Annals of the South African Museum.

tubercle to the tongue. The pubic tongue extends far forward,

thereby producing a great symphyseal length. Its upper surface is

flat, but as the symphysis thickens greatly towards the ischio-pubic

suture its surface is directed more anteriorly than dorsally.

TExt-FIG. 34.—Femur of Hmbrithosaurus schwarzi. x 4. (8034.)

(a) Ventral. (b) Dorsal. (c) Anterior. (d) Proximal.

Text-Fic. 35.—Tibia of Embrithosaurus schwarz. ~X< 4. (8034.)

(a) Dorsal. (b) Ventral. (c) Posterior. (d) Proximal. (e) Distal.

The acetabulum is large, shallow, rounded, and directed laterally.

Femur (text-fig. 34).—The femur is large and massive. The width

across the external trochanter is fairly large, and across the distal

epicondyles large. Proximally the external trochanter is not very

distinctly demarcated, but distally the postaxial edge cuts in sharply.

The flange for the ilio-femoralis is therefore narrow and straight.

The ridge separating the surfaces of insertion of the ilio-femoralis and

Parevasaurian Studies.—Part VI. 329

pubo-ischio-femoralis internus is strong, and distally ends abruptly.

The internal trochanter is not very strong, but prominent and slightly

curved and directed preaxially. The stem of the primitive Y is

strong and directed much medially, but dies out quickly. The shaft

is fairly long, but fairly much constricted. The dorsal curvature is

weak and the proximal surface directed wholly proximally, but the

tibial facets are directed fairly much distally. The intertrochanteric

fossa is of medium depth.

Tibia (text-fig. 35)—The tibia 1s long and massive; the proximal

end is long and very broad; the postaxial facet is smaller than the

preaxial facet; the distal end is long and

narrow ; the eminence on the postaxial border

of the shaft is moderately developed and some-

what dorsally situated.

Fibula (text - fig. 36).— The fibula is of

medium length and it is only a very little

longer than the tibia. It is straight, of

medium strength, with a weak ridge on the

postaxial border of the shaft. The expansion

of the ends is medium. een ie 86 = RibalA

Foot (Pl. XX XVIII, c).—In 9148 there are of Hmbrithosaurus

two complete feet. The dorsal ridge on the SPE) ee 808)

proximal element is sharp. Only the fourth (2) nee

distal is ossified as a pebble-like bone. The

segments of the first digit are not much broader than the second,

third, and fourth.

Genus DoticHoPpaREIA Haughton and Boonstra.

The pelvis and femur are unknown.

Tibia (text-fig. 37).—The tibia is long and massive; the proximal

Text-F1G. 37.—Tibia of Dolichopareia angusta. X 4. (6238.)

(a) Dorsal. (b) Ventral. (c) Posterior. (d) Proximal. (e) Distal.

VOl, XXVIII, PART 3. 23

330 Annals of the South African Museum.

end is very long and very broad; the postaxial facet is smaller than

the preaxial one; the distal end is long and narrow; very weak

indication of eminence on the postaxial border of the shaft.

Fibula.—The fibula is presumably of medium length and presum-

ably only a very little longer than the tibia. It is straight, strong.

with a prominent but not very strong ridge on the postaxial border

of the shaft. The ends are greatly expanded.

Foot.—Unknown.

Genus BRACHYPAREIA Haughton and Boonstra.

Pelvis (text-figs. 38, 39).—The pelvis is small and light. Itis broad

and very low. In anterior view it is broadly U-shaped.

Sacrum.—Unknown.

Ilium.—The ilium is much forwardly directed, with its blade

diverging anteriorly. The antero-ventral edge is slightly everted, but

TExt-FIG. 38.—Lateral view TextT-Fic. 39.—Antero-internal

of the right side of the view of the right half of

pelvis of Brachypareia the pelvis of Brachypareia

rogerst. X 5. (8953.) rogersi. X zy. (8953.)

the whole blade is concave antero-posteriorly. The posterior process

is slightly weathered, but it undoubtedly did not project far. The

blade is short and high, which gives it an upstanding appearance.

The outer surface is moderately concave. The shaft is short and

fairly constricted. The swelling of the anterior border is prominent

and moderately broad, whilst that on the posterior border is weak.

Ischium.—The end of the tuberosity is wanting, but it appears that

the “ shaft’ was moderately long. It is downwardly directed and

parallel to its fellow. At the pubo-ischial suture the symphysis is

moderately thick. In antero-posterior direction the dorsal ischial

surface is practically flat. The medial portion lies almost horizontally.

The posterior brim appears to have been concave and there is a notch

at the symphysis.

Pareivasaurian Studies.—Part VI. 331

Pubis.—The anterior lateral border of the pubis continues the

swelling of the iium. The swelling is broad and prominent, is ven-

trally directed, but ends abruptly in a rounded step. The pubic

tubercle is not differentiated from this swelling. Ventral to the ill-

defined step the pubic border is directed downwards and forwards to

form a forwardly directed bulwark, which is medially continuously

rounded by the upper pubic brim. The bulwark does not reach the

symphysis. A deep and broad notch separates it from the well-

developed, downwardly directed tongue with a flat anteriorly directed

upper surface.

The acetabulum is small, rounded, deep, and slightly downwardly

directed. ,

Femur (text-fig. 40).—The femur is small and light and relatively

long and slender, with a fairly long and much constricted shaft. The

width across the external

trochanter and the distal

condyles is small. The pos-

terior flange for the ilio-

femoralis is fairly broad but

shaft-like. The external tro-

chanter is fairly distinct

from the proximal border.

The internal trochanter is

prominent and curved, with

its distal half turned pre-

axially. The stem of the

primitive Y is fairly strong, Ppxr-rie. 40.—Femur of Brachypareia rogersi.

but dies out before reach- x g- (8953.)

ing the preaxial facet. The (a) Ventral. (0) Dorsal. (c) Anterior.

(d) Proximal.

proximal expanded end is

rotated (in the right femur) in anti-clockwise direction through fully

15° in relation to the distal end. In addition, the proximal surface

is much pre-axially inclined. The dorsal curvature is very great.

This produces a somewhat dorsally directed proximal surface, which

appears like what the incipient stage in the femur of the tortoises

with its dorsally divided facet probably was. The intertrochanteric

fossa is deep.

Tibia (text-fig. 41)—The tibia is very short and very light; the

proximal end is short and narrow; the postaxial facet is nearly equal in

size to the preaxial one; the distal end is long and broad ; the post-

axial border of the shaft shows only a weak indication of a tubercle.

332 Annals of the South African Museum.

Fibula (text-fig. 42).—The fibula is short and only a very little

longer than the tibia. It is curved and weak, with the postaxial

Trext-Fic. 41.—Tibia of Brachypareia rogersi. Text-Fic. 42.—Fibula of

x ¢. (8953.) Brachypareia —_rogersi.

(a2) Dorsal. (b) Ventral. (c) Posterior. x g- (8953.)

(d) Proximal. (e) Distal. (a) Dorsal. (6) Ventral.

border of the shaft with small rugosities. The expansion of the ends

is small.

Foot.—In 5012 the foot is nearly complete. The dorsal ridge on

the proximal element is sharp. Only the fourth distal is present—it

is pebble-like. The elements of the first digit are not broad.

Brachypareva rogersi (Broom).

The above applies to this species. The pelvis and limb of Brachy-

pareva watsont are unknown.

Genus PRopapPus Seeley.

Pelvis.—The pelvis is small and light. It is broad and low or

narrow and fairly high. In anterior view the inner edge is broadly

U-shaped or broadly V-shaped.

Sacrum.—Four vertebrae with ankylosed centra are concerned in

the attachment of the pelvis. In keeping with the lightness of the

pelvis the sacral ribs are light. The first is the strongest. The second

is only slightly less strong. The third and fourth are in comparison

weak.

Ilium.—The ilium is upstanding, and this produces the appearance

of a high pelvis. The blade is parallel to the median vertical plane.

The anterior edge is very strongly everted. The posterior iliac process

is large and projects far posterior to the vertical through the ace-

tabulum. The blade is short and high. The outer surface 1s concave

or concave with a slight medial swelling. The iliac shaft is of medium

Pareiasaurian Studies.—Part VI. S38)

length and moderately or much constricted. The swelling of both the

anterior and posterior border is medium or weak and narrow.

Ischium.—The ischial “ shaft” is short, with the tuberosity turned

upwards so that the two “shafts” run parallel. The symphyseal

thickness is not very great. The medial portion of the ischial plate

is weakly or strongly oblique, thereby producing a broad or fairly

narrow pelvis. The posterior brim appears to be fairly straight,

converging medially so that the ischial tuberosities are more posterior

than the symphysis. .

Pubis.—The antero-lateral border of the pubis has a medium or

weak swelling. The pubic tubercle is indistinct. The pubic border

is rounded continuously into the downwardly and forwardly directed

pubic lip. The pubic rampart is prominent but medially thick or

thin, separated from the slight symphyseal tongue by a lateral notch.

The acetabulum is small, oval or very compressed dorso-ventrally,

fairly deep and outwardly directed.

Femur.—The femur is small, medium, light, and relatively long and

slender, with a long and much constricted shaft. The width across

the greater trochanter and over the distal condyles is small. The

posterior edge of the flange for the ilio-femoralis is narrow and curved

inwards. The external trochanter is distinctly demarcated both

proximally and distally, and situated very much proximally. The

internal trochanter is fairly strong, prominent, curved, with its distal

half turned preaxially. The stem of the original Y is fairly weak

and dies out medially. The dorsal curvature is strong. The proximal

expansion is rotated preaxially in anti-clockwise direction in the right

femur through about 15° in relation to the distal expansion. In

addition, the proximal surface is much preaxially inclined. These

characteristics both conduce to a more upward pose of the limbs.

The intertrochanteric fossa is deep.

Trbia.—The tibia is short and light; the proximal end is short and

narrow ; the postaxial facet is equal in size to the preaxial one; the

distal end is long and broad; the tubercle of the postaxial border of

the shaft is very strong.

Fibula.—The fibula is unknown, as is the foot, with the exception

of the proximal element.

Propappus omocratus Seeley.

Pelvis (text-fig. 43)—The pelvis is fairly broad, rounded posterior

and anterior swellings of the ilio-ischiadic and ilio-pubic borders.

TExtT-FIc. 43.—Lateral view of the left Trxt-Fic. 44.—Dorsal view of

half of the pelvis of Propappus the femur of Propappus

omocratus. xX }. (Type in British omocratus. xX 4. (Type

Museum.) in British Museum.)

TExtT-FIc. 45.—Femur of Propappus omocratus. xX }. (1058.)

(a) Dorsal. (6) Anterior. (c) Ventral. (d) Proximal.

TExtT-FIc. 46.—Tibia of Propappus omocratus. xX }. (1058.)

(a) Dorsal. (6) Ventral. (c) Posterior. (d) Proximal. (e) Distal.

334

Pareiasaurian Studies.—Part VI. 305

The ischial tuberosity is not very strongly turned upwards. The

pubic lip is more or less horizontal and moderately thick. The

acetabulum is oval.

Femur (text-figs. 44, 45).—As for the genus.

Tibia (text-fig. 46).—As for the genus.

Fibula.—Unknown.

Foot.—Unknown.

Propappus parvus Haughton.

Pelvis (text-fig. 47)—The pelvis is fairly narrow; weak, thin,

posterior and anterior swelling of ilio-ischiadic and ilio-pubic borders.

The ischial tuberosity is very strongly turned

upwards. The pubic lip is oblique, with the

posterior half higher than the anterior. The

limb is unknown.

Genus PAREIAsucHUS Broom and Haughton.

Pelvis.—The pelvis is small and its con-

stituent bones light. It is very low but broad.

The breadth is made to appear still greater by

: : : . Text-Fic. 47.— Lateral

its lowness. In anterior view the inner edge ~ View of right half of the

is broadly U-shaped. With the sacral ribs an pelvis of Propappus

: é : : : parvus. X py. (2351.)

situ the opening is oval, with the lateral dia- Mirror image.

meter the greater.

Sacrum.—Four vertebrae are concerned in its attachment, and these

all have their centra coalesced. Concomitant with the lightness of

the pelvis the sacral ribs are also light and comparatively slender.

The first is the strongest, and its articulatory surface to the surface

anterior to the ridge is large. The other three progressively decrease

In size.

Ilium.—The ilium is much forwardly directed, with its blade

diverging anteriorly. The antero-ventral edge is greatly everted.

The posterior iliac process is large and projects far posteriorly, but

the ilium is so greatly forwardly directed that it also lies anterior to

the vertical plane, passing through the centre of the acetabulum. The

iliac blade is long and very low. The outer surface is concave. The

iliac shaft is either long or very long and moderately or very much

constricted. The anterior swelling of the iliac border is strong and

broad. The posterior swelling is only indicated by a prominence at

the ilio-ischial suture.

336 Annals of the South African Museum.

Ischium.—The ischial “ shaft” is fairly long and directed down-

wards, with a strong ridge on the dorso-lateral surface of the tuber-

osity. This gives the appearance of an outwardly diverging ischial

“ shaft.” At the ilio-ischial suture there is a slight prominence. At

the pubo-ischial suture the symphysis is moderately thick. The

dorsal surface of the ischial plate 1s slightly concave in antero-posterior

direction. The medial portion of the ischium is practically flat,

thereby producing the great pelvic breadth. The posterior brim is

straight, z.e. the posterior edge of the symphysis and the tuberosity

are in one plane.

Pubis.—The antero-lateral border of the pubis continues the

swelling of the ilium. The distance from the ilo-pubic suture to the

pubic tubercle is very short and strongly convex. From the pubic

tubercle a step runs slightly obliquely to the symphysis. From this

step the pubis runs downwards but also forwards, so that the pubic

bulwark is not rounded continuously with the upper true pubic border.

This pubic bulwark extends to the symphysis. There is no marked

pubic tongue and no notch lateral to the symphysis. The acetabulum

is of medium size, oval, with its dorso-ventral diameter the smaller,

deep and directed outwardly and downwardly.

Femur.—The femur is light and small in comparison to those of the

genera Bradysaurus, Nochelesaurus, Platyoropha, and Embrithosaurus.

The width across the greater trochanter and the distal condyles is

small. The posterior edge of the flange for the insertion of the ilio-

femoralis is moderately or strongly curved. The greater trochanter

is situated fairly distally. It is proximally continuous with the

proximo-postaxial edge or is distinctly demarcated from it. Distally

it is continuous with the preaxial edge. The internal trochanter is

not very strong, fairly straight, but prominent. The “fourth tro-

chanteroid ”’ ridge is fairly strong, and continues to the preaxial facet.

The dorsal curvature appears to be fairly weak. This may, however,

in part be due to crushing. The proximal surface is moderately pre-

axially inclined. The intertrochanteric fossa is of moderate depth.

The shaft is long and fairly constricted.

Tibia.—The tibia of Pareiasuchus nasicornis is preserved, but too

crushed to describe. The fibula is unknown, and only a few dis-

articulated phalanges of the foot of P. peringueyt.

Pareiasaurian Studies.—Part VI. Sau

Parevasuchus peringueyt Broom and Haughton.

Pareiasuchus nasicornis Haughton and Boonstra.

Pelvis (text-figs. 48-50, 52).—The pelvis is higher, the posterior

iliac process is stronger, the ischial “ shaft” is longer and more con-

TExtT-FIc. 48.—Lateral view of TExtT-FIG. 49.—Posterior view of the

the left side of the pelvis of pelvis and the sacrum of Pareia-

Pareiasuchus peringueyt. suchus peringueyt. X yy. (2337.)

ites (2837!)

Text-FIc. 50.—Anterior view of the pelvis

and sacrum of Pareiasuchus peringueyt.

X 25. (2337.)

Text-Fic. 51.—Femur of Pareiasuchus

peringueyt. Xz. (2337.)

(a) Anterior. (b) Dorsal. (c) Ventral.

(d) Proximal.

TExt-FiG. 52.—Lateral view of the right half

of the pelvis of Pareiasuchus nasicornis.

X zy. (3016.)

338 Annals of the South African Museum.

stricted, and the blade is more diverging in peringueyr than in

nasicornis.

Femur (text-fig. 51).—The description of the femur of the genus is

that of peringueyt ; that of nasicornis is too crushed to compare. It

appears, however, to be shorter than that of peringueyi.

Genus PAREIASAURUS Owen.

Pareiasaurus serridens Owen.

Pelvis (text-fig. 53).—The pelvis is large but not very massive. It

is broad and very high. In anterior view it is broadly U-shaped.

Sacrum.—According to Owen there are six

ankylosed vertebra. Watson maintains there

are four. Although Haughton sketched the

ilium no notes were made of the sacrum.

In comparison with the other species it seems

probable that four vertebrae only are con-

cerned in the sacrum.

Ilium.—The ilium is moderately upstand-

ing, with the blade diverging anteriorly. The

antero-ventral edge is very greatly everted ;

fal 0s eee eee although the edge is weathered it appears

view of the left half of that the posterior process was large and

the pelvis of Parevasawrus qyite prominent. The blade is long and fairly

serridens. X -+y. (Type :

in British Museum.) low. The outer surface is moderately con-

cave. The shaft is long and constricted.

Both the anterior and the posterior swelling are prominent and broad.

Ischvum.—The ischial tuberosity is wanting, but it appears that

the ischial shaft was moderately long and horizontal. At the pubo-

ischial suture the symphysis is moderately thick. In antero-posterior

direction the dorsal ischial surface is moderately concave. The

medial portion is practically horizontal. The posterior brim is slightly

concave, with the tuberosity more posterior than the symphysis.

Pubis.—The anterior border is swollen, and passes without a well-

defined step on to the ventrally and forwardly directed bulwark.

The pubic tubercle is not very distinct. The swelling passes medially

and fades away. The fairly forwardly directed bulwark extends

practically to the symphysis, where a slight notch separates it from

a weak pubic tongue.

The acetabulum is of moderate size, rounded, fairly deep, and

directed outwards and only slightly downwards.

The hind limb is unknown.

Parerasaurian Studies.—Part VI. Y 3e9

From the above detailed account it would appear that—taking the

hind limb into account only—the South African Pareiasaurs fall into

three main groups. In the first group occur the forms Bradysaurus,

Platyoropha, and Dolichopareia. These are the large forms from the

Tapinocephalus Zone. They have a forwardly directed ilium, with

a strong anterior process to the iliac blade; the proximal articulatory

surface of the femur is terminal. The second well-defined group

contains the forms Brachypareia, Propappus, Pareiasuchus, and

Parevasaurus. These are the small forms chiefly from the Endothiodon

and Crstecephalus Zones. One specimen of Brachypareia is a Tapino-

cephalus Zone form. These all have the proximal articulatory

surface of the femur situated somewhat preaxially and dorsally. In

Brachypareia and Propappus parvus the ilium is upstanding, with the

anterior process not prominent. In the others the ilium is forwardly

inclined with a strong anterior process. In the third group occur forms

with narrow pelves and upstanding ilia, viz. Embrithosaurus and

Nochelesaurus. In the former the proximal surface of the femur is

somewhat preaxially inclined. These forms are from the Tapino-

cephalus Zone. We refrain from expressing any views on the phylo-

genetic relations of the various forms within the group. This we

hope to do after the study of the whole skeleton is completed.

MEASUREMENTS.

Nearly all the measurements were taken with Martin’s beam-

compass, and are therefore direct lengths, 7.e. they are either pro-

jections on to a vertical or a horizontal plane. The length from the

post-acetabular buttress on the ischium to the posterior border of the

tuber ischii is termed the ischial length. Similarly, the iliac height is

the distance from the supra-acetabular buttress on the ilium to the

most dorsal point on the iliac blade. The total pelvic height is the

true height measured from the lowest point of the plate to the highest

point of the ilium. The distance from the lowest point of the pelvis

to the dorsal border of the tuber ischii is termed the ischial height.

The iliac length is the greatest antero-posterior length of the iliac

blade. The width between the two tubera ischii is termed the ischial

width, and the iliac width is the greatest distance between the iliac

blades. The distance between the highest and the lowest point on

the symphysis is styled the symphyseal height. The acetabular width

is the distance between the centres of the two acetabuli. This

measurement was taken by Martin’s beam-calipers.

340 Annals of the South African Museum.

Pelvic Measurements.

Pro- |Pareia-

No. 5127/5624/8941) 3718 | 5590 |8953|5002/|8034/2337| 3016 | 2351 ee pecs

cratus.| dens.

Symphyseal length . | 320 | 300 | 235 | 220?) 200 | 186 | 290 | 210/155/190 | .. | 165 | 180

Ischial length . . | 195 | 160 | 122} 160 | 160 | 105}180/1385)120; 95?) 85 | 120 Be

Iliac length . . | 290 | 307 | 215 | 290 | 230?) 160 | 278 | 255 | 230 | 225 | 14a 22a

Acetabular length . | 160 | 185/140 | 160 | 145 | 110/170 | 160 | 100 | 120 86 | 140 | 125

Width of acetabulum | 155 |110/115|140 | 145 75 |150|140] 85] 75 50 | UTS ase

Total pelvic height . | 467 | 405 | 277 | 380 | 585 | 260 | 450 | 530 | 285 | 265 | 2757?) 410 | 429

Ischial height . . |100|}140} 90] 90 |190 | 95) 160|146; 60} 55 | 702} 90 | 120

Iliac height. . | 265 | 235 | 160 | 245 | 315 | 160 | 285 | 255 | 237 | 145 | 165 | 235 | 184

Ischial width . . | 353 | 325 | 360 | 500 2] 265 | 250 | 390 | 350 | 320 | 250 | .. i se

Iliac width . . | 562 | 675 | 620 | 520 | 370 | 380 | 600 | 560 | 550 | 500

Symphyseal height . | 80} 90) 82] 90 |118 794 OO) si aul sO Sales

Interacetabular width | 280 | 340 | 350} .. | 220 | 240 | 320 | 300 | 300 | 240

Vote.—For specific names of specimens see list on pp. 311 and 312.

Femoral Measurements.

Width Width Length Width Min

: over over of over Zap

Name. No. oe Ext. Prox. Prox. Dist. ee

8-| Tro- | Articula- | Articula-| Articula- Shaft

chanter. tion. tion. tion.

Bradysaurus baint a |e 450 245 120 195 180 125

am seeleyt . | 5624 440 240 100 205 180 134

x vanderbyli . | 9169 350 180 ie 155 145 102

Nochelesaurus strubeni . | 5590 417 200 80 160 155 115

Bs alexanderi | 6239 417 200 80 155 135 125

Brachypareia rogersi —.. |: 8953 280 153 75 110 110 ' 65

ee = 5 | a0 295 150? 60 105 105 77

Platyoropha broomi . | 5002 412 200 104 152 160 120

Embrithosaurus schwarz | 8034 414 222, 100 180 175 107

Pareiasuchus peringueyt | 2337 297 165 58 125 125 75

As nasicornis 3016 280 176 63 1a, 120 95

Propappus omocratus .| type 350 140 Bs 116 120 68

- . | 1058 SoDN eG Toes Ne Slee 65

Tibial Measurements.

. . : Distal

Aen Proximal Articulation. ee an

= = Max. | Width

Name. No. | tTength.| of Length | Length

Shaft. |Breadth.| Length. eae eee Breadth.} Length.

Facet. | Facet. :

Bradysaurus baini ee! Par 245 85 120 187 80 100 80 150

a seeleyt : 5624 284 72 123 aa 65 102 95 140

Z vanderbyli . 9169 230 65 90 ESF 58 82 64 107

Nochelesaurus strubeni . 5590 260 64 102 (ay 79 87 70 128

se alexanderi | 6239 250 TE 102 156 75 90 80 135

Brachypareia rogersi . | 8953 179 50 78 110 53 53 60 90

Dolichopareia angusta . 6238 265 70 126 178 88 110 87 140

Embrithosaurus schwarz | 8034 257 58 122 150 65 85 Die 90

Propappus omocratus .| 1058 205 52 87 110 56 56 62 120

Pareiasaurian Studies.—Part VI. 341

Fibular Measurements.

Proximal End. Distal End.

: Max.

Name. No. eneehe

Length. | Width. | Length. | Width.

Bradysaurus baini 5s ey 295 75 50 90 63

a seeleyt 2 5624. 300 80 45 100 60

a vanderbyli . 9169 220 62 S57 68 47

Nochelesaurus strubeni . 5590 He V5) 43 Sec a

Brachypareia rogersi x 8953 200 60 33 62 4]

Dolichopareia angusta . 6238 Le 72 40 95 60

Embrithosaurus schwarz 8034 260 70 38 80 56

COMPARATIVE OSTEOLOGY.

A. Pelvis.—Little is known of the pelvis of the Osteolepidae (32).

It seems reasonable to assume that it was plate-like without a dorsal

iliac extension, as is generally the case among fishes, with the

exception of Dipnor.

In the rhachitomous Amphibia (35) (e.g. Eryops, Trematops, etc.)

the ilium is a vertically directed rod without either anterior or posterior

processes. The dorsal portion of the external iliac surface served as

surface of attachment for axial muscles. The pubis is long and so

is the ischium. The plate is broad and flat, and only perforated by

a pubic foramen. The acetabulum is directed outwards. There are

one or two sacral vertebrae. The main ilio-sacral suspension is supra-

acetabular.

The embolomerous Amphibia (32) (Hogyrinus), which are more

closely connected with the reptilian ancestry, have a backwardly

inclined ilium with a very long though narrow posterior process.

The anterior process is directed more dorsally than anteriorly. As

in the Rhachitomi, the dorsal portion of the external iliac surface was

occupied by axial muscles. The pubis is short, but the ischium is

long. The plate is broad and flat, with only a pubic foramen. The

acetabulum is directed outwards. There are one to three sacral

ribs. The main ilio-sacral suspension is post-acetabular, and is

ligamentous. |

The somewhat heterogeneous assembly grouped under the order

Cotylosauria (12) is best considered in some detail.

In the Seymouriamor pha (Seymouria, Kotlessia) the ilium is vertically

342 Annals of the South African Museum.

directed. Seymouria (33) has no anterior process, and Koélassva (1)

only a slightly projecting anterior process. In both the posterior

process is long and narrow. In Kotlassia (and probably also in

Seymouria) there is a dorsal keel on the outer surface of the ilium.

The ilio-femoralis does not therefore extend to the dorsal border of

the iliac blade. In Seymourza the pubis is short, whereas in Kotlassia

it is fairly long. In both the ischium is long, and the plate is broad

and flat and only perforated by a pubic foramen. The acetabulum

is directed downwards and outwards. In Seymouria there is one

sacral vertebra and a subsidiary pre-sacral vertebra with a rib, whereas

in Kotlassia there are two—the first having the shorter rb. In

Kotlassia the main iliac-sacral suspension is supra-acetabular, and in

Seymourva it is slightly post-acetabular. In contradistinction to the

embolomerous Amphibia the sacro-iliac articulation is firm.

In the Diadectomorpha, comprising the Diadectidae, Parerasauridae,

and Procolophonidae, the pelves are very different. In the Diadectidae

both Diadectes and Diasparactus (13) have vertical iia. Diasparactus

has a strong anterior process to its ilium, whereas in Diadectes it is

very slightly developed. In both the posterior process is long and

strong. Both have a dorsal shelf on the ilium, so that the ilio-

femoralis does not reach the dorsal border of the ilium. The pubis

is long in Deadectes and short in Diasparactus. In both the ischium

is long. The pubo-ischiadic plate is broad and flat in Diadectes with

a not prominent carina, whereas in Diasparactus the two sides of the

plate meet at an angle and the carina is prominent. The only per-

foration 1s a pubic foramen. The acetabulum is directed outwards.

In Diadectes there are two coalesced sacral vertebrae—the rib of the

first is the stronger; in Duasparactus there are two non-coalesced

sacrals—the rib of the first is also the stronger. In both the main

sacro-illac suspension 1s situated high up and is supra-acetabular.

In the Paretasauridae the ilia are practically vertical in Nochele-

saurus, Brachypareia, andin Propappus ; slightly forward in Embritho-

saurus; and forward in Bradysaurus, Platyoropha, Pareiasaurus,

“ Parevasaurus ” karpinsku, and Parevasuchus. The anterior process

projects moderately in Nochelesaurus, Brachypareia, and Propappus ;

strongly in Hmbrithosaurus ; and very strongly in the other genera.

The posterior process 1s very short but strong in EHmbrithosaurus ;

short and strong in Bradysaurus and Brachypareia; medium and

strong in Nochelesaurus, “ Parevasaurus ” karpinskw, and Platyoropha ;

medium but only moderately strong in Propappus, Pareiasaurus, and

Pareiasuchus. In all the ilio-femoralis extends to the dorsal iliac

Pareiasaurian Studies.—Part VI. 343

edge. The pubis is short and the ischium of medium length in all

the genera. The two sides of the pubo-ischiadic plate meet at an

angle in Nochelesaurus and Embrithosaurus ; in the others the plate

is broad and flat. The only perforation is the pubic foramen. The

acetabulum is directed more outwardly in Nochelesaurus and Embritho-

saurus, and more downwardly in the rest. The number of coalesced

sacral vertebrae varies from two to four; in the same genus—Brady-

saurus—baint has two, seeleys three, and vanderbylc four. The first

sacral rib is strongest in Nochelesaurus, Embrithosaurus, Pareiasaurus,

Propappus, Parevasuchus, and Bradysaurus vanderbyli ; the second is

strongest in the two other species of Bradysaurus and Platyoropha.

The main ilo-sacral suspension is practically supra-acetabular in

Nochelesaurus, Embrithosaurus, Propappus, and Brachypareia ; pre-

acetabular in the rest.

In the Procolophonidae (4, 20, 19) the pelvic girdle of Procolophon,

Teleperton, and Koiloskiosaurus is very similar. The ilium is directed

very slightly forward. It is rod-like with very slight anterior and

posterior projections. The ilio-femoralis extends to the dorsal edge.

The pubis is short; the ischium long. The plate is broad and flat,

with a large pubic foramen. The acetabulum is directed downwards

and outwards. There are two coalesced sacrals with the second rib

shghtly stronger than the first. The main sacro-iliac suspension is

only slightly pre-acetabular.

In the Captorhinomorpha, comprising the Captorhinidae and Limno-

scelidae, there are differences in the pelvis. In the Captorhinidae the

ilium is inclined backwards in both Labidosaurus and Captorhinus.

There is no anterior process to the ilium ; the posterior process is long

and strong. The ilio-femoralis extends to the dorsal border. The

pubis is short in Labidosaurus and long in Captorhinus, whereas the

ischium is long in both. The pubo-ischiadic plate is broad and flat,

and is perforated only by a pubic foramen. The acetabulum is

directed outwards. There are two non-coalesced sacrals; the first

rib is the stronger. The main sacro-iliac suspension is high and post-

acetabular. In the Limmnoscelidae (36) the ilium is vertical, with no

anterior process but with a long narrow posterior process. There is

a dorsal shelf on the ilium, so that the ilio-femoralis does not extend

to the dorsal edge. The pubisis short; theischium long. The plate

is broad and flat, with a not prominent carina. The pubic foramen

is the only perforation. The acetabulum is directed outwards.

There is only one sacral vertebra. The sacro-iliac suspension 1s low

and supra-acetabular. Considering the pelvis, Lamnoscelis more

pr eter =

- 7 =

344 Annals of the South African Museum.

nearly approaches the Seymourzamorpha and the Diadectidae than it

does the Captorhinidae, with which Watson classes it.

In the Pelycosauria (11) the pelvis varies considerably. In the

Caseidae (36) the ilium is forwardly directed. The anterior process

of the iliac blade is strong; the posterior of medium strength. The

ilio-femoralis doubtfully extends to the upper border. The pubis is

long; the ischium moderate. The plate is apparently broad and

flat, and is perforated by a small pubo-ischiadic fenestra as well as

by a pubic (obturator nerve) foramen. The acetabulum is directed

downwards and outwards. There are three non-coalesced sacrals.

The ribs of the first two are equally strong, but all three are only

ligamentously attached to the ilium. The main suspension is slightly

pre-acetabular.

In the Ophiacodontidae (13) the ilium is inclined backwards. There

is no anterior process, but a very long and narrow posterior process.

There is a shelf on the external surface, so that the ilio-femoralis does

not extend to the dorsal edge. The pubis is very long and the ischium

long. The plate is broad and flat, and is perforated by a pubo-

ischiadic foramen as well as a pubic foramen. The acetabulum is

directed outwards. There are two non-coalesced sacrals. The rib of

the first is the stronger and the main ilio-sacral suspension 1s slightly

post-acetabular.

In the Poliosauridae (Varanosaurus (36), Theropleura) the ilium is

directed backwards. There is no anterior process, but the posterior

process is long and narrow. The ilio-femoralis does not reach the

dorsal iliac border. The pubis is very long; the ischium long. The

plate is broad and flat (broader anteriorly than posteriorly), and is

perforated by a pubo-ischiadic fenestra and a pubic foramen. The

acetabulum is directed outwards. There are two non-coalesced

sacrals ; the rib of the first is the stronger. The ilio-sacral suspension

is post-acetabular.

In the Clepsydropidae the ilium is directed slightly backwards. The

anterior process is not prominent in Clepsydrops and only moderate

in Dimetrodon. In both the posterior process is long and narrow.

The ilio-femoralis extends to the dorsal edge of the ilium. The pubis

is very long; the ischium long. The sides of the plate meet at an

angle, producing a high and narrow cavity and a deep symphysis. The

plate is pierced by a pubic (obturator) foramen. The acetabulum is

directed outwardly. There are three coalesced sacrals. The first

and second ribs are stronger than the third. The ilio-sacral suspen-

sion is post-acetabular,

Pareiasaurian Studies.—Part VI. 345

In the Edaphosauridae the ilium is directed backwards in Naosaurus.

There is a moderate anterior but a strong posterior process to the

iliac blade. It is uncertain whether the ilio-femoralis extends to the

dorsal edge or not. The pubis is very long; the ischium long. The

plate is apparently broad and flat, and pierced by a pubic foramen.

The acetabulum is directed outwards. There are two non-coalesced

sacrals. The rib of the first is the stronger and the ilio-sacral sus-

pension is slightly post-acetabular.

In the Dinocephalia there are some differences between the various

members of the Tapinocephalidae. The ilium is vertical. The

anterior process is moderately strong in Moschops; long but not

strong in Struthiocephalus ; and the ischium in the former is long, in

the latter short. In both the plate is broad and flat. The only

perforation is the pubic foramen. The symphysis is much weaker

than in the Cotylosaurs. The acetabulum is directed downwards

and outward in Moschops; outward in Struthiocephalus. There are

three coalesced sacral vertebra—the rib of the first is the strongest.

The main ilio-sacral suspension is supra-acetabular.

In the Anomodontia the Lystrosaurian (34) pelvis is adapted to

aquatic habits. In the Dicynodontia the ilium is directed forwards.

The anterior process is long and strong. The ilio-femoralis extends

to the dorsal edge. The pubisis short; theischium long. The plate

is broad and flat. It is perforated by a pubic foramen only. A

cartilaginous prepubis was probably present. The acetabulum is

directed outwards. The symphysis is fairly weak. The ilo-sacral

suspension is pre-acetabular. |

In the Therocephalia the pelvis is not very well known. In Choero-

saurus (17) the ilium is directed forwards. The anterior process is

very long and very strong. The posterior iliac process is strong. The

pubis is short ; the ischium long. The pubo-ischiadic plate is broad

and flat, and is perforated by a pubo-ischiadic fenestra and by an

obturator foramen, which lies between the pubis and ischium and not

through the pubis only. A cartilaginous prepubis was probably

present. The acetabulum is directed outwards. The symphysis is

fairly weak. The ilo-sacral suspension is pre-acetabular.

In the Gorgonopsia the pelvis is imperfectly known. In Lycaenops

(17) the ilium is vertical. Both the anterior and posterior processes

are strong but not long. The pubis and ischium are long. The plate

is apparently broad and flat, and pierced by a large foramen appar-

ently through the pubis alone. The ilio-sacral suspension is supra-

acetabular. The acetabulum is directed outwards.

VOL. XXVIITI, PART 3. 24

346 Annals of the South African Museum.

In the Cynodontia the pelvis is also not very well known. In

Cynognathus (30) and Diademodon (9) the ilium is vertical, with a very

strong anterior process and a very strong and long posterior process.

The gluteal muscles extended to the dorsal iliac edge. The pubis is

short in Drademodon and medium in Cynognathus. In both the

ischium is very long. The plate is broad and flat, and is perforated

by a median pubo-ischiadic fenestra and an obturator foramen lying

between the ischium and pubis. A cartilaginous prepubis was present

in life. The acetabulum is directed outwards. In Cynognathus there

are four sacrals—the first is probably a modified lumbar vertebra.

The ilio-sacral suspension is pre-acetabular. The symphysis is weak.

In the Monotremes the ilium is directed forwards with a very long

and very strong anterior process. There is no posterior process. The

gluteals extend to the dorsal edge of the ihac blade. The pubis is

short ; the ischium is of moderate length. The plate is broad and

flat, and is perforated by an obturator foramen lying between ischium

and pubis. A prepubis is present. The acetabulum is directed out-

wards. There are three sacralsin Echidna and two in Ormithorhynchus.

The ilio-sacral suspension is pre-acetabular.

In a generalised Mammalian scheme the ilium is directed forwards.

The anterior process is strong ; the posterior also fairly strong, and this

receives the sacral ribs. The ilium diverges strongly anteriorly. The

gluteals extend to the dorsal edges. The pubis is very short; the

ischium long. The symphysis is weak. The plate is broad and flat,

perforated by a large obturator foramen. There is no prepubis or

cartilaginous edge. The acetabulum is directed downwards and out-

wards. The ilio-sacral suspension is acetabular.

B. Femur.—Unfortunately the femur of the embolomerous

Amphibia is not very well known. In Diplovertebron and EHogyrinus

(32) the femur is of medium size and fairly slender. The shaft is

long and constricted. Both the proximal and distal ends are not

greatly expanded. The articulatory faces are proximal and distal

respectively. The fibular facet is lateral. Distally the dorsal and

ventral grooves separating the condyles are shallow. Watson thinks

that the adductor crest was probably not well marked.

In the Rhachitomi the femur is well known in Trematops, Eryops,

Lydekkerina, Rhinesuchus (18), and others. The bone is long and

fairly slender. The shaft is long and constricted, with hardly any

ilio-femoral flange in Eryops and a moderate flange in Trematops.

Both the proximal and distal ends are greatly expanded. The proxi-

mal articulatory surface is terminal; the distal semi-ventral. The

Parevasaurian Studies.—Part VI. 347

fibular facet is semi-terminal. Distally the intercondylar grooves are

not deep. The adductor ridges form an intact Y. The anterior

branch has a strong, preaxially directed “internal trochanter”; the

posterior branch has an indistinct or slightly differentiated “‘ external

trochanter”’; at their junction there is a strong eminence for the

coccygeo-femoralis; the stem of the Y is strong and medially

directed.

In the Seymouriamorpha the femur is short and massive. The

“ shaft’ is short, constricted, but obscured by the development of

an ilio-femoral flange. Both the ends are greatly expanded. The

proximal articulatory surface is terminal; the distal semi-ventral in

Kotlassia, but less so in Seymouria. The fibular facet is terminal in

Seymouria ; lateral in Kotlassia. The intercondylar sulcus is deep

in Kotlassia but apparently very shallow in Seymouria. The adductor

ridges form an intact Y. The anterior branch has a strong and

prominent preaxially directed ‘“‘ internal trochanter ”’ ; the posterior

branch has no distinct “ external trochanter” separated from the

posterior margin of the bone; at their junction there is a moderate

eminence for the coccygeo-femoralis; the stem is strong, running

obliquely to the tibial condyle.

In the Diadectomorpha the Diadectidae have a short and massive

femur. The shaft is short and constricted, slightly obscured by an

ilio-femoral flange in Diasparactus. Both ends are greatly expanded.

The proximal articulatory surface is terminal ; the distal semi-ventral.

The fibular facet is lateral. The ventral intercondylar sulcus is very

shallow. The adductor ridges form an intact Y in Dradectes, but in

Diasparactus the posterior branch is cut off. The anterior branch

has a strong, prominent, and postaxially directed “ internal

trochanter’; the posterior branch has no distinct “ external tro-

chanter”’; at their junction the coccygeo-femoralis is moderate in

Diadectes, strong in Diasparactus ; the stem is strong, running obliquely

to the tibial condyle.

In the Parewasauridae the femur is fairly short and massive in

Bradysaurus, Platyoropha, Nochelesaurus, “‘ Parevasaurus” karpinska,

and Embrithosaurus ; short and moderately stout in Brachypareia,

Propappus, and Pareiasuchus. The shaft is short and constricted,

but obscured by an ilio-femoral flange. Both ends are greatly ex-

panded. The proximal surface is terminal in Bradysaurus, Platy-

oropha, Nochelesaurus, and Embrithosaurus ; slightly dorsally and pre-

axially directed in Brachypareia, Propappus, and Pareiasuchus ; the

distal articulatory surfaces are semi-ventral. The fibular facet is

348 Annals of the.South African Museum.

lateral. The ventral intercondylar sulcus is deep. The adductor

ridges form a broken-up Y. The anterior branch has a strong, promi-

nent, sometimes sigmoidally curved

“internal trochanter,” directed vent-

rally and slightly preaxially; the pos-

terior branch is cut off, but only forms a

very indistinct “ external trochanter ”’ ;

at their junction there is no coccygeo-

femoralis eminence; the stem is a con-

tinuation of the anterior branch; it is

weak, and either dies out quickly or

nearly reaches the tibial condyle.

In the Procolophonidae the femur is

long and slender. The shaft is long and

sigmoidally curved — the antero-dorsal

surface being concave. The ends are

not greatly expanded. The proximal

articulatory surface is terminal, but ap-

Tuxt-r1c. 54.—Dorsal view of Peats to have a neck because of the cur-

the femur of “ Pareiasaurus” vature of the shaft; the distal surface

se Sy ECG is terminal. The fibular facet is terminal.

Ventrally the popliteal fossa is deep.

The adductor ridges do not form a Y. Anteriorly there is a promi-

nent internal trochanter ; posteriorly there appears to be no external

trochanter ; there is no eminence for a coccygeo-femoralis ; the stem

of the primitive Y is absent.

In the Captorhinomorpha the femora of the Captorhinidae vary. In

Captorhinus the femur is long and fairly massive ; in Labidosaurus it

is short and massive. In both the shaft is constricted and unobscured

by an ilio-femoral flange. The ends are greatly expanded. The

proximal articulatory surface is terminal; the distal semi-ventral.

The fibular facet is terminal. The intercondylar sulcus is weak in

Labidosaurus ; fairly distinct in Captorhinus. The adductor ridges

form an intact Y. The anterior branch has a strong, prominent,

preaxially directed, “internal trochanter”; the posterior branch

has an “ external trochanter,” not distinct from the posterior border ;

the stem is prominent, but partially separated from the anterior and

wholly from the posterior branch in Labidosaurus, continuing with

the anterior branch—Cuaptorhinus. In neither is there a distinct

coccygeo-femoralis eminence.

In the Lamnoscelidae the femur is short and massive. The con-

Pareiasaurian Studies.—Part VI... 349

striction of the shaft is not obscured by an ilio-femoral flange. The

ends are greatly expanded. The proximal and distal articulatory

surfaces are terminal. The fibular facet is terminal. The inter-

condylar sulcus is deep. The adductor ridges form an intact Y; the

anterior branch has a strong, prominent, preaxially directed, “‘ internal

trochanter”; the posterior branch has an “ external trochanter,”

not distinct from the posterior edge; at their junction there is no

coccygeo-femoral eminence; the stem runs obliquely towards the

tibial condyle.

In the Pelycosauria the femur in the Casezdae is of medium length

and stoutness. The shaft is long and constricted, and not obscured

by an ilio-femoral flange. The ends are greatly expanded. The

proximal articulatory surface is terminal; the distal surfaces are

semi-ventral. The fibular facet is terminal. The intercondylar

sulcus is deep. The adductor ridges form a nearly intact Y. The

anterior branch has a strong, preaxially directed, “internal tro-

chanter’”’; on the posterior branch the “ external trochanter ”’ is not

distinct from the posterior border; the posterior branch only in-

distinctly joins the anterior and there is no coccygeo-femoralis

eminence; the stem is a continuation of the anterior branch, is

strong, and runs obliquely to the fibular condyle.

In the Ophiacodontidae the femur is large and massive. The shaft

is fairly long and constricted ; it is hardly obscured by an ilio-femoral

flange. The ends are greatly expanded. The proximal and distal

articulatory surfaces are terminal. The fibular facet is terminal.

The intercondylar sulcus is deep. The adductor ridges form an

intact Y; the anterior branch has a strong, prominent, preaxially

directed, “internal trochanter”; the posterior branch has a fairly

strong and distinct “ external trochanter”; at their junction there

is no coccygeo-femoralis eminence ; the stem is a strong ridge running

to the fibular condyle, or a continuation of the posterior branch.

In the Polrosauridae the femur is large and massive. In Thero-

pleura the shaft is of medium length and constriction; in Varano-

saurus it is long and constricted ; in neither is it obscured by an ilio-

femoral flange. The ends are fairly greatly expanded. The proximal

articulation is terminal but also slightly preaxial in Theropleura ;

the distal is semi-ventral. The fibular facet is terminal. The inter-

condylar sulcus is deep. The adductor ridges form an intact Y ; the

anterior branch has a strong, prominent, preaxially inclined, “ internal

trochanter’; the posterior branch has an “ external trochanter ”

not distinct from the posterior border; at their junction there is an

——S *

350 Annals of the South African Museum.

indistinct coccygeo-femoralis eminence; the stem is a fairly strong

ridge running to the postaxial condyle. In Varanosaurus it is a con-

tinuation of the anterior, in Theropleura of the posterior branch.

In the Clepsydropidae the femur is long and fairly slender in Clepsy-

drops but massive in Dimetrodon. The shaft is long and constricted

in both, and is somewhat obscured by a flange for the ilio-femoralis

and femoro-tibialis. The ends are greatly expanded. The proximal

articulatory surface is terminal, but slightly preaxial in Clepsydrops ;

the distal is semi-ventral. The fibular facet is terminal. The inter-

condylar sulcus is deep. The adductor ridges form a practically

intact Y; the anterior branch has a strong, prominent, preaxially

inclined, “‘ internal trochanter”’; the posterior branch has an “ ex-

ternal trochanter” not very distinct from the posterior edge; at

their junction there is a strong coccygeo-femoralis eminence; the

stem running from the anterior branch forms an indistinct ridge.

In the Edaphosauridae the femur is large, but fairly slender in

Naosaurus. The shaft is quite long and constricted, and is unobscured

by an ilo-femoral flange. The ends are greatly expanded. Both

the proximal and distal articulatory surfaces are terminal. The

fibular facet is terminal. The intercondylar sulcus is deep. The

adductor ridges form an intact and strong Y; the anterior branch

has a strong, prominent, preaxially inclined, “ internal trochanter ”’ ;

the posterior branch has an “ external trochanter ” not distinct from

the posterior border ; at their junction there is a strong eminence for

the coccygeo-femoralis ; the stem is a strong ridge running obliquely

to the fibular condyle from the anterior branch.

In the Dinocephalia the femur is, in general, short but large and

massive (Moschops, Struthiocephalus, Tapinocephalus, etc.). Recently

we obtained from the Gouph a number of limb-bones associated with

Moschopid skulls that are long and very slender. The femora, in

fact, closely approach those of the Dinosaur Plateosaurus cullingwortht.

These will be treated of elsewhere. In the more typical Dinocephalia

the shaft is short and moderately constricted, and is obscured by the

ilio-femoralis flange. The ends are greatly expanded. The proximal

articulation is terminal, but through the curvature of the preaxial

surface it is also slightly preaxially inclined. The distal surfaces are

semi-ventral. The fibular facet is semi-lateral. The intercondylar

sulcus is deep. The primitive Y system of adductor ridges is broken

up and the intertrochanteric fossa is much shallower than in the

Rhachittomi, Cotylosauria, and Pelycosauria. The remnant of the

original anterior branch is a strong, ventrally directed, internal tro-

Pareiasaurian Studies.—Part VI. 351

chanter ; the posterior branch is not yet developed into an external

trochanter distinct from the posterior border of the bone; at the

place of their original junction there is in Moschops a slight eminence

for the coccygeo-femoralis ; the stem of the primitive Y is represented

by a ridge running to the tibial condyle.

In the Anomodontia the Lystrosauridae have a femur with carti-

laginous ends and weak adductor insertions as a result of aquatic

adaptations. In the Dicynodontidae the femur is, proportionately

to the body, large and massive. The shaft is constricted, but ob-

scured by an ilio-femoral flange. The ends are greatly expanded.

The proximal surface is terminal, but through the curvature of the

preaxial surface also slightly preaxially inclined. The fibular facet

is terminal. The distal surfaces are terminal. The intercondylar

sulcus is weak with the condyles not well separated. The primitive

Y system of adductor ridges is broken up. The internal trochanter

is the remnant of the anterior branch and is fairly strong and promi-

nent, and is preaxially inclined ; the external trochanter is represented

by the rugose posterior edge ; the intertrochanteric fossa is shallow ;

there is no coccygeo-femoralis eminence and no remnant of the stem

of the Y.

The Therocephalian femur is very imperfectly known. In Choero-

saurus it is long and slender. The shaft was probably long and

slender, and unobscured by a gluteal flange. The proximal end is

expanded. Its articulatory surface is terminal, but because of the

curvature of the anterior surface is also slightly preaxial. The ex-

ternal trochanter is a not very prominent ridge on the dorso-posterior

surface. The internal trochanter is more prominent than the

external.

The Cynodont femur is not well known. In Diademodon it is a long

and slender bone, with a long constricted shaft unobscured by a gluteal

flange—the gluteals being inserted on the external trochanter. The

proximal end is much expanded in Diademodon and Cynognathus.

Through the curvature of the preaxial surface of the bone the head

is fairly much preaxially inclined, although there is no neck. In

Diademodon the distal end was probably moderately expanded, with

its articulatory surfaces terminal and with a moderate intercondylar

sulcus. The external trochanter is a prominent ridge on the post-

axial corner of the head; on the upper third of the mid-ventral

surface of the bone in the internal trochanter is developed as a promi-

nent ridge.

In the Monotremes the femur of Echidna is short and flattened,

302 Annals of the South African Museum.

with a constricted shaft unobscured by a gluteal flange. .The ends

are greatly expanded. The proximal articulatory surface is terminal

but situated on a neck; the distal is terminal, with a deep inter-

condylar sulcus. The external trochanter on the dorso-posterior

surface is not very great and the internal trochanter is directed

inwards.

In the general Mammalian condition the external trochanter is on

the dorsal posterior surface and the internal trochanter on the ventral

anterior surface. The proximal head is carried on a distinct neck.

The shaft is long and slender, and bears the lateral lip of the linea

aspera, which represents the stem of the primitive Y.

C. Epipodial—The epipodial of the Embolomeri is not well known.

In Diplovertebron both the bones are short and stout with expanded

ends. The tibiais the more massive bone. The fibula is articulated

with the distal end of the femur. Nothing is known about the

cnemial crest.

In the Rhachitomi the epipodial is short and stout, with expanded

ends. The tibia is the more massive bone, with its proximal end more

expanded than the distal. The reverse obtains in the fibula. The

fibula appears to articulate with the distal end of the femur. The

cnemial crest is moderate, with a medium groove postaxially.

In the American Cotylosaurs the epipodial is short and massive.

The tibia is the more massive, with a greatly expanded bipartite proxi-

mal surface. The cnemial crest is strong and the postaxial cnemial

groove is deep. The fibula is fairly massive, with a great distal

expansion. The fibular articulation is terminal in all except the

Diadectidae. The spatium interosseum is great.

In the South African and Russian Pareiasaurs the epipodial is

short and quite massive. The tibia, though relatively slightly longer,

is very similar to that of the American Cotylosaurs. The fibula is

less massive, but the spatium interosseum is equally great. The

fibular articulation is lateral.

In the Procolophonidae the epipodial is long and slender. The

cnemial crest is weak, with practically no groove. The spatium

interosseum is not great. The fibular articulation is terminal.

In the Pelycosauria the epipodial is very similar to that of the

Pareiasaurs. On the whole it is longer and more slender, especially

in Clepsydrops and Varanosaurus. The cnemial crest and groove are

well developed, as in the Pareiasaurs. The spatium interosseum

appears narrower. The fibula has a lateral articulation with the

femur,

Pareiasaurian Studies.—Part VI. 353

In the Dinocephalia the epipodial is shorter than and as massive as

in the Pareiasaurs. The cnemial crest is weaker and the groove very

shallow. The proximal end of the tibia is not so clearly bipartite.

The proximal end of the fibula is more expanded and its articulation

semi-lateral.

In the Dicynodontia the epipodial is longer and more slender than

in the Pareiasaurs. The proximal end of the tibia is only moderately

expanded and not bipartite. The cnemial crest is weak and there is

practically no groove. The fibula is very weak and straight, so that

the spatium interosseum is narrow. Its proximal articulation is

terminal.

In the Cynodontia the epipodial is imperfectly known. It appears

to be long and slender. The cnemial crest is long and has a sharp

edge. The groove is absent. The fibula is very slender and straight,

and its proximal articulation terminal.

Little is known of the epipodial of either the Therocephalia or

Gorgonopsia.

In the Monotremes the epipodial is long and slender, with a sharp

cnemial crest but no groove on the tibia. The fibula is straight.

In the Mammals the epipodial is long, straight, and slender. The

tibia has a long, sharp cnemial crest with no groove. The fibula is

weak.

D. Foot.—In Trematops the tarsal formula is 3, 1, 3, 5; the

phalangeal formula probably 2, 3, 4, 4, 2. The foot is long and

fairly broad. The third digit was the longest.

In Limnoscelis the tarsal formula is 2, ?, ?, ?; the phalangeal

formula is 2, 3, 4,5, 4. The footis long and broad. The fourth digit

is the longest. The fifth is also long.

In Seymoura the tarsal formula probably was 2, 0, 1, 5, and the

phalangeal 2, 3, 4, 5, 4. The foot was probably long and fairly

narrow, with the fourth toe the longest. The fifth was also probably

long.

The tarsal formula of Diasparactus apparently was 2, 0, 0, 4, and

the phalangeal 2, 3, 4, 5, 4. The foot was short and broad. The

fourth digit was the longest.

In Diadectes the probable tarsal and phalangeal formulae were

2, 0, 0, 5 and 2, 3, 4,5, 4.. The foot was broad and short. The fourth

digit probably was the longest.

In Captorhinus and Labidosaurus the tarsal formula appears to

have been 2, 0, 1, 5, and the phalangeal 2, 3, 4,5, 4. The foot is long

and fairly narrow. The fourth digit was probably the longest.

354 Annals of the South African Museum.

In Casea the tarsal formula was 2, 0, 1, 5, and the phalangeal

probably 2, 3, 4, 5, 4. The foot was fairly narrow and long. The

fourth digit was the longest. The fifth also long.

In Ophiacodon the tarsal and phalangeal formulae were 2, 0, 2, 5

and 2, 3, 4, 5, 4. The foot was broad and fairly long. The fourth

digit was the longest. The fifth also long.

In Varanosaurus the tarsal and phalangeal formulae were 2, 0, 0, 5

and 2, 3, 4, 5, 4. The foot was narrow and long. The fourth digit

longest ; the fifth short.

In Clepsydrops the tarsal formula was 2, 0, 2, 4. The phalanges

have been restored to produce a long and slender foot.

In the Dicynodontia the tarsal formula was 2, 0, 1, 4, and the phalan-

geal 2, 3, 3, 3, 2.

MYOLOGY.

As the Pareiasaurian pelvis is immovably fixed to the sacral verte-

brae by the sacral ribs there are no muscles, as in the pectoral girdle,

to suspend it. The muscles attached to the pelvis are of two classes

—the axial and appendicular. Both these subdivisions consist of a

dorsal and ventral mass.

I. AxtaL Muscuss.

A. Dorsal Group.

This mass of muscle consists of the median spinalis and semi-

spinalis, and the more lateral longissimus dorsi and _ilio-costalis,

which extend to the tail as an extensor caudae. The median spinalis

and semispinalis are not attached to the pelvis, but serve as sub-

sidiary locomotor muscles by altering the curvature of the vertebral

column. There is some difficulty in determining the area of origin

of the anterior longissimus dorsi and ilio-costalis. In all the pelves

examined there is an area on the internal surface of the ilium anterior

to where the first rib abuts on it. This area faces medially or is

directed forwards. In some pelves there is also a small area dorsal

to this rib (e.g. 3533 Bradysaurus bainr). It would therefore appear

that these two muscles are attached to the anterior surface or the

anterior plus the dorsal surface, and an attachment to the anterior

surface of the first rib is not excluded. These muscles—the longis-

simus dorsi and ilio-costalis—are inserted respectively on the trans-

verse processes of the thoracic vertebrae and the thoracic ribs. They

Pareiasaurian Studies.—Part VI. 355

assist locomotion in so far as they alternately pull the two sides of

the pelvis forwards. These muscles extend posteriorly where they

form the extensor caudae. The posterior internal surface of the ilium

served for the attachment of this muscle. This surface may extend

on to the internal surface dorsally of the last two ribs (e.g. 3533).

In one pelvis (6242) at least, the ribs occupy the whole of the posterior

internal surface, and the extensor caudae could only have been

attached to the dorsal edge of the ilium or (and) to the last rib. The

other end of attachment is to the transverse processes of the tail

vertebrae. This muscle is the antagonist of the two anterior ones—

it tends to pull the pelvis backwards.

B. Ventral Group.

This group consists of the obliquus, rectus, transversalis, pyra-

midalis, and quadratus lumborum anteriorly, and the ilio- and ischio-

caudalis posteriorly. The obliquus, transversalis, and pyramidalis

are attached to the ilio-pubic ligament, which runs from the anterior

dorsal extremity of the ilium to the pubic tubercle. The rectus con-

tinues ventrally back of the pubo-ischiadic ligament and is attached

to the ischiadic symphysis. The roughened antero-dorsal edge of

the ilium is well-marked in most pelves. As a prominence the pubic

tubercle is not well-defined, but at the most antero-lateral corner of

the pubis there is some indication of the attachment of the ligament.

From these pelvic attachments the obliquus stretches to the posterior

ribs, the rectus to the sternum, the pyramidalis along the linea

alba, and the quadratus lumborum to the lumbar vertebrae. Con-

tracting alternately these muscles assist locomotion by exerting an

anterior pull on the pelvis, which furthers the transverse rocking of

the pelvis. Posteriorly the edge of the posterior iliac process shows

in some pelves a distinctly roughened surface for the attachment of

the ilio-ischiadic ligament, which then runs to the rugose extremity

of the ischial tuber. From the upper end of this ligament a part of

the ilio-caudalis arises, and from the tuber ischii and the ligament the

ischio-caudalis. The other end of these muscles are attached to the

transverse processes of the caudal vertebrae. By lateral movements

of the tail these muscles may act as antagonists to the anterior

muscles.

306 Annals of the South African Museum.

Il. APPENDICULAR MUSCLES.

A. Dorsal Group.

(a) Short Deep Dorsals (text-figs. 55, 56 (6) and (c)).

This mass consists of the pubo-ischio-femoralis internus, the ilio-

femoralis, and the ischio-trochantericus.

Pubo-ischio-femoralis internus.——Three muscles and one ligament

were attached to the anterior, antero-internal, and antero-external

surfaces of the pelvis, viz. pubo-ischio-femoralis internus, pubo-

tibialis, ambiens, and the pubo-iliac ligament. The pubo-ischio-

femoralis internus is attached most internally. Its origin was prob-

ably spread over a large portion of the internal surface of the internal

Text-Fic. 55.—Lateral view of the right half of the pelvis of Embrithosaurus

schwarzi, showing the attachment of the muscles.

REFERENCES TO TEXT-FIGS. 55-60.

a—=ambiens. ipl=ilio-pubic ligament.

ad—adductor. is.c=ischio-caudalis.

ecd=extensor communis digitorum. i.is.1=ilio-ischial ligament.

eh=extensor hallucis. is.t=ischio-trochantericus.

eta—extensor tibialis anticus. it=ilio-tibialis.

fdcl—fiexor digitorum communis oae=obliquus abdominis externus.

longus. pb=peroneus brevis.

ft—femoro-tibialis. pife = pubo-ischio-femoralis externus.

fl.t—flexor tibialis (i=internus ; pifi—pubo-ischio-femoralis internus.

e=—externus). pist=pubo-ischio tibialis.

g—gastrocnemius. pl=peroneus longus.

ic=interosseum cruris. po= popliteal.

if =ilio-fibularis. pt=pubo-tibialis.

il.fe =ilio-femoralis. ql=quadratus lumborum.

is.c=ischio-caudalis. ra=rectus abdominis.

Pareiasaurian Studies.—Part VI. 357

swollen surface of the iium and pubis. There is no indication of a

shelf as in Dradectes, Cryptobranchus, etc. The muscle is inserted on

the proximal third of the femur on its dorsal and preaxial surface.

As in Dimetrodon, it appears that the muscle was divided at its in-

sertion into (1) a dorsal, and (2) a preaxial part. The two surfaces of

insertion are separated by the dorso-preaxial edge.

(1) The dorsal surface is a shallow groove running from the proximal

edge and fading away distally on the rounded dorso-preaxial surface

of the shaft, where it is confluent with the area of origin of the femoro-

tibialis. Postaxially the surface is bounded by a ridge situated in

Q. 6.

Trext-Fic. 56.—Femur of Hmbrithosaurus schwarz, showing muscle attachments.

(a) Ventral. (6) Dorsal. (c) Anterior.

the medio-longitudinal plane of the femur. This ridge ends distally

at the proximal third of the bone, where it forms a prominent rugosity

—to which, asin some extant lizards, a tendon was probably attached.

Preaxially it is proximally separated from the preaxial surface by a

triangular surface, which is more distinct in some femora than in

others. More distally the two areas are confluent.

(2) This portion of the insertion apparently occupies the whole of

the proximal third of the preaxial surface. Distally it is confluent

with the dorsal portion, where they are not clearly demarcated from

the area of origin of the adductor. Ventrally this surface does not

stretch on to the preaxial surface of the internal trochanter. In

Propappus, Brachypareia, and Pareiasuchus, where the proximal end

of the femur projects much more preaxially, both the surfaces lie

more preaxially than in the other forms.

398 Annals of the South African Museum.

The pubo-ischio-femoralis internus raises and advances the femur.

In the Pareiasaurs the raising of the femur is mostly done by the ilio-

femoralis, with the result that the pubo-ischio-femoralis internus is

not a very strong muscle.

Ilio-femoralis.—Most of the outer surface of the iliac blade and part

of the shaft shows signs of muscular attachment. With the exception

of two fairly distinctly separate surfaces anteriorly and posteriorly,

the whole of this surface right up to the dorsal ridge serves for the

origin of the ilio-femoralis. Postaxially of the dorso-medio-longi-

tudinal ridge mentioned above, the femur has developed a strong

flange of bone which stretches from the sloping proximal edge to just

proximal of the facet for the fibula situated on the postaxial condyle.

With the exception of the proximal portion, the whole of the dorsal

surface of this flange serves for the insertion of the ilio-femoralis.

Proximally the ilio-femoralis insertion is in some femora clearly, in

others indistinctly, separated from a small area for the insertion of

the ischio-trochantericus. At the middle third of the femur it is

sometimes not clearly demarcated from the general dorsal surface of

the shaft, which is also occupied by the femoro-tibialis. This area

also includes the dorsal surface of the external trochanter. It has

already been noted that the dorsal edge of this flange is broadly

rounded or straight in different femora. The ilio-femoralis was pre-

sumably stronger in the femur with the rounded edge.

As the surface of origin of the ilio-femoralis is to a greater or lesser

extent pre-acetabular, in addition to abducting and extending the

limb it pulls the femur up at the external trochanter, and consequently

the distal end of the femur is forced downwards and backwards. The

antagonistic action between the ilo-femoralis and the adductor results

in the rotation of the femur on its long axis.

Ischio-trochantericus.—On the inner surface of the ischium, just

below the ilio-ischiadic suture, there is a small shallow depression.

It is probable that the ischio-trochantericus originated from here.

It is inserted on the proximal portion of the postaxial flange. In

some femora there is a smaller concave area on the flange proximal

to the external trochanter, demarcated distally from the rest of the

flange by a not very definite irregular ridge. In other femora this

area is not distinct from the area for the insertion of the ilio-femoralis.

The ischio-trochantericus acts partly as antagonist to the pubo-

ischio-femoralis internus by pulling the femur backwards towards

the tail.

Pareiasaurian Studies.—Part VI. 359

(b) Long Superficial Dorsals (text-figs. 55, 57, 58).

Ilvo-tabialas.—The anterior external corner of the iliac blade bears

a not very distinct surface of origin of the ilio-tibialis. The eversion

of this part of the iliac blade in some forms must result in a difference

in the action of the muscle. The ilio-tibialis is inserted on the proxi-

mal dorsal surface of the tibia. Here a well-developed cnemial crest

with a deep groove on its postaxial side is developed, which serves as

surface of insertion of the triceps, which is composed of the ilio-tibialis,

ambiens, and femora-tibialis, which unite and form a tendon which

distally passes through the intercondylar sulcus and is inserted on

the cnemial crest and groove. The ilio-tibialis pulls the distal end

of the femur up and extends the lower leg.

TEXxT-FIG. 57.— Tibia of Hmbritho- Text-Fic. 58.—Fibula of Hm-

saurus schwarzi, showing muscle brithosaurus schwarzi, show-

attachments. ing muscle attachments.

(a) Dorsal. (6) Ventral. (a) Dorsal. (6) Ventral.

Ambiens.—The anterior pubic surface between the ilio-pubic suture

above and the step below shows an unmistakable area of muscular

attachment. To this surface the ambiens was presumably attached.

As noted above, the ambiens unites with the ilio-tibialis and the

femoro-tibialis, and inserts on the cnemial crest of the tibia. The

ambiens tends to abduct the femur and extend the lower leg.

Femoro-tibialis.—It originates on the middle third and proximal

part of the distal third of the dorsal surface of the femur. Just above

the intercondylar sulcus there is a groove—narrow and deep in some

forms and broader and shallower in others. This groove probably

served as the main area of origin, but the muscle also apparently had

a wide diffused area of origin without clear limiting edges. Proximally

it probably reached up to the distal plane of the proximal third of

the bone, where it bordered on the insertion of the pubo-ischio-

360 Annals of the South African Museum.

femoralis internus. Ventro-preaxially on the shaft there is an in-

distinct limiting edge between it and the dorsal edge of the area of

insertion of the adductor. Postaxially it probably did not extend

on to the flange occupied by the ilio-femoralis. To what extent the

pre- and postaxial condyles were occupied is not certain, but it is

probable that these surfaces were only slightly included. Joining

the ilio-tibialis and the ambiens, the femoro-tibialis inserts on the

cnemial crest. By contracting, the femoro-tibialis would tend to

extend the lower leg in relation to the upper.

Ilio-fibularits.—The triangular external surface on the posterior

process of the iliac blade shows an area fairly distinctly demarcated

from the ilio-femoralis surface. From this surface the ilio-fibularis

originated. This muscle is inserted on the roughened ventro-postaxial

part of the head of the fibula. The ilio-fibularis assisted slightly in

the abduction and extension of the limb; but, in addition, it rotated

the fibula, which carried this rotating force on to the foot.

Peronaeus longus et brevis.—The peronaeus longus originates partly

on the dorso-posterior surface of the postaxial condyle of the femur.

On the condyle thereis a very shallow longitudinal depression, bounded

ventro-posteriorly by a ridge lying dorsal to the fibular facet, and

bounded dorso-anteriorly by a ridge running from the shaft towards

the end of the condyle just posterior to the intercondylar sulcus.

The peronaeus longus partly originates from the fibula. On the

postaxial border of the fibula there are two long, parallel, flattened

surfaces. At the proximal third of the bone a ridge or a longitudinal

tubercle separates the two surfaces. The more ventral is shorter and

narrower than the more dorsal surface. On the former the longus

partly originates. The peronaeus brevis originates on the latter and

also on the ridge or tubercle. These two muscles are inserted on the

fifth metatarsal.

Extensor digitorum communis longus.—This muscle originates on

the roughened distal extremity of the postaxial condyle, distal to the

origin of the peronaeus longus and dorsal to the fibular facet. It

inserts by means of a number of tendons on the toes.

Tibialis anticus.—This muscle originates on the proximal postaxial

surface of the tibia. Its area of origin is not separate from that of

the popliteal muscle. It is inserted on the preaxial proximal surface

of the first metatarsal.

Pareiasaurian Studies.—Part VI. 361

Extensor hallucis.—On the distal dorsal surface an undefined area

served as surface of origin for this muscle, which inserted on the first

metatarsal.

As their name implies, the lower leg extensors extended the distal

segments of the limb in relation to the more proximal segments.

(d) Short Deep Ventrals (text-figs. 55, 56 (a) and (c)).

Pubo-ischio-femoralis externus.—The ventro-external surface of the

pubo-ischiadic plate from the symphyseal carina to the lower border

of the acetabulum affords a very good surface of origin for the deep

ventral muscles. The widespread disposition of the limbs made it neces-

sary for these muscles to be very strong. The pubo-ischio-femoralis

externus is the deepest ventral muscle. Its area of origin appears to

have been most of the surface ventral of the ridge continued from the

lower acetabulum border to the ischial tuber. Ventrally towards the

carina its area of origin is not separated from that of the adductor.

It was inserted into the fossa lying between the remnants of the proxi-

mal branches of the primitive Y-shaped adductor ridges. This fossa

is deeper in some femora than in others. Distally the fossa has no

distinct limiting edge. The muscle apparently does not insert more

distally than the plane of the distal end of the internal trochanter.

Postaxially the muscle does not extend on to the posterior edge of

the external trochanter. Preaxially it inserts on both the pre- and

postaxial surfaces of the internal trochanter. Proximally the muscle

inserts right up to the edge—especially just proximal to the external

trochanter, where the fossa has no distinct edge. The insertion on

the preaxial surface of the internal trochanter was stronger in the

forms with a sigmoidally curved trochanter. The preaxial limiting

edge proximal to the internal trochanter is distinct in some forms

(Brachypareia, Propappus, Embrithosaurus), whereas in the others

the surface here is rounded and passes imperceptibly into the preaxial

surface of the attachment of the pubo-ischio-femoralis internus.

Adductor.—The middle layer of the deep ventral muscles is formed

by the adductor. Its surface of origin is not differentiated from that

of the deeper pubo-ischio-femoralis externus. It apparently originated

along the whole pubo-ischiadic plate ventral to the pubo-ischio-

femoralis externus. It is inserted on the preaxial surface of the shaft

of the femur, and probably extended on to the preaxial surface of the

internal trochanter. Ventrally it is bounded by the remnant of the

stem of the primitive Y ridges. Proximally, dorso-proximally, and

VOL. XXVIII, PART 3. 25

362 Annals of the South African Museum.

dorso-distally it is not clearly limited from the area of insertion of the

pubo-ischio-femoralis internus and the area of origin of the femoro-

tibialis.

The adductor partly draws the femur downwards towards the pubo-

ischiadic plate; partly it tends to draw the femur backwards; and

thirdly, by acting in opposition to the ilio-femoralis it produces

rotation of the femur. The pubo-ischio-femoralis externus acts

mainly as an adductor—pulling the femur ventrally down towards

the pubo-ischiadic plate. When the femur is advanced it, however,

also tends to pull the femur backwards; if the femur is directed back-

wards, it again tends to pullit forwards. As the Pareiasaur limbs are

very widespread these two muscles must have been very strong ; both

their large surfaces of origin and insertion support this assumption.

(e) Long Superficial Ventrals (text-figs. 55, 57, 58).

Pubo-tibialts.—From the external corner of the pubis to the border

of the ischium there stretched a pubo-ischiadic ligament. Along this

ligament and the symphyseal carina the long, superficial, ventral

muscles originate. The most anterior is the pubo-tibialis, which

originates from the pubic tubercle. It inserts, together with parts of

the flexor-tibialis internus and externus, on the distal depression on the

ventral surface of the tibia. On the postaxial surface of these is a

tubercle well-developed in Propappus omocratus. The pubo-tibialis

partly imserted on it.

Pubo-ischio-tibialis.—This muscle is the medial one of the long

superficial ventrals. It originates from the carina and the pubo-

ischiadic ligament. It inserts on the tibia. On the ventral surface

of the tibia a ridge runs diagonally from the proximal postaxial corner

to the distal preaxial corner. Proximal and distal to this ridge lie

two depressions. The pubo-ischio-tibialis inserts on the proximal

depression and descends down the bone on the preaxial side of the

ridge. With it are associated parts of the flexor tibialis internus.

Flexores tabsalis internus and eaxternus.—Posteriorly these muscles

originate around the tuber ischu and the adjacent pubo-ischiadic and

ilio-ischiadic ligaments. As already mentioned, a part of the flexor

tibialis internus inserts on the diagonal ridge and proximal ventral

depression in association with the pubo-ischio-tibialis. A part, and

also the flexor tibialis externus, inserts on the distal depression and the

postaxial tubercle of the tibia. These long superficial ventrals acting

directly from the pelvis to the tibia act as flexors of the lower leg.

Pareiasaurian Studies.—Part VI. 363

(f) Lower Leg Flexors (text-figs. 56, 57, 58).

Gastrocnemius.—This muscle originates in two parts from the sur-

face proximal to the preaxial and postaxial tibial facets of the femur.

The two facets are lower than the ventral surface just proximal to

them, and on these curved surfaces leading up to them there are

indications of muscular attachment. The gastrocnemius inserts on

the sole, where its tendons form the superficial fascial layer of the

sole.

Flexor digitorum communis longus.—This muscle originates from

most of the ventral surface of the fibula. Together with the tibialis

posticus this muscle inserts on the sole, where it forms the deeper

layer of the sole.

Tibialis posticus.—It originates from the distal ventral surface of

the fibula, and together with the flexor digitorum communis forms

the deeper layer of the sole.

Popliteal Muscle.—It originates on the fibula. On the ventro-

preaxial border of the fibula, just below the proximal articulatory

surface, there is a fairly large but low elevation, from which the

popliteal muscle originates. It is inserted on the proximal postaxial

corner of the diagonal ventral ridge of the tibia.

Interosseus cruris.—On the ventral preaxial border of the fibula,

just above the distal articulatory facet, a ledge is developed, from

which the interosseus cruris originates. It is inserted on the distal

postaxial corner of the ventral surface of the tibia.

These lower leg flexors, as their name implies, serve to bend the

lower leg medially towards the femur and to hold the lower leg in

position when the body is pulled forwards over them.

(g) Caudal Muscles.

The stem of the primitive Y system of adductor ridges is reduced,

and as there is no connection with the posterior limb there is no

eminence at the joint where the two limbs primitively joined. The

internal trochanter is generally continuous with the remains of the

stem. There is, however, no eminence on this ridge, and if there

were a coccygeo-femoralis muscle its area of insertion must have

been diffusely spread along this ridge. As the Pareiasaur tail is not

very long this muscle was probably not strong.

To describe the movement of the hind limb in walking let us take

up the account at the moment that the right limb has been put

forward. In this position the propodial is directed forwards and out-

364 Annals of the South African Museum.

wards, and is more or less horizontal; the epipodial is directed verti-

cally but also somewhat obliquely forwards; the foot is directed

forwards and somewhat outwards.

With the strong fourth toe directed

outwards the foot is firmly implanted

to prevent any slipping out sideways.

The body is now pulled over the fixed

foot. In doing so the right side of

the pelvis is raised and displaced

towards the left. When the right side

of the pelvis is at its highest point

the femur is directed obliquely out-

wards; the epipodial is vertical and

the foot directed outwards and for-

wards. This necessitates a twisting

movement of the tibia over the proxi-

mal tarsal element. The left side of

the pelvis is then sagging. The femur

is now rotated slightly in clockwise

direction when viewed from the side.

The right side of the pelvis then con-

tinues its forward motion, with the

result that the propodial is now

directed obliquely backwards and

outwards, the epipodial obliquely

Text-Fic. 59.—Left side of pelvis backwards, and the foot also obliquely

and limb, showing the attach-

mention cholehicnnnmenice backwards. For the next step the

limb describes an are and reaches a

position in front of the pelvis, from whence we started our account.

At the forward position the anterior right axial muscles pull the right

side of the pelvis forwards, this being augmented by the antagonistic

pull of the posterior left axial muscles and the left caudal muscles, and

by the pull of the pubo-ischio-femoralis externus and ischio-trochan-

tericus. The total pull, therefore, could have been considerable, and

rapid movement seems possible. In walking, the rump and tail are

therefore to some extent simuously curved in lateral direction. The

sides of the pelvis are alternately raised and lowered, producing a

lateral rocking. To make this clumsy swaying mode of progression

possible it is necessary that there be strong muscles holding the limbs

in position and preventing them slipping away sideways. The short

ventral muscles to the femur as well as the longer ventrals, and even

Pareiasaurian Studies.—Part VI. 365

the lower leg flexors, must therefore have been very strong, whereas

the extensors were only moderately developed. The Pareiasaurian

Bes ,

SEES

(bs

TEext-Fric¢. 60.—Anterior view of the pelvis and limbs, showing the attachment

of the chief muscles.

tail was only moderately long and the coccygeo-femoralis weak, so

that the tail did not play a great role in moving the limb; being

short and massive its action on the pelvis must, however, have been

considerable.

366 Annals of the South African Museum.

LITERATURE CITED.

(1) Amatitzky, V.—‘‘ Seymouridae,”’ Akademii Nauk. Petrograd, 1921.

(2) Amanitzky, V.—‘‘ On a New Cotylosauria of the Family Seymouridae from the

Permian of North Russia,’’ Ann. Mag. Nat. Hist., 9th series, vol. xiii, 1924.

(3) Boonstra, L. D.—‘‘ On the Pareiasaurian Pes,’’ Ann. S. Afr. Mus., vol. xxviii,

pt. i, 1929.

(4) BouLencER, G.—‘‘ On the Characters and Affinities of the Triassic Reptile

Telerpeton elginense,” Proc. Zool. Soc., vol. i, 1904.

(5) Broom, R.—‘‘ On an Almost Perfect Skeleton of Paretasaurus serridens,

Owen,” Ann. S. Afr. Mus., vol. iv, pt. ii, 1903.

(6) Broom, R.—“ On the Pareiasaurian Genus Propappus,” Ann. S. Afr. Mus.,

vol. iv, pt. vii, 1908.

(7) Broom, R.—‘‘ On a New Species of Propappus, ete.,’ Ann. S. Afr. Mus.,

vol. vii, pt. v, 1912.

(8) Broom, R.—‘‘ On the Manus and Pes of Pareiasaurus,” Ann. 8. Afr. Mus.,

vol. vii, pt. vi, 1913.

(9) Broom, R.—‘‘ On Some Points in the Anatomy of the Theridont Reptile

Diademodon,”’ Proc. Zool. Soc., vol. i, 1905.

(10) Broom, R., and Haucuron, 8. H.—‘‘ On the Skeleton of a New Pareiasaurian

(Pareiasuchus peringueyt),’ Ann. 8. Afr. Mus., vol. xii, pt. i, 1913.

(11) Cass, E. C.—‘‘ Revision of the Pelycosauria of North America,” Publ. Carnegie

Inst., Washington, 1911.

(12) Casz, EB. C.—“ Revision of the Cotylosauria of North America,” Publ. Carnegie

Inst., Washington, 1907.

(13) Casz, E. C., Wituiston, 8. W., and Ment, M. G.—‘‘ Permo-carboniferous

Vertebrates from New Mexico,” Publ. Carnegie Inst., Washington, 1913.

(14) Gregory, W. K., and Camp, C. L.—‘‘ Studies in Comparative Myology and

Osteology,” Bull. Am. Mus. Nat. Hist., vol. xxxviii, 1918.

(15) Hartmann-Wetnsera, A.—< Uber Carpus und Tarsus der Pareiasauriden,”’

Anat. Anz., Bd. 67, Nr. 19/21, 1929.

(16) Havenron, 8. H.—‘‘ On a New Species of Propappus,’ Ann. 8. Afr. Mus.,

vol. xii, pt. i, 1913.

(17) Haveuton, 8S. H.—‘‘ On some New Therapsid Genera,” Ann. S. Afr. Mus.,

vol. xxviii, pt. i, 1929.

(18) Haventon, 8. H.—“ Investigations in South African Reptiles and Amphi-

bians,”’ Ann. S. Afr. Mus., vol. xxii, 1925.

(19) Huznr, F. von.—‘‘ Ueber die Procolophoniden u.s.w.,’’ Centrblatt fur

Mineralogie, p. 78, 1911.

(20) Huxtey, T. G.—‘‘ On a New Specimen of Telerpeton elginense,” Quart.

Journ. Geol. Soc., vol. xxiii, 1867.

(21) LypEKKER, R.—Catalogue of Fossil Reptilia and Amphibia, pt. iv, 1890.

(22) MantrEeLL, G.—“‘ Description of the Telerpeton elginense,” Quart. Journ. Geol.

Soc., vol. viii, 1852. ;

(23) OwEN, R.—“‘ Report on the Reptilian Fossils from South Africa,” Trans.

Geol. Soc. (London), ser. ii, vol. vii, 1856.

(24) OwEN, R.—“ On the Dicynodont Reptilia, etc.,’’ Phil. Trans. Roy. Soc., B,

vol. 152, 1862.

Ann. 8. Afr. Mus., Vol. XXVIII. Plate XX XVIII.

HIND FEET OF SOUTH AFRICAN PARETASAURS.

: Neill & Co., Lid.

Pareiasaurian Studies.—Part VI. 367

(25) OwEN, R.—Catalogue of the Fossil Reptilia of South Africa, 1876.

(26) Romer, A. S.—* The Locomoter Apparatus of Certain Primitive and Mammal-

like Reptiles,’ Bull. Am. Mus. Nat. Hist., vol. xlvi, 1922.

(27) SpELEY, H. G.—“‘ Researches on the Structure, Organisation, and Classifica-

tion of the Fossil Reptilia,’ Phil. Trans. Roy. Soc., B, vol. 179, 1888.

(28) Srmtey, H. G.—Jbid., Proc. Roy. Soc. (London), vol. xliv, No. 267, 1888.

(29) Srmtny, H. G.—Jbid., Phil. Trans. Roy. Soc. (London), B, vol. 183, 1892.

(30) Srrtey, H. G.—Jbid., Phil. Trans. Roy. Soc. (London), B, vol. 186, 1895.

(31) Watson, D. M. S.—“ On the Nomenclature of the South African Pareia-

saurians,’ Ann. Mag. Nat. Hist., ser. viii, vol. xiv, 1914.

(32) Watson, D. M. S.—‘‘ The Evolution and Origin of the Amphibia,” Phil.

Trans. Roy. Soc. (London), B, vol. 214, 1926.

(383) Watson, D. M. S.—“‘ On Seymouria, the Most Primitive Known Reptile,”

Proc. Zool. Soe. (London), vol. ii, 1918.

(34) Watson, D. M. S.— The Skeleton of Lystrosaurus,” Rec. Albany Mus., vol. ii,

1912.

(35) Wituiston, S. W.—“‘ Little-known Permian Vertebrates,” Journ. Geol., vol.

xvii, Chicago, 1919.

(36) Wittiston, S. W.—American Permian Vertebrates, Chicago, 1911.

EXPLANATION OF PLATE.

PLATE XX XVIII.

a. Dorsal view of the right foot of Bradysaurus seeleyr (9168).

b. Dorsal view of the left foot of Bradysaurus seeleyt (9137).

c. Dorsal view of the left foot of Hmbrithosaurus schwarzt (8034).

ie ANNALS.

OF THE

Pom APRICA wUsHt

VOLUME XXVIII.

"DESCRIPTIONS: OF THE ‘PALABONTOLOGICAL MATERIAL

COLLECTED sy tHe SOUTH AFRICAN MUSEUM anp

“THE GEOLOGICAL SURVEY OF SOUTH AFRICA

“PART. IV. re —

10. Some Fossil Plants from the Karroo System of South Africa.

_ By Avex. L. pu Torr, D.Sc., F-G.8. (With Plates XX XIX

_ and XL and 2 Text-figures.)

— ao ae Fosal Plants from the roe Beds of Uganda. By

= ALEX. L. pu Tort, D.Sc., F.G.8. (With Plate XLL)

42. The Fossil Bguidec of South Africa. By S. H. Haveuron,

B.A., D.Se., Hon. Curator, Palaeontological Collections.

(With s Text- oe &

“13. ‘Pareiasaurian Studies.

ee VUL.—On the Hind Limb of the Two Little-known

- Pareiasaurian Genera: Anthodon and Pareiasaurus.

= iby Iiguwe D. Boonsrra, D.Sc. (With Text-

figures ae ) :

Pee VILL. —The Oe and Mu yology of the Locomotor

ae Apparatus. B.—Fore Limb. By L. D. Boonsrra,

—DSe.,. Assistant in Palaeontology. (With Plates

-XLII-XLIV and Text- -figures 1-41.)

: Tite Page and Index to Volume XXVIII.

«ISSUED FEBRUARY 1982. PRICE 10s.

: = PRINTED FOR THE

TRUSTEES OF THE SOUTH AFRICAN MUSEUM

See -__ AND THE :

GROLOGIOAL SURVEY OF SOUTH AFRICA

( 369 )

10. Some Fossil Plants from the Karroo System of South Africa.—

byeAtEX. pu Tom, D.S8c., F.G.S.

(With Plates XX XIX and XL and 2 Text-figures.)

In the following pages a number of plants are described that come

from various horizons within the Karroo System and from widely

separated places, the bulk of the material examined belonging

to the Union Geological Survey. Certain of them were provision-

ally identified some years ago, though no descriptions were

published.

While the number of new forms is small, several are of uncommon

interest, more particularly Hretmonia, a new genus, which is regarded

as probably the sporangiophore of Glossopteris, while in the case of

Stigmaria this structure is recorded from South Africa for the first

time. In addition, several known species are redescribed in the light

of more complete examples, while reference is made to certain new

localities of palaeobotanical importance.

Stratigraphically the plants can be arranged as follows :—

A. Ecca Series—Lower Permian.

Stigmaria ficordes, Brongn.

Gangamopteris cyclopteroides, Feist.

B. Beaufort Series (Lower stage)—Upper Permian.

Schizoneura gondwanensis, Feist.

Phyllotheca australis, Brongn.

Sphenophyllum speciosum (Royle).

Sphenopteris alata, Brongn.

Glossopteris cordata, Feist.

Eretmomea natalensis, gen. et sp. nov.

C. Stormberg Series (Molteno stage)—Upper Triassic.

Baiera tenurfolia, Johnston.

Johnstonia coriacea, Walkom.

Pterophyllum (Anomozamites) inconstans (Braun).

VOL. XXVIII, PART 4. 26

MAY 6 1982

Sa ey Teg NE ees, ath

—7

—— . 2 = Sa _ eS eae ———— ae

Se pares are | Eck SM enemas PL We Pale ee ee Cama 2 ee = Pare vi

a WAG

7 ee ae Saree $F

Se ee Le ee

a ee oe

—

Tana

370 Annals of the South African Museum.

A.—ECCA SERIES.

LYCOPODIALES.

Genus STIGMARIA, Brongniart, 1822.

STIGMARIA PICOIDES, Brongniart.

Plate XXXIX.

1882. Mém. Mus. d’hist. Natur. Paris, pl. xu, fig. 7, p. 228.

The interesting specimen about to be described was kindly loaned

me by Dr. T. N. Leslie of Vereeniging, Transvaal. It was picked up

not far from a waste dump at the central coal mine at Vereeniging by

Mr. R. H. Smith of that town, and is presumed to have come from

the workings of that colliery, though unfortunately such presumption

cannot be established. Various features, such as the nature of the

adhering sandstone, the mode of fossilisation and the fact that pieces

of rather similar fossilised wood have actually been got in mining,

strongly suggest that this fossil is not of foreign origin, but has

actually been derived from the workings of this mine. Admitting

such, it would have come from the Middle Ecca Beds or “ Coal

Measures,”’ which formation has locally yielded by quarrying so many

good specimens of the Gangamopteris Flora. It is therefore to be

hoped, in the interests of science, that some day confirmation may be

obtained through the finding of identical remains 7 sztw.

The specimen in question is in a fairly good state of preservation,

being only slightly flattened ; its elliptical cross-section measures

7 by 5 cm., while its length is 31 cm., the uniformity being such that

one cannot be certain which is the proximal or distalend. A goodly

proportion of the spirally arranged appendages are still in attachment,

wrapped around the rhizome and closely pressed against the surface

thereof, but without any semblance of regularity; in places they have

left impressions on the rhizome itself. On one side of the specimen

a small amount of the pale sandstone matrix is still adhering to the

rhizome, either filling the scar-depressions, or surrounding the bases of

the rootlets, or presenting casts of the latter. Occasionally a rootlet

has partly encircled another and has left its partial impression upon

the latter. Such casts show the appendages to have been nearly

cylindrical in shape with slightly swollen bases, though the finely

striated rootlets themselves are usually wrinkled, such being appar-

ently due to the disappearance of the interior tissue. Such rotting,

Some Fossil Plants from the Karroo System of South Africa. 371

coupled with the twisting of the soft organs and subsequent pressure,

‘has tended to obscure the originally cylindrical nature of these

rootlets. At their base a diameter of fully 5 mm. is attained, while the

greatest length measurable is 45 mm. In a few favourable cross-

sections the central vascular bundle can be distinguished.

Where the surface of the fossil is exposed, it is seen to be very

faintly and longitudinally ribbed and wrinkled, while in some favoured

areas a finely pitted face, still coated with carbonaceous matter,

indicates some of the cortical exterior. The regular, spirally set scars

are deep and slightly oval, this last feature being due to compression.

In the middle of each is an umbilicus with a central depression marking

the position of the vascular bundle. No progressive difference in the

average size of the scars along the length of the fossil is discernible.

One feature, that is typical of Stigmaria, is particularly well pre-

served, namely, the eccentric cast of the pith-cavity and its small

diameter. The specimen, as it fortunately happens, has a conspicuous

external groove, which is purely accidental, that runs the full length

of the fossil and broadens slightly towards the one end. Situated

immediately beneath this furrow, and enveloped by the sandstone

which fills the cylinder of cortical tissue, is the flattened pith-cast

16 mm. wide and 3 mm. thick. Where the groove becomes widest the

cylinder must have been ruptured, and the surface of the pith-cast in

consequence now stands revealed and can be examined over a short

distance. It shows shallow longitudinal flutings of the intermittent

type so characteristic of Stigmaria. An accidental though lucky

transverse break across the fossil, severing it into two portions,

displays the internal fluted cast retaining a coating of coaly matter.

Whitish sandstone fills the pith-cavity as well as the space between the

latter and the cortex, such having been rendered possible by the

decay of the woody tissue.

These various features enable the fossil to be referred Bramestonebly,

to Stigmaria, while the nature of the cortical surface, the uniformity in

size of the scars and several minor characters indicate an identity

with that very well known “ species’ S. ficoides ; indeed the rhizome

can be closely compared with some of the many figures of that plant

published by Brongniart, Géppert, Schimper, and Williamson.*

S. ficoides is considered by most palaeobotanists to represent the

underground rhizome of Sigillaria and very probably too of Lepido-

dendron, wherefore it is important to know that both of these genera

* See Solms-Laubach, Fossil Botany, chap. xii, 1891; and Seward, Fossil

Plants, vol. ii, chap. xvii, 1910.

a a a ee ae a al

<< - y. “2 -

t= —_ - = SRST SS EET poner

~~ .-

™

— 35 =

woke

= — ¥ » —*S =

PS SSI STE

372 Annals of the South African Museum.

are represented in the fossiliferous beds at Vereeniging. It is never-

theless admitted that under the name Stigmaria have perforce been

placed fossils indistinguishable from one another anatomically,

though belonging perhaps to distinct species or even genera.

S. ficoides (sensu lato) is known from strata ranging in age from the

Devonian to the Lower Permian, but is most common in the Carboni-

ferous and is especially abundant in the European Coal Measures,

though not hitherto recorded from the Gondwana System of South

Africa, so that a further name has now to be added to the list of the

“northern ”’ forms known to occur in the Lower Karroo Flora of this

country. S. ficoides has, however, been identified by Zeiller * and

Halle t from Shansi, China, from beds of Upper Carboniferous and

Lower Permian age with a flora containing certain Gondwana elements,

while it has also been recorded by Walkom t from the (probably)

Lower Carboniferous of New South Wales.

The photograph reproduced in Plate XX XIX was kindly taken by

Mr. A. M. Duggan-Cronin of Kimberley, while the specimen itself is

preserved in the private collection of Dr. T. N. Leslie at Vereeniging,

Transvaal.

PLANTS OF UNCERTAIN POSITION.

Genus GANGAMOPTERIS, McCoy, 1861.

GANGAMOPTERIS CYCLOPTEROIDES, Feistmantel.

(For synonomy see Arber, The Glossopteris Flora, p. 104, 1905.)

The number of places in South Africa where this typically Lower .

on record an additional locality, namely, Artesia Siding in the Bechu-

analand Protectorate, which incidentally is the northernmost spot in

South Africa where this genus has been discovered.

The material was collected by the writer from ground thrown out

of a well that had been deepened to 132 feet, the strata concerned

being grey and sometimes micaceous shales, whitish or yellowish sand-

stones, and thin bands of grits and arkoses. In 1930 a borehole was

sunk alongside, which proved these beds to extend to the depth of

* R. Zeiller, Annales des Mines, p. 16, pl. vii, fig. 8, 1901.

+ T. G. Halle, Palaeontologica Sinica, Ser. A, vol. ii, fasc. 1, p. 181, pl. xlix,

figs. 11, 12, 1927.

t A. B. Walkom, Proc. Linn. Soc. N.S.W., vol. lili, pt. 3, p. 259, pls. xxi (fig. 4),

xxiii (fig. 4), 1928.

Some Fossil Plants from the Karroo System of South Africa. 373

590 feet at least, their general character remaining the same, though

the sandstones became on the whole a bit finerin grain. The lithology

is typically that of the Middle Kcca “ Coal Measures ”’ and corresponds

closely with that found to the east in the coalfield of the North-

western Waterberg.

The majority of the examples consist of wide fronds of G. cyclo-

pteroides, but there are a couple of examples (Nos. 8663, 8668) of

the narrower forms belonging to Feistmantel’s variety attenuata.

Curiously, Glossopteris is unrepresented.

The specimens (Nos. 8661-8) are preserved in the South African

Museum.

B.—BEAUFORT SERIES.

The majority of the plants described below were obtained from

a road-cutting on the ascent to Bezuidenhout’s Pass, from the farm

Zand Drift to the north-west of Bergville in Natal. They occurred in

a 5-foot bed of bluish shale resting upon a thin impure coal and over-

lain by sandstone. At a level about 100 feet higher is a thick body

of sandstone containing much silicified wood, probably Dadoxylon

arbert. The fossiliferous shale is situated 420 feet below the base of

the Middle Beaufort Beds and the horizon is therefore well up in the

upper part of the Lower Beaufort Group, the thickness of which not

far to the east is estimated at 1500 feet.

EQUISETALES.

Genus SCHIZONEURA, Schimper and Mougeot, 1844.

SCHIZONEURA GONDWANENSIS, Feistmantel.

Plate XL, figs. 1, 2.

1876. Feistmantel, Rec. Geol. Surv. India, vol. ix, pt. i, p. 69.

(For synonomy see Arber, The Glossopteris Flora, p. 5, 1905.)

The three specimens from Natal collectively display all the essential

characters of this typically Indian form. In No. 269w, fig. 1, the

narrow striated stem gives off from each node a pair of curving leat-

sheaths, that tend to become split up distally. It can be compared

with Feistmantel’s pl. ia, fig. 5; pl. itia, fig. 2; pl. xa, fig. 7.

Specimen 283w, fig. 2, shows the impression of a larger leaf-sheath with

eight strong longitudinal veins that converge towards both base and

374 Annals of the South African Museum.

apex, while a more delicate strand runs along the middle of each of the

interspaces. It resembles Feistmantel’s pl.i a, fig. 2; pl. 14, fig. 1;

pl. iv 4, fig. 1; pl. via, figs. 1 and 3.*

Certain broad striated stems are also represented in the collection,

that could well belong to this genus and species, but in the absence of

sheaths in attachment they might perhaps be portions of Phyllotheca.

Remarkable is the finding by the writer of a specimen of this form,

showing a narrow stem with bilateral leaflets, in a boulder of dark

fine-grained sandstone included as a foreign block at the depth of

1000 feet in the kimberlite of the Jagersfontein Diamond Mine, Orange

Free State.

The specimens are preserved in the South African Museum, Cape

Town.

Locality.—A little to the north-west of Bergville, Natal.

Horizon.—Near the top of the Lower Beaufort Beds.

Genus PHYLLOTHECA, Brongniart, 1828.

PHYLLOTHECA AUSTRALIS, Brongniart.

Plate XL, figs. 3, 4.

(For synonomy see Arber, The Glossopteris Flora, p. 17, 1905.)

Several shale specimens from Natal carry impressions of detached

equisetaceous leaf-sheaths belonging to this form and possessing a

collar about 6 mm. in width from which arise from 22 to 26 narrowing

and slender leaves that vary somewhat in their length and are occasion-

ally as much as 10 mm. long. Through each leaf runs a strong nerve

right to the apex, and another one midway between each pair of

leaves up to the edge of the collar. These sheaths appear to have

fitted the stem closely in their lower parts (No. 276w, fig. 3), but to

have belled out thereafter, so that the leaves spread out at a wide

angle to the stem, sometimes to as much as aright angle. The differ-

ences in appearance due to the direction in which flattening has taken

place can be judged by comparison with No. 280w, fig. 4, where the

sheath, pressed down on the slab, is viewed from below. In cer-

tain examples seen from above pressure has not obliterated the

cup-like character. One small stem on No. 275w is 5 mm. wide and

has a leaf-sheath still attached to it; the internodal distance is

20 mm.

* QO. Feistmantel, Palaeontologica Indica, vol. iii, 1880.

Some Fossil Plants from the Karroo System of South Africa. 375

A good agreement is to be found between these examples and

Phyllotheca indica, Bunbury, from the Raniganj Group of India,*

which species Seward } regards, and with correctness, so far as can be

judged, as identical with Brongniart’s P. australis. Having had the

opportunity of instituting comparisons with specimens of the latter

from the type locality at Newcastle, New South Wales, it can be said

that the Natal examples differ from the Australian only in the shorter

length of the delicate leaves and the more open nature of the sheaths,

and they can notwithstanding be ascribed to that well-known and

earlier-described species ; their venation is certainly the same. The

differences between P. indica and P. australis have largely been based

upon the degree of closeness of the sheaths on the stem, a character

which, from the published figures, would scarcely appear to be a

constant or reliable one. Furthermore, this character would be

influenced by the nature and direction of the compression of the

structure during fossilisation.

Almost indistinguishable are sheaths with filiform leaves from the

Lower Beaufort Beds of Somkele, Zululand, that were described by

Etheridge, jun.,[ as P. zedlleri, and those from the same locality

similarly named by Seward,§ who, however, drew attention to the

strong resemblance to P. australis. For making his new species,

Etheridge laid stress upon the coarse ribbing and long internodal

distances in certain associated stems, but it is not clear from the

text that the sheaths in question were actually found in attachment

to those stems. Seward’s figures are so like those given here, that

I have taken leave to doubt the validity of Etheridge’s species so

far as the sheaths are concerned, and have accordingly placed the

Zululand sheaths along with these from Bergville under P. australis.

On the same slabs as the latter (Nos. 275w and 280w) are various

flattened stems from less than 10 mm. up to nearly 30 mm. wide,

and with internodal distances approximately equal to the width.

They are smooth or faintly and finely striated, save at the nodes,

where they show the impress of the leaf-sheaths; the number of

leaf-scars for the half-surface may be as many as 32. Where the

surface layer has been removed the fluted cast of the pith cavity is

seen with the narrow grooves running directly through the nodes.

* Bunbury, Quart. Journ. Geol. Soc., vol. xvii, p. 335, pl. x, figs. 6-9, 1861.

+ Seward, Fossil Plants, vol. i, p. 288, 1898.

{ R. Etheridge, First Rep. Geol. Surv. Natal and Zululand, pp. 72-75, pl. xiii,

figs. 4-6, 1902.

§ A. C. Seward, Trans. Geol. Soc. S.A., vol. x, p. 82, pl. viii, figs. 2, 3, 1908.

376 Annals of the South African Museum.

From the fineness of the ribbing these are very probably stems

of P. australis.

The Natal plants differ only shghtly from P. griesbachi, Zeiller,*

from the Barakar Group of India, but do not agree with any of those

described by Zeiller from the Permian of Siberia.

Locality.—A little to the north-west of Bergville, Natal.

Horizon.—Near the top of the Lower Beaufort Beds.

Genus SPHENOPHYLLUM, Brongniart, 1822.

SPHENOPHYLLUM SPECIOSUM (Royle).

Plate XL, figs. 5, 6.

1833. Trizygia speciosa, Royle, Illust. Bot. Himal. Mounts., p. xxix,

(OG 7, He, ‘e).

(For synonomy see Arber, The Glossopteris Flora, p. 35, 1905.)

This well-known Indian form is represented by several specimens.

The largest of these (No. 283w, fig. 5) consists of a single, complete

whorl of six leaves attached to the slender, striated stem, and can

be compared both as to size and shape with Feistmantel’s examples

(pl. xia, fig. 8, and pl. xia, fig. 1).— No. 285w, fig. 6, shows

several whorls of smaller leaflets in attachment to a slender, jointed

axis, and can be paralleled with the forms figured by Feistmantel

in his pl. xia, fig. 2, and pl. xia, fig. 2. In view of the precise

agreement of the Natal with the Indian examples, further detailed

description becomes unnecessary. The fossils in question were

collected and identified by the writer so far back as 1916, and con-

stituted the first proof of the existence of this characteristic Gond-

wana plant in the Karroo System, but in 1929 Walton t recorded it

from the Wankie Coalfield of Southern Rhodesia.

The above, together with other specimens (Nos. 284w, 286w, and

287w), are preserved in the South African Museum, Cape Town.

Locality.—A little to the north-west of Bergville, Natal.

Horizon.—Near the top of the Lower Beaufort Beds.

* R. Zeiller, Mem. Geol. Surv. India, New Ser., vol. ii, pt. 1, p. 30, pl. vii, fig. 1,

1902.

+ O. Feistmantel, Flora. Gondwana System, vol. iii, 1880.

t J. Walton, 8. Rhodesia Geol. Surv., Bull. No. 15, p. 64, 1929.

Some Fossil Plants from the Karroo System of South Africa. 377

PLANTS OF UNCERTAIN POSITION.

Genus SPHENOPTERIS, Brongniart, 1822.

SPHENOPTERIS ALATA (Brongniart).

Plate XL, figs. 7, 8.

1834-36. Pecopteris alata, Brongniart, Hist. Végét. foss., p. 361,

pl. exxvil.

(For synonomy see Arber, The Glossopteris Flora, p. 138, 1905.)

Several portions of fronds, one of which (No. 369w), is figured, can

be referred to the species S. alata rather than to the similar S. lobifolia,

Morris,* both from the Newcastle Series of New South Wales. All

three possess in common a frond with alate rachis, secondary pinnae

set rather open, and at a wide angle, but in the South African form

the alternate character is persistent up to the apex, just as in S.

alata, whereas in S. lobifolia the secondary pinnae become opposite

in the apical portion of the frond.

The pinnules are membraneous, alternate, contracted at base, and

decurrent, having two and more generally three lobes on either side

and a terminal lobe, while the primary nerve is flexuous, giving off

alternating secondary branches to the lobes, these dividing in turn

into two or three branches, sometimes four, as indicated in the

enlarged drawing, fig. 8. In their shape and venation these pinnules

are exceedingly like those of S. hughesz (Feist.),t from the Raniganj

Stage of India, but the rachis in the latter is not recorded as being

alate. They are definitely unlike the pinnules of S. lobifolia, which

are less numerously lobed—see Arber’s figures.

Our specimens seem indeed indistinguishable from the figures of

S. alata, var. exilis, given by Morris (pl. vu, figs. 4, 4at) (the

varietal name is superfluous), and of S. alata, recorded by Dun,§

for example his pl. vi, fig. 1.

Interest centres in the fertile character of certain apical pinnae

(Nos. 267w, 268w, and 269w), the relatively few sori being situated

generally on certain of the secondary veins within, though not at

the very edges of, the lobes. The arrangement differs from that

recorded by Shirley for S. morrisianum of Tasmania—ascribed by

* See Arber, The Glossopteris Flora, pp. 135-138, 1905.

tT Ibid., pp. 133-135.

t Morris in Strzelecki’s New South Wales, 1845.

§ Dun, Rec. Geol. Surv. N.S.W., vol. vi, pt. 1, p. 48.

378 Annals of the South African Museum.

Arber to S. lobsfolia—in which the sorus is situated on the first fork

of the primary vein of each pinnule. As the matrix is not very fine-

grained, the details cannot be made out, but a few of the knobs seem

to be approximately kidney- eRe this and their position recall

the living Dicksonia.

Locality.—A little to the north-west of Bergville, Natal.

Horizon.—Near the top of the Lower Beaufort Beds.

Genus GLOSSOPTERIS, Brongniart, 1822.

GLOSSOPTERIS CORDATA, Feistmantel non Dana.

Text-figure 1, A, B, C, D.

1878. Glossopteris cordata, Feistmantel, Palaeontographica, Suppl.,

pa o2:

1882. Glossopteris cordata, Feistmantel, Palaeont. Indica., Gondwana

Flora, vol: iv, pt-.1, p: 35, pl acme ie

1890. Glossopteris cordata, Feistmantel, Mem. Geol. Surv. N.S.W.,

Palaeont., No. 3, p. 124.

1904. Glossopteris nephroerdicus, Etheridge, jun., Rec. Geol. Surv.

N.S. W., vol. vi, pt. 4, ps 315, pls. ving ix.

1926. Glossopteris nephroeidicus, Du Toit, The Geology of South

Africa, p. 274.

Among the specimens from Inhluzani, not far from Nottingham

Road, Natal, abundant in one particular layer, are several fronds,

few of them, however, complete, with a distinctive bi-auriculate base,

the auricles being clearly separated from the strong leaf-stalk and

forming two lobes of greater or less prominence. The midrib is

strong and striated, and tapers rapidly to the apex, while there is a

short petiole. The venation is open and coarse, with oblong-polygonal

meshes becoming somewhat smaller towards the margins, and is rather

like that of G. ampla or even G. retifera. In the upper part of the

frond the venation is inclined at an angle of from 60 to 80 degrees

to the midrib, but near the base it issues at a wide or even a right

angle, while in the lobes themselves it may even bend backwards

quite noticeably. In size, outline, and in the degree of lobing there

is a considerable variation, as the instances represented in fig. 1 will

show (A and B, No. 226w; C, No. 238w; D, No. 208w). In the

smallest leaves a reniform outline is approached with wide curved

spaces between the lobes and the stalk. The larger ones are more

“Sekt =

LS =

ae a SSS

eae

a a ES | | | Ll LA ao ee ae = =

= SS SS eee SS - : —— = = —o-

a 2 = —— st a ch} eee er = 2 = : ~ = = aa — = —

PN a eS SSS a a : SS yee So Se SSE SSS ET -- - <= —

Some Fossil Plants from the Karroo System of South Africa. 379

A a] eS)

\\ UI RAYaH

\ A wN

A\ \ NN YY

ASE

——

Baiera tenuifolia.

TEXT-FIc. 1.—A-D, Glossopteris cordata; E,

OG, di

/T|

CA Vy) TH) Wy Yt

MINS

Lite

Soy

380 Annals of the South African Museum.

elongated, while the gaps between the auricles and the stalk get

quite narrow. The apex varies from broadly ovate to obtusely

pointed.

The study of these curious forms, which in 1926 were provisionally

identified by me with Etheridge’s-G. nephroeidicus, suggests that the

auriculation is essentially a juvenile character and that it becomes

less prominent with age, being accompanied by corresponding changes

in the outline of the frond and also in the course of the venation.

The behaviour is nevertheless wholly different to that in G. retcfera,

which has its sides tapering regularly towards the base—impressions

of which species actually occur on the Natal slabs—otherwise their

general habit is similar.

As Etheridge * has pointed out in describing fronds of this character

from Illawarra, N.S.W., such forms because of their unique basal de-

velopment would not strictly fall within the current definition of the

genus Glossopteris. In 1849 Dana ft had described as ? Glossopteris

cordata a similar type of frond, also from Newcastle, but Etheridge

states that the apparent incurving at the base is merely due to a

fractured edge; Arber,t on the other hand, regards the plant as a

basal portion of G. ampla. Since the whereabouts of Dana’s type

appears to be unknown, and as so much doubt has been expressed

by two competent authorities concerning its interpretation, it would

perhaps be better to pass over this early determination, though the

re-examination of the original might possibly justify Dana’s views.

Fortunately in 1882, Feistmantel, following Dana’s account, described

under the same name from the Raniganj Group of the South Rewah

Coalfield of India an imperfect frond having a width of 60 mm., in

which one of the sub-auriculate basal lobes is rather poorly preserved ;

the venation is precisely like that to be observed in the Natal specimens.

Although Arber has tried to cast doubts on Feistmantel’s interpreta-

tion, the description and the accompanying drawing seem quite

definitely to uphold that original identification. Feistmantel further-

more in 1890 recorded this form from Illawarra, though he un-

fortunately gave no figure thereof. The fronds described by Etheridge

HT as G. nephroerdicus from the same locality differ only from the most

14 complete of the Natal examples (fig. 1, A), in their greater size, more

rounded outline, and in a slightly deeper basal lobing.

| Reviewing the evidence, I am inclined to regard G. nephroeidicus,

* Ktheridge, loc. cit., p. 316.

t+ Dana, in Wilke’s U.S. Explor. Exped., vol. x, p. 718, pl. xiii, fig. 5.

t Arber, The Glossopteris Flora, p. 52, 1905.

Some Fossil Plants from the Karroo System of South Africa. 381

Eth., and G. cordata, Feist., as most probably stages in the develop-

ment of the same species or at most as mere varieties thereof. Under

such circumstances Feistmantel’s title will, because of priority, have

to be adopted.

It should not be overlooked that the taxonomic relationship of

these several auriculate forms is actually of less importance than

the fact that such types, so different from the other well-known

species of Glossopteris, have now been placed on record from strata

in Australia, India, and South Africa, that because of their respective

floras must be of equivalent age. Their presence in these three

formations must therefore be regarded as constituting a link of no

small stratigraphical value.

The specimens are preserved in the South African Museum.

Locality Glandisrock (Parkside), near the Inhluzani Mountain,

Natal.

Horizon.—Lower part of the Lower Beaufort Beds.

It is instructive to note that the Durban Museum contains a

specimen of dark, highly fossiliferous shale containing fragments of

this form in addition to other species of Glossopteris ampla. It came

from a farm near Muller’s Pass over the Drakensberg, which is situated

17 miles to the south-west of Newcastle, the precise spot being un-

known, but the formation from which it came must unquestionably

be the Lower Beaufort Beds.

Genus ERETMONIA,* gen. nov.

ERETMONIA NATALENSIS, Du Toit.

Plate XL, figs. 9-12.

‘““ Small, leaf-like, pedicellate expansions of a narrow stalk, that

show a faint venation of Glossopteris character, and have spoon-

shaped ends in the shallow hollows of which were contained groups

of minute sac-like bodies resembling sporangia, of somewhat irregular

elongated outline and twisted shape, which exhibit an irregular

longitudinal] veining.”

In 1905 Arber + described in considerable detail from the Permo-

Carboniferous Beds of Port Stephens, New South Wales, certain

* From the Greek ‘‘ Hretmon ’’—a paddle.

+ E. A. N. Arber, Catalogue of the Fossil Plants of the Glossopteris Flora,

Brit. Mus., pp. 39-45, figs. 12-15, 1905; Quart. Journ. Geol. Soc., pp. 324-338,

pls. xxx, xxxi, 1905.

382 Annals of the South African Museum.

lal il ie

minute sac-like bodies occurring in groups in association with fronds

and scale-leaves of Glossopteris, and gave reasons for believing that

they represented the sporangia (microsporangia) of that genus. He,

indeed, figured a portion of a scale-leaf displaying what seemed to

be scars of attachment on its concave surface, and also showing

fragments of these small bodies adhering to the latter and possibly

in continuity therewith.

Similar bodies were recorded by Seward * from the Glossopteris-

bearing strata of Zululand, by Lundquist ¢ from those of Brazil,

and by Walkom t from those of New South Wales, while Thomas

has observed them in material from Natal. Structures in every way

identical are present in shales from near Bergville, Natal, but although

undoubted scale-leaves of Glossopteris occur on the same slabs, the

tiny bodies are in very close relationship instead, with small, spoon-

like leaves, or as they can perhaps better be described, pedicels, and

indeed would appear with a certain amount of probability to have

belonged to the latter. When the base of one of these pedicels is

missing, this curious structure closely resembles a scale-leaf of Glosso-

pteris, turned upside down and with the apex wanting. The further

fact that these impressions lie in the same planes with, and some-

times touching or even crossing tapering fronds of Glossopteris, hints

at a close affinity with that plant, a supposition that is strengthened

by the nature of the faint venation traceable on the pedicel surfaces.

It is indeed believed that the evidence to be submitted rather

favours the view that these structures are of a staminate character,

and represent reproductive organs belonging to the genus Glossopteris.

Morphology.—The eleven examples that are available for study

show slight differences among themselves, which may largely be due

to growth, though it is also possible that such may in part be specific,

since more than one species of Glossopteris occur in association, the

most common forms present being G. browniana and G. indica. Four

of the specimens display the convex, the rest the concave side. The

three examples that are figured here will have to be regarded as

co-types.

Their length ranges from 15 to 35 mm., and their maximum width

from 5 to 8 mm. In general the narrow stalk expands distally in

regular fashion and ends in a rather bluntly rounded termination,

* A.C. Seward, Trans. Geol. Soc. S.A., vol. x, p. 84, pl. viii, figs. 7, 7a, 1908.

+ G. Lundquist, “ Fossile Pflanzen der Glossopteris Flora aus Brasilien,”

Kungl. Svensk. Vet. Akad. Handl., 60, No. 3, p. 8, 1919.

{ A. B. Walkom, Proc. Linn. Soc. N.S.W., liii, p. 561, pl. xxxvi, fig. 4, 1928.

Some Fossil Plants from the Karroo System of South Africa. 383

sometimes in an obtuse angle. Hxtending back from the apex, and

having a diameter of from one-half to three-fourths of the maximum

width of the lamina, is a shallow depression, deepest on the distal

edge and of a shape best described as corresponding to the mould

that would be made by a seed of the Samaropsis type. In what is

regarded as being a youthful example (fig. 9), the lamina is narrow

and the depression therein shallow and of relatively small diameter ;

in the broader and presumably older example (fig. 10), the depression

is relatively wider, longer, and deeper, while the lamina is reduced

‘to a narrow wing on either side.

The strong veining of the stalk becomes fainter in the expanded

portion, but under oblique illumination it can be made out that the

venation, on both exterior and interior of the lamina, is precisely like

that seen in the youthful leaflets of Glossopteris, of which there are

examples scattered over the slabs, for the secondary veins anastomose

in the paddle-shaped expansion and show also a certain amount of

convergence towards the apex. They are, furthermore, not deviated

in passing from the flat part of the lamina into and through the spoon-

like hollow. Where the film of coaly matter has been removed the

cell structure is visible as an impression on the shale, though the finest

details cannot be made out. The concave surface in nearly all the

specimens has lost the organic film, but in specimen 264w some of the

material is still left as a cracked skin of coal and upon it occur several

slight prominences, that may perhaps represent points of attachment

of the supposed sporangia ; on 255w, fig. 11, however, no irregularities

could be made out on the interior cast. It is nevertheless not improb-

able that the presumed “ scale-leaf ”’ of Glossopteris figured by Arber

(his fig. 15) was merely an incomplete example of one of these pedicels.

The general resemblance to some of the associated narrow and

gradually expanding Glossopteris leaflets has already been mentioned,

which is even more marked in the case of the small scale-leaves ascribed

to that plant, for these are frequently markedly convex and sometimes

spade-like. The sharp-edged spoon-like depression serves, however,

to discriminate the form under description. If, on the other hand,

these organs are viewed as constituting the sporangiophores of

Glossopteris, a further possibility can be suggested, namely, that, after

having shed their sporangia, they might nevertheless have continued

to increase in size, to have broadened and flattened, and ultimately

have developed into normal Glossopteris fronds. Such a suggestion is

not altogether speculative, since one specimen (294w) shows a paddle-

shaped leaflet 20 mm. long and 9 mm. wide, which from its typically

384 Annals of the South African Museum.

developed venation would be identified as a youthful Glossopteris were

it not for the presence of a sharp-edged “‘spoon”’ near the apex ;

furthermore, its stalk is short and relatively broad (1-5 mm.).

In several instances the support of the pedicel is remarkably long

and slender ; in one case it is 24 mm. long and less than 1 mm. wide.

In two instances two pedicels precisely similar in size, shape, and

colour lie close together on the one slab, 262w, which fact indeed sug-

gests that those organs were borne not singly, but in pairs at least,

quite possibly in a cluster, either from a common axis or from a

strobilus. In none, unfortunately, is the extreme base preserved,

though in several there are signs of a faint expansion.

Sporangia.—While some isolated sac-like sporangia or else groups

of two or three occur here and there on the shale, they are found in

several instances densely clustered in most suspicious fashion alongside

pedicels, a closer study showing that in one example (262w, fig. 10)

some are resting upon the concave surface of the pedicel, and that in a

second (255w, fig. 11), which is viewed from the convex side, the lamina

has been pressed down over a few of these bodies, so that their out-

lines and even some of their ornamentation are still discernible

through the covering lamina. In more than one instance, e.g. fig. 11,

they are spread out along the one side of the pedicel, just where the

latter commences to widen, and next the edge of the hollow.

The circumstances indeed strongly suggest that the sporangia

have been squeezed out of the spoon-like hollows of the pedicels,

though there is unfortunately nothing to indicate how the tiny sacs

could normally have been retained in that position.

The bodies are from 1 to 2 mm. in length, rudely triangular, oval

or somewhat irregular in shape or else twisted, though details of the

surface can only be made out where the carbonaceous matter has been

removed, leaving a sharp cast of the exterior, which shows the peculiar

crinkly groovings so well recorded by Arber and Lundquist. Arber,

Walkom, and Dr. and Mrs. H. H. Thomas all failed to obtain spores upon

treatment of such bodies, though the two latter recovered seeds and

winged pollen-grains from Glossopteris-bearing shales in Natal. Some

of the coaly impressions of these bodies from the specimens under

description were digested in hydrofluoric acid, etc., and yielded a

few firm aggregations of yellow, ovate, but not winged spores about

0-02 mm. in size, though admittedly they might have come from the

small quantity of adherent matrix that perforce could not be elimi-

nated from the sample treated. Unfortunately further tests could not

be made save by damaging the types, and for the present the matter

Some Fossil Plants from the Karroo System of South Africa. 385

cannot be settled. Thomas * has indeed suggested that these bodies

are more like ramental scales, but their sometimes regular grouping

and their peculiar positions in relation to the pedicels in question are

more suggestive of their being, as Arber and Lundquist have supposed,

microsporangia.

Relationships.—The supposed sporangia were found by Arber

associated with Glossopteris browniana, but by Lundquist with G.

undica, while Seward’s examples from Zululand and those from Natal

under description occurred not only with G. browniana but with several

other species. The specific connection must for the present remain

open ; though either or both G. brownana and G. indica are appar-

ently signalled.

In 1931 + Walkom described under the name of Nummulospermum,

in intimate association with fronds and scale-leaves of Glossopteris, a

seed which he regarded as being the megasporangium of that plant.

Its anatomical features seem to favour its reference to the Trigono-

carpales, but its mode of attachment to the sporangiophore remains

unknown. The association of seeds of the Samaropsis type with

Glossopteris has already been noted and remarked on by several

palaeobotanists, a fact that gives some support to his views. Nummulo-

spermum has not yet been found in the material from Natal. Arber,

Seward, and Walkom are in general agreement that the twisted bodies

studied by them are with some probability the microsporangia of

Glossopteris, which the new data secured from the Natal specimens

distinctly support.

Assuming that the structures called Hretmonia represent the stami-

nate sporangiophores of Glossopteris, we can indulge in a little specu-

lation concerning their mode of arrangement and their botanical

affinities. Certain features rather suggest analogies with the Cycade-

oidea, though of course the analogy should not be pressed too far.

It is interesting to note that in his comparison of the supposed micro-

sporangia Arber { drew attention to the resemblances shown to the

living cycad, Stangeria paradoza. We can, for instance, visualise the

pedicels as packed closely together with their concave faces inwards,

while around them or between them were set the bract-like scale-leaves

(of Glossopteris), and outside the latter the normal fronds. The

* H. H. Thomas, Rept. Brit. Assn. for 1925, p. 364.

+ A. B. Walkom, Quart. Journ. Geol. Soc., Ixxvii, p. 289, pl. xxi, 1921 ; Queens-

land Geol. Surv., Publ. No. 270, p. 24, pl. vii, 1922.

+ E. A. N. Arber, “On the Sporangium-like Organs of Glossopteris browniana,

Brong.,’’ Quart. Journ. Geol. Soc., Ixi, p. 324, 1905.

VOL. XXVIII, PART 4. 27

SS nee ne

= SS eR

= —— SS :

a = — ae wpe : ——

a a eS IS —_ = -

: eae

~aen 7S

= \

—— ETRE

386 Annals of the South African Museum.

alternative view that the so-called microsporangia are ramental scales

is manifestly not inconsistent with the above.

It has already been pointed out that the shape of the spoon-like

hollows in the pedicels would just fit seeds of the Samaropsis type,

or even, save for size, of Nummulospermum ; this seems more than

mere coincidence. It is indeed conceivable that Nummulospermum

may have been borne on ovuliferous scales of somewhat similar

pattern to Hretmonia, either on the same strobili that carried the

microsporangia or on separate structures. It is nevertheless not easy

to explain why under such an hypothesis more than dismembered

fragments of the fructification should not have been discovered long

ago, taking into consideration the abundance of Glossopteris-bearing

material that has been available for examination.

Granted the correctness of Walkom’s interpretation of the nature

of Nummulospermum, the additional evidence furnished by Hretmoma,

if rightly interpreted, would go far towards indicating the cycado-

filicean affinities of that distinctly puzzling plant Glossopteris, a view

which as a matter of fact is not at all out of harmony with recent

ideas on the subject. The Natal locality is hence worthy of a further

careful search for more and better material.

The specimens concerned (Nos. 255w, 262w, 264w, 265w, 274w, 276w,

and 294w) are all preserved in the South African Museum, Cape

Town.

Locality.—A little to the north-west of Bergville, Natal.

Horizon.—Top of the Lower Beaufort Beds.

Relationships of the Natal Lower Beaufort Flora.

Of the forms described above :

1. Schizoneura gondwanensis, Feist., has been noted by Arber * in

material collected from the Mount Waller region of Nyasaland, but

none of the specimens was figured. On the other hand, the ribbed

stems described and figured by von Brehmer tf from the Uluguru

region of Tanganyika Territory have, owing to the absence of attached

leaves, only doubtfully been ascribed by him to this species.

It is common in India in the Karharbari and Damuda Series, and

particularly in the Ranigan] Stage (Permian), though smaller but

otherwise identical leaflets are preserved in the Panchet Series (Lower

* KE. A. N. Arber, A. R. Andrew, and T. E. G. Bailey, Quart. Journ. Geol. Soc.,

Ixvi, p. 238, 1910.

+ W. v. Brehmer, Botan. Jahrb., 51, p. 409, fig. 3, A-C, 1914.

Some Fossil Plants from the Karroo System of South Africa. 387

Triassic). It has been recorded as S. australis by Etheridge,* from

the Newcastle Series of New South Wales (Permian)—redetermined

as being that Indian species by Arber t—and also, though with some

doubt, by Lundquist { from the Bonito Coal Measure of Brazil (Upper

Carboniferous or Lowest Permian). Zeiller § mentions a very similar

form under the name of S. paradoxa from beds in Madagascar that

are probably Triassic.

2. Phyllotheca australis, Brong. (=P. indica, Bunb.), comes from

_ the Raniganj Stage of India, and typically from the Newcastle Series

of New South Wales, both Permian, and has been recorded by Halle,||

Seward, and Walton { from the Lower Gondwana Beds of the Falkland

Islands, but has been reported by Walkom ** from the Ipswich Series

of Queensland and from the equivalent beds of Tasmania (both

Upper Triassic).

3. Sphenophyllum speciosum (Royle) is known from the Upper

Wankie Sandstone of Southern Rhodesia, considered by Walton tT as

Permo-Carboniferous or Permian, from the Barakar and Raniganj

Stages of India (Permian), from the Lower Bowen Series of Queens-

land by Walkom ti (“ Permo-Carboniferous ’’), and from Korea by

Kawasaki,§§ while a closely allied species, S. seno-coreanum, Yabe,||||

characterises the Upper Shihotse Series of China (Permian). Signi-

ficant is the absence from Natal of any of the associated Hurasian

species such as S. thonw and S. oblongifolium, that have recently

been discovered at Wankie.

4. Sphenopteris alata (Brong.) comes from the Newcastle Series

of New South Wales, but is also very like S. hughes: from the Ranigan]

Stage of India (both Permian).

5. Glossopteris.—While the species G. browniana, indica, angustr-

* R. Etheridge, jun., Rec. Geol. Surv. N.S.W., iti, pt. 3, p. 77, pl. xii, 1893;

Ibid., iv, pt. 1, p. 32, pl. vii, fig. 1, 1894; zbid., vii, pt. 3, p. 234, pls. xlvu, xlix, 1903 |

+ E. A. N. Arber, The Glossopteris Flora, p. 9, 1905.

t G. Lundquist, Kungl. Svensk. Vet. Akad. Handl., 60, No. 3, p. 6, pl. viii, fig. 2,

1919.

§ R. Zeiller, Compt. rendus, 153, p. 230, 1911.

| T. G. Halle, Bull. Geol. Inst. Upsala, xi, p. 115, 1911.

q A. C. Seward and J. Walton, Quart. Journ. Geol. Soc., lxxix, p. 318, pl. xix,

figs. 3, 4, and 6, pl. xxi, fig. 16, text-fig. 2, 1923.

** AB. Walkom, Queensld. Geol. Surv., Publ. No. 252, p. 32, pl. i, fig. 5, 1915;

Roy. Soc. Tasmania, Pap. and Proc., p. 74, 1925.

+t J. Walton, Geol. Surv. S. Rhodesia, Bull. 15, p. 64, pl. A, fig. 2, 1929.

tt A. B. Walkom, Queensld. Geol. Surv., Publ. No. 270, p. 7, pl. i, figs. 3, 4, 1922.

§§ S. Kawasaki, Bull. Geol. Surv. Chosen, iv, pt. 1, 1925.

\||| I. G. Halle, Palaeont. Sinica, Ser. A, ii, fasc. i, p. 47, 1927.

388 Annals of the South African Museum.

folca, and ampla characterise beds extending from Upper Carboniferous

to Lower Triassic in various parts of Gondwanaland, there are three

species with a more restricted vertical distribution.

(a) G. cordata, Feist., is so far only known in the Newcastle Series

and Raniganj.

(b) G. retifera, Feist., characterises the Damuda Series of India,

but curiously does not seem to have yet been found in India, though

it has been recorded by Kurtz * from Argentina in beds considered

as Permian. This species is widespread in South Africa, having been

recorded from Graaff Reinet (Cape), Molen River, north of Harrismith

(Orange Free State), Glencalder, near Newcastle, and Glandisrock,

near Nottingham Road, both in Natal,+ and at Somkele in Zululand ft

—in each case from Lower Beaufort Beds.

and has been described by Walkom § from Belmont, near Newcastle,

New South Wales. It occurs at Molen River, near Harrismith, and

Glandisrock, near Nottingham Road,]|| but also occurs sporadically in

the Molteno Beds (Upper Triassic) of the Upper Umkomaas Valley,

Natal.

To sum up, these common species tend to confirm the correlation

of the Lower Beaufort Beds with the Raniganj Stage (Upper Damuda)

of India, and the Newcastle Series of New South Wales, a view that

is supported in the case of the first-mentioned by the fact that the

vertebrate remains contained in the Middle Beaufort Beds are closely

allied to those obtained from the Panchet Series of India (Lower

Triassic).

The alliances are unquestionably Permian, and on the whole rather

with the Lower than the Upper division thereof. On the other hand

the vertebrate remains would indicate pretty strongly an Upper

Permian (Zechstein) age. This is once more an illustration of the

peculiar observation that, judged by its contained fossil animal life,

a formation would tend to appear somewhat younger than would

be deduced from the study of its vegetation.

* F. Kurtz, Act. Acad. Nac. Cien. Cordoba, vii, pl. x, fig. 90, 1921-22.

+ Arber, loc. cit., p. 85, 1905.

t A.C. Seward, Trans. Geol. Soc. 8.A., x, p. 85, pl. viii, figs. i, 10, pl. ix, figs. 5, 6,

1908.

§ A. B. Walkom, Proc. Linn. Soc. N.S.W., liii, pt. 5, p. 558, figs. 6, 6a, 6b, 7, 1928.

|| Arber, loc. cit., p. 87, 1905.

q A. L. du Toit, Ann. 8. Afr. Mus., xxii, pt. 2, p. 364, fig. 15a, 1927.

Some Fossil Plants from the Karroo System of South Africa. 389

C.—STORMBERG SERIES.

GINKGOALES.

Genus BAIERA, Braun.

BAIERA TENUIFOLIA, Johnston.

Text-figure 1, HE, F (see p. 379).

1887. Johnston, Pap. Proc. Roy. Soc. Tasmania, 1886, p. 176,

pl. i, fig. 2, a-e.

1888. Johnston, Geol. Tasmania, pl. xxvii, fig. 2, a-e.

1925. Walkom, Pap. Proc. Roy. Soc. Tasmania, 1924, p. 85,

figs. 14-16.

In 1929 a few specimens from the Molteno Beds of Natal exhibited

in the Durban Museum were provisionally identified by Dr. A. B.

Walkom with this Tasmanian species. Thanks to the kindness of

the Curator of that institution, Mr. E. C. Chubb, F.Z.S., I have been

enabled to study this material and to confirm the diagnosis, two

fairly complete specimens being reproduced as the result.

The smaller (No. 3163) well shows (fig. 1, F) the regular and repeated

dichotomy, one median vein traversing each of the leaves, which are

about one millimetre wide. In the larger one (No. 3176, fig. 1, E)

some subsidiary strands can in places be made out in addition to the

median vein that runs along the narrow coriaceous leaves. These

veins have been omitted from the accompanying figures.

In every respect these plants agree with the species B. tenuifolia

from the Mesozoic of Tasmania associated with Thinnfeldia as

described by Johnston, a copy of whose description and drawings

was kindly made for me by Dr. Walkom of Sydney, by whom this

form was redescribed from Tasmania.

It is not known elsewhere, and its discovery in South Africa is

hence of importance, though not unexpected, in view of the close

similarity of the Thinnfeldia Floras of the two countries.

The specimens in question come from the highly fossiliferous zone

of black shale in the Molteno Beds of The Waterfall, Upper Umkomaas

Valley, Natal, from strata regarded as of Upper Triassic age.

A small slab (No. 1953R) collected by Dr. A. W. Rogers, Director

of the Geological Survey, from the farm Ziska, N’Wanedzi River,

to the east of Messina, Transvaal, examined by me in 1929, was found

to carry examples of the same species, the leaves being a little broader,

390 Annals of the South African Museum.

about 2 mm. wide (and hence more like those of Johnston’s type),

traversed by one or more veins. These form impressions on a mere

film of reddish shale in a fine-grained sandstone well below the

Bushveld Sandstone, and more than 200 feet above the local base

of the Karroo System.

The find is of importance since it indicates a Stormberg age for

the containing strata at Ziska, this being furthermore the first

example of identifiable plant remains from that series in the

Transvaal.

This and a second slab (No. 1952R) are preserved in the Geological

Survey collections at Pretoria.

PLANTS OF UNCERTAIN POSITION.

Genus JOHNSTONIA, Walkom, 1925.

JOHNSTONIA CORIACEA (Johnston).

Text-figure 2, A, B, C, D, E.

1886. Rhacophyllum coriaceum, Johnston, Pap. Proc. Roy. Soc.

Tasmania (1887), p. 170; Geol. Tasmania, pl. xxvi, fig. 9,

1888.

1925. Johnstonia coriacea, Walkom, Pap. Proc. Roy. Soc. Tasmania,

pp. 79-81, figs. 6-8.

1927. Johnstonia coriacea, Du Toit, Ann. 8. Afr. Mus., xxu, p. 360,

figs. 12 D, 13 B.

Slabs of shale sent by Mr. G. Pyke from the Matatiele Commonage,

East Griqualand, are crowded with impressions of this uncommon

plant, unfortunately none of them complete, though several show the

base, a portion not hitherto known in the type.

The fronds are dichotomously branched and strap-shaped, with

margins entire, though a faint sinuosity is occasionally noticeable ;

| within the fork one of the laminae overlaps the other slightly. Some

of them must have exceeded 12 cm. in length and had a width of just

over 1 cm., but small ones only 6 mm. wide are also present. The

portion below the fork is usually much shorter than either of the

arms, and, contracting proximally, ends in a very delicate base

(fig. 2, A (9525) and B (9539)). Distally each branch tapers slightly

and terminates in a rounded or sub-pointed end (fig. 2, C (9521)).

Instructive is the case recorded in fig. 2, D (9505), of a second dicho-

3 = en

| "4

all

if i

rel

} i]

oA —— =

ates

Some Fossil Plants from the Karroo System of South Africa. 391

tomy of one of the branches just below the apex. The mid-vein is

prominent and pitted, and gives off secondary veins at an acute

angle (fig. 2, E (9520)), that usually bifurcate twice in their curving

passage to the margin. In the broader fronds the veining is markedly

clearer and more open.

A slight asymmetry in the tapering base, coupled with a twisting

Trext-FIc. 2.—A-K, Johnstonia coriacea.

of the lamina at the very end, suggest that the fronds grew from

a stalk in the form of a verticil, a presumption that obtains support

from the occurrence in a few cases of two or more fronds lying side

by side on the same slab or partially overlapping, but none were

discovered in actual attachment to a stem. The delicate character

of the base is a matter of surprise when the extent of the lamina is

considered, and it may be conjectured that the leaf was very flexible

or that it grew in a drooping position.

300 oh Annals of the South African Museum.

The fossils occur as impressions, and all attempts to obtain prepara-

tions from them with hydrofluoric acid failed, while the collodion

method produced nothing of real value. Nothing fresh was therefore

found that could throw any light upon the affinities of this plant,

which remain problematical.

J ohnstonia corvacea has been recorded from the Mesozoic of Tasmania

and the Molteno Beds of the Upper Umkomaas Valley, Natal, but the

genus is stated by Mr. W. M. Edwards of the British Museum in a

letter to the writer to be represented in some material from Cacheuta,

Argentina. Such would not be unexpected, for one of Kurtz’s draw-

ings (pl. xx, No. 285%*) strongly resembles the apical portion of

a branch of Johnstonia.

The specimens are preserved in the South African Museum, Cape

Town. 7

Horizon.—Molteno Stage of the Stormberg Series.

CYCADOPHYTES-BENNETTITALES.

Sus-cenus ANOMOZAMITES, Schimper.

PTEROPHYLLUM (ANOMOZAMITES) INCONSTANS (Braun), 1843.

(For synonymy, see Du Toit, Annals 8. Afr. Mus., vol. xxu, pt. 11,

Pos ple xexay. hice le)

In 1927 a small piece of a frond in the South African Museum,

labelled 399u, was described under this name from the black Molteno

shales of The Waterfall, Upper Umkomaas Valley, Natal (Du Toit,

p. 379). A slab of similar shale coming from apparently the identical

horizon, though from a spot two and a half miles to the north on

the farm Vergelegen, carries impressions of two more-complete fronds

belonging to the same cycadean species, one of them 13 cm. long, and

only wanting the apex, and the other 18 cm. long, with the extreme

base missing. In the mid-section they are each just over 1 cm.

wide, and thus narrower than the majority of described specimens,

but display in precise fashion the characters of this well-known

species, that is to say, short, nearly opposite, approximately rec-

tangular segments with squarish or rounded ends, unequal in breadth

and with confluent bases, while the strong veins emerge more or less

at right angles to the rachis, and generally fork on or just after

leaving the latter. In the basal part of the frond the segments

* F. Kurtz, Act. Acad. Nac. Cien. Cordoba, vii, 1921-22.

Some Fossil Plants from the Karroo System of South Africa. 393 |

and are directed slightly forward. The rachis is stout and strongly

ribbed.

These two all-but-complete fronds show that a correct determina-

tion was made of the fragmentary specimen from The Waterfall.

The slab (No. 4291) is preserved in the McGregor Museum, Kimberley.

Horizon.—Molteno Stage of the Stormberg Series.

become irregular or vestigial, while in the apical part they get narrower

EXPLANATION OF PLATES. |

Puate XXXIX.

Stigmaria ficoides, Brongn. ; halt natural size.

PuatTE XL.

FIG. f !

1. Schizoneura gondwanensis, Feist. ; natural size. }

ae ie me oA single leaf, natural size. )

3. Phyllotheca australis, Brongn. ; leaf sheath, lateral view, x 14 times. |

4, ae HS Pe A seen from below, x 14 times. |

5, 6. Sphenophyllum speciosum (Royle); x 14 times. f

7. Sphenopteris alata, Brongn. ; natural size. \

8. - o - pinnule, x 2 times. iH

9,10. Hretmonia natalensis, gen. et sp. nov. ; interior view, x 2 times. \

te A aS Be a exterior view, x 2 times.

12. a a i s microsporangia, x 15 times. I

Plate XXXIX,

Ann. 8. Afr. Mus., Vol. XXVIII.

STIGMARIA FICOIDES

Lid.

Neill & Co.,

XXVIII. Plate Xi. .

: ill & Co., Lid.

A. L. du Toit del. Neill Oro

' KARROO FOSSIL PLANTS.

( 395 )

11. Some Fossil Plants from the Gondwana Beds of Uganda.—

By Auex. L. pu Torr, D.Sc., F.G.S.

(With Plate XLI.)

In 1920 the Director of the Geological Survey, Mr. E. J. Wayland,

A.R.C.S., reported that the sedimentary strata underlying the town of

Entebbe could be ascribed to the Karroo or Gondwana System because

of the finding in them of fragile plant remains. In 1923 certain of

these impressions were submitted to Professor A. C. Seward, F.R.S.,

of Cambridge, who provisionally identified the following forms :— *

Glossopteris indica, Brongniart, Psygmophyllum sp., and Noeggera-

thiopsis (Cordaites) sp., which, he remarked, were all genera character-

ising the Ecca Beds of Vereeniging, Transvaal.

By 1922 various prospecting shafts coupled with a borehole had

proved the presence beneath the thick mantle of lateritic material of

a deep synclinal basin made up mainly of clay-shales together with

some arenaceous layers and thin partings of coal.t Further boring at

Entebbe showed the maximum thickness of these beds to exceed 1070

feet, but, as the machine was unfortunately of the percussion type, no

cores were obtained of the strata pierced.

In 1929 the Director kindly forwarded to me a number of specimens

mostly of highly weathered ochreous shales from scanty surface

exposures, collected either by himself or by Mr. T. Hirst, A.R.C.S.,

though it was not found possible to examine them carefully until

early in 1930, when a request for additional material met with a prompt

response. The writer accordingly desires to proffer his thanks to

Mr. Wayland for presenting these specimens and for permitting the

publication of the results of such examination.

Although the collections are not large, the pieces of shale usually

very small and the fossils contained therein rather fragmentary, the

remains are deemed to be worthy of a full description, since the deposits

that have yielded them are situated right on the Equator and con-

* Ann. Rept. Geol. Surv. Dept. Uganda for 1923, p. 8, 1924.

+ Jbid., 1926, p. 17, 1927.

_ 3 Po

396 Annals of the South African Museum.

stitute the most northerly ones in Africa within which the “ Gloss-

opteris ” or “Southern Flora’ is yet known to occur. Doubtless

very much better material will become available when the Entebbe

basin has been explored by means of a core-drill.

To the forms mentioned above have now to be added the following :

Gangamopteris cyclopteroides, var. attenuata, Feistmantel, Cyclo-

dendron Leslit (Seward), and Cornucarpus sp.

Affinities of the Flora.—Including those plants identified by A. C.

Seward, the general correspondence of the Uganda assemblage with

the flora of the Ecca Series of the south is unquestionable. In South

Africa the genus Gangamopteris, while characteristic of the Ecca

(Lower Permian), is not as yet known to extend into the succeeding

Lower Beaufort Beds (Upper Permian). The variety attenuata has

been identified at Kimberley * from either the uppermost Dwyka

Shales or the extreme base of the overlying Ecca Shales, and by the

writer at Komati Poort in the Transvaal from coal-bearing Lower

Karroo rocks (Ecca) and at Artesia in the Bechuanaland Protectorate

from the Middle Ecca “‘ Coal Measures.” Outside of Africa it has been

recorded from the Lower Gondwanas (Talchir and Karharbari) of India,

from the Mersey Series of Tasmania (which is the equivalent of the

Ecca Series), from Bajo de Velis in Argentina, and from strata at

Ankazomanga in the south-west of Madagascar.f

Cyclodendron Lesliz, Psygmophyllum, and Cordattes are representative

of the Ecca “‘ Coal Measures ”’ of the Transvaal. Glossopteris indica,

on the other hand, is a widespread species with far too great a vertical

range for close zoning. The seed Cornucarpus has its nearest repre-

sentatives among forms occurring in the Upper Carboniferous and

Permian of Europe.

To sum up, an age not younger than Lower Permian is indicated for

the flora. The specimens in question are preserved in the South

African Museum, Cape Town.

* Q. Feistmantel, Abh. bdhm. Ges. Wiss. Prag., vol. vii, pt. iii, p. 37, pl. iv,

fig. 2, 1889.

7 P. H. Fritel, Comp. Rend. Acad. Sci., vol. clxxi, p. 963, 1920.

Some Fossil Plants from the Gondwana Beds of Uganda. 397

SYSTEMATIC DESCRIPTIONS.

Genus GANGAMOPTERIS, McCoy, 1861.

GANGAMOPTERIS CYCLOPTEROIDES, var. ATTENUATA, Feistmantel.

Plate XLI, fig. 1.

1879. Feistmantel, Palaeontologica Indica, vol. iii, pt. i, p. 14.

(For references see Arber, The Glossopteris Flora, pp. 104-5, 1905.)

This small frond No. 1 with its counterpart No. 2 is sufficiently com-

plete to enable the fossil to be ascribed definitely to this genus and

not to Glossopteris. Lacking the apex, it shows the lower two-thirds

of a frond narrowing gradually towards the base, where there is a

fairly rapid contraction. There is also a slight asymmetry and a

general curving of the leaf, while the upper surface is convex.

Although preserved as a pink impression in a shale that has become

decomposed to a soft ochreous material, the venation is quite distinct.

There is no true mid-rib, the median nerves, somewhat stouter towards

the base, running in a more or less parallel fashion, but clearly an-

astomosing. The lateral nerves arise by repeated dichotomy from the

(compound) median axis, forming narrow elongate meshes, that become

smaller towards the edges of the frond.

Such a venation is typical of Gangamopteris and distinguishes the

specimen from the only species of Glossopteris for which it might be

taken, G. indica, var. decupiens. Since the apex is missing, its reference

to Gangamopteris angustifolia, McCoy, is not excluded, but it appears

to agree excellently with the variety attenuata from India, figured by

Feistmantel in 1879 (pl. xvi, fig. 5), but is smaller than the example

described by him from Kimberley, South Africa,* ten years later

(pl. iv, fig. 2). It accords well with a frond in the South African

Museum (No. 8663) collected by the writer along with larger fronds of

G. cyclopteroides from a well at Artesia Siding in the Bechuanaland

Protectorate.

From the viewpoint of strict nomenclature it might be remarked

that the species G. cyclopteroides ought to be designated G. obovata

(Carruthers), as has been done by Lundquist,f but the former name

has been so long established, and is so universally in use, as to have

made all other palaeobotanists reluctant to enforce the rule of strict

+ G. Lundquist, ‘‘ Fossile Pflanzen der Glossopteris Flora aus Brasilien,’’ Kungl.

Svensk. Vetens. Akad. Handl., Bd. 60, No. 3, p. 14, 1919.

398 Annals of the South African Museum.

priority. The title here accepted is accordingly that conferred by

Feistmantel. .

Locality.—From beach below Geological Survey Office, Entebbe.

Genus GLOSSOPTERIS, Brongniart, 1822.

GLOSSOPTERIS cf. INDICA, Schimper, 1869. ~

A small fragment (Nos. 42 and 43) embedded in a hard reddish shale

of a folded leaf that must have been at least 32 mm. wide can be

ascribed to the species indica rather than to G. browniana.

The interest centres in the fact that this was collected by Mr. T.

Hirst from Bugiri in the Eastern Province, approximate latitude

0° 30’ north and longitude 33° 45’ east and about 100 miles to the east

of Entebbe.

Genus CYCLODENDRON, Krausel, 1928.

Under the name of Bothrodendron Leslii Seward * described certain

spirally pitted stems from the Ecca “ Coal Measures ” of Vereeniging,

Transvaal, with which he subsequently compared a fragment from the

*“‘ Coal Measures ”’ of Newcastle, Natal. The assignment of those stems

to that well-known Huropean genus was, however, not universally

accepted, being called into question by Gothan some years later. In

1928 Krausel recorded from the (presumed) Ecca Beds of South-West

Africa stems almost indistinguishable superficially from the above,

and, as in a few places certain of the scars seemed to have attached to

them what he took to be leaves of a spinous character, he instituted

a new genus, Cyclodendron, for these forms, to which he incidentally

allotted the Vereeniging plants.

When investigating the material under description, the writer found

that, while the pitted stems from Uganda were practically indistin-

guishable from those figured by Krausel, the leaves discovered attached

to certain of them were on the contrary flat and lanciform. For that

reason the Uganda plants could not in strictness be referred to Cyclo-

dendron, though otherwise they agreed excellently with the types

thereof. In view of this unsatisfactory state of affairs Dr. Krausel was

asked kindly to re-examine his specimens, whereupon he reported in

a letter dated the 8th of May 1931 that the supposed spines were only

very poorly to be made out, and that a certain amount of doubt now

arose as to their precise nature, wherefore he was agreeable to such

* For references see synonymy given below.

Sa ee

‘ib

ae ae

= tied

Some Fossil Plants from the Gondwana Beds of Uganda. 399

alterations being made as were necessary for the re-description of his

genus. It fortunately happened that Dr. W. Gothan of Berlin and

Mr. W. N. Edwards of the British Museum were also enabled to examine

the types from South-West Africa, and both of these palaeobotanists

have concurred in the above expressed opinion. In consequence the

Uganda plants have now been allotted to Krausel’s genus, while his

diagnosis has been considerably amended. At the same time it should

be explained that, while there seems to be hardly any doubt as to the

generic and specific identity of the plants from South-West Africa and

from Uganda, there are certain differences in the case of those from

Vereeniging, as will be pointed out in the sequel.

Diagnosis.—‘ Stems bearing spirally arranged, raised leaf-cushions,

oval or slightly polygonal along their upper borders, each pierced by a

nearly central vascular pit and supporting a polygonal or diamond-

shaped area for the attachment of the leaves. Leaves, flat throughout

their length, nearly parallel-sided with lanciform points, traversed for

their full length by a single medial vein. Leaves attached to the

cushions in planes transverse to the latter and set out at a wide angle

to the stem.”

CYCLODENDRON LeEsuit (Seward).

Plate XLI, figs. 2-10.

1903. Bothrodendron Lesli, A. C. Seward, Ann. 8. Afr. Mus., iv,

pps cet-92, pl. xi, igs. 1-6.

1905. Bothrodendron Lesliu, EK. A. N. Arber, The Glossopteris Flora,

pp. 166-8, fig. 36.

1907. Bothrodendron sp., cf. Kidstoni, A. C. Seward, Trans. 8. Afr.

Geol. Soc., x, p. 83, pl. vii, figs. la, 1b.

1914. Bothrodendron Lesli, W. Gothan, Sonder. a. d. “ Branca-

Festschrift,” p. 15.

1928. Cyclodendron Leslii, R. Krausel, Beit. geol. Hrfor. deut.

Schutz, Heft 20, pp. 18-21, pl. i, figs. 3-10; pl. u1, fig. 1;

text-figs. 1-3.

1929. Cyclodendron Leslit, J. Walton, Geol. Surv. 8. Rhod., Bull. 15,

Pp. 66.

The material available for study is admittedly most fragmentary,

but there are fortunately included enough pieces to show that the

plant in question differs from any of the well-known palaeozoic

lycopodalean and coniferalean genera. While the bits of shale are

ageravatingly small, their thoroughly weathered nature has enabled a

> Ea. GS

—_

Ln = \ —=_ ~- —

————————

4 > oa

=e {Saree STS

=z -

2 SS a ee : SSS = SS = — = = = - =

> - = = = - - DASA S 5 >

400 Annals of the South African Museum.

good deal of plant structure to be made out by careful dissection of the

friable matrix. Several tiny specimens from the same source have

kindly been supplied by Mr. W. N. Edwards, together with sketches and

photographs of other fragments catalogued in the British Museum,

two of which (Nos. V 20944 and V 20958) are reproduced here (figs. 5

and 6). The writer desires at this stage to express his great indebted-

ness to Mr. Edwards for the interest that has throughout been shown

and for the help that he has given in discussing the relationships of

the plant under description.

Stems.—The largest and best preserved of these fragments (No oy)

from the beach below the Geological Survey Office, Entebbe, is a sub-

cortical mould (Plate XLI, fig. 2),50 mm. long and 12 mm. broad, show-

ing externally regular, smooth, elongate-cordate, protuberant leaf-

cushions that are spirally arranged, oval or slightly polygonal along

their upper edge, which is separated from the rest of the stem by a

shallow groove, but which slopes away evenly in the opposite or down-

ward direction. Each cushion is pierced, not quite centrally, but

more towards its upper side by a single pit, which marks the point of

emergence of the leaf bundle. On either side there may be a faint

grooving with downward convergence, that gives the stem a lepido-

dendroid aspect. There are no traces, however, of parichnos scars or

of any ligule. Although the matrix is not specially fine-grained, the

impression of the cuticle can be made out under magnification on

certain of the cushions as a fine vertical striation and even in places

as a network formed by rectangular to polygonal cells.

In its external characters the stem is in full agreement with those

figured by Krausel (see his pl. i, figs. 3-6). In none of the speci-

mens represented are there any indications of irregularity in the

spacing of the cushions, such as has been noted by Seward in the case

of the Vereeniging examples (see his pl. xi, fig. 1).

Where the fracture in the slab has passed sufficiently deep below

the surface of the fossil, the interior of the cushion can be made out ;

when, as viewed from within, it shows as a nearly circular depression

divided symmetrically by a narrow, vertical, medial ridge thickened

towards its upper end, which rather recalls a boy’s kite (Plate XLI, fig.

2at A). Structurally the vertical bar would function as a girder, and

from it evidently sprang the vascular bundle. Where the fracture has

not passed so deeply downwards the surface of the stem is found to be

covered with the diamond- or eye-shaped scars that are so well shown

in Krausel’s specimens (pl.i, figs. 3, 4, 6, 9, 10) and in the examples

Nos. V 20944 and V 20958, pictured in figs. 5 and 6. :

Some Fossil Plants from the Gondwana Beds of Uganda. 401

An important feature, which would otherwise have remained

almost unnoticed, is revealed by the study of the counterpart,

specimen No. 4, which shows inside the hollow replica of each cushion

a fairly deep depression, symmetrical and polygonal or eye-shaped,

which manifestly represents the cast of certain soft tissues that

composed the base (pulvinus) to which the leaf was attached (fig. 3).

This drawing is not too good, as it has been found difficult to depict

satisfactorily the depression situated right at the bottom of the

mould of the cushion. An impression of this hollow made in plasti-

cene is more illuminating in that it shows a raised structure that

terminates in a diamond-shaped area (fig. 4), 3 mm. in length, set

transversely to the stem, on which can be made out, though faintly,

the horizontal line of attachment of the leaf-base, and in a few favour-

able cases the point of emergence immediately below it of the vascular

bundle.

In one of the fragments (No. 37) some of these raised diamond-

shaped areas are quite well preserved, but in the majority of the

examples the stems have been much pressed and the cushions or

pulvini flattened, so that one observes merely diamond or eye-shaped

markings spirally arranged on the flat surface. The roughly circular,

seal-like impressions shown on the left of fig. 3 and the right of fig. 4

have evidently been formed by squashing of the corresponding

cushions. Undoubtedly differences in the mode of preservation of

the stems have been responsible for great differences in the character

of the markings on the impressions in the shale. It has been noticed,

furthermore, that in the case of the broader stems the cushions are

fainter and set farther apart, though retaining nearly the same

average size as those of the smaller shoots. These facts suggest that

the leaves were shed with age and that the pulvini then readily

decayed or scaled off. As regards size the widest stem is about

20 mm. across, but the majority are less than half that in breadth.

Leaves.—In the example under description (No. 3) a few of the

leaves are still in attachment along the sides of the stem (Plate XLI,

fig. 2), but their preservation in this specimen is poor and their habit

not easy to make out. Owing to the transverse compression of the

lamina, the leaves are no longer quite flat and give a false impression

of rigidity. Fortunately, among the many fragmentary leaflets that

are scattered through the shales a number have been found attached

in continuous spirals to small stems, which in a few cases agree in

habit with the larger ones and that described above. Once the

fact was established that these leaves actually belonged to such

VOL. XXVIII, PART 4. 28

402 Annals of the South African Museum.

Cyclodendron stems, an explanation was forthcoming for the fre-

quency with which both occurred in juxtaposition, though the state

of preservation was just insufficient to prove definitely any actual

attachment.

The leaves are from 25 to 35 mm. long (Plate XLI, fig. 7), lanciform

in outline, widening rather gradually from the base, but contracting

rapidly at the apex, which is pointed. The base is 25 mm. wide,

with faintly rounded angles and with a central shallow indentation,

while on the lower side is situated a tiny protuberance from which

arises the principal vein, that is usually traceable right to the apex.

While in certain specimens the primary vein is scarcely detectable,

in the majority of instances it is quite distinct, and it was noticed

further that there was a tendency for the strand to be more marked

on the lower than on the upper surface, a feature characteristic of the

lepidophylla of the northern hemisphere, as Mr. Edwards has pointed

out ina letterto me. The lamina is also traversed by a very delicate,

almost parallel or slightly diverging striation, that in a few favourable

cases can be made out not to converge apically, but to run out at

the distal edges in the manner of a Cordaites frond.

Where the leaves are lying at an angle with or are set on edge to

the planes of stratification, they are traversed by crinklings due to

the subsequent compression of the matrix. In certain instances,

indeed, this transverse crumpling has given to some of the leaves

_ the deceptive appearance of being spinous. The fragments of shale

show numerous cross-sections of these leaves, but even in their

extreme basal portions they display no signs of thickening, and are

hence quite unlike the trigonal, spinous appendages so characteristic

of the palaeozoic lycopods. Unfortunately, no traces remain of the

cuticle, but the lamina seems to have been delicate.

Mode of Attachment.—In three examples the fossilisation has been

such that a group of leaflets lie spread out in the plane of lamination

of the shale, yet firmly attached to an axis that is set obliquely to

the latter. In Plate XLI, fig. 8, of a fine specimen (No. 5) from the

Kuropean Hospital Quarry, at least two whorls or spirals are present

directly superposed, with the edges of the two sets of leaves over-

lapping one another next the axis. In another example (No. 7) from

the same spot (fig. 9) two whorls are preserved, and beneath the lower

one a leaf can be seen belonging to a third spiral, while the bending

down of the leaves at their very bases and their manner of attach-

ment to the cushions of the stem can sometimes be made out, which

is by means of the whole base and transverse to the stem; this is

Some Fossil Plants from the Gondwana Beds of Uganda. 403

more clearly indicated in the case of the tiny specimen, No. 8, fig. 10.

The slight twisting next the axis, that is often to be noticed, is clearly

due to pressure induced during fossilisation.

The regular disposition of the leaves in the upper whorl directly

over those in the lower one should be noted, as this is in accord with

the phyllotaxy of the associated stems. Ina third fragment (No. 6)

from this quarry eye-shaped leaf-scars can just be made out on the

narrow stalk to which the leaves are spirally attached. In a fourth

(No. 19) a single leaf is in attachment to a cushion belonging to a stem

9 mm. in breadth. In a fifth (No. 9) the leaves have been flattened

in the same plane as the narrow stem, and are diverging from the

latter at angles of about 45 degrees. Some overlapping naturally

occurs, while those leaves, originally set more or less transversely to

the lamination of the shale, have suffered compression and show an

irregular wrinkling, which tends to obscure their true nature in such

positions. Proximally some of them have been squeezed up until

they are channelled, and have come to look rather like spines, though

towards their apices the flat nature of the lamina is usually obvious.

Not improbably some distortion of this character in one of his speci-

mens may have misled Krausel.

There is some evidence that, as in the conifers, the leaves emerged

at acute angles in the younger shoots and formed tufts at the apices

of the latter, whereas in the larger and older stems the leaves stood

out almost at right angles, though apparently soon shed.

In the McGregor Museum, Kimberley, there is a specimen (No. 3312)

of a stem from the Vierfontein Colliery, Bothaville district, Orange

Free State, 80 mm. long and 36 mm. wide, which in the character of

the cushions and in their spacing is like the form under description,

though more closely resembling some of Krausel’s examples; the

cushions are from 3 to 4 mm. across. In regard to the form from

Vereeniging, described originally as Bothrodendron Leslw, nothing more

has been learned from an inspection of Seward’s types, which are

unfortunately without leaves, as are all the other specimens from that

locality. It was possible, however, to examine a small specimen,

30 mm. long by 20 mm. wide, from Vereeniging, kindly loaned by

Dr. T. N. Leslie, which showed unusual characters. The scars are small

and set 9 mm. apart, measured vertically ; seen from within they show

the peculiar, typical kite-like ridgings, but externally the cushions are

different, consisting of a regular annulus 2-5 mm. in diameter, out of

which projects a strong umbilicus. In this they are different from

those of the Uganda examples, though recalling certain described and

i. —eag — a

we >=

~~ ae a

——= Se

oe TE es

aie ee

| =a

———————— ESS -s : . . ——— = ~ —

— —— ——s ™: — —— es

ee . ja a ar : , = ee

OS a = — — == - -

404 Annals of the South African Museum.

figured by Seward (pl. xi, fig. la) showing a depression with a central

umbo. It is not unlikely that these variable features are largely due

to differences in the degree of decortication of the stem. It is not

unlikely, too, that with the shedding of the leaves the cushions tended

to alter in shape and to become more circular. The local crowding

of the scars in some of Seward’s types (pl. xi, figs. 1, 16) has not,

however, been noticed in the Uganda fossils. .

A ffinities.—Reference of this unusual form to any of the lycopod

genera such as Lepidodendron, Sigillaria, or Bothrodendron is definitely

excluded, first, by the absence of signs of parichnos scars or ligular

pits, and, secondly, by the peculiar nature of the leaves. So far as

the author has been able to discover from the available literature,

the northern palaeozoic lycopods were all characterised by rather

spinous foliar organs, swollen at the base, where the tendency

was towards a markedly trigonal cross-section. The only clear

lepidodendroid feature is the single primary vein in the leaf of

Cyclodendron.

In certain respects the Uganda plant recalls decidedly the Upper

Carboniferous and Permian genus Dicranophyllum,* in which each leaf-

cushion possesses a single vascular bundle-scar near its upper edge ;

each leaf, furthermore, consists of a narrow lamina that is traversed

by one or more veins to the apex, but, on the other hand, the leaf forks

into one or more branches at a certain distance from the base. The

systematic position of Dicranophyllum is uncertain, though this plant

is supposed to belong to the Coniferales and is thought to be allied to

the Cordaitales. In Cordaites itself, while the leaf-scars (with decurrent

bases) are spirally arranged, there is nothing to show that a single

vascular bundle existed, while the falciform leaves were traversed by

numerous sub-parallel veins.

Comparison can also be instituted with the casts of stems from the

Upper Carboniferous and Permian with spirally disposed cushions

having vertical slits in the apical portion due to leaf traces, described

as Tylodendron and Schizodendron by Weiss and EHichwald, and a rather

similar one from the Permian as Hapaloxylon by Renault.f Zalessky {

has, it is important to note, described from the Permian of Siberia as

Caenodendron and Angaradendron two kinds of spirally pitted stems,

which recall in very many respects Cyclodendron, but, as no leaves

* A.C. Seward, Fossil Plants, vol. iv, p. 93, 1919.

+ Ibid., pp. 282 and 286, 1919.

t M. D. Zalessky, “‘ Flore paléozoique de la série d’Angara,’”’ Mém. Comité

géol. Petrograd, livr. 174, pls. xiii, 1x—Ixiii, 1918.

Some Fossil Plants from the Gondwana Beds of Uganda. 405

occur in attachment with either, and as no detailed descriptions have

yet been given, closer comparisons cannot be made.

Reviewing the very slender evidence it would seem that, while

Cyclodendron shows certain affinities to the Lycopodiales, it belongs

with more probability to the Coniferales ; unfortunately the absence

of any woody tissue leaves us in ignorance of the internal structure of

the stem.

Distribution.—Cyclodendron has so far been found in the Coal

Measures of Vereeniging (Transvaal), Vierfontein (Orange Free State),

and (probably) Newcastle (Natal), at Goamus (South-West Africa),

Wankie (Southern Rhodesia), eastern portion of Tanganyika,* and in

the Lukuga Valley (Congo), on the western side of L. Tanganyika,

according to Renier.t In 1913 Mercenier { described very briefly

certain plants from the last-named locality, among which are appar-

ently some of this form, for example, his Srgillariostrobus (pl. vii,

fig. 1), Husigillaria (pl. vu, fig. 2), and Subsigillaria (pl. vii, fig. 3) ;

the fragments of narrow leaflets with median vein, indicated in fig. 2,

certainly resemble those described here. Future work will doubtless

show a wide distribution through the Lower Karroo Beds of the

African Continent.

So far as is known, Cyclodendron Lesli characterises strata ranging

from the Middle Ecca (Union) up to the Lowest Beaufort Beds

(Wankie), and is hence typically Lower Permian. The form is there-

fore of considerable stratigraphical value.

Genus CORNUCARPUS, Arber, 1914.

CORNUCARPUS §p.

Plate XLI, fig. 11.

The platyspermic seed, 3-5 mm. in length (No. 18), associated with

fragments of Cyclodendron shown enlarged in Plate XLI, fig. 11, has

a narrow wing that is all but absent towards the rounded base, but

widens along the sides and is prolonged at the apex into two slightly

diverging, delicate cusps or horns. One of these has largely been

* W. Janensch, “ Beit. Kennt. Karru-Schichten dst. Deut. Ost-Afrika,”’

Palaeontographica, sup. vii, p. 114, 1927.

+ A. Renier, in “ A. Jamotte,” Bull. Acad. R. Belge, 5° sér., tom. xv, pp. 655-8,

1929.

+ M. Mercenier, “‘ Le Bassin permien de la Lukuga,” Ann. Soc. géol. Belg.

Publ. rel. Congo Belge, Annexe, tom. xl, fase. iii, p. 172, 1913.

406 Annals of the South African Museum.

broken away, but the other is fairly distinct, its length being well over

twice its width. The seed therefore falls within the definition of

Arber’s genus * Cornucarpus, as re-defined by Halle.t Near the apex

the medial ridge can just faintly be made out. The surface of the seed

is covered by a fine elongate reticulation, but no details of the interior

can be discerned.

While it does not agree in its characters with any of Halle’s species

from the Permian of China, it shows appreciable resemblances to

C. (Cardiocarpus) acutus (Lindley and Hutton) t from the Coal

Measures of Scotland, but is smaller, relatively narrower, and with

a less distinct medial ridge.

The affinities of this seed are quite uncertain, but Halle § is inclined

to place it among the Pteridosperms.

Locality.—F rom clifi below Geological Survey Office, Entebbe.

* EH. A. N. Arber, Ann. Botany, vol. 28, No. 109, p. 96, 1914.

+ T. G. Halle, Palaeontologia Sinica., ser. A, vol. ii, fasc. 1, p. 201, 1927.

t A. C. Seward, Fossil Plants, vol. iii, p. 171, fig. 444, 1917.

§ T. G. Halle, loc. cit., p. 200.

EXPLANATION OF PLATE XLI.

FIG.

1. Gangamopteris cyclopteroides var. attenuata, Feist.,1; x 14 times.

2. Cyclodendron Leslii (Sew.), 1; x 1% times, showing exterior of stem.

3. 55 x »» »> mould of the above.

> > +> 35 Plasticene impression of mould.

>» a5 V 20944; stem, natural size.

y 20958; Ne J

> > ; leaflet, x 14 times.

ea Be 2 whorl of leaflets, x 14 times.

A. be 7; whorl of leaflets attached to stem, x 14 figs:

10. Z ibe 8; leaflets attached to stem, x 14 times.

11. Cornucarpus sp.; x 5 times.

6. 5 5,

Plate XLI.

Neill & Co., Lid.

A. L. du Toit del.

GONDWANA PLANTS FROM UGANDA.

( 407 )

12. The Fossil Equidae of South Africa.—By 8. H. Haucuton,

B.A., D.Sc., Hon. Curator, Palaeontological Collections.

(With 6 Text-figures.)

THE identification of isolated teeth of members of the Horse family

presents considerable difficulties. In his description of the first fossil

horse discovered in South Africa (Equus capensis) Broom relied upon

the size of the teeth and the pattern of the enamel folding as distin-

guishing characters. Gidley, however, has well pointed out that the

degree of complexity of the enamel folding is greatly affected by

both age and individual variability, and considers that size, especially

the transverse diameters of the molars and premolars, is more constant

in a species than any other character.

Of recent years, additions to the number of species of extinct South

African horses have been made by Broom, van Hoepen, and Dreyer,

whilst a number of teeth collected by the McGregor Museum, Kim-

berley, have been sent to me for study. The species so far described

can be listed as follows :—

Equus capensis Broom (Type—lower teeth).

Equus harrist Broom (Type—two lower molars).

Equus cawoodi Broom (Type—upper premolar).

Equus kuhni Broom (Type—upper molar).

Equus gigas van Hoepen (Type—series of upper teeth).

Equus platyconus van Hoepen (Type—upper premolar).

Equus simplex van Hoepen (Type—upper molar).

Equus simplicissimus van Hoepen (Type—upper molar).

Equus lylet Dreyer (Type—upper teeth).

Equus helmei Dreyer (Type—upper and lower teeth).

Equus westphali Dreyer (Type—lower molars).

Sterrohippus robustus van Hoepen (Type—upper premolar).

Kraterohippus elongatus van Hoepen (Type—upper premolar).

Kolpohippus plicatus van Hoepen (Type—lower molar series).

Equus louwi van Hoepen (Type—upper premolar).

Hipparion steytleri van Hoepen (Type—upper molar).

Eurygnathohippus cornelianus van Hoepen (Type—anterior end of

lower jaw).

408 Annals of the South African Museum.

This is a somewhat formidable list of supposed species from an

area whose present-day equine fauna—apart from the introduced

horse and donkey—consists of the mountain zebra (Equus zebra

zebra), the various varieties of the zebra (Equus quagga vars.), and

the recently exterminated quagga (Hquus quagga quagga). Any

attempt at estimating the value of these extinct forms must be

preceded by a study of the dental characters and variability of the

present-day species.

Dreyer has formulated certain conclusions on this point, and a

study of the recent skulls in the South African Museum by the present

writer supports some of these. The number of these skulls is not

great, but there are undoubted specimens of Equus quagga burchelli,

E. quagga wahlbergi, and EL. zebra zebra. It is not possible to distin-

guish the dentition of the two varieties of quagga from one another,

but, taken as a group, quagga can be readily differentiated from zebra

by three characters—viz. the shape of the two halves of the ectoloph,

and the presence or absence of folds on the anterior plate of the

anterior island and the posterior plate of the posterior island. It is

possible, too, that the shape of the parastyle and mesostyle is a further

point of differentiation, although these vary slightly in an individual

with age. Thus, with respect to the upper grinders, we can differentiate

two groups :

(a) Equus quagga group. Halves of ectoloph concave from with-

out, passing by easy curves into styles; a fold present on anterior

plate of anterior island (pli-protoloph), and one on posterior plate

of posterior island (pli-hypostyle), smaller in the molars than in the

premolars ; caballine fold usually present; mesostyle of premolars

usually grooved, parastyle sometimes grooved.

(b) Equus zebra group. Halves of ectoloph straight or slightly

convex, abruptly marked off from styles; folds of outer plates of

islands absent ; caballine fold absent; parastyles broad, becoming

sharper in posterior part of series; mesostyles sharply rounded, and

tending to overhang posterior half of ectoloph.

In addition, Dreyer relies on the inclination of the end-plate of the

median inlet to the antero-posterior length of the tooth as a character

of considerable importance, expressing the inclination as what he

calls the “‘pillar-index.”’ The difficulty, to which he points, of

determining this index accurately, and the variation to which it is

subject in teeth of one and the same species, make this character

of little use in the differentiation of closely allied teeth, whilst the

obvious differences between the shape and nature of this “‘ end-plate ”’

The Fossil Equidae of South Africa. 409

in forms such as £. harrist and EF. quagga make such an index

unnecessary.

Considerable caution must be exercised in the consideration of the

presence or absence of folds in the fossette walls and elsewhere. Gidley

has clearly shown that the effect of wear on an upper tooth is to

decrease the degree of complexity of the folding, and that there is

even great individual variation of this character within a species.

Thus the degree of wear must be taken into account in deciding the

TEext-FIG. 1.—Patterns of premolar—molar series in two skulls of Hquus quagga

wahlbergi from Zululand, to show variation in members of a single herd. A,

Skull No. 14883; B, Skull No. 14884.

value of the absence of the protoloph, hypostyle, and caballine folds

in any tooth.

There are two skulls (a male and a female) of E. quagga wahlbergr

from Zululand in the South African Museum, one of which shows

a well-developed caballine fold in each of the upper teeth, whereas the

other—save in pm.2—has the end-plate of the valley similar to that

of van Hoepen’s FE. platyconus.

Gidley’s main conclusion, after a study of individual and age

variations, is that “ size, especially the transverse diameters of the

molars and premolars, although least account has been taken of it,

is more constant in a species than any other tooth character hitherto

used.”’

As far as the characters of the lower teeth are concerned, they

410 Annals of the South African Museum.

seem to be of even less value in the determination of species. Never-

theless, four of the South African extinct species are founded upon

lower teeth only.

Equus capensis Broom.

1909. Broom, Ann. 8. Afr. Mus., vol. vu, p. 281.

1913. Broom, Bull. Amer. Mus. Nat. Hist., vol. xxxu, p. 487, fig. 1.

1928. Broom, Ann. 8. Afr. Mus., vol. xxu, p. 441, fig. 2, A.

1931. Dreyer in Dreyer and Lyle, New Fossil Mammals and Man

from South Africa, p. 36.

The first reference to the existence of a large extinct form of horse

in South Africa was made by Broom, who in 1909 described a portion

of a worn series of lower teeth embedded in calcareous sandstone

from the beach at Yzerplaats, Maitland, C.P.

To this species in later years the same author also assigned a

number of upper teeth, suggesting in 1928 that he may have been

wrong in referring these specimens to the same species as the

type. Dreyer has suggested that #. capensis is related to the

quaggas.

The type series is incomplete, both as a series and as individual

teeth, and it has only recently been figured for the first time by

Broom. In his restoration of the missing portions he has shown the

metaconid, metastylid, and entoconid as being rounded in outline,

his restoration of the metastylid in particular differing considerably

from the somewhat triangular or pointed form seen in Equus caballus.

The dominating characteristic of the type of EZ. capensis is the great

size compared with that of any living South African species of Hquus,

and the greater depth of the outer pillars. Broom has also drawn

attention to differences in the enamel folding ; but, as Gidley pointed

out in his discussion of the North American Equidae, the degree

of complexity of the enamel folding is greatly affected by both age

and individual variability.

If Gidley’s conclusion regarding the constancy of transverse

diameter in a species be accepted, then certain other lower molars

are known which must be considered as belonging to #. capensis.

From Saldanha Bay the South African Museum possesses a fourth

premolar of the left side which has a height of 92 mm., an antero-

posterior diameter of 33 mm., and a transverse diameter of 18°5 mm.

It corresponds very closely with the type of H#. capensis, and, from

its better state of preservation, may be taken as a neotype. From

the figure given it will be seen that Broom’s restoration of the

The Fossil Equidae of South Africa. 411

metastylid is incorrect, and that the shape of this pillar corresponds

to that in Equus harrist. The specimen is numbered 2821.

A worn first right lower molar from Bloembosch, Darling (S.A.M.,

Cat. No. 2718), 76 mm. high, must also be assigned to the same form.

Its antero-posterior diameter is 30-5 mm., and its transverse diameter

18:5 mm. It is a much worn tooth, and the valley between the

anterior and posterior outer pillar is impressed to such an extent

. that the enamel impinges on that bordering the valley between the

metaconid and metastylid and causes a slight fold in it.

TexT-FIc. 2.—Enamel patterns of teeth of Equus capensis Broom. A, No. 604,

Second upper left molar, Bloembosch, Darling ; B, No. 2821, Fourth lower

left premolar, Saldanha Bay, C.P. ; C, No. 2718, First right lower molar,

Bloembosch, Darling.

Since the lower teeth from Bloembosch and Saldanha Bay seem

definitely referable to E. capensis, we can reasonably infer that the

upper premolars and molars of large size from the same localities

should be assigned to the same species. Broom has already done

this, and has described and figured (Bull. Amer. Mus., vol. xxxu,

1913, p. 438) a worn upper fourth premolar from Saldanha Bay—not

from Darling—in the collection of the South African Museum (Cat.

No. 2717). The height of this tooth is 70 mm. We have no good

upper first molar, but the second upper left molar is represented

by specimen No. 654 from Bloembosch, Darling. This has an antero-

posterior diameter of 30 mm., a transverse diameter of 30 mm.,

and an antero-posterior length of the protocone of- 17 mm. The

412 Annals of the South African Museum.

enamel of the fossettes is slightly crimped, and there is no caballine

fold visible. The height of the molar is 90 mm.

The third upper left molar is represented by specimen No. 2820,

from Saldanha Bay. This is a much worn tooth, with an antero-

posterior diameter of 35 mm.,a transverse diameter of 28 mm.,

and a protocone whose length is 18 mm. It shows no caballine

fold.

On account of the decreased antero-posterior diameter, the fossettes

in the molar are shorter and relatively deeper than in the premolar,

and, owing to differences in age, the end-plate of the valley between

the protocone and metaloph is higher and more transverse in the

molar. But the teeth agree in the following essentials: transverse

diameter, shape of the walls of ectoloph, presence of very small proto-

loph and hypostyle folds, presence of prefossette and postfossette

folds, and absence of caballine fold, and may thus be considered as

members of the Hquus quagga group, in which the caballine fold has

not been developed.

Some teeth from the Skildegat, Fish Hoek, loaned me by Mr.

A. J. H. Goodwin and said by him to be associated with Middle

Stone Age implements, must also be assigned to this species. Of the

permanent teeth I have seen a right fourth lower premolar, a left

lower first or second molar, and an incisor. The antero-posterior

diameter of the premolar is 34-5 mm., and its transverse diameter

19 mm.; the corresponding measurements in the molar are 31 mm.

and 17-5 mm. A deciduous right second premolar has an antero-

posterior diameter of 38-5 mm. and a transverse diameter of 13 mm.

Associated with these are some smaller teeth whose measurements

fall within the limits of the living Equus zebra.

Equus harrist Broom.

1928. Ann. 8. Afr. Mus., vol. xxu, p. 441, fig. 2, B14, B?, B2.

1930. Van Hoepen, Pal. Navors. Nas. Mus., vol. i, p. 7.

1981. Dreyer, in Dreyer and Lyle, p. 23, pl. vi, figs. 13, 14; pl.

vii, figs. 1, 2.

The type of this species consists of two lower teeth from the gravels

of the Middle Terrace at the Bend near Barkly West, and Broom has

placed in the genus a fourth upper premolar found in association.

The species is slightly smaller than H. capensis, and agrees in

showing no folding on the anterior border of the outer posterior pillar,

although an incipient fold is seen in Broom’s figure of the second lower

The Fossil Equidae of South Africa. 413

molar. In the fourth lower premolar the outer surface of the posterior

pillar is also slightly folded. The entoconid is strongly developed.

The premolar is much worn, only 45 mm. high. Its length is 30 mm.,

and its transverse diameter 17-5 mm.

The chief feature in the upper tooth, which is very worn, is the

small size of the posterior portion of the protocone. Broom’s figure

pictures the protocone as being more rounded than it actually is;

the inner margin is straight and the antero-inner angle more marked

than in the figure. Broom refers to the fact that the hypocone is

only slightly marked off from the rest of the tooth by a shallow

posterior fold; but this is a feature of an advanced degree of wear,

the fold disappearing altogether near the roots in some species.

Length, 33 mm.; transverse diameter, 31 mm.; height as preserved,

57 mm.

Van Hoepen is desirous of putting the upper tooth into his species

Sterrohippus robustus, which is founded on a single second premolar.

The two agree in the transverse position of the end enamel of the

valley between the hypocone and protocone, and in the comparative

straightness of the wall of the posterior half of the ectoloph, but

they differ in the shape of the protocone, and in the absence of the

hypostyle fold in F. harrisi.

The teeth assigned by Dreyer to this species (from Pniel) are smaller

than the premolar described by Broom, but otherwise seem to agree

fairly closely with it.

Equus cawoodi Broom.

1928. Ann. S. Afr. Mus., vol. xxu, p. 443, fig. 3, A.

1930. Van Hoepen, loc. cit., p. 3, fig. 2 (later as #. loww2, 1930,

p- 19).

1931. Dreyer, loc. cit., p. 26, pl. vi, figs. 8-12; pl. v, fig. 2.

The type is a fourth right upper premolar from the gravels at

Winter’s Rush, Vaal River (Kimberley Museum, No. 3711). The

enamel on the anterior face of the postfossette is more crimped than

in Broom’s figure. Hypostyle, protoloph, and caballine folds present ;

postfossette and protoconule folds deep. Tooth curved. Outer wall

of parastyle broad and flat. Angle in front of mesostyle slightly

re-entrant.

Specimen No. 4276 of the Kimberley Museum, collected by Mr.

W. Fowler in the bed of the Riet River, below Koffiefontein, is a left

molar that is very close to the type in its details. Its antero-posterior

414 Annals of the South African Museum.

diameter is 33 mm., its transverse diameter 34:5 mm., and the diameter

of its protocone 14 mm. Tooth is slightly curved and is 80 mm.

high, lacking the roots.

It is doubtful whether the tooth figured by van Hoepen belongs to

this species or is not closer to H. kuhni; and none of the small teeth

figured by Dreyer seems to be appropriately placed here, although

obviously members of the quagga group, as is LH. cawoodv.

Equus kuhni Broom.

1928. Ann. 8. Afr. Mus., vol. xxu, p. 444, fig. 3, B.

1931. Dreyer, loc. cit., p. 25.

The type is a well-fossilised right upper premolar from Pniel, Vaal

River. The parastyle and mesostyle are strongly developed; the

bifurcation of the mesostyle disappears as the tooth is further ground

down, but that of the parastyle continues almost to the root. Tooth

slightly curved.

Equus gigas van Hoepen.

19502 Loch cit, p. 2, igel.

The type is a series of two deciduous molars and two permanent

molars. The walls of the ectoloph of the deciduous teeth are of the

E. zebra type, those of m.1 and m.? of the EF. quagga type. Teeth

high, curved, and large. Caballine fold present in m.1, but not in

deciduous teeth. Hypostyle and protoloph folds absent or very

inconspicuous. Parastyle and metastyle rather broad and flat, the

latter overhanging the ectoloph wall. Teeth curved. ~

The teeth figured by van Hoepen agree closely with the upper

teeth that are here assigned to H. capensis, save in the presence of

a caballine fold in m.' of gigas and in their greater size.

Equus platyconus van Hoepen.

1930. Loc. cit., p. 4, fig. 3.

Type, a right upper fourth premolar, 77 mm. high, 27 mm. long,

and 23 mm. broad. Parastyle and mesostyle rather sharply rounded ;

walls of ectoloph concave from without. Caballine fold absent.

Protoconule fold deep. Hypostyle and protoloph folds prominent.

Protocone long.

Van Hoepen considers this tooth to be sharply distinguished from

that of any living South African species ; but it approximates fairly

)

fal

it i

(

i] 1

‘a

i We

hit

‘ t

tx] :

Wath

| |

has

{ |

1 hal

anh a

ia)

" i

i Ai

ah |

Hee

iy || tt

. i

Ail (it

any i

ole?

D. q i

yee

fh |

op |

Mall cast

| i

me |

pe | i

aoa tl:

j he

1 t

‘a ig

Wei,

ao |!

Twit!

' i

mat

' }

ui 4

inl

ge ia}!

hess)

whys

Lea

C i ei!

ity

ball

4 7

‘ yl

led

Ory |

s] {

j |

\ Hi

ia Phi

1 ay

iW pel

ie) eG

ilLve

} i

}

\ {

ra SS

Se eee Ue a ee TE

. aS > SS

———

er Aue

2s =

The Fossil Equidae of South Africa. 415

closely to the teeth of an Equus quagga wahlbergi in the South African

Museum Collection, and can easily fall within the limits of variation

shown by that form, especially if it be looked upon as a first molar

rather than as a fourth premolar.

Equus simplex van Hoepen.

£320. Loc. cit, p. 5, fig. 6.

Type—a right upper first molar. Height, 89 mm.; length,

29 mm.; breadth, 30 mm. Van Hoepen considers the type to be

similar to the tooth described as from Darling by Broom, and con-

sidered here to be a first molar of Equus capensis. It differs from

A B

TExt-FIc. 3.—Hquus capensis Broom var. simplex van Hoepen. A, Third upper

left molar, Koffiefontein, O.F.S. (Kimb. Mus., No. 4062); B, Fourth lower

right premolar, same locality (Kimb. Mus., No. 4066).

that tooth, however, in the absence of a hypostyle fold and in being

somewhat smaller.

The type locality is Koffefontem, and from that locality the

present writer has examined some teeth presented to the Kimberley

Museum by Mr. W. Fowler. Among them is a third upper molar

of the left side (K.M., No. 4062), which seems definitely to belong to

the same species as the type of H. somplez. The length is 32 mm.,

the breadth 25-5 mm., and the length of the protocone 15 mm. In

the shape of the valley between hypocone and protocone, shape of

the ectoloph wall between the mesostyle and metastyle, and the details

of the fossette walls, the tooth compares closely with H. simplez,

differing only in the presence of a hypostyle fold.

Associated with this are two lower teeth—a right fourth premolar

(K.M., No. 4066), and a right first molar (K.M., No. 4047). The former

is 28-5 mm. long and 15-5 mm. broad, the latter 25-5 mm. long and

13-5 mm. broad. Both are much worn. In pattern both agree

fairly well with H. harrisz, and to a less extent with H. capensis.

416 Annals of the South African Museum.

The posterior outer pillar is distinctly notched in the middle, but

has no anterior fold. Both teeth are much worn.

From the evidence available it would seem desirable to regard

E. simplex as a variety of EB. capensis.

Equus sumplocissimus van Hoepen.

1930. Loc. cu; p. 6; ea; “ps2 ness 2h 43.

Founded on a solitary small tooth, considered to be an upper

fourth premolar, 25 mm. long and 24 mm. broad, from Koffiefontein.

Mesostyle rounded, walls of ectoloph concave, no caballine fold,

protocone short, hypostyle fold absent, protoloph fold rudimentary.

Considered by van Hoepen to be close to the tooth from Middelburg

described by Broom as E. capensis, but smaller. Insufficiently

known. An upper m.? assigned to this species by van Hoepen has

a long protocone, and the parastyle and metastyle have flattened

outer anterior faces.

Sterrohippus robustus van Hoepen.

1930. Loe. cit., p. 6, fig. 8.

1931. Dreyer, loc. cit., p. 23.

Founded on a second upper premolar 65 mm. high, 38 mm. long,

28 mm. broad. Tooth straight. Metastyle inconspicuous, mesostyle

fairly broadly rounded ; protocone short and rounded, larger than

hypocone. Fossettes long and shallow; protoloph and hypostyle

folds present. Walls of ectoloph between styles flattened.

Van Hoepen considers that the upper tooth assigned by Broom to

E. harrisi should be included here, pointing out that the occurrence

of upper and lower teeth at the same locality is no justification for

putting them into the same species. At the same time it must be

noted that no lower teeth of robustus are known, and that van Hoepen

therefore has no justification for separating the type lower teeth of

harrist from the upper tooth assigned to that species. If the type

of robustus is of the same form as the assigned specimen of harrist,

then the latter specific name must stand for this form until such time

as it is proved to differ from the type.

Further, van Hoepen’s genus Sterrohippus seems to be unnecessary.

The features of the only known tooth are not sufficiently distinct

from those of even modern South African species of Equus to be

The Fossil Equidae of South Africa. 417

reckoned as of generic value—even the short round protocone can

be paralleled in some skulls of Equus zebra, especially in the anterior

premolar.

It must be concluded that Dreyer was right in sinking this species

into Equus harris.

Kraterohippus elongatus van Hoepen.

1930. Loc. cit., p. 8, fig. 9.

1931. Dreyer, loc. cit., p. 27.

This new genus and species was founded by van Hoepen for a single

incomplete upper tooth which he deemed to be a second premolar,

30 mm. long and 24 mm. broad. The reasons given for separating

the tooth from the genus Hquus or from the E. quagga group do not

seem to be cogent.

Kolpohippus plicatus van Hoepen.

I9a02 Loc, ct.,.p. 9, fie. 10.

1931: Dreyer, loc. ci., p. 33.

Founded on a series of four lower teeth, which are larger than the

teeth of the living zebras. The chief features are the presence of a

strong fold on the anterior face of the outer hypolophid wall of the

premolars and the strong crimping of the inner wall of the hypolophid.

The latter feature seems to suggest affinities with EH. cawoodi and

E. kuhni, the lower teeth of neither of which are known for certain.

Again, the necessity for the erection of a new genus does not seem

obvious. Dreyer places the species “‘ very near to EL. quagga quagga.”’

Equus lyler Dreyer.

ial loc. ei. p..o0, pl. vi, figs. 6, 7; pl. vu, fig, 8.

Founded on upper teeth which are considered by Dreyer to be

from one skull from Floris Bad. A series of lower teeth from the

same locality is associated with the type in the description of the

species.

Dreyer has pointed out the resemblance of the molars to certain

teeth of E. quagga burchelli. To the present writer it seems that

the teeth of F. lylei definitely lie within the limits of variation shown

by either the variety burchelli or the Zululand form wahlbergr. That

these limits are wide are shown by the drawings of the molar series

VOL. XXVIII, PART 4. 29

‘| M

Shean

in)

i j

hi +

i ;

i H

va A

liuet

}

FEES

—— >.

——

enero

OTe So awe

418 Annals of the South African Museum.

of two skulls in the South African Museum, a male and a female of

the same herd shot in Zululand (Nos. 14883 and 14884) (see text-

fig. 1). Comparison of the drawings given by Dreyer of m.’ and

m.” of £. lylec with those given here shows that one or other of the

two recent sets of teeth contain all the characters listed by Dreyer

as being distinguishing features of this species, and the measurements

are very close to those given by him. The name Hquus lylev, there-

‘fore, seems to be a synonym for #. quagga wahlbergi.

Equus helmer Dreyer.

1931. Loc. evt., p. 30, pl. vi, figs. 1-5 ; pl. vu, figs. 1, 7, 8.

Founded on upper and lower grinders and on milk teeth from

Floris Bad. Enamel of upper molars fairly strongly folded ; hypo-

_style and protoloph folds present in premolars but absent in molars ;

caballine fold present; postfossette and protoconule folds deep ;

outer wall of parastyle broad and flat, except in m.?; outer wall of

mesostyle broad and flat ; protocone large, with its inner face grooved.

Recognising the variation possible within any one recent species,

there seems little to distinguish this form from EF. cawoodi. The

elongate protocone of the type of cawoodi (which may be either pm.*

or m.?) is paralleled in the m.1 of helmez, and the other features are

seen in one or other of the teeth figured by Dreyer. An even closer

parallel is seen between the pm.? of helmei and the tooth from

Koffefontein referred by the present writer to cawoodi. As Dreyer

has pointed out, forms such as these are very distinctly of the guagga

type, and are distinguishable from the living species mainly by their

greater size.

Dreyer has also pointed out the resemblance of the lower teeth

assigned by him to L. helmez to the type of van Hoepen’s Kolpohippus

plicatus, although he has no hesitation in placing them in different

species.

Equus westphali Dreyer.

1931. Loe. cit., p. 36 (pl. vii, fig. 5 2).

Founded on two lower teeth from Pniel which are considered by

Dreyer to differ from EH. capensis, but of which the measurements

given by Dreyer must be incorrect. In the possession of a slight

ridge on the anterior face of the posterior outer pillar, the premolar

agrees with the premolar from Fish Hoek assigned to EL. capensis.

The Fossil Equidae of South Africa. 419

Equus louwi van Hoepen.

1930. Loc. cit., p. 19, figs. 6-11.

Founded on an upper fourth premolar which lacks the parastyle

and metastyle. To this species van Hoepen assigns the tooth pre-

viously called H. cawoodi by him. He distinguishes the form from

EH. kuhni by a number of characters that can scarcely be considered

as of specific value, and in view of the close similarity in size and

general form between the type and that of EH. kuhni, it would seem

unnecessary to retain this new specific designation.

Hipparion steytlery van Hoepen.

1930. Loc. cit., p. 21, figs. 14-19.

Founded on an upper molar, with a third upper molar and two

lower deciduous molars as paratypes. Sharply marked off from other

known South African forms by the isolation of the protocone in the

upper teeth and the accessory isolated outer column of the lower

milk molars.

Eurygnathohippus cornelianus van Hoepen.

1930. Loc. cit., p. 23, figs. 20-22.

Founded on the front part of a lower jaw. Two large incisors only

on each side, the four almost in a straight line. Canine small, directly

behind second incisor.

Equus sandwithi sp. nov.

Some years ago the South African Museum received from Mr.

H. Sandwith, of Usakos, South-West Africa, a consignment of bones

and teeth which he had obtained in digging a well. They were dis-

covered in a bed of ‘“ hard clay and sand’ 10 feet thick, underlying

8 feet of surface limestone. The teeth consist of a number of equine

grinders of two distinct types, and fragments of molars of a species

of Archidiskodon. 7

Of the two varieties of Equus teeth, one—allied to the H. zebra

- group—shows characters that mark it off very definitely from any-

thing yet described from South Africa, and it is this which is described

here as a new species. As co-types are taken 5 upper cheek teeth

and 4 lower teeth bearing the catalogue numbers 6577 and 6578 in

the South African Museum Collection.

~~ + —

Cw=(—V—[email protected]_ ia

—

“ta an =

420 Annals of the South African Museum.

Upper Teeth._—These are all well-worn teeth, characterised definitely

by the simplicity of the enamel pattern, by the strong development

of the parastyle and metastyle, by the shape of the outer walls of

the paraloph and metaloph, and by the shortness of the protocone.

The protocone pillar has a definite median groove. The pre- and

post-fossettes are large. The prefossette has a small protoconule fold,

a trace of a prefossette fold, and no protoloph fold except in the second

premolar. The postfossette has a minute anterior fold and no hypo-

style fold. No caballine fold is present.

The teeth consist of a right pm.?, a right and a left pm.®, a left

pm.*, and an incomplete right m.?. Measurements in millimetres

are given in the following table :—

enone Pm Pm Pmes

right. right. left. left.

Length . 5 elo 30 30 29-2

Breadth ; . | 27-5 | about 29 30 33

Length of protocone 7-5 10 10 10-5

Height of crown ./| 60 56 58 60

The shortness of the protocone is thus a well-marked characteristic,

and, as will be seen from the figures, it is due to the lack of development

of the anterior prolongation. This is a character that is displayed by

Protohippus and is not seen in the true Equus, although visible in the

tooth referred by Broom to £. harrisz.

Lower Teeth.—The lower teeth associated with the upper grinders

consist of an incomplete third premolar and a first molar of the right

side, and the first and third molars of the left side. These are definitely

characterised by their general shape, and particularly by the obliquity

of the anterior face of the parastylid to the external face of the anterior

pillar.

The third premolar had a probable length of 28°55 mm. The meta-

conid and metastylid are both rounded, and are separated by a narrow

| V-shaped valley. The ectoconid is large, and the external and anterior

_ walls of the valley between it and the metastylid are strongly crimped.

i The first molar of the left side has a length of 30 mm. and a trans-

SS a

—

verse diameter of 16 mm. Here again the groove between the meta-

conid and metastylid is deep and narrow; the entoconid is smaller

than in the premolar.

=e _

E > a

1 eet, ee

=—

SSS

Lhe Fossil Equidae of South. Africa. 421

The third molar is 29°5 mm. long and 11 mm. wide. Compared with

Equus caballus, therefore, the posterior molars of both upper and lower

Jaws are small.

The form is definitely distinct from E. zebra, to which among South

African forms it is most closely allied.

Two teeth from Sheppard Island, Vaal River, belonging to Mr. C.

van Riet Lowe, can be compared closely with the type lower teeth of

this species. They consist of a third molar of the left side and what is

Text-Fic. 4.—Hquus sandwithi sp. nov. Type teeth from Usakos, S.W. Africa.

A, Third left upper premolar; B, fourth left upper premolar; C, first left

lower molar; D, third left lower molar.

probably the first molar of the right side. The resemblance to the

type is particularly seen in the acute angle which the anterior face of

the parastylid makes with the plane tangential to the outer face of the

tooth. The acuteness of this angle is rather more pronounced than in

the teeth from Usakos.

In the third molar the hypoconulid of the Sheppard Island tooth

is stronger than in the typical sandwithi, and the crown is considerably

more curved. In the anterior molar there is a close approximation

to the type.

Notohipparion namaquense gen. et sp. nov.

The lower teeth upon which this new genus is founded indicate a

stage of equine evolution not hitherto discovered in South Africa.

422 Annals of the South African Museum.

Type lower teeth from Namaqualand.

Trxt-ric. 5.—Notohipparion namaquense gen. et sp. nov.

They consist of a series from pm.?

to m.1, together with m.? of the left

side, and pm.*, m.1, m.?, and a part

of m.® of the right side, obviously

of the same individual. The teeth

came into the writer’s hands through

the agency of Dr. A. Young, who had

received them from a former student

of his, Mr. P. van der Heever. The

latter obtained the material during

the course of sinking a well on a

farm 40 miles east of Springbok in

Namaqualand, the teeth being re-

covered from a granite gravel 60 feet

down in the surface limestone.

As preserved, the teeth are all

low, the height not being much

greater than the length. The main

features can be distinguished from

the drawings given. In all the teeth

the metastylid column is separated

from the metaconid column to the

base of the crown, and the former

projects further inwards than the

metaconid or the entaconid, whilst

its posterior flange overlaps the an-

terior border of the entaconid. The

antero-external cingulum fold is pro-

minent to the grinding surface, being

fused with the parastylid in pm.?

and m.}, but still separated from it

in pm.?,m.*, and om-2). Ini pmitsgen

the right side (but not of the left),

in m.1 and m.2, there is an ectostylid

pillar, which near the root fuses with

the wall of the hypolophid to form

an ectostylid ridge. Enamel fairly

wavy. Teeth heavily cemented.

The following are the measurements of the teeth of the left side,

with the exception of m.?, which is from the right side. The height

given is exclusive of the roots, and is measured from the top of the

The Fossil Equidae of South Africa. 423

roots to the highest point of the crown. Measurements are in

millimetres.

Pn. Pm. Pm.?. . Mee, Nace Je Ie

Length . 29-5 27 26 23 24 29

Breadth 16 17 lea 16 14:5 12:5

Height . 14-5 17-5 25 23 30 o4

SYSTEMATIC SUMMARY.

Summarising the conclusions reached in the foregoing pages, it can

be seen that the various species described fall into five groups. These

are :

(1) The peculiar Eurygnathohippus, represented by one form Euryg-

nathohippus cornelianus ;

(2) the Hipparion group, into which fall Hipparion steytleri and

Notohipparion namaquense ;

(3) the harrist group, consisting of the one species, characterised by

the primitive shortness of the protocone ;

(4) the quagga group; and

(5) the zebra group.

Group (3) contains, in addition to the type, the specimen described

by van Hoepen as Sterrohippus robustus. Specifically this is synono-

mous with Broom’s Equus harrisi, but it would seem convenient to

retain van Hoepen’s generic name, giving it subgeneric rank. The

single known species would thus be known as Equus (Sterrohippus)

harris. :

A large number of the described forms fall within the limits of

Group (4), the qguagga group. For this group Dreyer has suggested the

revival of the generic name Hippotigris. Such a proceeding, however,

is scarcely possible. Hippotigris was founded by Hamilton Smith in

1841 to include five species, of which H. zebra stands first. In 1900

W. L. Sclater definitely recognised this species as the type of the group,

and was followed by Lydekker in 1916. Pocock, in 1902, restricted

Hippotigris to the quagga group and placed H. zebra in the subgenus

Asinus. In this he seems to have been guilty of the same error as

Dreyer. The difficulty of nomenclature is further increased by the

fact that Lydekker has placed both quagga and zebra in the sub-

genus Hippotigris. At the risk, therefore, of using a cumbersome

——-— ee ll

424 Annals of the South African Museum.

nomenclature, the course is adopted here of using the terms “ quagga

group’ and “ zebra group ’”’ until such time as an attempt is made to

correlate the grouping of the various African horses based on external

features with that based on the characters of the dentition.

Tabulating the ideas already expressed, we arrive at the following

results :—

Old Name.

Eurygnathohippus cornelianus

van Hoepen.

Hipparion steytlert van Hoepen.

Equus harrisi Broom.

Sterrohippus robustus

van Hoepen.

Equus capensis Broom.

Equus simplex van Hoepen.

Equus westphali Dreyer.

Equus cawoodi Broom.

Equus helmei Dreyer.

Equus kuhni Broom.

Equus louwi van Hoepen.

Equus platyconus van Hoepen.

Kraterohippus elongatus

van Hoepen.

Kolpohippus plicatus

van Hoepen.

Equus lylec Dreyer.

Equus simplicissimus

van Hoepen.

_ Equus gigas van Hoepen.

Name proposed.

Group lL.

Eurygnathohippus cornelianus

van Hoepen.

Group 2.

Hipparion steytlert van Hoepen.

Notohipparion namaquense

Haughton.

Group 3. Subgenus Sterrohippus.

Equus (S.) harrist Broom.

Group 4. guagga Group.

Equus capensis Broom.

Equus capensis Broom,

simplex van Hoepen.

Equus capensis Broom.

Equus cawoodi Broom.

Equus kuhni Broom.

Equus quagga var.

var.

Equus quagga var.

Equus quagga wahlbergi (*).

Equus gigas van Hoepen (tran-

sitional to zebra group).

Group 5. zebra Group.

Equus sandwitht Haughton.

The Fossil Equidae of South Africa. 425

GEOLOGICAL AGE.

The most ancient horses in South Africa seem to be those of the

Hipparion group. Unfortunately, nothing is known of the sequence

of gravels and limestone encountered in the well from which Noto-

hipparion namaquense was obtained, and but little light can be thrown

upon its age. The valleys of Namaqualand seem to have suffered a

progressive infilling with sand from Upper Cretaceous times onwards ;

but the process was, in all probability, not a continuous one, and

further study will probably reveal breaks in the sedimentation. In so

far as comparison is possible, Notohipparion would seem to represent

an early Pliocene stage of equine evolution.

Hipparion steytlera was found by van Hoepen in the lowest bed of

the exposure of clays and gravels on the farm Uitzoek in the eastern

part of the Orange Free State. This lowest bed was given the name

Cornelia Beds, and assigned by van Hoepen to the Lower Pliocene.

The evidence for age is based on the discovery of the type tooth. In

America, Hipparion ranges from the Upper Miocene possibly into the

early Pleistocene ; it dies out in Europe in the Pliocene, but Pomel re-

cords it from what he considers to be early Pleistocene in North Africa.

From the evidence afforded by Equus gigas, it would appear that the

zebra group of horses is of greater antiquity than the quagga group,

- and this conclusion seems to be borne out in part by the geological

evidence.

The type of Equus sandwithi was found in definite association at

Usakos with teeth of a species of Archidiskodon. The teeth referred by

me to E. sandwithi from Sheppard Island, Vaal River, are thoroughly

fossilised and slightly rolled, and come from the base of what Mr. van

Riet Lowe calls the D1 gravels. The accompanying sketch section

supplied by Mr. van Riet Lowe illustrates the approximate location of

the two teeth. The D! gravels are the tail-end of the D gravels,

which have yielded the types of Archidiskodon sheppardi and A. trans-

vaalensis, and artefacts belonging to the Upper Stellenbosch culture.

With regard to the horse teeth Mr. Lowe considers, from their rolled

condition and their location at the base of the gravels, that “ we are

entitled to assume :

(1) that they are older than the D! gravels, and therefore

(2) at least as old as the D gravels, and so

(3) from the time horizon of the mammoth.”

Mr. Lowe is inclined to regard the D gravels as Pleistocene in age—

probably Late Pleistocene.

— ee ce c=

426

Annals of the South African Museum.

, Vaal River (by C. van Riet Lowe), showing relation of C, D, and D!

TrExtr-FiG. 6.—Sketch section through Sheppard Island

h of Equus sandwithi.

The arrow indicates the approximate position of the teet

gravels.

The C gravels of Sheppard Island

have yielded two lower horse grinders,

one of which agrees closely with the

fourth premolar of the type of Equus

capensis. Associated with these is a

tooth of Bubalus sp. (possibly baint),

and implements of the Fauresmith and

Middle Stone Age culture.

The type of Equus capensis occurs in

a calcareous block washed up on the

beach north of Cape Town. Along the

shore thereabouts water-worn blocks of

sandy limestone bored by marine ani-

mals are of fairly common occurrence.

They must represent a former land de-

posit now below sea-level. The most

recent event in the history of the area

is an uplift; but evidence of previous

local sinking as a result of warping

is afforded by the older raised beach

of the western coast from Saldanha Bay

northwards, and has resulted in the

drowning of some of the river mouths.

Unfortunately, it is not possible to fix

an exact date for this episode, but Equus

capensis must have existed in the area

before it.

Specimens referred to Equus capensis

have been recorded from the Darling

area (associated with Bubalus baini, and

possibly with Wilton type of imple-

ments), from the Skildegat, Fish Hoek

(with implements of the Still Bay type),

from Pniel (completely fossilised), and

from Koffiefontein (completely fossilised

tooth—E. simplex of van Hoepen—and

teeth in the Kimberley Museum). The

tooth from Tuinplaats, Transvaal, found

in association with the Springbok Flats

man, is a first left lower molar almost

identical in all its features with the tooth

The Fossil Equidae of South Africa. 427

numbered 2718 in the South African Museum Collection from Bloem-

bosch, Darling. Dreyer does not mention whence came the teeth

which he made the types of westphali, but it is to be presumed that

they came from the diamondiferous gravels at Pniel which, according

to Goodwin, occur in the present river-bed and yield rolled artefacts

of Stellenbosch type—the basis of van Hoepen’s Pniel culture. They

may thus possibly be as old as the fossils from the D gravels of Sheppard

Island. Equus capensis, therefore, has a wide distribution both in time

and space.

The types of Equus harrist come from “ the gravels of the Middle

Terrace at The Bend near Barkly West,” whilst van Hoepen’s specimen

is from the Old Mission Stationat Pniel. The latter locality has yielded

abundant implements of the Stellenbosch (Pniel) culture. The so-

called Middle Terrace at The Bend also gives evidence of the same

culture. From the Pniel site van Hoepen has recorded a small tooth

belonging to the guagga group and a deciduous molar which has all the

features of the zebra group. From gravels in the river-bed at Barkly

West comes also the deciduous molar described by Fraas as Equus cf.

zebra, which may belong to van Hoepen’s species Equus gigas, and

from Pniel comes the type of E. kuhni. This species is also recorded

from Cornelia, but its precise level in the sequence there seems

uncertain.

The other described species, together with a number of teeth from

various places near Koffiefontein collected by Mr. W. Fowler and sub-

mitted for examination by Miss Wilman, show—as Dreyer has pointed

out—that the quagga group and the zebra group were differentiated at

the time of deposition of the later river deposits, and that they were

contemporaneous with Early Man in South Africa. The older gravels,

those containing no human artefacts, have yielded little evidence of

the existence of horses, but the work done by Dr. van Hoepen at

Uitzoek, and the discovery of Notohipparion in Namaqualand, lend

hope to the view that a far better succession of forms will be discovered

than that known to us at present.

( 429 )

13. Parevasaurian Studies.

Part VII.—On the Hind Limb of the Two Litile-known Pareiasaurian

Genera: Anthodon and Pareiasaurus. By Lizuwe OD.

Boonstra, D.Sc.

(With Text-figures 1-7.)

Since the publication of Part VI of the “ Pareiasaurian Studies,’ I

have had the good fortune to obtain portions of the hind limb of both

Anthodon and Pareiasaurus. On the farm Dunedin, in the Beaufort

West Division, the sacrum, pelvis, right femur, tibia, and fibula

(together with a partial skull and a humerus) of Anthodon were

found (S.A.M. Cat., No. 10074), and on the farm Zwaluw Krantz, in

the Murraysburg Division, the femur, tibia, and fibula (together with

scutes and teeth) of Pareiasaurus serridens were recovered (S.A.M.

Cat., No. 10032). Both these specimens have been identified on the

portions of the skulls preserved—the types of both genera being

skulls. Both are Crstecephalus—zone forms—Anthodon low down in

the zone and Parevasaurus high up.

The descriptions given here are intended as an appendix to the

systematic part of the paper on “ The Osteology and Myology of the

Locomotor Apparatus. A.—Hind Limb,” and the form in which

the descriptions are drawn up is similar to that applied in the paper

cited.

Genus PAREIASAURUS Owen.

Pareiasaurus serridens Owen.

Femur (fig. 1).—The femur is of medium size and moderately

massive. The width over the external trochanter and also over the

distal epicondyles is very small (this is partly due to a slight post-

mortem distortion).

Proximally the external trochanter is moderately well demarcated,

but distally the postaxial edge runs down evenly and straightly.

The flange for the ilio-femoralis is therefore narrow and straight.

The ridge separating the surfaces of insertion of the ilio-femoralis and

430 Annals of the South African Museum.

the pubo-ischio-femoralis internus is strong, and distally ends abruptly.

The internal trochanter is not very strong, but is prominent, sharp-

edged, straight, and directed ventrally. The stem of the primitive

Fie. 1.—Left femur of Pareiasaurus serridens. x. (10032.)

(a) Dorsal. (b) Ventral. (c) Anterior.

Y-shaped adductor ridge is fairly strong, directed somewhat medially,

but dies out before reaching the distal facets. The shaft of the femur

Fig. 2.—Left tibia of Parevasaurus serri- Fic. 3.—Left fibula of Pareiasaurus

dens. xt. (10032.) serridens. x. (10032.)

(a) Dorsal. (6) Ventral. (c) Posterior. (a) Dorsal. (6) Ventral. (c) Anterior.

}

is fairly long and moderately constricted. The dorsal curvature is

moderate, but the anterior part of the proximal articulation is directed

much preaxially and also dorsally. The tibial facets are directed

moderately distally. The intertrochanteric fossa is deep. The

insertion of the pubo-ischio-femoralis internus is strongly bipartite.

Parevasaurian Studies.—Part VIT. 431

A remarkable feature of the dorso-distal end is the presence of a deep

sulcus analogous to the intercondylar sulecus—to which it lies post-

axially. It appears to be pathological.

Trbra (fig. 2).—The tibia is short and light ; the proximal articula-

tion is fairly short and narrow ; the preaxial distal facet is slightly

larger than the postaxial one; the distal end is short, but broad ;

there is no indication of a tubercle on the postaxial border of the

shaft. The tibia figured by Seeley (Pl. XXIII), Phil. Trans.

Roy. Soc., vol. clxxx, B, 1889, probably belongs to Pareiasaurus

serridens.

Fibula (fig. 3).—The fibula is short and equal in length to the

tibia ; it is curved and only moderately strong ; it possesses a strong

muscle-ridge (peronaeus) on the dorso-postaxial border. The prox-

imal end is greatly expanded, whilst the distal expansion is only

moderate.

Genus ANTHODON Owen.

Pelvis (fig. 4).—The pelvis is small and its constituent bones light.

It is high and fairly narrow. The height tends to give an increased

appearance of narrowness. In anterior view

the inner border is U-shaped—with the sacrum

and sacral ribs in position it is suboval.

Sacrum.—Four vertebrae are attached to the

pelvis, of which the two anterior ones are

coalesced with each other. The first sacral rib

1S Massive, whereas the other three are slender,

becoming progressively more so in posterior

direction. The massive anterior sacral rib is

attached to the partially forwardly directed

internal face of the iium. The other three

abut on the posterior, wholly inwardly directed,

internal face of the ilium. The ridge dividing

these two faces is very strong.

. ee : : : é Fic. 4.—Lateral view of

Ilum.—The ilium is upstanding, with its “the left side of the pel-

blade not diverging much anteriorly. The an- ae eel x

terior edge is strongly everted as in Brachy-

pareia rogersi. The posterior iliac process is not strong, and projects

only slightly. The iliac blade is short. The outer surface is flat except

for the anterior eversion. The iliac shaft is long and moderately

slender. The anterior swelling of the iliac border is weak and narrow

—so is the posterior swelling. From the nature of the antero-dorsal

432 Annals of the South African Museum.

corner of the ilium it is evident that the ilio-tibialis and quadratus

lumborum were well developed.

Ischium.—The ischial “ shaft ’’ is short, inclines down, and diverges

posteriorly. The posterior border continues the weak and narrow

swelling initiated by the ilium. At the pubo-ischial suture the

symphysis is thick. The dorsal surface of the ischial plate is convex

in antero-posterior direction. The medial portion of the ischial plate

approaches the vertical, thereby causing the narrowness of the

pelvis. The posterior edge is slightly concave, but at the symphysis

lies in the same plane as the ischial tuberosities.

Pubis.—The anterior-lateral border of the pubis continues the weak

swelling of the ium. The distance from the ilio-pubic suture to

the not very prominent pubic tubercle is short. From the tubercle

a rounded “step” runs obliquely towards the’ symphysis. From

the “ step” the pubis continues downwards but also forwards, so that

the pubic bulwark is not rounded continuously with the upper true

pubic border. The pubic bulwark does not extend to the symphysis,

but is separated from it by a deep notch. The pubic symphysis,

therefore, has a tongue developed between the two notches. This

tongue is directed downwards and forwards. The acetabulum is

small, oval with the dorso-ventral border the smaller, shallow, and

directed outwards.

Femur (fig. 5).—The femur is light and small—the smallest Pareia-

saurian femur known. The width across the greater trochanter and

Fic. 5.—Right femur of Anthodon. xz. (10074.)

(a) Ventral. (6) Dorsal. (c) Anterior. (d) Proximal.

the distal epicondyles is small. The posterior edge of the flange for

the insertion of the ilio-femoralis is very slightly curved. The greater

trochanter is situated in the proximal third of the bone. Proximally

it is not continuous with the proximo-postaxial edge. Distally it is

Parevasaurian Studies.—Part VII. 433

continuous with the postaxial edge. The internal trochanter is not

strong, sigmoidally curved, and prominent. The “fourth trochan-

teroid ’ ridge is moderately strong, and does not continue to the pre-

axialfacet. The dorsal curvatureis weak. The proximal articulatory

surface is fairly much preaxially directed. The intertrochanteric

fossa is deep and situated much proximally. The shaft is long and

slender. There is a deep pit on the dorso-distal surface which indi-

cates a very secure origin for the femoro-tibialis ; the insertion surface

of the pubo-ischio-femoralis externus also indicates a strong muscle,

and this, in conjunction with the outwardly directed acetabulum,

seems to indicate that although the proximal articulatory surface of

the femur is preaxially directed, the femur was still far from assum-

ing an upright stance.

Tibia (fig. 6).—The tibia is extremely short and very light; the

proximal end is fairly short and broad; the postaxial and preaxial

facets are about of equal size; the distal end is long and narrow ;

weak indication of a tubercle on postaxial border of the shaft.

Fic. 6.—Right tibia of Anthodon. Fic. 7.—Right fibula of Anthodon.

x4. (10074.) x4. (10074.)

(a) Ventral. (6) Dorsal. (c) Proximal. (a) Ventral. (6) Dorsal.

(d) Distal. (c) Anterior.

Fibula (fig. 7).—The fibula is short and slender, and only slightly

longer than the tibia. It is curved and weak, with a strong muscle

ridge (for the peronaeus) on the postaxial border. The expansion

of the ends, in particular the distal one, is small.

Foot.—Nothing is as yet known of the hind feet of these two genera.

Recently Broom (2) has named a new species of Anthodon. Very

little description is given, and it has been considered advisable not to

name the material here described specifically until more is known of

Broom’s new type. In a subsequent paper the three partial skulls

of this genus in the collection of the South African Museum will be

studied taxonomically.

If one refers to the summing up (p. 339) in the detailed paper

VOL. XXVIII, PART 4. 30

434 Annals of the South African Museum.

on the hind limb, the following additional considerations are

apparent :—

That Parevasaurus and Anthodon must be included in the second

well-defined group, viz. that of Crstecephalus and Endothiodon zone

forms. '

The ilium of Anthodon agrees fairly well with that of Brachypareia

rogersi, and slightly less with that of Propappus omocratus.

The proximal articulatory surface of the femur of Parevasaurus

(and slightly less in the case of Anthodon) is situated somewhat pre-

axially and dorsally.

The tibiae and fibulae are remarkably short, particularly those of

Anthodon, so that a short epipodial is a characteristic of the latter

genus.

MEASUREMENTS.

For the method applied in the measurements, see Annals South

African Museum, vol. xxviii, pt. lil, p. 339.

Pelvis. Anthodon

(10074).

Symphysial length . ~ d28

Ischial length . * Ot

Thac length : : : ’ Peo)

Length of acetabulum oO

Width of acetabulum oe)

Total pelvic height . : . 300

Ischial height . : 4 8D

Tliac height : ; ; . 160

Ischial width . ; . 2160

Tliac width ‘ ; : . $260

Symphysial height . : mee)

Interacetabular width : : SS)

Femur. Anthodon Pareiasaurus

(10074). (10032).

Maximum length . ; 205 310

Width over external trochanter. A) 137

Width over proximal articulation . yO 86

Length of proximal articulation. 785 100

Width over distal articulation x | 380 155

Minimum width of shaft i : ~ 50 63

—_—

Parevasaurian Studies.—Part VII. 435

Tibia. Anthodon Pareiasaurus

(10074). (10032).

Maximum length . 7 lon 195

Minimum width of shaft ) 2B 37

Breadth . : Aen 83

Proximal | Length . 2) aS 120

articulation | Length of postaxial facet af ao 63

Length of ee facete M4 158 70

Distal f Breadth . 2 a8 65

articulation | Length . 280 80

Fibula. Anthodon Pareiasaurus

(10074). (10032).

Maximum length . . 145 95

{ Length . ; at wid5 76

Proximal end ae ett a Fes £90 30

Length . ; yn 40 70

ee Width 88 33

LITERATURE.

1. Boonstra, L. D.—‘‘ A Contribution to the Cranial Osteology of Pareiasaurus

serridens Owen,”’ Ann. Univ. Stellenbosch, vol. viii, sec. A, No. 5, 1930.

2. Broom, R.—‘‘ On a New Species of Anthodon (A. gregory), Amer. Mus. Novi-

tates, No. 448, 1930.

3. Haucuton, 8. H., and Boonstra, L. D.—“‘ Pareiasaurian Studies, Part VI,

The Osteology and Myology of the Locomotor Apparatus. A.—Hind Limb,”

Ann. 8. Afr. Mus., vol. xxviii, pt. iii, 1930.

4. Owrn, R.—Catalogue of the Fossil Reptilia of South Africa, 1876.

( 437 )

Part VIII.—The Osteology and Myology of the Locomotor Apparatus.

B.—Fore Limb.—By L. D. Boonstra, D.Sc., Assistant in

Palaeontology.

(With Plates XLII-XLIV and Text-figures 1-41.)

INTRODUCTORY.

A CONSIDERABLE period has elapsed since the publication of the

companion paper dealing with the Hind Limb. Collecting expedi-

tions and the rehousing of the whole exhibition of the Palaeontological

Department have taken up most of the intervening time. In the

preparation of this paper I am indebted to Dr. 8. H. Haughton and

Mr. J. Drury—the former for critical discussion and the latter for the

plaster reconstructions of the feet of Bradysaurus seeleyi and

Parevasuchus péringueyr.

HISTORICAL.

In 1856 Owen (32) described (p. 246) and figured (pl. xxxiv, fig. 4)

the distal end of a humerus of Propappus omocratus from Kast Brak

River, Fort Beaufort, in error as that of Dicynodon tigriceps.

With the material on which Owen (33), in 1876, founded the species

Parevasaurus serridens there was, according to Watson, associated

a scapula. This has, however, never been figured or described.

To Owen’s (33) type of “ Parevasaurus”’ bombidens there were

referred, without their association with the skull being established,

the right and left humerus and the right ulna. These were all

described (pp. 11, 12) and the left humerus figured (pl. xii, figs. 1-5).

These bones are Dinocephalian and not Pareiasaurian at all.

According to Haughton and Boonstra (28) the Palmiet Fontein

specimen, described by Seeley (37) as Pareiasaurus bombidens, cannot

be so determined from the imperfect material which constitutes

Owen’s type. It was proposed as the type of a new species—Brady-

saurus seeleyi. In this specimen the shoulder-girdle is imperfectly

438 Annals of the South African Museum.

preserved. Parts of the scapulo-coracoid, clavicles, and interclavicle

are described (pp. 89-90) and figured (pls. xii and xx). The imperfect

interclavicle was erroneously described as V-shaped, but the relations

of clavicles with interclavicle were correctly determined. Seeley

recognised a precoracoid and coracoid, and also the foramen passing

obliquely through the bone in advance of the articulation. The

latter is described as shallow and transversely oval.

The Tamboer specimen of Bradysaurus seeleyt included parts of the

scapulo-coracoid, which were very briefly described (p. 321) by

Seeley (88).

The type of Bradysaurus baini includes a good shoulder-girdle,

part of the left humerus, and a right fore limb, of which parts of the

manus alone are missing. Seeley (38) described (pp. 334-338) and

figured (fig. 6, p. 336, pl. xvi) a scapula,.coracoid, precoracoid (with

foramen), and a cleithrum (or, as he called it, a “‘ supra” or “ epi-

clavicle’). A suture between the scapula and coracoids was noted.

The right humerus was described (pp. 358-360) and figured (pls.

Xvi, xxii, fig. 1), but being fractured at the shaft the two ends when

joined were in error made to lie in the same plane. The right radius

and ulna were described (pp. 360-362) and figured (fig. 14, p. 361,

pl. xvu), and the presence of the olecranon noted. The imperfect

manus was described (pp. 362-363) and figured (figs. 15-17, pl. xvii), but

owing to its incompleteness nothing could be determined definitely.

In 1903 Broom (7) described (pp. 130-133) and figured (pl. xv,

figs. 1-11, pl. xvi) the fore limb of Embrithosaurus schwarz. Im-

portant are the following facts mentioned: the Pareiasaurian scapula

must stand much more upright than Seeley figured it for Bradysaurus

baina ;; the peculiar bipartite nature of the glenoid, whose two cavities

are separated by a vertical ridge; and concerning the humerus Broom

said: “‘ As a result of the complexity of the joint, the movements of

the humerus must have been limited to those in a vertical direction

and to rotatory movements. There can have been little or no antero-

posterior movement.’ Furthermore, “‘ the precoracoid almost cer-

tainly forms no part of the glenoid cavity.” Broom drew attention

to Seeley’s error in the repairing of the humerus of Bradysaurus

baint, and shows that in his specimen the two ends lie in planes

subtending an angle of 70 degrees. An entepicondylar foramen was

noted. Broom stated that there was no olecranon process to the

ulna as in Bradysaurus baim. The manus was incomplete, and

Broom’s reconstruction (pl. xvi) quite erroneous.

In 1904 Boulenger (4) criticised Broom’s reconstruction of the

Parevasaurian Studies.—Part VIII. 439

position of the pectoral girdle of Hmbrithosaurus schwarz, maintaining

that it is not placed far enough forward.

In Broom’s (9) specimen of Propappus omocratus only a fragment of

the girdle was preserved. The right humerus, though imperfect,

could be described (pp. 354-355) and figured (pl. xlv, fig. 1). With

the exception of the weathered proximal end of the right ulna, nothing

further of the fore limb was preserved.

In 1912 Broom (11) described (pp. 324-326) and figured (pl. xix,

figs. 5-11; pl. xxi, figs. 24-25) parts of the pectoral girdle and fore

limb of Brachypareia rogersi. The salient points are: Broom con-

sidered that a cartilaginous sternum articulated with the coracoid ;

the ends of the humerus make an angle of only 30 degrees with each

other; and he definitely stated that there is no ectepicondylar

foramen, this error being solely due to the fact that the bone was

not freed from matrix as well as it should have been. The pose of

the limbs was also dealt with by Broom, without, however, any

reference to the musculature.

In 1913, in describing (pp. 353-357, and 1st text-fig.) the carpus

of Bradysaurus baini, Broom (12) gave the incorrect carpal formula

3, 2, 5, and phalangeal formula 2, 3, 3, 4, 3.

In the same year Broom and Haughton (17) described (pp. 22-24)

and figured (pl. iii, figs. 1-6) the well-preserved pectoral girdle of

Pareiasuchus péringueyi, and the less perfect humeri, right radius,

and proximal ends of the ulnae. It is noted that the olecranon process

is not so well marked as in Bradysaurus baini. Other specific differ-

ences are also noted.

In his classification of the Pareiasaurian genera Watson (49) con-

sidered the presence or absence of an olecranon process as of generic

value. He also drew attention to the fact that there is a scapula

associated with the type material of Pareiasaurus serridens.

In 1917 Watson (46), in his ‘The Evolution of the Tetrapod

Shoulder Girdle and Fore Limb,” gave a comparative account of the

known pareiasaur girdles and fore limbs, and in fig. 27 arranged a

number of scapulo-coracoids in a series. As will become clear in the

sequel, not all his conclusions were correct.

In 1922 Romer (34) summarised (pp. 588-589) our knowledge of

the limbs of the Pareiasaurs, and thinks that ““some forms very

probably possessed an ectepicondylar foramen.” On pl. xxxvil

he gives a figure of the humerus of Brachypareva rogerst, drawn from

a cast, indicating the probable muscle attachments.

In 1929 Hartmann-Weinberg (26) described and figured the carpus

= se

——

— ~~ === i =

—

440 Annals of the South African Museum.

of a number of Russian and South African Pareiasaurs. In the same

year Haughton and Boonstra (28) proposed a preliminary classifica-

tion of the South African Pareiasaurs, and in a further paper Boonstra

(3) described and figured the manus of a number of South African

Parelasaurians.

In 1930 Hartmann-Weinberg (27), in ignorarice of the proposed

classification of Haughton and Boonstra (28), published a paper on

the systematics of the Russian Pareiasaurs. This paper is of interest

in so far as the elements of the fore limb are used in the classification.

MATERIAL.

The material on which this paper is based is presented here in

tabular form. This method is more economical of space than refer-

ring to individual specimens in the text. The specimens identified

specifically by means of skulls are indicated by an asterisk. With

the exception of the partial girdle of Parezasaurus serridens, lent for

study by Professor 8. J. Shand of the University of Stellenbosch,

all the specimens are in the collection of the South African Museum.

Inter-

clavicle.

Name. No. Scapulo- Clavicle. Humerus. Ulna. Radius. | Foot.

coracoid.

Bradysaurus 1204

part right

baini.

baini. and left and left and left and left left

Bradysaurus 4347* a - = left

baini.

Bradysaurus 4999* part right

baini.

Bradysaurus 5015 | part left | parts right + left left left part

baini. and left left

Bradysaurus 5127*| right and si right and | right right part

bain. left left right

. Bradysaurus 9001 | right and right

Hi baint. left

ki Bradysaurus 9089 uM left

i baini.

Bradysaurus 9104*| part right | right and + right and | right and | right left

i baini. and left left left left and left

th fi Bradysaurus 9105* ors A 2 i oe part

I’ bain. right

e Bradysaurus 9136 part left | right and | left left

{ baini. left

i f Bradysaurus 5013 part left | left part

‘J baini ?. left

if Bradysaurus 5624*| right and | right and “ right and | right and | right, parts

fl seeleyt. left left left left left right,

Pareiasaurian Studies.—Part VIII.

Name. No. Sa es Clavicle. see Humerus. Ulna Radius. | Foot.

Bradysaurus 9115*| part left right part part

seeley?. right right

Bradysaurus 9121*| parts parts parts

seeleyt. h

Bradysaurus 9137*| right and | left + left left left left

seeleyt. left |

Bradysaurus 9144*| part left left left part left i

seeleyt. left if

Bradysaurus 9165 left left 1M

seeleyt ?. |

Bradysaurus 9168*| right and | right and + right and | right and | right, left bh

seeleyt. left left left left left |

Bradysaurus 3718*| parts right et parts right} parts right| parts part

vanderbyli. and lett and left | and left | right, | left

left

Bradysaurus 8941*| right and | right + right and | right and | right right

vanderbyli. left left left

Bradysaurus 9169*| right and | right and sf parts right left left

vanderbyli. left left and left

Bradysaurus, 4348 | part right | part right | part ee right and | right

part left

Bradysaurus. 8934 | part right | parts + part left ae part

and left right

Bradysaurus. 9142 | part right | parts part part right | right and | right part

left right

Bradysaurus. 9143 right right part

right

Bradysaurus ? 4351 | parts right| right and + left left left

and left left

Bradysaurus ? 5011 | part right | parts part

Platyoropha 5002*| parts right] parts parts

broom. and left

Nochelesaurus 5590 | part left | parts part right and \

strubent. left i

Nochelesaurus 6239*| parts right] right and + left ih

alexanderv. and left left ;

Nochelesaurus 8944* ne right and | right, | right,

iesudenn left left left {

Dolichopareia 3717*| part left | parts parts left \

angusta. :

. Dolichopareia 6238*| right and | right + right and | right right [

angusta. part left lett |

Dolichopareia 6535 | parts right] right and + right parts right -

angusta. and left left and left | He

Embrithosaurus 6242 | right and | right st parts right right and | right i

schwarz. part left and left | lett i

Embrithosaurus 8034*| right and | parts right + right right and | right, th

schwarz. left and left left left

Embrithosaurus | 9148 sh; part right | left left left th

schwarzt. and left it

Embrithosaurus ? | 9116 se a das eis left ae

Embrithosaurus ? | 9128 Us; ty, - Piece right, left | lett t h

Brachypareia 4350*| parts right} parts parts right and . . es

rogers. and left left i |

ee eT |

442

Annals of the South African Museum.

Scapulo-

coracoid.

Inter-

clavicle.

Ulna.

Humerus.

Radius.

Brachypareia 5012*| parts part right right right right

rogerst.

Brachypareia 8953 | part right right, part| part right right.

rogerst. left

Brachypareia ? 9156 a6 she part part left

Pareiasuchus 2337*| right and | right and = right and parts right} right parts

peringueyt. left left left and left

Pareiasuchus 3016*| parts right] left + right and | right right part

nasicornis. and left left right

Pareiasaurus Stellen-| part left part part left

serridens. bosch*

Propappus omo- | 1058 | part right right part right

cratus.

Anthodon _ ser- |10074* right

rarvus.

Anthodon _ ser- |10026 part part right

rarvus. :

Anthodon minu- parts right

sculus. and left

Indet. | 4346 Ke a +

Indet. | 5000 oe aes os right

Indet. | 5597 | part left

Indet. | 9098 5 left right

Indet. | 9139 left and | left left

right

Indet. | 9151 right right

Indet. | 9158 56 part . part

right right

GENERAL OSTEOLOGY.

Girdle (figs. 1, 2, 3, 4, and 5).—The breast-shoulder apparatus of

the Pareiasaurs consists, as is the case in all primitive tetrapods, of

two sets of bony elements which have been shown to differ in their

method of origin—both phylogenetically and ontogenetically. They

are the scapular (sometimes erroneously referred to as the “‘primary’’)

and the clavicular (also sometimes erroneously called the “‘secondary’’)

girdles. The scapular girdle consists in the Pareiasaurs of three bony

elements—the scapula, forming the shoulder blade, and two coracoids

forming, in part, what may be called the “breast plate.” The

coracoidal plate is built up of two elements, an anterior one, with a

perforating foramen, called the pre-coracoid, and a posterior one

the coracoid. It is not necessary to enter into any discussion on

the homologies of the two coracoidal elements (10, 23, 35). Whatever

may be the difficulties in other groups, in the Pareiasaurs there is no

doubt that the two coracoidal elements are the pre-coracoid with its

Pareiasaurian Studies.—Part VIII. 443

foramen and the coracoid. These three bones of the scapular girdle

are ossa substitventua—t.e. are derived from the skeletogenous layer of

the sclerotome and are pre-formed in cartilage. Phylogenetically,

it appears that they arose secondarily and are not the first bony

elements of the breast-shoulder apparatus to develop. The clavicular

girdle is in the Pareiasaurs composed of the primitive reptilian number

Fic. 1.—Lateral view of right limb and girdle of Bradysaurus seeleyt (5624). x

of three bony elements. The clavicles and cleithra are paired, whereas

the interclavicle is an unpaired median bone. Whilst the clavicles

and the interclavicle are robustly developed, the cleithrum is a

degenerate bone which is functionally, to all appearance, redundant.

The bones of the clavicular girdle are dermal bones ossa wnvestiiventia,

i.e. are derived from the cutigenous layer of the miotome and are not

pre-formed in cartilage. Phylogenetically these bones are older than

those of the scapular girdle.

Contrary to Broom’s (11) expressed opinion, 1t does not appear

444 Annals of the South African Museum.

possible that a cartilaginous sternum could have been present

in life.

The breast-shoulder apparatus differs in a number of important

aspects from the pelvic girdle. Of these the chief are—dorsally the

scapula (analogous to the ilium) is not attached to the vertebral

column by any bony connection ; the whole connection is muscular

(serratus series), and although the scapular blade lies on the anterior

dorsal ribs there is no special method of attachment as, for instance,

the development of uncinate processes on the ribs, as is the case in

some forms (Diadectes, Eryops, etc.). Ventrally the pre-coracoid and

coracoid (analogous respectively to the pubis and ischium) enter into

no median symphysis; the two coracoidal plates practically touch

at the medial plane (firmistern), and in life, with the development

of a cartilaginous edge, the two halves must have rubbed against each

other; this “‘joint’’ is, however, strengthened by the development of

the medial interclavicle ; this brings us to another major difference

between the girdles of the posterior and anterior limb, viz. that in

the shoulder girdle dermal bones, with no analogues in the pelvic

girdle, are developed.

The two halves of the shoulder girdle are thus much more capable

of movement relative to each other and to the rump than is

the case in the pelvic girdle, and this would have been much more

so were it not for the presence of the clavicular part of the girdle

(this greater freedom is obtained in the later reptiles leading up to

the mammals by the gradual elimination of most of the clavicular

elements).

The relation of the clavicular to the scapular girdle can profitably

be discussed at this point. The median stem of the T-shaped inter-

clavicle receives 1n two not clearly differentiated grooves on its dorsal

surface the two coracoidal plates—thus strengthening the medial

ventral “‘joint ’’—and as the coracoidal plates practically abut and the

stem of the interclavicle extends some distance posteriorly to the

coracoidal plate, the presence of a supposed (Broom 11) cartilaginous

sternum is excluded. Anteriorly the cross-bar of the T-shaped

interclavicle curves dorsally, clasping the curved clavicles, and is still

closely applied to the coracoidal plate. Curving still further dorsally,

the clavicles articulate with the acromion process of the scapula.

The clavicular girdle thus, in anterior view, forms practically a

semicircle, and this arch together with the median stem of the inter-

clavicle forms a moderately strong reinforcement to the two halves

of the girdle without (on account of the looseness of the articulations)

Parevasaurian Studies.—Part VIII. 445

producing too great a rigidity with the resultant loss of freedom of

movement. The cleithrum is a rudiment.

Scapular Girdle (figs. 2, 3, 4, and 5).—The scapular girdle can roughly

be considered as two triangles abutting obliquely at their bases and

with their vertices diverging so that in anterior view a V is seen.

The broad base is formed by the coracoidal plate of great antero-

posterior length which tapers up towards the vertex as the scapula.

In lateral view the base of the coracoidal plate lies horizontally, from

which the anterior border extends upwards in a slightly convex

curve which is interrupted by a projection—the acromion (an im-

portant feature dealt with later) ; the posterior border, unlike the

anterior, is generally fairly strongly concave. This curvature of the

scapular girdle is variable amongst the members of the family.

Another feature remarkable in lateral view is the glenoid cavity

situated near the posterior border in the ventral third of the girdle ;

it is sometimes directed more laterally, in other cases more posteriorly;

it is obviously bipartite, and its nature will be discussed more fully

later. In anterior, as well as posterior, view another curvature

becomes apparent, viz. in that the internal border is concave and the

external border convex ; this is of course necessarily to fit around

the rump. In these two views it is often difficult to determine the

extent to which the coracoidal plate and the scapula le in different

planes as this relation is affected by post-mortem crushing. In

general it would appear that at the scapulo-coracoidal suture the

coracoidal plate lies in a plane directed obliquely medially at about

30°, and, higher up at the thinner part of the scapular shaft the dorsal

part of the scapula is again bent in a broad curve over the rump.

Another curvature that is of interest is that at the thinnest part of the

scapular shaft the dorsal part of the scapular, 7.e. the blade, is

rotated more or less so that the anterior border is directed somewhat

inwardly and the posterior border somewhat outwardly.

The scapular girdle of the different forms of Pareiasaurs is of the

same general plan, but differences can be easily enumerated, viz.

they vary in size, in posterior and internal curvature, in the position

of the acromion, in the nature of the glenoid cavity, in the size of the

coracoidal plate, and the part played therein by its two elements.

All these differences, to be treated of later, are, however, all of degree.

There is only one foramen penetrating the scapular girdle, and this

is the precoracoidal (supra-coracoid) foramen. In outer view it

appears to lie mostly in the precoracoid, but its dorsal border is

formed by the scapula. This external opening lies just anterior and

&§

| 4

446 Annals of the South African Museum. :

CL. |

Wi ay, j j a) |

GL.

/j YY SU

i, SSE Hh

! A

! \

| \ \ if |

\\WVA

ae \ : Why Yi) ey

arcana

Ac. =acromion process. Cor. =coracoid. [ Cla, Pr Cor

Cla. =clavicle. Cl. =cleithrum.

Gl. =glenoid cavity. I. Cla. =inter- Fic. 3.—Internal or medial view of the

Pr. Cor. = precoracoid. clavicle. left half of the girdle of Bradysaurus

Pr. Cor. For.=precoracoid or supra-coracoid seeleyt (9137). xy.

foramen.

Se. =scapula. Sub. Sc. Fos. =subscapular

fossa.

Fic. 2.—Lateral view of left half of the pec-

toral girdle of Bradysaurus seeleyi (9137).

x zy:

Prlovfor

—

—.

TH wis)

| SS Ac.

hil

ii [ Cla.

Lier} 1]

fl Hi ita

i eG

aa Z

ty ;

I | |

j a Fia. 4.—Ventral view of the left half of the girdle of Bradysaurus seeleyi (9137). xy.

Wah

‘3

Pareiasaurian Studies.—Part VIII. 447

ventral to the notch in the glenoidal border, and is large and of oval

shape. It penetrates the girdle obliquely to open high up on the

internal surface of the scapula in a groove (subscapular fossa) which

is anteriorly limited by a dorso-ventrally directed broad, but low,

ridge (supra-glenoid buttress). On the rudimentary triangular fossa

postero-dorsal of the glenoid cavity there is a pit which may be the

rudimentary remains of the primitive supra-glenoidal foramen.

The three elements of the scapular girdle are very closely and

Fic. 5.—Anterior view of the pectoral girdle of Bradysaurus

seeleyt (9137). x +4.

strongly fused, and it is not always possible (particularly in old

individuals) to determine the sutures with exactitude. The scapulo-

coracoidal suture is more often clearly visible than the coraco-

precoracoidal. On the external surface the former can be seen com-

mencing anteriorly at a slight notch half-way between the medial

coracoidal border and the acromion process; from there it runs

slightly obliquely downwards to the upper border of the precoracoidal

foramen ; then obliquely upwards to the notch on the antero-ventral

border of the glenoid cavity ; through the glenoid it passes along the

deepest part, 7.¢e. running obliquely upwards along the line of demarca-

tion of the two surfaces of the glenoid articulation ; emerging from

the glenoid it passes round the posterior border through a slight

notch to the internal face, where it extends to the anterior border

along a downwardly sagging line marked by a low ridge which limits

448 Annals of the South African Museum.

the two surfaces of the scapula and the coracoidal plate, which lie in

slightly different planes. Externally the coraco-precoracoidal suture

extends downwards from the notch on the antero-ventral border of

the glenoid, where it meets the scapulo-coracoidal suture ; it stands

practically vertically on the medial plane, and lies along a slight

ridge which indicates where the suture should lie if, as is the case

in old individuals, the suture itself is not visible; internally the

suture curves medially from a point nearly a third of the distance of

the scapulo-coracoidal suture from the posterior border ; this indicates

that the suture lies obliquely.

This is in general the nature of the three constituent bones of the

scapular girdle, but in the more detailed individual treatment of these

bones variations in their relations amongst the members of the family

will become evident.

Scapula.—In proportion to the body the scapula is a tall bone,

with its upper extremity expanded. It does not appear likely that

any separate supra-scapular cartilage was developed. In some forms,

e.g. Bradysaurus baini (9001), the blade is also much thickened. At

the shaft, just above the acromion, the bone is generally much con-

stricted, and it 1s at this point that two types of curvature commence,

viz. a posteriorly and a medially directed curvature. Here also the

blade is rotated on the shaft, viz. in such a manner that the anterior

border is directed somewhat medially and the posterior externally.

From the dorsal edge of the acromion the cleithrum is loosely applied

to the anterior edge of the scapula. Proof of the not intimate relation

of cleithrum and scapula is given by the fact that the cleithrum is

frequently disarticulated, and not recovered in the field. The most

conspicuous element in the structure of the scapula is, however, the

strong everted acromion process, which is more typically a mam-

malian than a reptilian structure. The strength and the amount of

eversion is variable in the family, but is always a prominent feature

of great diagnostic value. Ventral and internal to the acromion, the

scapular surface stretches somewhat upwards and inwards of the

acromion—this must represent an incipient supra-spinatus fossa as

is found typically in mammals. The acromion is not continuous

ventrally with the anterior border of the scapula, as is the case, for

instance, in Dicynodonts. Neither is it continuous with the dorsal

anterior scapular border as is typical of mammals. On the posterior

border a triangular rudimentary surface representing the supra-

glenoid fossa is developed above the glenoid, and contains what

appears to be a rudiment of the supra-glenoid foramen. The scapula

;

Pareiasaurian Studies.—Part VIII. 449

forms, in one specimen, half, but in all others the greater part, of

the glenoid ; and not only is the scapular portion a part of the total

glenoid cavity, but one of two distinct surfaces of the whole cavity.

The scapular surface of the glenoid articulatory surface varies in the

direction it faces, viz. it may combine any of the following three

directions—outward, backward, and downward. At the scapulo-

coracoidal suture the glenoid is deepest, and antero-ventrally its

border is deeply notched. Internally the scapula has a plane surface

with only the dorso-ventral groove (subscapular fossa) with the pre-

coracoidal foramen opening into it, and with the buttress that

demarcates the groove anteriorly.

Coracoidal Plate-——As the coracoid and precoracoid functionally

represent one plate, it is unnecessary to treat them separately. The

Pareiasaur coracoidal plate is long and high. The borders apparently

had a thin cartilaginous edge, but no real prolongations of any kind

(epicoracoids). The precoracoid is wholly excluded from the glenoid.

The coracoid forms the ventral part of the bipartite glenoid. This

surface varies In its direction, viz. it may be any combination of the

three—upwards, outwards, and backwards. To hold this surface the

coracoid is raised into a ridge at the ventro-posterior border of the

glenoid, just as is the case in the scapular surface of the glenoid. Just

ventro-posterior to this buttress there is a well-developed surface for

the origin of the coracoidal head of the triceps. The length and height

of the coracoidal plate varies within the family, and the proportional

size of the precoracoid to the whole plate varies from large to small.

Clavicular Girdle (figs. 16, 23,24, 25, and 37).—In anterior view the

two clavicles and the interclavicle form a quadrant of a circle. The

median interclavicle forms the ventral part of the quadrant, and

receives in its deeply grooved anterior surface the two clavicles.

The hollowing-out of the anterior surface of the cross-piece of the

T-shaped interclavicle is very deep—so much so that the edges clasp

the medial ends of the clavicles both ventrally and dorsally. The

stem of the T-shaped bone lies in the median plane. It is long, and

extends beyond the ends of the coracoids. The medial edges of the

two coracoidal plates fit in to the dorsal surface of the stem, and

practically touch in the medial plane. The interclavicular stem thus

acts as a bar of bone reinforcing the medial coracoidal * joint.”

Anteriorly the coracoidal plate fits tightly on to the cross-bar of the

T, but the edge anteriorly partly extends over it. The ventral

surface of the stem carries surfaces for muscular attachment (pec-

toralis), and the slightly bulbous end of the stem is rugose (linea alba).

VOL. XXVIII, PART 4. 31

i aaeall

SS AS Se

EE = pL mete

EEE ees

450 Annals of the South African Museum.

The clavicle is a more or less flat, longitudinally curved strip of bone,

extending from the median plane to a little distance dorsal of the

acromion process on the scapula.. Medially it is somewhat pointed,

and it also has this appearance dorsally, but this is due to the internal

border being twisted edgeways to end in a slightly bulbous end

(trapezius). The upper third is excavated for a muscular attachment

(deltoid). This surface is limited internally by the rounded edge that

has been turned edgeways. Externally this surface broadens out

into a flange. On the medial third of the bone there is also a similar

flange-like expansion, and here, and on the edge of the interclavicle

to which the clavicle is here applied, some of the ventral axial muscles

were attached (hyoids). Medially on the posterior surface the clavicle

carries a strong keel, which fits firmly in the groove of the cross-piece

of the interclavicle. For its medial third the clavicle les external to

the anterior part of the coracoidal plate, and, curving upwards internal

to the everted acromion, it lies applied to the internal scapular surface

and projects dorsally of the acromion.

The cleithrum, thickest at the acromion, extends from there in

some cases to the dorsal edge of the scapula; it is generally loosely

applied to the antero-external surface of the scapula, but in some

cases it appears to be partially fused ; functionally it is practically a

rudiment, although the whole attachment of the deltoid and trapezius

is probably not completely transferred to the scapula.

The Inmb (fig. 1, Pls. XLII, XLITI, and XLIV).—Considered 7n toto,

the Pareiasaur fore limb is short, massive, and clumsy. This even

applies to the smaller, lighter forms such as Parevasuchus and An-

thodon. The total height of the fore limb is less than that of the hind

limb, notwithstanding the fact that the radius, which accounts for most

of the height, is longer than the tibia. This is due to the horizontal

position of the humerus. If the pelvis hangs between the hind legs,

the breast-shoulder apparatus does this to a still greater extent. The

epipodial of the fore limb stands more vertical than that of the

hind limb—the mechanical reason for this is easily found in con-

sidering the difference of function, viz. the fore limb acts as a

supporting structure, whereas the hind limb is the propelling agent.

As in the hind limb, the fore limb 1s composed of five segments—

propodial, epipodial, basipodium, metapodium, and acropodium.

The first segment—the propodial—consists of one bone—the humerus.

This is a massive, complex element, with widely expanded ends. The

proximal articulation is a curious, apparently bipartite, strip, which

fits none too accurately into the equally complex glenoid cavity.

Parevasaurian Studies.—Part VIII. 451

This joint will be discussed fully in the sequel. It is sufficient to

state here that it is of such a nature that the humerus practically

always remains in a horizontal position. Ventro-distally the articu-

lating surface receives the combined proximal articulation of the

epipodial elements to form a modified trochlear joint. The epipodials

—consisting of the radius and ulna—are apparently firmly applied to

each other proximally so as to form one composite articulatory

surface. This is achieved by the proximal radial end fitting tightly

into the sigmoid notch of the ulna. The amount of movement at

this joint must have been very small indeed, being further limited by

the development of a strong olecranon process in some forms, with

only a shallow trochlear fossa on the humerus. The only movement

apparently was a slight flexion and extension. The spatium interos-

seum between ulna and radius is narrow. Distally the elements of

the epipodial are also held together, viz. by the development of

an element of the basipodium between the ulna and radius.

The basipodium consists of three subsections—a proximal, central,

and distal series of carpal bones. The proximal series consists of

three elements—ulnare, intermedium, and radiale. There is one

centrale, and only four distal carpals are present.

The metapodium consists of five short, squat metacarpals—the

first of complex shape.

The acropodial formula is 2, 3, 3, 3, 2 in all well-known forms.

The movement in these three distal segments appears to be chiefly

located at the distal carpals, which are round, pebble-like bones.

The whole foot must have stood on a slant, with well-developed pads

under the basipodium.

Humerus (figs. 1, 7, 9, 11, 14, 19, 27, 31, 34, and 39).—The Pareia-

saurian humerus is a short, massive bone, with greatly expanded ends

and a very short shaft. It can be roughly divided into six surfaces,

viz. proximal and distal articulatory surfaces, dorsal and ventral

proximal surfaces, and dorsal and ventral distal surfaces. The

proximal and distal expansions make an angle with each other,

because of a relative twisting round the shaft. The amount of this

twisting varies within the family.

In the large Tapinocephalus Zone forms, the angle subtending the

two expansions is generally about 35° ; only in one specimen of

Nochelesaurus strubent (5590) and one of Embrithosaurus schwarzt

(8034) is the angle much greater, viz. 70°-80°. Now, in the former

the humerus has undoubtedly suffered from post-mortem distortion,

whereas in the latter the bone appears normal, but in this case a

~~. _ £2 ks ss | DA

452 Annals of the South African Museum.

discrepancy arises in the fact that in another specimen (9148) referred

to this species the angle is only about 35°. The condition in these

two specimens must therefore be cited with caution as examples of a

primitive greater amount of “ twisting.” In the smaller Endothiodon-

Cistecephalus Zone forms the angle is in general much less, 15°-30°.

Thus in general terms it can be established that there is an untwisting

of the humerus from the older to the younger Pareiasaurs; but

unfortunately the humerus of Pareiasaurus serridens is practically

unknown. The dorso-proximal surface is roughly halved by the

anterior dorso-ventral line (ADVL) (fig. 7). Anterior to this line

there les the triangular upper surface of the delto-pectoral crest

(DPCR) ; posteriorly lies a triangular surface generally well marked

by muscle-scars, and with a transverse line, and limited posteriorly

by the lateral-medial line (LML). Distally this bipartite surface

extends on to the short shaft.

The dorso-distal surface is also roughly triangular, and is distally

greatly expanded by the development of epicondylar flanges. Distally

the surface is terminated by the curved border of the articulatory

surface ; proximally to this there lies a shallow triangular hollow,

forming in part a trochlear fossa for the ulnar olecranon process (when

present). The anterior border of this trochlear hollow is quite sharply

demarcated by a ridge continuous with the proximal lateral median

line (LML); anterior to this ridge lies the confluent ectepicondylar

and supinator flange with its rounded ectepicondylar foramen ;

distally this flange thickens and terminates abruptly—the thickened

truncated end forming the surface for the extensor muscles. This

projects distally of the articulatory surface. Similarly, but less

clearly limited from the trochlear fossa, the entepicondylar flange

hes posteriorly, housing the slit-like entepicondylar foramen, and

distally of the foramen it is thickened to form the flexor flange ; this

also extends distally of the articulatory surface.

Ventro-proximally a hollowed-out triangular surface lies between

the delto-pectoral crest, the proximal border, and the sharp posterior

edge. The delto-pectoral crest extends distally from the antero-

proximal corner of the articulatory surface as a thin edge which

thickens to a knob-like surface for muscular attachment (pectoralis).

The ventro-distal surface houses chiefly the bulbous swelling

representing the radial condyle or capitellum. This lies more in the

preaxial half of the surface. Postaxially it is not well differentiated

from the ulnar condyle which passes as a trochlear surface round the

distal end slightly on to the dorsal surface. In general, it seems to be

Pareiasaurian Studies.—Part VIII. 453

true that the distal humeral facets lie more distally in the geologically

younger than in the older forms.

The lateral portions of the ventro-distal surface form the under-

surfaces of the ent- and epi-condylar flanges. These flanges are

fairly weak and are not bent downwards; their distal ends bear muscle

scars for the flexor and extensor muscles respectively. In this view

the entepicondylar foramen appears different in the two big groups

of Pareiasaurs, e.g. in Bradysaurus one can see through the foramen

in ventral view, whereas in Anthodon one has to look from an antero-

ventral direction.

Ulna (figs. 1, 12, 20, 29, 32, and 35, Pls. XLII, XLII, and XLIV).—

- The ulna is a long bone—larger than the radius by the whole length of

the articulatory surface and the olecranon where this is developed.

Its proximal third contains on the proximal surface the articulatory

surface, which is partly bipartite, being indistinctly divided by a

longitudinally oblique ridge—this surface is limited distally by the

two strong processes that limit the sigmoid notch dorsally and ven-

trally (fig. 12); posteriorly it is limited by a short ridge, or in the

forms without an olecranon passes over on to the proximal end. The

fact that the proximal surface is situated preaxially and not at all

terminally is a point worthy of notice. Some forms have an olecranon

process which is more or less sharply separated from the articulatory

surface. The postaxial border opposite the articulation is thickened

into a more or less broad curve which has a rugose surface and served

for muscular attachment, e.g. triceps. Between this thickened

border and the sigmoid processes the bone is scooped out on both its

ventral and dorsal surfaces. The shaft is comparatively thin but

expands again at the distal end. Here, on its preaxial border, a

process is developed which fits into the postaxial border of the radius.

The distal articulatory surface is flattened dorso-ventrally with sharp

edges, and in life was certainly covered by a thick cartilaginous pad.

With regard to the olecranon it may be said that, of the larger

Tapinocephalus Zone forms, it is very strong in Bradysaurus, with

one exception, and absent in the other genera of this zone, whereas

in the higher forms as far as they are known it is always moderately

developed.

Radius (figs. 1, 21, 28, 33, Pls. XLII, XLII, and XLIV).—The

radius is a straight, rounded rod, with expanded ends. Both articula-

tory surfaces are concave—the proximal are more so—and both con-

tained much cartilage. The shaft is long and comparatively slender.

On the ventral surface there is a distal and a proximal hollow for

454 Annals of the South African Museum.

muscular attachment. Postaxially the radius proximally fits securely

in the sigmoid notch and distally receives the preaxial process of

the ulna (Pl. XLII, fig. 1). This, in addition to the presence of the

intermedium, must cause the two bones to move as one, with

little power of rotation about each other.

The Carpus and Manus (Pls. XLIJ-XLIV).—Although a number of

fore feet have been collected during recent expeditions, hardly any

new considerations need be added to my first account. Since this

publication two papers have appeared by Hartmann- Weinberg (26, 27).

In the first paper two different types of carpus are described from the

i | Russian material, viz. one with all the elements fused into a single

, mass called the “carpus” (sectioning this mass apparently reveals |

rit the constituent elements), and, secondly, one consisting of separate

aah elements which are maintained to be eleven in number, but in the

enumeration only nine are mentioned. In her second paper the two

types of carpus are again enumerated, but here only seven elements

compose the carpus. A pisiform is maintained to be present. Atten-

tion is called to the apparently bipartite nature of the first metacarpal.

I have not been able to determine this, but do not consider it wholly

improbable ; for the rest, my previous description (3) holds good as

far as the known South African forms are concerned.

The Shoulder Joint.—This is a joint not very easy to understand,

owing chiefly to the fact that the articulatory facets are not clearly

| i | demarcated because of the amount of cartilage which was undoubtedly

\ developed. In the Rhachitomous Amphibia (e.g. Evryops), the

primitive Cotylosaurs and the Pelycosaurs, the shoulder joint presents

an intricate position which is now generally referred to as “ screw-

shaped.” The Pareiasaurian shoulder joint has been derived from

such an elongated “‘ screw-shaped”’ joint by a shortening of the

glenoid cavity and an alteration of the direction in which the cavity

faces. Within the family two stages in this change can be roughly

determined. In the higher Pareiasaurs the glenoid is shorter,

situated further posteriorly, and directed more backwards than in

the lower forms, but the ground-plan is evident in both.

The nature of the glenoid cavity can profitably be re-stated here—

the coracoid carries a face directed upwards, outwards, and more or

less backwards, especially at the posterior end; the scapula carries

a face directed downwards, outwards, and more or less backwards.

In life the cavity between these faces must have been filled to a

great extent with cartilage. The humerus carries a “strap or

boomerang-shaped ”’ articulatory facet. This surface is roughly

—

eta. > 4

Pareiasaurian Studtes.—Part VIII. 455

bipartite—the anterior portion is directed upwards, inwards, and

forwards with respect to the axis of the body, i.e. it is a mirror image

of the scapular part of the glenoid with which it articulates; the

posterior part is directed downwards, inwards, and forwards, i.e.

a mirror image of the coracoidal part with which it articulates. The

amount and nature of the movement of which the humerus is capable

is now Clear, viz. at the most anteriorly directed position the humerus

articulates with the scapular part of the glenoid, the longitudinal

axis of the humerus is directed forwards and upwards (i.e. the

distal end is slightly raised), and the proximal end of the bone lies in

a horizontal plane ; as this is turned to its most posteriorly directed

position it articulates with the coracoidal part of the glenoid, the

longitudinal axis is directed backwards and downwards (7.e. the

distal end is lowered), and the proximal end lies in a slightly

oblique plane, viz. in that the anterior end is raised somewhat. The

movement of the humerus is therefore chiefly in a horizontal plane

with only a slight movement in a vertical plane and a slight rotation.

Fundamentally the above applies to both stages of Pareiasaurs,

but from the above-described nature of the joint it becomes apparent

that the shorter the glenoid becomes, as it does in the later forms

(except Pareiasaurus serridens) the more the amount of movement

in the vertical plane becomes and the more posteriorly directed

glenoid causes the humerus to lie less at right angles and more

parallel to the animal’s sagittal plane.

The Elbow Joint.—The distal portion of the humerus carries two

articular facets which are not always very clearly distinct from each

other, especially so in the later forms. These are the radial and ulnar

facets. The radial facet (capitellum) is a large hemispherical swelling

situated ventrally on the anterior or preaxial portion of the distal

end. In some specimens its posterior or postaxial extension 1s

indicated by a ridge, and distally it terminates in a curved line before

reaching the distal end of the bone. In the higher forms, particularly

in the one humerus of Anthodon, the hemispherical shape is lost and

the condyle is directed more distally than ventrally. The ulnar

facet lies postaxially and distally of the radial. Ventrally it tapers

to zero in proximal direction, lying as a groove between the ridge of

the radial facet and a ridge on the ectepicondylar flange. This

trochlear ulnar groove widens distally and has most of its facet

directed distally. The facet does not, however, curve round on to

the dorsal surface. There it ends abruptly and is distinct from the

shallow dorsally situated trochlear fossa.

ff? SS ee

——

_— =

_ “a>

mm”

All |

}

1 |

aii

| |

aif 1

Sah

Salle |

NRL |

pei

AMAT

} i}

vO}

} !

" |

u tT

alt

be) ug

Pea pay t |

Pia t

J ia

ii |

' 4

ibuitt,!

ope!

Dit

| th}!

on) Hi}

bei

: Sean |

| i

~*~

: oi

7 1

wei

1 /

; Wi

ah)

| |

aed Mi

Peay

it!

Phey

| Hil

' 7 }

Me I Wit

aah |

A. i

}

1 ae

{

pa!

}

-.-_ —

= ~ or

> SO

456 Annals of the South African Museum.

The epipodial presents two facets—one simple and the other

composite. The radial surface is a simple cup-shaped hollow and

fits over the hemispherical facet (capitellum) on the humerus.

The articulatory facet on the ulna lies preaxially. Proximally

it is abruptly terminated by a ridge in the forms with an olecranon

process or merges with the proximal surface in the former without

an olecranon. Distally the surface is terminated by a ridge lying

before the sigmoid notch and process. From the sigmoid process a

ridge runs obliquely across the facet to the proximo-dorsal corner.

This divides the facet into two surfaces with faces whose planes

subtend an angle of about 30°.

As the radius fits into the sigmoid notch of the ulna the epipodial

moves more or less as one mass in relation to the humerus. The

extent of the movement at the elbow joint is as follows: At the moment

of extreme flexion the cup-shaped surface of the radius fits the hemi-

spherical condyle on the humerus, and the ventral half of the ulnar

surface is applied to the tapering ventral portion of the trochlear

ulnar condyle ; in this position the epipodial is directed inwards and

forwards, making an angle of about 70° with the longitudinal axis of

the humerus ; as the lower leg is extended the radius is removed from

its good articulation and is less firmly applied to the more distally

directed portion of the radial condyle; the dorsal half of the ulnar

surface now articulates with the distal wider ulnar trochlear condyle

of the humerus, with the result, of course, that the direction of the

lower leg is changed, viz. from forwards and inwards to backwards

and outwards. The amount of extension possible is not great because

the ulnar trochlear surface ends before reaching the dorsal surface

of the humerus, and also because the radius would be wholly dislocated.

In the older forms this would be about 80°—-90° ; in the younger forms

(e.g. Anthodon) the extension is greater because of the more terminally

situated condyles on the humerus.

The Carpal Joint.—Although the articulation between the radius

and the radiale is a restricted ball-and-socket joint, and there probably

was a certain amount of rotation of the manus on this joint, the whole

wrist joint, owing to the interlocking nature of the intermedium, was

fairly rigid. The centrale also precluded any intercarpal joint,

analogous to the intertarsal joint of the hind limb. Between the

distal carpals and the metacarpals movement was fairly free; but

between the phalanges not much flexion was possible. The whole

manus was a fairly rigid mass, standing in an oblique semi-digitigrade

position.

Pareiasaurian Studies.—Part VITI. 457

SYSTEMATIC.

Genus BRADYSAURUS Watson.

Girdle.—The whole breast-shoulder-apparatus is large, strong, and

massive. The scapula is tall or fairly tall, with a wide blade, inclined

moderately backwards, and with the acromion situated fairly low

down or fairly high up and moderately everted. The coracoid plate

is long or fairly long, high or fairly high; the precoracoid is large,

and is greater or much greater than the coracoid. The glenoid is

fairly deep, the scapular part is very long and is much greater than

the coracoidal part. The direction of the two parts of the glenoid

differs in the three species ; the scapular part is directed downwards,

slightly backwards and outwards, or only slightly outwards, the

coracoidal part upwards or slightly upwards, outwards or slightly

outwards. The cleithrum is a fairly strong bone and partly fuses

with the scapula; it lies either on the anterior border or well away

from the border on the external surface. In the former case it is a

fairly straight bone, in the latter it is strongly curved. In some cases

it reaches right up to the top of the scapula.

Humerus.—The humerus is large, long, or very long and massive,

with a medium or great proximal, and a great or very great distal

expansion, standing at 35°-50° to each other; the delto-pectoral

crest is long or very long, with the proximal corner (for the supra-

coracoideus) on the anterior dorso-ventral line moderately prominent ;

the proximal articulatory surface is a partially widened ribbon

imperfectly separated into two surfaces, which respectively face

slightly downwards and slightly upwards; the trochlear fossa is

weak or fairly deep, and the bulbous distal articulatory surface 1s

fairly strongly or strongly swollen.

Ulna and Radius.—The radius and ulna are large, massive, long

or moderately long bones; the ulna has a very strongly marked

olecranon process, which may in one species, however, be absent ;

the bulging of the postaxial proximal half of the ulnar border is

strong or moderate ; the border of the proximal articulatory surface

is limited by a clear-cut ridge.

Foot.—The fore-foot is wholly or partially known in at least eleven

specimens. In structure it is typically Pareiasaurian, with the

carpal formula 3, 1, 4, and the phalangeal formula 2, 3, 3; 3, 2.

In this genus the foot attains the greatest size known in this

family.

—

458 Annals of the South African Museum.

Bradysaurus baini (Seeley).

Girdle (fig. 6).—The scapula is tall, with the acromion situated

fairly high up. The coracoidal plate is long and high, the pre-

coracoid is much greater than the coracoid. The scapular part of

the glenoid is directed downwards, slightly backwards, and slightly

outwards ; the coracoidal part upwards and slightly outwards. The

Fic. 6.—Left scapulo-coracoid of Bradysaurus baini (from cast of type in British

Museum). x ,.

(a) Lateral. (6) Internal. (c) Posterior.

The lateral and internal views of the scapulo-coracoids are taken at right

angles to the coracoidal plate, the posterior parallel.

cleithrum is fairly strong and lies wholly on the external surface ; it

is curved and lies some distance from the anterior border; the

distal end tends to fuse with the scapula; for the rest it is only

loosely applied to the scapula.

Humerus( fig. 7).—The humerus is long, with a great proximal

and a very great distal expansion standing at 35° to each other;

the delto-pectoral crest is very long; the trochlear fossa is fairly

deep, and the bulbous swelling of the distal articulatory surface is

strong.

Pareiasaurian Studies.—Part VII De 459

Ulna and Radius (PI. XLII, fig. 1).—These two bones are large,

long, and massive; the ulna has a strong olecranon process; the

bulging of the postaxial proximal half of the ulnar border is strong.

Fic. 7.—Left humerus of Bradysaurus baini (51

(a) Dorsal. (6) Ventral. (c) Anterior. (dZ) Proximal.

Din Bes,

The dorsal and ventral views of the humeri are taken at right angles to

the plane of the dorso-distal surface, the anterior parallel to this plane and

the proximal at right angles to the proximal articulatory surface.

ADVL=anterior dorso-ventral line. DPCR =delto-pectoral crest.

Ect. =ectepicondylar foramen. Ent. =entepicondylar foramen.

LML = lateral median line.

Foot (fig. 1, Pl. XLI).—The foot is wholly or partially known in

Six specimens, and in 9104 the greatest size in foot is reached.

Bradysaurus seeleyi (Haughton and Boonstra).

Girdle (figs. 1, 2, 3, 4, 5, and 8).—The scapula is fairly tall, with the

acromion situated fairly high up. The coracoidal plate is fairly long,

the precoracoid is greater. than the coracoid. The scapular part of

the glenoid is directed downwards, slightly backwards and outwards ;

the coracoidal part upwards and outwards.

The cleithrum is a long, straight, rod-like bone lying on the anterior

border of the scapula, to which it is partially fused along its whole

length ; it reaches to the top of the scapula.

Humerus (figs. 1, 9).—The humerus is very long, with great proximal

-+43:-8 -B Ass

460 Annals of the South African Museum.

Fic. 8.—Left scapulo-coracoid of Bradysaurus seeleyi (9137). x 445.

(a) Lateral. (6) Internal. (c) Posterior.

Fic. 9.—Left humerus of Bradysaurus seeleyi (5624). x4.

(a) Dorsal. (b) Ventral. (c) Anterior. (d) Proximal.

Pareiasaurian Studies.—Part VIII.

46]

hy | Zn ‘ /

lj Za ii\s\ eens

Ura

] Wp Ht at

wile

ates /

=> —_— } — pak ee ta

Fic. 10.—Right scapulo-coracoid of Bradysaurus vanderbyli (3718). x +5.

(a) Lateral. (6) Internal. (c) Posterior.

Fic. 11.—Left humerus of Bradysaurus vanderbyl (8941). x &-

(a) Dorsal. (b) Ventral. (c) Anterior. (d) Proximal.

462 Annals of the South African Museum.

and distal expansions standing at 35° to each other; the delto-

pectoral crest is long; the trochlear fossa is fairly deep, and the

bulbous swelling of the distal articulation is strong.

Ulna and Radius (fig. 1, Pl. XLIV, fig. a).—Both bones are large,

long, and massive; the olecranon process is strong; the bulging of

the postaxial proximal half of the ulnar border is nearly as strong

as in B. bane.

Foot (fig. 1, Pl. XLIV, fig. a).—Three feet are partially or wholly

known. They are of moderate size.

Bradysaurus vanderbyli (Haughton and Boonstra).

Gordle (fig. 10).—The scapula is fairly tall, with the acromion

situated fairly low down. The coracoidal plate is fairly long, with

the precoracoid greater than the coracoid. The scapular part of the

glenoid is directed downwards, slightly backwards, and outwards,

the coracoidal part upwards and outwards.

The cleithrum appears to be a straight, rod-like bone lying on the

anterior border of the scapula, to which it is partially fused along its

whole length.

Humerus (fig. 11).—The humerus

is long, with a medium _ proxi-

ss’ mal and a great distal expansion

standing apparently at 50° to each

other ; the delto-pectoral crest is

long ; the trochlear fossa is shallow,

and the bulbous distal articulation

fairly strongly swollen.

Ulna and Radius (fig. 12).—In

the type (3718) both bones are large,

' strong, and massive; the ulna hasa

strong olecranon process; the bulg-

Fic. 12.—Left ulna of Bradysaurus ing of the ulna is quite strong. In

vanderbyli (8941). x4. two other specimens (9169 and 8941)

(2) Doral (0) Ventral. the bones are not very massive :

If the limb were stretched out at

the ulna has no olecranon, and the

right angles to the body the outer or

anterior surface of the epipodial lies bulging is not very strong.

dorsally, and the inner or posterior

carface cea Foot (Pl. XLIII, fig. 1).—Of the

two feet known one is the original

complete Pareiasaur foot described (3). They are quite moderate

in size.

Parevasaurian Studies.—Part VITTI. 463

Genus NocHELESAURUS Haughton and Boonstra.

Girdle.—The whole breast-shoulder-apparatus is large, strong,

and fairly massive. The height and width of the scapular blade e

not accurately known; in one specimen (6239) it is fairly upright,

the acromion is situated fairly low down, and is weakly or moderately

everted. The coracoidal plate is fairly long and fairly high; the

precoracoid is large and is much greater than the coracoid. The

glenoid is fairly deep or shallow, the scapular part is long or very

long, and is greater or much smaller than the coracoidal part. The

direction of the scapular part of the glenoid is downwards, backwards

or slightly backwards, outwards or slightly outwards, the coracoidal

part upwards, backwards, or outwards.

The cleithrum appears to be a rod-like bone partially fused to the

antero-external border of the scapula.

Humerus.—The humerus is fairly large, long, and fairly massive,

with a medium or great proximal and distal expansion, standing at

20°-70° to each other; the delto-pectoral crest is long, with the

proximal corner on the anterior dorso-ventral line weak or prominent ;

the proximal articulation is a narrow ribbon very indistinctly divided

into two surfaces; the trochlear fossa is shallow or deep, and the

bulbous distal articulatory swelling is moderately swollen.

Ulna and Radius.—Unknown in the types, but preserved in another

specimen (8944) identified as N. alexandert. The two bones are of

moderate size and length; the ulna has no olecranon process, and

the proximal articulatory surface passes gently over on to the

proximal end without a limiting ridge ; the bulging of the proximal

half of the postaxial border is weak.

Foot.—Two extremely well-preserved feet of this genus are known.

They are moderate in size and of typical structure, viz. carpal

formula 3, 1, 4, and phalangeal formula 2, 3, 3, 3, 2.

Nochelesaurus strubent (Broom).

Girdle (fig. 13).—The acromion is weakly everted. The glenoid is

fairly deep ; the scapular part is very long and much greater than the

coracoidal. The scapular part is directed downwards, slightly back-

wards, slightly outwards, the coracoidal part upwards and slightly

backwards.

The cleithrum is not known in this species.

Humerus (fig. 14).—The one humerus known has obviously been

464 Annals of the South African Museum.

deformed post-mortem. The proximal end is greatly, but the distal

only moderately, expanded, and as preserved they stand at 70° to

Fic. 13.—Left scapulo-coracoid of Nochelesaurus strubeni (5590). x4).

(a) Lateral. (6) Internal. (c) Posterior.

Fic. 14.—Left humerus of Nochelesaurus strubent (5590). x $.

(a) Dorsal. (6) Ventral. (c) Anterior. (d) Proximal.

each other; the proximal corner on the anterior dorso-ventral line

is prominent ; the trochlear fossa is deep.

Ulna, radius, and foot are unknown.

Pareiasaurian Studies.—Part VIII. 465

Nochelesaurus alexanderi Haughton and Boonstra.

Gurdle (figs. 15, 16).—The acromion is moderately everted. The

glenoid is shallow ; the scapular part is long and much smaller than

Fie. 15.—Right scapulo-coracoid of Nochelesaurus alexanderi (6239). 1,

(a) Lateral. (6) Internal. (c) Posterior.

= =

Fie. 16.—Anterior view of the clavicular girdle of Nochelesaurus alexanderi

(6239). x4.

Cla. =clavicle. I.Cla. =interclavicle.

the coracoidal ; the former is directed downwards, backwards, and

outwards ; the latter upwards and slightly outwards.

VOL. XXVIII, PART 4. 32

466 Annals of the South African Museum.

Cleithrum as for genus.

Humerus.—This humerus is also imperfect. The proximal end is

moderately and the distal end greatly expanded, and the two ends

stand at 20° to each other ; the proximal corner on the anterior dorso-

ventral line is weak ; the trochlear fossa is shallow.

Ulna, radius, and foot as for genus (Pl. XLII, fig. 2).

Genus PLatyoropHa Haughton and Boonstra.

Platyoropha broom: Haughton and Boonstra.

Girdle (fig. 17).—The whole breast-shoulder-apparatus is large,

strong, and massive. The scapula is tall, with a wide blade, inclined

a Bee oa ee C

Fic. 17.—Left coraco-scapula of Platyoropha broomi (5002). x 35.

(a) Lateral. (6) Internal. (c) Posterior.

moderately backwards, with the acromion situated fairly low down

and moderately everted. The coracoidal plate is practically un-

known. The scapular part of the glenoid is directed downwards,

backwards, and outwards.

The cleithrum is a straight, rod-like bone lying on the external

surface of the scapula; except for its distal end, which fuses with the

scapula, it is loosely applied ; it does not reach the top of the scapula.

Pareiasaurian Studies.—Part VIII. 467

Genus EMBRITHOSAURUS Watson.

Embrithosaurus schwarz Watson.

Girdle (fig. 18).—The whole breast-shoulder-apparatus is large

strong, and fairly massive. The scapula is fairly tall and inclined

much backwards, with the acromion low down and moderately

everted. The coracoidal plate is long and fairly high; the pre-

Fie. 18.—Left coraco-scapula of Hmbrithosaurus schwarzi (8034). x 25.

(a) Lateral. (6) Internal. (c) Posterior.

coracoid is small and only slightly greater than-the coracoid. The

glenoid is fairly deep, the scapular part is very long, the coracoidal

short, the former being much longer than the latter ; the scapular part

is directed downwards, backwards, and outwards; the coracoidal

upwards, backwards, and outwards.

The cleithrum is a straight, rod-like bone partially fused to the

antero-external border of the scapula. Owing to the fusion it is

difficult to determine its dorsal extension.

Humerus (fig. 19).—The humerus is large, long, and massive, with

the proximal end moderately, and the distal greatly, expanded ;

the ends stand at 80° to each other; the delto-pectoral crest is of

468 Annals of the South African Museum.

medium length, and the corner on the anterior dorso-ventral line is

prominent ; the posterior part of the ribbon-shaped proximal articu-

Fic. 19.—Right humerus of Embrithosaurus schwarzi (8034). x4.

(a) Dorsal. (6) Ventral. (c) Anterior. (dq) Proximal.

latory surface is greatly widened and more or less distinct from the

anterior part ; the trochlear fossa is fairly shallow, and the bulbous

Fie. 20.—Right ulna of Embritho- Fic. 21.—Right radius of Embrithosaurus schwarz

saurus schwarz (8034). x. (8034). x.

(a) Dorsal. (b) Ventral. (a) Ventral. (b) Dorsal. (c) Distal.

distal articulatory swelling is strong, and the ectepicondyle does not

protrude distally to such an extent as in other forms.

Pareiasaurian Studies.—Part VIII. 469

Ulna and Radwus (figs. 20, 21, Pl. XLIII, fig. 2).—The radius and

ulna are medium-sized bones ; the ulna has no olecranon process, and

the articulatory surface passes gently over on to the proximal surface

without being marked by a limiting ridge ; the bulging of the proximal

half of the postaxial border is weak.

Foot (Pl. XUIII, fig. 2).

One specimen (9148) has been identified as belonging to this genus.

The foot is typical of the family, but quite moderate in size.

Genus DoticHoparrEtA Haughton and Boonstra.

Dolichopareva angusta Haughton and Boonstra.

Girdle (figs. 22, 23, 24, 25, and 26).—The whole breast-shoulder-

apparatus is large, strong, but only fairly massive. The scapula is

Fic. 22.—Right coraco-scapula of Dolichopareia angusta (6238). x75

(a) Lateral. (6) Internal. (c) Posterior.

very tall, with presumably a wide blade, moderately inclined back

wards, with the acromion situated fairly low down and strongly

everted. The coracoidal plate is apparently fairly long and 1s ser

high ; the glenoid is fairly deep, the scapular part 1s very long, the

470 Annals of the South African Museum.

Fic. 24.—Ventral view of interclavicle of Dolichopareia angusta (6238).

Fic. 26-—Posterior view of right clavicle of Dolichopareia angusta (6238).

>|t—t

Olt

Pareiasaurian Studies.—Part VIII. 471

coracoidal part long, and the former much longer than the latter ;

the scapular part is directed downwards, backwards, and uamrdnal

Fie. 27.—Right humerus of Dolichopareia angusta (6238). x4.

(a) Dorsal. (6) Ventral. (c) Anterior. (d) Proximal.

and the coracoidal part upwards and outwards. The cleithrum is

unknown.

Humerus (fig. 27).—The humerus is large, long, and massive, with a

very great proximal and a great distal expansion standing at 30° to

Fic. 28.—Right radius of Dolichopareia angusta (6238). x7.

(a) Ventral. (b) Dorsal. (c) Distal. (d) Proximal.

each other; the delto-pectoral crest is long, with the corner on the

anterior dorso-ventral line prominent ; the posterior part of the

proximal articulatory surface is greatly widened, but not distally

separated from the anterior part ; the trochlear fossa is deep, and the

472 Annals of the South African Museum.

bulbous distal articulation is strongly swollen; there is a strong

muscle-scar near the shaft just ventral to the anterior dorso-ventral

line (brachio-radialis).

Ulna and Radius (figs. 28, 29).—They are large, long, and quite

massive bones; the ulna has hardly any olecranon process, and the

Fie. 29.—Right ulna of Dolichopareia angusta (6238). x.

(a) Ventral. (6) Dorsal. (c) Anterior or Preaxial.

articulatory surface is proximally limited by a fairly sharp ridge ;

the bulging of the proximal half of the postaxial surface is only

moderately strong.

Foot.—Unknown.

Genus BrRacHYPAREIA Haughton and Boonstra.

Brachypareva rogerst Broom.

Girdle (fig. 30).—The whole scapulo-coracoid is of moderate size,

strong, and fairly light. The scapula is very low, with a narrow

blade, inclined somewhat backwards, with the acromion situated

fairly low down and moderately everted. The coracoidal plate is

than the coracoid. The glenoid is deep; the scapular part is very

Parevasaurian Studies.—Part VIII. 473

short, the coracoidal short, and they are of nearly equal size; the

scapular part is directed downwards, slightly backwards, and out-

wards ; the coracoidal upwards, backwards, and outwards.

Fic. 30.—Right scapulo-coracoids of Brachypareia rogerst. x q'y-

(a), (b), and (c) of 4350. (d), (e), and (f) of 8953.

(a), (d), Lateral ; (5), (e), Internal; (c), (f), Posterior.

The cleithrum is not associated with the scapula (4350) ; it would

therefore appear that it was very loosely applied to the scapula.

xX >

(d) Proximal.

Fic. 31.—Right humerus of Brachypareia rogersi (8953).

(a) Dorsal. (b) Ventral. (c) Anterior.

of moderate size, short, and

Humerus (fig. 31).—The humerus 1S

nd a moderate distal

fairly light, with a fairly small proximal a

474 Annals of the South African Museum.

expansion standing at 35° to each other; the delto-pectoral crest is

short, with the proximal corner on the anterior dorso-ventral line only

moderately prominent ; the proximal surface is a fairly narrow ribbon

indistinctly bipartite ; the trochlear fossa is shallow, and the bulbous

distal swelling is moderate and lies more distally than ventrally.

Ulna and Radius (figs. 32, 33).—Both bones are small, short, and

light ; the ulna of the type has a strong olecranon process, and the

Fic. 32.—Right ulna of Brachy- Fic. 33.—Right radius of Brachypareia

pareva rogersi (5012). x2. rogerst (8953).

(a) Ventral. (b) Dorsal. (2) Ventral. (6) Dorsal. (c) Distal.

articulatory surface is clearly limited proximally by a sharp ridge

(in specimen 5012 the olecranon is not so strong and the proximal

ridge is indistinct) ; the bulging of the proximal half of the postaxial

border is weak.

Foot.—The right manus is described in a previous paper (3) ; appar-

ently differs from all the Pareiasaurs in having the phalangeal

formula: 2.353.055 12

Brachypareia watsont Haughton and Boonstra.

Nothing is known of the breast-shoulder-apparatus and fore-limb

of this species.

Genus PRoPAPPUS Seeley.

_ Propappus omocratus Seeley.

Gurdle.—The girdle is practically unknown.

Humerus (fig. 34).—The humerus is of moderate size and length

and is fairly light, with a medium proximal and distal expansions

Pareiasaurian Studies.—Part VIII. 475

standing at 30° to each other; the delto-pectoral crest was

apparently of medium length, and the proximal corner on the

Fic. 34.—Right humerus of Propappus omocratus (1058). x2.

(a) Dorsal. (6) Ventral. (c) Anterior. (d) Proximal.

anterior dorsal-ventral line was apparently not prominent; the

proximal articulation is a fairly narrow ribbon with the typical

double direction ; the trochlear fossa is

very shallow and the distal swelling is

moderate.

Ulna and Radius (fig. 35).—Only the

proximal end of one ulna is present but

is, unfortunately, weathered. The olec-

ranon was fairly weak, but a ridge

apparently limited the articulatory sur-

face proximally; the bulging of the

postaxial border is weak.

Foot.—Unknown.

Propappus parvus Haughton. Fic. 35.—Right ulna of Pro-

ld t at pappus omocratus (1058). x +.

The breast-shoulder-apparatus and fore Cy veiaal, ¢(0) ool

limb of this species is unknown. (c) Anterior or Preaxial.

Genus Paretasucuus Broom and Haughton.

Girdle (figs. 36, 37).—The whole breast-shoulder-apparatus is of

moderate size, moderately strong, and light. The scapula is low,

with a narrow blade, upright, with the acromion situated high up

and moderately everted. The coracoidal plate is apparently very

short and is fairly low ; the precoracoid 1s apparently small and appears

476

Annals of the South African Museum

to be very slightly greater than the coracoid. The glenoid is deep,

short, situated and directed posteriorly.

Fic. 36.—Left coraco-scapula of Pareiasuchus péringueyi (2337). x zy.

(a) Lateral.

(5) Internal.

The cleithrum is not preserved, although the surface for its articu-

lation is apparent on the antero-external border of the scapula ; being

Fic. 37.—Anterior view of clavicular

girdle of Pareiasuchus péringueyr

(233i) ee

Cla. =clavicle. 1.Cla. =interclavicle.

lost, it would appear as if it were

but loosely applied.

Humerus.—The humerus is of

medium size, short and lght,

with a fairly small proximal and

a medium distal expansion stand-

ing as preserved at 20° to each

other ; the delto-pectoral crest 1s

short and the proximal corner on

the anterior dorso-ventral line is

prominent ; the proximal surface

is very much distorted, so that

its nature can, unfortunately, not

be accurately determined. The trochlear fossa is fairly deep and

the distal articulation only moderately swollen.

Ulna and Radius (Pl. XLIV, figs. b and c).—The olecranon process

was apparently quite strong, the proximal limiting ridge to the

articulatory surface present, and the postaxial bulging weak.

Foot.—Only part of the carpus and scattered phalanges are

known.

Pareiasaurian Studies.—Part VIII. 477

Parevasuchus péringueyt Broom and Haughton.

The description of the genus is applicable to this species.

Parevasuchus nasicornis Haughton and Boonstra.

The shoulder girdle, humerus, radius, ulna, and part of the carpus

are too much affected by post-mortem distortion to attempt any

specific diagnosis between the two species. In any case, the differ-

ences would not appear to be very great.

Genus PAREIASAURUS Owen.

Pareiasaurus serridens Owen.

Girdle (fig. 38).—Judging from the parts preserved, the whole

breast-shoulder-apparatus was of moderate size, moderately strong,

and fairly light.

' The coracoid plate is of moderate length and fairly shallow; the

precoracoid is small and is slightly greater than the coracoid. The

y) Hy, yy} SS-— SAS

We > NO) Sess .

Y fi it ; bis SSN NS

[ARRON

y MIAN WL) NY

Fic. 38.—Left coraco-scapula of Pareiasaurus serridens (specimen from

the University of Stellenbosch). x 7-

(a) Lateral. (6) Internal. (c) Posterior.

glenoid is fairly shallow but fairly long, the scapular part is very

short, the coracoidal short, and they are nearly of equal size; the

direction of the scapular part is downwards, backwards, and out-

wards, of the coracoidal upwards, backwards, and outwards. The

only part of the clavicular girdle known is the stem of the inter-

clavicle. This, however, is useless for systematic diagnosis. .

Humerus.—Only the impertect proximal half of the humerus 1s

478 Annals of the South African Museum.

known. In the proximal articulatory surface the posterior part is

widened anid fairly distinct.

Ulna, radius, and foot are unknown.

Genus ANTHODON Owen.

Girdle.—All that is preserved in one specimen (10026) is the stem

and part of the cross-bar of the interclavicle. Only its very small

size differentiates it from the rest of the Pareiasaurs.

Humerus (fig. 39).—The one humerus was fractured at the shaft,

and the joint cannot be vouched for, although the lines of the two

Fic. 39.—Right humerus of Anthodon serrarius (10074). x4.

(a) Dorsal. (6) Ventral. (c) Anterior. (dq) Proximal.

ends afford an indication for the reconstruction. It is a small, short,

and light bone, with fairly small distal and proximal expansions

which apparently stand at only 15° to each other. The delto-

pectoral crest is very short, and the proximal corner on the anterior

dorso-ventral line is prominent; the proximal surface is a long

narrow ribbon, very imperfectly divisible into separate surfaces ; the

trochlear fossa is fairly shallow, but long and different in shape to the

other Pareiasaurs; the distal swelling is weak and directed more

obliquely forwards than in all other humeri, and it is also more

terminal than ventral; the ectepicondyle does not extend distally

to this surface, with which it is distally confluent. The entepicondylar

foramen is situated much ventrally.

Ulna, radius, and foot are unknown.

Anthodon serrarius Owen.

Anthodon minusculus Haughton.

This generic description is based chiefly on specimens which have

been referred to A. serrarius ; but is also applicable to the two distal

Parevasaurian Studies.—Part VIII. 479

ends of humeri from the Ruhuhu area in Tanganyika, named A.

minusculus.by Haughton, the specific differences being apparently

of size only.

MEASUREMENTS.

All the measurements have been taken by Martin’s beam-compass

and are therefore direct lengths, 7.e. projections on to a horizontal

or vertical plane. These measurements must be taken to refer only

to the type specimens or the specimens mentioned, and only in a

general way as indicative of the size of the elements of the species.

There is a marked variation in size between specimens referred to

the same species. The chief value of these tables is that they clearly

mark two main groups differentiated roughly by size.

Clavicular Girdle.

Clavicle. Interclavicle. Cléitties m.

Length :

‘Lene Max 2 Length Max.

dthe eee: of stem. Length. Width.

Br. barna 5127 | 440 90 450 2202 | 235 45

Br. seeleyi 9137 | 430 85 460 320 275 35

Br. vanderbyli . | 3718 | 340 80? 420 220? | 250 30

N. strubena 5590 q q 440 uy Q 2

N. alexanderi 6239 | 445 90 440 4 160? 30

D. angusta 6238 | 475 15 430 285 a a

Pl. broomi 5002 2 a q a 255 45

: (8034 375 65 450 Q 230? 30

= (2642 | 3602 | 70 390 | 280 ? ?

Par. serridens . |\Stellen-| ? 2 Q 250 Q a

bosch

P. péringqueyt 2337 | 305 55 260 Q Q ;

P. nasicornis 3016 | 340 60 390 U Q

Anthodon _ ser-

rarius . 2 a Q 145 -

Annals of the South African Museum.

480

‘qystoyH] (9)

*ploov1od01g jo YAsuoT ‘6 ‘ yysuoT (n)

*plooeloy Jo yysuo'yT *g : plous[y °9 “unUITUTyy (9) “UOTULOIOe WOT} (q)

‘qaed [eplooe10( (q) ‘yysueT (9) : UNUIxeyy, (2) ‘ plous,s wor (7)

‘qaed aepndeog (n) ‘qystoy] (”) > epndeog Jo YYPIM ‘fF : eyndeog jo yYySIoFT ‘Zz

"plously) Jo YyoUOT */, : oyeTq [eploowsoy *¢ “UOTMOIOW Jo YISueT *¢ ‘qysToY [PIO], “T

é é $9 | 98 | 06 | OOT |Z0LT | OTT | 08 é G8 é u é é 9L0€ ; ; UE IESTONS al

668) | 89 | 09) | OL> S85 oss COSIe SOOR SLe OGL 1 5o2) OFG OSE) Sce | SOP LEES ; } , "dad

qosoq

eo) os. Nowmleeememioraite eos lercpocl cr oe pe | eel

Sil SS Ges SOv es SOG TOO Te ReSae ecole Oss OLIN Cis) OLE OLP= O85) 079 PE08 : ; | GIES. ING

é é é é é é é é G8 | OGT | OOT | OYE | OLF GSP é GOOG i : i (LOAN ial

v é CLE sy) ESTE | Oizik ST! eae Ov ie sO6 é O€T | O8€ | OIG OLF | O89 8EC9 : : mysnbun “(7

6 é O6T | Gc | O€L | OZI é é OOT é OOT é é é es 6E¢9 : ; 5 RIO TID INL

SLi_| Sél | SOL | OVI | O€L | SOL | SOE | SFI |) G6 é COT é é é é 069¢ ; : j (Uaqnys * NT

é é KOE Wiz SAIL} totaal é é OIL | O8T | OGI | OLE | OOS O€F é SILE : : * hqvapuna “ag

Ceci COL s Sri ecOie SOS ie O0Ce SeleecO5 | Gols 0cl) Ore | O96. Sr. 029 LE16 ‘ : : whajaas “4g

0OG | SHI | OOT | OFT | OTT | GOT | GFE | OOT | GOT | GOT | GIT | O€€ | OLF OOF | OOL ANE NN mg “ag

é é | OOT | SET | SZT | OST | & | GET | OTL | S6I | OFT | OSE | Gor OSF | O19 | edd) a

q Si 9 ii “9 a) 9 “4 9 ci

6 ‘8 L 9 G v 6 G ll

‘app soyndpay

Pareiasaurian Studies.—Part VIII. 48]

Humerus.

: :

1 2 3 4, 5 Op az 8 i)

a b.

Br. baini : ; . | 5127 | 440 | 235 | 290 | 95 | 145 | 115 | 90 | 85 | 215 | 35°

Br. seeleyi . . . | 5624 | 465 | 255 | 2602] 80 | 115 | 125 | 75 | 80 | 205 | 35

Br. vanderbyli. . . | 8941 | 390 | 195 | 235 | 65 | 120 | 95. | 70 | 65 | 1802] 50°|

N. strubeni . . . | 5590 | 4202] 260 | 2102| 75 | 1052| 1202| 60 | 65 | 200 | 70°|

N. alexander: . ; . | 6239 | 420 | 2002) 265 | ? | 120 | 105 | 80 | 70 | 180?) 20°)

D.angusta .° . . | 6238 | 420 | 280 | 2502] 95 | 130 | 130 | 80 | 70 | 190 | 30°|

Em. schwarz . : . | 8034 | 405 | 230 | 245 | 90 | 120 | 115 | 70 | 65 | 175 | 80

Br. rogersi : . | 8953 | 260 | 175 | 180 | 54 | 100 75 | 55 | 48 | 1388 | 35°]

P. péringuey . : . | 2337 | 290 | 180 | 1402} 602} 65 | 70 | 45 | 55 | 145 | 20°

P.nasicornis . ; . | 3016 | 310 | 160 | 140?) 602) ? 85 | 50 | 50 | 145 | ?

Pr. omocratus . » = . | 1058 | 315 | 205?) 210 | 55 | 120 85 | 60 | 45 | 1607) 30

Anthodon serrarius . . {10074 | 220 | 130 | 128 | 40 80 }) 50:1 45°) 30 | 100 | 16°

Anthodon minusculus ie ‘ 87 60 | 30 | 35 | 25

1. Length. : : ‘ a. length.

2. Maximum width of proximal expansion. De Ibe articulation { b. width,

3. Maximum width of distal expansion. 6. Maximum width of shaft.

4. Width of proximal articulation. 7. Minimum width of shaft.

8. Length of delto-pectoral crest.

9. Degree of “‘ twist” on the shaft.

Ulna.

ES en OPE, eer ee

Width Prox. Art. Surface. | Distance Tee Distal End.

i f Proc. g

Length. ae sll na enna | pi Bi Ne

——_

Br. baina OS Par 380 172 110 90 205 50 50 90

Br. seeleyi .| 5624 380 180 115 95 200 55 45 85

Br. vanderbyli | 8941 335 143 115 00? 153 0 50 75

3718

N. alexandert | 8944 360 120 120 85 160 0 50 OU

D. angusta . | 6238 370 165 sty 95 195 0 45 oi

E. schwarzi . | 8034 | 305 130 107 80 140 0 45 80

B : 5012 245 100 80 535) 130 35 35 mic

r. rogerst 8953 9 1052 67 67 120 35 Q 2

P. nasicornis | 3016 295 115 90 ? 150 35 : e

P. omocratus | 1058 ? 115? 85 75 130? 20 ? :

VOL. XVIII, PART 4.

482 Annals of the South African Museum.

Radius.

Proximal End. Distal End. Shaft.

Length. : :

Length. | Breadth. | Length. | Breadth. | Maximum. Minimum.

Br. baini ae ay br 270 125 90 128 90 57

Br. seeleyi . | 5624 297 130 95 130 85 65

Br. vanderbyli | 9169 200 85 80? 100 70 45

N. alexanderi . | 8944 260 110 80 155 90 60

D.angusta . | 6238 260 130? 112 120 110 57

Em. schwarz . | 8034 230 115 90 105 86 51

Br. rogersi —._| 8953 165 85 58 80 60 40

P. péringueyt . | 2337 190 80 60 87 67 38

P. nasicornis . | 3016 205 ? ? ? 2 ?

COMPARATIVE OSTEOLOGY.

A. Breast-shoulder-apparatus.—The fish Eusthenopteron (48) of the

primitive Osteolepidae is nearly related to the Tetrapod ancestry.

The shoulder-girdle consists of a small, irregular “‘ scapulo-coracoid ”’

with no glenoid cavity, lying internal to the four large dermal bones—

clavicle, cleithra, and supracleithrum. No interclavicle is as yet

developed, and the girdle articulates with the skull.

The primitive condition of the embolomerous Amphibia is ex-

emplified by the fish-like form Hogyrinus (48). This form still retains

the primitive four pairs of dermal bones with a small “scapulo-

coracoid ’’ lying internally. The girdle is still attached to the skull

(post-temporal). The glenoid cavity is very indefinitely demarcated

and faces directly outward, and the motion of the humerus was

unlimited. One more dermal element is added, viz. the median

interclavicle. No perforating foramina are known. The terrestrial

forms, e.g. Diplovertebron (48), show some changes in the shoulder-

girdle, viz. that the glenoid cavity has become of the specialised form

called “‘ screw-shaped,” so that the humerus is defined to movement

along a definite track ; the connection with the skull is lost and the

post-temporal and supra-cleithrum disappear ; parts of the cleithrum

and clavicles become reduced, whilst the interclavicle becomes strong _

and is firmly attached to the clavicles.

In some of the rhachitomous Amphibia, e.g. Hryops (30), the girdle

has become specialised. The massive pectoral girdle consists of

paired scapulo-coracoids, clavicles and a median interclavicle; a

Pareiasaurian Studies.—Part VIII. 483

strong cleithrum caps the scapulo-coracoid, and the coracoidal portion

is apparently small, whereas the scapula is tall; strong supra-

glenoid buttress and fossa. The glenoid cavity is antero-posteriorly

elongated and “ screw-shaped.” The interclavicle is rhomboidal and

1S dorsal to the medial clavicular ends. The cleithrum is of the

2 capping ” type. Miner (30) considers that a sternum and carti-

laginous epicoracoids were present in life. There are three perforating

foramina—supra-coracoid, supra-glenoid, and glenoid.

For our purposes it is necessary to treat the order Cotylosauria in

some detail. In the most primitive group—the Seymouriamorpha

[Seymourva (50), Kotlassia (1, 2)|—the scapulo-coracoid is a composite

bone divided by sutures. In Seymouria there is only one ventral

element—the precoracoid (Watson)—whereas in Kotlassia there are

two—coracoid and precoracoid. In both the coracoidal plate is not

very large, and the scapula is low and broad with a quite strong

supra-glenoid buttress and fossa. Seymouria, like Eryops, has three

perforating foramina, of which the supra-coracoid is large, whilst in

Kotlassia nothing appears to be known about them. In Seymouria

‘the glenoid is “ screw-shaped,” but this part is not described for

Kotlassia ; in the former the scapula and precoracoid, in the latter

scapula and coracoid, form the glenoid. In both the long inter-

clavicle is anteriorly rhomboidally expanded and is ventrally clasped

by the two expanded ends of the strong clavicles. In Kotlassia there

is a strong cleithrum which is absent in Seymouria, where the clavicle

stretches high up on the scapula; according to Watson Seymouria

possessed an ossified sternum. The dermal bones still retain the

sculpturing typical of such bones lying near the surface.

The next group—the Dviadectamorpha—comprises the families

Diadectidae, Pareiasauridae, and Procolophonidae. In the Diadectidae

the two well-known genera—Dvadectes (20) and Diasparactus (22)—

have very similar pectoral girdles. The scapulo-coracoid is apparently

a composite bone consisting of scapula, coracoid, and precoracoid,

although no sutures are visible ; the coracoidal plate is large, and the

scapula is moderately tall and broad in Diasparactus, and tall and

fairly broad in Diadectes. There is a strong supra-glenoid buttress

and fossa. The three perforating foramina are present—in Draspar-

actus the glenoid foramen is exceptionally large. The glenoid is of

the typical primitive bipartite “‘screw-shape.’ In both the inter-

clavicle is anteriorly expanded—T-shaped in Diadectes where the

stem is long, and rhomboidal in Diasparactus, where the stem is of

moderate length. The clavicles are strong, ventrally expanded and

484 Annals of the South African Museum.

dorsally narrow—and clasp the interclavicles anteriorly in a strong

sutural joint. In both the cleithrum is large, with an expanded

upper erid and clasping the anterior border of the scapula—this type

is generally referred to as a “‘capping’’ cleithrum. Case considers

that a cartilaginous epicoracoid may or may not have been present.

No sculpturing of the dermal bones.

‘In the Pareiasauridae the scapulo-coracoid is a composite bone

suturally separated into scapula, coracoid, and precoracoid. The

coracoidal plate is large and the scapula tall and fairly broad in the

genera Bradysaurus, Nochelesaurus, Embrithosaurus ; in Brachypareva,

Parevasaurus, Parevasuchus, and Scutosaurus the former is fairly

small and the latter comparatively low and narrow. The supra-

glenoid buttress is moderately strong, but there is no supra-glenoid

fossa. Only the precoracoidal foramen is present; a pit is possibly

a rudiment of the supra-glenoid foramen. In the first three genera

mentioned the glenoid is fairly typically “screw-shaped”’ and

elongate, whereas in the latter three it becomes shortened and

directed more backwards, then directly outwards. The interclavicle

is T-shaped, with a long stem. The clavicles are strong but slender

and ventrally contracted—they fit into the antero-dorsally deep

groove on the cross-piece of the T-shaped interclavicle. The cleithrum

is rudimentary—more so in the later forms. Features not known

in the forms as yet considered are: the development of a strong

everted acromion process, and associated with this the incipient

development of a supra-spinatus fossa.

In the Procolophonidae the scapulo-coracoid is composed of three

bones—scapula, coracoid, and precoracoid, loosely connected with

each other; von Huene maintains that in Kozloskiosaurus (24) they

are separate elements. The coracoidal plate is relatively large, and

the scapula is low and fairly broad, with a moderately developed

supra-glenoid fossa ; von Huene mentions a small acromion, whereas

Watson says that it is not present in Procolophon (44). There is

only one foramen in the precoracoid. The glenoid cavity is formed

of all three bones and is a trihedral outwardly directed pit, and is not

at all ‘‘ screw-shaped.” The interclavicle is T-shaped, with a long

stem. The clavicles are long, reaching the top of the scapula—

dorsally they are narrow but ventrally expanded where they are

applied to the anterior surface of the cross-piece of the interclavicle.

There is no cleithrum.

The third group of Cotylosaurs—the Captorhinomorpha—contains

the Captorhinidae and Limnoscelidae. In Limnoscelis (50) the

Pareiasaurian Studies.—Part VIII. 485

scapulo-coracoid is composed of three elements suturally joined

scapula, coracoid, and precoracoid. The coracoidal plate is quite

large and horizontal, and the scapula is low and broad, and Williston

is of the opinion that a cartilaginous supra-scapula was developed

in life. The supra-glenoid fossa is large. All three foramina are

present, the supra-coracoid being very large. The glenoid is formed

by the three bones and is typically “‘ screw-shaped.” The inter-

clavicle is anteriorly expanded, with a short cross-piece and a long

stem. The clavicles are long, dorsally narrow but ventrally expanded,

and applied to the anterior face of the interclavicle. There is a small

rudimentary cleithrum articulated to the dorsal end of the clavicle

and the dorsal corner of the scapular edge.

In the Captorhinidae the shoulder-girdle of Labidosaurus (20) is

well known. ‘The three bones are present; the coracoidal plate is

large, and the scapula low and broad, with a very large supra-glenoid

fossa. It appears that the glenoid foramen is not developed. The

glenoid is formed by the three bones, and is typically long “ screw-

shaped.” The interclavicle is long and anteriorly expanded. The

clavicle is large and ventrally expanded ; it does not reach the top of

the scapula, and it is possible that a small cleithrum was developed.

The clavicles clasp the anterior face of the interclavicle.

In the Pelycosauria are included the five families : Casedae, Ophia

codontidae, Poliosauridae, Clepsydropidae, and Edaphosauridae. Con-

ditions are fairly similar within the group—the three elements are

present except in Varanosaurus (50) where the coracoid is unossified,

and in Casea (50) where no sutures are visible. In Casea and Ophia-

codon (50) the coracoidal plate is large, in Dimetrodon (19) and Vara-

nosaurus medium, and in Edaphosaurus (19) very large. The scapula

is tall and narrow in Casea and Dimetrodon, low in the others, and

fairly broad and very broad in Varanosaurus and Edaphosaurus and

Ophiacodon respectively. The supra-glenoid fossa is medium in

Casea, Ophiacodon, and Varanosaurus, large in Dimetrodon and

Edaphosaurus. The glenoid foramen is not present, the supra-

coracoid is always present; the supra-glenoid is absent in Casea, in

the primitive position in Ophiacodon and Edaphosaurus, anterior to

the supra-glenoid fossa in Dimetrodon, and still further forward in

Varanosaurus. The glenoid is formed by the three bones, except in

Varanosaurus, where the coracoid is unossified ; it is typically of the

long, narrow, bipartite “‘ screw-shape.” In Ophiacodon the pre-

coracoidal participation in the glenoid is much reduced. The

interclavicles are very long, anteriorly expanded, spatulate, or slightly

486 Annals of the South African Museum.

T-shaped. The clavicles are large, ventrally expanded, generally

long and dorsally narrowed, and ventrally clasp the interclavicle.

The cleithrum is small, rudimentary in Edaphosaurus and Clepsy-

drops (19), and was possibly present in the others, although it has not

as yet been found. A further point of interest is the process for the

coracoidal head of the triceps, which is very strong in Dimetrodon,

strong in Casea and Edaphosaurus, and moderate in the others.

In the order Deinocephalia a higher developmental niveau is

reached. Of the two families Tapinocephalidae and Titanosuchidae,

the genera Moschops (25) and Jonkeria (13) are well known. In

all there are three bones in the scapular girdle—scapula, coracoid, and

precoracoid—which are firmly ankylosed (Jonkeria, Rhophalodon (46)),

moderately so (Moschops, Phocosaurus (43)), or separate (Struthio-

cephalus). The coracoidal plate is large in Struthiocephalus and

Jonkeria, but only moderately so in‘the others. The scapula is tall

and narrow (Moschops, Rhophalodon, Phocosaurus, and Struthio-

cephalus), or tall and fairly narrow (Jonkeria). In all the supra-

glenoid fossa, buttress and foramen have disappeared. The

precoracoid enters the glenoid hardly or not at all (Phocosaurus).

The glenoid is still a, albeit much modified, “‘ screw-shaped ”’ cavity

in Rhophalodon and Struthocephalus ; it 1s, however, short, situated

much posteriorly and directed somewhat backwards; in Moschops,

Phocosaurus, and Jonkeria it is no longer “ screw-shaped’”’ and

directed backwards. In all only the supra-coracoid foramen passing

through the precoracoid is retained. The interclavicle is long, with

the anterior end expanded and directed upwards; the posterior end

is also much widened. The clavicles are expanded both dorsally

and ventrally, and meet (Moschops) ventrally underneath the inter-

clavicle, which they clasp. The cleithrum isin all a small rudimentary

bone lying on the dorso-anterior border of the scapula. Gregory

considers it probable that in Moschops a cartilaginous supra-scapula

and sternum were developed. An acromion is indicated to a very

slight extent in some forms.

In the order Anomodontia it is only necessary to consider the

terrestrial Dicynodontia (e.g. Dicynodon and Kannemeyeria (46)) in a

generalised scheme. The two coracoids are suturally well attached

to each other, but fairly loosely to the scapula. The coracoidal plate

is quite small, the scapula tall and slender ; there is no supra-glenoid

fossa orforamen. The glenoid is formed by the coracoid and scapula ;

it is not “‘ screw-shaped,”’ but simple, situated far posteriorly and

directed backwards. The supra-coracoid is the only foramen present.

Pareiasaurian Studies.—Part VEL: 487

An acromion, hitherto only encountered in the specialised Pareiasaurs,

is present; it is not strong and hardly everted, and is continuous

with the antero-ventral border. Of importance is, however, the

presence of a surface under the acromion for the “ spinatus ”’

muscles (supra-coracoideus). The interclavicle is a short, squat

bone expanded or T-shaped anteriorly and expanded posteriorly.

The clavicle is expanded ventrally, but by the time it reaches the

acromion it is reduced to a point. The clavicles support the inter-

clavicle ventrally. Both these bones are weak. There is a fairly

well-developed, slender cleithrum applied to the anterior border of the

scapula. It is less rudimentary than in the Deinocephalia. A

pentagonal ossified sternum is here encountered for the first time.

In the Therocephalia the shoulder-girdle is not very well known.

This summary is based on Pristerognathus (14) and Ictidosuchus (16).

The three elements of the scapulo-coracoid are quite firmly united

suturally. The coracoidal plate is small, and the scapula is compara-

tively tall and slender. There is no supra-glenoid fossa, and the

supra-coracoidal is the only foramen. The glenoid is not “ screw-

shaped,” but simple, situated far posteriorly and directed backwards.

The precoracoid takes no part in the glenoid. No acromion is

developed. The interclavicle is relatively long, and expanded at

both ends. The clavicles are expanded ventrally and support the

ventral anterior expansion of the interclavicle. The cleithrum is

small and lies on the dorso-anterior border of the scapula. Broom (14)

sketches a sternum in Pristerognathus minor.

In the Gorgonopsia the shoulder-girdle is fairly well known in

Scymnogathus (18), Lycaenodontoides, and Aelurognathus (16). The

three elements are joined by firm sutures. The coracoidal plate is

small, and the scapula is comparatively tall and slender. There is

no supra-glenoid fossa, and the supra-glenoid is the only foramen.

The glenoid is no longer “ screw-shaped,’ but simple, situated far

posteriorly and directed backwards. The precoracoid does not enter

into the glenoid. No acromion is developed. The interclavicle is

relatively long and expanded at both ends; its stem lies ventral to

the coracoidal plate, not in between as in Pristerognathus ; on its

antero-ventral end it has two sockets in which lie the expanded

ventral ends of the clavicles. In Lycaenops (16) Broom sketches a

small rudimentary cleithrum lying on the dorso-anterior edge of the

scapula. In Aelwrognathus Broom draws a sternum lying dorsal to

the interclavicle and posterior to the coracoids ; medially there also

appears cartilage along the coracoidal edges. -

488 Annals of the South African Museum.

In the Cynodontia (39) the pectoral girdle is fairly well known.

There are still three elements, the coracoidal plate is small, and the

scapula is tall and slender. The supra-glenoidal is the only foramen.

The glenoid is simple, situated far back posteriorly, and directed

outwards and backwards; the precoracoid does not enter into it.

The outer surface of the scapula has now started on its way to the

mammalian complexity, viz. there is a strong acromion process,

and the anterior scapular border is everted to form the beginnings of

the scapular spine, separating an anterior from a posterior spinatus

fossa.

In the Bauriamorpha * the little that is known of the pectoral girdle

indicates a condition very similar to that in Cynognathus.

Within the Mammalia the development continues. In Ornitho-

rhynchus of the Monotremata the precoracoid, for the first time in its

history, loses connection with the scapula. In the ontogeny of

Marsupials the history of the gradual diminution of the coracoid to

a final coracoid process can be traced. For the rest, the differentiation

of the scapular blade initiated by the Cynodonts is increased to the

typical mammalian condition with spine, pre- and postspinous fossae,

and strong acromion.

B. Humerus.—In the Embolomeri the humerus is imperfectly known

in Diplovertebron, and in the bone assigned by Case to Cricotus. It

is a massive bone with expanded ends, not very much greater than

the wide, flattened, very short shaft, and they are in nearly parallel

planes. The proximal articulatory surface appears to have been

‘“strap-shaped.”’ There is no entepicondylar foramen, and the

ectepicondylar foramen is represented by a notch lying between a

supinator crest and an extensor ectepicondylar flange. Both epi-

condyles are fairly small.

In the Rhachitom: the humerus is much better known, especially so

in Eryops. The bone is short, massive, with hardly any shaft, and

greatly expanded ends which stand nearly at right angles to each

other. The proximal surface is “strap-shaped.” There is no

entepicondylar foramen, and the ectepicondylar one is represented by

a deep notch lying between a strong stub-like supinator crest and an

extensor flange. The epicondyles are large, especially the ectepi-

condyle. There is a ventral process subdivided into two separate

crests—a pectoral and a deltoid. There is also a strong processus

* Since the above was written an account of the breast-shoulder-apparatus

of a Bauriamorph—ZE#riciolacerta parva—has appeared. (Watson, Proc. Zool.

Soc., part iii, 1931.)

Pareiasaurian Studies.—Part VIII. 489

latessima dorsi. Distally there is a strong bulbous antero-ventral

radial condyle and a distal trochlear condyle for the ulna.

In the Cotylosaurian Seymouriamorpha the humerus has not been

adequately figured or described in either of the genera Seymouria

or Kotlassia. It would appear that in both it is very similar to that

of the Rhachitomi—in fact Williston described some of the bones of

Seymourva as that of a genus belonging to this group. The humerus,

however, is undoubtedly reptilian in that it possesses a large entepi-

condylar foramen unknown in Amphibia (? Diplocaulus). There does

not appear to be any ectepicondylar foramen. For the rest the bone

is very much of the same type as that of Eryops, viz. short, stout,

without shaft, expanded ends at right angles; “ strap-shaped ”’

proximal end, distal end with ulnar and radial condyle, strong delto-

pectoral crest, probably a supinator process and large epicondyles.

The nature of the humerus differs considerably in three families of

the Diadectamorpha. In the Diadectidae it closely approaches the

primitive condition. In both Diadectes and Diasparactus it is short,

strong, massive, without distinct shaft, and with greatly expanded ends

standing nearly at right angles. The proximal end is ““strap-shaped.”’

The entepicondylar foramen is large, and in Dvadectes the ectepi-

condylar foramen is represented by a deep notch lying between a

strong supinator and ectepicondylar process, whereas in Diasparactus

there is no such notch, as there is no supinator process. The ectepi-

condyle carries a strong process, and the entepicondyle a very broad,

strong flange. In Diadectes, as in Eryops, the anterior proximo-

ventral process is partially subdivided to form separate pectoral and

deltoid processes ; there is also a moderate processus latissima dorsv.

Distally there is a strong bulbous radial condyle and a terminal ulnar

condyle. ]

In the Pareiasauridae, although built on the same ground-plan, it

is possible to consider two indistinctly separated types of humeri.

In all there are two foramina—an oval, slit-like entepicondylar, and a

small, round ectepicondylar ; the possession of the latter is unique

amongst the Cotylosaurs; the supinator and extensor flanges are

confluent. Both epicondyles are only moderately developed. The

delto-pectoral crest does not carry two processes as is typical in

Eryops and Diadectes. There is no distinct processus latissimi dorst.

The shaft is short, distally the ulnar and radial surfaces are partly

‘confluent ; the proximal surface is “‘ strap-shaped ” or ‘ boomerang-

shaped,’ and proximally directed. On the dorso-distal surface

there is a distinct trochlear fossa. In the genera Bradysaurus,

490 Annals of the South African Museum.

Nochelesaurus, Dolichopareca, and Embrithosaurus, the humerus is

long, massive, with fairly great expansions of the ends which stand

at 20°-80° to each other. The distal articulatory surface is hardly

directed distally. In the genera Brachypareia, Parevasuchus, Pro-

pappus, Anthodon, and probably Parevasaurus, the humerus is of

medium size and moderately slender, with only moderately expanded

ends standing at 15°-35°. The distal surface is somewhat more

distally directed, and appears to lie somewhat more anteriorly.

In the Triassic Procolophonidae the humerus is distinctive. It is

long, slender, with a well-developed shaft, quite small expansion of

the ends only slightly twisted on the shaft. The proximal end is

no longer “ strap-shaped,” but triangular. There is an entepicondylar

foramen, but no ectepicondylar, and also no separate supinator and

ectepicondylar (extensor) crests. A slight ridge represents the delto-

pectoral crest. The epicondyles are quite small. The distal ulnar

and radial surfaces are distinct and situated far terminally.

In the Captorhinomorpha, Limnoscelis has a humerus closely ap-

proaching that of Diadectes, and is therefore quite primitive. This

cannot be said of Captorhinus or Labidosaurus. Here the humerus

is more advanced—it is fairly long, with distinct and fairly slender

shaft, although the ends are moderately expanded and still stand at

about right angles to each other. There is no ectepicondylar foramen,

and it is only indicated by a slight notch between small supinator

and extensor crests. The entepicondylar flange is wide. The radial

condyle is practically situated anteriorly. The delto-pectoral crest

is apparently weak.

The humerus in the Pelycosauria is in general a grade higher than

that of the Cotylosauria, and they form, as far as this feature is

concerned, a fairly homogeneous group. There is, however, one

remarkable feature, viz. only in Edaphosaurus is there a ectepi-

condylar foramen. Hdaphosaurus, in this instance, occupies the same

unique position amongst the Pelycosaurs as does the Pareiasauridae

amongst the Cotylosaurs. In the other Pelycosaurs this foramen is

represented by a notch (situated more distally than in Pareiasaurs)

between moderately developed supinator and extensor crests. In

general the humerus is fairly short, strong, and massive, with a

moderately long, slender shaft, and ends not expanded very greatly,

standing at 70°-90° to each other, except in Ophiacodon, where they

are nearly parallel. The proximal end is typically “ strap-shaped,”’

but the distal articulatory surfaces are situated more terminally

than in Cotylosaurs, and the radial condyle or capitellum is situated

Parevasaurian Studies.—Part VIII. 49]

more on the anterior surface. The entepicondylar foramen is always

present, and is very large in Casea.

In the Deinocephalia the humerus of the Titanosuchidae and the

Tapinocephalidae differs in that in the former there is no ectepi-

condylar foramen; they agree in being short and massive (except

Moschops), with expanded ends standing at 45°, or less, to each

other ; the proximal articulation is a long modified “ strap,” and

the distal surface is (especially in Moschops) situated much ter-

minally ; the delto-pectoral crest is very long; the supinator and

extensor processes are confluent and not prominent, and the shaft

is short. On the whole, the Titanosuchids appear more like the

Cotylosaurs, but in the Tapinocephalids this condition is approached

by Struthvocephalus and Tapinocephalus, whereas the more slender

humerus of Moschops appears more like that of the higher Therapsids,

particularly in the more terminal situation of the distal articulatory

surface, indicating a greater power of extension of the epipodial.

In the Anomodontia the humerus of Dicynodon is well known.

The ends are fairly much expanded and stand at about 60° to each

other; the shaft is fairly short and thick, the delto-pectoral crest is

enormous; only the entepicondylar foramen is present; the radial

condyle is ventral, whereas the ulnar is very much more terminal ;

the entepicondyle is moderate; the supinator and extensor flanges

are confluent and fairly weak ; the proximal end is long, but only the

middle part entered the glenoid.

The humerus of Scymnognathus of the Gorgonopsia is well preserved.

It is a light, fairly slender bone, with ends little expanded (especially

the distal end) standing at 30°; the shaft is quite long and slender ;

the delto-pectoral crest is still strong; only the entepicondylar

foramen is present; there is a fairly strong supinator crest; the

proximal condyle is restricted to the widened medial part of the

proximal end, the distal condyles are situated much terminally, the

radial is anterior and somewhat ventral, whereas the ulnar is posterior

and terminal—this indicates a great power of extension of the

epipodial. Here the humerus has definitely become capable of

movement in a vertical plane in contradistinction to the Cotylosaurs

and Pelycosaurs, with their motion practically restricted to a hori-

zontal plane.

The Therocephalian humerus does not differ very much from that of

the Gorgonopsia, except in a greater slenderness and increased length of

shaft, with articular facets typical of a nearly vertical pose of the limb.

The humerus of Cynognathus of the Cynodontia is best known in

492 Annals of the South African Museum.

this group. The ends are moderately expanded and nearly parallel,

the shaft not very long or slender ; the proximal end is long, but the

articular facet is confined to the middle part; the delto-pectoral

crest is long and strong; both foramina are present; the distal

epicondyles are weak ; the small ectepicondylar and supinator crests

are confluent; the distal facets are confluent—the radial situated

antero-ventrally, but the ulnar disto-dorsally, so that the epipodial

could be much extended. The whole humerus indicated a more or

less vertical position with ability of motion in a vertical plane.

The humerus of the Bauriamorpha * is not adequately known, but

appears to agree with that of the Cynodonts.

In the final mammalian generalised scheme the humerus is long,

slender, with lttle-expanded ends lying in the same plane, long

delto-pectoral ridge; entepicondylar foramen; terminal, separate,

distal facets, which indicate an upright position and fully developed

movement in a vertical plane.

C. Ulna and Radius.—Amongst the Embolomeri the epipodial of

Diplovertebron consists of two featureless bones. The ulna is longer,

with the humeral facet terminal.

In Eryops the radius has expanded ends and is short and stout.

The ulna is much longer, with a strong sigmoid notch and process

to hold the proximal end of the radius ; a strong olecranon ; the deep,

concave humeral facet is situated preaxially ; distally the interven-

tion of an intermedium would limit the amount of movement between

radius and ulna.

In the Seymouriamorpha the epipodial differs but little from that of

Eryops.

In the Diadectidae there is no olecranon, and the humeral facet is

not so concave, and was therefore only applied to the humerus, which

it did not clasp.

In the Pareiasauridae an olecranon is present or absent. As in the

forms already considered, there is a deep sigmoid notch into which

the radius fits. The humeral facet is proximo-distally more or less

concave, and thus clasps the humerus more or less. Distally the

intermedium locks the two bones. Except for the inconstancy of the

olecranon the Pareiasaur epipodial is similar to that of Eryops.

In the Procolophonidae both bones are long and slender, the humeral

facet of the ulna is terminal, and there is no sigmoid notch with its

process.

In the Captorhinomorpha the two bones are shorter and more

* Since described by Watson (Proc. Zool. Soc., part iii, 1931).

Parevasaurian Studies.—Part VIII. 49;

massive. The ulna has no olecranon, and a not very deep sigmoid

notch. The humeral facet did not clasp the humerus.

In the Pelycosauria the epipodial is built up of long, fairly light

bones. In the Casecdae, Ophiacodontidae, Poliosauridae there is no

olecranon process, the sigmoid notch and its process are not well

developed, and the humeral facet does not clasp the humerus. In the

Clepsydropidae and Edaphosauridae the olecranon is strong, the

sigmoid notch deep, and its process strong; the humeral facet is

deeply concave, with a strong proximal ridge, so that the humerus was

firmly clasped by the ulna and the power of extension of the foramen

very limited.

In the Deinocephalia the epipodial was extremely short. In the ulna

no olecranon extends to the facet, the sigmoid notch and process

are strongly developed. A great degree of extension was possible.

The Anomodont epipodial is more slender than that of the Deino-

cephs, but for the rest does not differ much.

In the Gorgonopsia and Therocephalia the ulna-proximal facet is

semi-terminal as in Anomodonts and the sigmoid notch and process

are only moderately developed. A great degree of extension was

therefore possible, being aided by the facet on the humerus also being

more terminal.

In mammals the facet of the ulna clasps the humerus in a typical

trochlear joint, but here the great power of extension is due to the

terminally situated facet on the humerus and the trochlear fossa or

foramen.

D. Foot.—In Diplovertebron the phalangeal formula is 2, 3, 3, 3, 4.

Third and fourth digits are of equal length.

In Eryops the carpal formula is 4, 1, 3, 5, the phalangeal 2, 2, 3, 2.

The third digit is the longest.

Diasparactus has the carpal formula 3 (4), 0, 1, 3 (4), and phalangeal

2, 3, 4, 5, 3.

In the Paretasauridae the carpal formula is 3, 0, 1, 4, and the phalan-

geal 2, 3, 3, 3, 2, with one aberrant form, Brachypareia rogersi, with

Bed sichs a5) J.

In the Procolophonia the carpal formula is 3, 1, 0-1, 4-5, the phalan-

geal 2, 3, 4, 5, 3.

In the Limnoscelidae the carpal formula is 4, 0, 1, 2, and the

phalangeal 2, 3, 4, 5, 3; whereas in the Captorhinidae it is 3, 1, 1, 4

and 2, 3, 3 (4), 2.

In Casea and Ophiacodon the carpal formula isp Abuden ds Shy the

phalangeal 2, 3, 4, 5, 3.

494 Annals of the South African Museum.

In Varanosaurus the carpal formula is 4, 1, 0, 4, the phalangeal

2,3 %,47%, 5%, 3.

In Dimetrodon the carpal formula is 4, 1, 1, 5, and the phalangeal

probably 2, 3, 4, 5, 4.

In the Anomodontva the carpal formula is 4, 1, 5, and the phalangeal

2D as OZ .

In Scymnognathus the carpal formula is 3, 1, 1, 4, and the phalangeal

2 ations:

In the Cynognathus the carpal formula is 4, 1, 1, 4, and the phalan-

geal 2, 3, 3, 3, 3.

In the primitive mammalian hand the carpal formula is 4, 1, 4, and

the phalangeal 2, 3, 3, 3, 3.

Summary.—From the above tabular presentation of the salient

comparative facts it is hoped that the morphological position of the

Pareiasaurs, in so far as the breast-shoulder-apparatus and fore limb

are concerned, will be seen in true perspective. It is clear that the

Pareiasaurian ancestry can be traced from a piscine Crossopterygian

beginning through a Stegocephalian stage—and here the Embolomert

are directly concerned, whereas the Rhachitomi exemplify the evolu-

tionary tendencies innate in the Hmbolomeri—then through the

primitive Cotylosaurian niveau to their specialised position within

this group. The rest of the facts presented clearly show that the

Pareiasaurs have nothing whatsoever to do with the Cotylosaur-

Therapsid-Mammalian evolutionary series, and this notwithstanding

the three facts which superficially may give this appearance, viz. the

development of the acromion with its accompanying incipient supra-

spinous fossa, and the development of a simple glenoid cavity, with

the concomitant more upright gait and the mammalian-like digital

formula. These are purely Pareiasaurian reactions to special conditions.

The two fairly distinct morphological stages within the Pareia-

saurian family agree to a great extent with the geological time-

sequence, but although Parevasaurus is found in younger rocks than

Pareiasuchus (and also apparently Scutosauwrus) the former form has

a glenoid cavity more nearly related to the geologically much older

forms than to the only shghtly older Pareiasuchus. Thus, whatever be

the genealogical] relations of the Endothiodon-cistecephalus-zone forms

to the Tapinocephalus-zone forms, Pareiasuchus and Pareiasaurus

cannot have the same descent. This aspect, viz. the phylogenesis

within the family, will, at a future date, be dealt with more in detail.

For the present it suffices to state that there appear to be two morpho-

logical stages which agree roughly with the geological time-sequence.

Paretasaurian Studies.—Part VIII. 495

MYOLOGY.

(Figs. 40-41.)

As the pectoral girdle is not attached to the vertebral column by

any osseous connections, as 1s the case in the pelvic girdle, its suspen-

sion must be muscular. The muscles having this function are the

Lev sc Sup

Fic. 40.—Lateral view of pectoral girdle of Bradysaurus seeleyi (9137). x qo.

To show muscular attachments.

bi. =biceps.

cor. br. =coraco-brachialis.

delt. = deltoideus.

hy. =hyoideus.

lev. sc. sup. =Ilevator scapulae superficialis.

p- = pectoralis.

S. COF. =supra-coracoideus.

sc. hum. ant. =scapulo-humeralis anterior.

sc. hum. post. =scapula-humeralis posterior.

ser. ant. sup. =serratus anterior superficialis.

sub. cor. sc. _=subcoraco-scapularis.

trap. = trapezius.

tric. cor. =coracoidal head of triceps.

tric. sc. =scapular head of triceps.

axial muscles, which must be distinguished from the other muscles

also attached to the girdle. The latter are the appendicular muscles

which move the fore limb. Both the axial and appendicular sets of

muscles are subdivided into a dorsal and a ventral mass.

Annals of the South African Museum.

496

‘sdoorsy JO peoy [eiouNy [eIpow= ,°19 ‘sdoola} JO peoy [esourny [eroyr~= “19 ‘10yeurdns = ‘dns

‘ISIOp SNUIISSTIZV]= “pT ‘SIOXOJ= ‘Pf *S10SU0}X0 = "4x9 "SI[VIper-orpoeig = *1q “sITerpoRag = “eaq

‘SPUOUTTORIYY TBTMOSNUT MOYS OT, “TX (EGG) 1s1abou vrawvdhyooug yo snrowuny yYsrYy—' [TP “O1yT

ag doD Jue WNY IS

“OS AOD ‘VS

‘Of 10) qs pen.

‘sod ung as

Pareiasaurian Studies. —Part VITI. 197

I. AxtaL Musc es.

A. Dorsal Group.

Lhe levator scapulae superficialis originated from the skull, cervical

ribs and transverse processes, and was inserted on the ites intiee:

external surface of the scapular blade and also partly on the edge

and on the cleithrum. The insertion was diffuse. _

The serratus anterior superficialis originated from the ribs and w

inserted on the dorso-postero-external corner of the scapular blade.

This rugose surface is much more distinctly marked than that of the

levator scapulae superficialis.

The levator scapulae profundus and serratus anterior profundus

form a median group. They originate from the transverse processes

on the ribs and are inserted diffusely on the upper portion of the

internal surface of the scapula dorsal to the subscapular fossa.

B. Ventral Group.

Little can be determined of these muscles. On the anterior edge

of the clavicle and at the lateral termination of the cross-piece of the

interclavicle there are muscle scars—probably for the omo-hyoideus.

The internal surface of the coracoidal plate is somewhat hollowed

out, and this was the area of the diffuse attachment of the sterno-

costo-coracoideus group of muscles.

Trapezius.—This muscle belongs to the visceral arch series, and

originates from the skull and antero-dorsal fascia of the back, and is

inserted on the acromion and on the rugose dorsal tip of the clavicle.

Both these areas are well marked on most Pareiasaur girdles.

I]. APPENDICULAR MUSCLES.

A. Dorsal Group.

This group consists of the sub-coraco-scapularis, scapulo-humeralis

posterior, latissimus dorsi, triceps, supinators and extensors.

Sub-coraco-scapularts and scapulo-humeralis posterior.—The area

of origin of these two muscles is not very clearly marked. It

apparently lies on the postero-external surface of the scapula on the

triangular area just above the posterior half of the glenoid, and pos-

terior to the origin of the scapular head of the triceps on the supra-

glenoidal ridge.

The area of insertion is very clearly marked on the dorsal proximo-

VOL XXVIII, PART 4. : 34

498 Annals of the South African Museum.

posterior corner of the humerus. It is a roughened, raised, nearly

circular area proximally reaching nearly to the edge of the bone ;

postero-distally a groove and sharp ridge separates it from the

insertion of the latissimus dorsi; anteriorly a groove separates it

from the area of insertion of the scapulo-humeralis anterior and

supra-coracoideus ; the distal terminating ridge is part of the trans-

verse line on the humerus.

Latissimus dorsi.—This sheet of muscle originates from the pos-

terior dorsal fascia, and converges to a small area of insertion lying

between the transverse line and the latero-median line, and it probably

also includes the rather rugose proximal termination of the latero-

median line.

Triceps.—Posterior to the glenoid there is a shelf on the coracoid,

and it is from here that the long coracoidal head of the triceps arose.

(This appears to be only true for the lower Pareiasaurs ; in the higher

ones this shelf has disappeared, and with it this origin of the triceps.)

A short head of this mass arises from the rectangular area on the

proximo-posterior surface of the humerus lying between the latero-

median line and the postero-ventral edge. The above represent the

medial portion of the triceps. The lateral portion has a long head

from the scapula—the area of origin being the ridge above the pos-

terior portion of the glenoid. This is joined by a short head which

originates from a triangular area on the proximo-dorsal surface of the

humerus lying between the transverse, the lateral median, and the

anterior dorso-ventral lines. The insertion is on the roughened

postaxial convex border of the ulna and in the forms with an olecranon

chiefly on this structure.

Brachio-radialis and Supinator.—The primitive separate supinator

process and the ectepicondylar flange have fused around the ectepi-

condylar foramen in the Pareiasaurs. It is from the dorsal surface

of this compound flange that the brachio-radialis and supinator

originated. It appears probable that the former may have originated

somewhat proximal of the supinator. These muscles are inserted

on the distal outer edge of the radius and the radiale.

Fore-arm Extensors.—Distal to the area of origin of the brachio-

radialis and supinator muscles there is an area facing more or less

antero-distal. This is separated from the area of origin of the two

muscles mentioned by a strong ridge. It is from here that the

various extensor muscles originated and stretched to their various

points of insertion on the distal ends of the epipodial, carpus, and

manus.

Pareiasaurian Studies.—Part VIII. 499

B. Ventral Group.

This group consists of the deltoid, pectoralis, scapulo-humeralis

anterior, supra-coracoideus, coraco-brachialis, biceps, brachialis,

pronators, ard flexors.

Deltoid.—The deltoid primitively originated as a single mass from

the clavicular girdle. In the Pareiasaurs with their rudimentary

cleithrum this muscle apparently originates as a divided mass, the

upper or scapular part originating partly from the cleithrum but

having most of its fibres transferred to the dorso-external part of the

scapular surface. The lower or clavicular part originates from a

well-marked hollow on the dorso-anterior surface of the clavicle.

This part of the deltoid was apparently strong.

The area of insertion on the humerus is very distinct, owing to the

strong development of the delto-pectoral crest. The deltoid inserted

on the anterior face of this flange of bone, which is limited dorsally

by the anterior dorso-ventral line and proximally and ventrally by

the edges of the flange itself.

The chief function of the deltoid was to move the limb up and

forwards and also to rotate it slightly, and had apparently nothing

to do with the formerly supposed digging habits of the Pareiasaurs.

The strong pectoralis had also no “digging function,’ but was

simply correlated with the heavy body and the disadvantageous

horizontal position of the humerus.

Pectoralis.—In the Pareiasaurs the area of bony origin of the

pectoralis is never in doubt, viz. from each half of the ventral surface

and the edge of the stem of the interclavicle. On the dorso-medial

surface of this stem there are grooves to receive the edges of the

coracoidal plate. This has, however, nothing to do with the ventral

and lateral area of origin of the pectoralis. The muscle was inserted on

the ventral knob which lies on the distal end of the delto-pectoral crest.

Scapulo-humeralis anterior, Supra-coracoideus, and Coraco-brachialis.

—The origin of the scapulo-humeralis anterior and supra-coracoideus

muscles from the external surface of the girdle is not very clear.

The scapulo-humeralis anterior would appear to arise dorsally of the

glenoid and anterior to the scapular head of the triceps, whereas

the supra-coracoideus occupied the area ventral to the acromion and

posterior to the clavicle. How far dorsally these muscles extended

cannot be determined. The supra-coracoideus would, however, be

limited by the acromion, as the ventral edge of the scapula is not

continuous with that of the outer surface of the acromion.

500 Annals of the South African Museum.

The scapulo-humeralis anterior is inserted on the proximo-anterior

dorsal surface of the humerus proximal to the transverse line, limited

anteriorly by the anterior dorso-ventral line and posteriorly by a groove

from the area of origin of the sub-coraco-scapularis.

The supra-coracoideus inserted on the proximal extremity of the

anterior dorso-ventral line.

A third muscle is associated with the group of short, deep ventral

muscles. This is the coraco-brachialis, originating from the anterior

external surface of the coracoidal plate and inserted on the proximo-

ventral surface of the humerus between the delto-pectoral crest and

the postero-ventral edge of the humerus.

Biceps and Brachialis——The biceps shares its area of origin with

the coraco-brachialis, 2.e. from the external surface of the coracoidal

plate. The brachialis originates from an area on the anterior face

of the shaft of the humerus distal to the deltoid insertion and proximal

to the supinator-extensor flange. This scar is very strong on some

humeri. These muscles have a common area of insertion on the

proximal ends of the ulna and radius.

Pronators and Flexors.—These muscles originate from the entepi-

condyle and proximal end of the ulna. On the entepicondyle there

can be distinguished a proximal tuberosity for the radial flexors and

a distal one for the ulnar flexors.

Although our knowledge of the Pareiasaur manus, based on

numerous specimens, is very complete, it has not been considered

advisable to treat of the various small toe-muscles in detail, as

Miner (3), for instance, does in the case of the foot of Eryops—a form

in which the osteology of the foot is not even known with any degree

of certainty. To our mind, no useful results can possibly accrue from

any such myological considerations based on an uncertain osteological

basis, and, furthermore, no advance in our knowledge is made by

transferring the minute muscles of Sphenodon or Megalobatrachus on

an Eryops or a Pareiasaur when there is not a clear osteological

‘indication of the muscular attachments. The power of flexion and

extension of the toes can be gauged by considering the toe-bones

without going into the details of the small muscles.

Pareiasaurian Studies.—Part VIIT. 501

LITERATURE.

(1) Amatirzxy, V. P.—‘‘ Seymouridae,”” Akademii Nauk. (Petrograd), 1921.

(2) Amatitzky, V. P.—‘ On a New Cotylosauria of the Family Seymouridae

from the Permian of North Russia,” Ann. Mag. Nat. Hist. (Lond.), ser. ix,

vol. xiii, pp. 64-77, 1924.

(3) Boonstra, L. D.—* Pareiasaurian Studies. Part IIJ.—On the Pareiasaurian

Manus,” Ann. S. Afr. Mus., vol. xxviii, pt. i, 1929.

(4) BouLENGER, G.—‘‘ On the Characters and Affinities of the Triassic Reptile

Telerpeton elginense,”’ Proc. Zoo. Soc. (Lond.), vol. i, 1904.

(5) Brot, F.—“‘ Permische Stegocephalen und Reptielen aus Texas,’ Palaeonto-

graphica, Bd. 51, pp. 1-120, 1904.

(6) Broom, R.—‘“‘ On Ictidosuchus primaevus,” Trans. Phil. Soc. 8. Afr., vol. xi,

1901.

(7) Broom, R.—‘ On an Almost Perfect Skeleton of Pareiasaurus serridens,

Owen,” Ann. S. Afr. Mus., vol. iv, pt. ii, 1903.

(8) Broom, R.—‘‘ On a New Cynodont Reptile (Aelurosuchus browni),” Trans.

S. Afr. Phil. Soc., vol. xvi, p. 376, 1906.

(9) Broom, R.—‘‘ On the Pareiasaurian Genus Propappus,” Ann. 8. Afr. Mus.,

vol. iv, pt. vill, 1908.

(10) Broom, R.—‘‘ The Morphology of the Coracoid,” Anatomischer Anzeiger,

Bd. 41, No. 23/24, 1912.

(11) Broom, R.—‘‘ On a New Species of Propappus, etc., Ann. S. Afr. Mus.,

vol. vii, pt. v. 1912.

(12) Broom, R.—‘‘ On the Manus and Pes of Pareiasaurus,” Ann. 8. Afr. Mus.,

vol. vii, pt. vi, 1913.

(13) Broom, R.—‘‘ On the Carnivorous Mammal-like Reptiles of the Family

Titanosuchidae,” Ann. Trans. Mus., pt. i, vol. xiii, 1929.

(14) Broom, R.—‘‘ On some Recent Light on the Origin of Mammals,” Proc.

Linn. Soc. N.S.W., pt. v, 1929.

(15) Broom, R.—‘‘ On a New Species of Anthodon (A. gregoryi),’ Amer. Mus.

Novitates, No. 448, 1930.

(16) Broom, R.—‘‘ On the Structure of the Mammal-like Reptiles of the Sub-

Order Gorgonopsia,” Phil. Trans. Roy. Soc. (Lond.), B, 457, vol. ccxviii,

pt. viii, 1930.

(17) Broom, R., and HavGHTON, S. H.—<< On the Skeleton of a New Pareiasaurian

(Pareiasuchus péringueyt),’ Ann. S. Afr. Mus., vol. xii, pt. i, 1913.

(18) Broom, R., and Haveuton, 8. H.—<‘ On a New Species of Scymnognathus

(S. tigriceps),” Ann. S. Afr. Mus., vol. xii, pt. i, 1913.

(19) Casz, E. C.—* Revision of the Pelycosauria of North America,’ Carnegie

Publ., No. 55, Washington, 1907. .

(20) Casz, E. C.—“‘ A Revision of the Cotylosauria of North America,” Carnegie

Publ., No. 145, Washington, 1911. .

(21) Casz, E. C.—* The Permo-carboniferous Red Beds of North America and

their Vertebrate Fauna,” Carnegie Publ., No. 207, Washington, 1915.

(22) Casz, E. C., WILLISTON, S. W., and Mest, M. G.—* Pemmo-carboniferous

Vertebrates from New Mexico,” Carnegie Publ., No. 181, Washington, 1913.

502 Annals of the South African Museum.

(23) GELDEREN, CHR. vAN.—‘‘ Over de Ontwikkeling van Schoudergordel en

Episternum by Reptilién,”’ Verslag. Vergad. Wis-Natuurk. Afd. Kon.

Akad. Wet. Amsterdam, vol. xxxi, 1922.

(24) Grucory, W. K., and Camp, C. L.—“‘ Studies in Comparative Myology and

Osteology, No. III,” Bull. Amer. Mus. Nat. Hist., vol. xxxviii, art. xv, 1918.

(25) Grecory, W. K.—*< The Skeleton of Moschops capensis Broom, a Dinocephalian

Reptile from the Permian of South Africa,’’ Bull. Amer. Mus. Nat. Hist.,

vol. lvi, art. iii, 1926.

(26) Harrmann-WernBERG, A.—‘ Uber Carpus und Tarsus der Pareiasauriden,”’

Anatomischer Anzeiger, Bd. 67, No. 19/21, 1929.

(27) HanTMANN-WEINBERG, A.—‘‘ Zur Systematik der Nord-Diina Pareiasauridae,”

Palaeontologische Zeitschrift, Berlin, Bd. 12, 1930.

(28) Haveuton, 8. H., and Boonstra, L. D.—‘‘ Pareiasaurian Studies. Part IL—

An Attempt at a Classification of the Pareiasauria based on Skull Features,”

Ann. 8. Afr. Mus., vol. xxviii, pt. i, 1929.

(29) Hurnnsz, F. von.—‘‘ Ueber die Procolophoniden, ete.,’’ Centrblatt fiir Miner-

alogie, p. 78, 1911.

(30) Miner, R. W.—*‘ The Pectoral Limb of Lryops and other Primitive Tetrapods,”’

Bull. Amer. Mus. Nat. Hist., vol. li, art. vii, 1925.

(31) Nopsca, Fr. v.—‘‘ The Genera of Reptiles,’ Palaeobiologica I., 1928.

(32) OwEN, R.—** Report on the Reptilian Fossils from South Africa,” Trans.

Geol. Soc. (Lond.), ser. ii, vol. vu, pt. iv, 1856.

(33) OwENn, R.—‘‘ Catalogue of the Fossil Reptilia of South Africa, ete.,” Publ.

British Museum (Lond.), 1876.

(34) Romer, A. S.—*‘ The Locomotor Apparatus of Certain Primitive and Mammal-

like Reptiles,’ Bull. Amer. Mus. Nat. Hist., vol. xlvi, 1922.

(35) Romer, A. 8.—*‘ The Comparison of Mammalian and Reptilian Coracoids,”

Anat. Rec., vol. xxiv, No. 2, 1922.

(36) Romer, A. S.—*‘ Pectoral Limb Musculature and Shoulder-girdle Structure

in Fish and Tetrapods,” Anat. Rec., vol. xxvii, No. 2, 1924.

(37) SpELEY, H. G.—‘ Researches on the Structure, Organisation, and Classifica-

tion of the Fossil Reptilia. IJ.—On Pareiasaurus bombidens (Owen), ete.,”’

Phil. Trans. Roy. Soc. (Lond.), B, vol. clxxix, 1888.

(38) SELEY, H. G.—‘ Researches on the Structure, Organisation, and Classifica-

tion of the Fossil Reptilia. VII.—Further Observations on Pareiasaurus,”

Phil. Trans. Roy. Soc. (Lond.), B, 183, 1892.

(39) SrnLey, H. G.—‘ Researches on the Structure, Organisation, and Classifica-

tion of the Fossil Reptilia,” Phil. Trans. Roy. Soc. (Lond.), B, vol. clxxxvi,

1895.

(40) Susxry, P. P.—‘‘ On the Representatives of the Seymouriamorpha, supposed

Primitive Reptiles, from the Upper Permian of Russia, and on their Phylo-

genetic Relations,’’ Occ. Papers, Boston Soc. Nat. Hist., vol. v, pp. 179-181,

1925.

(41) Watson, D. M. S.—* The Limbs of Lystrosaurus,” Geol. Mag., vol. x, 1913.

(42) Watson, D. M. S.—*‘ On the Primitive Tetrapod Limb,’ Anatomischer

Anzeiger, Bd. 44, No. 1/2, 1913.

(43) Watson, D. M. 8.—* The Deinocephalia, an Order of Mammal-like Reptiles,”

P.Z.S., p. 749, 1914.

Pareiasaurian Studies.—Part VIII. 503

(44) Watson, D. M. S.—‘ Procolophon, a Cotylosaurian Reptile from South

Africa,” P.Z.S., p. 735, 1914.

(45) Watson, D. M. 8.—‘‘ On the Nomenclature of the South African Pareia-

saurians,’ Ann. Mag. Nat. Hist., ser. viii, vol. xiv, 1914.

(46) Watson, D. M. 8S.—‘ The Evolution of the Tetrapod Shoulder Girdle and

Fore Limb,” Jour. Anat., vol. lii, 1917.

(47) Watson, D. M. S.—*‘ A Sketch Classification of the Prejurassic Tetrapod

Vertebrates,’ P.Z.S. London, 1917.

(48) Watson, D. M. S.—* The Evolution and Origin of the Amphibia,” Phil.

Trans. Roy. Soc. (Lond.), B, 416, vol. cexiv, vi, 1926.

(49) Wittiston, S. W.—*‘ Cacops, Desmospondylus; New Genera of Permian

Vertebrates,” Bull. Geol. Soc. Amer., p. 280, vol. xxi, 1910.

(50) Winuiston, S. W.—‘‘ American Permian Vertebrates,” Chicago, 1911.

EXPLANATION OF PLATES.

Puate XLII.

Fig. 1. Ventral view of right epipodial and part of carpus of Bradysaurus baini

(5127). x.

Fig. 2. Dorsal view of right epipodial and manus of Nochelesaurus alexanderi

(8944). x3.)

Puate XLIII.

Fig. 1. Dorsal view of part of left epipodial and manus of Bradysaurus

vanderbyli (9169). x =.

Fig. 2. Ventral view of left epipodial and manus of Hmbrithosaurus schwarzi

(9148). x

-ajbe

Puate XLIV.

(a) Anterior view of reconstructed right epipodial and manus of Bradysaurus

seeleyi (5624). x4.

(b) Anterior view of reconstructed right epipodial and manus of Pareiasuchus

péringueyi (2337). x t-

? *

oy =

Ann. 8S. Afr. Mus., Vol. XXVIII.

‘On

‘Toor

ee PEF AS hse

Sige es

Ann. S. Afr. Mus., Vol. XXVIII.

ria.

Ann. S. Afr. Mus., Vol. XXVIII. Plat XLIV

ate Al :

Neill & Co., Ltd.

+? See

Ley

(ny

a Gee oS Index, Title, etc., 1/6. PA ees .

Ss Vol. IV (containing Palacontologieal papers published in conjunetion with

ee ‘Vol. VI—Part, 1, 13/6; Part | 2, 4,6; Part 3, 3/6; Part 4,303 Index, z. :

ae FJ pen i oof a 2 7/-. Complete. —

ees | Part 6, 5/-; Part 7, 30/-3 Part 8, 1/-; Index, Title, etc., 1/6.

. Vol. XXVI.—Complete, 25/—.,

ee ee 10, 7/-3 ‘Part 11, 3/-

“Vol. 31.11. —Part 1, out of print; Part 2, 1/6;

BRS ee = — Part 5, 5/6; Part 6, Tas Part 1; 1/6; Part 8,1

ee Geological Survey).— =——

a -- Part 1, .11)-; Part 2, T=; Part = 4/6; ‘Part 4, 4/6 Part 5, 2

= Part 6, 4/6; Part 7, 14/-; Part 8, 8/-. Ss

Vol. V.—Part 1, 4/6; Part 2, 86 Part 3, 2/6; Part 4, 1/6; Part 5, Qs os

Part 6, 5/-; Part 7 » 3/3 ie 8, 4/6; Part 9,. 4/6; Index, :

Title, ete:, 1 (62355

- Title, etc. ~ L /6.- oa

--Vol.. oo (containing Palaeontological papers published i in 1 conjunction with the g

Se _ Geological survey). _— =

Se Part -t, 3/-; Part: 2, out a prt ark. 3, i z Part 4, 8s

; Part 5, 5/6; Part 6, 1/6; Index, Title, etc., 1/6. pe eee

Vol. Vii. —Complete, out of print. Index, Title, etc., 1/6. _ Bese

Vol. 1 —Part 1, 4/6; Part 2, 5/6; Part 3, 10/-; Part 4, 6/6 ; Part 5, 3/85 2 .

ieee Part 6, lif-; Part a 9/-; Index, Title, ete., 1/6. = Se

ea Vol. X.—Part. 1, 3/—; Part 2; 2/6; Part 3,2/—; Part 4, 3/-; ‘Part 5, 20-3

a SS. a Part 6, 3/-; Part 7, 10/-; Part 8, 2/6; Pait 9, 5/-3 = Part 10, 2/65

Vol. XE —Part 1, 3/6; Part 2, 2/-; Part 3, 13/6; Part 4, 1/6; Part 5, Wes ~

: Part 6, 11/- ; Index, Title, etc., and Plate III, 3/-.

Vol. xu (containing Palaeontological papers published i in conjunction with the

2: Geological Survey).— = =

Sate. = Part hy es Part 2, 3/6; Part 3, 4/6; ‘Part. 4, S/=; Part 5, Was a oe

. Part 6, 6/-; Part 7, 20/-; Part 8, 20/-; Index, Title, etc., 1/6.

‘Vol. XIII. —Part 1, 6/—: Part 2 , 2/6; Part 3, 3/-; Part 4, 8/6;-Part 5, 1/6;

Vol. ‘XIV. —Part 1, 8/6; Part 2, 8; Part 3, 6/-; Part 4, 21/-5 Part 5, Sree a

Part 6, 9/—-: Index, Title, ete:_1/6;2 == eee

BES Vol. XV. —Part 1, LTS Part 2, 17/-; Part 3, aes Part 4, 12/6; Part 5, 5/65 “ete

=e - Part 6, 3/6; Index, Title, ete., 1/6. = 5a

7 Vol. XVI.—Part 1, 30/6; Part 2, 4j—. J aise:

Vol. ‘XVU. —Part 1, 12/-; Part 2, 9/6; Part 3, SI; Part 4, li/-; Part 5, 1): = =

re Part 6, 2/6; Index, “Title, ete. 1j6-- = eee

Vol. XVI. —Part 1, 20/-; Fark 2; 7/65 Part 3, ae Part 4,- 12/6 ; Index, =

- Title, ete., 1/6.=—> -- | Seg

‘Vol. ‘XIX—Part l; 29/-s : Part 2 17/65 Part S AL; Part 4, 5/6: Index, f

: ‘Title, etce., 1/6. - x

Vol. aa —Part 4, 8/6; Part 2, 12/6; ‘Part 3, eee Part 4, 10-5 Part 5, Ans “Seas

- Part 6 (with Title, etc.), 4/6. 3

Vol. XXL Feat 1, 25/-; Part 2 (with Title, etc.), 30/-. so a ren

Vol. XXII (containing Eslacontological papers published in conjunction with the Sh

5 Geological | Survey).— ;

; co Part ee 20/-; Part. 2, 10/-3 Part 3 (with Title, ete. ), 3/6. - ee

Vol. XXIII ——Part 1, ‘12/6; Part 2, 85 Part 3 (with Index, Title, ete. )s 9 :

Vol. XXIV.—Part 1, 10. => = =

- Vol. XXV.—Part 1, 12/6: 5 Part 2, 10-3 ‘Part 3 gaat < Title, ote. )s 10 =

Vol. XXVIT.—Complete, 25; ee eee a

“Vol. XXVIII (containing Palacontologeal papers published j in conjunction with

_ the Geological Survey .— er 2

Part 1, 10/—; Part 2, MW (6: Part 3, 6/-; Part 4 ‘(vith Indes, = =

ee - Title, ete.), 10/-. == Se as

1S Vol. XXIX. —Part 1, 20/-; Part 2 (with Index, Title, etc.), 20). = a 3 ae Z =

Vol. XXX.—Part 1, 15/6. Steers

~ The Annals of. the South Apter Mo useum will be issued at sig intervals ee

as See for publication 4 available. = % a

Fees a ~ Copies may be obtained from—_

: Messrs. WHELDON & 2 Lap.