I'l B RAR.Y

OF THE

U N I VERSITY

OF ILLINOIS

BSO.'S

FB

RfQLQCf

T

STUDIES OF THE AMERICAN FLORA— I

BY

JULIAN AJ STEYERMARK

ASSISTANT CURATOR OF THE HERBARIUM

THE LI3RARY OF THE

J UN 2 01938

UNIVERSITY OF ILLINOIS

BOTANICAL SERIES

FIELD MUSEUM OF NATURAL HISTORY

VOLUME XVII, NUMBER 5

MAY 27, 1938

PUBLICATION 414

^atu^

STUDIES OF THE AMERICAN FLORA— I

BY

JULIAN A. STEYERMARK

ASSISTANT CURATOR OF THE HERBARIUM

THE LIBRARY OF THE

JUN201938

UNIVERSITY OF ILLINOIS

BOTANICAL SERIES

FIELD MUSEUM OF NATURAL HISTORY

VOLUME XVII, NUMBER 5

MAY 27, 1938

PUBLICATION 414

PRINTED IN THE UNITED STATES OF AMERICA

BY FIELD MUSEUM PRESS

580,5

STUDIES OF THE AMERICAN FLORA— I

JULIAN A. STEYERMARK

MORTONIODENDRON, A NEW GENUS OF TILIACEAE

Mortoniodendron Standley & Steyermark, gen. nov. Tiliacea-

rum. — Arbores magnae, pube stellata, ramulis alternis. Folia alterna,

subcoriacea, penninervia, Integra, subtus parce pube stellata con-

spersa. Flores cymoso-paniculati, paniculis terminalibus, ramulis

minute et dense stellato-tomentellis. Sepala 5, valvata, crasso-

subcoriacea, extus dense stellato-tomentosa, intus glabra. Petala

valvata, sepalis paullo breviora. Stamina numerosa, omnia fertilia,

inaequalia, fasciculata, in phalanges disposita, exterioribus brevis-

simis, petalis opposita. Antherae 2-loculares, longitrorsum dehis-

centes, extrorsae, apice mucronatae. Filamenta brevia, basi libera.

Ovarium e calyce liberum, superum, 5-loculare, dense stellato-

tomentosum, longitrorsum rugosum, ovulis numerosis, placentis

protrusis affixis. Fructus capsula globosa, 5-locularis, rugulosa,

4." basi et apice late rotundata, valvis crassis. Semina in quoque

loculo 2-3.

Type species, Mortoniodendron anisophyllum (Standley) Standley

& Steyermark. Panama.

At the time of its description, it was considered improbable

^ that this Panama tree was properly referable to the genus Sloanea.

3^ However, no further attention was given it until recently, when

Professor Samuel J. Record called attention to the fact that its wood

was not that of Sloanea. He, in fact, considers the wood referable to

> the Sterculiaceae, and quite similar to that of the genus Guazuma,

t> with which, of course, the tree can not be considered to be closely

> related.

Mortoniodendron is dedicated to Mr. Conrad V. Morton of the

• United States National Herbarium, who also has made a preliminary

v study of the plant. It is a pleasure to be able to give recognition,

slight although this may be, to the excellent work that he has done

/ upon Mexican and Central American plants, particularly those of

Costa Rica, and in such groups as Viburnum and Gesneriaceae.

Mortoniodendron anisophyllum (Standley) Standley &

Steyermark, comb. nov. Sloanea anisophylla Standley, Field Mus.

Bot. ±: 228. 1929.

411

412 FIELD MUSEUM OF NATURAL HISTORY — BOTANY, VOL. XVII

The conclusion by the authors that this new genus is a member

of the Tiliaceae was reached only after long and careful consideration.

The combination of stellate pubescence with numerous stamens

in definite clusters, with the latter character the first approach to

cohesion of filaments to form a staminal tube, is characteristic of

the Mai vales as defined by Engler and Prantl. The reasons for as-

signing this genus to the Tiliaceae rather than to the Sterculiaceae,

which family was at first suggested, are several. The stamens are

all fertile, no staminodia being present as in most sterculiaceous

genera. The condition of the stamens being opposite the petals,

a situation duplicated by some members of the Sterculiaceae, such

as Theobroma and allied genera, might be somewhat indicative of a

relationship with the latter family were it not for the fact that in

such sterculiaceous genera the stamens are either accompanied by

staminodia or are few in number. In the Tiliaceae numerous

stamens are prevalent and are all fertile, and their grouping into

definite clusters or bundles is found in a number of genera, while

in the genus Mollia of the Tiliaceae some of the stamen clusters

may be opposite the petals as well as the sepals. The valvate sepals

and petals found in Mortoniodendron are further characteristic of

some members of the Tiliaceae, whereas in Sterculiaceae, although

valvate sepals are the rule, no genus is known with valvate petals.

Moreover, the larger flower buds with the thickened sepals covered

by a grayish buff tomentum and the occurrence of the flowers in a

terminal, cymose panicle are much more easily duplicated in a number

of Tiliaceae, such as Tilia, Mollia, Apeiba, and Sloanea, than in most

sterculiaceous genera.

Although Mortoniodendron thus shows affinity to the Sterculi-

aceae, its numerous stamens with no co-existing staminodia would

indicate that this genus has not yet approached the degree of staminal

cohesion and partial sterilization which characterizes most of the

Sterculiaceae. The fact that the numerous fertile stamens occur

in definite groups seems to indicate that Mortoniodendron has

advanced a step toward the sterculiaceous condition, while the fact

that its numerous stamens are all fertile establishes its connection

with the Tiliaceae. Thus, in respect to its staminal characters, the

genus may be said to be a connecting link between the Tiliaceae and

the Sterculiaceae.

Explanation of Figure 26. — Photograph of type collection of

Mortoniodendron anisophyllum (Standley) Standley & Steyermark,

with a dissected portion of the flower drawn (X9) to show pistil,

Tree 40-50 ft. by 18 it. #ith spreadin)

drooping full oroim ana lo* tuttresees

FIG. 26. Photograph of type collection of Mortoniodendron anisophyllum

(Standley) Standley & Steyermark. Flower drawn to show pistil, stamen cluster

opposite petal, and sepal, and a single stamen unattached.

413

414 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

stamen cluster opposite petal, and sepal, and a single stamen

unattached.

Explanation of Figure 27. — Photograph of type collection of

Mortoniodendron anisophyllum (Standley) Standley & Steyermark,

showing the fruit.

NEW SOUTH AMERICAN EUPHORBIACEAE

In attempting to determine several South American Euphorbi-

aceae, a number of novelties have been encountered, of which the

following represent a preliminary study.

Conceveiba magnifica Steyermark, sp. nov. — Arbor 25 m. alta;

ramulis junioribus pilis minutis stellatis conspersis; petiolo 3-6.5 cm.

longo, minute stellato-pubescente, apice obscure glandulari; limbis

late ellipticis vel oblongo-ellipticis, 12-22 cm. longis, 7.5-12 cm.

latis, apice obtusis, basi obtusis, leviter crenatis, membranaceis,

supra viridibus, subtus pallidioribus et prominente reticulatis,

costis 12-16, parce et minute stellato-pubescentibus; inflorescentia

feminea 6.5-9 cm. longa, simplici, elongata, ramulo robusto, 2-3 mm.

crasso, dense pubescente; floribus femineis in pedicellis crassis,

glandulis basi magnis nigris 1.5 mm. longis, 1 mm. latis; sepalis

femineis lanceolatis, acuminatis, 2.5 mm. longis, extus et basi intus

dense tomentosis, basi eglandulosis; ovario ovoideo, dense tomentoso;

columna stylari 1-2 mm. longa; ramulis stylaribus crassis, bilobatis,

recurvato-patentibus, 3 mm. longis. — Brazil: State of Amazonas,

Municipality Sao Paulo de Oliven^a, basin of Creek Belem, basin

of Rio Solimoes, October 26-December 11, 1936, Krukoff 8698

(type in New York Botanical Garden herbarium).

Differs from the other species in its very densely tomentose ovary,

much larger and membranaceous leaves, and in the lack of glands

at the base of the sepals.

Conceveiba Krukoffii Steyermark, sp. nov. — Arbor 15 m. alta,

trunco 10 cm. diametro, saltern ramulis junioribus pilis minutis

stellatis conspersis; petiolo 1.5-4.5 cm. longo, minute stellato-

pubescente, apice indistincte glandulari; limbo oblongo-ovato vel

late oblongo-elliptico, obtuse et abrupte caudato-acuminato, apice

obtuso, distante et inconspicue dentato, coriaceo, reticulato, supra

nitido, concolore, costis et superficie inferiore parce minute stellato-

pubescentibus; inflorescentia mascula non visa; sepalis femineis 5

ovato-lanceolatis, acuminatis, 2-2.2 mm. longis, stellato-pubescen-

tibus similiter ut pedicellis; glandulis basi calycis feminei et basi

Tree 40-50 ft. by 18 in. -with spreading

drooping full crown and low buttresses.

Oreaay rhlte fls. and brown globular fr.

are conspicuous. FT. pendulous on long sta)

ks .splits open on "searee" ae It drlee on

FIG. 27. Photograph of type collection of Mortoniodendron anisophyllum

(Standley) Standley & Steyermark, showing fruit.

415

416 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

pedunculi nigris oblongis, 1.5-1.8 mm. longis, ca. 1.5 mm. latis;

ovario ovoideo, obtuse trigono, tomentello; stylis inferne in columellas

ca. 3 mm. longas connatis, ramulis stylaribus carnosis, bilobatis,

recurvato-patentibus, 4-5 mm. longis; capsula (immatura) 1.5 cm.

longa, 1.3 cm. lata, obtuse trigona, angulis carinatis, 1.3 cm. lata.—

Brazil: State of Amazonas, Municipality Sao Paulo de Olivenca,

near Palmares, basin of Rio Solimoes, September 11-October 26,

1936, Krukoff 8396 (type in New York Botanical Garden herbarium) ;

same locality, on terra firma, high land, Krukoff 8533 (N. Y. Bot.

Gard. herb.).

This species differs from C. guyanensis principally in the longer

style column, 3 mm. long.

Conceveiba simulata Steyermark, sp. nov. — Arbor 20-25 m.

alta, trunco 17 cm. diametro, saltern ramulis et inflorescentiis

firmiter et minute stellato-pubescentibus, ramulis vetustioribus mox

glabris; petiole 1-3 cm. longo, tenui, dense minute stellato-pubes-

cente; limbo elliptico-oblongo vel oblongo, 7-12 cm. longo, 2.5-5

cm. lato, obscure et distante serrulate, apice eaudato-acummato,

basi obtuso vel acutiusculo, supra laevi et nitido, subtus minute et

parce stellato-pubescente, costis eminentibus et ad marginem

adscendentibus, reticulate, concolore; inflorescentia mascula late

pyramidali, stellato-pubescente; inflorescentia feminea non visa;

calyce masculo 3-4-partito, lobis ovatis, acutis, glabris; staminibus

exterioribus plerumque 9, interioribus sterilibus, plerumque 9-10,

flexuosis, exteriores excedentibus, filamentis latis — Brazil: Terra

firma, high forest, State of Amazonas, Municipality Sao Paulo de

Olivenca, basin of Creek Belem, basin of Rio Solimoes, October 26-

December 11, 1936, Krukoff 8616 (type in New York Botanical

Garden herbarium); same locality and date, Krukoff 8609 (N. Y.

Bot. Gard. herb.).

This species is similar to C. guyanensis, from which it differs in

its much shorter staminate inflorescence, and in possessing a greater

number of fertile and sterile stamens, usually 9 fertile and 9-10 sterile.

Mabea Klugii Steyermark, sp. nov. — Arbor 6 m. alta; ramulis

inferne glabris, superne furfuraceo-pubescentibus; petiolo 0.4-0.6 cm.

longo, cano-furfuraceo, puberulento; foliis oblongo-lanceolatis, 7-14

cm. longis, 2-3.8 cm. latis, basi rotundatis, apice abrupte caudatis,

coriaceis, supra laevibus nitidisque, subtus furfuraceis, costa media

et basi limbi densius furfuraceis, minute cano-puberulentis vel fuscis,

subtus pallidis et glaucescentibus, subintegris vel minute crenulato-

STUDIES OF THE AMERICAN FLORA — I 417

denticulatis, costis 15-20 paribus, divaricate patentibus; stipulis

conspicuis, in ramulis florescentibus persistentibus, lineari-lanceolatis,

setaceis, 1.7-1.9 cm. longis, furfuraceis; paniculis 1.8 dm. longis,

in parte mascula 2.5 cm. latis, in parte feminea 3 cm. latis, axe

dense furfuraceo-tomentoso; pedicellis masculis umbelliformibus,

cinereo-puberulentis, 1.2-1.4 cm. longis, basi firmiter valde biglandu-

losis, glandulis nigris, oblongis, 2 mm. longis, 1 mm. latis; sepalis

masculis apice rotundatis, firmiter cinereo-puberulentis; staminibus

25-30; antheris subsessilibus; bracteis stamina subtendentibus

ovatis 4-5 mm. longis, acuminatis; sepalis femineis cum 3 sepalis

exterioribus magnis 4.5-6 mm. longis, 3 sepalis interioribus leviter

longioribus, 5.5-6.5 mm. longis, ovatis, valde acuminatis, intus

dense glandularibus, extus firmiter cinereo-puberulentis, bracteis

subtendentibus conspicuis, ovatis, acuminatis, 0.8-1 cm. longis,

3-4 mm. latis; columna stylari 0.8-1 cm. longa, ramulis liberis 6-7

mm. longis. — Colombia: Forest, Comisaria del Putumayo, Umbria,

0° 54' N., 76° 10' W., alt. 325 meters, January-February, 1931,

Klug 1969 (type in the Herbarium of Field Museum).

This is the second species of the section Intermediae to be known,

both species of the section being thus far limited to Colombia. The

only character common to both the members of this section (M.

Klugii and M. Trianae) is the glandular inner surface of the pistillate

sepals. Mabea Klugii is at once distinguished from M. Trianae by

its much larger pistillate bracts and sepals, by the occurrence of

glands at the base of the much longer staminate pedicels, and by the

peculiar pubescence, especially on the lower surface of the leaves.

In general habit it combines characters found in several species

of section Umbelluliferae. It is related in appearance to M. Taquari

but differs from this species in the basal glands, the longer panicles,

and longer pistillate bracts and sepals; it differs from M. Piriri in

similar characters and in pubescence, and from M. fistuligera it

differs in having a less ferruginous pubescence and an umbelluliform

inflorescence.

Mabea Standleyi Steyermark, sp. nov. — Frutex 2 m. altus;

ramulis glabris, olivaceo-fuscis; petiolo 0.8-1 cm. longo; foliis oblongis

vel elliptico-oblongis, 12-20 cm. longis, 5.5-6.5 cm. latis, subintegris,

apice abrupte caudatis, basi obtusis, omnino glabris, concoloribus,

subtus fuscis, costis subtus conspicuis arcuato-ascendentibus cum

11-12 paribus 1-1.5 cm. distantibus, ante margines extendentibus

cum arcis magnis conjungentibus; stipulis conspicuis in ramulis

418 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

florescentibus persistentibus, lineari-setaceis, 1-1.2 cm. longis; pani-

culis 11 cm. longis, 2.5 cm. latis, axe firmiter cinereo-puberulente,

pedicellis masculis 1-1.3 cm. longis, firmiter cinereo-puberulentis,

basi juxta axem valde biglandulosis, glandulis nigris, globosis, 2-2.3

mm. longis; sepalis masculis apice rotundatis, purpureo-puberulentis,

late rotundatis; staminibus 60-65, breviter stipitatis; sepalis femineis

2.5-3.5 mm. longis, purpureo-puberulentis, ovatis, acuminatis, intus

eglandulosis; columna stylari 1.8-2 cm. longa, cinereo-puberulenta;

stylis recurvatis 6 mm. longis. — Peru: Dept. Loreto, Florida, forest,

Rio Putumayo, at mouth of Rio Zubineta, alt. 200 meters, March-

April, 1931, King 2064 (type in Herbarium of Field Museum).

This new species adds another member to the section Umbelluli-

ferae. The number of stamens (60-65) and narrowly oblong panicles

relate M. Standleyi to M. speciosa, from which species and from

all other species of this section it is distinguished by its suborbicular

glands being directly next to the rachis at the very base of each

staminate branch of the panicle.

Mabea elata Steyermark, sp. nov. — Arbor 10 m. alta, ramulis

furfuraceo-fuscis; foliis oblongis, apice abrupte caudato-acuminatis,

basi obtusis, 8.5-11.5 cm. longis, 2.5-4.5 cm. latis, subintegris,

subcoriaceis, supra in costa media et prope apicem furfuraceo-

puberulentis, in superficie inferiore tota pube fusco-cana puberulentis,

subtus fuscis; costis 15-17 paribus, in superficie inferiore non promi-

nentibus nee ante margines conspicue confluentibus; stipulis absenti-

bus; paniculis 11.5-16 cm. longis, 4-5.5 cm. latis, axe firmiter pallido-

fusco pubescente; pedicellis masculis in racemis umbelluliformibus,

1.5 cm. longis, 2-3 mm. supra basim biglandulosis, glandulis nigris,

pube fusca conspersis, magnis, 3-4 mm. longis, 2 mm. latis, dense

cinereo-puberulentis; pedicellis masculis conspicue bracteatis, bracteis

ovatis, acuminatis, 4-5 mm. longis, 2.5 mm. latis; sepalis masculis

apice rotundatis, cinereo-puberulentis; staminibus 50-60; sepalis

femineis 4.5-6 mm. longis, interioribus longioribus quam exterioribus,

acuminatis, ovatis, pallido-fusco-cinereo-puberulentis, infra nigris

et intus dense cinereo-glandularibus; columna stylari 8 mm. longa,

fusco-tomentosa; stylis 1.2 cm. longis. — Peru: Dept. Loreto, forest,

between Balsapuerto and Moyobamba, alt. 600-1,200 meters,

August-September, 1933, Klug 3206 (type in Herbarium of Field

Museum).

The glandular inner portion of the pistillate sepals allies this

species to the section Intermediae, it being the first of this section

from Peru. In appearance the plant simulates some species of

STUDIES OF THE AMERICAN FLORA — I 419

section Umbelluliferae having numerous stamens, i.e., M. caudata,

M. speciosa, and M. pulcherrima.

Mabea Piririoides Steyermark, sp. nov. — Arbor 12 m. alta

omnino glabra; petiolo 6-7 mm. longo; foliis oblongo-lanceolatis vel

elliptico-lanceolatis, basi acutis, apice abrupte acuminatis, 4-5.5 cm.

longis, 1-2 cm. latis, concoloribus, firme membranaceis vel subcori-

aceis, glabris, obscure crenulato-serrulatis, costis 8-10 paribus

divaricato-adscendentibus, subtus prominentibus; axe et pedunculis

fere omnino glabris, fuscis; paniculis 5-6 cm. longis, 1.5-2 cm.

latis, pedicellis masculis cano-tomentulosis, 2.5-5 mm. longis;

glandulis supra basim minutis, subglobosis vel oblongis, fusco-nigris;

staminibus 27-33; antheris oblongis, sessilibus; pedicellis femineis

cano-tomentulosis; sepalis femineis 6, 2 mm. longis, ovatis, acutis,

cano-tomentulosis, intus eglandulosis; columna stylari ca. 1 cm.

longa, ramulis liberis 4-5 mm. longis. — Brazil: Terra firma, Campo

de Boa Esperanca, State of Maranhao, Maracassume" River region,

October 24, 1932, Krukoffl983 (type in Herbarium of Field Museum).

This new species is a member of the section Umbelluliferae, and

is most closely related to M. Piriri, although also related to M. parvi-

folia and M. subserrulata. In several respects it may be distinguished.

The glands on the staminate pedicels are smaller and less conspicuous,

and are brownish black, not as dark as in M. Piriri or M. parvifolia.

In M. parvifolia the leaves are obtuse at the base, whereas in

M. Piririoides they are acute. In the new species the leaves on

the lower surface are not grayish blue as they are in both M. parvi-

folia and M. Piriri, nor are they finely reticulated as in M. parvi-

folia. The leaves of M. Piririoides are shorter, more coriaceous,

and less serrulate than in M. Piriri and smaller than in M. sub-

serrulata. Finally, the style column in M. Piririoides is longer than

in M. parvifolia.

Richeria submembranacea Steyermark, sp. nov. — Arbor 17 m.

alta, trunco 12 cm. diametro, ramulis glabris; petiolo 0.5-1.2 cm.

longo, minute strigilloso vel glabro; foliis obovato-ellipticis, 7.5-

11.5 cm. longis, 3.5-5.5 cm. latis, supra glabris et obscuris, subtus

glabris, concoloribus, integris, basi acutis, apice abrupte caudatis,

membranaceis vel submembranaceis, costis 6-8 paribus; inflores-

centiis masculis tenuibus, 9-11 cm. longis, dense pallido-fusco-

hirtellis; floribus masculis in pedunculis brevibus 1.5 mm. latis,

sepalis 5, late ovatis, obtusis, extus paullo hirtellis; staminibus 5,

antheris introrsis, glandulis 5 alternantibus; inflorescentia feminea

420 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

non visa. — Brazil: State of Amazonas, Municipality Sao Paulo de

Olivenca, near Palmares, basin of Rio Solimoes, September 11-

October 26, 1936, Krukoff 8513 (type in New York Botanical

Garden herbarium).

This new species of Richeria differs from all the others in its

submembranaceous leaves. From R. grandis it may be distinguished

by its pedunculate staminate flowers, while from R. laurifolia it

at once differs in its thin, entire, abruptly caudate leaves. The

male flowers are smaller than in R. obovata and the inflorescence

and leaves are shorter than in R. racemosa.

Drypetes amazonica Steyermark, sp. nov. — Frutex 6-11 m.

altus; ramulis glabris cinereis; foliis subcoriaceis, ellipticis, basi

acutis apice sensim acuminatis, integris, 12-24 cm. longis, 3-7 cm.

latis, per omnes partes glabris, concoloribus; petiolo 0.4-1 cm.

longo, glabro; floribus masculis 7-numerosis in axillis foliorum

fasciculatis, pedicellis 3-4 mm. longis, tenuibus, sepalis 4, subrotund-

atis, extus sparse strigillosis, margine minute ciliatis, 2.6-2.9 mm.

longis, 2.1-2.3 mm. latis; staminibus 8-10, discum centralem cingen-

tibus; ovarii rudimento nullo; floribus femineis 1-3 in axillis folio-

rum, sepalis 5, rotundatis, 3-3.5 mm. longis, 2.2 mm. latis, extus

pilosulis, rubris; fructu drupaceo, pyriforme, tomentoso, immaturo

1.5-1.7 cm. longo. — Brazil: Amazonas, on varzea land, basin of Rio

Madeira, Municipality Humayta, near Tres Casas, September 14-

October 11, 1934, B. A. Krukoff 6210 (type in New York Botanical

Garden herbarium; fragment in Herbarium of Field Museum); same

locality and date, Krukoff 6176 (New York Botanical Garden her-

barium); same locality and date, Krukoff 6196 (Herbarium of Field

Museum) ; Amazonas, on terra firma, basin of Rio Madeira, Munici-

pality Humayta, near Livramento, on Rio Livramento, October

12-November 6, 1934, Krukoff 6630 (New York Botanical Garden

herbarium); same locality and date, Krukoff 6711, 6734 (New York

Botanical Garden herbarium).

This new species of Drypetes belongs to the section Hemicyclia

(Wight & Arn.) Pax & K. Hoffm. It is related to D. variabilis Uitt.,

from which it differs in the 7-many-flowered instead of 1-3-flowered

staminate fascicles, in the number of stamens, which are 8-10,

rather than 4-7, and in the greater leaf length and long-acuminate

apex. It is also related to the West Indian D. keyensis, from which

it differs in leaf shape and apex, much larger number of staminate-

flowered fascicles with longer filaments, and 4 instead of 5 staminate

STUDIES OF THE AMERICAN FLORA — I 421

sepals, and to the West Indian D. Picardae, which has a different

leaf shape and apex.

A NEW BRAZILIAN SPECIES OF MENDONGIA

Mendoncia Mello-Barretoana Steyermark, sp. nov. — Planta

volubilis, caulibus teretibus, sulcatis, pilis patentibus flavidis hirsutis;

foliis late ovatis vel elliptico-ovatis, apice abrupte cuspidato-acu-

minatis, basi rotundatis vel acutis, 7-11 cm. longis, 4-7.5 cm. latis,

pagina superiore dense pubescentibus, pilis brevibus basi strumosa

nascentibus, pagina inferiore dense molliterque flavido-pubescentibus;

petiolo 1.7-3 cm. longo, dense pilis patentibus flavidis pubescente;

bracteolis oblongo-ovatis, apice rotundatis mucronatis, mucrone

fere 2 mm. longo, 2.5 cm. longis, 1.7-2 cm. latis, pilis brevibus

patentibus flavidis pubescentibus, 0.5-0.8 mm. longis; pedicellis

4-5 cm. longis, dense pilis patentibus pubescentibus; calyce cupulare,

2.5-3 mm. longo, dense adpresso-pubescente; corolla infundibuli-

formi, extus glabra, fere 4.5 cm. longa, tubo fere 2.2 cm. longo,

basi 0.6 cm. diametro, alba, fauce cum rubro lineato; ovario 5 mm.

alto, dense flavido-hispidulo; stylo 2.3 cm. longo, inferne leviter

puberulo, superne glabro. — Brazil: Suyo, Jardim Botanico, Muni-

cipio Bello Horizonte, Minas Geraes, December 1, 1934, H. L. Mello

Barreto 226 (type in Herbarium of Field Museum).

This new species of Mendoncia is characterized by its dense,

short, spreading, yellowish brown pubescence on all parts, large,

infundibuliform corolla 4.5 cm. long, whose throat is striped with

reddish, and densely hispidulous, cupuliform calyx. This last char-

acter separates it at once from the related M. albida, M. puberula,

and M. hirsuta. From the Peruvian M. Killipii it is distinguished

by its much larger and differently colored corolla.

A NEW VIGUIERA FROM MEXICO

Viguiera Shrevei Steyermark, sp. nov. — Herbacea erecta, superne

ramosa, caule valido per omnes partes praeter partes supremas fere

glabro; foliis fere oppositis superne interdum alternis anguste lanceo-

latis acutis distante et obscure denticulatis sessilibus penninerviis,

supra viridibus et minute tuberculato-hispidulis, infra pallidis et dense

canescenti-hispidulis, 7-14 cm. longis, 0.8-2 cm. latis, membranaceis;

pedunculis pluribus, nudis, monocephalis, cano-hirtellis, 8 cm. longis;

capitulis majusculis, ca. 5 cm. latis luteis homochromis; disco 1.2 cm.

alto, 1.7 cm. lato, involucre 2-3-seriato, bracteis exterioribus foliaceis

discum excedentibus lanceolatis acutis 19-27 mm. longis, 3-5 mm.

422 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

latis, 1-nervatis, canescenti-hirtellis, herbaceis; radiis speciosis 15-18,

oblongo-obovatis, apice rotundatis tridentatis 2-2.2 cm. longis,

10 cm. latis; pappo nullo; achaenio glabro. — Mexico: Chihuahua,

stream bank near San Juanito, District of Bocoyna, alt. 2,400

meters, July 26, 1937, Forrest Shreve 8035 (type in Herbarium of

Field Museum).

It is a pleasure to dedicate this new species to Dr. Forrest Shreve,

who has contributed such valuable studies on the flora of the south-

western United States and adjacent Mexico.

Viguiera Shrevei, with its herbaceous, 2-3-seriate involucre and

epappose achene, is obviously a member of the section Heliomeris

(Nutt.) Blake. It appears most closely related to V. multiflora

(Nutt.) Blake, var. genuina Blake, from which it differs in its much

stouter and mostly glabrous stems which are canescent only at the

top and on the peduncles, larger leaves and heads, longer and broader

rays, and conspicuously foliaceous involucral bracts.

ANOTHER NEW SPECIES OF SELENIA FROM TEXAS

Selenia oinosepala Steyermark, sp. nov. — Herba acaulis vel

caulescens annua hiemans glabra; foliis pinnatim dissectis, divisuris

primariis et secondariis pinnatifidis, segmentis ultimis oblongis vel

obovatis apice obtusis; sepalis roseo-purpureis vel vinaceis, 8-9 mm.

longis, oblongo-lanceolatis, caudatis; petalis luteis, 10-11 mm.

longis; stylo 3 mm. longo; siliquis immaturis; seminibus orbicularibus

alatis. — Texas: San Benito, February 28, 1930, Eula Whitehouse

(type in herbarium of the University of Texas; fragment in Herb.

Field Mus.); San Benito, February-April, 1931, Mrs. Paul Cottr ell

(herbarium of the University of Texas).

This new species is the sixth one of the genus to be described,

the most recently published ones being S. mexicana Standley,1

S. Jonesii Cory,2 and S. grandis Martin3 (see p. 443). From all spe-

cies of the genus S. oinosepala differs in its remarkable rose- or pink-

purplish sepals of delicate, petaloid texture. In habit, cut of foliage,

caudate sepals, and length of style it resembles and is most closely

related to S. dissecta Torr. & Gray. Besides the colored sepals, S.

oinosepala differs from S. dissecta in its smaller petals, 10-11 mm. long

instead of 13-15 mm. as in S. dissecta, and in its obtuse rather than

1P. C. Standley. Studies of American Plants. Field Mus. Bot. 17: 191. 1937.

2 V. I. Cory. A New Selenia from the Edwards Plateau of Texas. Rhodora

33: 142. 1931.

3R. F. Martin. A New Selenia from Texas. Rhodora 40: 183. 1938.

STUDIES OF THE AMERICAN FLORA— I 423

acute or acutish ultimate leaf segments. From S. mexicana Standley

and S. Jonesii Cory it differs in its entirely distinctive foliage and

larger flowers.

Selenia oinosepala has been collected thus far only in the region

of San Benito, Cameron County, in the extreme southeastern corner

of Texas. This portion of the Gulf Coast has a number of other

interesting isolated species, among which may be mentioned Grin-

delia oolepis Blake.

GROSS MORPHOLOGY OF THE GENUS GRINDELIA

This paper concludes the monographic treatment of the genus

Grindelia which was started in 1931, and is a part of the series of

"Studies in Grindelia I, II, and III" published respectively in

volumes 21 and 24 of the Annals of the Missouri Botanical Garden.

Monographic studies are now in progress on the genera Chrysopsis,

Gutierrezia, and L/iatris.

HABIT

The genus Grindelia comprises annuals, biennials, and perennials.

Some of the Mexican and Texan species are annuals, whereas most,

if not all, the Californian species are perennials. The perennial

habit is particularly evident in the Pacific coastal species, many of

which frequent salt marshes and tidal estuaries. Species, like

G. lanceolata and G. squarrosa, may be both biennial and perennial.

Most of the biennials in Grindelia produce a many-leafed rosette

the first season and flower the next. Grindelia lanceolata and

G. squarrosa well illustrate two different types of biennials. Grindelia

lanceolata, a late-flowering species, from July to early November,

matures its seeds late in autumn, and during the ensuing winter the

seeds fall to the ground. The following spring, germination normally

occurs and the rosette of leaves which forms carries the plant over

until the following spring when the second season's growth, with

flowering stems, occurs. This species is normally a biennial, dying

at the close of the second season. Sometimes, however, a basal

rosette of leaves remains attached to the stem, endures the winter,

and the plant flowers a third or fourth season or sometimes longer,

as sometimes occurs when plants growing along railroad tracks are

cut close to the ground and the roots are thereby stimulated to

perennate. The same process of cutting the flowering stems close

to the ground will stimulate G. squarrosa to act as a perennial.

This last species, usually a biennial, blooms earlier than G. lanceolata,

commonly June-August, ripens its seeds in late summer, and germina-

424 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

tion mostly takes place the same season, in early or late autumn.

These seedling plants produce a few leaves which tide the plant over

the winter, and the following spring and summer more rosette leaves

are produced, which endure the succeeding winter. Then, the next

year the first flowering stems are produced.

ROOTS

The root system in all cases is that of a tap root rarely extending

more than 6 decimeters below the surface of the ground. The

perennial and often some biennial species develop a semi-ligneous

root, which becomes more and more lignescent with age. The tap

root is usually vertical, but sometimes develops curvatures when

the plant occurs in rocky or gravelly soil or on steep hillsides or in

other habitats where obstructing masses may prevent a normal

vertical growth.

STEMS

The plants are all caulescent, but the stems may vary from such

dwarf types as G. microcephala var. pusilla and G. squarrosa f.

depressa, only a few centimeters tall, to very tall plants, like G.

camporum and G. procera, the latter commonly reaching a height

of 2-3 meters. In the majority of the species the stems are herbace-

ous, arising directly from the crown at the surface, or the crown

may produce short, simple or branched basal caudices which remain

above ground, become semi-ligneous in age, and in turn give rise

to herbaceous stems of the season. These herbaceous stems die down

at the close of each growing season. A number of species, such as

G. humilis, G. Blakei, G. arenicola and G. stricta var. Hendersoni,

but only those inhabiting salt marshes, tidal estuaries, coastal and

sand dunes or beaches, and similar habitats, are suffrutescent or

even suffruticose. In G. arenicola and G. stricta var. Hendersoni

the basal portion of the stem is ligneous, the ligneous portion is

comparatively slender and attains at most a length of 2-3 decimeters,

whereas in G. humilis, especially in older plants, the ligneous portions

may be very conspicuous and occupy over half the length of the

entire plant, the woody cylinder attaining a length up to a meter

or so and a surprising thickness of 6 centimeters. These aerial

ligneous stems are perennial and remain above ground throughout

the duration of the life of the plant, giving rise each season usually

from the upper portions to herbaceous or semi-herbaceous stems

of the growing season. This development of persistent, aerial, suf-

fruticose stems in these littoral species seems to be directly correlated

with the environmental conditions of such habitats.

STUDIES OF THE AMERICAN FLORA— I 425

The stems are simple and monocephalous in the South American

G. globularifolia and G. chiloensis, and in the Mexican G. inuloides

var. glandulosa. The other North American species are practically

always branched, the heads being arranged on mostly corymbosely

to paniculately branched, floriferous branchlets. The stems are

mostly terete and without prominent striations or grooves. In

G. stricta the stems may be slightly flattened and oval or elliptic in

cross section. Some species, like G. integrifolia and G. Howellii, have

deeply grooved or furrowed stems, especially in the upper portions.

The stems may be entirely glabrous or strongly villous or sublanugi-

nose. The hairs are multicellular and often slender, weak, and

crooked. Often the pubescence, when present, is confined toward

the heads, as in G. stricta, or various species may be pubescent

throughout in the early stages of growth and in age become glabrous

below. In the case of species which develop pubescence on the

stems, a short, glandular, multicellular puberulence often occurs,

especially near the heads. In addition to the glandular hairs being

mixed with the non-glandular hairs, the stems of certain species,

like G. inuloides var. glandulosa, may become almost predominantly

glandular.

Erect or ascending stems are the general rule in the genus. In

G. stricta var. procumbens the stems are procumbent, widely and

almost horizontally spreading. Grindelia arenicola and var. pachy-

phylla often develop procumbent stems from the suffrutescent

basal axes.

LEAVES

The leaves in the same species are tremendously variable in

outline and margin but at the same time very characteristic, and

often serve as taxonomic criteria, at least in part, in the matter of

the differentiation of the various entities. The radical and lower

cauline leaves may be conspicuously pinnatifid, lyrate, laciniate, or

incised-dentate. As the middle cauline leaves are approached, the

pinnatifid or laciniate condition becomes less marked, the divisions

become reduced and more crowded, and with the gradual reduction

in size of the upper cauline leaves and those on the floriferous branch-

lets the margin becomes more regularly serrate, dentate, or crenulate.

Such a series of changes is to be noted in G. squarrosa var. nuda,

G. grandiflora, G. texana, and G. microcephala var. adenodonta.

In some of the South American species, like G. chiloensis and

G. globularifolia, the stems are more leafy near the base, whereas

in most of the North American species this condition does not exist,

426 FIELD MUSEUM OF NATURAL HISTORY — BOTANY, VOL. XVII

the leaves being more or less scattered on the stem from base to

apex; only in some Mexican species, like G. inuloides and G. inuloides

var. glandulosa, which often have very reduced leaves in the upper

portion of the stem, is there any suggestion of a condition such as

exists in the South American species mentioned.

The leaves are always alternate, and all, except the basal and

lowermost cauline, which are mostly petiolate, are sessile. The

main cauline leaves may be conspicuously narrowed below the mid-

dle to the base, as in G. nana, G. subalpina, and G. stricta, or they

may be strongly ampliated and amplexicaul or subcordate as in

G. rubricaulis var. latifolia and G. rubricaulis var. robusta. The mid-

dle and upper cauline leaves of many of the species are amplexicaul

or subamplexicaul.

The shape of the leaves varies considerably even on the same

plant. The basal and lowermost are usually attenuate into a petiole-

like base, making the leaf spatulate to obovate. As the middle and

upper cauline leaves are approached, the basal portion tends to

become broader and less attenuated, and when the uppermost

cauline leaves and those on the floriferous branchlets are reached,

the basal portion in the majority of the species has come to be as

broad or much broader than the breadth about the middle or apex.

Such transitional changes in general shape and outline of the leaves

require comparison of leaves from approximately the same areas on

the main stem or branchlets if differentiation is to be made.

The texture of the leaves is exceedingly characteristic of many

species. Most of the species have firmly membranaceous to sub-

coriaceous leaves, but there is much diversity. Grindelia integrifolia

has membranaceous leaves, quite thin for the genus. Many of the

coastal species show interesting correlations between leaf texture

and environment. For example, G. Blakei and G. humilis and

varieties of salt marshes and tidal estuaries develop leaves of very

thick, coriaceous or leathery-fleshy texture, whereas G. stricta,

G. stricta var. procumbens, and G. stricta var. macrophylla of tidal

estuaries, salt marshes, or sandy beaches along the seashore have

soft, fleshy leaves which are quite thin when dried and pressed.

Grindelia arenicola var. pachyphylla of coastal sand dunes also

develops thick, coriaceous or leathery-fleshy leaves.

The color of the leaves is a remarkably constant and definite

taxonomic-genetic character for the different species. The dull

or pale grayish or bluish green shades definitely mark such species

as G. perennis, G. squarrosa, G. squarrosa var. nuda and var. serrulata,

STUDIES OF THE AMERICAN FLORA — I 427

pale yellow and light grass-green stamp G. integrifolia, G. lanceolata,

G. nana, G. arizonica, and G. texana, whereas various types of

dark greens characterize G. humilis, G. Blakei, and G. rubricaulis

var. robusta.

Also of distinct taxonomic significance are the relative abundance

and conspicuousness of resinous secretion in the leaves. In G. cam-

porum, G. squarrosa, and its var. nuda the leaves are conspicuously

punctate with resin-secreting areas, and the abundant resinous

secretion frequently gives the surface a lustrous or almost varnished

appearance. In other species, namely G. hirsutula var. brevisquama,

G. stricta var. macrophylla, G. lanceolata, G. inuloides, G. Greenmanii,

and G. oolepis the resinous secretion is so slight or negligible and the

punctate areas so inconspicuous that the leaf does not appear at all

resinous-punctate. Some species, like G. squarrosa and its var.

nuda and G. microcephala var. adenodonta, have the closely and

finely pectinate-serrulate or denticulate margins conspicuously resini-

ferous at the apex of the teeth, whereas many of the species lack

resiniferous- tipped teeth.

ARRANGEMENT OF HEADS

The real inflorescence, as in all Compositae, is a capitulate

one with a centripetal plan. In defining the type of inflorescence

treated, the writer has regarded a single head as if it were a single

flower. The mode of branching in the genus is always centrifugal or

determinate, with the central head, which terminates the first

developed axis, being the first to flower. The heads are sometimes

disposed terminally on simple, unbranched stems, as in G. globulari-

folia, G. chiloensis, and G. inuloides var. glandulosa, but more often

are arranged on few- to many-branched stems. The mode of branch-

ing (not as to flowering) is most often corymbose or subcorymbose,

typically exemplified by G. squarrosa and G. lanceolata. By elonga-

tion and irregularity in branching of the lateral axes a paniculate

or subpaniculate type of branching may develop, as in G. inornata,

G. procera, G. humilis, and G. fastigiata. When the lateral axes

become more approximate and closely ascending they may produce

a fastigiate appearance, as in G. arizonica var. stenophylla, G. nana

and var. turbinella, and G. fastigiata. In G. aggregata the heads are

agglomerated into a compactly arranged cluster, terminating sim-

ple, short, subracemosely or subcorymbosely branched floriferous

branchlets.

428 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

HEADS

The heads are always homochromous and yellow; both radiate

and discoid heads occur. Among the South American species only

one discoid species, G. discoidea, is known. In Mexico G. oxylepis

var. eligulata is discoid, whereas in the rest of North America discoid

examples are met in G. columbiana, G. rubricaulis var. bracteosa,

G. oolepis, G. aphanactis, G. fastigiata, G. inornata, and G. squarrosa

var. nuda. Those species of Grindelia possessing radiate heads are

by far in the majority.

When the heads are discoid all the flowers are tubular, acti-

nomorphic, and perfect. When the heads are radiate and hetero-

gamous, all the tubular disk florets are perfect, and the radiate

florets are ligulate and pistillate.

The heads in all the species contain numerous florets, and a

direct correlation exists between the size of the head and the number

of its florets; such species as G. oolepis, G. decumbens, G. laciniata,

G. oxylepis L capitellata, G. nana var. integerrima, G. tenella, and G.

Robinsonii with comparatively small heads possess far fewer florets

than such large-headed species as G. rubricaulis var. robusta, G.

lanceolata, G. subalpina var. erecta, G. Blakei, and G. camporum.

While the heads of a given species average a more or less constant

number of florets per head, yet the use of such a character has not

been found of significance in the taxonomy of the group.

DISK

The shape of the disk is a definite character to be relied upon in

Grindelia. The genus as a whole is typified by a subhemispheric or

campanulate-hemispheric type of disk. In a number of species like

G. squarrosa, G. camporum, and G. rubricaulis var. robusta the disk

is depressed-hemispherical, in G. arizonica, G. Hallii, G. decumbens,

and others it is campanulate or campanulate-hemispherical, and in

well developed heads of G. fastigiata and G. nana var. turbinella it

becomes deeply campanulate.

INVOLUCRE

After many attempts toward a natural classification of the

genus, working with different criteria, the author has had to rely upon

the characters of the involucre as a basis for major or minor divisions.

As a criterion of definite taxonomic value the involucre serves not

only as a convenient and readily usable character but also as an

extremely natural one; moreover, the involucral characters in Grin-

STUDIES OF THE AMERICAN FLORA — I 429

delia are as fundamental as those of the pappus and achene. The

involucre is always more or less graduated; sometimes the outer-

most bracts subtending the heads become so foliaceous as to obliterate

the graduated appearance, as in G. hirsutula L patens. The involucre

is mostly 4-6-, sometimes 7-8-seriate.

Perhaps as a result of an evolutionary adaptation to environ-

mental conditions, the resinous character of the involucre has become

definitely tied up in the genetic constitution of the various species.

Species of Grindelia growing in arid, dry, wind-swept habitats,

where paucity of rainfall exists and where evaporation is great, as

on prairies, plains, dry places along streams, semi-desert areas, and

the like, often develop large quantities of resinous substances on the

involucre. Examples of this type of correlation are found in G. nana,

G. squarrosa, G. fastigiata, and G. perennis. On the other hand,

species growing under conditions of more abundant rainfall or where

the habitat is otherwise associated with more moisture often show

much less quantities of resinous exudate, as in G. integrifolia, G.

stricta and var. macrophylla, and others. The amount of resin

present differs according to the species. It is very conspicuous and

abundantly secreted in G. nana, G. squarrosa, G.camporum, G.Blakei,

and G. fastigiata, whereas in G. oolepis, G. Greenmanii, G. Palmeri, G.

hirsutula var. brevisquama, G. scabra var. neomexicana, G. stricta var.

macrophylla, and others it is so scarce as to appear almost lacking;

in many of these cases only the innermost bracts or the youngest

heads produce any obvious resin secretion.

Many of the species, preceding and during the first stages of

anthesis, exude a large amount of resinous substance which accumu-

lates in excess at the summit of the head in the form of a white or

cream-colored, sticky, viscous mass. This sticky mass diminishes

in extent more or less in proportion as the centripetal evolution in

anthesis progresses, some of it probably evaporating. This phe-

nomenon is well observed in such resinous species as G. squarrosa,

G. subalpina, G. nana, G. camporum, and many others.

The position of the involucral bracts is a very definite and useful

character in distinguishing the various species. Whether the upper

portions of the bracts are erect-appressed, ascending, spreading,

reflexed-squarrose, or revolute is of fundamental taxonomic impor-

tance, but it is a character which must be used and studied carefully,

and due allowance must be given for slight variations or for

abnormally developed involucres. Interpretation of the position of

the bracts on herbarium material may sometimes lead to confusion,

430 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

particularly when one is not experienced with the group. For exam-

ple, a species which has the upper portions of the bracts revolute or

strongly recurved or reflexed-squarrose may, when pressed and dried,

have those bracts in the plant press straightened or curved out of

the normal position. On the other hand, the upper portions of

bracts which in nature are ascending or spreading, with straight

tips may, upon being placed in the plant press, appear to be squar-

rose or abruptly bent back. If in such cases one will examine those

portions of the involucre and bracts which have been least disturbed,

such as near the base of the involucre or the sides, which have not

had direct pressure applied to them, one usually will be able to decide

what is the normal position of the bracts.

For the sake of clarity, the author desires at this point to explain

the terms used in the keys and descriptions of the monograph for

the various positions of the involucral bracts. Assuming that the

original position of the upper portion of a bract were erect-appressed,

it would pass through the following changes in the natural sequence

of turning down, namely: erect-appressed, erect-ascending, ascend-

ing, spreading, horizontally spreading with straight tips, simply

squarrose, slightly reflexed-squarrose, moderately reflexed-squar-

rose, strongly reflexed-squarrose, or recurved, then loosely and

finally strongly revolute. Theoretically, all these terms denote

definite positions; practically, only a few of them have been

maintained.

Erect, ascending, and spreading are clear enough, but the stages

between squarrose and revolute types grade into one another, and

are sometimes difficult to construe. The bracts are squarrose only

when the upper portions are projecting or spreading horizontally

outward with the tips slightly sloping gradually downward or when

slightly descending; in other words the bracts describe a more or

less low-convex arc. As the tips of the bracts become more abruptly

bent downward they pass into the reflexed-squarrose condition.

At times, as often in G. camporum, the flexure at the tip is so abrupt

downward in respect to the horizontally spreading portion as to

give almost a geniculate or deflexed condition. For the sake of

clarity and because the deflexed condition is often inseparable from

a moderately or strongly reflexed one, only the term reflexed-squar-

rose is here used. The reflexed-squarrose condition may be only

slightly marked or may become strongly bent downward. From

this position, by the tip becoming more and more turned down-

ward and then inward and upward, a recurved condition arises.

STUDIES OF THE AMERICAN FLORA — I 431

When the tip becomes farther curved upward it assumes a revolute

appearance. A loosely or openly revolute bract is one in which the

curvature described is rather gradual, and in which the revolute

part of the bract shows obvious spatial separation between proxi-

mate portions. This condition is well exemplified by G.Blakei. A

closely revolute condition is here used to express a further degree of

revoluteness, in which the curvature described is more abrupt, and

in which the revolute part of the bract is more tightly fitted within

the curvature, giving a crowded, compact appearance to the figure.

This is well observed in G. revoluta and G. nana and var. integerrima.

The upper portion of the involucral bract may be involute, the tip

becoming upwardly and inwardly recurved. This condition is best

seen in G. grandiflora.

The upper portion of the involucral bract then is quite character-

istic and for the most part constant. All the Mexican and South

American species have either erect, ascending, spreading, or slightly

reflexed-squarrose tips (the outermost may be slightly recurved at

tip), but they are never revolute nor conspicuously recurved as in

many North American ones.

In G. oolepis, G. arizonica, and G. hirsutula var. brevisquama the

bracts as a rule are broadly lanceolate or ovate-lanceolate with short,

acute to slightly acuminate tips. In most of the species the apex

becomes more or less elongated, often slenderly filiform-subulate.

The latter condition is very accentuated in G. integrifolia, G. stricta

var. macrophylla, and G. lanceolata. In the last two species the free

portion of the tip may be as long as 10-12 mm. and the outer and

middle bracts may be free for one-half to three-fourths their total

length, whereas in species like G. humilis, G. arizonica, G. nana var.

integerrima, and others, in which the uppermost portion only is

loose, the erect-ascending or revolute portion may be free only

one-fifth or less the total length. There are all gradations between

these extremes.

The texture of the bracts is mostly correlated with that of the

leaves. Thin, membranaceous-leafed species, like G. integrifolia,

possess similar types of bracts, whereas species possessing subcoria-

ceous or coriaceous leaves, like G. squarrosa and G. arenicola var.

pachyphylla, show a similar development in texture in the involucral

bracts. The tips of the involucral bracts may be conspicuously

thickened and terete, as in G. squarrosa, G. perennis, G. fastigiata,

G. nana, and G. revoluta, or may be only scarcely or not at all thick-

ened and mostly flattened, as in G. stricta var. macrophylla, G. inte-

432 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

grifolia, and G. hirsutula var. brevisquama. In most of the species

the upper portions of at least the inner and often the middle and

outer bracts are resinous-punctate, either conspicuously and abun-

dantly to slightly or less marked. In some of the thin, membra-

naceous, flattened, and scarcely or not at all thickened types the

resinous-punctate areas are almost negligible.

In most of the species the involucral bracts are entirely glabrous.

A short, villosulous or hirsutulous type of pubescence is found on

the bracts in G. hirsutula and var. brevisquama and f. tomentulosa,

G. rubricaulis, G. stricta var. Hendersoni, and var. lanata. Sometimes

a short, glandular puberulence occurs on the outer or middle bracts,

as in G. inuloides var. glandulosa.

In some species, like G. arizonica, the tips of the outer and middle

bracts are conspicuously greener than the lower portions, whereas

in G. scabra, G. stricta, G. lanceolata, and many others the bracts are

uniformly green throughout.

RECEPTACLE

The receptacle of Grindelia is mostly flat or very slightly convex.

In some species, such as G. squarrosa, the sides distend conspicuously

downward in fruit, but most of the species do not exhibit this

property. The receptacle goes through a definite sequence of de-

velopment from the period of an thesis to that of fructification.

The author has been able to observe from field study the series of

changes which take place in G. squarrosa and G. lanceolata.

At least in G. squarrosa and G. lanceolata the obvious changes

are as follows: the centripetal sequence of anthesis, beginning with

the ray florets, then the outer disk florets, and subsequently more

and more centripetally, allows the outer portion of the head to

flower earlier and to be pollinated sooner than the inner and central

part, and just as soon as the first ray and outer disk florets have

come into anthesis, various dipterous, hymenopterous, and coleop-

terous insects ensure pollination and resulting fertilization. Conse-

quently, these outer florets begin the process of ripening the seed

even before the central and inner ones have even reached anthesis.

The centripetal development of the florets in G. squarrosa goes on

at a faster rate, with consequently a shorter period of anthesis for

the entire head, than the corresponding development of the florets

in G. lanceolata.

During the centripetal evolution of the florets in each species

the receptacle increases in size. The outer margins of the receptacle

STUDIES OF THE AMERICAN FLORA— I 433

grow and gradually extend upward. During these changes the in-

volucral bracts increase in size and elongate somewhat. Gradually

the involucral bracts, especially the inner ones, together with the

vigorous upwardly growing margin of the receptacle bend upward

and inward centripetally. The inner bracts extend more and more

centripetally and begin to close over the center of the head. Many

of the central disk florets have not reached anthesis at this time or,

if they have reached it, have been in that stage only a relatively

short time.

Finally, the head becomes practically closed above by the growth

inwardly of the receptacle and inner involucral bracts. At this stage

the shape of the disk has become changed from its original appear-

ance at the beginning of anthesis; it has now taken on a broadly

conical or turbinate shape, the sides at the middle or base distending

conspicuously outward and narrowed conspicuously at the top.

When the head is examined at this stage the innermost disk florets

frequently are found not to have been pollinated nor even to have

reached anthesis. The heads remain closed over in this condition

during the ripening and maturing of the seeds.

After a considerable lapse of time, as much as two or three months,

the inner involucral bracts and the receptacle begin to show signs of

losing their turgidity and original position; the inner bracts begin

to bend upward and gradually outward while the margins of the

receptacle are going through similar outwardly directed movements.

Gradually the central and outer portion of the head again become

exposed. In time the entire head is open and, except for its obvious

increase in size, takes on the appearance which it possessed at the

beginning of anthesis.

The opening of the head occurs only after the seeds are fully

ripe. We see that the most striking phenomenon has taken place,

namely the ray and outer disk florets have ripened seed, but prac-

tically all the numerous central and innermost disk florets (always

hermaphroditic) have failed to ripen. This fact the author believes

to be explained as follows: the ray and outer disk florets, being the

first in the head to be in anthesis, require sufficient nourishment to

ripen their seeds. The more central disk florets are placed at a dis-

advantage in flowering much later, for by the time they have gone

into anthesis and have been pollinated the outermost ones have

already begun to ripen seed and have monopolized the food supply,

with the result that the central disk florets are gradually starved at

the expense of the outer ones.

434 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

Also correlated with the lack of formation of seed in the central

disk florets is the fact that the head becomes closed over by the

involucral bracts before the central disk florets have gone into

anthesis, or when they have been in anthesis only a short time.

While these observations were made in the field only on G. lanceolata

and G. squarrosa, the author has observed among the thousands of

herbarium specimens that most, if not all, of the other species in the

genus show heads which have produced ripe seeds on the ray and

outer disk florets, whereas the central and innermost disk florets

have failed to ripen seed.

The receptacle in most of the species is deeply foveolate. In

G. oolepis there is scarcely any indication of a foveolate receptacle.

The nature of this foveolation is worthy of note. Each foveola is

bounded more or less on four sides by fleshy processes, which in

anthesis surround the entire ovary, and in fructification subtend the

base or lower part of the achene. These fleshy processes are often

thick and turgid during anthesis as in G. lanceolata. Moreover, the

foveolae occur in rows spirally arranged in clockwise fashion, each

floret being subtended by a foveola, and consequently also arranged

spirally in rows. Moreover, the fleshy processes protruding upward

from the four sides of a foveola may become remarkably elongated,

and attenuated to a sharply acute or acuminate apex. In this latter

condition they appear very much as if they were reduced or degener-

ate paleae.

COROLLA

The disk florets, as has been stated previously, are always

tubular and perfect, and the ray florets when present are always

ligulate and fertile. The corollas are always glabrous.

RAY FLORET

The number of ray florets varies in each species, but for each

species averages a more or less constant number. For example, in

G. tenella the number of ray florets is ordinarily 15-18, whereas in

G. maritima and G. squarrosa it is usually 25-35. The ligule generally

is quite conspicuous and usually oblong or elliptic-oblong, or it may

be narrowly linear-oblong to broadly oblong-spatulate. It varies

from 5-7 mm. long in G. Robinsonii to 17-22 mm. long in G. Nelsonii,

and in G. Blakei up to 25 mm. long. It is quite slender, only 1.5-2.5

mm. broad, with usually revolute margins, as in G. squarrosa, or

relatively broad, 4-5 mm., with non-revolute margins, as in G.

stricta var. macrophylla. Moreover, the ligule is obtuse or acutish,

STUDIES OF THE AMERICAN FLORA — I 435

obscurely 3-lobed or practically subentire at the apex and narrowed

below into a slender, straight tube 3.5-7 mm. long. The color of

the ligule is usually quite characteristic for the different species.

Some species, G. squarrosa and G. nana, are pale or lemon-yellow-

colored. Others, like G. perennis, G. humilis, G. stricta and varieties,

are bright to orange-yellow, while G. grandiflora possesses orange-

colored ligules. Those species inhabiting salt marshes, estuaries,

and found in general near the coast develop a deeper, brighter orange-

yellow than the ones inland. This deeper pigmentation may be

associated with a higher salt content of the soil resulting in a denser

protoplasm associated with higher osmotic pressure, or there may

be some relation between the deeper pigmentation and the ability to

absorb certain light rays which occur in coastal habitats.

DISK FLORET

The disk corollas are always some shade of yellow. They all are

tubular, slender, and gradually or abruptly constricted about half-

way down into a slender, straight tube. The lobes of the 5-toothed

limb are short, 0.5-1.2 mm. long, lanceolate, oblong to ovate-lanceo-

late, acute or acutish. The veins are sometimes more prominent in

certain species than in others, where, for example, as in G. grandi-

flora the veins are rather conspicuously brownish contrasted with

the yellow color on the other parts.

The corolla is always 5-nerved and presents the usual condition

found in the Compositae. Miss Koch,1 in her anatomical studies

concerning the composite flower, showed that the floral anatomy of

the floret in Grindelia is very similar to that found typically in the

Heliantheae, and not at all like those studied in the Astereae.

The disk corolla is too homogeneous and stereotyped in plan to

offer any practical service in the differentiation of species or varie-

ties, although the author has attempted many times to find characters

which might be of use to segregate his entities as Wiegand did for the

allies of Aster lateriflorus and A. paniculatus.

ANDROECIUM

The androecium of Grindelia is yellowish throughout and consists

of five connate anthers with more or less elongate-oblong, oblong-

lanceolate, or ovate, acutish or obtuse, terminal appendages and

short, broadly deltoid or ovate bases. It is very homogeneous

1 M. F. Koch. Studies in the anatomy and morphology of the Composite

flower II. The corollas of the Heliantheae and Mutisieae. Amer. Journ. Bot.

17 : 1001-1002. 1930.

436 FIELD MUSEUM OF NATURAL HISTORY— BOTANY, VOL. XVII

throughout the genus, the only variations occurring in the relative

length and shape of the terminal appendage, but there is not suf-

ficient noticeable variation in this character to apply it to practical

use as a taxonomic criterion.

GYNOECIUM

The elongated arms of the style are quite different in the ray

and disk florets. In the ray florets the style has two erect or ascend-

ing, slender branches, the papillate stigmatic lines following along

the margins. The branches are glabrous or practically so on the

inner and outer faces. In the disk florets the style also bears two

erect to slightly spreading, straight branches, but each branch

terminates in an ovate-oblong to narrowly linear-lanceolate, acute

or obtuse appendage which is sparsely to densely hispidulous- or

hirsutulous-pubescent on the external dorsal surface and along the

margins, and glabrous on the internal ventral surface. The papillate

stigmatic lines extend along the margins of the stylar branches to

the base of the terminal appendage. The shape of the terminal

appendage has proved to be of some limited use as a taxonomic

criterion. Some species, G. camporum, G. procera, and G. nana,

possess a relatively elongated, narrowly linear-lanceolate, acute

or acuminate terminal appendage, in contrast to others, such as

G. rubricaulis var. robusta, G. squarrosa, G. maritime,, G. texana,

G. glutinosa, G. scabra var. neomexicana, and many other species have

relatively shorter and broader, oblong or oblong-lanceolate, obtuse

to acute terminal appendages. In some species the terminal append-

age is much more densely hirsutulous-pubescent on the outer dorsal

surface than in others.

PAPPUS

In many ways the pappus characters permit of ready aid in the

differentiations of species. The pappus in Grindelia is always

strongly caducous, and this property distinctly separates it from all

genera in the Astereae.

The pappus of a single floret consists of comparatively few awns

or bristles, usually 2-10. According to Cabrera,1 G. buphthalmoides

may have as many as 15 or even more bristles. A number of South

American species have 6-10 bristles to the florets, whereas the

Mexican species possess mostly 2-3, sometimes 4 on the ray and outer

disk florets. Some of the North American species, such as G. texana,

1 A.. Cabrera. Revision de las especies sudamericanas del genero "Grindelia."

Rev. Mus. La Plata 33 : 224. 1931.

STUDIES OF THE AMERICAN FLORA — I 437

G. microcephala var. adenodonta, G. lanceolata, G. nana and G. cam-

porum have only two bristles to the floret (G. camporum sometimes

has three), whereas others, G. subalpina and var. erecta, usually have

4-8. Grindelia rubricaulis var. latifolia of the Santa Barbara Islands

and the adjacent California coast may have as many as nine to the

floret. When two bristles are present, they are found at opposite

angles. When 5-9 or more are present, they are borne at and be-

tween the angles of the achene. Often two or more are found about

each angle.

The number of bristles to the floret, though often variable,

constitutes an important taxonomic character, since each species is

characterized by a certain definite number of bristles to the floret.

Great care is required, however, in determining this accurately.

The number of bristles per floret may be the same throughout the

head in some cases, whereas in many other instances the number on

the ray floret will exceed that on the disk floret. It is often possible

to find florets which have all the bristles in place, and to obtain a

certain count, but the bristles are so easily caducous that their nu-

merical determination is a very difficult problem. At the mere

touch of a dissection needle the bristles in a head which is in anthesis

readily disengage themselves from the achene. Furthermore, the

florets in a head are closely packed together and they as well as the

bristles are covered with resinous secretion; consequently, when one

attempts in a dissection to separate one floret from another or a

group of florets from another group, the bristles become easily

separated from the achene and fall off or become transferred by their

sticky surfaces to another floret. The author has often spent tedious

hours making dissections of many heads of a plant in his attempt to

be certain of the number of pappus bristles in one floret.

The best method found by which one may obtain more or less

certain numerical determinations of the number of bristles to a

floret in herbarium is to boil very young heads or those which, at

least, have not reached anthesis, and then make the dissections of

this material. In the young heads and in those which have not

reached anthesis the connection between the achene and the bristle

is still weakly intact.

Another method which is sometimes the only other alternative

and which with careful dissection often yields more or less certain

results is that of counting the total number of bristles in a head as

well as the total number of florets and dividing the former by the

latter. This latter is sometimes the only method available where

438 FIELD MUSEUM OF NATURAL HISTORY — BOTANY, VOL. XVII

all the heads are in a mature condition. It is the very caducous

nature of the pappus which has caused previous inaccuracies con-

cerning the number of pappus bristles.

The relative length of the pappus bristle compared with that of

the entire corolla of the disk floret is another character of taxonomic

significance. In some species, G. texana, G. lanceolata, G. scabra var.

neomexicana, G. inuloides var. adenodonta, G. littoralis, and G. campo-

rum, the bristle is as long or nearly as long as the length of the corolla

of the disk floret, whereas in G. integrifolia, G. integrifolia var.

virgata, G. rubricaulis var. elata, and a few others it is only one-half

to two-thirds the length of the corolla of the disk floret.

The relative width and thickness of the pappus bristle is a

character of importance. In all Mexican species, in G. microcephala

and varieties, and in G. scabra and var. neomexicana, the bristles are

very delicate and capillary-like or filiform in appearance. Although

they are not terete as true hairs, nevertheless they have the delicate,

filiform appearance of hair, and are of the same narrow breadth from

base to apex. All the other species have subpaleaceous, paleaceous,

or strongly paleaceous bristles. The subpaleaceous type occurs in

G. Havardii, G. arizonica, G. aphanactis, and others. They represent

transitions from the capillary-like or filiform examples of the Mexican

species to the more truly paleaceous types. As such transitions they

are on the border line and are sometimes difficult to place. However,

close comparison of these transitional, subpaleaceous forms with the

real capillary-like types reveals the finer distinctions between the

two. The strongly paleaceous types of bristles are best developed

in the maritime or coastal species, such as G. maritima, G. Blakei,

G. rubricaulis var. latifolia, and G. camporum.

A very interesting characteristic shown as regards the bristles

is the relative curvature or straightness exhibited in the dried con-

dition of herbarium material. All the Mexican species, and those

showing close relationship to the Mexican types, such as G. scabra

and var. neomexicana, G. microcephala and varieties, G. grandiflora,

G. oolepis, G. Havardii, G. arizonica and varieties, G. aphanactis, and

G. decumbens var. subincisa, possess fairly straight bristles when they

are dry. Others, including the large proportion of species of the

United States and Canada, possess bristles which are slightly to

strongly curved or twisted when dry. This is well shown in G. nana

and varieties, G. maritima, G. Blakei, G. rubricaulis var. robusta, G.

arenicola, and many others. Those species with strongly paleaceous

bristles or those which grow in proximity to the sea invariably ex-

STUDIES OF THE AMERICAN FLORA— I 439

hibit strongly curved or twisted bristles when dry, whereas the types

with delicate, capillary or filiform bristles invariably possess straight

bristles in the dried state.

One of the most important diagnostic characters for the dif-

ferentation of the species is the margin of the pappus. This is a

taxonomic criterion which has either been entirely neglected by

previous students or has been misused or erroneously treated.

Several species, among them G. texana, G. lanceolata, G. microcephala

and varieties, G. nana and varieties, G. camporum and varieties,

G. integrifolia, G. integrifolia var. virgata, G. Greenei, G. Greenmanii,

G. tenella, G. Robinsonii, G. oxylepis, G. inuloides, G. inuloides var.

glandulosa, G. oolepis, G. Nelsonii, G. Palmeri, and G. subdecurrens,

have the margins of the pappus bristles entire or subentire (only

remotely marked with a few short projections). Many others pos-

sess all degrees of serrulations.

For the sake of clarity, the author desires to explain at this time

the terms used in his monograph in describing the pappus margins.

The use of the terms "entire" and "subentire" is obvious and re-

quires no explanation. Three main types of dentation are found on

the margin of the pappus bristles in Grindelia: namely, first, the

serrulate type with relatively short or minute, deltoid, broadly or

ovate-lanceolate teeth, i.e. those about as long as broad or only 2 or

3 times longer than broad; second, the setulose type with relatively

elongated, linear or subulate teeth, i.e. those many times longer than

broad; and, third, the setulose-serrulate type — an intermediate con-

dition between the first and second in which there may be both types

(serrulate and setulose teeth) present or the teeth themselves may

be intermediate in length between the serrulate and the setulose type.

Among the three types themselves all gradations are found. There

are again three general categories under each of the previous types:

namely, first, remotely, second, moderately, and third, numerously

toothed. In the first the teeth are few and rather distant from one

another; the second type is intermediate in its degree of serration as

to the proximity and number of teeth; in the third the teeth are

abundant and numerous and relatively very close together. There-

fore, we may have either remotely, moderately, or numerously

serrulate margins, or remotely, moderately, or numerously setulose-

serrulate margins, or again, remotely, moderately or numerously

setulose margins. In G. subalpina and var. erecta, G. scabra and var.

neomexicana, and G. glutinosa the bristles are numerously setulose.

Between these types and those with entire margins are all stages.

440 FIELD MUSEUM OF NATURAL HISTORY — BOTANY, VOL. XVII

The bristles are opaque, cream-colored, or pale brown, smooth,

with shining surfaces, and are composed of many small, oblong cells.

They are mostly compressed-flattened in the upper half or through-

out or may be triquetrous below or in the lower half. They usually

taper gradually from base to apex to an acute or an acuminate tip,

but in some species, G. microcephala, G. microcephala var. adenodonta,

G. texana, and a few others, they are dilated or enlarged at the tip.

ACHENE

The achenes of Grindelia are usually narrowly or broadly oblong

and two to several times longer than broad. They vary from rela-

tively small, approximately 1.5-2 mm. long and 0.75-1.5 mm. broad,

to relatively large, 5-7.5 mm. long and 2.5-3.5 mm. broad. In G.

microcephala var. adenodonta they are depressed rhomboid or cuboid,

therefore as broad as or sometimes broader than long. The achenes

have a firm, glabrous, crustaceous or corky-thickened pericarp.

They vary in color from stramineous or pale brown to dark brown,

seal brown, or burnt sienna. The angles are often more thickened

than the rest of the achene. The angles may be rounded or sharply

acute, slightly winged, with a narrow, membranous portion, or not

at all winged.

The achenes of the ray florets ordinarily differ in cross section

from those of the disk florets. The latter usually being more com-

pressed, at least on four sides, and being more crowded, are usually

semi-flattened or compressed, therefore much broader than thick,

appearing 2-sided only, and oblong, elliptic-oblong, or fusiform in

cross section. In some species, G. squarrosa, G. oxylepis, G. micro-

cephala var. adenodonta, G. subdecurrens, G. inuloides, G. sublanugi-

nosa, and a few others, they are quadrangular or subquadrangular

and rectangular or squarish in cross section. The ray achenes may

be triquetrous, quadrangular, or subquadrangular, and are fre-

quently much broader or have a greater diameter than those of the

disk florets, particularly of the latter toward the center of the disk.

The triquetrous ray achenes occur in the majority of species and

are triangular or subtriangular in cross section, whereas the quad-

rangular or subquadrangular types which are squarish or rectangu-

lar in cross section are found mostly in many of the Mexican species,

and some of those obviously closely related to or derived from

the Mexican species, such as G. microcephala, and varieties, and

G. squarrosa. In the triquetrous ray achenes the two inner faces

are mostly straight, or sometimes slightly convex.

STUDIES OF THE AMERICAN FLORA — I 441

The surfaces of the achenes, particularly in the Mexican species,

show diversities which are characteristic and definite for the differ-

ent species, and thus serve as a very important diagnostic character

in specific differentiation. In G. nana and varieties, G. Havardii,

G. lanceolata, and G. littoralis, the surface of the faces is smooth and

mostly free from surface irregularities. In G. grandiflora the achenes

are conspicuously 9-10-costate or ribbed with slender, longitudinal

ribs alternating with rather deep sulcations. In G. aphanactis the

faces are as a rule deeply longitudinally ribbed and furrowed. In

other species the achenes are irregularly or slightly ribbed only

about the angles. Some of the maritime species, G. humilis and

G. Blakei, have the faces grooved irregularly and longitudinally, or

ribbed or thickened, the outer convex face in G. humilis often being

irregularly roughened or thickened. Some of the Mexican species,

G. tenella, G. oxylepis, G. sublanuginosa, and also the Texan G.

microcephala var. adenodonta, have the faces finely or coarsely and

irregularly rugose-wrinkled or convolutely rugose. In G. oxylepis

this convolute rugosity is very fine, while in G. sublanuginosa and

G. microcephala var. adenodonta it is much coarser and deeper.

Diversity in the type of apex about the areola of the achene at

maturity offers differentiating specific characters of fundamental,

deep-seated importance. Two main types are present, one in which

the apex of the achene is horizontally or obliquely truncate, the

other in which the areola is bordered at the angles with little, tooth-

like processes or knobs, or the angles and margins between the

angles may develop a shallow or prominulous, irregularly undulate

or thickened ridge or upraised rim. This rim may be more promi-

nently developed at one of the angles and thus appear like a knob

or toothlike process, or it may be developed at all the angles. The

knobbed or toothlike angles at the apex are commonly encountered

in many of the Pacific Coast species, such as G. camporum and

varieties, G. rubricaulis var. robusta, G. hirsutula and varieties, G.

procera, G. maritima, and G. humilis, and in some other species

like G. columbiana. Grindelia nana and varieties have 1-3 very

short projections or knobs, and G. texana and G. lanceolata of the

south-central United States and Texas have 1-2 very- short processes

at the apex or only a shallow, upraised rim. All the Mexican species

have horizontally truncate apices. Grindelia squarrosa, G. arizonica

and varieties, G. subalpina and var. erecta, G. microcephala and

varieties, G. perennis, G. inornata, and other interior land types

have the apices of the achenes horizontally or obliquely truncate.

442 FIELD MUSEUM OF NATURAL HISTORY — BOTANY, VOL. XVII

It should be emphasized, however, that as in Carex neither the

surface markings on the achene, nor the shape and size, nor the type

of apex can be used as critical characters unless the achenes of the

specimens examined are fully mature. Otherwise confusion is

likely to result. The achenes of G. microcephala var. adenodonta,

for example, are smooth to slightly rugose in the early stages, but at

maturity, and not until then, are conspicuously and convolutely

rugose- wrinkled. In some of the Mexican species like G. inuloides,

G. subdecurrens, G. Robinsonii, G. Palmeri, and G. Greenmanii the

faces are mostly smooth in the beginning and do not show signs

of shallow wrinkling or other rugosities until late maturity. Great

care must be taken in using achenial characters in the key to species,

varieties, and forms in the monograph.

A NEW GRINDELIA FROM MEXICO

Grindelia confusa Steyermark, sp. nov. — Herba perennis,

caulibus tenuibus, pluribus, adscendentibus, monocephalis vel pauci-

cephalis, dimidio superiore villosiusculis, 1.5-2 dm. altis; foliis per

omnes partes plerumque aequalibus firmiter membranaceis, salienter

vel spinulose dentatis, 1.5-3.3 cm. longis, 0.2-0.5 cm. latis, linear-

ibus, acutis, minute glandulosis; capitulis radiatis, 2.5-3 cm. latis;

disco campanulato-hemispherico, 0.7-0.8 cm. alto, 1.3-1.5 cm. lato;

involucre parce resinoso, involucri bracteis apice patentibus vel

paullo squarrosis, exterioribus et mediis 5-6 mm. longis, ca. 0.5

mm. latis, tenuibus, lineari-subulatis, glabris; ligulis 25-31, 8-9

mm. longis, ca. 2 mm. latis; achaenio immaturo laevi; aristis 3-5,

tenuibus, modice setulosis, 3-4 mm. longis, ca. % longitudinem

disci floris aequantibus. — Mexico: Chihuahua, swale, Namiquipa

plains, Harde LeSueur 1016, August 17, 1936 (type in Herbarium

of Field Museum).

Grindelia confusa from the Chihuahuan plains, with its spinulose-

dentate, linear leaves and moderately setulose awns, differs from all

the other Mexican species. It is most closely related to Grindelia

laciniata of Arizona and southeastern Utah, differing chiefly in its

involucral bracts, which are linear-subulate with free, spreading,

slender tips, in its more spreading-ascending stems which are pubes-

cent instead of glabrous, in its more setulose pappus awns, and in

its leaf margins, which are spinulose-dentate rather than pinnatifid.

It is somewhat of a surprise and rather disconcerting to encounter

such a distinct new species, not only because the author has already

monographed the genus, but also because the characters of this new

STUDIES OF THE AMERICAN FLORA — I 443

species spoil some of the key characters based upon geographical

distinctions and awn characters. It actually combines the habit,

pubescence, and involucral bracts of some of the Mexican species

with the leaves and pappus awns of some southwestern United

States species.

As this paper goes to press, it is somewhat unusual to record

another species of Selenia, S. grandis Martin, newly described from

the same county in Texas as that from which S. oinosepala has

been found. That the two species are totally distinct, however,

is borne out by the much smaller and entirely different colored

sepals of S. oinosepala and by the twice as large petals of S. grandis.

UNIVERSITY OF ILLINOIS-URBANA