Annals

of the

Missouri Botanical
Garden

Volume XLVIII
1961

Published quarterly at Galesburg, Illinois, by the Board of Trusi
Entered as second-class matter at the post-office at Galesburg, I

Annals

of the

Missouri Botanical Garden

A quarterly journal containing scientific contributions from the
Missouri Botanical Garden and the Henry Shaw School of Botany of
Washington University in affiliation with the Missouri Botanical

Information

February, May, September, and November. Four

TABLE OF CONTENTS

Morphology and Anatomy of the Saururaceae. I. Floral Anatomy and
Embryology M. V. S. Raju

Index of Orchid Names — 1960..

Preliminary Revision of the Genus Drymari

Sanango: New Amazonian Genus of Loganiaceae

George S. Bunting and Jai

Two New Members of the Genus Scolecopteris R. Bradley Ewart 275-

A Synopsis of Poinsettia (Euphorbiaceae)-
General Index to Volume XLVIII

Robert L. Dressier

MISSOURI BOTANICAL GARDEN

BOARD OF TRUSTEES

President

Robert Brookings Smith

Vice-President

Howard F. Baer

Henry Hitchcock

EX-OFFICIO

Raymond R. Tucker,

Mayor of the City of St. Louis

• — - , ..... . • .

Annals

of the

Missouri Botanical
Garden

ARY, 1961

-i f PiP.ama F IV Fasd . 4

Annals

Missouri Botanical Garden

A Quarterly Journal containing
Missouri Botanical Garden and the Henry Sh.iv School i

Contents o£ previ
Company.

FLORA OF PAN AM A

ROBERT E. WOODSON, Jr.

AND

ROBERT W. SCHERY

AND COLLABORATORS

PART IV
Fascicle 4

CHENOPODIACEAE (J. A. Duke)
AMARANTHACEAE (J. A. Duke)
NYCTAGINACEAE
BATIDACEAE

PHYTOLACCACEAE (K. Raeder)
AIZOACEAE (L. I. Nevling, Jr.)
PORTULACACEAE (L. I. Nevling, Jr
CARYOPHYLLACEAE (J. A. Duke)

Annals

OF THE

Missouri Botanical Garden
FEBRUARY, 19(51

FLORA OF PANAMA

Part IV. Fascicle 4*

CHENOPODIACEAE

By JAMES A. DUKE

Flowers perfect or unisexual, monoecious, polygamous or dioecious, sessile or
shortly pedicellate, often bracteate and bibracteolate, the bracts mostly herbaceous.
Perianth uniseriate, of (0-) 2-5 lobes, usually hypogynous, the lobes discrete or
basally connate, usually greenish. Corolla absent. Stamens as many as or fewer
than the sepals, the filaments mostly discrete, the anthers 2-4-locellate, usually
introrse and dorso-medially attached. Ovary superior, unilocular, uniovulate, the
ovule erect on a short funicle or pendulous from an elongate funicle; styles 1-3,
the stigmata capitate or elongate or the stigmata 2-5 and elongate. Fruit a
1 -seeded indehiscent or circumscissile utricle, the perianth often and the sepals
occasionally adherent to the seed; seeds erect, horizontal or inverted, lenticular,
cochlea te, subglobose or ellipsoid; embryos circular or hippocrepiform and more
or less encircling the endosperm or spirally coiled and nearly filling the seed. Herbs,
rarely shrubs or small trees, with simple alternate or opposite exstipulate leaves,
these often farinose or glandular. Flowers glomerulate in axillary or terminal
spikes, racemes, panicles or cymes, or rarely solitary and axillary, or embedded in
the strobiloid axis of the inflorescence.

Embracing about a hundred genera and more than a thousand species, the
chenopods are a rather cosmopolitan family containing many weeds and halophytes
in addition to a few vegetables such as the beet, spinach and swiss chard, the first
of which is sometimes grown around the Canal Zone. Three weedy species of
Chenopodium occur in Panama, one probably native, the other two known only
as weedy advents on San Jose Island.

1. CHENOPODIUM L.
Chenopodium L. Sp. PL 218. 1753.
Morocarpus Adans. Fam. 2:261. 1763.
Anserina Dumort. Fl. Belg. 21. 1827.
Teloxys Moq. in Ann. Sci. Nat. 2 T :289. 1834.

2 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Agathophytum Moq. loc. cit. 2 1 :291. 1834.

Roubieva Moq. loc. cit. 2 1 :292. 1834.

Oligandra Less, in Linnaea 9:199. 1834. non Oligandra Less. 1832.

i T. Ness, Gen. Fl. Germ. Dicot. 1: pi. 57. 183 5.
Ambrina Spach, Hist. Veg. 5:295. 1836.
Botrydium Spach, loc. cit. 5:298. 1836.
Lipandra Moq. Chenop. Enum. 19. 1840.
Gandriloa Steud. Nom. Bot. 2 1 :662. 1840.
Oligatitbera Endl Gen. 1377. 1841.

Oxybash Kar. & Kir. in Bull. Soc. Nat. Mosc. 1841 :738. 1841.
Orthospermum Opiz, Seznam 70. 1852.
Vulvar* Bubani, Fl. Pyren. 1:174. 1897.
Botrys Nieuwl. in Am. Midi. Nat. 3 :274. 1914.

Flowers perfect or rarely unisexual, sessile or subsessile, ebracteate. Sepals (3-)
5, hypogynous, free or basally united, herbaceous, subequal, often strongly 1-ribbed
and cucullate. Stamens 5 or fewer, occasionally varying in number in different
flowers of the same inflorescence, the flattened filaments free or basally connate,
the anthers mostly suborbicular, introrse, dorso-medially attached. Ovary sub-
globose, the stigmata 2 (-5), filiform or subulate, mostly sessile or subsessile. Fruit
an indehiscent utricle, ovoid to subglobose, the pericarp membranaceous to carnose,
free or adherent to the single seed; seeds mostly cochleate-lenticular, smooth to
roughened, vertically or horizontally oriented. Annual or perennial often strong-
scented herbs. Leaves alternate, entire to pinnatifid, the lowermost at least usually
petiolate, frequently glandular or farinose. Inflorescences of terminal or axillary
glomerules, the glomerules variously arranged.

Consisting of about a hundred weedy species, Chenopodium is a rather cosmo-
politan assemblage of meager economic importance. Various species are used in
Central America and elsewhere as potherbs, condiments and vermifuges. Three
easily differentiated species are presently known to occur in Panama, but several
others occur in Central America north of Panama.

linear to oblong, entire; sepals

Chenopodium murale L. Sp. Pi. 219. 1753.

triplex muralh Crantz, Inst. 1:206. 1766.
lenopodium guineense Jacq. Coll. 2:346. 1788.
levnpodium cartbagencnse Zucca^ni, in Roem. Coll. 133. 1:
wnopodtum murale p albescens Moq. Chenopod. Enum. 32.
icnopodtum murale 7 carthageneme Moq. loc. cit. 32. 1840

flora of Panama (Chenopodiaceae)

Chcnopodium murale sphsidentatum Murr, in Magyar Bot. Lap. 2:11. 190).

Erect or ascending annuals to as much as 6 dm. tall, the branches ascending,
often deeply sulcate, occasionally farinose, especially toward the extremities.
Leaves somewhat lustrous and occasionally bullate above, glabrous to rather
densely farinose below, deltoid to rhomb-ovate, marginally irregularly and .\cuteh
dentate, apically acuminate to attenuate and often mucronate, basally acute to
truncate, 2.5-8 cm. long, 2-5 cm. broad, the petioles mostly 3-70 mm. long,
sometimes quite as long as the blades. Flowers irregularly disposed in glomerules,
the glomerules sessile along the dichotomously branching rhachises, these axillary
and terminal, distally farinose. Sepals 5, subequal, ovate, obtuse, cucullate,
1-ribbed, slightly if at all carinatc, farinose, basally connate, 1-1.5 mm. long;
stamens 5, caducous, the filaments flattened, the anthers orbicular, exserted;
ovary subglobose, the style quite short, the 2 (-3) stigmata ca. twice as long,
spreading, irregularly glandular. Utricle partially enclosed by the sepals, the seed
lenticular-cochleate, minutely tuberculate, dull reddish-brown, 1.2-1.5 mm. broad,
horizontal, the pericarp adherent.

Presumably a native of Europe, Asia and Africa, now widely advent ive in
America as far north as Canada; in Panama known only as a lawn weed in San
Jose Island.

Champodium
( 'htuopodium

Nov. 26:1
Chcnopodium
Chenopodium

64. 1943.

petiolare var. Uptophylloides Murr. in Bull. Herb. Boiss. 2':994. 1904.

bdfang & Aellen, in Feddo, Rep. Spec.
34. 1929.

pratericola var. leptopbylloiJes Aellen, in Ostenia 100. 1933.
Pratericola ssp. pratericola var. hptopbylloides Aellen, in Am. Midi. Nat. 3 0:

ascending ai

muals to as much as 3 dm. h

iigh, the br

inches ascending,

usually mull

:isulcate, bi<

:olored, bullate or farinose,

especially

distally. Leaves

yellowish-gre
late, basally

en above, whitish and densely farinose
itire or rarely subhastatulate, apically ob

below, narrowly oblong to
-10 mm. broad, the petioles

2-10 mm. long. Flowers sessile in dense glomerules, the glomerules in panicled
interrupted spikes, the rhachises densely farinose. Sepals 5, subequal, ovate, sub-
acute, cucullate, strongly carinate, farinose, basally connate, ca. 1 mm. long;
stamens 5, the flattened filaments included, the anthers orbicular; ovary subglobose,
the 2 stigmata sessile or subsessile, spreading, irregularly glandular. Utricle com-
pletely embraced by the sepals, the seed lenticular-cochleate, smooth, lustrous,
reddish-brown, 1-1.2 mm. broad, horizontal, loosely embraced by the pericarp.

Western United States to Mexico; Argentina; probably adventive on San Jose
Island as it is in the eastern United States.

(345)

THE MISSOURI BOTANICAL GARDEN

046)

FLORA OF PANAMA (ChetlOpodh

corral, San Jose Island, Perlas Archipelago, Gulf of Panama, ca.

: of this species as a weed on San Jose Island would tend to sub-
stantiate I. M. Johnston's claim (in Sargentia 8:30. 1949) that the weeds were
introduced from Texas in unclean hay. The two related taxa [C. pratericola
ssp. pratericola var. thellungianum Aellen, with linear leaves, and C. pratericola ssp.
desiccatum (A. Nelson) Aellen, a semi-prostrate form] were unknown in Texas at
the time of Aellen's monograph (in Fedde, Rep. Spec. Nov. 26: 1 19. 1929) . It must
be noted that Aellen (in Am. Midi. Nat. 30:64. 1943) has reduced C. dessi-
catum to a subspecies of C. pratericola, a very peculiar mishap since C.
desiccatum antedates C pratericola by 10 years and seems not to be homonymous.
If they do indeed prove to be mere subspecies of the same species, several new
combinations will be required, but I will leave these for the monographer.

3. Chenopodium ambrosioides L. Sp. PL 219. 1753.
Chenopodium anthelminticum L. Sp. PL 220. 1753.
Atriplex ambrosioides Crantz, Inst. 1:207. 1766.
Atriplex anthelmintica Crantz, loc. cit. 1:207. 1766.
Chenopodium suffruticosum Willd. Enum. 290. 1809.
Chenopodium spathulatum Sicber, Fl. Martin. 92. 1825.
Chenopodium santa-maria Veil. FL Flum. 126. 1825.

Chenopodium chilense Schrad. Ind. Sem. Hort. Gotting. 1832:2. 1832. non Pers. 1805.
Orthosporum ambrosioides Kostel. Allg. Med.-Pharm. FL 1433. 1835.
Orthosporum suffrutim urn Kostol. loc. cit. 1433. 1835.
\mhrina chil >, k Spach. Hi«. Veg. 5:297. 1836.
Ambrina ambrosioides Spach, loc. cit. 5:297. 1836.
Ambrina anthelmintica Spach, loc. cit. 5:298. 1836.
Roubieva anthelmintica H. & A. Bot. Beech. Voy. 387. 1840.

:37. 1856.

'icum A. Gray, Manual 2:364. 1856.
Ambrina andicola Phil, in Anal. Univ. Sant. 91:442. 1895.
Chenopodium querciforme Murr, in Mag. Bot. Lap. 3:37. 1904.
Blitum ambrosioides G. Beck, in Reichenb. Ic. FL Germ. 24:118. 1908.
Botrys ambrosioides Nieuwl. in Am. Midi. Nat. i:275. 1914.
Botrys anthelmintica Nieuwl. loc. cit. 3:275. 1914.
Chenopodium vagans Standi, in N. Am. FL 21:26. 1916.
Also numerous other infraspecific names; see Aellen (in Am. Midi. Nat. 30:51. 1943).

Erect or ascending ill-scented perennials to as much as 15 dm. high, the
branches ascending, usually multisulcate, bicolored, often lignescent, glabrous to
tomentulose or villose about the inflorescence. Leaves yellowish green, puberulent,
villosulous or glabrate, glandular, with amber-colored secretions, lanceolate to
rhomb-elliptic, entire to coarsely and irregularly sinuate-dentate or sinuate-
pinnatifid, apically acute, basally attenuate, 2-12.5 cm. long, 0.5-5.5 cm. broad.
Flowers sessile, usually in dense glomerules, these contiguous or not and
paniculately disposed, interspersed with reduced or rather large leaves, the rhachises
glandular and villosulous to glabrate. Sepals (3-) 5, subequal, narrowly ovate,
cucullate, very slightly carinate if at all, connate for about one third their length,

047)

6 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ca. 1 mm. long; stamens (3-) 5, the flattened filaments about as long as the sepals,
the anthers orbicular, slightly exserted; ovary subglobose, the sessile or subsessile
stigmata spreading, irregularly glandular. Utricle completely embraced by the
sepals, the seed lenticular-cochleate, smooth, lustrous, reddish-brown, 0.6-0.8 mm.
broad, horizontal and vertical, loosely embraced by the pericarp.

A cosmopolitan weed, perhaps indigenous in Mexico and Central America.

chiriqui: Alto Lino, 4200 ft., Bro. Maurice 882. Panama: Juan Diaz, Standley
30529-

Polymorphically perplexing, this species, or some of its subspecies, varieties or
forms, is quite cosmopolitan. Colloquial names applied to the wormseed in Central
America are almost as numerous as the Latin names. In Panama only the name
"paico" has been encountered. The plant is used medicinally as a vermifuge, a
poultice and reputedly even as a soporific; it is also used as a culinary flavoring.

AMARANTHACEAE

By JAMES A. DUKE

Flowers perfect or unisexual, monoecious, polygamous or dioecious, mostly
pentamerous, usually bracteate and bibracteolate, sessile or short-pedicellate. Calyx
uniseriate or biseriate, commonly hypogynous; sepals (2-) 5, discrete or partially
connate, scariose, whitish or variously colored. Corolla absent. Stamens (2-) 5,
rarely more, hypogynous or perigynous, the filaments discrete or flattened and
united below into a tube, the tube often with filamentous pseudostaminodia
inserted between the filaments; anthers 2- to 4-locellate, usually introrse and
dorso-medially attached. Ovary superior, unilocular, uniovulate (multiovulate in
Pleuropetalum, Celosia and allied genera) ; ovules campylotropous on mostly
elongate flattened funicles; styles 1-8; stigmata capitate or filiform. Fruit a
1 -seeded utricle or rarely a several-seeded capsule, often circumscissile; seeds usually
cochlea te-orbiculate, the embryo excentric in a mealy endosperm. Herbs, shrubs,
small trees or clambering vines with alternate or opposite mostly entire exstipulate
leaves. Inflorescences of solitary flowers, spikes, glomes, glomerules, or thyrses
simply, racemosely or corymbosely disposed.

This largely tropical and subtropical family, of some 50 genera and 500 species,
is represented in Panama by eleven of the twelve genera thus far reported for
Central America. The twelfth genus, represented in Central America by Froe-
lichia interrupta, has not yet been reported between Guatemala and Colombia.
Many of the Panama species are cultivants or anthropochorous weeds of disturbed
habitats. Several species in several genera are employed as potherbs and species of
Amaranthus are a source of grain. Species of Iresine, Celosia, Amaranthus and
Gomphrena are planted as ornamentals in Central America and other places.

(348)

flora of Panama ( Am araii thaceae)

(in Panama) 2. CtLOMA

reddish brown, I

dd. Flowers perfect; filaments basally connate; seeds ai

. Leaves opposite; fruit a 1 -seeded indehiscent utricle; si

e. Inflorescences of elongate sj

4-loceIlate: style 1, with a,

>m>„,I> s

5. glomerules or short congested
spp.); anthers 2-loceItate; ityks

nflorescences of spikes in mucr

■: -
-3, if 1, bilabiate; erect herbs, s
. Flowers perfect; stigma 1,

broader than tin

tube between b
or exceeding the

^han'These'al

B

(349)

8 ANNALS OF THE

C35o;

flora of Panama (Amaranthaceae) 9

1. PLEUROPETALUM Hook. f.
Pleuropetalum Hook. f. in Proc. Linn. Soc. 1:278, and in Lond. Jour. Bot. 5:

l Hook. Gen. 3:24. 1880.

Flowers perfect, subsessile to pedicellate, unibracteate and bibracteolate. Sepals
5, discrete, hypogynous, concave, subequal, many-ribbed. Petals absent. Stamens
5-8, hypogynous, connate about half their length forming an exappendiculate
tube shorter than the ovary; filaments flattened; anthers 4-locellate, introrse,
medially attached. Ovary subglobose, 1-locular, multi-ovulate; ovules cam-
pylotropous; funicles flattened or filiform, exarillate, with free central placenta-
tion; style 1; stigmata mostly 3-6, at anthesis scarcely distinguishable, ultimately
reflexed. Fruits at first baccate, later developing into circumscissile or irregularly
dehiscent capsules; seeds cochleate-orbiculate, black, lustrous, reticulate. Glabra te
sparingly branched erect suffruticose herbs or shrubs. Leaves alternate, entire,
petiolate, lanceolate to rhombic-ovate, glabrous to minutely strigillose, deciduous
Inflorescences of terminal and axillary racemes, panicles or corymbs, with or
without foliar leaves.

Inhabiting moist forests of low elevations, thil genus, of perhaps three species,
ranges from Jalisco to Peru and the Galapagos Islands. The type species, P. daruinii
Hook, f., endemic to the Galapagos Islands, was reluctantly assigned to the
Portulacaceae by Bentham & Hooker (Gen. 1:157. 1862.). The two other species
ally unimportant genus are both found in Panama.

1. Pleuropetalum sprucei (Hook, f.) Standi, in N. Am. Fl. 21:96. 1917.

Melanocarpum sprucei Hook. f. in Benth. & Hook. Gen. 3:24. 1880.
Pleuropetalum costaricense Hort. Kew, ex Hemsl. Biol. Centr. Am. Bot. 3:12. 1882.
Pleuropetalum tucurriquense Donn. Smith, in Bot. Gaz. 61:387. 1916.
Pleuropetalum calospermum Standi, in Jour. Wash. Acad. Sci. 13:368. 1923.

Erect sparingly branched subglabrous suffruticose herbs or shrubs to 3 m. high.
Leaves glabrous to minutely strigillose on the veins below, broadly lanceolate to
rhombic-ovate, often falcate, apically attenuate, basally acute to subrounded, 3-15
cm. long, 1-6 cm. broad; petioles 5-20 mm. long. Inflorescences of mostly
terminal corymbose racemes, the rhachises usually glandular. Flowers perfect, on
pedicels 1-10 mm. long; bracts and bracteoles subequal, ovate to orbicular, carinate,
0.5-1 mm. long; sepals 5, subequal, ovate, concave, glabratc or scurfy, rounded to
subacute, ultimately spreading, 2-4 mm. long, 1-2 mm. broad, with 7-13 (-17)
ribs; stamens 5-8; filaments 2-3 mm. long; filament tube 1-2 mm. long; anthers
oblong; ovary globose; style 1, about 0.5 mm. long; stigmata mostly 3-4, longer
than the style, ultimately reflexed. Fruit an irregularly dehiscent capsule exceeding

10 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the sepals, 4-7 mm. long, 3-6 mm. broad; seeds cochleate-orbiculate, black,
lustrous, reticulate, 1.5-2.2 mm. long, on filiform or flattened funicles becoming
1-3 mm. long.

bocas DEL toko: vicinity of Chiriqui Lagoon, von WeJel 1028. chiriqui: Bajo
Mono, Boquete District, 4500 ft., Davidson 481; R. Chiriqui Viejo Valley, on island near
River, White 15a; valley of the upper Rio Chiriqui Viejo, 1300-1900 m., White 6
White 12.

This woodland species ranges from Jalisco to Peru. The Jalisco specimen
(Mexia i<Sgg) is atypical in having a circumscissile capsule, in that respect resem-
bling the Galapagos species P. daruinii Hook. f. Standley (in Jour. Wash. Acad.
13:368. 1923.) describes P. calospcrmum from Salvador, and ascribes to it larger
seeds, sepals and capsules than occur in P. sprucei. The larger seed size is not
always correlated with the larger sepals. Variations in sepal and capsule size are
frequently great enough in single specimens to embrace the descriptions of both
species, so they are here concluded to be conspecific.

2. Pleuropetalum pleiogynum (O. Ktze.) Standi, in Jour. Wash. Acad. 13:
369. 1923.

Celosia pleiogyna O. Ktze. Rev. Gen. 541. 1891.

Pleuropetalum staitdleyi Suessenguth, in Rep. Spec. Nov. 44:41. 1938.

Erect sparingly branched subglabrous suff ruticose herbs or shrubs to 3 m. high.
Leaves glabrous to minutely strigillose on the veins below, lanceolate to rhombic-
ovate, often falcate, apically attenuate, basally acute to subrounded, 3-15 cm.
long, 1-5 cm. broad; petioles 5-2 5 mm. long. Inflorescences of mostly terminal
racemes or panicles, the rhachises usually glandular. Flowers perfect, on pedicels
1-5 mm. long; bracts and bracteoles subequal, broadly ovate, carinate, 0.5-2 mm.
long; sepals 5 (-6), subequal, ovate, concave, glabrate or scurfy, rounded to
acute, with 19-27 ribs, 2.5-6 mm. long, 1-4 mm. broad; stamens 6-8; filaments
2-3 mm. long; filament tube 1-2 mm. long; anthers oblong; ovary globose to
ellipsoid; styles 1, sulcate, about 0.5 mm. long; stigmata 5-6, rarely more or less,
usually longer than the style, ultimately reflexed. Fruit a circumscissile capsule
equaling or exceeding the sepals, red becoming black, 5-7 mm. long, 4-7 mm.
broad; seeds cochleate-orbiculate, black, lustrous, reticulate, 1-1.2 (-1.5)
mm. broad, on filiform or flattened funicles becoming 1-3 mm. long.

canal zone: vicinity of Las Cruces, 26-40 m., Seibert 578; drowned forest of Que-
brada Culebra, Dodge <5 Allen 17051; forests along the banks of Quebrada Fea, Q. Pura
and Canon of R. Chagres, Dodge 2f Allen 17405; drowned forest of Quebrada Bonita and
along its branches 7,,-su m., Steye, m* > forest of Quebrada Ancha,

70 m Steyerr.unk * Allen ,. «. CHH | ,...!,,, ,!r.

1000 ft., Allen 5054. darien: Rio Sabana, Leopold 106; Cana and vicinitv. WHIimiU
727. Panama: Cerro Campana, alt. 800 m., Allen 4024.

This species is apparently confined to Costa Rica, the type locality, and Panama.
It occurs in rather deep woods, especially along streams in ravines and drowned

FLORA OF PANAMA (An

2. CELOSIA I..

Cfxosia L. Sp. PI. 205. 1753.

Lcstibudesia Thouars,

sionally sterile, subsessile or short pedicel-
late, unibracteate and bibracteolate. Sepals
5, discrete, hypogynous, concave, variously

Petals absent,
mate less than
appendiculate

: strongly ribl
Stamens 5, hypogynous,
half their length forming
or exappendiculate tube shorter than the
ovary; filaments flattened; anthers 4-
locellate, introrse, medially attached.
Ovary subglobose to ellipsoid, 1-locular,
multi-ovulate; ovules campy lotropous;
funicles flattened, exarillate, with free
central placentation; styles 1; stigmata 2-
3. Fruit a 2- to many-seeded circumscissile
capsule; seeds cochleate-orbiculate, dark
reddish brown to black, reticulate. Gla-
brate or pubescent herbaceous of suffruti-
cose perennials or annuals. Leaves alternate,
entire, subsessile to petiolate, glabrous or
pubescent, lanceolate to rhombic-elliptic,
deciduous. Inflorescences of terminal and
axillary panicles or spikes, with or without
foliar leaves.

A large genus of some fifty species, this
seems to have reached its best development
in subtropical regions of America, Africa
and Asia. Only one species, the cultivated
cockscomb, is so far reported from Panama.

12 ANNALS OF THE MISSOURI BOTANICAL GARDEN

1. Celosia argentea L. Sp. PI. 205. 1753.
Celosia cristata L. loc. cit. 205. 1753.
Celosia margaritacea L. Sp. PI. ed. 2:297. 1763.
Celosia coccinea L. loc. cit. 297. 1763.
Celosia pyramidalis Burm. Fl. Ind. 65. 1768.

Lmdica Retz. Obs. Bot. 3:27. 1783.
Celosia pallida Salisb. Prodr. 145. 1796.
Celosia htittonii Mast, in Card. Chron. 215. 1 872.
Amaranthus purpureus Nieuwl. in Amer. Midi. Nat. 3 :279. 1914.

Erect glabrous simple or much branched annuals to 1 m. high. Leaves
glabrous, linear-lanceolate to rhombic or ovate, apically acuminate to acute, 3-12
cm. long, 0.5-6 cm. broad; petioles 1-30 mm. long. Inflorescence of simple (quite
complex, often fascia ted, in cultivated varieties) pedunculate cylindric spikes 2-20
cm. long, 1-2 cm. broad. Flowers perfect, the uppermost occasionally sterile,
sessile; bracts and bracteoles subequal, ovate, mucronate, 2-7 mm. long; sepals 5,
subequal, ovate, concave, mucronate, white to pinkish (variously colored in
cultivated varieties), erect in fruit, 6-10 mm. long; stamens 5, 3-5 mm. long, the
tube shorter than the free portions of the filaments; pseudostaminodia minute and
deltoid or absent; anthers oblong; ovary ellipsoid; style 1, 3-6 mm. long, usually
exceeding the sepals; stigmata 2 (-3), minute. Fruit a circumscissile capsule,
shorter than the calyx, 3-4 mm. long; seeds (1-) 3-6 (-9), cochleate-orbiculate,
dark reddish brown, about 1.5 mm. broad.

Cultivated in many regions, possibly natives of Africa, the cockscombs are
reported by Standley (in Contr. U. S. Nat. Herb. 27:172. 1928) to be cultivated
in gardens in Panama, where they are called abanico. The cultivants often escape
and morphologically approach the natural variety. Backer (in Fl. Mai. 4 2 :74.
1949.) notes that in some escapes, one branch may bear the mark of cultivation
with another resembling the wild form. He informs us that the seeds are used by
the Chinese for poultices and for adorning cakes, while the leaves furnish an
inferior vegetable. In India, where it is used medicinally and as a vegetable, it
often appears spontaneously in paddy and ragi fields. Among vernacular names
reported from Mexico and Central America are mono, san jose, cresta de gallo, amor
seco, mano de leon, flor de mano, amaranto and boria.

3. AMARANTHUS L.
Amaranthus L. Sp. PI. 989. 1753.
B//7o« Adans. Fam. PI. 2:506. 1763.

Koemeria Moench, Meth. 341. 1794.

Glomeraria Cav. Descr. 319. 1803.

DimeiandraWzi. Neogen. 2. 1825.

Scleropus Schrad Ind. Sem. Hort. Gott. 1835.; in Linnaea 11: Litt.-ber. 89. 1837.

Dimetanthus Raf. Fl. Tell. 3:41. 1837.

Amblogyna Raf. loc. cit. 42. 1837.

Euxolus Raf. loc. cit. 42. 1837.

Pentrius Raf. loc. cit. 42. 1837.

flora of Panama ( Amaranthaceae)

Mengea Schauer, in Meyer, Nov. Acta Acad. Leop. 19: Supply. 1 :405. 1843.
Pyxidium Moq. in DC. Prodr. 13 2 :262. 1849.
Sarratia Moq. loc. cit. 262. 1849.
Gall/aria Bubani, Fl. Pyren. 1:184. 1897.

Flowers monoecious, dioecious or polygamous, sessile or subsessile, unibracteate
and bibracteolate. Sepals 3-5, rarely 1, discrete, hypogynous, concave, occasionally
basally clawed, equal or subequal, membranaceous, erect and persistent in fruit.
Stamens 3-5, rarely 1, hypogynous, discrete; filaments filiform; anthers 4-locellate,
introrse, medially attached, oblong. Ovary ovoid to lenticular, 1-locular, 1-
ovulate, the ovule campylotropous on a short exarillate funicle, the placentation
basal; styles 1-3; stigmata 2-3, usually exceeding the styles. Fruit a utricle,
seeds cochleate-orbiculate, smooth to minutely
itire to minutely crenulate, long-petiolate. Erect
or prostrate, glabrous to pubescent annual herbs. Inflorescences of terminal
and/or axillary spicate or paniculate thyrses, in Panama monoecious or polygamous
with female flowers basally, hermaphroditic flowers medially, and male flowers
distally, the latter tending to have longer, narrower sepals.

Widespread in both temperate and tropical regions throughout the world,
Amarantbtts consists of perhaps fifty species. Many of them arc inhabitants of
open disturbed areas where they may become rather annoying weeds. Some are
cultivated for their edible seeds and leaves, and others, with brightly colored
inflorescences, are planted as ornamentals. Sauer (Ann. Missouri Bot. Gard. 37:
561-632. 1950) reports on the detailed ethnological history of the grain-amaranths.
Seven species have been reported in Panama, but one, adventive on San Jose Island,
apparently no longer persists.

cc. Utricle compressed -ovoid; stylopodium 2-3-lobulate in cross section;

d. Plants unarmed; axillary inflorescences cylindric.
e. Thyrses less than 6 mm. broad; bracts of the pistillate flowers

mostly shorter than the sepals, some of the sepals spatulate 4. A. OUBlUs

mostly longer than the sepals, giving the spikes a bristly appear-

dd. Plants armed; axillary inflorescences globose 6. A. spinosus

In addition to the aforementioned, Johnston (in Sargentia 8:124. 1949) adds
that A. retroflexus L. was collected on San Jose Island in 1945 (Erlamon l88) ,
but was unsuccessfully searched for in 1946. I have not seen the specimen. The

055)

ANNALS OF THE MISSOURI BO TANK A

species is separated from the closely allied species A. hybr
of the pistillate flowers.

[acq. Coll. 2:325. 1788.
gracilis Desf . Tabl. Bot. 43. 1804.
Mbersia gracilis Webb. & Berth. Phyr. Canar. 3:287. 1836.
Euxolns cnudatus Moq. in DC. Prodr. 13 2 :274. 1849.

t. 1 J 8 :274. 1849.
F.uxohs caudatm 7 maximus Moq. loc. cit. 13 2 :274. 1849.

Rather delicate glabrous annuals to as much as 1 m. high, the branches ascend-
ing. Leaves glabrous, entire to minutely crenulate, deltoid- to rhombic-ovate,
apically emarginate to rounded and mucronate, basally truncate to subacute, 1.5-6
cm. long, 1-4 cm. broad, the petioles 0.5-4 cm. long. Inflorescences of thyrses
racemosely disposed, the terminal and axillary thyrses cylindric, mostly less than
7 mm. broad. Flowers polygamo-monoecious or monoecious; bracts and bracteoles
subequal, lanceolate to ovate, ca. 1 mm. long; sepals 3, linear-oblong to obovate,
rounded, mucronate, 1-1.5 mm. long, the midrib dark green, the margins scariose;
stamens 3, discrete, 1-1.5 mm. long; ovary compressed-globose; style 1, minute,
stigmata 3, longer than the style. Fruit an indehiscent, strongly rugose utricle,
1-1.5 mm. long, 1-1.5 broad; seeds cochleate-orbiculate, dark red to black, dull,
minutely pebbled, 1-1.2 mm. broad.

Presumably a native of the Old World tropics, this weedy species is adventive
around Balboa (fide Standley, in Contr. U. S. Nat. Herb. 27:173. 1928). It is
also adventive in the eastern United States. The names bledo and calalu are
applied to all the species of Panamanian Amarantims, any of which may serve
as potherbs.

2. Amaranthus californicus (Moq.) S. Wats. Bot. Calif. 2:42. 1880.

Mengea calif ornica Moq. in DC. Prodr. 13 2 :270. 1849.

Amaranthus carneus Greene, in Pittonia 2:105. 1890.

Amaranthus albomarginatus Uline & Bray, in Bot. Gaz. 19:318. 1 894.

Prostrate, radially spreading delicate annuals, the branches to as much as 5 dm.
long, often tinged with anthocyanins. Leaves glabrous, bullate, entire, often
white-margined, ovate to obovate and spatulate, apically emarginate to rounded
and mucronate, basally attenuate, 3-25 mm. long, 2-12 mm. broad, the petioles
3-25 mm. long. Flowers in axillary, often leafy, few-flowered clusters, with every
node often bearing an inflorescence, these ultimately shorter than the petioles.
Flowers monoecious; bracts and bracteoles lanceolate, apically aristate, 1.5-2.5 mm.
long; sepals 3 and quite unequal, or 1-2 in the pistillate flowers, linear to lanceolate,
acute to obtuse, 0.4-1.0 mm. long; sepals of the staminate flowers 3, subequal;
stamens 2, discrete, ca. 1 mm. long; ovary ellipsoid, the stigmata 3, ca. as long as
the ovary at anthesis. Fruit a utricle, circumscissile about the middle, rugose,
1-1.5 mm. long, 1-1.5 mm. broad; seeds cochleate-orbiculate, dark reddish-brown,
lustrous, minutely pebbled, 0.8-1.1 mm. broad.

FLORA OF PANAMA (A*

Although the cited specimen was labeled as A. albus L., the prostrate habit,
few stamens, and reduced sepals in the pistillate flowers clearly separate it from
that species. The small seeds and rugulose utricles also serve to separate it from
A. blitoides S. Wats., a very closely related prostrate species.

3. Amaranthus cruentus L. Syst. Veg. 10:1269. 1759.

Amaranthus paniculatus L. Sp. PI. 2:1406. 1763.

Amaranthus flavus L. loc. cit. 2:1406. 1763.

Amaranthus sanguineus L. loc. cit. 1407. 1763.

Amaranthus parisiensis Schkuhr, Handb. 3:249. 1808.

Amaranthus speciosus Sims, Bot. Mag. pi. 222J. 1821.

Amaranthus patiicuUtu* « /',•/•;.>-.,,■./. Moq. in DC. Prodr. 13 L ':257. 1849.

Amaranthus paniculatus /3 cruentus Moq. loc. cit. 257. 1849.

Amaranthus hybridus paniculatus Uline & Bray, in Mem. Torr. Club 5:145. 1894.

Amaranthus dussii Sprenger, in Bull. Soc. Tosc. Ort. 3 1 : 178. 1896.

Galliaria patula Bubani, Fl. Pyren. 1:187. 1 897.

Amaranthus hybridus subsp. cruentus Thcllung, Fl. Adv. Montp. 205. 191 2.

Coarse usually pubescent annuals to 2 m. high, simple or with ascending
branches. Leaves glabrous above, pubescent or glabrate below, entire to minutely
crenulate, ovate to rhombic, apically rounded to acute and mucronate, basally
acute to cuneate, 3-15 cm. long, 1-6 cm. broad; petioles 1-8 cm. long. Inflores-
cences of thyrses racemosely disposed, the terminal and axillary thyrses cylindric,
often drooping, 5-15 mm. broad. Flowers polygamo-monoecious or monoecious;
bracts and bracteoles subequal, lanceolate to ovate, 1.5-2.5 mm. long; sepals 5,
oblong to ovate, rounded to mucronate apically, 1.5-2 mm. long, the midribs dark
green, the margins scariose and tinted with anthocyanins; stamens 5, discrete, 1-2
mm. long; ovary compressed globose, capped by a circular stylopodium; stigmata 3
(-4) , longer than the stylopodium. Fruit a smooth to rugulose subglobose utricle,
circumscissile near the middle, about 2 mm. long, usually exceeding the sepals;
seeds cochleate-orbiculate, dark reddish brown to black, minutely reticulate,

This is the commonly cultivated grain amaranth of Central America. In
Panama it is called abanico rhino and calalu, and it is cultivated both as an
ornamental and as a potherb. In Mexico and Guatemala the seeds are used in making
sweetbreads and mushes. Standley & Steyermark (in Field Mus. Bot. 24*: 153.
1946) point out that A. caudatm L., A. leucosperma Wats, and A. cruentus L.
are best treated as one species "rather doubtfully distinct from A. hybridus"
Sauer (Ann. Missouri Bot. Gard. 37:561-632. 1950) on the other hand contends
that there are four major grain amaranth species in the Americas, A. leucocarpus
Wats., chiefly in Mexico; A. cruentus L., chiefly in Guatemala; A. caudatus L.,
chiefly in the Andes; and A. ednlfs Spegazzini, chiefly in Argentina. Of the
cultivated amaranths, whether they comprise several or one species, he cites only
A. cruentus from Panama.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

us Mart. PL Hort. Erlang. 197. 1814.

Hist. Amaranth. 21. 1790. not A. tristis L.

illd. Enum. Hort. Berol. suppl.: 64. 1813. hyponymn.
Amaranths t\ Moq. in DC. Prodr. 13 2 :260. 1849.

Amaranthus iristis ~, ficsuosus Moq. loc. cit. 260. 1849.
Amaranths t\ q. loc. cit. 260. 1849.

Amaranthus dubius B xanthostachys Thellung, in Asch. & Graebn. Syn. Mittel-Eur. Fl. 5:

.!;«.■ antb:r luluus C Hexuosus Thellung, loc. cit. 266. 1914.
Amaranthus dubius D lepiosiachys Thellung, loc. cit. 266. 1914.

Rather delicate mostly glabrous annuals to 1 m. high with ascending branches.
Leaves glabrous or glabrate, entire or minutely crenulate, deltoid- to rhombic-ovate,
apically rounded to acute, often mucronate, basally rounded to acute, 2-8 cm.
long, 1-6 cm. broad; petioles 1-6 cm. long. Inflorescences of thyrses racemosely
disposed, the terminal and axillary thyrses mostly cylindric and less than 6 mm.
broad. Flowers polygamo-monoecious or monoecious; bracts and bracteoles sub-
equal, lanceolate to ovate or obovate, acute to acuminate, 1.5-2 mm. long, the
midribs dark green, the margins scariose; stamens discrete, 1-3 mm. long; ovary
ovoid; styles (2-) 3, fimbrillate. Fruit a slightly rugose compressed ovoid utricle,
circumscissile about the middle, 1.5-2.5 mm. long; seeds cochleate-orbiculate,
reddish brown to black, reticulate, 0.8-1.1 mm. broad.

bocas del toro: Isla Colon, vicinity of Chiriqui Lagoon, von Wedcl 2854. canal
zone: near Gorgas Memorial Laboratory, viciniiv « f Mn it) res, (;. V. bid I IQ; in govern-
ment forest along Las Cruces Trail, Hunter & Allen 712; Bohio Soldado, Couell 235.
darien: vicinity of Boca de Cupe, ca. 40 m., Allen 874. Panama: Agricultural Exp.
Sta. at Matias Hernandez, Pittier 6839.

Reputedly the common weed amaranth of the Caribbean, this species extends
from Mexico through tropical South America and is adventive in Europe. In
Panama, where the leaves and young shoots are employed as potherbs, the names
bledo and calalu are applied. The Mayas call it xetz and chactex; Mexicans call
the amaranths quelite, a name derived from an Aztec word and used for any pot-
herb. The specific epithet is strongly suggestive of its specific status. It is weakly
separated from A. hybridvs L., from which it is frequently keyed by the length of
the sepals in proportion to the length of the utricles. In individuals of both species,
however, the utricles vary from shorter to longer than the sepals.

5. Amaranthus hybridus L. Sp. PL 990. 1753.

Amaranthus bypocondriacus L. Sp. PL 991. 1753.

Amaranthus beet:. A h. 25. 1790.

Amaranthus strictus Willd. loc. cit. 27. 1790.

Amaranthus laetus Willd. loc. cit. 28. 1790.

Amaranthus chlorostachys Willd. loc. cit. 34. 1790.

Amaranthus retroflexus var. hybridus A. Gray, Man. ed. 5:412. 1867.

Amaranthus retroflexus var. chlorostachys A. Gray, loc. cit. 412. 1867

Amaranthus chlorostachys var. hybridus S. Wats, in A. Gray, Man. ed. 6:428. 18 89.

Amaranthus hybridus var. bypocondriacus B. L. Robinson, in Rhodora 10:32. 1908.

G. Beck, in Reichenb. Tc. Fl. Germ. 24:175. 1908.
bnda Nieuwl. in Am. Midi. Nat. 3:278. 1914.

flora of Panama (Amaranthaccac) \j

18 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Rather coarse often pubescent annuals to 2.5 m. high with ascending branches.
Leaves pubescent or glabrous, entire to minutely crenulate, deltoid- to rhombic-
ovate, apically rounded to acute, often mucronate, basally acute to rounded, 2-15
cm. long, 1-7 cm. broad; petioles 1-8 cm. long. Inflorescences of congested
thyrses racemosely disposed, the terminal and axillary thyrses mostly cylindric, 6-12
mm. broad. Flowers polygamo-monoecious or monoecious; bracts and bracteoles
subequal, lanceolate to ovate, 2-4 mm. long, often conspicuously longer than the
flowers; sepals lanceolate to ovate or obovate, acute to acuminate, 1.5-2.5 mm.
long, the midribs dark green, the margins scariose, rarely tinged with anthocyanins;
stamens 5, discrete, 1-3 mm. long; ovary ovoid; styles (2-) 3, conical; stigmata
(2-) 3, fimbrillate. Fruit a slightly rugose compressed ovoid utricle, circum-
scissile near the middle, 2-2.5 mm. long; seeds cochlea te-orbiculate, reddish brown
to black, minutely reticulate, 1-1.3 mm. broad.

chiriqui: vicinity of Boquete, 1200-1500 m., Woodson & Schery ?22.
This weedy species, frequently forming large stands in old fields, is found
throughout temperate and tropical North and South America, and is adventive in
many parts of the Old World. Vernacular names reported in Mexico and Central
America are quintoniles, calcte, ses, huisquilete and quiec tes, bledo, xtez, quelite
and hutsquelite. The leaves and young shoots are often cooked as potherbs.

6. Amaranthus spinosus L. Sp. PI. 991. 1753.

Amaranthus diacanthus Raf. Fl. Ludov. 31. 1817.
Amaranthus caracasanus H.B.K. Nov. Gen. & Sp. 2:195. 1817.
Amaranthus sph:- . Flora 25: litt. 20. 1842.

Amaranthus spin | M -. c ; t . litt. 20. 1842.

Amaranthus spinosus f pygmaeus Hassk. Cat. PI. Bogor. 83. 1844.
Amaranthus spinosus (9 purpurascens Moq. in DC. Prodr. "
Amaranthus spinosus i. i

269. 1914.
Amaranthus spinosus 77 basiscissus Thellung, loc. cit. 269. 1914.
Amaranthus spin i lung, loc. cit. 269. 1914.

Galliaria spinosa Nieuwl. in Am. Midi. Nat. 3 -.278. 1914.

Rather coarse, mostly glabrous armed annuals to 2 m. high, usually with erect
branches. Leaves glabrous or pubescent below, narrowly rhombic-ovate, apically
rounded to acute, basally acute to cuneate, 1-12 cm. long, 0.5-5 cm. broad;
petioles 0.5-8 cm. long, with 2 thorns in their axils. Inflorescences of terminal
and axillary thyrses racemosely disposed, the lower axillary thyrses globose and up
to 15 mm. broad. Flowers polygamo-monoecious or monoecious, the bracts and
bracteoles rather variable, lanceolate to acicular, 1-5 mm. long, often resembling
the thorns, some conspicuously longer than the flowers; sepals 5, lanceolate to
oblong, 1-2.5 mm. long, the midribs dark-green, the margins scariose; stamens
discrete, 1-2 mm. long; ovary ovoid; styles 3, conical; stigmata 3, fimbrillate.
Fruit a slightly rugose compressed ovoid utricle, irregularly to regularly circum-
scissile, 1.5-2.5 mm. long; seeds cochleate-orbiculate, reddish brown to black,
minutely reticulate, 0.7-1.0 mm. broad.

(360)

flora of Panama ( Amaranthaceae) 19

canal zone: Monkey Hill and vicinity, Couell IQ. chiriqui: vicinity of Boqucte,
1200-1500 m., Woodson 8 Schery 723; vicinity of Puerto Armuelles, 0-75 m., Woodson
d Schery 833. col6n: Culebra, Couell 214. herrera: vicinity of Ocu, 100 m., Allen
4086.

Presumably a native of America, this anthropochorous species now has a
circumpolar distribution in temperate and tropical countries. In Panama it is
probably called bledo and calalii. Other Mexican and Central American names
are xtez, kix-xtez, huisqiielite, blcdo macho, ixtez, tsetz, labtzetz and nigna. In
spite of the rather formidable thorns this species is also utilized as a potherb.

4. CHAMISSOA HBK.

Chamissoa HBK. Nov. Gen. & Sp. 2:196. 1817.
Ko*<?r« Adans. Fam. Pi. 2:269. 1763. hyponym.

Flowers perfect, subsessile, with 1-3 bracts subtending 1-5 bracteate fertile
and 0-4 bracteate sterile flowers. Sepals 5, discrete, hypogynous, concave, greenish
white, with conspicuous midribs and membranaceous margins. Stamens 5,
hypogynous or perigynous, basally connate forming an exappendiculate tube
shorter than, but occasionally basally fused with, the ovary; free portions of the
filaments longer than the tube, usually somewhat flattened; anthers 4-locellate,
introrse, medially attached. Ovary globose or ovoid, 1-locular, 1-ovulate; ovules
campylotropous, the filiform arillate funicles with central basal placentae; styles
1, stigmata 2 (-3). Fruit a utricle circumscissile near the middle; seeds cochleate-
orbiculate, black. Erect or clambering glabrate or pubescent herbs or shrubs.
Leaves alternate, entire, petiolate or subsessile, linear-elliptic to broadly ovate,
attenuate, deciduous. Inflorescences of axillary and terminal glomerules, these
racemosely or paniculately disposed, with or without foliar leaves.

This genus of perhaps five species, is chiefly a South and Central American
genus with only one species ranging as far north as Mexico. The closely related
oriental genus Allmania has occasionally been treated as a subgenus but morpho-
logically seems to deserve generic status. Two species of Chamissoa occur in

1. Chamissoa altissima (Jacq.) HBK. Nov. Gen. & Sp. 2:197. 1817.
Celosia paniculata L.Sp. PI. ed. 2:298. 1762. not L. 1753.
Achyranthes altissima Jacq. Enum. PI. Carib. 17. 1762
Chamissoa macrocarpa HBK. Nov. Gen. & Sp. 2:197. 1817.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

flora of Panama ( Amaranthaceae) 21

Celosia tomentosa Willd. in Roem. & Schult. Syst. Veg. 5:531. 1819.

Achyranthes linkiana Roem & Schult. Syst. Veg. 5:545. 1819.

Cbamissoa martij Moq. in DC. Prodr. 13 2 :252. 1849.

Cbamissoa altissima laxiflora Moq. loc. cit. 251.1 849.

Cbamissoa altissima y densifiora Moq. loc. cit. 251. 1849.

Cbamissoa altissima var. glabrata Seub. in Mart. Fl. Bras. 5^242. 1875

Kokera paniculata O. Ktze. Rev. Gen. 542. 1891.

Cbamissoa altissima var. rubella Suesseng. in Rep. Spec. Nov. 35:306. 1934.

Cbamissoa altissima subsp. albo-grisea Suesseng. in Rep. Sp. Nov. 39:6. 1935.

Cbamissoa altissima var. gran J t ;. 6. 1935.

Cbamissoa altissima var. gra?u: SoMKng. loc. cit. 6. 1935.

Cbamissoa altissima var. densi- paniculata Suesseng. in Lilloa 4:129. 1939.

Clambering subglabrous shrubs or vines to 3 m. high. Leaves glabrous or
slightly pilose, broadly lanceolate to ovate, apically acute to acuminate and
mucronate, basally trvincate to acute, 3-14 cm. long, 1-6 cm. broad; petioles 0.5-
3.5 cm. long. Inflorescences of glomerules paniculately disposed, the pubescent
primary rhachis usually visible at intervals. Flowers perfect, some usually sterile,
subsessile; bracts 1-3, broadly deltoid, carinate, mucronate, with membranaceous
margins, 1—1.5 mm. long, 1—1.5 mm. broad; sepals 5, subequal, ovate, concave,

stamens 5; filaments 1.5-3 mm. long; filament tube 0.5-1 mm. long, often basally
adnate to the ovary; anthers globose to oblong; ovary at anthesis tympaniform to
globose, usually as broad as long and truncate or flanged at the summit; style 1,
0.5-1 mm. long, usually shorter than the stigmata; stigmata 2 (-3), erect at
anthesis, reflexed in fruit. Fruit a utricle, exserted, circumscissile near the middle,
compressed ovoid, the summit conspicuously operculate and truncate to incon-
spicuously operculate and rounded, 3-5 mm. long, 2-3.5 mm. broad, aril bivalvate,
becoming wrinkled, brownish white, completely investing the seeds; seeds smooth
but punctate, black, lustrous, 2-2.5 mm. broad.

bocas del toro: Changuinola Valley, Dunlop 309; vicinity of Chiriqui Lagoon, von
Wedel 1380 & 1799; Water Valley, vicinity of Chiriqui Lagoon, von Wedel 634A, 840 &
1641; Old Bank Island, vicinity of Chiriqui Lagoon, von Wedel 1896 & 2141; Garay
Creek, vicinity of Chiriqui Lagoon, von Wedel 2637. canal zone: between Summit and
Gamboa, Greenman tf Greenman 5258. cocle: vicinity of El Valle, 600-1000 m., Allen
1205; El Valle de Anton, trail near Finca Tomas Arias, 600 m., Allen 4230. darien:
Tucuti, Chepigana, Terry & Terry 1382. Panama: Taboga Island, on more or less forested
slopes, MacBride 2823.

This most commonly collected species in the genus ranges from Mexico to
Argentina, usually in thickets where it tends to clamber over adjacent plants. Of
limited occurrence in amazonian South America is a rather distinct type often
regarded as a separate species, C macrocarpa HBK., with rounded exoperculate,
frequently 3-stylate utricles much longer than the sepals. Kuntze (Rev. Gen.
542. 1891), perhaps the first lumper, regards C. macrocarpa as conspecific with
C. altissima. Annotation labels reveal that Schinz at one time considered C.
"macrocarpa as only a variety of C. altissima, and finally Suessenguth (in Rep. Sp.
Nov. 35:306. 1934) admits that a sharp distinction between the two is impossible.

(363)

22 ANNALS OF THE MISSOURI BOTANICAL GARDEN

2. Chamissoa maximiliana Mart, ex Moq. in DC. Prodr. 13 2 :251. 1849.

Cbamissoa celosioides Griseb. in Goett. Abh. 19:79. 1874.

Chamissoa maximiliana var. /3 procumbens Seub. in Mart. Fl. Bras. 5 1 :243. 1875.

Kokera celosioides (Griseb.) O. Ktze. Rev. Gen. 543. 1891.

Kokera acuminata (Mart.) O. Ktze. loc. cit. 543. 1891.

Cbamissoa maximiliana var. pubescens Chod. in Bull. Herb. Boiss. 7:63. 1899.

Chamissoa maximiliana i. celosioides (Griseb.) Suesseng. in Rep. Sp. Nov. 3 5:306. 1934.

Clambering subglabrous herbs, shrubs or vines to 2 m. high. Leaves glabrous
or slightly pilose, narrowly to broadly ovate, apically acuminate to mucronate,
basally acute to rounded, 2.5-10 cm. long, 1-4 cm. broad; petioles 1-2.5 cm.
long. Inflorescences of glomerules racemosely disposed, the pubescent primary
rhachis usually obscured. (In the typical form the glomerules are not all closely
approximated and the rhachis is visible at intervals.) Flowers perfect, some
usually sterile, or tending to monoecism, subsessile; bracts 1-3, narrowly deltoid
to lanceolate, carinate, mucronate, with membranaceous margins, 1.5-2 mm. long,
1-1.5 mm. broad; sepals 5, subequal, lance-oblong, concave, acuminate to mucro-
nate, obscurely 3-7 nerved, greenish white, 3-4 mm. long, 0.5-1 mm. broad;
stamens 5; filaments 1.5-3 mm. long; filament tube about 0.5 mm. long; ovary
at anthesis doleiform, 1-2 times as long as broad, attenuate to the base of the style,
often lobulate or flanged near the summit; style 1, 1-1.5 mm. long, longer than
the stigmata; stigmata 2, erect at anthesis, reflexed in fruit. Fruit a utricle, usually
included, circumscissile near the middle, globose to ovoid, the summit operculate
and usually truncate, 2-3.5 mm. long, about 2 mm. broad; aril minute; seeds
verrucose, black, often pleiochroistic, 1.5-2 mm. broad.

Chambers etal.6 77 .

Inhabiting thickets and forest clearings, this species ranges from Costa Rica
to Argentina. The Panama specimens seem referable to the forma celosioides
(Griseb.) Suesseng. which differs from the typical form in possessing compact
inflorescences. The type of C. maximiliana has lax inflorescences with subglobose
operculate utricles. Three other species with minute arils occur in South America.
C. acuminata Mart., with smooth seeds and petiolate, narrowly ovate falcate leaves,
occurs in Brazil and northern Argentina. C. blanchetii Moq., with smooth but
punctate seeds and subsessile linear-elliptic leaves, occurs in Brazil. C. brasiliana
(Moq.) R. E. Fr., native to Brazil and Argentina, differs from all other species
in having axillary verticillate glomerules.

5. ACHYRANTHES L.
Achyranthes L. Sp. PL 20. 1753.
Centrostacbys Wall, in Roxb. Fl. Ind. 2:497. 1824
Cadelaria Raf. Fl. Tell. 3:39. 1837.
Stachyarpagopbora Vaill. in Maza, Fl. Haban. 92. 1897.

Flowers perfect, bracteate and bibracteolate, subsessile in elongate spikes.

flora of Panama (Amaranthaceae) 23

24 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Sepals 4-5, discrete, hypogynous, concave, subequal, obscurely 3-5 nerved, occa-
sionally pungent. Stamens 2-5, the filaments flattened and united below forming
an appendiculate tube; anthers 4-locellate, introrse, medially attached. Ovary
obovoid, 1 Jocular, 1 -ovulate; ovules campy lotropous on elongate flattened
funicles; style 1, filiform; stigma 1, capitate. Fruit a membranaceous turbinate
indehiscent utricle; seeds cochleate-orbiculate, reddish brown. Glabrous to
pubescent erect or decumbent annual or perennial herbs. Leaves opposite, short-
petiolate, scantily appressed strigose to sericeous, deciduous. Inflorescences of
terminal and axillary spikes, the flowers deflexed in age, the spinose tips of the

This weedy genus, of about five species, mostly of the Old World, is repre-
sented by one species in Central America. The spines of the inflorescences are
often a source of discomfort. Two varieties of one species occur in Panama.

1. ACHYRANTHES ASPERSA L. Sp. Pi. 204. 1753.

Achyranthes aspersa p indica L. loc. ck. 204. 1753.
Achyranthes indica Mill. Gar. Diet. ed. 8: no. 2. 1768.
Achyranthes argaiti a Lara, l.ncvcl. 1:545. 1785.
Achyranthes obtnufnlia 1 urn. W. cit. 545. 1785.
\ch\nmth cuh R.,th, Cat. Bot. 1:39. 1797.
Cadelaria sicula Raf. Fl. Tell. 3 :39. 1837.
Cadelaria indica Raf. loc. cit. 39. 1837.
Stachyarpagophora aspersa Maza, Fl. Haban. 93. 1897.
Achyranthes aspersa simplex M 2:3 6. 1900.

Centrostachys indi •. Wash. Acad. 5:75. 1915.

Centrostachys ay it. 75. 1915.

Erect or procumbent pubescent annual
stems to 2 m. long. Leaves scantily to densely pubescent, ovate to orbicular,
apically acuminate to rounded, basally cuneate to rounded, 2-25 cm. long, 2-9
cm. broad; petioles 2-25 mm. long. Inflorescences of terminal and axillary
pedunculate spikes, 3-40 cm. long, 6-12 mm. broad, the flowers deflexed. Flowers
perfect; bracts and bracteoles subequal, 2-3.5 mm. long, ovate, spinescent, the
tips rarely arcuate but not uncinate; sepals 4-5, greenish white, subequal, hypogy-
nous, obscurely nerved, 4-7 mm. long; stamens 5, united below into a short tube;
pseudostaminodia lacerate, exceeded by the 4-locellate anthers; ovary obovoid;
style 1, filiform, at anthesis longer than the ovary; stigma 1, capitate. Fruit an
indehiscent turbinate utricle; seeds cochleate-orbiculate, reddish brown, about 1
mm. broad.

bocas del toro: H. Wedel 418. canal zone: Gatun, Cowell 300; near Frijoles,
Pittier 6834; around Culebra, 50-150 m., Pittier 4068. province unknown: Sutton
Hayes 320.

A rather ubiquitous pantropical weed, this species is perhaps adventive in
America. In India the dried inflorescences are used by orthodox Hindus in sacred
pyres. Standley & Steyermark (in Field Mus. Bot. 24 4 :145. 1946), who maintain
A. indica as specifically distinct from A. aspersa, say that A. indica is scarce in
Central America and known only from the Atlantic Coast, in contrast to the

(366)

flora of Panama (Amaranthaceae) 25

widely distributed A. aspersa. The variety indica seems to be rather regularly
distinct from the variety aspersa in having rounded leaves and shorter sepals. All
the cited specimens except Hayes 320 seem referable to the variety indica. Many
vernacular names are reported; zacpaiche and zorillo bianco in the Yucatan, abrojo
in Salvador, mozote in Salvador and Honduras, mozotillo and rabo de chanco in
Costa Rica, cola de arm ado, penegato, pije de gato and chile de perro in Guatemala
where the typical variety is apparently distinguished by the natives from the variety
indica to which the names pegapega, goincilla and mozotlexc are applied.

6. CYATHULA Lour.
Cyathula Lour. Fl. Cochinch. 1:101. 1790.
Polyscalis Wall. Cat. no. 6939. 183 2.

Flowers perfect, some sterile and highly modified, bracteate and bibracteolatc,
subsessile in short-pedunculate glomerules. Sepals 5, discrete, hypogynous, concave,
subequal, 1-3 ribbed, occasionally pungent. Stamens 5, the filaments flattened and
united below into an appendiculate tube; anthers 4-locellate, introrse, medially
attached. Ovary ovoid, 1-locular, 1 -ovulate; ovules campylotropous on elongate
flattened funicles; style 1, filiform; stigma 1, capitate. Fruit an ovoid indehiscent
utricle; seeds cochlea te-orbiculate, succineous. Pubescent branched ascending to
prostrate occasionally frutescent herbs. Leaves opposite, entire, petiolate, strigose,
deciduous. Inflorescences spicate, composed of terminal and axillary thyrses of
complex glomerules containing some highly modified flowers with some of the
bracts and sepals uncinate-spinescent.

Of this rather weedy genus of some twenty species, most of the species are
centered in Africa and the Orient. Only two occur in Central America, usually
in disturbed habitats of low elevations, where they may become annoying weeds.

a. Terminal spikes 2-10 times as long as the uppermost leaves, interrupted
over half their length; lower glomerules with mostly 3 fertile flowers

broad; sepals 2-3 mm. long; pseudostaminodia 3-toothed 1. C. prostrata

1. Cyathula prostrata (L.) Blume, Bijdr. Ned. Ind. 549. 1826.

Achyranthes prostrata L. Sp. PI. ed. 2:296. 1762.

Cyathula geniculate Lour. Fl. Cochinch. 1:102. 1790.

Desmochaeta prostrata DC. Cat. Hort. Monsp. 102. 1813.

Piipalia prostrata Mart. Nova Acta Acad. Leop.-Carol. 13 1 :321. 1826.

Erect or decumbent geniculate pubescent perennials to 1 m. high, often rooting
at the lower nodes. Leaves strigose on both surfaces, entire or undulate, ovate to
obovate, apically acute, basally acute to rounded, 2-6 cm. long, 1-4 cm. broad;
petioles 2-8 mm. long. Inflorescences of terminal and axillary spikes of glomerules,
the mature glomerule usually <

067)

ANNALS OF THE MISSOURI BOTANICAL GARDEN

lateral perfect flowers, each of the lateral
flowers subtended on each side by a
cluster of about 20 hooks representing 3
modified flowers, the hooks not conspicu-
ously longer than the glomerule, the
peduncle of the glomerule articulated near
the pubescent primary rhachis; spikes 5-
30 cm. long, 4-7 mm. broad, with only
the distal half continuous. Unmodified
flowers perfect; bracts and bracteoles

sepals 5, subequal, ovate,
throughout
their length, with 1-3 ribs strongly ex-
serted dorsally, 2-3 mm. long; stamens
5, 1—2 mm. long, united below to form
a tube bearing 5 intercalated 3 -dentate
pseudostaminodia; ovary at first tympani-
form, becoming ovoid; style 1, stigma 1,
at anthesis about equaling the stamens.
Utricle ovoid, inconspicuously operculate,
1-2 mm. long; seeds cochleate-orbiculate,
succineous, smooth 1-1.5 mm. long.

bocas del toro: vicinity of Chiriqui
Lagoon, von Wedel 1357; Water Valley,
vicinty of Chiriqui Lagoon, von Wedel
l6g6A; Bastimentos, Mariano Creek, vicinity
of Chiriqui Lagoon, von Wedel 2898.

Probably adventive in the Americas

ru. .1

thts species is presumed to be native in
tropical Asia and Africa. The glomerules
are similar in complexity to the spikelets
of some grasses, and they probably cling
rather tenaciously to fur and clothing

flora of Panama ( Amaranthaceae)

2. Cyathula achyranthoides (HBK.) Moq. in DC. Prodr. 13*:326. 1849.

Desmochaeta achyranthoides HBK. Nov. Gen. & Sp. 2:210. 1817.
Desmochaeta densiflora HBK. loc. cit. 211. 1817.
Desmochaeta uncinata R. & S. Syst. Veg. 5:5 54. 1819.
Pupalia densiftora Mart. Nov. Gen. & Sp. 2:61. 1826.
Cyathula achyranthoides j3 glabrescens Moq. in DC. Prodr. 13-:327. 1849.
Cyanthula achyranthoides y densi flora Moq. loc. cit. 327. 1849.
Achyranthes hirtiflora A. Rich, in Sagra, Hist. Cuba 11:175. 1850.
Cyathula prostrata p achyranthoides O. Ktze. Rev. Gen. 542. 1891.

Erect or decumbent geniculate pubescent or glabrate annuals or perennial-
1 m. high, often rooting at the lower nodes. Leaves usually strigose on b
surfaces, entire or slightly undulate, ovate to rhombic-ovate, apically i
to attenuate, basally acute to cuneate, 4-15 cm. long, 1.5-5 cm. broad; petioles
2—10 mm. long. Inflorescences of terminal and axillary spikes of glomerules, the
mature glomerule usually consisting of a fertile lower flower and a rudimentary,
rarely perfect, terminal flower with 2 of its sepals uacinate-aristate, the terminal
flower subtended on each side by a cluster of 3-7 hooks representing 1 modified
flower, some of the hooks conspicuously longer than the glomerule; spikes 3-15
cm. long, 5-7 mm. broad, interrupted less than half their length; bracts and
bracteoles subequal, ovate, concave, mucronate or aristate, 1.5-3 mm. long; sepals
5, subequal, ovate, concave, scariose, usually pubescent only at the extremities,
with 1-3 ribs occasionally strongly exserted dorsally, 2.5-4 mm. long, 1-1.5 mm.
broad; stamens 5, 1-2 mm. long, united below forming a tube bearing 5 inter-
calated irregularly lacerate pseudostaminodia; ovary at first tympaniform, becoming
ovoid; style 1, stigma 1, at anthesis about equaling the stamens. Utricle ovoid,
rather conspicuously operculate, 1.5-2.5 mm. long; seeds cochleate-orbiculate,
, smooth, 1.5-2.5 mm. long.

bocas del toro: vicinity of Nievecita, ca. 0-50 m., Woodson, Allen # Seibert 1806;
Little Bocas, vicinity of Chiriqui Lagoon, von Wedel 2525; Peach Creek, vicinity of
Chiriqui Lagoon, von Wedel 2650; Water Valley, vicinity of Chiriqui Lagoon, von Wedel
610 & II42. canal zone: shrubby grassy trail-border, MacBride & Featherstone 2779;
Balboa Heights, Greenman & Greenman 5038. col6n: Juan Mina plantation, Rio Chagres,
region above Gamboa, 25 m., Allen 4119. darien: trail between Paya and Pucro, Stern,
Chambers et al. 402 A; vicinity of Boca de Cupe, ca. 40 m., Allen 8ot. veraguas: be-
tween R. Tabasara and Sona on highway, Woodson, Seibert (3 Allen 504.

Extending from Mexico and the West Indies to Brazil, this species is rather
common in low-elevation disturbed habitats of tropical America. In Guatemala
it is called cola de armado and in Honduras, mozote. The stouter hooks of the
glomerules of this species may perhaps make them more painful and tenacious
dingers than those of the preceeding species.

7. PFAFFIA Mart.

Pfaffia Mart. Nov. Gen. & Sp. 2:20. 1826.
Serturnera Mart. loc. cit. 36. 1826.
Hebanthe Mart. loc. cit. 42. 1826.

28 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Flowers normally perfect, bracteate and bibracteolate, sessile, often with an
unpleasant odor. Sepals 5, discrete, hypogynous, concave, subequal, the outer 3
slightly larger, often 3 -ribbed, greenish white, becoming scariose, with a con-
spicuous tuft of intertwined whitish hairs arising mostly within and longer than
the outer sepals. Stamens 5, the filaments flattened and united below forming an
entire or appendiculate tube; anthers 2-locellate, introrse, medially attached.
Ovary ovoid to obovoid, 1-locular, 1- ovulate; ovules campylotropous on elongate
flattened funicles; style 1, very short; stigma 1, bilabiate, exceeding the style.
Fruit an indehiscent utricle; seeds cochleate orbiculate, reddish brown. Subglabrous
or copiously pubescent clambering vines or erect shrubs. Leaves opposite, entire to
undulate, stibsessile to short petiolate, glabrate to densely pubescent. Inflorescences

paniculate with opposite branches or verticillate branches ultimately bearing spikes.

(In some South American species, the inflorescence consists of condensed terminal

heads.)

Of this predominantly South American genus of some twenty species, two

members of the clambering section hebanthe are found in Panama, one apparently

reaching its northernmost station here. Stiitzer in a generic monograph (in Rep.

Spec. Nov. Beih. 88:2. 1935) has noted that the section hebanthe is closer

related to Iresine § trommsborffia than to the other two sections of the genus

Pfaffia.

a. Mature sepals 2-3.5 mm. long, the ribs often not obvious to the naked
eye; rhachises and lower leaf surfaces densely rufous-pubescent; stamen

1. Pfaffia grandiflora (Hook.) R. E. Fr. in Ark. Bot. 16 12 :10. 1921.

Iresine grandiflora Hook, in Icon. PI. 2: tab. 102. 1837.

Gomphrena panicuhita var. y hookeriana Seub. in Mart. Fl. Rr.is. 5^1 92. 1875.
Hebanthe deciph m H< k. f. B nth. & Hook. f. Gen. PI. 3:41. 1880.
Hebanthe hookeriana Hems. Biol. Cent. Am. 3:19. 1882.
tkms hookerianus O. Ktze. Rev. Gen. 543. 1891.
/ 'piens O. Ktze. loc. cit. 543. 1891.
Pfaffia hookeriana (Hems.) Greenm. in Field Mus. Bot. 2:330. 1912.

Uflora (Hook.) R. E. Fr. in Stiitzer, in Rep. Spec. Nov. Beih. 83:9. 193 5.
Pfaffia grandiflora var. hookeriana Stiitzer, loc. cit. 10. 1935.

Suffrutescent, usually pubescent, clambering perennials to about 3 m. long.
Leaves rufous- strigose, at least below, lanceolate to broadly ovate, apically acute
to attenuate, basally rounded to acute, 2-10 cm. long, 1-5 cm. broad; petioles
2-10 mm. long. Inflorescence a terminal panicle of pedunculate, opposite or
verticillate, simple or compound spikes, the lower often subtended by reduced
leaves, the rhachises and peduncles densely rufous-pubescent. Flowers perfect,
often infertile, sessile; bracts and bracteoles subequal, broadly ovate, rufous-
tomentose, 1-2 mm. long, 1-2 mm. broad; sepals 5, subequal, elliptic, the ribs
obscured by the pubescence, 2-3.5 mm. long, 1-2 mm. broad; stamens 5, 1.5-2.5
mm. long, united below into an exappendiculate tube about 0.5 mm. long; ovary

(370)

FLORA OF PANAMA (AlflcildU I J.hlCCih' )

s ovoid; stigma 1, bilabiate, subsessile. Fruit an inch
2 mm. long; seeds cochleate-orbiculate, reddish brown,
i funicle up to 1 mm. long.

This species ranges from Mexico to British Guiana and Peru. Standley (in
Contr. U. S. Nat. Herb. 27:174. 1928) reports this species, as P. hookeriana
(Hems.) Greenm., to have been collected at Gatuncillo. His identification is
questionable however since he keys the species as having lobed stamen tubes. The
reliability of pseudostaminodial characters is thrown open to suspicion by examina-
tion of Williams 828 from Darien. In this specimen, possibly a hybrid between
P. grandiflora and P. paniculate, although Luning strongly toward the former in
other characters, the stamen tube in a single flower has pscudostaminotli 1 between
some filaments and none between others. Both Fries (in Ark. Bot. 16 12 :10. 1921)
and Standley (in Field Mus. Bot. 13:490. 1936) have erroneously associated the
name P. grandiflora (Hook.) R. E. Fr. with P. bracbiata Chod. (incl. P. bangii
R. E. Fr.), a species differing by its striking dendroid pubescence.

2. Pfaffia paniculata (Mart.) O. Ktze. Rev. Gen. 542. 1891.

//, bunt Ik- {anil /.'./,■ M.:n. Nov. Gen. & Sp. 2:140. 1826.

jthe tirgata Mart. Ioc.
? paniculata

x Spreng. Syst. Veg. Cur. Post. 103. 1827.
Gompbrena paniculata Moq. in DC. Prodr. 13 2 :385. 1849.

ulata O. Ktze. Rev. Gen. 545. 1891.
Pfaffia eriantba (Poir.) O. Ktze. Ioc. cit. 543. 1891.

Suffrutescent subglabrous clambering perennials to about 3 m. long. Leaves
glabrous to appressed strigose below, lanceolate to ovate, apically acute, basally
rounded to cuneate, 2-8 cm. long, 1-4 cm. broad; petioles 2-7 mm. long. Inflores-
cence a terminal panicle of pedunculate, opposite or compound
spikes, the lowermost often subtended by reduced leaves, the rhachises and
peduncles slightly cinereous-pubescent. Flowers perfect, often infertile, sessile,
bracts and bracteoles subequal, broadly ovate, translucent except for the carinae,
1-1.5 mm. long, 1-2 mm. broad; sepals 5, subequal, elliptic, the outer 3 con-
spicuously 3-ribbed, 1.5-2.5 mm. long, 1-1.5 mm. broad; stamens 5, 1.5-2 mm.
long, united below into a tube with 2 rounded lobes between adjacent filaments,
the tube about 0.5 mm. long; ovary at anthesis obovoid, occasionally distally
4-lobate; stigma 1, bilabiate, subsessile. Fruit an included indehiscent utricle about
1 mm. long; submature seeds cochleate-orbiculate, reddish brown, about 0.5 mm.
broad, on funicles about 1 mm. long.

cocLi: El Valle, floor, 1800 ft., Allen 4475.

This species, far from its reported range of Brazil to Peru and Paraguay, is
possibly a recent introduction in Panama. The conspicuous cottony mass of hairs
are certainly an aid to fruit dispersal. Three recent works on Pfaffia (Fries, in Ark.
Bot. 16 12 :1. 1921; Suessenguth in Rep. Sp. Nov. 35:325. 1934; and Stiitzer, in

(371)

30 ANNALS OF THE MISSOURI BOTANICAL GARDEN

fe I m

*jii *

Rep. Sp. Nov. Beih. 88:1. 1935) all recognize that P. eriantba was described from
a monstrosity. This monstrosity, which occurs in at least three species, has many or
all of the flowers reduced to small sterile clusters of short stout hairs. However any
flowers that may develop normally are adequate for identification. Moquin (in
DC. Prodr. 13 2 :386. 1849), who examined the type of P. eriantba, gives a
description adequate to identify it with P. paniculata. International rules of
nomenclature recommend that names based on monstrosities be rejected; therefore
the name P. paniculata is upheld.

flora of Panama ( Amaranthaceae)

8. IRESTNE P. Br.
Iresine P. Br. Hist. Jam. 358. 1756.
Trommsdorffia Mart. Nov. Gen. & Sp. 2:40. 1826.
Rosea Mart. loc. cit. 58. 1826.
Xerandra Raf. Fl. Tell. 3:43. 18 37.
Ireneis Moq. in DC. Prodr. 13 2 :349. 1849.

Flowers perfect, polygamous
in spikes or glomes. Sepals 5, d
scariose, often with conspicuous tufts of hairs between the sepals and the bracteoles.
Stamens 5, the filaments flattened and united below into an entire or appendiculate
tube; anthers bilocellate, oblong, introrse, medially attached. Ovary 1-locular,
1- ovulate; ovule campylotropous on an elongate flattened funicle; stigmata 2-3,
filiform or deltoid. Fruit an indehiscent utricle; seeds cochlea te-orbiculate, whitish
or reddish brown, smooth. Glabrous to copiously pubescent herbs, shrubs, trees or
clambering vines. Leaves opposite, entire to denticulate, petiolate, glabrous to
sericeous, herbaceous to coriaceous. Inflorescences of spikes or glomes, paniculately
disposed, the branching alternate, opposite or verticillate.

This heterogeneous genus, of some forty species, is largely if not wholly native
to the Americas. I have seen no specimens, except of the cultivant Iresine herbstii
Hook., from outside the Americas and the West Indies. Iresine herbstii is culti-
vated in many places for its attractive variegated red and yellow foliage, but I have
seen no specimens from Panama. The genus, and especially the polygamous and
dioecious species, are sorely in need of revision. Four species are known to occur
in Panama.

polygamous; spikes ovoid or globose, pedi

32 ANNALS OF THE MISSOURI BOTANICAL GARDEN

herbaceous to coriaceous, appressed pubescent below, glabrate above, narrowly
elliptic to ovate, apically acute to attenuate and mucronate, basally rounded to
acute, 2-15 cm. long, 1-8 cm. broad; petioles 3-15 mm. long. Inflorescence a
panicle of opposite or verticillate pedunculate glomes, the rhachises cinereous-
pubescent. Flowers perfect, the bracts and bracteoles subequal, orbicular, apically
rounded, concave, entire, transparent, 1-1.5 mm. long, with a conspicuous tuft of
hairs about equaling the sepals in length and arising between the bracts and the
sepals; sepals 5, subequal, oblong-ovate, acute, obscurely 3-5 nerved, 1.5-2 mm.
long; stamens 5, united below into a shallow tube bearing two rounded lobes
between adjacent filaments; ovary subglobose; style 1, shorter than the 2 deltoid
stigmata. Fruit an indehiscent ovoid utricle, 1-1.3 mm. long; seeds whitish,
cochleate-orbiculate, about 1 mm. long.

province unknown: Sutton Hayes 934.

Apparently reaching its northern limits in Panama, this species is also found in

flora of Panama (Amaranthaceae) 33

Peru, Bolivia, Brazil and Colombia. Specimens from the northern half of the range
tend to have smaller less coriaceous leaves, hemic domingensh Urban, described
from Haiti, is a very closely related species, which Suessenguth (in Rep. Sp. Nov.
35:320. 1934) reports from Trinidad and Venezuela.

2. Iresine completa Uline & Bray, in Bot. Gaz. 21:349. 1896.

Erect or clambering shrubs, the younger branches cinereous-pubescent. Leaves
herbaceous, glabrate or scantily strigose below, ovate, apically attenuate and
mucromte, basally acute, 1.5-11 cm. long, 1-5.5 cm. broad; petioles 4-10 mm.
long. Inflorescence a panicle of opposite or verticillate pedunculate glomes, the
rhachises cinereous-pubescent. Flowers perfect (or polygamous?); bracts and
bracteoles subequal, broadly ovate, acute, concave, entire, transparent, about 1 mm.
long, with a conspicuous tuft of septate hairs about twice as long as the sepals
arising between the bracts and the sepals; sepals 5, subequal, oblong-ovate, acute,
obscurely 3-nerved, 2-2.5 mm. long; stamens 5, united below into a shallow tube
bearing obscure broadly deltoid pseudostaminodia between adjacent filaments;
ovary subglobose; style 1, slightly shorter than the 2 deltoid stigmata. Fruit an
indehiscent ovoid utricle, 1.5-2 mm. long; seeds whitish, cochleate-orbiculate,
about 1.5 mm. long.

dari£n: above Paca and near Cana, Williams Jo6.

Ranging from Panama to Guatemala, this species is nowhere commonly col-
lected. The cited specimen, annotated by Standley, agrees with the original
descriptions in all respects except that the pseudostaminodia are not prominent.
Uline & Bray (in Bot. Gaz. 21:349. 1896) have emphasized that the type, Thieme
33S, collected in 1888, is completely hermaphroditic. Donnell Smith (Enum.
Plant. Guat. 5:72. 1899) cites two of Thieme's collections, omitting Thieme's
numbers, and using his own numbers, 544.3 collected in 18 89 and 5447 collected
in 1890, as Iresine completa. Donnell Smith 5447 is however the staminate plant
of a dioecious or polygamous species and has obvious pistillodes. It is obviously
not the specimen used in the original description. Nevertheless Suessenguth (in
Rep. Sp. Nov. 35:320. 1935) offers us his "Erganzung zur Originaldiagnose"
citing Donnell Smith 5447 as the type. This cannot be the holotype because the
flowers are not perfect, they lack the basal tuft of hairs and they lack the orbicular
utricles. It is possible, but presently strictly conjectural, that 1. completa is
polygamous with the pistillate flowers perfect and the staminate flowers bearing
pistillodes.

3. Iresine angustifolia Euphr. Beskr. St. Barth. 165. 1795.

Iresine elatior Rich, in Willd. Sp. Pi. 4:766. 1805.

Iresine racemosa Poir. in Lam. Encycl. Suppl. 3:180. 1813.

Rosea elatior Mart. Nov. Gen. & Sp. 2:59. 1826.

Xerandra elatior Raf. Fl. Tell. 3:43. 1837.

Alternanthera linearis Bello, in Anal. Soc. Esp. Hist. Nat. 12:107. 1883.

34 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Erect or reclining suffruticose perennials to 1 m. high, the younger branches
glabrous or scantily strigose below, lanceolate to narrowly ovate, apically attenuate
and mucronate, basally acute to attenuate, 5-10 cm. long, 0.5-3 cm. broad;
petioles 0.5-2 cm. long. Inflorescence a panicle of alternate pedunculate ovoid
spikes, these occasionally subtended by reduced linear leaves, the rhachises glabrate
or puberulent. Flowers perfect, the bracts and bracteoles subequal, broadly ovate,
mucronate, concave, entire, transparent, 0.5-1 mm. long, with a conspicuous mass
of tawny hairs arising between the bracts and the sepals; sepals 5, subequal, oblong-
ovate, acute 3-nerved, 1-1.5 mm. long; stamens 5, united below into a shallow
tube with obsolete pseudostaminodia; ovary obovoid; style 1, slightly shorter than
the 2 papillose filiform stigmata. Fruit an indehiscent utricle, 0.5—1 mm. long;
seeds reddish brown, cochleate-orbiculate, 0.6-0.8 mm. broad.

darien: Rio Sabana, Leopold 131; Boca de Cupe, Williams 670.

This lochmocolous species extends from Mexico through Central America to
Brazil and Ecuador. Standley (in Contr. U. S. Nat. Herb. 27:174. 1928) reports
that the name cadillo, a name applied to many plants with bur-like fruits, is used
on Taboga Island.

4. Iresine celosia L. Syst. ed. 10:1291. 1759.
Celosia paniculata L. Sp. Pi. 206. 1753.

/:' L. So. PL ed. 2 1456. 1763.
Iresine diffusa Humb. & Bonpl. in Willd. Sp. PI. 4:765. 1 805.
Iresine elongata Humb. & Bonpl. loc. cit. 765. 1805.
Iresine parviflora HBK. Nov. Gen. & Sp. 2 : 1 98 . 1818.
Iresine havanensis HBK. loc. cit. 199. 1818.
Iresine mutisii HBK. loc. cit. 200. 1818.
Iresine verticillata Spreng. Syst. 1:821. 1825.
Iresine polymorpha Mart. Nov. Gen. & Sp. 2:56. 1826.
Iresine polymorpha a. alopecuroidea A

ne polymorpha effusa Mart. loc.
ne polymorpha y v

Y verticillata Mart. loc. cit. 56. 1826.
Xerandra celosioides Raf. Fl. Tell. 3 :43. 1837.
Iresine floribunda Mart. & Gal. in Bull. Acad. Brux. 101:347. 1843.
Iresine celosioides var. eriophylla Benth. Bot. Voy. Sulph. 156. 1 844.
Iresine hookeri Moq. in DC. Prodr. 13 2 :344. 1849.
Iresine acuminata Moq. loc. cit. 345. 1849.
Iresine celosioides j3 pubescens Moq. loc. cit. 347. 1849.
Iresine eriophylla Moq. loc. cit. 347. 1849.

Iresine gossypiantha A. Rich, in Sagra, Hist. Cuba 11:177. 1850.
Iresine eriophora Peyr. in Linnaea 30:21. 1850.

Altemanthera paniculata Bello, in Anal. Soc. Esp. Hisp. Nat. 12:106. 1883.
Iresine paniculata (L.) O. Ktze. Rev. Gen. 2:542. 1891. not I. paniculata Poir.
Achyranthes lanata Sesse & Moc. Fl. Mex. ed. 2:67. 1894. not A. lanata L.
Iresine paniculata var. floridana Uline & Bray, in Bot. Gaz. 21:3 53. 1896.

Erect or clambering herbaceous annuals or perennials, to 3 m. high, the younger
branches glabrous or pubescent. Leaves herbaceous, glabrate to densely pubescent,
ovate, apically acute to attenuate and mucronate, basally rounded to cuneate, 4-
15 cm. long, 1-7 cm. broad; petioles 0.5-6 cm. long. Inflorescence a panicle of
:s, the rhachises glabrate to densely pubescent.

076)

flora of Panama (Amaranthaceae)

Flowers dioecious; bracts and bracteoles subequal, broadly ov
concave, often auriculate, transparent, 0.5-1 mm. long, with a conspicuo
of hairs longer than the sepals arising between the bracts and the sepals
female flowers; stamens 5, united below into an entire or minutely denticula
ovary obovoid; style minute, n
Fruit an indehiscent utricle 0.
orbiculate, 0.5-0.6 mm. broad.

bocas del toro: Water Valley, vicinity of Chiriqui Lagoon, von Wedel i;6l tf 1849.
canal zone: Colon to Empire, Panama Railroad, Crawford 453; between Mt. Hope and
Santa Rita trail, Cowell 8l; Chagres, Isthmus of Panama, Fendler J<V; between Summit
and Gamboa, Grccnman d Greenman 5235. chiriqui: Potrero Muleto to Summit. Vul-
can de Chiriqui, 3 500-4000 m., Woodson & Scbery 466; Rio Chiriqui to Remedios, 15-50
m., Woodson, Allen 8 Seibert 1192; Finca Lerida to Boquete, ca. 1300-1700 m., Wood-
son, Allen # Seibert 1131; Bajo Mono, mouth of Quebrada Chiquero, along Rio Caldera,
155-2000 m., Woodson, Allen 8 Seibert 1012; Rio Chiriqui Viejo Valley between II Vole.in
and Cerro Punta, G. White 21; rain forest, Bajo Chorro, Boquete, 6000 ft., Davidson $30\
Volcan de Chiriqui, Boquete Distr., 8000 ft.. Dwidson p»; vicinity of New Switzerland,
central valley of Rio Chiriqui Viejo, 1800-2000 m., Allen 14II. cocLE: no specified
locality, Macbride 2717. Panama: Taboga Island, up to 300 m., Allen 117; near mouth
of R. Chagres, Allen 875; hills between Capire and Potrero, Dodge & Hunter 8645.
province unknown: Sutton Hayes 722, 928 C* 929.

This species is rather common in the warmer regions of North, Central and
South America. Standley & Steyermark (in Field Mus. Bot. 24 4 :170. 1946) report
that the sap is employed as a remedy for erysipelas around Coban in Guatemala.
Maya names for this plant are zactezxiu and zacxiu. Other vernacular names
reported are pie de palonia, velo de princesa, adorno dc niiio, chancanil, tabudo,
mosquito, hierba de gato, sietc pcllcjos, coyotUur*, coyontur* de polio, taba de
giieguecho and camaron. Iresine spiculigera Seub. and Iresine acicularis Standi, were
first lumped by Suessenguth (in Rep. Sp. Nov. 35:319. 1934), but later separated
again (in Rep. Sp. Nov. 39:13. 1935). Standley (in Field Mus. Bot. 13 2 :517.
1937) apparently accepts the lumping and further states that I. spiculigera "is
doubtfully distinct from J. celosia". Analysis of /. spiculigera with cinereous
spicular pubescence, I. acicularis, with ochraceous spicular pubescence, and 7.
frutescens Moq., with small flowers, reveals that they are best considered conspecific
with I. celosia L.

9. PHILOXERUS R. Br.
Philoxerus R. Br. Prodr. 416. 1810.
Caraxeron Vaill. in Raf. Fl. Tell. 3:38. 1837.
Blntaparon Raf. New Fl. 4:45. 1838.

Flowers perfect, bracteate and bibracteolate, subsessile in spikes. Sepals 5,
discrete or basally slightly connate, hypogynous, the outer 3 flat and broader, the
inner 2 conduplicate and narrower. Stamens 5, the filaments flattened and united
below to form a very shallow tube; pseudostaminodia absent; anthers 2-locellate,
introrse, medially attached. Ovary ovoid to orbiculate, 1-locular, 1 -ovulate;

(377)

36 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ovules on elongate flattened funicles; stigmata 2, on 1 short style, often erect at
an thesis and recurving in fruit. Fruit an orbiculate indehiscent utricle; seeds
cochleate-orbiculate, reddish brown. Prostrate or decumbent glabrous to pubes-
cent herbaceous perennials. Leaves opposite, entire, often succulent, the sessile
bases usually amplexicaul. Inflorescences of terminal and axillary pedunculate
congested spikes or glomes often subtended by slightly reduced foliar leaves.

This small genus of two or three species seems to be largely restricted to beaches
and salt flats in tropical America and west Africa. Schinz (in Engler & Prantl,
Pflanzenf. 16c:83. 1934) has suggested that many Australian species formerly
referred to this genus perhaps belong to Gomphrena. Only one species reaches
North America.

1. Philoxerus vermicularis (L.) R.Br. Prodr. 416. 1810.

Gomphrena vermicularis L. Sp. PI. 224. 1753.
lllecebrum vermiculatum L. Sp. Pi. ed. 2:300. 1762.
Gomphrena aggregata Willd. Enum. 294. 1809.

ablegatus HBK. Nov. Gen. & Sp. 2:203. 1817.
Philoxerus crassifolius HBK. loc. cit. 203. 1817.
Achyranthes vermicularis Ell. Bot. S. C. & Ga. 1 :3 10. 1821.
Gomphrena crassifolia Spreng. Syst. 1:824. 1825.
Caraxeron vermis ulan R ,t . Fi. rdl. 3:38. 1837.
Uhu'afuuon bwiifolium Rat. New. 11. 4:45. 1838.
Blutaparon repens Raf. loc. cit. 46. 1838.
Iresine vermicularis Moq. in DC. Prodr. 13 2 :340. 1849.
Iresine crassifolia Moq. loc. cit. 340. 1849.
Iresine aggregata Moq. loc. cit. 340. 1849.
Cruzeta crassifolia Maza, Fl. Haban. 94. 1897.
Lithophila vermiculata Uline, in Field Mus. Bot. 2:39. 1900.

Prostrate or decumbent perennials, the much branched stems glabrous except
at the nodes. Leaves glabrous, succulent, linear to narrowly oblanceolate, apically
acute to rounded, often mucronate, basally cuneate, 1-5.5 cm. long, 2-10 mm.
broad, the leaf bases encircling the swollen nodes, with an axillary tuft of hairs.
Inflorescences of terminal and axillary pedunculate bracteate white to stramineous
spikes or glomes, 5-30 mm. long, 5-10 mm. broad, solitary or aggregated, the
rhachises villous. Flowers perfect; bracts and bracteoles subequal, ovate, acute to
acuminate, concave, translucent, 2-3 mm. long, slightly shorter than or equaling
the sepals; sepals 5, the outer 3 almost flat, usually glabrous, apically rounded, 3-4
mm. long, 1-1.5 mm. broad, the inner 2 conduplicate, pubescent, apically acute,
3-4 mm. long, 1 mm. broad; stamens 5, 1-2.5 mm. long, united below into a very
shallow or obsolete tube; pseudostaminodia absent; ovary orbiculate to elliptic,
style 1, about as long as the 2 papillose stigmata. Fruit an indehiscent utricle,
about 2 mm. long; seeds cochleate-orbiculate, reddish brown, 0.8-1.2 mm. broad.

bocas del toro: Old Bank Island, vicinity of Chiriqui Lagoon, von Wedel 1948.

A rather characteristic sea beach plant, this species ranges from Florida and
Texas to Mexico and British Honduras, and from Panama to Brazil in the Amer-
icas; it also occurs on the west coast of Africa. As the name implies, the plant

(378)

(379)

38 ANNALS OF THE MISSOURI BOTANICAL GARDEN

reputedly possesses vermifuge qualities. P. portulacoides St.-Hil. differing in
having glabrous white sepals, and P. littoralis Suesseng., differing in having glabrous
white inner sepals and pilose foliar midribs are weakly denned species occurring

10. ALTERNANTHERA Forskal
Alternanthera Forskal, Fl. Aeg.-Arab. 28. 1775.

Flowers perfect, bracteate and bibracteolate, sessile or subsessile in spikes or
glomes. Sepals 5, discrete, hypogynous, concave, subequal or strongly dissimilar,
white to purple, the bases often indurate. Stamens 3-5, the filaments flattened
and united below into an entire or appendiculate tube; anthers 2-locellate, oblong,
introrse, medially attached. Ovary globose to obovoid, often compressed, 1-locular,
1-ovulate; ovule campylotropous on an elongate funicle; stigma 1, capitate or
obscurely bilabiate, on a short or elongate style. Fruit a membranaceous utricle,
indehiscent; seeds cochleate-orbicular, smooth. Glabrous to copiously pubescent
erect, scandent or prostrate annuals or perennials. Leaves opposite, sessile or
petiolate, entire to denticulate, glabrous to densely pubescent, mostly herbaceous.
Inflorescences of axillary and terminal sessile or pedunculate spikes or glomes.

This genus of over 100 species is best developed in the Americas, although
many species are ubiquitous weeds. Some species are cultivated for their ornamental
inflorescences and others are planted as tufted border plants and soil binders.
Alternanthera axillaris Hornem. and A. repens (L.) Gmel. are rather obnoxious weeds
with pungent sepals. Working under the American Code of nomenclature, Standley
(in Jour. Wash. Acad. 5:72. 1915) explains the rather complicated reasons which
led him to transfer the Alternanthera species to Achyranthes, a treatment which
he later abandoned. Hitchcock & Green (in Brittonia 6:114. 1947) have proposed
that Alternanthera repens (L.) Gmel. be the lectotype of the genus. Eleven
species are reported from Panama.

as the inner (sen !c m Z Tessi™*) ;

prostrate or decumbent herbs.

b. Utricle longer than the sepals; sepals subequal acute, 1 -nerved,

glabrous; pubescence of simple hairs ... 1. A. sessilis

bb. Utricle shorter th,r rphic or trimorphic,

s long as the sepals; stigrn

flora of Panama ( Amaranthaceae)

. Sepals 5-6 mm. long, 3-4 times as long as the bracts; glomes white,
10-18 mm. broad; peduncles unbranched, mostly less than 5 cm.

Sepals 2-6 mm. long, no more than twice as long as the brae
2 mm. broad; peduncles brand

mm. broad, if narrower,
lostly simple and exceeded

. Leaves ovate to ellip
peduncle branched or

1. Alternanthera sessilis (L.) R. Br. Prodr. 417. 1810.

Gompbrena sessilis L. Sp. PI. 225. 1753.
Illecebrum sessile L. Sp. PL ed. 2:300. 1762.
lllecebrum indicum Houtt. Nat. Hist. 2:7. 1777.
Alternanthera triandra Lam. Encycl. 1:95. 1783.
Alternanthera denticulata R. Br. Prodr. 417. 1810.
Alternanthera nodiflora R. Br. loc. cit. 417. 1810.
Allagantbera forskalli Mart. PL Hort. Erlang. 69. 1814.
Paronychia sessilis Desf. Tabl. Bot. ed. 2:54. 1815.
Adenogramma oppositifolia Hassk. in Flora 31:754. 1851.
Alternanthera tenuissima Suesseng. in Bot. Arch. 39:382. 1939.

Prostrate or decumbent puberulent or glabrous annuals or perennials, th
sparsely branched stems to 60 cm. long. Leaves glabrous or sparsely villous below

40 ANNALS OF THE MISSOURI BOTANICAL GARDEN

elliptic, apically acute and mucronate, basally cuneate to attenuate, 1-4 cm. long,
0.5-1.5 cm. broad; petioles 1-5 mm. long, often winged. Inflorescences of white
or pinkish globose heads, 2-4 mm. long, 2-4 mm. broad, sessile in the axils.
Flowers perfect, the bracts and bracteoles subequal, acute, 1 mm. long or less;
sepals 5, subequal, transparent, not indurate, 1 -nerved, slightly concave, 1-1.5
mm. long; stamens 5, united below into a short tube, the pseudostaminodia entire,
exceeded by the anthers, about equaling the filaments; ovary obovoid, style 1,
3-4 times as long as the minute capitate stigma. Fruit an indehiscent obcordate
yellowish utricle, 1.5-2 mm. long, usually slightly exceeding the sepals; seeds yel-
lowish or reddish brown, cochleate-orbicular, 1-1.2 mm. broad.

This pan-tropical weed seems to range from Honduras to Brazil in the Amer-
icas. It inhabits open moist areas and may grow as an aquatic, in which case the
vegetative portions are unusually large.

2. Alternanthera polygonoides (L.) R. Br. Prodr. 417. 1810.

Gomphrena polygonoides L. Sp. Pi. 225. 1753.

lllecebrum polygonoides L. Sp. PI. ed. 2:300. 1762.

Achyranthes polygonoides (L.) Lam. Encycl. 1:547. 1785.

Buchokia polygonoides Mart. Nov. Gen. & Sp. 2:51. 1826.

Alternaritbt. til. Voy. Distr. Diam. 2:439. 1833.

Steiremis repens Raf. Fl. Tell. 3:41. 1837.

Telanthera polygonoides Moq. in DC. Prodr. 13 2 :363. 1849.

Alternanthera ficoidea Griseb. Fl. Brit. W. Ind. 67. 1859. not Gomphrena ficoidea L.

Procumbent villous or glabrate perennials, the much branched stems to 80 cm.

long. Leaves villous or glabrate, elliptic to ovate or obovate, apically acute to

rounded and mucronate, basally cuneate to attenuate, 1-3 cm. long, 0.5-1.5 cm.

broad; petioles 2-10 mm. long, often winged and with basal tufts of dentate hairs.

Inflorescences of white globose to cylindric heads, 4-10 mm. long, 3-6 mm. broad,

sessile in the axils. Flowers perfect, the bracts and bracteoles subequal, acute, ovate,
1-1.5 mm. long; sepals 5, the outer 3 broader, 3-nerved, slightly indurate, pilose

below, acuminate, 3-4 mm. long; stamens 5, united below into a short tube, the

pseudostaminodia dentate, much exceeded by the anthers; ovary obovoid, style 1,

about half as long as the capitate stigma. Fruit an indehiscent obcordate to

orbicular utricle, 1-1.5 mm. long; seeds reddish brown, often strongly reticulate,
1-1.2 mm. broad.

This species, occasionally adventive in eastern United States, ranges from Mexico
to Argentina in open dry to moist habitats. Standley (in Contr. U. S. Nat. Herb.
27:174. 1928) reports it from Panama but since he keys it as having simple
pubescence, the identity of the specimen or specimens is questionable. Backer (in
Fl. Mai. 1<:93. 1949) seems to have lumped this species with A. ficoidea (L.)
R. Br., but the staminodial differences between the two species appear to be con-
stant. Standley (in Field Mus. Bot. 18:419. 1937) reports the vernacular name
golondrina in Costa Rica. The South American species Alternanthera boliviana
Rusby is a very closely allied species, with acute sepals and unusually long spikes.

0*2)

flora of Panama (Amaraiithaccae) 41

lns (L.) Gmel. Linn. Syst. Nat. ed. 13 2 :106. 1791.
Acbyranthes repens L. Sp. Pi. 205. 1753.
Alternanthera achy rant h. Forsk. Fl. Aeg.-Arab. 2 8. 175 5.
lllecebrum achyrantha L. Sp. PI. ed. 2:299. 1762.

Acbyranthes radio S-.27. 1801.

Paronychia achyrantha Desf. Tabl. Bot. ed. 2:54. 1815.
Pityranthus crassifolius Mart. Denks. Akad. Munch. 5:179. 1817.

Alternanthera liWfhna Scheie, in 1 i,, 11: uu 22:149. 1849.
Alternanthera parviflora Fawc. & Rendle, Fl. Jam. 3:139. 1914.

Procumbent villous or glabrate perennials, the much branched sums up to 50
cm. long. Leaves appressed villous or glabrate, ovate to obovate, apically acute to
rounded and mucronate, basally rounded to attenuate, 5-25 mm. long, 3-15 mm.
broad; petioles 2-10 mm. long, often winged. Inflorescences of yellowish or stra-
mineous usually cylindric heads, 5-12 mm. long, 4-8 mm. broad, sessile in the
axils. Flowers perfect, the bracts and bracteoles subequal, 3-4 mm. long; sepals 5,
trimorphic, the outer 2 broader, strongly 3 -nerved, indurate, pilose below and with
rigid awns, the third sepal narrower, shorter and merely acuminate, the inner 2
conduplicate; stamens 5, united below into a shallow tube, the pseudostaminodia
entire or dentate, exceeded by the anthers; ovary obovoid, style 1, about as long
as the minute capitate stigma. Fruit an indehiscent obcordate utricle, 1.2-1.5 mm.
long; seeds reddish brown, 1-1.5 mm. broad, cochleate-orbicular.

Presumably a native of tropical America, this weedy species is now also well
established in Asia, Europe and the East Indies. Standley's recount of the distribu-
tion (in N. Am. Flora 21:136. 1916) indicates that the species has been
collected in Panama. It certainly is to be expected there. In the Yucatan the
Mayas apply the name cabalxtez; in Guatemala the names sanguinaria, sacacbiquim
and hierba de toro are reported. This species is doubtfully distinct from A. pungens
HBK., which differs in having subrotund leaves and longer less villous sepals.

4. Alternanthera ficoidea (L.) R. Br. Prodr. 1:417. 1810.

Gompbrena ficoidea L. Sp. PI. 225.

lllecebrum ficoideum L. Sp. Pi. i

Paronychia ficoidea Desf. Tabl. Bot. ed. 2:14. 1815.

Bucbohia poly^onohU < a erecta Mart. Nov. Gen. & Sp. 2:51. 1K26.

Bucholzia polygonoides p diffusa Mart. loc. cit. 51. 1826.

Bucholzia polygonohh; ; .-./,//«■.»»/< Mart. loc. cit. 51. 1826.

Bucholzia ficoidea Mart. loc. cit. 52. 1826.

Steiremis ficoidea Raf. Fl. Tell. 3:41. 1837.

/ . /. m-uiu ,j mviulieinis Walp. in Nov. Act. Ac. Nat. Cur. IV: Mjppi. I : *04. 1843.

Telanthera polygonoides o diffusa Moq. in DC. Prodr. 13 2 :364. 1849.

Telanthera polygonoides P brachiata Moq. loc. cit. 364. 1849.

I !. Brit. W. Ind. 67. 1859. not Gompbrena polygono-
ides L.
Alternanthera polygonoides P glabrescens Griseb. loc. cit. 67. 1859.
Alternanthera versicolor Hort. ex Regel, Gartenfl. 101. 1869.
AltentHtherA I Diet, ed. 1:59. 1884.

ie»s Back. & Sloot. Handb. Thee. 108. 1924.

(383)

42 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Procumbent or decumbent hispidulous or glabrate perennials, the branching
stems to 100 cm. long. Leaves slightly pubescent but soon glabrate, elliptic to
broadly ovate or obovate, apically acuminate to acute and mucronate, basally
rounded to cuneate, 2-6 cm. long, 0.5-2 cm. wide; petioles 2-10 mm. long, usually
winged. Inflorescences of stramineous to white globose to ovoid heads, 3-10 mm.
long, 3-6 mm. broad, sessile in the axils. Flowers perfect, bracts and bracteoles
subequal, ovate, arista te, 1.5-3 mm. long; sepals 5, the outer 3 broader, 3 -ribbed,
basally indurate and hispidulous, 3-5 mm. long; stamens 5, united below into a
tube, the pseudostaminodia lacerate and exceeding the filaments; ovary obovoid;
style 1 , 2-3 times as long as the capitate stigma. Fruit an indehiscent suborbicular
utricle, 1-1.5 mm. long; seeds reddish brown, cochleate-orbiculate, 0.8-1.2 mm.

canal zone: vicinity of Monkey Hill, Cornell 33; Gatun, Sutton Hayes 185; near
Culebra, 50-200 m., Pittier 2289; Chagres, Isthmus of Panama, Fendler 261; Ancon,
Greenman & Greenman 5018; low ground, outskirts of Ancon, Greenman 6 Greettman
5047; near Corrosion Laboratory, vicinity of Miraflores Locks, Stern, Chambers et al. 65.
chiriqui: vicinity of San Bartolome, Peninsula de Burica, 0-50 m., Woodson & Schery
941. province unknown: Sutton Hayes 940.

The native range of this species is perhaps from Mexico to Argentina, but
various forms are cultivated elsewhere as ornamentals and soil binders.
5. Alternanthera halimifolia (Lam.) Standi, in Pittier, PI. Us. Venez. 145.

Acbyranthes halimifolia Lam. lincycl. 1 :547. 1785.

lllecebrum frutescens L'Her. Stirp. Nov. 75. 1788.

lllecebrum limense Dum.-Cours. Bot. Cult. 1 :646. 1 802.

Paronychia frutescens Desf. Tabl. Bot. ed. 2:54. 1815.

Telanthera crucis Moq. in DC. Prodr. 13 2 :362. 1849.

Telanthera frutescens \f q. I,,c. at, 365. 1849.

Telanthera frutescens £ acutifolia Moq. loc. cit. 366. 1849.

Alternanthera ficoidea-y halimifolia (Lam.) O. Ktze. Rev. Gen. 539. 1891

Telanthera flavogrisea Urban, Symb. Ant. 1:300. 1899.

Alternanthera asterotricha Uline, in Field. Mus. Bot. 1 :419. 1899.

Alternanthera flavogrisea Urban, Symb. Ant. 5:340. 1907.

Alternant hers Dutch W. I. 1:58. 1909.

Telanthera halimifolia A. Stewart, in Proc. Calif. Acad. 4*:58. 1911

Alternanthera ficoidea var. flavogrisea Fawc. & Rendle, Fl. Jam. 3:140. 1914.

Procumbent or decumbent sericeous perennials, the sparingly branched stems
to 20 cm. long. Leaves densely pubescent with stellate hairs, rarely glabrate,
elliptic to ovate, apically rounded and mucronate, basally rounded to cuneate, 1-5
cm. long, 0.5-2.5 cm. broad; petioles 2-10 mm. long. Inflorescences of stramineous
ovoid heads, 4-12 mm. long, 3-5 mm. broad, sessile in the axils. Flowers perfect;
bracts and bracteoles subequal, ovate, acuminate to aristate, 1-3 mm. long; sepals
5, the outer 3 broader, 3 -ribbed, basally indurate and hispidulous, 3-5 mm. long;
stamens 5, united below into a tube, the pseudostaminodia lacerate and exceeding
the filaments; ovary globose to obovoid; style 1, about twice as long as the capitate
stigma. Fruit an indehiscent globose utricle 1-1.5 mm. long; seeds reddish brown,
cochlcate-orbicular, 0.8-1.1 mm. broad.

flora of Panama ( Amarantbaccac) 43

Inhabiting woods, thickets, pond borders and sea beaches, mostly at low eleva-
tions, this species extends from Mexico and the West Indies to Colombia, Chile and
the Galapagos Islands. Standley (in N. Am. Flora 21:140. 1917) indicates that
it occurs in Panama. The obvious silvery pubescence usually separates this from
the preceding species with which it reputedly intergrades.

6. Alternanthera philoxeroides (Mart.) Griseb. Abh. Ges. Wiss. Gott. 24:
36. 1879.

Bucholzia philoxeroides Mart. Nova Acta Acad. Leop.-Carol. 1 3 1 :3 1 5. 1826.

Telanthera philoxeroides Moq. in DC. Prodr. 13 2 :362. 1849.

Telanthera philoxeroides a obtusifolia Moq. loc. cit. 363. 1849.

Telanthera philoxeroides p acutifolia Moq. loc. cit. 363. 1849.

Telanthera philoxeroides var. S phyllantha Seub. in Mart. Fl. Bras. 5 X ;169. 1875.

Telanthera philoxeroides var. -y dcnticulata Seub. loc. cit. 170. 1875.

Achyranthes philoxeroides (Mart.) Standi, in Jour. Wash. Acad. Sci. 5:74. 1915.

: or ascending glabrate aquatic perennials, the simple or branched,
often fistulose stems to 100 cm. long. Leaves glabrous or glabrate, lanceolate to
narrowly obovate, apically rounded to acute, basally cuneate, rarely denticulate,
2-10 cm. long, 0.5-2 cm. broad; petioles 1-3 mm. long. Inflorescences of terminal
and occasionally axillary white glomes, 10-18 mm. long, 10-18 mm. broad, the
usually unbranched peduncles 1-5 cm. long. Flowers perfect, bracts and bracteoles
subequal, ovate, acuminate, 1-2 mm. long; sepals 5, subequal, oblong, apically
acute and occasionally denticulate, neither indurate nor ribbed, 5-6 mm. long,
1.5-2.5 mm. broad; stamens 5, united below into a tube, the pseudostaminodia
lacerate and exceeding the anthers; ovary reniform, the style about twice as long
as the globose capitate stigma. Fruit an indehiscent reniform utricle 1 mm. long,
1-1.5 mm. broad; mature seeds not seen.

Standley (in N. Am. Fl. 21:142. 1917) gives the range of this species as from
Colombia to Brazil and Argentina, adventive in the southeastern United States.
I have seen no specimens from Panama but it could easily be adventive there.
Suessenguth (in Rep. Spec. Nov. 39:4. 1935) reports a teratological specimen
from Pearl Islands in San Jose in which the stamens are replaced by carpels.

7. Alternanthera williamsii (Standi.) Standi, in Jour. Wash. Acad. Sci. 15:

458. 1925.
Achyranthes williamsii Standi, in Contr. U. S. Nat. Herb. 18:89. 1916.

Erect or clambering pubescent perennials, the sparingly branched stems to 5 m.
long. Leaves appressed pubescent, sometimes variegated with purple, ovate,
apically acute to acuminate, basally rounded to acute, 2.5-10 cm. long, 1-4 cm.
broad; petioles 2-8 mm. long. Inflorescences of terminal and axillary white to
stramineous, globose to ovoid heads, 10-30 mm. long, 10-12 mm. broad, the usually
unbranched peduncles 1-7 cm. long. Flowers perfect; bracts and bracteoles
subequal, ovate, carinate, aristate, 2-4 mm. long; sepals 5, subequal, oblong,
long, 1-1.5 mm. broad; stamens 5, united
i lacerate and exceeding the anthers; ovary

44 ANNALS OF THE MISSOURI BOTANICAL GARDEN

globose; style 1, 1-2 times as long as the globose stigma. Fruit an indehiscent
obovoid utricle, the seeds reddish brown, cochleate-orbicular, 1.2-1.5 mm. long.

canal zone: between Summit and Gamboa, Greenman & Greenman 5244- col6n:
Ahorca Lagarto, Cowell 255. darien: vicinity Yape, Rio Tuira, Allen 4302; Tucuti,
Chepigana, Terry & Terry 1 388; Cituro, Williams 675. province unknown: Sutton
Hayes g4Q\ Halsted in 1850.

An inhabitant of dry thickets, this species ranges from Panama to southern
Mexico. It is very closely related to the Mexican species A. gracilis (Mart. & Gal.)
Standi, which supposedly differs in having globose heads, recurved sepals, trichoto-
mous peduncles and black seeds. Intermediate forms do occur however. A.
wHliamsii f. purpurea Standi., with purplish red leaves, is cultivated in the Canal

8. Alternanthera laguroides (Standi.) Standi, in Standi & Cald. Lista PI.
Salvador 74. 1925.

Achy rant lie U&uruides Si.mdl. in Cunu. U. S. \ T ar. Herb. 18:90. 1916.

Erect or clambering pubescent perennials, the much branched stems to 4 m.
long. Leaves densely sericeous, especially below, lanceolate to narrowly ovate,
apically attenuate, basally attenuate, 2-10 cm. long, 1-3.5 cm. wide; petioles 1-4
mm. long. Inflorescences of terminal and axillary stramineous to whitish ovoid
heads, 5-15 mm. long, 4-12 mm. broad, the peduncles to 9 cm. long, branched or
unbranched, naked or leafy (when leafy, the heads may appear sessile). Flowers
perfect; bracts and bracteoles subequal, ovate, acuminate to aristate, 2-3.5 mm.
long; sepals 5, narrowly oblong, 1-3 nerved, scarcely indurate, 4-5.5 mm. long,
with basal hairs almost exceeding the sepals; stamens 5, united below into a tube,
the lacerate pseudostaminodia exceeding the anthers; ovary obovoid to globose;
style 1, 2-3 times as long as the globose stigma. Fruit an obovoid utricle to 1 mm.
long; mature seeds not seen.

According to Standley & Steyermark (in Field Mus. Bot. 24:149. 1946) this
species ranges from Guatemala to Panama in dry thickets of the Pacific slope. In
Guatemala it is called botoncito. Possibly Moquin was using this species in his
description of Telantbera pubifiora a monocephala, the type of which was collected
in Panama. I have not seen the type, but believe the description fits A. laguroides
more closely than any of the other Central American species.

9. Alternanthera stenophylla (Standi.) Standi, in Field Mus. Bot. 8:9.
1930.

Achyranthes stenophylla Standi, in Contr. U. S. Nat. Herb. 18:90. 1916.

Ascending or erect sparsely strigose or glabrate perennials. Leaves sparsely
pilose or glabrate, linear to linear-elliptic, apically acute, basally acuminate, 2.5-5
cm. long, 3-6 mm. broad; petioles 2-6 mm. long. Inflorescences of axillary and
terminal stramineous cylindric heads, 6-10 mm. long, 5-6 mm. broad; peduncles
unbranched, 2-5 cm. long. Flowers perfect, the bracts and bracteoles subequal,
ovate, aristate, 1-2 mm. long; sepals 5, oblong, obscurely 3-nerved and indurate

(386)

flora of Panama ( Amaranthaccae) 45

basally, 2-2.5 mm. long; stamens 5, united below into a tube, the lacerate pseudo-
staminodia exceeding the anthers; ovary subglobose; style 1, 2-3 times as long
as the globose stigma. Mature fruit not seen.

province unknown: Sutton Hayes 941.

The type for this species was collected by Sutton Hayes in the year 1859-1860
and the only given locality was Panama. Apparently it is known only from the
type collection. Except for the linear leaves, this species very strongly resembles
the following species.

4EXICANA (Schlecht.) Hieron. Bot. Jahrb. 20: Beibl. 49:8.

A ,. in DC. Prodr. 13 2 :371. 1849.
Telanthera mexicana Moq. loc. cit. 372. 1849.

Alternanthera lehmannii Hieron. in Bot. Jahrb. 20: Beibl. 49:8. 1895.
Achyrantbes lehmannii (Hieron.) Standi, in Jour. Wash. Acad. 5:74. 1915.
Achyranthes mexicana (Schlecht.) Standi, loc. cit. 74. 1915.

Ascending or erect pilose or glabrate perennials, the scantily branched stems
to 1 m. long. Leaves appressed pilose or glabrate, elliptic to ovate, apically attenu-
ate, basally acute to cuneate, 2.5-10 cm. long, 1-5 cm. broad; petioles 2-15 mm.
long. Inflorescences of axillary and terminal white to stramineous subglobose to
cylindric heads, 5-10 mm. long, 4-7 mm. broad; peduncles mostly unbranched, to
9 cm. long. Flowers perfect, the bracts and bracteoles subequal, ovate, acuminate
to aristate, 1-2 mm. long; sepals 5, oblong, obscurely 3 -nerved and slightly
indurate basally, 2-3.5 mm. long, 1 mm. broad; stamens 5, united below into a
tube, the lacerate pseudostaminodia exceeding the filaments; ovary globose to
obovoid; style 1, about twice as long as the globose stigma. Fruit an indehiscent
obovoid utricle, 1-1.5 mm. long; seeds reddish brown, cochleate-orbiculate, about

Sutton Hayes 94$.

Inhabiting mesic forests, this species extends from Mexico south locally to
Colombia. It is of more than casual interest that Sutton Hayes 945 is A. mexicana,
Sutton Hayes 944 is the type of A. panamensh and Sutton Hayes Q41 is the type
of A. stenophylla. The flowers of all these specimens are very similar indeed and
would suggest that the latter two were mere vegetative variants of A. mexicana.

ensis (Standi.) Standi, in Field Mus. Bot. 8:9.

Contr. U. S. Nat. Herb. 18:89. 1916.

: perennials, the stems much branched. Leaves

11. Alterna
1930.

Acbyn

tnthes par

lamenm Standi, in 1

Er,

ect or (

rlambering pilose

appressed pilose, ovate to ellipt

2-5.5

ig, 0.5-2 cm. v

termir

lal and a:

icillary stramineoi]

long,

about 7

mm. broad; pe(

apically acute to acuminate, basally subacute,
le; petioles 1-4 mm. long. Inflorescences of
to purplish globose to cylindric heads, 7-1 1 mm.
ncles mostly branched, 1-6 cm. long, villous.

flora of Panama (Amaranthaceae) 47

Flowers perfect; bracts and bracteoles subequal, ovate, acuminate to aristate, 1-2
mm. long; sepals 5, oblong, strongly 3-nerved and indurate below, 2.5-3.5 mm.
long; stamens 5, united below into a tube, the lacerate pseudostaminodia exceeding
the filaments; ovary obovoid to subglobose; style 1, about as long as the globose
stigma. Fruit an indehiscent subglobose utricle; seeds dark reddish brown to black,
cochleate-orbiculate, 1 mm. broad.

province unknown: Sutton Hayes Q44.

This species, represented solely by the type collection, is perhaps just a casual
variant of A. mexicana (Schlecht.) Hieron. It seems to bear the same relation to
A. mexicana that the branched-pedunculate A. gracilis (Mart. & Gal.) Standi,
bears to the unbranched-pedunculate A. uilliamm (Standi.) Standi.

11. GOMPHRENA L.
Gomphrena L. Sp. PI. 224. 1753.

lh>i;i.mtLi WmA. Fast. Pi. Nov. 6. 1771.

WadapusKii.Vl. Tell. 3:77. 1837.

Xerosiphon Turcz. in Bull. Soc. Nat. Mosc. 16:55. 1843.

Xeraea O. Ktze. Rev. Gen. 545. 1891.

Amarantoides Maza, Fl. Haban. 94. 1897.

Flowers perfect, bracteate and bibracteolate, sessile in spikes or heads. Sepals 5,
basally connate or discrete, hypogynous, concave, subequal, colorless or brightly
colored. Stamens 5, hypogynous, the filaments connate forming a deep stamen
tube, the anthers sessile at the summit of the tube; pseudostaminodia 2-lobate;
anthers 2-locellate, introrse, medially attached. Ovary globose to obovoid, 1-
locular, 1 -ovulate, the ovule campylotropous on an elongate flattened funicle;
style 1, filiform, about as long as the 2 filiform stigmata. Fruit an indehiscent
utricle; seeds cochleate-orbiculate. Pubescent erect or decumbent annual or
perennial herbs. Leaves opposite, sessile or petiolate, entire, usually pubescent,
apparently deciduous. Inflorescences of terminal and occasionally axillary globose
to cylindric solitary or clustered heads, usually subtended by foliar leaves.

This genus of nearly a hundred species, with some native to tropical America
and others to tropical Austral-Asia, seem to reach its best development in South
America. Several species are widely cultivated, as is the "bachelor's button",
G. globosa, a rather handsome ornamental with edible leaves. Four species are
known from Central America but only two are reported from Panama.

a. Heads 5-14 mm. in diameter; bracteoles 4-6 mm. long; leaves 1-5

aa. Heads 15-25 nun 7-12 nm long; leaves 2-

10 cm. long 2. G. globosa

1. Gomphrena decumbens Jacq. Hort. Schoenbr. 4:41. 1804.

Xeraea decumbens O. Ktze. Rev. Gen. 545. 1891.

Gomphrena decumbens f. albiflora Chod. & Hassl. in Bull. Herb. Boiss. 2 3 :389. 1903.

Gomphrena decumbens f. roseiflora Chod. & Hassl. be. cit. 389. 1903.

(589)

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Gomphrena ixiamensis Rusby, in Bull. N. Y. Bot. Gard. 6:502. 1910.
Gomphrena / Rep. Spec. Nov. 11:156. 19

Gomphrena decumbem var. grandifolia Stuchlik, loc. cit. 157. 1912.
Gomphrena perennh subsp. pseudodecumbens Stuchlik, loc. cit. 153. 1912.

, IoC. or. 153. 1912.
Gomphrena dispersa Standi, in Contr. U. S. Nat. Herb. 18:91. 1916.
Gomphrena decumbem var. carinata Suesseng. in Rep. Sp. Nov. 39:8. 193 5.

Prostrate or decumbent pilose annual or perennial herbs, the much branching
stems to 50 cm. long, often rooting at the nodes and forming mats. Leaves seri-
ceous to strigose, occasionally glabrate, oblong to ovate or obovate, apically acute
to rounded and mucronate, basally acute to attenuate, 1-5 cm. long, 0.5-2.5 cm.
broad; petioles 1-5 mm. long. Inflorescences of terminal and axillary ovoid to
cylindric whitish or purplish heads, occasionally clustered, 4-20 mm. long, 5-14
mm. broad, subtended by ovate leaves 0.5-2 cm. long. Flowers perfect; bracts

(390)

flora of Panama (Amaratithaceae) 49

persistent, ovate, acuminate, 1.5-3 mm. long; bracteoles narrowly to broadly
cristate, equaling or exceeding the sepals, 4-6 mm. long; sepals 5, subequal, basally
connate, cinereous- to rufous-lanate, ultimately indurate and closely embracing
the fruit; stamens 5, subequal, the anthers sessile between the 5 2-lobate pseudo-
staminodia; stamen tube 3-6 mm. long; ovary globose; style 1, 1-1.5 mm. long;
stigmata 2, 1-1.5 mm. long. Fruit an indehiscent areolate utricle 1-2 mm. long;
seeds cochlea te-orbiculate, yellowish to reddish brown, ocellate, 1.5-1.7 mm. broad.

bocas del toro: Research grounds, region of Almirante, Cooper 138. canal zone:

open sunny area next to road, Corrosion Laboratory, vicinity of Miraflores Locks, Stern,

mm & Greenman 5015; vicinity of Miraflores Lake,

White 8 White 192; Pedro Miguel, Bro. Heriberto 19. chiriquI: vicinity of Puerto

Armuelles, 0-75 m., Woodson tf Schery 819.

This species ranges from the "West Indies and Mexico to Panama and from
Brazil and Bolivia to Argentina. Common names in Guatemala are botoncillo,
sangrinaria, sanguinariu, siempreiha and •■<<<! hu; ,ni dv montc; other names in the
Yucatan and adjoining regions are chacmol, tmiuil, amor scco and secicante.
Standley (in Contr. U. S. Nat. Herb. 18:91. 1916) has separated a variant, with
the crests of the bracteoles conspicuously widest at the apex, as G. dispersa, which
seems to occur throughout the range of G. decumbent.. No correlated differences
exist and transitions occur, so G. dispersa is reduced to synonymy. As Suessenguth
(in Rep. Spec. Nov. 39:8. 1935) has noted Greenman tf Greenman 5015 is
intermediate between G. deenmbens and G. nitida Rothr., a questionably distinct
species supposed to differ in having lacinate-dentate crests, larger rarely purple
heads and an erect habit. Palmer 401 from Mexico has crests both dentate and
laciniate-dentate, or even absent; the specimen also has both included and exserted
stamen tubes. This particularly illuminating example serves to demonstrate that
G. nitida is hardly more deserving of specific status than is G. dhpersa.

2. Gomphrena globosa L. Sp. Pi. 224. 1753.

Gomphrena globosa P carnea Moq. in DC. Prodr. 13 2 :409. 1 849.

Gomphrena globosa y albiflora Moq. loc. cit. 409. 1849.

Xeraea globosa O. Ktze. Rev. Gen. 545. 1891.

Amarantoides globosus Maza, FI. Haban. 94. 1897.

Amarantoides globosus albiflorus Maza, loc. cit. 95. 1897.

Gomphrena globosa var. aureiflora Stuchlik, in Rep. Spec. Nov. 12:340. 1913.

Erect or ascending pilose annual herbs, the branching stems to 60 cm. high.
Leaves sericeous, occasionally glabrate, oblong to ovate or obovate, apically acute
and mucronate, basally rounded to attenuate, 2-10 cm. long, 0.5-5 cm. broad;
petioles 1-15 mm. long. Inflorescences of terminal and axillary reniform to
cylindric variously colored heads or spikes, occasionally clustered, 10-25 mm. long,
15-25 mm. broad, subtended by ovate leaves 5-20 mm. long. Flowers perfect;
bracts persistent, ovate, acuminate, 3-4 mm. long; bracteoles dentate-cristate,
exceeding the sepals, 7-12 mm. long, 2-3.5 mm. broad; sepals 5, subequal, basally
connate, cinereous- to rufous-lanate, ultimately indurate and closely embracing
the fruit, 5-6.5 mm. long; stamens 5, the anthers sessile between the 5 2-lobate

(3*1)

50 ANNALS OF THE MISSOURI BOTANICAL

pseudostaminodia; stamen tube 4-8 mm. long; ovary globose; style 1, 1-1.5 mm.
long; stigmata 2, 1-1.5 mm. long. Fruit an indehiscent areolate utricle 1.5-2.5
mm. long; seeds cochleate-orbiculate, reddish brown, ocellate, 1.5-2 mm. broad.

Cultivated and escaped in many regions of the world, this species may be a
native of America although it was originally described from India. Standley (in
Contr. U. S. Nat. Herb. 27:173. 1928) reports that it occurs in Panama as a
cultivant and an escape. In Panama it is called siempreviva and suspiro. The
names amor seco, inmortal and boton are also used in Guatemala. Some of these
vernacular names suggest the persistence of the flowerss which makes them popular
in wreaths and dry floral arrangements. That the differences between this and the
preceding species are quantitative may suggest a polyploid derivation.

flora of Panama (Nyctaginaccac) 51

NYCTAGINACEAE*

Herbs, shrubs, or trees, sometimes scandent, the stems frequently swollen at
the nodes, particularly in the herbaceous genera, and sometimes armed with
axillary spines. Leaves alternate, opposite, whorled, or approximate, simple,
estipulate, entire. Flowers monochlamydeous, perfect or unisexual, in various
cymose inflorescences, occasionally solitary, the bracteoles often enlarged and
involucrate, herbaceous or petalaceous and enclosing one or more flowers. Perianth
monochlamydeous, herbaceous or petalaceous, tubular to infundibuliform, cam-
panulate, urceolate, or salverform, involute-plicate in aestivation, the base of the
tube persistent in fruit, the limb persistent or deciduous. Stamens 1 to many,
hypogynous, the filaments unequal, usually united at the base, the anthers included
or exserted, dorsifixed near the base. Ovary 1 -celled, containing a solitary basal
ovule; style elongate, short, or lacking. Fruit a fleshy, coriaceous, or woody
anthocarp, occasionally bearing viscid stipitate glands.

Nyctaginaceae are an essentially tropical and subtropical family, particularly
in the New World. The Four-o'clock Family is represented in Panama by five
indigenous genera, but more familiar to the general public are the introduced
genera Bougainville a and Mirabilis. Bougainville a, almost ubiquitously planted as
an ornamental woody vine because of its handsome clusters of three showy magenta,
red, pink, white, or yellow bracts subtending the inconspicuous flowers, is a South
American genus which does not appear to naturalize readily in Panama. The
plants variously are referred to B. glabra Choisy and B. spectabilis Willd. It
usually is known by its Latin name, or by the vernacular veranera and ftor de
verano because of its habit of blooming in the dry season. Mirabilis jalapa L., the
common four-o'clock of the English, is a herb with showy red, white, yellow, or
variegated flowers set in individual green calyx-like involucres. It is cultivated
in gardens and escapes as a vigorous weed in waste ground and roadsides. It is
known locally as maravilla and buenas tardes because of its vespertine blooms.

The taxonomy of the Nyctaginaceae, particularly of the woody dioecious
genera, is very confused. It seems entirely possible that interspecific hybridization
may be one factor responsible for obscuring the lines of the species. The species
bibliographies which follow are frankly provisional; to some, additional synonyms
will have to be added from the South American literature and some such names
possibly will have to take precedence over the Central American which are adopted
here. The detailed study essential to a competent treatment unfortunately is
beyond the scope of this flora.

52 ANNALS OF THE MISSOURI BOTANICAL GARDEN

1. BOERHAVIA L.
Boerhavia L. Sp. PI. 3. 1753.

Annual or perennial herbs. Leaves opposite or subopposite, petiolate, frequently
unequal at the nodes, the blade undulate or sinuate, conspicuously raphidulous.
Inflorescence terminal, paniculiform to racemiform, many-flowered. Flowers small,
perfect, subtended by 1-3 minute bracteoles, the perianth tube persistent, more or
less herbaceous, frequently glandular, the perianth limb deciduous, petalaceous,
obscurely 5-lobed; stamens 1-5, included or exserted; ovary shortly stipitate, the
stigma capitate, more or less exserted. Anthocarp obpyramidal to obovoid or
subellipsoid, glabrous to glandular, 3- to 5-angulate.

Numerous species in the tropics and subtropics of both hemispheres. Both
species in Panama are weeds and possibly

nthocarps obpyramidal,

anthocarps obovoid to subellipsoid, obtuse

Boerhavia erect a L. Sp. 3. 1753.

Boerbaavia discolor HBK. loc.

Boerhaavia atomaria Raf. Aut. Bot. 40. 1840.

Valeriana latifolia Mart. & Gal. in Bull. Acad. Brux. II 1 : 124. 1844.

Boerhaavia paniculata jS subacuta Choisy, in DC. Prodr. 13 2 :451. 1849.

Boerhavia thornberi M. E. Jones, Contr. West. Bot. 12:72. 1908.

Boerhaavia erecta thornberi (M. E. Jones) Standi, in Contr. U. S. Nat. Herb. 12:3 81.

1909.

Annual herbs; stems erect or ascending, 2-8 dm. tall, rather slender, glabrous
or essentially so. Leaves opposite or subopposite, often unequal at the nodes,
petiolate; blade rather irregularly ovate-rhombic, obtuse to rounded, or infrequently

flora of Panama (Nyctaginaceae) 53

t diffusa; insert: fruit of B. t

(395)

54 ANNALS OF THE MISSOURI BOTANICAL GARDEN

acute, at the apex, broadly obtuse to rounded at the base, undulate or sinuate, 2-9
cm. long, 1-4 cm. broad, glabrous or essentially so, yellowish green above, paler
and more or less conspicuously bracteate, glabrous or essentially so. Flowers in
pedunculate 2- to 3 -flowered dichasia, distinctly pedicellate, the perianth tube
1.0-1.5 mm. long, about 0.5 mm. broad, herbaceous, glabrous, the perianth limb
campanulate, about 0.5-0.7 mm. long, white or pink; stamens 2-3, slightly
exserted. Anthocarps obpyramidal, truncate, glabrous, 3-4 mm. long, 1.0-1.5 mm.
broad, borne upon accrescent pedicels of about equal length.

A small weed of beaches, open thickets, fields, roadsides, and wasteland;
southern United States, Antilles, Central and South America; a rare sporadic
ruderal in tropical Africa.

canal zone: Gamboa, Bro. Heriberto 28, Standley 28461; Monte Lirio, Maxon 68fo\
b- , Si' tdl y f"5 G tun, Courlt 20; 1 r j te< Ptjn i J , Nov Ir oles, Pith
6836; Balboa, Bro. Celestine IO; Parai*,, -/,'?; Monkey

Hill, OnveU 28. CHiRiQui: Puerto Armuelles, Woodson & Scbery 838. colon: Colon,
Macbride 6 Featberstone I, Piper 5892, Rose 22084.

The glabrous anthocarps and inflorescences of B. erecta have made this species
less widely dispersed as a ruderal than B. diffusa.

2. Boerhavia diffusa L. Sp. PL 3. 1753.

Boerhaavia coccinea Mill. Gard. Diet. ed. 8. Boerhaavia no. 4. 1768.

Boerhavia caribaea Jacq. Obs. Bot. 4:5. tab. 84. 1771.

Boerhavia diandra Aubl. Hist. PI. Guian. 1 :4. 1775.

Boerhaavia paniculata A. Rich. Act. Soc. Hist. Nat. Paris 1:105. 1792.

Boerhaavia polymorpba A. Rich. loc. cit. 185. 1792.

Boerhavta hirsuta Willd. Phytog. 1:1. 1794.

Boerhaavia adscendens Willd. Sp. PI. 1 : 19. 1798.

Boerhaavia viscosa Lag. & Rodr. in Anal. Ci. Nat. 4:256. 1801.

Boerhaavia laxa Pers. Syn. PI. 1:36. 1805.

Boerhavia squamata Raf. Aut. Bot. 40. 1840.

Boerhaavia sonorae Rose, in Contr. U. S. Nat. Herb. 1:111. 1891.

Boerhaavia diffusa var. hirsuta (Jacq.) O. Ktze. Rev. Gen. 533. 1891.

Boerhaavia diffusa var. paniculata (Rich.) O. Ktze. loc. cit. 1891.

Boerhaavia diffusa var. viscosa (Lag. & Rodr.) Heimerl, Beitr. Syst. Nyct. 27. 1897.

Boerhaavia viscosa forma oligadena Heimerl, in Ann. Cons. Jard. Geneve 5:189. 1901.

Boerhavia ramulosa M. E. Jones, Contr. West. Bot. 10:40. 1902.

Boerhaavia viscosa apiculata Standley, in Contr. U. S. Nat. Herb. 12:383. 1909.

Boerhaavia coccinea forma' par cebirsuta Heimerl, in Urb. Symb. Ant. 7:212. 1912.

Perennial herbs; stems decumbent to ascending, 4-12 dm. long, rather slender,
more or less viscid-puberulent. Leaves opposite or subopposite, often unequal at
the nodes, petiolate; blade orbicular to rhombic-ovate below, frequently grading
to lanceolate above, obtuse to rounded or acute at the apex, obtuse to rounded at
the base, undulate or sinuate, 2-6 cm. long, 2-5 cm. broad, more or less puberulent
to glabrate, yellowish green above, paler beneath, not punctate; petiole 1-4 cm.
long. Inflorescences paniculiform, lax and many-flowered, Ieafy-bracteate at least
below, viscid-puberulent to glabrate. Flowers in pedunculate glomerulate clusters,
sessile or subsessile, the perianth tube 0.7-1.0 mm. long, somewhat less than 0.5

(396)

flora of Panama (Nyctaghiaceae) 55

mm. broad, herbaceous, minutely glandular, the perianth limb campanulate, about
0.5 mm. long, red to purple, rarely white; stamens 3, slightly exserted. Anthocarps
obovoid to subellipsoid, obtuse to obscurely beaked, 3-4 mm. long, 1 mm. broad,
glandular, sessile or subsessde.

A weed widely dispersed throughout the tropics and subtropics of America,
Africa, Asia, and Oceania. Carasola (Taboga Isl.).

bocas del toro: Lincoln Creek, Chaguinola valley, Dunlap 377. canal zone: Fri-
joles, Wilson 5/; Empire, Hunter d Allen 786; Aneon. Gu-emuan 3 Grcemnan 3042;
between Corozal and Ancon, Pittier 2188; Ahorca Lagarto to Culebra, Cowell 369;
Balboa, Macbrid .. Waxon 6847; Gamboa, Bro. Hcriberto

29, Standley 28344, P'P e r 5&49; Gatun, Standlcy -;.'/: D.incn. Standley 31520, Mac-
bride 2684. chiriqui: Puerto Armuelles, Woodson & Scbery 820. Panama: Taboga
Wand, Pittier 3534, Killip 3177, Allen 101, Standley 27114, 27845; San Jose Island, Pearl
Archipelago, Harlow 47. san blas: Perme, Cooper 222.

Boerhavia erecta and B. diffusa frequently grow together, and hybridization
must occur occasionally but insufficiently to blur the outline of the species. In
Standley's treatments of Boerhavia for the North American Flora and the Flora of
the Panama Canal Zone, the parental species would be called B. erecta and B.
caribaea, respectively, and the putative hybrids B. coccinea. Such plants will
usually key readily to B. diffusa in the present treatment. A rather extensive
examination of herbarium specimens has revealed no tangible differences between
B. caribaea and B. coccinea of the New World and B. diffusa of the Old World.
The species appears plainly to be an effusive pantropic weed, the dispersal of which
has been greatly facilitated by the viscid anthocarps. It is probably, but not
certainly, of American origin.

2. CEPHALOTOMANDRA Karst. & Triana
Cephalotomandra Karst. & Triana, ex Karst. in Linnaea 28:429. 1856.

Dioecious trees. Leaves opposite to irregularly approximate. Inflorescences
terminal, corymbosely thyrsiform. Flowers small, immediately subtended by 1-3
minute bracteoles. Staminate flowers broadly campanulate; stamens 25-30, unequal,
included, the short tube of the filaments hypogynous and free from the sessile
pistillode. Pistillate flowers unknown. Anthocarps woody, eglandular, longi-
tudinally costate, the limb of the perianth persistent.

A single species, Panama (?) and Colombia.

1. Cephalotomandra fragrans Karst. & Triana, ex Karst. in Linnaea 28:429.

1856.
Cephalotomandra panamensis Standi, in North Am. Flora 21:179. 1918.

Dioecious trees to 25 m. tall. Branches rather stout, glabrous, with a longi-
tudinally striate, scarcely lenticellate, yellow periderm. Leaves rather sparse, long-
petiolate, the blade broadly elliptic to ovate-elliptic, Obtusely acuminate to acute,
base obtuse or rounded, 8-2 5 cm. long, 2.5-15.0 cm. broad, glabrous, firmly
membranaceous or subcoriaceous, yellowish green, venation prominent but rather

($97)

THE MISSOURI BOTANICAL GARDEN

distant; petiole 1-12 cm. long. Inflorescence 10-20 cm. long, long -pedunculate,
many-flowered, minutely puberulent, the ultimate cymules typically dichasial with
the terminal flower sessile and 1-bracteolate and the lateral definitely pedicellate
and 3-bracteoIate. Stamina te perianth broadly campanulate, yellowish, about
4 mm. long and 3 mm. broad at the conspicuously inflated, slightly spreading limb,
minutely puberulent-papillate. Anthocarps oblong-fusiform, crowned by the
persistent perianth limb, about 1 cm. long and 3 mm. broad, glabrous.

■ (?):

, Sutton Ha :

flora of Panama (Nyctaginaceac) 57

of C. fragrans (Triana Qg8) deposited in the herbarium of the Missouri Botanical
Garden and originally distributed by the British Museum. It may be merely a
peculiar coincidence, but it seems more than strange that both sheets bear a small
fruiting twig and a small twig with a staminate inflorescence in identical stages of
development, and that both the Triana specimen and that ascribed to Sutton Hayes
bear the collection number QgS. Karsten's description of CcphalotomanJra is
extremely vague in details of the pistillate inflorescence, and the account of the
genus in Bentham & Hooker's Genera Plantarum, although describing the fruit and
the staminate flowers, expressly states that the pistillate flowers were not seen by
them. I think that it is altogether probable that the New York sheet is actually
a second duplicate of Triana's original collection of C. fragrans and that its
ascription to Sutton Hayes is in error. It is doubtful, therefore, that the genus
Cephalotomandra can be included in the flora of Panama without serious question.

Neea R. & P. Fl. Per. Prodr. 52. 1794.

Neeania Raf. Princ. Som. 30. 1814.

Mitu A, Aichia Kunth, in Abh. Berl. Akad. 1831:209. 1831.

Eggerua Hook. f. in Hook. Ic. 15:1. /. 1401. 1883.

Dioecious trees and shrubs. Leaves opposite to irregularly approximate. In-
florescences terminal, thyrsiform. Flowers small, immediately subtended by 1-3
minute bracteoles. Staminate flowers urceolate; stamens 5-10, usually 8, unequal,
included, the short tube of the filaments surrounding the sessile pistillode. Pistillate
flowers urceolate or tubular-urceolate, the limb of the perianth erect or spreading,
the staminodes usually somewhat surpassing the ovary and with enlarged sterile
anthers, the pistil sessile. Anthocarps fleshy, eglandular, the limb of the perianth
withering and at length abscissing.

Numerous species in the Antilles, and Central and South America; southern
peninsular Florida.

a. Cymules strikingly dichasial, the terminal iiov.tr- -^ile, the lateral
with relatively long slender pedicels; leaves predominantly rather small
or of moderate size, and only very rarely somewhat pubescent beneath,
b. Inflorescences usually quite glabrous, the pedicels of lateral flowers as

1. Neea delicatula Standi, in Contr. U. S. Nat. Herb. 18:!
Torrubia panamemh Standi, in North Am. Fl. 21:185. 1918.

Dioecious shrubs or small trees to 6 m.
or minutely puberulent when very young,

(399)

58 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Leaves petiolate, the blade predominantly obovate to obovate-elliptic,
obtusely cuneate at the base, 4-12 cm. long, 2-4 cm. broad, firmly membranaceous,
glabrous, the petiole 2-1 5 mm. long. Inflorescences 4-7 cm. long, usually glabrous,
rarely minutely puberulent, rather few-flowered, the peduncles slender, the cymules
strikingly dichasial, the terminal flowers sessile, the lateral with slender pedicels as
long as the flowers or somewhat longer. Staminate perianths ovoid-urceolate, about
6 mm. long and 3 mm. broad, glabrous. Pistillate perianths tubular, about 3 mm.
long and somewhat less than 1 mm. broad, minutely puberulent to glabrous.
Anthocarps broadly ellipsoid or ellipsoid-obovoid, 8-10 mm. long, 4-5 mm. broad,
pink or red.

Panama, in forests and thickets near sea level to 100 m. alt.

canal zone: Balboa, Standley 32114. Panama: around Alajuela, Chagres valley,
Ptiiicr ^-2; Tumba Muerto Road, near Panama, Standley 20805; near Matias Hernandez,
Standley 28880; Rio Tapia, Standley 2820I; Taboga Island, Pittier 3602; Trapiche Island,
Perlas Islands, Allen 2422. province unknown: Seemann s. n. pro parte.

I believe it quite possible that N. delicatula and N. aniwphylla Ernst, of
Colombia and Venezuela (= N. wiemeri Heimerl) may be conspecific. The leaves
of the latter, however, appear to be somewhat larger than those of the former, as
well as of a somewhat different pattern of shape; the staminate perianths, also,
seem to be somewhat narrower than those from Panama.

The type specimen of Torrubia panamemis definitely is Neea and differs from
the other specimens cited only in the thin minute pubescence of the peduncles.

Comments on the Seemann collection are included in the discussion of Guapira
standleyana. Seemann's illustration of Pisonia pacurero (in Bot. Voy. Herald, pi
34. 1854) may have been prepared from a specimen of Neea delicatula, but is not
definitely diagnostic.

2. Neea laetevirens Standi, in Field Mus. Publ. Bot. 4:204. 1929.
Neea xanthina Standi, in Ann. Missouri Bot. Gard. 30:86. 1943.
Neea psychotrioides auct. pro parte, non Donn. Sm.

Dioecious shrubs or small trees to 7 m. tall. Branches rather slender, minutely
puberulent to glabrate, inconspicuously lenticellate at maturity. Leaves petiolate,
the blade elliptic to obovate- or oblong-elliptic, rarely lanceolate, acuminate, obtuse
to cuneate, 5-23 cm. long, 3-9 cm. broad, glabrous or rarely very inconspicuously
puberulent beneath particularly on the midrib, the petiole 0.5-4.0 cm. long. In-
florescences 8-20 cm. long, minutely puberulent, many-flowered, the cymules
strikingly dichasial with the terminal flowers sessile and the lateral with slender
pedicels as long as the flowers or somewhat shorter. Staminate flowers obiong-
urceolate, 6-9 mm. long, 2-3 mm. broad, thinly and minutely puberulent, white,
yellow, or red. Pistillate flowers tubular, the lobes erect or suberect, conspicuously
thickened at the orifice of the tube, 4-5 mm. long, 1.0-1.5 mm. broad, densely
and minutely puberulent. Anthocarps narrowly ellipsoid, 7-10 mm. long, 2-3
mm. broad, red.

Panama and Costa Rica, in moist forests and thickets near sea level.

(400)

bocas del toro: Old Bank Island, Von Wedel 1885, 1970; Little Bocas,
2533; Fish Creek Hills, Von Wedel 2433; Shepherd Island, Von Wedel 2695; Rio Crica-
mola, Woodson, Allen & Seibert 1914; region of Almirante, Cooper 340, 544, 572a; Isla
CohSn, Von Wedel 2793, 2842; Water Valley, Von Wedel 730, 979, 1678, tilO\ Changui-
nola Valley, Dunlap 94, 136, $11, 519, 562, 488, 329; Cooper & Slater 106; Farm Six,
Stork 94; vie. on, Hart 168; Von Wedel UOfj, 1 182, 1298, 1 1 45.

san blas: Perme, Cooper 639.

This is the species which has been determined almost invariably as Neea psycho-
trioides in Panama. The specimens enumerated above are relatively uniform; but
the species apparently is capable of hybridization with N. amplifolia, although such

(401)

60 ANNALS OF THE MISSOURI BOTANICAL GARDEN

specimens will usually key to the latter under the discussion of which they will
be noted.

There probably are scarcely more than a half dozen valid species of Neea in
Central America. These seem to form two vicarious series upon the Pacific and
the Atlantic slopes, recognizable by the staminate perianths which are ovoid- or
oblongoid-urceolate respectively. True N. psychotrioidcs is a representative of
the former series, as is N. delicatula of Panama. Both Panamanian species are
distinguishable from their related species to the north by their longer-pedicelled
flowers and proportionally longer lobes of the pistillate perianths. True N. psycho-
trioides extends on the Pacific slope from southern Mexico to Nicaragua and is not
yet known to occur in Costa Rica, although it might be expected in Guanacaste.

In Costa Rica, N. laetevirens has been collected at Puerto Limon, and up the
Reventazon and Parismina watersheds to the highlands of the provinces of Alajuela
and Cartago.

3. Neea ampliiolia Donn. Sm. in Bot. Gaz. 61:386. 1916.

Neea pittieri Standi, in Contr. U. S. Nat. Herb. 13:3 83. 1911.
Neea orosiana Standi, in Journ. Wash. Acad. Sci. 15:473. 1925.
Neea uropbylla Standi, in Field Mus. Publ. Bot. 4:203. 1929.
Neea pycnantha Standi, in Ann. Missouri Bot. Card. 30:85. 1943.
Neea psychotrioides auct. pro parte, non Donn. Sm.

Dioecious shrubs and small trees to about 7 m. tall. Branches moderately
stout, densely ferruginous-pubescent to glabrate. Leaves petiolate, the blade ovate-
to obovate-elliptic, usually more or less caudate-acuminate, obtuse to cuneate at the
base, 7-36 cm. long, 3-15 cm. broad, firmly membranaceous, glabrous above but
frequently more or less puberulent beneath, the petiole 0.5-5.0 cm. long.
Inflorescences 4-15 cm. long, essentially glabrous to densely ferruginous-pubescent,
rather few-flowered to many-flowered and occasionally greatly proliferous and the
flowers and bracts becoming phylloid; cymules obscurely dichasial, both terminal
and lateral flowers subsessile or with essentially similar short stout pedicels. Stami-
nate perianths tubular-urceolate, 5-10 mm. long, 2-3 mm. broad, essentially
glabrous to minutely puberulent to essentially glabrous. Anthocarps ellipsoid to
ovoid, 7-12 mm. long, glabrous to minutely puberulent, red.

Panama and Costa Rica, in forests and moist thickets, from near sea level to
about 700 m.

bocas del toro: region of Almirante, Daytonia Farm, Cooper S45> 373 1 Almirante
region, Cooper & Slater 58; Changuinola valley, Cooper & Slater 60, 140; Cricamola,
Cooper 491 ■; Fish Creek Mts., Von Wedel 2267; vicinity of Chiriqui Lagoon, Von Wedel
roil, 1017; Water Valley, Von Wedel 1 57 4, 1609, 1503, r43 3 , 927. canal zone: hills
north of Frijoles, Standley 27427, 27509; Margarita Swamp, south of France Field, Haxon
tS Valentine 7048; Gatun, Sutton Hayes 626; railroad relocation between Gorgona and
r 2265, 2271 ■; Barro Colorado Island, Frost r23, Kenoyer 491, 638, Aviles 30,
Bailey & Bailey 120, 522, Wetmore & Abbe 22, Standley 31310, 31336, 40905, 40955'
4*023, 41013. chiriqui : Progreso, Cooper & Slater 184. cocle: region north of El
Valle de Anton, Allen 3732. colon: between France Full -n ! ( *'i\ .1, standley 30182;
along Rio Fato, Pittier 3912. darien: trail between Pinogana and Yavisa, Allen 268;
forests around Pinogana, Pittier 6347. san blas: forests around Puerto Obaldia, Pittier
4397-

(402)

flora of Panama (Nyctaginaceae) 61

Neea amplifolia usually can be keyed without difficulty from the two preceding
species, particularly by means of the inflorescence characters, the predominantly
larger leaves and staminate perianths, and the more frequent and abundant indu-
ment. In all respects intermediates occur, nevertheless, and the relative abundance
of disparate specimens, particularly from Bocas del Toro, where N. laetevirens also
is abundant and variable, provides grounds for strongly suspected hybridization.

The proliferation and phyllody of the inflorescence is a frequent and puzzling
propensity of this species, the agent of which is obscure. In addition, both in
JV. amplifolia and in N. laetevirens, the terminal staminate flowers occasionally
develop at the expense of the lateral.

4. GUAPIRA Aubl.
Guapira Aubl. Hist. PI. Guian. 1:308; 3: /. Iig. \77">.
Gynastrum Neck. Elem. 1:224. 1790.
Torrubia Veil. Fl. Flum. 139. 1825; Icon. 3: /. 150. 1827, nun Tul.

Dioecious trees and shrubs. Leaves opposite to irregularly approximate. In-
florescences terminal, frequently at the tips of short lateral branches, corymbosely
thyrsiform. Flowers small, immediately subtended by 1-3 minute bracteoles.
Staminate flowers campanulate; stamens usuall} 6-8, unequal, widely exserted, the
short tube of the filaments adnate to the stipe of the pistillode. Pistillate flowers
tubular to tubular-campanulate, the limb of the perianth erect or spreading, the
staminodes about as long as the ovary and with enlarged sterile anthers, the short
tube of the filaments adnate to the stipe of the pistil. Anthocarps fleshy, eglandular,
the limb of the perianth persistent and rather fleshy.

Numerous species in the Antilles, and Central and South America.

There can be little doubt that Guapira Aubl. is congeneric with Torrubia Veil,
as suggested by Heimerl (in Engl. Nat. Pflanzenfam. ed. 2. 16 ( :127. 1934) and
not an indecipherable verbenacea as previously held by many (although questioned
by Bentham & Hooker). The specimen illustrated by Aublet obviously is pistillate
and, with the understandable error of the four to five minute "leaflets of the calyx
[i.e. bracteoles]", is a good representation of Torrubia for its time, particularly
with respect to the fruit with its persistent perianth limb ("Drupa monosperma . .
coronata denticulis corollae, quae evadit pulposa") , and the persistent style crowned
by a penicillate stigma. With the exception of the supernumary bracteoles, the
figure of the flower also is good, showing the spreading perianth limb characteristic
of so many South American Torrubias, and the widely exserted style and penicillate
stigma. The single species described by Aublet, G. guianensis, perhaps is con-
specific with Torrubia eggersiana (Heimerl) Standi.

Perhaps too many transfers from Pisonia to Torrubia already have been made
by Britton and by Standley, and one might be inclined to submit Torrubia as a
candidate for conservation over Guapira in order to avoid additional transfers to
the latter. I am inclined not to do so, however, for the following reasons. Tor-
rubia has not generally been considered as distinct from Pisonia, and consequently

(401)

62 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the transfers to the former are not in wide use; transfer to Guapira therefore would
entail little mental adjustment and inconvenience at this date. Both Torrubia and
Guapira are based upon rather inadequate antique illustrations, but those of the
latter are definitely more capable of critical interpretation than the former; it
furthermore will be much easier to associate G. guianensis with a modern species
than T. opposita Veil.

. (Standi.) Woodson, comb. nov.
Torrubia costaricana Standi, in Contr. U. S. Nat. Herb. 13:385. 1911.

Small dioecious trees to 10 m. tall. Branches moderately stout, glabrous,
prominently lenticellate at maturity. Leaves rather crowded, petiolate, the blade
predominantly elliptic to oblong-elliptic, acuminate, obtuse at the base, 2.5-10.0
cm. long, 1.5—4.5 cm. broad, subcoriaceous, glabrous, the petiole 0.5—1.0 cm. long.
Inflorescences 3-6 cm. long, rather shortly pedunculate, many-flowered, minutely

anth campanulate, about 3 mm. long, about 0.5 mm. in diameter at the base and
2 mm. at the orifice, minutely puberulent, greenish white or yellow, the stamens
about half exserted. Pistillate flowers tubular-campanulate, the limb slightly
spreading, about 2 mm. long and 0.5 mm. in diameter, minutely puberulent-
papillate. Anthocarps globose or subglobose, about 7-8 mm. long, black or purple
when ripe.

Costa Rica and Panama, in forests and savannas, sea level to 1000 m.

chiriqui: Boquete, Davidson 68q; David, Pittier 3370. cocle: Vallc <!< Anton,
Bro. Maurice 777; north rim of El Valle, Alston tf Allen J 8 '47; region north of El Valle
de Anton, trail to Las Minas, Allen 24.72.

Brother Maurice reports the vernacular names of llanto and mala sombra.
Guapira costaricana is widely distributed also on the Pacific slope of Costa Rica.
In both countries the leaves of plants growing at higher elevations seem broader
and more elliptic than those of lower elevations. The fruit, however, is relatively

2. Guapira standleyana Woodson, spec. nov.

Arbor dioica usque ca. 10 m. alta. Ramuli graciliusculi indistincte lenticellati
juventate tenue puberuli mox glabrati. Folia densiuscula sed mox decidua petiolata
lamina praecipue obovata vel obovato-elliptica brevissime acuminata vel obtusa
basi latiuscule cuneata 3-15 cm. longa 1-6 cm. lata firmiter membranacea glabra
petiolo 0.5-1.0 cm. Iongo. Inflorescentiae 4-8 cm. longae minute puberulae in
fructu paulo accrescentes floribus sessilibus vel subsessilibus. Perianthium mas-
culum campanula turn ca. 3.5 mm. longum faucibus ca. 2 mm. diam. minute

(404)

flora of Panama (Nyctaginaceae) 63

puberulo-papillatum

ca. tertio exsertis. Perianthium femineum tubu-
losum limbo erecto vel fere erecto ca. 3 mm. longum et 0.5 mm. diam. minute
puberulum. Anthocarpia oblongo-ellipsoidea 10-13 mm. longa ca. 6-7 mm. lata

[ Laboratory [Cocoli], P. White IOQ; bank of Rio

Cocoli, P. White J04. Panama: sabanas north of Panama City, Bro. Paul 467; Taboga

nd, Pittier 3544 (US, holotype; F) ; vicinity of Juan Franco Race Track, near Pana-

, Standley 27747; between Las Sabanas and Matias Hernandez, Standley 31823. san

vs: forests around Puerto Obaldia, Pittier 4316. province unknown: Seemanns.n.

64 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Sharply distinguished from G. costaricana because of the rather narrow oblong-
ellipsoid fruit. Two dupla of the Seemann collection seen in the herbarium of the
New York Botanical Garden and the Gray Herbarium consist of mixtures; the
former of Guapira standleyana and Neea delicatula, and the latter of Guapira
standleyatta, Neea delicatula, and Pisonia aculeata. This mixture may be responsible
for Seemann's remarks on the "hermaphrodite" flowers of his collection, which he
cited under Pisonia pacurero HBK., and the difficulty of distinguishing Neea from
Pisonia (Seemann, in Bot. Voy. Herald, p. 192. 1854). Seemann's plate of P.
pacurero clearly is a pistillate Neea, although not definitely N. delicatula. Seemann
cited his collection from "dark forests, near Cruces."

5. PISONIA L.
Pisonia L. Sp. PI. 1026. 1753.

Bessera Veil. PI. Flum. 147. 1825; Icon. 4: t. 12. 1841, nee Schult, nee alior.
Columella Veil. loc. cit. 155. 1825; Icon. 4: /. If. 1841, nee Comm. nee Lour.
Palladia Veil. loc. cit. 151. 1825; Icon. 4: t. 12. 1841.

Shrubs or small trees, usually scandent and armed with stout axillary spines
in our species. Leaves opposite to irregularly approximate. Inflorescences usually
terminal on highly modified short shoots, umbelliform or densely corymbiform-
thyrsiform. Flowers small, the subtending bracteoles in a more or less contracted
spiral upon the pedicel. Staminate flowers campanulate; stamens usually 6-8,
unequal, widely exserted, the short tube of the filaments adnate to the stipe of the
pistillode. Pistillate flowers tubular, the limb of the perianth apparently always
erect, the staminodes reduced to a low, occasionally glandular-dentate disc adnate
to the stipe of the pistillode. Anthocarps pentagonal-clavate, coriaceous, armed
with longitudinal rows of stipitate glands upon the angles.

Numerous species in the Antilles, Central and South America; southern penin-
sular Florida; western Africa, southeastern Asia and Oceania. A single species is
known from Panama at present.
1. Pisonia aculeata L. Sp. PI. 1026. 1753.
Pisonia villosa Poir. in Lam. Diet. 5:347. 1804.
Pisonia rVim Schlecht. in Linnaea 12:876. 1822.
Pisonia loranthoides HBK. Nov. Gen. & Sp. 7: 197. 1825.

Pisonia monotaxadenia Wright, ex Sauv. in Anal. Acad. Cien. Habana 7:199. 1870.
Pisonia grandijolia Standi, in Contr. U. S. Nat. Herb. 13:391. 1911.

Divaricately branched shrub or woody liana, usually armed with stout recurved
axillary spines, glabrous to densely pubescent. Leaves opposite to irregularly
approximate, the blade very variable in shape and size, usually obovate to obovate-
elliptic, acute to acuminate, 2-10 cm. long, 1-5 cm. broad, the petiole 1-3 cm.
long. Inflorescences usually borne at the tips of greatly condensed short shoots,
umbelliform or densely corymbiform-thyrsiform, rather shortly pedunculate, 3-6
cm. long, greatly accrescent and expanded in fruit, many-flowered. Staminate
flowers campanulate, greenish yellow, about 3 mm. long, about 2 mm. in diameter
at the orifice of the limb, minutely puberulent-papillate without, the stamens

(406)

flora of Panama (Nyctaginaceae) 65

widely exserted. Pistillate flowers tubular to somewhat urceolate, about 2.5 mm.
long and 1 mm. in diameter, the limb erect, the stigma and style widely exserted.
Anthocarps borne upon greatly elongate pedicels, pentagonal-clavate, 7-9 mm.
long, 3-4 mm. broad, green, coriaceous, the stipitate glands uniseriate upon the five
angles, the sides minutely and densely puberulent.

Very widely distributed in tropical America, and apparently a ruderal introduc-
tion in western Africa and southeastern Asia; in rain forests and moist thickets
from sea level to about 1000 m.

north of El Valle de j

Tapia, Standley 28251.

This species must

Island, Standley 4

2835.

1 de gato, in allu

(407)

ra: Seemans.n
1 frequent thai

66 ANNALS OF THE MISSOURI BOTANICAL GARDEN

BATIDACEAE

Bath maritima L. is to be expected in Panama. This dioecious herbaceous to
fruticose plant with opposite, simple, entire and sessile leaves is the sole species of
the family. It is an inhabitant of the beaches of subtropical and tropical America
and the Hawaiian Islands.

PHYTOLACCACEAE

Herbs, shrubs or trees, often vines. Leaves simple, entire, alternate, petiolate
or subpetiolate, with or without stipules. Inflorescence in terminal or extra-axillary
racemes, although occasionally axillary. Flowers bisexual or unisexual by reduction,
actinomorphic ; perianth monochlamydeous; tepals generally 4-5, free or slightly
connate at the base, usually persistent in fruit; stamens 3-many, often varying in
number within the same species, frequently borne on a hypogynous disc in 1 or
2 cycles opposite or alternate with the tepals, the filaments free or basally connate;
ovary superior, rarely inferior, 1- to 16-carpellate, the carpels free to partly or
entirely connate, the styles when present as many as the carpels, usually terminal,
; subterminal, generally free but occasionally united, the stigma capitate,
>enici!late; ovules basal, campylotropous, solitary. Fruit a berry,
e, the seeds 1-many. Embryo coiled or bent around the perisperm.

1. PETIVERIA L.
Petiveria L. Sp. Pi. 342. 1753.

Tall herbs, sometimes woody at the base. Leaves alternate, simple, entire,
petiolate, minutely stipulate. Inflorescence an elongate terminal or axillary raceme,
the flowers small, subsessile, bracteate and bracteolate. Tepals 4, united into a
short tube, the lobes subequal, linear, persistent and erect in fruit. Stamens 8,
inserted irregularly on a hypogynous disc at the base of the ovary, the filaments
filiform and of varying lengths, the anther linear, dorsifixed. Ovary 1-carpellate,

(40 B)

flora of Panama (Phytolaccaceae)

(409)

68 ANNALS OF THE MISSOURI BOTANICAL GARDEN

oblong, tomentose, 4- to 6-uncinate, the stigma 1, sessile, penicillate, decurrent
along the ventral margin of the ovary, the ovule 1, basifixed. Fruit a linear achene;
pericarp coriaceous and adherent to the seed. Seed 1 , linear, the testa membranous,
the albumen scanty and mealy, the cotyledons foliaceous.
Two species in the Western Hemisphere.

1. Petiveria alliacea L. Sp. PI. 342. 1753.
Petiveria octandra L. Sp. Pi. 486. 1762.
Petivera foetida Salisb. Prodr. 214. 1796.

Moq. in DC Prodr. i5 :: :9. 1849.
Petiveria alliacea y octandra (L.) Moq. loc. cit. 1849.
Petiveria ochroleuca Moq. loc. cit. 1849.

Petiveria paraguayemis Parodi, in Anal. Soc. Cient. Argent. 160. 1878.
Petit eria hexandria Sesse & Moc. Fl. Mex. ed. 2. 90. 1894.

Tall herbs, often woody at the base, with slender sometimes angled stems, to
5-10 dm. tall, with a strong odor of garlic. Leaves elliptic to obovate, the apex
acute or acuminate, often mucronate, the base narrowed, 5.5-16.0 cm. long, about
2-6 cm. broad, slightly pubescent; petioles 1.5-2.0 cm. long. Inflorescence in
slender usually sparsely flowered racemes 15-40 cm. long; pedicels up to 1 mm.
long. Tepals white or greenish-white, sometimes pale pink, often basally pubescent,
about 3-5 mm. long. Stamens 8, free, up to 3 mm. long. Ovary 4-uncinate, the
hooks becoming elongate and quite prominent in fruit. Fruit linear, 8 mm. long.

United States, Florida to Texas; throughout the West Indies and Central Amer-
ica; South America, Colombia to Argentina; common in dry or moist fields and
forests, often found near habitations, especially on waste ground, up to 1500
Some of the common names are: Garlic-weed, Obeah-bush, Anomu, and

Lagoon, Von Wedel 1 323; Water Valley, Von Wedel 636.
, Voodson & Schery 841, OOO; between Remedios and David,
P. White 308; Rio Dupi, Pittier 5222. canal zone: Ancon Hill, Standley 26336; Las
Cruces Trail, Government forest, Hunter & Allen 73 J, 694. Panama: Taboga Island,
Allen 1291; Standley 27838, 27025; Woodson, Allen tf Seibert 1470; Macbride 2791.
darien: Pinogana, Allen 940.

Petiveria alliacea closely resembles P. tetrandra, a species limited in geographical
distribution to Brazil. However, the two may be separated fairly readily by the
presence in P. tetrandra of 6 hook-like processes in the ovary and fruit rather than
the 4 found in P. alliacea. Furthermore, the flower pedicels in P. tetrandra are
about 5 mm. long, whereas those of P. alliacea are so short that the flowers appear

2. PHYTOLACCA L.
Phytolacca L. Sp. PI. 441. 1753.

Sarcoca Raf . Tell. 3:55. 1837.

Pircunia Moq. in DC. Prodr. 13 2 :29. 1849.

Tall perennial herbs or shrubs. Leaves alternate, simple, entire, petiol;
stipulate, generally glabrous. Inflorescence a terminal, or extra-axillary 1

flora of Panama (Phytolaccaccae) 69

70 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the pedicels basally bracteate, often with 1 or more bracteoles above. Tepals 5,
equal, usually glabrous, persistent or deciduous in fruit. Stamens 8-22, inserted
at the base of the calyx usually on a hypogynous disc, in 1 or 2 cycles, the filaments
free, the anthers dorsifixed. Ovary subglobose, 5- to 16-carpellate, the carpels
completely or only partly united, the styles equal in number to the carpels, terminal,
generally connivent, often recurved, the ovule solitary in each cavity. Fruit a
globose 5- to 16-celled fleshy berry. Seed 1 in each cavity, the endosperm mealy.
About 26 species, tropical and subtropical, mostly in the Americas, some in
Africa, Asia, and Asia Minor.

aa. Carpels completely i
erect or spreading in f
b. Carpels 6-10; sts

to 55 cm. long 3. P. RivmoiDES

1. Phytolacca rugosa Br. & Bouche, Ind. Sem. Hort. Berol. 13. 1851. Linnaea

25:297. 1852.

Woody herbs up to 2 ^ m. tall, the glabrous branches erect and angled. Leaves
lanceolate-elliptic, the apex acuminate, the base attenuate, 2-5 cm. broad, 4-14
cm. long. Racemes suberect, terminal or extra-axillary, 4-15 cm. long, the
flowers fairly crowded; pedicels about 3.5-4.0 mm. long, the bracts about 4 mm.
long, the bracteoles about 1 mm. long. Tepals oblong-elliptic, the apex rounded,
white or pink, reflexed in fruit, 2-3 (-4) mm. long. Stamens 6-12 in 1 cycle,
inserted on a subhypogynous disc, about 1.5-2.0 mm. long. Ovary subglobose,
6- to 8-carpellate, the carpels free at the apex, the styles equal to the number of
carpels and free. Fruit a purple berry, 6-8 mm. in diameter.

In Panama, appearing at 1200-2000 meters; generally somewhat higher altitudes
throughout the rest of the range, Mexico southwards to Colombia. The common
name is jaboncillo.

bocas del toro: Robalo Trail, northern slopes of Cerro Horqueta, Allen 4998. chi-
riqui: trail from Paso Ancho to Monte Lirio, upper valley of Rio Chiriqui Viejo, Allni
'.-//; Libert J02; foot of Sierra del Boquete, Maurice 742; vicinity of Oita Alta, Volcan
de Chiriqui, Woodson, Allen & Seibert 975; north forested face of Cerro Copete, eastern
spur of Volcan de Chiriqui, Allen 4871 ; vicinity of Cillcjon Seco, Volcan dc Chiriqui,
Woodson tf Schery 480; trail from Bambito to Cerro Punta, Allen 311.

P. rugosa is distinguished from P. icosandra by the incomplete union of the
carpels, the shorter racemes, and the tendency towards fewer stamens and carpels.
The free apices of the carpels may be observed equally well in flower or in fruit.
The styles of the flower are fairly close together, but are definitely free and not
connivent; in fruit they are pulled farther apart by the expansion of the seeds
within the carpels. The staminal cycle is generally considered to be the inner one,
the outer cycle having been suppressed. As in most species of Phytolacca the
number of stamens is variable, fluctuating even between flowers of the same
inflorescence.

(412)

flora of Panama (Pbytolaccaceae) 71

2. Phytolacca icosandra L. Syst. ed. 10. 1040. 1849.
Phytolacca malabarica Crantz, Inst. 2:484. 1769.
Phytolacca mexicana Crantz, loc. cit. 1769.
Phytolacca mexicana Gaertn. Fruct. 1:377. 1788.
Phytolacca triquetra Moench, Meth. Suppl. 107. 1802.
Phytolacca bogotensis HBK. Nov. Gen. et Spec. 2:183. 1823.
Phytolacca sessiliflora Kunth & Bouchc, Ind. Scm. Hort. Berol. 15. 1848 Ann. Sci. Nat.

III. 11:230. 1849.
Phytolacca acuminata Hort. ex Moq. in DC. Prodr. 13 2 :33. 1849.
Phytolacca longespica Moq. loc. cit. 1849.
Phytolacca purpurascens A. Br. & Bouche, Ind. Sem. Hort. Berol. 13. 1851. Linnaea 25:

297. 1852.
Phytolacca macrostachya Wiiid. c.\ |. A. Schmidt, in Mart. Fl. Bras. 14 2 :344. 1872.

PnU. Bot. 2:41. 1900.
Phytolacca icosandra , ar. a , - Pflanzr. 4 M :«. 1909.

Phytolacca icosandra var. sessiliflora H. Walt. loc. cit. 1909.

Stout herbs 1-2 meters tall, sometimes rather succulent, with sharply angled
and often pubescent branches. Leaves narrowly elliptic or sometimes ovate-elliptic,
the apex acute, the base narrowed and decurrent along the petiole, 6.5-15 cm.
long, 2.3-5.0 cm. broad. Inflorescence an elongate pubescent raceme of densely
crowded flowers, 16-30 cm. long; pedicels 1-1.5 mm. long; bracts lanceolate-
linear, 3-4 mm. long, the bracteoles about 0.8 mm. long. Tepals broadly elliptic
or obovate, the apex rotund, pink, white, or sometimes greenish, occasionally
pubescent, 2.5-3.5 mm. long. Stamens 8-20, inserted in 2 cycles on a disc at the
base of the perianth, about 3 mm. long. Ovary 6-8 carpellate, the carpels com-
pletely united, the styles free but connivent, often recurved. Fruit a subglobosc
purple berry 6-8 mm. in diameter. Seed shiny black.

Mexico, the West Indies, Central America south to Ecuador. Most commonly
found up to 1450 meters in Panama. The common name is jaboncilb.

bocas del toro: Changuinoh Valley, Duiilap 401 .

The taxonomic controversy over this perplexing species has probably arisen
from the variability of the distinguishing characters. The single most distinctive
consideration which separates it from P. rugosa is that of whether or not the carpels
of the ovary are entirely connate. To the unpracticed eye this is a tricky character,
but it is the most reliable. The connivent styles are the best clue to the united
carpels, as even in fruit they are closely clustered at the center of the berry in
spite of the expansion of the ovary which would tend to pull them outwards. The
remaining separative characters are best expressed as strong tendencies. The pedicels
vary in length but in the Panamanian representatives are usually about 1 mm. long.
The number of carpels may vary from flower to flower in the same inflorescence,
but 8- to 9 -carpellate ovaries are the most common. The racemes tend to be
longer than those of P. rugosa, and the flowers are often more crowded.

Walter recognizes two varieties of which one, P. icosandra var. sessiliflora occurs
in a limited range. The subsessile flowers he considers to be important enough to
afford varietal status. By "subsessile" it is intended that the pedicel is about 1 mm.
in length, as no specimens with a shorter pedicel were found. Nearly all individuals

(413)

71 ANNALS OF THE MISSOURI BOTANICAL GARDEN

examined had pedicels of 1 mm. which would place them in this category, although
the typical variety with a pedicel of about 3 mm. is reportedly widespread through-
out Panama. Herbarium sheets from adjoining countries show a gradual tendency
toward longer pedicels, and it is probable that this character is expressed by gradual

Kunth & Bouche, Ind. Sem. Hort. Berol. 15. 1848.
Ann. Sci. Nat. III. 11:231. 1849.

Phytolacca icosandra Wright, Mem. 268. 1828, non Linn. ( 1759) .
Phytolacca bogotensis Miq. Ser. exot. t. 6. 3. 1842, non HBK. (1823).
Phytolacca icosandra /3 Frascri Moq. in DC. Prodr. 13-:3 3. 1849.
Phytolacca macrostachya Willd. ex Moq. loc. cit. 1849.
Phytolacca polystigma Benth. ex Moq. loc. cit. 1849.
Phytolacca acuminata Hort. ex Moq. loc. cit. 1849.
Phytolacca polystyla Schomb. ex Moq. loc. cit. 1849.

Woody herbs or weak shrubs up to 5 m. tall, with angled branches. Leaves

the base tending to be decurrent, 4.5-7.0 cm. broad, 9-17 cm. long, the petioles
0.9-4.5 cm. long. Racemes terminal or extra-axillary, the flowers uncrowded
along the axis, 30-55 (-70) cm. long; pedicels 7-12 mm. long, the basal bract
subulate, about 1.5 mm. long with 2 bracteoles above. Tepals white to red, some-
times cream, elliptic, about 2 mm. long, deciduous in fruit. Stamens 9-14 (-22),
inserted on a hypogynous disc, about 2 mm. long. Ovary globose, 12- to 16-
carpellate, the carpels united throughout their length, the styles as many as carpels,
connivent, and recurved. Fruit a purple berry 5-6 mm. in diameter.

Throughout the West Indies; Mexico southwards to Bolivia. Found as high
as 1680 meters, but generally below 1500 meters. The common name is jaboncillo.

bocas del toro: Water Valley, Von Wedel 833, 796; Chiriqui Lagoon, Von Wedel
2079, 2631, 1246, 2460, 2704- Ftf) I 2460; Garay Creek, 2631.

chiriqui: Bajo Chorro, Boquete district, Davidson IJ3. Panama: forests near Arraijan,
Woodson, Allen (S Seibert 1392. colon: around Dos Bocas, Rio Fato Valley, Pit tier 4205.
cocle: El Valle de Anton, along Rio Indio trail, Hunter (3 Allen 313.

Although P. rivinoides appears quite distinct from P. icosandra by virtue of its
longer pedicels, deciduous tepals, more numerous carpels, smaller fruit and strik-
ingly elongate racemes, nevertheless some specimens were examined which are
suspiciously a mixture of greater or lesser degree of the two species. The most
easily recognized clue to a questionable individual is a tendency for the pedicels
to be longer than normal for P. icosandra and shorter than P. rivinoides. In com-
bination with this character are usually found intermediate tendencies for raceme
length, fruit size, and carpel number. It is interesting to note that even in plants
showing extreme tendencies towards P. rivinoides the tepals are always persistent.
Furthermore, the leaves of many of these intermediates are sometimes larger than
found in either of the two species, although this is not always true.

(Vhytolaccaccac)

3. MICROTEA Sw

Microtea Swartz, Prodr. 5 3
Schollera Rohr, in Skim, Natu
Ancistrocarpus HBK. Nov. Gen
Potamophila Schranl

. Rar. Horc. Norn. 2. t. 63. 181

Aphanathe Link, Enum. Hort. Bcrol. 1:3 8 3. 1821.

Sprawling, decumbent or sometimes spreading annual herb;
simple, entire, petiolate or subpetiolate, exstipulate. Infloresi
or terminal, racemose, shortly pedicellate, bracteate. Flowers
oblong, persistent and erect in fruit. Stamens 5, inserted
perianth, alternate with the tepals, the filament filiform, th

Tepals 5,
dorsifixed.

74 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Ovary globose, 2-carpellate, unilocular, the styles 2, basally connate, the ovule 1,
basifixed. Fruit a minute drupe, tuberculate, the pericarp adherent to the seed.
Seed 1, the testa crustaceous, the embryo bent, the albumen scanty.
About 9 species in the American tropics.

1. Microtea debilis Sw. Prodr. 53. 1788.

Scbollera debilh Rohr, in Skirvt, Naturh. Selsk. Kjoeb. 2:210. 1792.

Microtea ovata Delile, Hort. Monsp. 1827. ex Moq. in DC. Prodr. 13 L ':17. 1849.

Microtea d eh. .:. 1849.

Microtea debilis /3 rhombifolia Moq. loc. cit. 1849.

Decumbent herbs to 50 cm., the stems sharply angled. Leaves elliptic to ovate,
the apex acute, sometimes rnucronate, the base attenuate, 1.0-2.3 cm. wide, 1.3-3.6
cm. long, glabrous. Inflorescence a many-flowered raceme 1.5-3.5 cm. long; bracts
membranaceous, persistent, about 1 mm. long; pedicels about 1.0 mm. long. Tepals
5, lanceolate, white, about 0.5-0.7 mm. long. Stamens about 0.4 mm. long.
Ovary globose, about 0.5 mm. in diameter. Fruit with the tubercles united into
a honeycomb like pattern, 1.0-1.5 mm. in diameter.

Throughout the West Indies; from Guatemala southward to Peru and Brazil.

canal zone: Chagres, Fendler IOO; hills between Rio Grande and Pedro Miguel on
road to Arraij . Standby 31470; vicinity of Summit, Standley

30145, 26968; Balboa, Standley 25826; Gatiin, Bro. Ueriberto 115. bocas del toro:
Changuinola Valley, Dunlap 402. Panama: Chepo, Pittier 4457; Panama City, Bro. Paul
166; Taboga Island, Standley 27088; L .-. ■;<> 277. darien: Boca de

Cupe, Allen 883.

Microtea is a striking genus not only because of the extremely minute flowers,
but also because the flowers seem to demonstrate a very much simplified flower
plan. The relationship of Microtea to the rest of the family is obvious but it also
suggests close affinity to some members of the Chenopodiaceae, particularly to the
genus Chenopodium. Although the number of stamens are usually equal to and
alternate with the petals, occasionally 8 stamens will be found irregularly inserted.
The tubercles vary in length, and although usually rather short, may be as long

4. RIVINA L.
Rivtna L. Sp. PI. 121. 1753.
Tithona L. Syst. ed. 1. 1735.

Solanoides Moench, Meth. 307. 1794.

Tithonia L. ex O. Kuntze, Rev. Gen. 2:552. 1891, sphalm.

Tall often woody herbs, frequently shrub-like, erect or straggling. Leaves
alternate, simple, entire, petiolate, exstipulate. Inflorescence a suberect many-
flowered terminal or axillary raceme. Flowers small, pedicellate. Tepals 4, sub-
equal, obovate-oblong, rounded or pointed at the apex, persistent and erect or
spreading in fruit. Stamens 4, inserted at the base of the perianth in 1 cycle
alternate with the tepals, the filaments free, the anther dorsifixed. Ovary

(416)

i-carpellate, the style subterminal, short, curved, the stigma capi
Fruit a globose red berry, the pericarp adherent to the seed, the

Three species in the
Australia.

i tropics and subtropics, introduced into Asia and

76 ANNALS OF THE MISSOURI BOTANICAL GARDEN

1. Rivina humilis L. Sp. PL 122. 1753.

Rjt ma humilis a canescens L. Sp. PI. 122. 1753.

Khina humilis 5 glabra L. loc. cit. 175^ .

Kit in ■. lacvis L. Mant. 41. 1767.

Piercea glabra Mill. Gard. Diet. ed. S. Piercea no. 1. 1768.

Piercca t entosa Mill. G.mt. Di cJ / no 2. 1768

R/i.'/'wfl viridis Schmidt, in Mayer, Samml. Phys. Aufs. 1:185. 1791.

'.an. Bot. 6:63. 1793.
SoKr, ,h . / c, Mo i M th 307. 1794.
S. lanoiib. ■ hiei is Moench, loc. cit. 1794.
Rli hia pallida Salisb. Prod. 67. 1796.

. Salisb. Ice. cit. 1796.
Solanoides undulata Moench, Meth. Suppl. 106. 1802.
Rivina Lireeulata Willd. Enum. Hort. Berol. Suppl. 8. 1813.
Rn ma tetrandra Desfl. Tabl. ed. 2. 49. 1815.
Rii ma puberula HBK. Nov. Gen. & Sp. 2:184. 1817.

Rivina tinctoria Ham. ex G. Don. in Loud. Hort. Brit. Suppl. 1:598. 1832.
Rivina poriulei coi it > Nutt. in i Vans. Am. Phil. Soc. II. 5:167. 1837.
Rivina acuminata Raf. New Fl. 4:13. 1838, non HBK. (1817).
Piercea acuminata Raf. loc. cit. 1838.

■ ■fa Raf. loc. cit. 183 8.
Piercea obliquata Raf. loc. cit. 183 8.

Rivina canescens G. Don. in Steud. Nom. Bot. ed. 2. 2:460. 1841.
Rit imt la: i is ft acuminata Moq. in DC. Prodr. 13 2 :12. 1849.
Rivina oricntalis Moq. loc. cit. 1849.
Riiiiu hrocuwbcus Ruiz, ex Moq. 1. .,. cit. 13. 1849.
Rivina humilis a puberula (HBK.) Moq. loc. cit. 1849.
Rii hid luivul '• < i tcscem (G. Don.) Moq. loc. cit. 1849.
Rii ind humilis ' plumbaginijolia Willd. ex Moq. loc. cit. 1849.
Rivina uurantiaca Warsz. ex Schenk, Ind. Sem. Werceburg. 1861.
Rivina lacvis var. pubescent Griscb. Fl. Brit. W. Ind. 59. 1864.
Rivina vhiJiflora Be!, An.ii. Soc. F.sp. Hist. Nat. 12:105. 1883.

'arodi, in Anal. Soc. Cient. Argent. 5:206. 1878.
Tithnnia humilis O. Kuntze, Rev. Gen. 552. 1891.
Titbonia humilis var. canescens f. albiflora O. Kuntze, loc. cit. 1891.
Tithnnia humilis var. glabra O. Kuntze, loc. cit. 1891.
Rii ma bvmilh m. Publ. Bot. 2:41. 1900.

Rivina humilis var. orientalis (Moq.) H. Walt, in Engl. Pflanzenr. 4 83 :105. 1909.

Straggling shrubs or tall herbs sometimes woody at the base, up to 10 meters
high. Leaves elliptic to ovate or lanceolate, the apex acute to acuminate, the base
rounded or truncate, 2-6 cm. wide, 4-12 cm. long; petioles 0.6-11.0 cm. long.
Flowers small, drooping; pedicels 3 mm. in flower, elongating to about 7 mm. in
fruit; bracts lanceolate, about 2 mm. long; tepals 4, white or pinkish white, 2.0-
3.5 mm. long; stamens 4, inserted at the base of the perianth, about 1.5 mm. long.
Fruit a scarlet or red berry, about 4.0-4.5 mm. in diameter.

Florida to Oklahoma and Texas; throughout Mexico and southward to Argen-
tina; West Indies. In Costa Rica found up to 1040 meters elevation. Particularly
common on waste ground, often on coastal rocks. Among the common names are
Wild Tomato, Bloodberry, and cartnin.

bocas DEL toro: region of Almirante, Cooper 77; location not stated, Carleton $2;
vicinity of Nievecita, Woodson 6 Schery 1023; Woodson, Allen (3 Seibert 1 807; Water

(418)

flora of Panama (Vhytolaccaceae) 77

Valley, Von Wedel 609; Chiriqui Lagoon, vicinity of Little Bocat, Vom Wtiti 2>20;
vicinity of Chiriqui Lagoon, Big Bight, Von Wedel jSSj. ruihiQui: v.illcv of upper Rio
( hiriqui Yiejo, P. .; C. V ■''•:/,■ SS. ( oc 1 1 : mountains bvvond 1 .. I'mt.uh. Hunter tf Allen
579- darien: near mouth of Rio Yapd, Allen .?.,V>; vicinity of Boo Jc Cu P c, .4//<« ,Wu.
Walter recognizes three varieties of Rivina humilis based on glabrity or degree
of pubescence. R. humilis var. orientalh is in Asia, presumably escaped from culti-
vation as the genus is probably not indigenous to the region. R. humilis var.
canescens is recognized by the tomentose stems and densely pilose leaves and petioles,
and is cited as found in Brazil and the lesser Antilles. R. humilis var. glabra is, as
the names implies, completely glabrous, and is widely distributed from Texas to
Argentina. Use of pubescence as a distinguishing varietal character is unconvinc-
ing; consequently the specimens examined have been treated as a single species with
a wide range of variation in the degree of relationship between pubescence and
glabrity of any part of a plant. The synonymy has likewise been regarded as
pertaining to a single species.

5. TR1CHOSTIGMA A. Rich.
Trichostigma A. Rich, in Sagra, Hist. Cuba 10:306. 1 845.
Villamilla R. & B. ex Benth. & Hook. Gen. Pi. 3:81. 1880.

Woody vines, shrubs. Leaves alternate, simple, entire, petiolate; stipules minute
and deciduous. Inflorescence a many-flowered terminal or extra-axillary raceme,
the bracts deciduous. Tepals 4, subequal, persistent and reflexed in fruit. Stamens
8-10 (—12), inserted on a hypogynous disc, the filaments free, the anther dorsi-
fixed. Ovary 1-carpellate, unilocular, subglobose, the stigma sessile or subsessile,
often penicillate, the ovule 1. Fruit a globose drupe, the pericarp adherent to the
seed. Seed 1, the testa crustaceous.

Three species in tropical America.

1. Trichostigma octandrum (L.) H. Walt, in Engl. Pflanzenr. 4 8S :109. 1909.

Rivina humilis p scan Jens L. Sp. PI. 122. 1753.

Rivina octandra L. Cent. PI. 2:9. 1756.

}Rivina dodecandra Jacq. Obs. Bot. 1 :6. 1764.

Rivina scandens Mill. Gard. Diet. ed. 8. Rivinia no. 2. 1768.

Rivina Mutisii WillJ. ex Schult. Mant. 3:305. 1827.

Rivina americana Raf. Fl. Tell. 3:56. 1837.

Trichostigma rh inoides A. Rich, in Sagra. Hist. Cuba. 10:306. 1845.

Rivina octandra £ obtusifolia Moq. in DC. Prodr. 1 3 2 : 1 1 . 1849.

Rivina Ehrenbergiana Klotzsch, ex Moq. loc. cit. 1849.

Rivina Moritziana Klotzsch, ex Moq. loc. cit. 1849.

Villamilla octandra Hook. f. in Benth. & Hook. Gen. Pi. 3:81. 1880.

Decumbent or suberect shrubs or woody vines up to 10 meters. Leaves oblong
to elliptic, the apex acute to acuminate, the base rounded or acute, 3.8-5.5 cm.
broad, 12.4-14.7 cm. long; petioles 2.4-3.7 cm. long, glabrous. Inflorescence a

(419)

78 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Fig. 131. Trichostigma polyandrum

flora of Panama (Phytolaccaceae) 79

fairly dense many-flowered raceme, about 5.0-6.5 (-11.0) cm. long; pedicels 3-9
mm. long; bracts lanceolate, 2 mm. long, deciduous, the bractlets triangular, 0.5
mm. long. Tepals white to whitish-green, 2-3 mm. long. Stamens 8-12, inserted
irregularly at the base of the perianth, 3-4 mm. long. Fruit a subglobose drupe,

United States (southern peninsular Florida and the Keys); West Indies;
throughout Mexico and Central America; South America (Venezuela to Argen-

canal zone: drowned forest of upper Rio Pequeni between Salamanca Hydro-
graphic Station and Rio Boqueron, Allen 17279. darien: Tucuti, M. E. tf R. A. Terry
1396.

The Terrys have stated on the specimen label that T. octandrum is a tree,
but there are no other indications that trees are even found in this genus.
T. octandrum occurs as a woody shrub or liana sprawling over adjacent shrubs

2. Trichostigma polyandrum (Loesener) H. Walt, in Engl. Pflan

1909.
Rivin,
Villa*

Weak shrubs up to 5 meters, or woody vines climbing on shrubs. Leaf blades
elliptic to ovate, the apex acute to attenuate, the base rounded or acute, 7.0-16.8
cm. long, 3.5-6.6 cm. broad, the petioles 1.0-3.0 cm. long, glabrous. Inflorescence
a long uncrowded raceme, about 8.0-21.0 cm. long; pedicels 8-16 mm. long;
bracts lanceolate, 1.5-2.0 mm. long, deciduous, the bractlets about 0.5 mm. long.
Tepals white in flower, red, purple or purplish-pink in fruit, 4.0-9.0 mm. long.
Stamens 20-25 inserted at the base of the perianth. Ovary subglobose, the style
very short. Fruit a drupe, purple or red, 4.5-5.5 mm. in diameter.

Central America (Nicaragua, Costa Rica and Panama).

bocas del toro: Pumpkin River near Chiriqui Lagoon, Von Wedel 2 57 1; Water
Valley, Von Wedel 926, 1439, 714, 1547, 940; Woodson, Allen & Scibert 1832; Almirante,
Cooper 133. Panama: Rio Juan Diaz above Juan Diaz, Allen 944.

Trichostigma appears to have close affinities with Rivina. However the two
may easily be distinguished by the more numerous stamens, the tepals reflexed in
fruit, the terminal stigma, the deciduous stipules, and the drupaceous fruit of
Trichostigma. The genera have been treated as one genus by some authors, but
recently they have been regarded as separate.

80 ANNALS OF THE MISSOURI BOTANICAL GARDEN

AIZOACEAE

By LORIN I. NEVLING, Jr.

Annual or perennial herbaceous or suffruticose plants, prostrate or upright,
often succulent. Leaves usually simple, alternate, opposite or pseudoverticellate,
often fleshy, sometimes reduced to scales, entire; stipules scarious or absent. In-
florescences axillary or terminal, modified cymes or the flowers solitary. Flowers
bisexual, polygamodioecious or unisexual, actinomorphic, often small. Perianth
monochlamydeous, usually 4- or 5 -parted, free or connate, sometimes appendaged
on the outer surface, often persistent in fruit. Stamens (3), 4, 5 or many, the
outermost often sterile and petaloid (but not in our species), the filaments free
or variously connate at the base into fascicles or into a monadelphous sheath, free
or adnate to the perianth, the anthers oblong or linear, small, dehiscing longi-
tudinally. Pistil 1, the ovary superior to inferior, 1-5 (-20) loculate, the placenta-
tion axile, parietal or basal, the ovules solitary to many per locule, anatropous or
campylotropous, the styles as many as the locules. Fruit a loculicidal or circum-
scissile capsule or indehiscent and either baccate or nut-like; seed with mealy
endosperm, sometimes strophiolate, the embryo curved.

A large weedy family particularly well-developed in South Africa. Three
genera are represented in Panama, each by a single species. A fourth genus, Glinus,
with a single species [G. radiatus (Ruiz & Pav.) Rohrb.] has been reported from
Panama. I have been unable to locate a voucher for this record but have included
the genus in the key to genera.

This study is based primarily on the work of F. Pax and K. Hoffman (in Natur-
Pflanzenf. 16 c :179-233. 1934) and P. Wilson (in North Amer. Fl. 21:267-277.
1932).

distinct to the base, without appendages;
spatulate; flowers sessile or short-pedicellat

-eaves opposite, fleshy; tepals connate, with appendages on i
urface beneath the apex; fruit a circumscissile capsule,
oculate; seeds estrophiolate.
. Leaves obovate to rounded-spatulate, strikingly unequal;

1. MOLLUGO L.
Mollugo L. [Gen. 356. 1737]; Sp. Pi. 89. 1753.
Galiastrum Heist, ex Fabricius, Enum. PI. Hort. Helmst. 1
Lampetia Raf. Fl. Tellur. 3:34. 1836.

(422)

FLORA OF PANAMA (AtZOaceae)

Herbaceous or suffruticose plants, glabrous throughout. Leaves simple, basal or
pseudoverticellate, generally linear; stipules deciduous. Inflorescences axillary,
cymose. Flowers bisexual, pedicellate. Perianth 5 -parted, distinct to the base,
imbricate, scarious at the margins, lacking appendages. Stamens 3-10, united at
the base. Ovary superior, 3- to 5-loculate, the ovules campylotropous, numerous
on the axile placenta. Capsule loculicidal; seeds generally numerous, small,
estrophiolate.

About 2 5 species in the temperate and tropical regions of both hemispheres.

L. Sp. PI.

I7«m

Pharnaceum hoffmannseggianum Roem. & Schult. Syst. Veg. 6:692. 1820.
Mollugo dichotoma Schrank, PI. Rar. Hort. Monac. /. 64. 1821.
Mollugo arenaria HBK. Nov. Gen. & Sp. 6:20. 1823.

Mollugo schrankii Ser. loc. cit. 391. 1 824.

Pharnaceum verticellatum (L.) Spreng. Syst. 1:949. 1825.

Pharnaceum arenarium (HBK.) Spreng. loc. cit. 1825.

Mollugo juncea Fenzl, in Ann. Wien Mus. 1:378. 1836.

Mollugo diffusa Willd. ex Fenzl. loc. cit. 377. 1836, nom. nud. in syn.

Mollugo spergulaefolia Willd. ex Fenzl, loc. cit. 378. 1836, nom. nud. in syi

Mollugo triphylla Schrank, ex Steud. Nom. ed. 2. 2:154. 1841, nom. nud. L

Mollugo gracillima Anders, in Vet. Akad. Handl. Stockh. 1853:226. 1855.

Pharnaceum cerviana Mart, ex Rohrb. in Mart. Fl. Bras. 14 2 :241. 1872, no

Mollugo axillaris Schlecht. ex Rohrb. loc. cit. 242. 1872, nom. nud. in syn.

82 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Annuals, prostrate to ascending, often falsely dichotomously branched. Leaves
pseudoverticellate, 3-6 per node, unequal, generally oblong-linear, 1-3 cm. long,
1-2 mm. broad, more or less acute at the apex, gradually tapered to the base; sessile.
Inflorescence axillary, modified cymose; pedicel 3-5 mm. long. Flowers with the
tepals elliptic, about 2.0-2.5 mm. long, 0.5-0.75 mm. broad, persistent and
subtending the fruit, glabrous; stamens 3, the filaments filiform, to 2 mm. long,
connate at the very base, the anthers ovoid, about 0.25 mm. long and broad; ovary
ovoid, 3-loculate, 1.0-1.5 mm. long, the styles 3, 0.5-0.75 mm. long, spreading.
Capsule ovoid, 2-3 mm. long, about 1.5 mm. in diameter; seeds numerous, reni-
form, about 0.5 mm. long and broad, the testa with several distinct ridges along
the back and sides, brown.

Cosmopolitan weeds in the temperate and tropical regions of the world. This
species has a rather large and complex synonomy involving numerous infraspecific
categories which are not included here in the synonymy.

canal zone: Chagres, Fendle r IT. Panama: San Jose Island, Johnston 1 232.

2. TRIANTHEMA L.

Trianthema L. Sp. PI. 223. 1753.

Rf/w? Adans. Fam. 2:245. 1763.

Zaleya Burm. f. Fl. Ind. 110, t. 31. 1768.

Papularia Forsk. Fl. Aegypt.-Arab. 69. 1775.

Rocama Forsk. loc. cit. 71. 1775.

Portulacastrum Juss. ex Medic. Phil. Bot. 1:99. 1789.

Zallia Koxh.Fl Ind. 3:74. 1832.

Ancistrostigma Fenzl, in Ann. Wien Mus. 2:293. 1839.

Kacoma Willd. ex Steud. Norn. ed. 2. 2:429. 1841.

Zaleia Steud. loc. cit. 795. 1841.

Pomatotheca F. Mull. Fragm. 10:72. 1876.

Herbaceous or suffruticose plants, branched, upright or prostrate. Leaves
simple, opposite, strikingly unequal, fleshy, obovate to rounded-spatulate, the
petioles connate into a petiolar sheath surrounding the stem; stipulate. Inflores-
cences axillary, cymose or the flowers solitary. Flowers bisexual, sessile or pedicel-
late. Perianth 5 -parted, connate below, the lobes appendaged on the outer surface
beneath the apex, imbricate. Stamens 5 or more, inserted on the perianth tube
or free and monadelphic. Ovary superior, 1- or 2-loculate, the ovules campylo-
tropous, few on the axile placenta, the styles 1 or 2. Capsule circumscissile; seeds
few, reniform, the testa often wrinkled, estrophiolate.

About 15 species of the tropics and subtropics.

1. Trianthema portulacastrum L. Sp. Pi. 223. 1753.
Trianthema monogynum L. Mant. 69. 1767.
Trianthema procumbens Mill. Gard. Diet. ed. 8. no. 1. 1768.
Portulacastrum monogynum (L.) Medic. Phil. Bot. 1:99. 1789.
Trianthema flexuosa Schum. & Thonn. Beskr. Guian. Pi. 241. 1828.

(424)

FLORA OF PANAMA (A'lZOCtCeac)

Herbaceous annuals, erect or prostrate, the young stems sometimes with a
decurrent line of hairs from the interpetiolar stipules to the node below and
glabrescent, often alternately branched by the development of the axillary bud
subtended by the smaller leaf of the nodal pair. Leaves of any pair strikingly
unequal, the larger at least twice as large as the smaller, obovate to rounded-
spatulate, 1-3 cm. long, 0.5-2.0 cm. broad, acute, obtuse, refuse and often
mucronulate at the apex, cuneate at the base, glabrous above and below; petiole
0.3-1.5 cm. long, at the base connate into a sheath surrounding the stem, the
interpetiolar stipule deltoid, 1-2 mm. long, remotely serrate or entire. Inflorescence
ssile. Flowers with the perianth tube campanulate,
iated with the petiolar sheath of the subtending leaves, the
n. long, 1 mm. broad, glabrous, appendage horn-like,
, long; stamens 10, inserted at the orifice of the perianth
filiform, about 1.5 mm. long, glabrous, the anthers ovoid,
about 0.5 mm. long and broad; ovary turbinate, about 1 mm. long, 1 mm. in
diameter, truncate and irregularly fleshy-lobed at the apex, glabrous, the style 1,
about 1 mm. long. Capsule almost enclosed in the petiolar sheath, turbinate, about

(425)

84 ANNALS OF THE MISSOURI BOTANICAL GARDEN

4.5 mm. long, 3.5 mm. in diameter, sessile, truncate and crested at the apex,
circumscissile at about the middle; seeds 2-5, reniform, 1.5-1.75 mm. in diameter,
the testa wrinkled, reddish-brown to black.

A pantropic species.

canal zone: Fort Clayton, Standley 29003.

3. SESUVIUM L.

Sesuvium L. Syst. ed. 10. 1058. 1759.

Squibbia Raf . New Fl. 4 : 1 6. 1838.
Diplochonium Fenzl, Nov. Stirp. Dec. 57. 1839.
Pyxipoma Fenzl, in Ann. Wien Mus. 2:293. 1839.
Psammanthe Hance, in Walp. Ann. 2:659. 1852.
Halimus O. Ktze. Rev. Gen. 1:263. 1891.

Herbaceous or suffruticose plants, upright or prostrate, sparsely branched.
Leaves simple, opposite, more or less equal, fleshy, generally linear, elliptic or
narrowly obovate, the petioles clasping or connate and surrounding the stem;
estipulate. Inflorescence axillary or rarely terminal, glomerules or the flowers
solitary. Flowers bisexual, sessile or pedicellate. Perianth 5 -parted, connate below,
the lobes generally appendaged on the outer surface beneath the apex, imbricate.
Stamens 5, free and alternitepalous, or numerous, and inserted on the perianth tube.
Ovary superior, 3- to 5-loculate, the ovules campylotropous, many on the axile
placenta, the styles 3 to 5. Capsule circumscissile; seeds numerous, reniform, the
testa smooth, estrophiolate.

Six to eight species, tropical and subtropical strand plants of both hemispheres.

1. Sesuvium portulacastrum (L.) L. Syst. Nat. ed. 10. 1058. 1759.

Portulaca portulacastrum L. Sp. PI. 446. 1753.

Sesuvium revolutt folium Ortega, Dec. 19. 1797.

Sesuvium ortegae Spreng. Bot. Gart. Halle Nachtr. 1:36. 1801.

Sesuvium pedunculatum Pers. Syn. PI. 2:39. 1806.

Sesuvium sessile Pers. loc. cit. 1806.

Sesuvium sessiliflorum Domb. ex Rohrb. in Mart. Fl. Bras. 14 2 :3 10. 1872, nom. nud. in syn.

Halimus portulacastrum (L.) O. Ktze. Rev. Gen. 1:263. 1891.

Herbaceous perennials, glabrous throughout, the branches trailing and often
rooting at the nodes, sometimes rather stout. Leaves linear, elliptic or narrowly
obovate, 1-5 cm. long, 0.2-1.0 cm. broad, subacute to acute at the apex, gradually
tapered to the base; petiole 1-5 mm. long, dilated at the base and clasping the
stem but the sheaths not connate. Inflorescence axillary, the flowers solitary, pink;
pedicel 2-11 mm. long. Flowers with the perianth tube obconic to subcampanu-
late, 1.5-3.0 mm. long, the lobes ovate, 4-7 mm. long, 3.0-4.5 mm. broad, some-
what auriculate at the perianth tube orifice, persistent and somewhat reflexed in
fruit, appendage horn-like, 1.0-1.5 mm. long; stamens numerous, inserted at the
orifice of the perianth tube, free or subconnate, the filaments 1.5-3.5 mm. long,
gradually tapering to the apex, the anthers oblong, 0.5-0.75 mm. long, 0.25-0.5
mm. broad; ovary ovoid to subglobose, 3.0-3.5 mm. long, 2.5-3.0 mm. in diameter,

(426)

flora of Panama (Portulacaceac) 85

1.5-3.5 mm. long spreading <

umscissile below the i

8-30, lenticular-renifo
Circumtropical.

bocas del toro: bar mouth, Changuinola Valley, Dunlap 139, Stork 139.
zone: Cristobal, Artamanofi s. n.; Balboa, Standby 25625, 30882; railroad and do
at Balboa, Macbriie & Featherstone 24; vicinity of Fort Sherman, Standley 31180.
vicinity of Palenque, Pittier 41 18. Panama: Pearl Islands, Johansen III; Sabog;
Allen 2634, Miller 1965; near Punta Paitilla Military Reservation outskirts of Panai
Bro. Maurice 754; San Jose Island, Erlanson 173. san blas: forests around
Obaldia, Pittier 4400.

PORTULACACEAE

By LORIN I. NEVLING, Jr.

Herbs, subshrubs or shrubs, more or less succulent, often glabrous. Leaves
alternate, opposite or in basal rosettes, simple, terete, subterete or plane, often
fleshy, pinnately veined; stipules scarious, fimbriate, tufted or rarely absent. In-
florescence terminal or axillary, paniculiform, racemose, cymose or the flowers
solitary. Flowers bisexual, often insignificant. Sepals (involucxal bracts) 2 or

86 ANNALS OF THE MISSOURI BOTANICAL GARDEN

rarely 4-8, connate at the base or free, persistent or caducous, scarious or herba-

or free, deciduous or rarely calyptrate, imbricate. Stamens often as many as the
petals and opposite them, sometimes fewer or more, free or basally adnate to the
corolla, the filaments filiform, the anthers 2 -celled, dehiscing longitudinally,
introrse. Pistil 1 ; ovary superior to inferior, becoming uniloculate, the placentation
central, basal, the ovules (1-) 2-many, campylotropous; styles and stigmas 1-9,
united below or free. Fruit capsular, circumscissily or loculicidally dehiscent,
rarely indehiscent and nut- like; seeds 1-many, generally reniform-round, com-
pressed, the embryo curved, endosperm mealy.

A family of weedy plants composed of 15-25 genera. Only 2 genera are
represented in Panama.

or estipulate, the stipules scarious,

1. TALINUM Adai
Talinum Adans. Fam. 2:245. 1763.

Krides Medik. Phil. Bot. 1:95. 1789.
Phemeranthus Raf. in Med. Repos. N. Y. 5:3 50. 1808.
Talinium Raf. in Amer. Monthly Mag. 175. 1818.
Litanum Nieuwl. in Amer. Midi. Nat. 4:90. 1915.

Herbs, subshrubs or shrubs, often succulent. Leaves alternate or approximate,
terete to plane, fleshy; estipulate. Inflorescence terminal or rarely axillary, paniculi-
form, cymose or the flowers solitary. Sepals 2, opposite, deciduous or rarely
persistent. Petals 5, rarely more, connate at the base or free, deciduous. Stamens
5-30, in antipetalous fascicles. Ovary superior, sessile or short-stipitate, the ovules
numerous, the styles 3, more or less united, filiform. Capsule loculicidal, 3-valved,
splitting from the apex to the base, chartaceous; seeds reniform-round, numerous,
the embryo incompletely annular, the testa smooth, striate or tuberculate, distinctly
or indistinctly strophiolate.

A genus of about 50 species of which only 1 is found in Panama.

1. Talinum paniculatum (Jacq.) Gaertn. Fruct. 2:219. 1791.

Portulaca paniculata Jacq. Enum. PI. Carib. 22. 1760.
Portulaca patens L. Mant. 242. 1771.
Koelingia patens (L.) Ehrh. Beitr. 3:135. 1788.
7 tlinum r« ; v«;,. Civ. k. 1:1. 1791.

1 chotomum Ruiz & Pav. Syst. Veg. 118. 1798.
Talinum patens (L.) Willd. Sp. Pi. 2:863. 1800.
Portulaca reftexa (Cav.) Haw. Misc. Nat. 141. 1803.
Talinum fruticosum Macfad. Fl. Jam. 2:169. 1850.
Talinum sarmentosum Englm. in Bost. Journ. Nat. Hist. 6:153. 1850.
Talinum spathulatum Englm. in Gray, PI. Wright. 1:14. 1852.
Talinum purpureum Hort. ex Englm. loc. cit. 1852.

(428)

Ii s . 13S. Talin

Mch, ex Rohrb. in Mart. Fl. Bras. 14*:295. 1872.
Talinum roseum Klotzsch, ex Rohrb. loc. cit. 1872.

Talinum patens var. sarmentosum (Englm.) Gray, in Proc. Amer. Acad. 22:275. 1887.
Claytonia patens (L.) O. Ktze. Rev. Gen. 1:56. 1891.
Claytonia paniculata (Jacq.) O. Ktze. loc. cit. 57. 1891.
Claytonia reflexa (Cav.) O. Ktze. loc. cit. 1891.
Claytonia sarmentosa (Englm.) O. Ktze. loc. cit. 1891.

Talinum chrysanthum Rose & Standi, in Contr. U. S. Nat. Herb. 13:288. 1911.

Talinum panicuhium var. wmentosum (Englm.) v. Poellnitz, in Deut. Bot. Gessel.

Herbaceous plants, erect, slender, glabrous throughout, the root often thick
and fleshy. Leaves alternate, broadly elliptic to obovoid, rarely oblanceolate, 4-13
cm. long, 1.5-5.5 cm. broad, acute to obtuse at the apex, cuneate or attenuate at
the base, plane, petiole 3-15 mm. long. Inflorescence terminal, compound, lax,
10-60 cm. long, 6-20 cm. broad, the lateral simple or compound cymes racemosely
arranged on the primary rhachis, bracteate; pedicel 6-25 mm. long, dilated distally.
Sepals ovate, 1.5-3.0 mm. long and broad, connate at the base, slightly keeled,
reflexing and deciduous; petals elliptic, 3.0-4.5 mm. long, about 1 mm. broad;
stamens 15 or more, the filaments 1.5-2.0 mm. long, free, the anthers ovoid, to
0.5 mm. long and broad; ovary sessile, globose, about 1 mm. long, 1 mm. in
diameter. Capsule ovoid to globose, 3-5 mm. long, 3-5 mm. in diameter; seeds

(429)

88 ANNALS OF THE MISSOURI BOTANICAL GARDEN

strongly reniform, 0.75-1.0 mm. in diameter, the testa minutely striate, black,
indistinctly strophiolate.

Southern United States, Central and South America, West Indies, southern
Asia and the East Indies.

Panama: Isla Taboga, Woodson, Allen & Seibert 1491.

2. PORTULACA L.

Portulaca L. [Syst. ed. 1. 1735]; Sp. Pi. 445. 1753.

Meridiana L. f. Suppl. 248. 1781.

Lemia Vand. Fl. Lus. & Bras. Sp. 3 5, /. 2. f. 75. 1788.

Merida Neck. Elem. 2:3 82. 1790.

PortulaccaHzw. Synop. 121. 1812.

Lamia Vandelli ex Endl. Gen. 949. 1 840.

Annual or perennial herbs, succulent. Leaves alternate (
to plane, often whorled about the flowers; stipules scarious, fimbriate or tufted,
sometimes very small, rarely absent. Inflorescence near the stem apex, crowded,
the flowers sometimes solitary, yellow to red. Sepals 2, opposite, connate below,
the anterior larger than the posterior, usually persistent. Petals 4-6, connate at the
base or free, somewhat auto-deliquescent. Stamens 4 to many, inserted at the base
of the petals, the filaments often hairy below. Ovary half to wholly inferior, the
ovules numerous, the styles 3- to 9-parted, rarely simple. Capsule circumscissile,
chartaceous; seeds reniform to cochleate, numerous, the testa smooth or minutely
tuberculate.

enus in tropical and subtropical areas of the world, about

Portulaca parvifolia Haw. Syn. Pi. Succ. 122. 1812.

Portulaca marginata HBK. Nov. Gen. & Sp. 6:58. 1823.

Portulaca oleracea p parviflora (Haw.) Griseb. Fl. Brit. W. Ind. 57. 1864.

Portulaca oleracea a macrantha Eggers, Fl. St. Croix 27. 1879.

Portulaca oleracea p micrantha Eggers, loc. cit. 1879.

Plants herbaceous, radially spreading and prostrate or somewhat ascending,
glabrous throughout, the roots fibrous. Leaves alternate, elliptic to obovoid, 1-3
cm. long, 0.5-1.0 cm. broad, generally obtuse at the apex, cuneate at the base,
plane; petiole 1-8 mm. long, the stipules fimbriate, inconspicuous. Inflorescence
with the flowers crowded, the flowers sessile or essentially so, yellow. Perianth tube
crateriform, 1 mm. long, 1.5 mm. in diameter; sepals ovate, 3.0-4.5 mm. long,
3-4 mm. broad, strongly keeled, connate at the base, generally persistent and
enclosing the operculum of the fruit; petals 4, 3.0-4.5 mm. long, 1.5-3.0 mm.
broad; stamens 6-15, the filaments 1.5-1.75 mm. long, the anthers globose, to
0.5 mm. long and broad; ovary half -inferior, short-conical, the style lobes 4-6.
Capsule ovoid to fusiform, 6-9 mm. long, about 2.5 mm. in diameter, circumscissile

(430)

flora of Panama (Portulacaceae)
slightly below the middle; seeds almost cochleate, about 0.5-0.75 i

Known as verdolaga in P:
Bot. 18:429. 1937) the plai

ding to Standley (Field Mus. Nat. Hist,
pot herb throughout Central America.

(431)

CARYOPHYLLACEAE

By JAMES A. DUKE

Annual or perennial herbs or subshrubs, often with swollen nodes. Leaves
opposite and decussate or verticillate, the petioles often amplexicaul, some with
conspicuous stipules. Flowers commonly bisexual, actinomorphic, 5- or 4-merous,
in dichasial cymes, or solitary in the axils, or solitary and terminal (Githago).
Sepals 5-4, separate or connate, often scarious, persistent. Petals as many as the
sepals, rarely absent, hypogynous or slightly perigynous, entire, bifid or lacerate,
usually white, pink or red. Stamens 2-10, mostly 5 or 10; filaments filiform or
flattened, occasionally united into a short tube below; anthers usually versatile,
2-celled, and longitudinally dehiscent. Carpels 2-5, the ovary 1-locular, with 1-
many campylotropous ovules on basal, central or free central placentae, styles 2-5,
free or united below. Fruit a utricle or a capsule longitudinally dehiscing into
as many entire or deeply emarginate valves as there are styles; seeds 1-many,
strophiolate or estrophiolate, smooth or tuberculate, the embryo curved about the

This family, with about 75 genera, mostly in temperate areas, is of little

Dianthus, which may ultimately be found in Panama gardens. Several weedy
genera, of more or less cosmopolitan distribution, despairingly need monographic
research, worldwide in scope. Five such weedy genera occur in Panama.

a. Leaves longer than broad; stipules absent; stamens usually 10, rarely 4
or 5; capsule splitting into 2-5 usually 2 -cleft valves; styles not united.

Leaves ovate or cordate; capsules ovoid, straight, with (

ves shorter or longer than broad; stipules present, o
to 3 entire val
ally united at least half their length,
-eaves broader than long, opposite, the ochraceous stip

'ugaceous; seeds 1-50, tuberculate in lines (in Panama)

.eaves longer than broad, verticillate, the argentate sti
istent; seeds 2-6, merely corrugated

1. ARENARIA L.

Arenaria L. [Rupp. ex L. Syst. ed. 1. 1735] Sp. PI. 423. 1753.

Moehringia L. [Syst. ed. 2:22. 1740] Sp. Pi. 359. 1753.

Uinuartia L. Sp. PI. 89. 1753

Gypsophytum Adans. Fam. 2:256. 1763.

Ammodenia Patrin, ex Gmel. Fl. Sibir. 4:160. 1769.

Alsine Scop. Fl. Cam. ed. 1:496. 1772. not L. 1753.

Honkenya Ehrh. Beitr. 2:180. 1788.

"irh. loc.cit. 4:147. 1789.

flora of Panama (Caryopbyllaccac)

Spergulastrum Michx. Fl. Bor. Am. 1:275. 1803. in pare.

Gouffeia Robill. '& Cast, in Lam. & DC. Fl. Fr. 5:609. 1815.

llalunthus Pries, Fl. Holland 75. 1817.

Adenarium Raf. in Dcsv. Jour. Phys. 89:249. 1818.

Bigelovia Raf. in Jour. Phys. 89:289. 1819.

Siebera Hoppe, in Flora 2:24. 1819.

Somerauera Hoppe, loc. cit. 2:26. 1819.

Merckia Reichb. ex Cham. & Schlecht. in Linnaca 1:59. 1826.

Brachystemma D. Don, Prodr. Fl. Ne P . 216. 182 8.

\\ //Vv/w ,u Reichb. Consp. 206. 1 828.

OJiHitostt mma Benth. in Wall. Cat. ;;. 64$. 1828.

Sabulina Reichb. Fl. Germ. Excurs. 78 5. 1832.

Eremogone Fenzl, Verbr. Alsin. 1 3 : 1 8 . 183 3.

Chetrophh Raf. Fl. Tell. 3:80. 1836.

Dolophragma Fenzl, in Ann. Wien. Mus. 1:63. /. ". 1836.

Merkia Reichb. Handb. 298. 1837.

Dufou rea Gren. in Act. Soc. Linn. Bord. 9:25. 1837.

/ />•> rodiclh Fenzl, in Fndl. Gen. 966. 1 840.

M'.iininthf Reichb. R. ! 1. Germ. 5:29. 1841.

facchinia Reichb. loc. cit. 5:29. 1841.

Neumayera Reichb. loc. cit. 5:30. 1841.

Pcttera Reichb. Norn. 205. 1841.

Tryphane Reichb. loc. cit. 205. 1841.

Wicrzbickia Reichb. loc. cit. 205. 1 84 1 .

Greniera J. Gay, in Ann. Sci. Nat. 3 ' :27. 1845.

Rbodahine J. Gay, loc. cit. 3 4 :25. 1845.

St, ■■:■:■■„, Dulac. Fl. Hautes-Pyr. 247. 1867.

Xeralsine Fourr. in Ann. Soc. Linn. n. s. 16:347. 1868.

F.uthalia Rupr. Fl. Caucas. 220. 1 869.

Ahinopsh Small, Fl. S. E. U. S. 419 & 1330. 1903.

Annual, biennial or perennial herbs, glabrous or pubescent, prostrate and
cespitose to diffuse and spreading, the stems somewhat resilient and occasionally
suffruticose below. Leaves opposite, petiolate to sessile and slightly amplexicaul,
exstipulate, subulate and coriaceous to broadly elliptic and membranaceous. Flowers
axillary or terminal, solitary or in dichasial cymes. Sepals 5, connate only at the
base if at all. Petals 5, rarely lacking, white or occasionally red, entire or shallowly
emarginate. Stamens 10, rarely 5; anthers versatile; 2-celled, longitudinally dehis-
cent; filaments flattened and usually united below to form a slightly perigynous
disk. Ovary superior, 3- (2-5-) carpellate with 3 (2-5) filiform styles free to
their bases; ovules numerous on basal placentae. Capsule dehiscing into as many
entire to deeply emarginate valves as there are styles; seeds numerous, cochleate,
smooth to tuberculate, estrophiolate, the embryo curved about the perisperm.

A weedy genus containing nearly 200 species, this nearly cosmopolitan genus
is in the tropics largely restricted to high altitudes. Fernald (in Rhodora 21:1.
1919) presents adequate reasons for retaining Arenaria sensu lato instead of main-
taining several poorly separated generic segregates. Three intergrading units, best
treated as subspecies of one polymorphic species, A. lanuginosa, are found in
Panama.

(433)

Maguire, in Am. Midi. Nat. 48:498.

Micropetalum lanuginosum (Michx.) Pers. Syn. Pi. 1:509. 1805.

in Mag. Gesells. Nat. Freund. Berl. 7:201. 11816.
Stellaria elongata Nutt. Gen. 1:289. 1818.
Arenaria diffusa Ell. Sketch. Bot. S. C. & Ga. 1 :519. 1821.
Arenaria nemorosa HBK. Nov. Gen. & Sp. 6:3 5. 1823.
Arenaria nemorosa a quitensis DC. Prodr. 1 -.409. 1824.

.\> en.ru >.< ^r. < < amb ex > 2:126. 1829.

Arenaria scabra Vahl, ex St. Hilaire, loc. cit. 2:126. 1829.

Arenaria paradoxa Bart. Rel. Haenk. 2:15. 1831.

SA llaria lanuginosa (Michx.) Torr. & Gray, Fl. Amer. Bor. 1:187. 1840.

Arenaria lanuginosa a genuina Rohrb. in Linnaea 37:260. 1872.

Arcnnir:.! lanuginosa p diffusa Rohrb. loc. cit. 37:263. 1872.

Arenaria lanuginosa diffusa (Ell.) Macloskie, Rep. Princeton Univ. Exped. Patag. 8:394.

1905.
Stellaria laxa Muschl. in Bot. Jahrb. 45:443. 1911.
Arenaria lanuginosa var. longipedunculata Duncan, in Phytologia 3 :282. 1950.

Herbaceous puberulent laxly spreading perennials. Leaves opposite, lanceolate,
oblanceolate or narrowly elliptic, apically attenuate and mucronulate, basally sub-
sessile and attenuate, 5-20 mm. long, 2-5 mm. broad, scantily to densely puberulent
with a whitish indumentum, often ciliate on the margins. Flowers solitary in the
axils, the puberulent pedicels much longer than the subtending leaves. Sepals
lanceolate to ovate, attenuate, puberulent 2-4.5 mm. long; petals 5, occasionally
absent, obovate, entire, 2-4.5 mm. long; stamens 10, the filaments 1-4 mm. long;
ovary sessile or short-stipitate, ovoid; styles 2-4. Capsule ovoid, 3-5 mm. long,
the 2-4 valves deeply emarginate; seeds numerous, on basal placentae, cochleate,
dark reddish brown to black, smooth or minutely tuberculate, ca. 0.75 mm. broad.

Southeastern U. S. to Peru and Bolivia, in Central America only at moderate

CHlRiQuf: vicinity of "New Switzerland", central valley of Rio Chiriqui Viejo, 1800-
00 m., Allen 1414; vicinity of Boquete, 1200-1500 m., Woodson 8 Schery 802.

The typical species was originally described as apetalous, as are most specimens
>m the southeastern United States, the type locality. In Mexico and Central
nerica however, where the ranges of several subspecies overlap, the large majority

(434)

flora of Panama (Caryopbyllaceae) 93

of specimens, obviously referable to this "apetalous" subspecies, have petals nearly
or quite as long as the sepals.

MacBride (in Field Mus. Bot. 13 2 :601. 1936-8) reports that in Peru, where
the plant is called ccledonia and tauchchalli, it is the source of an astringent used
for hemorrhages of the uterus.

2. Arenaria lanuginosa ssp. saxosa (A. Gray) Maguire, in Am. Midi. Nat. 48:
498. 1951.

Arenaria saxosa var. cinerascens Robinson, in Proc. Am. Acad. 29:293. 1894.
Arenaria confusa Rydb. in Bull. Torr. Club 28:275. 1901.

Arenaria mcarnsii Wool. & Standi, in Contr. U. S. Nat. Herb. 16:121. 1939.

29:475. 1939.

Herbaceous (occasionally subligneous at the base) puberulent cespitose
perennials, some of the branches occasionally proliferating and spreading. Leaves
opposite, mostly linear-lanceolate, apically attenuate and mucronulate, basally
subsessile and rounded to acute, 5-18 mm. long, 1-3 mm. broad, scantily to
densely puberulent with a whitish indumentum, often ciliate on the involute
margins. Flowers solitary in the axils, the puberulent pedicels longer or occasion-
ally shorter than the subtending leaves. Sepals lanceolate to ovate, attenuate,
glabrous, 3-5 mm. long; petals 5, obovate, entire, 4-6 mm. long; stamens 10, the
filaments 1—4 mm. long; ovary sessile or short-stipitate, ovoid; styles 2—4. Capsule
ovoid, 3-5 mm. long, the 2-4 valves deeply emarginate; seeds numerous, on basal
placentae, cochleate, dark reddish brown, usually tuberculate, ca. 0.75 mm. broad.

Southwestern U. S. to Mexico; Panama to northwestern South America, at
rather high elevations.

chiriqui: Potrero Muleto, Volcan Chiriqui, Boquete District, 10,400 ft., Davidson
1041 & 1051; Potrero Muleto to sumn -00-4000 m., Woodson &

Schery 405 8 434; Loma Larga to summit, Volcan Chiriqui, 2500-3380 m., Woodson,

3. Arenaria lanuginosa ssp. guatemalensis (Standi. & Steyerm.) J. Duke

comb. & stat. nov.
Arenaria guatemalensis Standi. & Steyerm. in Field Mus. Bot. 23:50. 1944.

Herbaceous puberulent laxly spreading perennials. Leaves opposite, lanceolate,
apically attenuate and mucronulate, basally subsessile and attenuate, 10-40 mm.
long, 2-6 mm. broad, irregularly puberulent with a whitish indumentum, often
ciliate on the margins. Flowers solitary in the axils, the puberulent ebracteate
pedicels 10-40 mm. long, usually exceeding the subtending leaves. Sepals lanceolate
to ovate, attenuate, strongly puberulent on the costa, 4-6 mm. long; petals 5,
obovate, entire, 5-8 mm. long; stamens 10, the filaments 3-6 mm. long; ovary
sessile or short-stipitate, globose or ovoid; styles 2-4. Capsule ovoid, 5-8 mm.
long, the 3 valves deeply emarginate; seeds numerous, on basal placentae, cochleate,

(435)

ANNALS OF THE MISSOURI BOTANICAL GARDEN

chiriqui: rain forest, Bajo Chorro, Boquete District, 6000 ft., Davidson J -
of Finca Lerida, 1750 m., Woodson (3 Schery 214; vicinity of Casita Alta, V
qui, ca. 1500-2000 m., Woodson, Allen & Seibert 8lO; valley of the upper R

•v of Monte Lirio, 1300-1900 m., Seibert 2QJ; in grassy sunny places, vicinity
of Cerro Punta, 1500-2000 m., Seibert 260.

Other variants of the lanuginosa complex are found in Mexico and Central
America and could conceivably occur in Panama. In the subspecies guatemalensis,
the variety ensifolia, with linear leaves and exerted petals, is so far reported only
from southern Mexico. Characterized by ovate leaves, cuneate to cordate leaf
bases and exerted petals, the variant currently passing as A. mcgalantha (Rohrb.)

flora of Panama (Caryophyllaccac) 95

F. N. Williams would probably better be treated as a lanuginosa subspecies, ranging
from southern Mexico to Guatemala. The original description of A. gut t e mgJe nsil
would have been more edifying had it been contrasted with A. lanuginosa instead
of A. megalantha, for although intergrades occur between all three, they are more
frequent and perplexing between the first two. The relative constancy of these
three taxa, where their ranges are distinct, coupled with limited intergradations
near the commissures of their ranges, seems to justify their relegation to the
subspecific level, in accord with the heirarchy established by Maguire (in Am.
Midi. Nat. 46:493. 1951) in his monograph of the North American species of

2. CERASTIUM L.
Cerastium L. Sp. PI. 437. 1753.
Prevotia Adans. Fam. 2:256. 1763.
Centunculm Adans. loc. cit. 2:256. 1761.
'.!••. nchit Ehrh. Beitr. J: 177. 1788.
Qiiatfrn t -lla Ehrh. loc. cii. 4:149. 1789.
DoerrU H» liorkh. in Rhein Mag. 1:528. 1793.
Myosotis Tourn. ex Moench, Meth. 224. 1794.
EsmarcbU Reichb. Fl. Germ. Excurs. 793. 1832.
Doerricra Steud. Nom. 2 1 :522. 1840.

I »UpU Rci.hh. l/li. Germ. 5:37. /. 22J. 1841.
Dicbodnn Bartl. ex Reichb. Nom. 205. 1841.

Annual or perennial, usually pubescent, erect or decumbent herbs. Leaves
opposite, usually sessile or subsessile, exstipulate, often viscid. Flowers few in
dichasial cymes, white. Sepals 5, rarely 4, not connate. Petals as many as the
sepals, rarely absent, emarginate or 2-cleft, white. Stamens 10, occasionally less;
anthers versatile, 2 -celled, longitudinally dehiscent; filaments flattened, scarcely if
at all connate to form an obscurely glandular annulus. Ovary superior; carpels 5,
rarely fewer, with as many distinct styles, the many campylotropous ovules arising
from basal or central placentae. Capsule often arcuate (hence the generic name),
dehiscing longitudinally into as many deeply emarginate valves as there are styles;
seeds numerous, cochleate, usually tuberculate, estrophiolate, the embryo coiled
about the perisperm.

Of this weedy cosmopolitan, temperate genus, only one of about fifty species

1. Cerastium viscosum L. Sp. Pi. 437. 1753.
Cerastium glomeratum Thuill. Fl. Paris 225. 1799.

and usually mucronulate, basally attenuate, 5-25 mm. long, 3-15 mm. broad, the
winged petioles less than 5 mm. long. Flowers few in terminal dichasial cymes with
reduced foliaceous bracts at the major dichotomies, the lanate to villous pedicels
becoming 5-10 mm. long, exceeding the subtending bracts. Sepals 5, ovate to
lance-oblong, acute to attenuate, villous, 3-5 mm. long; petals 5, rarely none,
oblong, deeply 2-cleft, 3-5.5 mm. long; stamens 10, the flattened filaments 2-4
mm. long; ovary sessile or short-stipitate, ovoid; styles usually 5, 1.5-2 mm. long,
glandular throughout; ovules many, campylotropous on central placentae. Capsule
subcylindric, arcuate, 6-8 mm. long, the 5 valves deeply 2-cleft; seeds numerous,
cochleate, light brown, minutely tuberculate, ca. 1 mm. broad.

Naturalized from Europe, now rather common throughout temperate and

chiriquI: Potrero Muleto to
Schery 472-, vicinity of Boq

Stellaria L. Sp. PI. 421

flora of Panama (CaryophyUaceae)

1814. In part.
Urbrea St. Hilaire, in Mem. Mus. Par. 2:287. 1815.
\Utlachium Fries, Fl. Halland. 77. 1817.
Ballarion Raf. in Am. Month. Mag. 266. 1818.
}.v.ti-o<tnuma Ik-nth. In Wall. Cat. /;. 642. 1828.
Krascbeninikovia Turcz. ex Bess, in Flora 17 1 Beibl.: 9. 1834.

.. Hunge, in Mem. Sav. Etr. Petersb. 2:448. 1836.
Scbizotrcbium Reichb. Norn. 205. 1841.
Hylebia Fourr. in Ann. Soc. Linn. Lyon n. s. 16:347. 1868.
Myosantbus Fourr. loc. cit. n. s. 16:348. 1868.

Annual or perennial herbs, often much branched and clambering, the stems
somewhat resilient and occasionally suffruticose below. Leaves opposite, petiolatc
or sessile and somewhat perfoliate, exstipulate, in Central America rather broad and
membranaceous. Flowers in few-flowered dichasial cymes or less frequently
solitary in the axils of slightly reduced leaves. Sepals 4-5, not connate. Petals 4-5,
rarely absent, white, deeply 2 -cleft. Stamens 4-10; anthers versatile, 2-celled, the
flattened hypogynous filaments connate below forming a brief glandular disk.
Ovary superior, carpels 2-4, with as many filiform styles free to the base; ovules
4-many, campylotropous on basal, central or free central placentae. Capsule deeply
or shallowly dehiscing into as many 2-cleft or rarely entire valves as there are
styles; seeds 4-many, cochleate, smooth or tuberculate, estrophiolate, the embryo
curved about the perisperm.

Composed of about a hundred species, this weedy genus is nearly cosmopolitan,
but is, as are most caryophyllaceous genera, largely restricted to high altitudes in
the tropics. Three of the six Central American species are found in Panama.

1. Stellaria nemorum L. Sp. PI. 421. 1753.

CerastiUM nervorum Crantz, Inst. 2:401. 1766.

Stellaria nemoralis Salisb. Prodr. 301. 1796.

Stellaria cuspidata Willd. ex Schlecht. in Ges. Natur. Freund. Berlin Mag. 7:196. 181

Stellaria reichenbachii Wierzb. in Reichb. Ic. Fl. Germ. 5:34. 1841.

$7. 1859-60.
Stellaria ciliata p cuspidata (Willd.) Rohrb. in Linnaea 37:278. 1872.
Stellaria nemorum var. japonic* Franch. & Sav. Enum. PI. Jap. 2 1 :295. 1876.

monUms Pierrat, in Soc. Bot. Roche 2:58. 1880.
Stellaria nemorum subsp. glochidosperma Murb. in Act. Univ. Lunds. 27:156. 1892.
Stellaria nemorum subsp. montana (Pierrat) Murb. Bot. Notiser 201. 1899.
Stellaria nemorum subsp. montana forma reichenbachii ("Wierzb.) Murb. loc. cit. 201.

D Bot. Mag. Tokyo 43:541. 1929.
Stellaria Ihmtanea Standi, in Field Mus. Bot. 22:74. 1940.

Herbaceous glabrous to pubescent profusely branching perennials, pro:

98 ANNALS OF THE MISSOURI BOTANICAL GARDEN

spreading or clambering. Leaves opposite, glabrous to viscid-villous, ovate, apically
acute to attenuate, basally truncate to cordate, 10-50 mm. long, 5-35 mm. broad,
all except the uppermost petiolate, the petioles 5-40 mm. long. Flowers solitary,
axillary or terminal, or in terminal few-flowered cymes, the glabrous to villous
pedicels 5-45 mm. long, often exceeding the subtending leaves. Sepals 5, lanceolate
to ovate, acute to attenuate, glabrous to villous, 4-6 mm. long; petals 5, obovoid,
deeply 2-cleft, 6-12 mm. long; stamens 10, the filaments flattened, 3-8 mm. long;
ovary sessile, globose to ovoid; styles usually 3, 2-3 mm. long, the numerous ovules
campylotropous on basal or free central placentae. Capsule ovoid, 4-7 mm. long,
the 3 valves deeply 2-cleft; seeds numerous, ca. 1 mm. broad, cochleate, dark
reddish brown, tuberculate, the spines usually broader than long.

In the New World, extending from Mexico to Patagonia.

bocas del toro: Robalo trail, northern slopes of Cerro Horqueta, 6000-7000 ft.,
Allen 4QQ?. chiriqui: Potrero Muleto to summit, Volcan Chiriqui, 3500-4000 m.,
Woodson & Schery jpj.

The New World material seems indistinguishable from the Old World material
except that seeds tend to be more uniformly tuberculate and leaves do not tend
to be so large in New World specimens. Phytogeographically it might seem that
New World specimens might at least deserve subspecific rank, but I am unable
to detect any constant differentia therein. The non-papillate tubercles on the seeds
of Panama representatives more closely resemble those of the European subspecies
viontana (Pierrat) Murb. than those of the subspecies glocbidosperma Murb.

2. Stellaria ovata Willd. ex Schlecht. in Ges. Naturf. Freund. Berl. Mag. 7:196.

Herbaceous mostly glabrous, branching perennials, prostrate or spreading.
Leaves opposite, glabrous or subglabrous, ovate to broadly ovate, apically rounded
and mucronulate, basally rounded but slightly attenuate, 10-40 mm. long, 5-30
mm. broad, the petioles 3-8 mm. long. Flowers solitary in the axils, the glabrous
to villous pedicels 10-40 mm. long, usually exceeding the subtending leaves.
Sepals 5, ovate to obovatc, obtuse to acute, glabrous to villous, 2.5-5 mm. long;
petals 5, obovate, deeply 2-cleft, 3-6 mm. long; stamens 10, the flattened filaments
3-4.5 mm. long; ovary sessile, ovoid; styles usually 3, 1-1.5 mm. long, the
numerous ovules campylotropous on basal placentae. Capsule ovoid, 3-5 mm. long,
the 3 valves deeply 2-cleft, occasionally entire; seeds numerous, ca. 1 mm. broad,
cochleate, dark reddish brown, tuberculate, the mature spines usually longer than

Mexico through Centi

valley of the upper Rio Chii

Alto, just e. of Boquete, 5000 ft., Stern, Chambers et al. IOOJ.

Standley and Steyermark (in Field Mus. Bot. 24 4 :238. 1946) report the
following vernacular names are applied in Guatemala: trips de polio, culantro de
monte and ctmrtillera.

(440)

a to northe

:rn and western South America.

Rio

Chiriqui Vi
ty of Monte

ejo, 1300-1900 m., White

Lin,,, n 00-1 900 m., Seibe

& White $i
•rt 275; Pal °

flora of Panama (Caryophyllaceae)

100 ANNALS OF THE MISSOURI

3. Stellaria irazuensis Donn. Smith, in Bot. Gaz. 23:236. 1897.

Herbaceous, mostly glabrous, much branching perennials, prostrate or spread-
ing. Leaves glabrous, ovate to deltoid, apically attenuate, basally cordate, 10-20
mm. long, 5-15 mm. broad, the scantily villous petioles 5-15 mm. long. Flowers
several in dichotomously branching cymes 10-25 cm. long, with minute bracts at
the major dichotomies, the glandular- villous pedicels 5-10 mm. long. Sepals 4,
narrowly ovate to elliptic, obtuse to acute, glabrous or glabrate, 1.5-2.5 mm. long;
petals 4, deeply 2-cleft, the segments linear, 1.5-3.0 mm. long; stamens 4-6 (-8),
the flattened filaments 1.5-2.5 mm. long; ovary sessile, ellipsoid; styles usually 2, ca.
1 mm. long, the 4-6 ovules campylotropous on basal placentae. Capsule flattened,
ellipsoid, ca. 2 mm. long, the 2 valves emarginate; seeds 6 or less, ca. 1 mm. broad,
cochleate, dark reddish brown, tuberculate.

Guatemala to Panama.

chiriqui: vicinity of Bajo Chorro, 1900 m., Woodson # Schery 640.

Further study of this distinctive tetramerous species may well prove it to be
conspecific with S. venezuelana Steyerm., which Steyermark rightly separated from
the heterogeneous assemblage described as S. micrantha Spruce ex Rohrb. Rohr-
bach's description is not at all applicable to Fendler 4.7 which was designated as
cotype of S. micrantha, and which is quite apparently referable to S. veneziielana.

4. DRYMARIA Willd. ex Roem. & Schult.
Drymaria Willd. ex Roem. & Schult. Syst. 5:31. 1819.

Annual or perennial, glabrous to pubescent herbs, occasionally subligneous
below, prostrate, spreading or erect. Leaves opposite, with persistent or fugaceous
small stipules, sessile or petiolate, glabrous to villose, the hairs often glandular.
Flowers in few-flowered racemes or in dichasial cymes, rarely solitary in the axils.
Sepals 5, not connate. Petals (0-) 5, white, usually 2-cleft, with the sinus of the
cleft occasionally laciniate. Stamens 2-5, the anthers versatile, 2-celled, the flat-
tened filaments slightly connate at the base, rarely with prominent staminodia.
Ovary superior, slightly stipitate; carpels 3, the 3 styles united below; ovules
few-many, campylotropous on free central placentae. Capsule ovoid to spheroid,
dehiscing into 3 entire valves; seeds 1-many, cochleate, foetiformor hippocrepiform,
usually tuberculate, the embryo curved about the perisperm.

Exceptional for the Caryophyllaceae in being an almost exclusively tropical
genus, Drymaria is represented in Panama by only two species. D. glandulosa Presl.
and D. palustris, both ranging from North to South America, have not yet been
collected in Panama. The genus consists of perhaps sixty species, all but three
confined to tropical and subtropical America. A peculiar group of species with
foetiform seeds centers about D. holosteoides Benth. which is a seriously toxic range
plant of the Sonoran Desert (Little, in Ecology 18:416. 1937). In the only

(442)

flora of Panama (Caryophyllaceae)

of the genus, Wiggins (in Proc. Calif. Acad. Sci.
4 25 :189. 1944), dealing with species on and near the Sonoran Desert, recognizes
that the seeds are distinctive but lays little emphasis thereon. In my soon forth-
coming revision of the whole genus, the seeds are one of the most important
diagnostic characters, and the petal and leaf shapes are next respectively in impor-
tance to an understanding of the relationships that exist in the genus.

a. Leaves and pedicels scantily to densely villose with long, spreading
non-glandular bai at* to emarginate, 2-4-nerved,

basally provided with linear auricles; seeds 0.5-0.9 mm. broad, with

of glandul

1. Drymai

. Cham. & Schlecht.

red, basally

i 5:232. 1830.

Fig. 140. Drymaria villos*

(443)

ANNALS OF THE MISSOURI BOTANICAL GARDEN

ordata var. pilosa Schlecht. in Linnaea 26:374. 1853.

ordata var. 5 villosa (Cham. & Schlecht.) Rohrb. in Mart. Fl. Bras. 14 2 :i

ull. Jard. B

Prostrate or ascending herbaceous annuals to as muc
nodes mostly longer than the leaves, villose to hirsute w
as 2 mm. long. Leaves opposite, scantily to densely villose or hirsute with cinereous
or ochraceous hairs, orbicular to reniform, apically rounded to acute and apiculate,
basally cordate to truncate, weakly 3-7-veined, 5-15 mm. long, 5-15 mm. broad;
petioles 1-10 mm. long, the stipules mostly entire, scarcely distinguishable from
the villosity, 0.5-1.5 mm. long. Inflorescences of terminal 5-many- flowered cymes,
the ultimate branches often tending to be racemose; peduncles 1-5 cm. long;
bracts 0.5-1.5 mm. long, the pedicels 2-20 mm. long, villose. Sepals 5, 2-3.6 mm.
long, 1-2 mm. broad, narrowly to broadly ovate or elliptic, apically acute to obtuse,
villose to glabrous, weakly 3 -nerved, with translucent borders; petals 5 (absent or
drastically reduced in forms passing as D. tepicana), 2-3.6 mm. long, bifid for half
their length or more, the lobes apically acute to deeply emarginate, 2-4-nerved,
basally provided with linear auricles, these very variable in number and orientation;
stamens usually 5, 2-3.5 mm. long, the oblong anthers 0.3-0.5 mm. long, the
filaments shallowly connate, devoid of staminodia; ovary at anthesis ovoid to
globose; styles 3, united for half their length or more, 1-1.5 mm. long; ovules
numerous, campylotropous on free central placentae. Capsule ovoid to ellipsoid,
2-3.5 mm. long, many-seeded (rarely as few as two in D. tepicana), the seeds
cochleate, 0.5-0.9 mm. broad, the dorsal tubercles filiform to capitate or subpinnate,
longer than broad, the facial tubercles stellate.

Central Mexico through Central America along the western coast of South
America to Peru; apparently introduced and widespread in the East Indies.

chiriqui: Finca Lerida to Boquete, ca. 1300-1700 m„ Woodson, Allen & Seibert 1162;
Bajo Chorro, Boquete District, 6000 ft., Davidson 2Q4; vicinity of Boquete, 1200-1500 m.,
Woodson & Schery 7 IO. cocle: lower portion of valley and marshes along R. Anton, El
Valle de Anton, ca. 500 m., Hunter & Allen 379; between Las Margaritas and El Valle,
Woodson, Allen tf Seibert 1771.

Examination of adequate material from Mexico, home of D. villosa, from Peru,
home of D. hirsuta, and Java, home of D. stylosa, clearly shows these are con-
specific, closely resembling the horticulturally propagated type of D. villosa, with
the petal lobes spatulate, retuse and basally provided with downwardly directed
linear auricles. The Indonesian material is marked, although inconstantly, by
glabrous sepals and stiff erect hairs, i. e. they are truly hirsute as contrasted with
the predominantly villose Mexican specimens. I have seen no specimens repre-
senting this species from the West Indies and believe that the epithet cubensis
stems from some error in transposition of labels.

flora of Panama (Caryophyllaccae) 103

Forms comparable to D. tcpicana M. E. Jones are not infrequent. In these, the
petals are reduced to mere vestiges or are completely absent, and the seeds are often
fewer in a capsule than is typical, but I am strongly inclined to doubt that these
forms warrant even formal recognition. It may be here appropriately mentioned
that similar reductions in the petals occur in the closely related species Drymaria
palustris Cham. & Schlecht., and such variants have been described as Drymaria

Standley and Steyermark (in Field N
and llovizna blanca as Guatemalan nam

lus. Bot. 24 4 :232. 1946) list p
es for this species.

2. Drymaria cordata (L.) Willd. ex

Roem. & Schult. Syst. Veg. 5

Holosteum cordatum L. Sp. PI. 8 8. 1753.
Holosteum diandrum Sw. Prodr. 27. 178S.
Drymaria diandra (Sw.) Macfadyen Fl. Jam
Drymaria cordata § diandra (Sw.) Griseb. F
Drymaria cordata p puberula Tr. & Planch.
Drymaria procumbem N. E. Rose, in Contr.
Drymaria adenopbora Urban, in Fedde, Repc
Drymaria cordata var. pacifica Mizushima, i

1:52. 1837. non Blume, 1825.
. Brit.W. Ind. 56. 1859.
n Ann. Sci. Nat. 4 17 :14R. 1862.
U. S. Nat. Herb. 1:304. 1895.
rt. Sp. Nov. 21:213. 1925.

Glandular-puberulent to glabratc ramifying annuals, prostrate and spreading or
erect, the internodes mostly longer than the nodes, glabrous to densely glandular,
often rooting at the nodes. Leaves glabrous to scantily pubescent, orbiculate
to reniform, apically rounded and occasionally mucronulate, basally rounded to
cordate, 5-25 mm. long, 5-30 mm. broad, the petioles 2-15 mm. long; stipules
mostly polylacerate, rather persistent, to 2 mm. long. Inflorescences of terminal
or axillary few-many-flowered dichasial cymes (flowers rarely solitary in the axils),
the bracteate pedicels locally girdled with a dense band of glandular pubescence,
rarely subglabrous, 2-15 mm. long, equaling to much exceeding the subtending
bracts. Sepals 5, lanceolate to ovate, acute, glandular-puberulent to glabrous,
obscurely to strongly 3 -nerved, 2.5-4 mm. long, the borders translucent; petals
5, 2-3 mm. long, deeply bifid, the lobes linear, acute, rarely obtuse, 1 -nerved,
basally exauriculate but rarely subdentate; stamens 2-3 (-5), the flattened fila-
ments 2-2.5 mm. long, the anthers suborbicular, 0.2-0.3 mm. long; styles 3, free
nearly to their bases or united for half their length, 0.5-1 mm. long; ovules few
to many, campylotropous on free central placentae. Capsule ovoid, 1.5-2.5 mm.
long, the 3 valves entire; seeds 1-12, 1.0-1.5 mm. broad, cochleate, dark reddish
brown, tuberculate in lines, all the tubercles low and domical, closely approximated

A nearly pantropical species, in America ranging from Florida and Mexico
through the West Indies and Central America along both coasts of South America
to Argentina.

Chagres, Isthmus c

Fendler Q was cited in the original description of D. cordata fi puberula Tr.
Planch, but differs little, if any, from typical cordata. Mizushima (in Jour. Ja

104 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Bot. 32:69. 1957), studying the varietal potentials of this polymorphic species,
described var. pacifica, which differs from typical cordata in having glabrous
sepals, more seeds, and a tendency to have a greater number of stamens and the
styles divided completely to the ovary. This rather striking variety seems to be
lacking in continental North America. Occurring in both var. pacifica and var.
puberula, for those who wish to recognize them, are variants in which the flowers
are subsolitary in the axils. Surprisingly and fittingly these have as yet received
no taxonomic status. Drymaria adenophora Urban represents a form with
diminutive leaves and flowers and apparently is to be expected throughout the range
of the species.

Mizushima resurrected D. diandra Blume from a long submergence in the
synonymy of D. cordata. D. diandra seems to be confined to the Old World and
differs in having pyriform flowers with the outer sepals strongly carinate and
three-ribbed, the ribs interconnected. The petal lobes are broader and more obtuse
and the seeds tend to be fewer and larger than in D. cordata.

Among the many colloquial Central American names for this species reported
by Standley and Steyermark (in Field Mus. Bot. 24 4 :228. 1946) are palitaria or
pelitaria, petatillo, comida de canario, trevcilla, comapa, compona and nervillo.
Because of its lush mat-forming properties, the plant is often cultivated as a
ground cover. According to Dickson (in Tea Quarterly 18:84. 1960) however,
the beneficial effects are nullified after a few years, the plant becoming so aggressive
as to notably reduce the yield of a tea plantation.

5. POLYCARPAEA Lam.
Polycarpaea Lam. in Jour. Hist. Nat. Paris 2:8. t. 2$. 1792. Nomen con-

servandum.
Antbyllis Adam. Fam. 2:271. 1763. In part.
Folia Lour. Fl. Cochinch. 164. 1790. Nomen reddendum.
Policarpaea Lam. in Jour. Hist. Nat. Paris 2:1. 1792,
Polycarpoea Lam. loc. cit. 25. 1792.
Hagaea Went. Tahl 3:240. 1799.
Hagea Pers. Syn. 1 :262. 1805.
Mollia Willd. Hon. Berol. II. A, //. 1806.
Labaya Roem. & Schult. Syst. 5:402. 1819.
Aylmeria Mart, in Nov. Act. Nat. Cur. 13:276. 1826.
Hyala L'Herit ex DC. Prodr. 3:375. 1828.

Poly car pia Webb, ex Berthel. Hist. Nat. Isles Canary 3 2 :156. 1836-50.
Plancbonia J. Gay, ex Benth. & Hook. f. Gen. 1:154. 1862.
Robbairea Boiss. Fl. Orient. 1:73 5. 1867.
Polycarpea Pomel, Nouv. Mat. Fl. Atl. 202. 1874.
Poly carpus O. Ktze. in Post, & O. Ktze. Lexicon 453. 1904. In part.

Annual or perennial, usually pubescent, erect or prostrate herbs. Leaves
opposite or verticillate, usually linear and sessile at the swollen nodes, with con-
spicuous scarious stipules. Flowers in terminal dichasial cymes, white or reddish.
Sepals 4-5, occasionally biseriate. Petals 4-5, shorter than the sepals, entire or
toothed. Stamens 4-5, the anthers versatile, the flattened hypogynous filaments

(446)

flora of Panama (Caryophyllaccae)

slightly connate below. Ovary superior, unilocular, carpels usually 3 ; sty
below, the stigmata subcapitate; ovules few, campylotropous on basal
Capsule ovoid to ellipsoid, dehiscing longitudinally into 2-3 entire val
1-few, cochleate, tuberculate, the embryo curved about the perisperm.
Primarily a tropical genus of some thirty species, with only one
species reaching as far north as Panama.

1. Polycarpaea corymbosa (L.) Lam. Illustr. Gen. 2:129. 1793.

Achyrantbes corymbosa L. Sp. PI. 205. 1753.
Polycarpaea spadicea Lam. Illustr. Gen. 2:129. 1793.
Polycarpaea indica Lam. Encycl. 5:483. 1804.

11. 1816.
Labaya corymbosa Roem. & Schult. Syst. 5:404. 1819.
Polycarpaea densiflora Wall. Cat. n. 1513. 1828.
Polycarpaea bra-: re. Fl. Bras. 2:132. 1829.

Polycarpaea b<. h Hilaire, loc. cit. 2:132. 18

Polycarpaea subulata Wight, & Am. Prodr. 35 8. 1834.
Polycarpaea atherophora Steud. in Flora 26:763. 1843.
Polycarpaea brevifolia Muell. in Rep. Babb. Exped. 9. 18 58.
Polycarpa corymbosa O. Ktze. Rev. Gen. 1:51. 1891.
Polycarpa brevifolia O. Ktze. loc. cir. 1:51. 1891.
Polycarpaea filifolia Muschl. in Bot. Jahrb. 45:453. 1911.

Polycarpaea corymbosa var. typica Domin, in Biblioth. Bot. Heft 89-:655. 1925.
Polycarpaea corymbosa var. brevifolia Domin, loc. cit. 89 2 :655. 1925.

(447)

106 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Herbaceous, argentate-pubescent, erect virgate or tufted annuals or perennials
to 1 m. high. Leaves verticillate, linear, aristate, 3-15 mm. long, mostly less than
1 mm. broad, villous or glabrate, the aristate, often lacerate, argentate stipules 2-5
mm. long. Flowers many in congested dichasial cymes, the puberulent pedicels
2-5 mm. long, mostly exceeded by the subtending bracts. Sepals usually 5,
occasionally biseriate, lanceolate to ovate, attenuate to aristate, argentate or rarely
rufous, often with a darker triangle at the base, 2-4 mm. long; petals 5, obovate,
apically undulate, drying yellowish or brown, 0.5-1 mm. long; stamens 5, the
flattened filaments 0.5-1 mm. long; ovary short-stipitate, obovoid, the style 0.25-
0.50 mm. long, the stigmata minute. Capsule ellipsoid, 1-2.5 mm. long, dehiscing
longitudinally into 3 entire valves; seeds 2-6, ca. 0.5 mm. broad, cochleate,
minutely corrugated, light brown, on basal placentae.

Usually in arenaceous savannas, cosmopolitan in the tropics, in the Americas
reaching its northern limits in Panama.

cocle: in savannas near sea level, Aguadulce, Pittier 4953.

cl BOTANICAL C

BOARD OF TRUSTEES

Volume XJLVIII

Annals

of the

Missouri Botanical
Garden

Morphology and Anatomy of the Saururaceae. L Floral Anatomy

and Embryology . M. V. S. Raju 107-124

■'•:■■ ■""■' ■■' .- '-■,: ' ■ .■. :. .": . .■".:■ ~

of New York Francis M. Hueber 125-132

Index of Orchid Names— 1960 . . . . Robert L. Dressier 133-136
Preliminar

Annals

of the

Missouri Botanical Garden

A quarterly journal containing scientific contributions from the
isouri Botanical Garden and the Henry Shaw School of Botany of
tshmgton University in affiliation with the Missouri Botanical

aunng the calendar y

numbers constitute a

Beginning with V,

Single 1

Garden are listed in i

I November. Four

Annals

of the

Missouri Botanical Garden

MORPHOLOGY AND ANATOMY OF THE SAURURACEAE.
I. FLORAL ANATOMY AND EMBRYOLOGY 1

M. V. S. RAJU

Introduction

The Saururaceae has been one of the long-neglected angiosperm families in the
field of morphology and anatomy. A perusal of the literature shows that earlier
investigations have been superficial. In several discussions of comparative mor-
phology and anatomy of the families of the order Piperales and other allied orders,
some of the old descriptions of the Saururaceae have been considered. Very often,
the family Saururaceae and its composition and affinities have been discussed
without much detailed work. In view of this lack of sufficient information, an
attempt has been made to investigate in detail some of the salient morphological
and anatomical features.

The Saururaceae, according to present knowledge, consists of 6 species — Sauru-
rus cernuus L., S. chinensis (Lour) Baill., Houttuynia cordata Thunb., Anemopsis
calif ornica Hook. & Am., Gymnotheca chinensis Decaisne., and G. involucrata P'ei.
No material of the last genus was available for this investigation. However, at
various stages of description and discussion, previous reports on Gymnotheca have
been taken into consideration. The present paper deals with floral anatomy and
embryology and other related features of the Saururaceae.

Members of the family Saururaceae are perennial aromatic herbs inhabiting
moist places. The spirally arranged simple leaves are stipulate (Figs. 1, 5, 8). In
Houttuynia the stipules are prominent, often enclosing a part of the internode

Washington University and submitted as part of a thesis in partial fulfillment of the requirements for

108 ANNALS OF THE MISSOURI BOTANICAL GARDEN

above the node (Fig. 8). In the other genera, they are not prominently visible but
are adnate to the petiole and form a sheath around the node as in monocots. The
leaves of Saururus, Gymnotheca and Houttuynia are auriculate to lanceolate and
palmately net-veined. In the case of Anemopsis, leaves are elliptic-oblong with a

The inflorescence of the Saururaceae is either a terminal spike or a raceme (Figs.
1, 5, 8). In Saururus the naked bracteate flowers are often arranged in pairs, a
feature which distorts their spiral arrangement. The flowers are adnate to the
subtending bract in earlier stages of growth. The flowers of S. chinensis are much
more sparsely arranged than those of S. cernuus, but in most other respects, the
inflorescences of the two species are similar (Figs. 12, 13).

The inflorescence of Houttuynia cordata has been described in detail by Nozeran
(1955). It is a spike with 4 conspicuous bracts at the base (some flowers, however,
have small pedicels). Bracts higher up are much reduced to either scales or simple
enations. The flowers show various degree of adnation of their pedicels with the

(Figs. 9, 15). In this species, Nozeran (1955) has reported a progressive trans-
spike. Similar transformation has been observed by the present author.

Structures of both flower and inflorescence of Anemopsis have been described
by Quibell (1941). The inflorescence is a highly condensed spike and is subtended
by 6-8 petaloid bracts. The young inflorescence axis in early stages resembles a
flower covered by bracts. The inflorescence axis is mainly made up of "sunken"
pistils (Fig. 16). The bracts, which vary in number, are foliaceous and gradually
decrease in size from base upwards (Fig. 5). Except for the lowermost bracts, the
rest are adnate to naked flowers.

The inflorescence of Gymnotheca chinensis is also a spike, if one follows the
terminology of Nozeran (1955). The bracts are not so conspicuous as those of
H. cordata and A. californica. However, the gradual reduction in size of bracts
from base upward is similar to the situation in H. cordata.

Flowers of Saururus, particularly S. cernuus show considerable variation in
structure and number. An average flower of S. cernuus is pedicellate with 4 simple
pistils and 6 stamens. Other combinations like 4-5, 4-7, 4-8, 5-7, 5-8, 3-4, and
3-6 have been observed. Serial sections of flowers of S. cernuus show the spiral
arrangements of carpels and stamens. Carpels are very closely approximated to one
another giving an impression of syncarpy. The stamens are hypogynous and are
opposite the carpels (Fig. 2). Pistils show extensive stigmatic crests on adaxial
sides of styles (Figs. 2, 4). The stamens are much longer than the styles. In
S. chinensis flowers show the same variation as S. cernuus, but the pistils have
shorter styles. Stamens in S. chinensis are perigynous and do not exceed the pistils
in height.

In Houttuynia cordata, the flowers are naked and have small pedicels. These

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 109

pedicels toward the inflorescence apex are further shortened or the flowers are
sessile. The compound pistil is made up of three carpels, and the styles are free
with wide stigmatic crests on their adaxial sides (Figs. 9, 10). In the arrangement
of flowers, Gymnotheca resembles Houttuynia. According to the diagram of
Decaisne (1845), the flower of Gymnotheca chinensh has a 4-partite pistil and 6
epigynous stamens (Fig. 14).

The inflorescence of Anemopsis californica has pistils buried in the "inflores-
cence axis". In these naked flowers, the compound pistil is made up of usually 3
carpels with free styles. The epigynous stamens are usually six in number. How-
ever, this number is subject to variation. The petaloid bract appears to be situated

110

ANNALS OF THE MISSOURI BOTANICAL GARDEN

» the pistil along

at the neck of the compound pistil. Develc

its length (Fig. 49). It is also interesting

bracts show a gradual change from base upwards from a petaloid bract to a petaloid

stamen. This also followed by a highly reticulate venation pattern in the lowest

bract to a highly reduced "dichotomous pattern" in bracts at the tip of inflorescence.

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 111

Saururus. — Anatomy of the flower has been recently described for S. cernuus
by Murthy (1959). Carpellary traces* in the flower originate spirally. Serial
sections also indicate the spiral arrangement of carpels. Although ventral margins
of carpels are closely approximated or partially fused at the lowermost regions of
'pistil', the ventral traces remain free throughout their course. In certain cases the
ventral and dorsal traces of each carpel merge at the apex and appear as though

ending near the base of styles. Ventral traces give off branches to the two ovules
present in each carpel; placentation in the carpel has been here interpreted as
laminar (Figs. 3, 4, 17-22).

The anatomy of the flower of Saururus cbiiu >n'n is similar to that of S. ccmuus
with, however, a few deviations. Stamen traces arise as branches from the dorsal
traces of carpels. The ventral trace invariably merges in the stigmatic region with
the dorsal trace to form a thick common trace. Unlike S. cernuus, there is no
conspicuous style in S. chinemh.

Houttuynia.—The naked flowers are trimerous and bracteate (Fig. 9). The
vascular bundle from the inflorescence axis extends into the cortex and becomes
the trace for the flower. It branches to form two traces, one leading into the bract
and the other into the flower proper (Figs. 32, 33). The bract trace may or may
not ramify depending on whether the bract is "leaf-like" or scaly. The remaining
trace branches at different levels to form three independent carpellary traces which
show a spiral arrangement (Fig. 32). Each carpellary trace gives rise to a dorsal
and two ventral traces (Figs. 32, 34-39). The ventral traces of adjacent carpels
fuse together to form "compound ventral trace" or placental trace. Each placental
trace gives off branches which supply the ovules. The placental trace when it
reaches the style bases separates and extends into the respective styles. Ventral
traces together with the dorsal trace go into the style and all three remain separate.
No instance of their fusion at the tip of style was found. Placentation is parietal
and ovules are present only in placental region where the ventral traces have fused
to form the placental trace. Stamen traces are formed much later than the ventral
traces and arise about the middle of the pistil from the carpellary dorsal trace (Figs.
32,40-48).

Anemopsis. — Quibell (1941) has given in sufficient detail the vascular anatomy
of the flower. As in other genera, a single trace from the inflorescence axis diverges
and branches to form a bract trace and a "flower trace" (Figs. 49-52). Although
the bract is adnate to the pistil, the bract-trace is free throughout (Figs. 49-66).
Each "flower trace" branches to form ventral and dorsal traces. Each carpel
receives two ventral traces and a dorsal trace. As in Houttuynia, ventral traces of
adjacent carpels immediately fuse to form the placental trace whose branches
supply the ovules. At about the base of style, the placental traces separate to give
ventral traces which pass into the respective styles. Each ventral trace, on its way

112 ANNALS OF THE MISSOURI BOTANICAL GARDEN

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 113

114 ANNALS OF THE MISSOURI BOTANICAL GARDEN

;

to the base of the style, is connected with the dorsal trace. The dorsal trace with
the other two ventral traces extend into the style (Figs. 58-66). At the base of
style the dorsal bundle gives off branches to one or two stamens. Placentation is
parietal as in Houttuynia.

Carpels and Placentation

A considerable variation has been found in the structure of carpels and in the
organization of pistils in the Saururaceae. Carpels in Saururus cernuus are free
throughout except at the base, a little above the short pedicel where they show a
lateral concrescence (Figs. 17-22). They arise in a spiral fashion, a fact clearly
seen in serial sections. There is considerable elongation above the ovule-bearing
region. In Houttuynia, Gymnotheca and Anemopsis, clearly recognizable stylar
regions are found. Furthermore, in all these three genera, the pistil is "compound"
(in the sense that the ovary is syncarpous and the styles are free) (Figs. 6, 7, 10,
11).

In Sauntrus cernuus, the folded laminae are conspicuous and the ventral traces
are present away from the carpellary margins. During organization of the pistil,
these laminar margins of adjacent carpels undergo concrescence which is to a
considerable extent mistaken for axile placentation; placentation here is actually
laminar. The lateral concrescence of adjacent carpels may be attributed to a case
of "incipient syncarpous tendency" (Bailey and Swamy, 1951).

The compound pistil of Houttuynia and Anemopsis is formed as a result of
lateral approximation and concrescence of carpellary margins of adjacent carpels.
This concrescence is also accompanied by fusion of ventral traces of adjacent
carpels. Only in this region of concrescence of carpellary margins and also of
lateral or ventral traces is parietal placentation organized. The sterile margins of
carpels also undergo concrescence to form septa; as a result the compound pistil
appears to be three-loculed (Figs. 23-31). This kind of incomplete septation is
found in the upper part of the pistil of Houttuynia cordata; it extends almost half
the depth of the pistil in Anemopsis.

Embryology

.e gametophyte. — The wall of a

.., hypodermal layer, middle layer

ither wall has an epidermis with

: fibrous thickening and

T «t Fig. 5. 67. A portion of anther wall showing cpid

dermal layer with crushed middle layer, and a tepetal layer, X 333. Figs.

- ■ . '

megaspore mother cell, X 666. Fig. 74. Same at a later stage with a dyad, X 666. Fig. 75. Shows
division of lower dyad cell, X 666. Fig. 7>,. r ■ "triad", the lowermost

being the megaspore, X 666. Fig. 77. Degeneration of middle cell of a linear "triad". Note vacu-
olation in the lowermost cell, X 666. Fig. 78. The upper iw„ cell of a "triad" have degenerated

embryo, cellular endosperm and a lower 'caecum containing" a hypertrophic^ nucleus, X 147- F *
ellular endosperm, embryo and perisperm (Emb, embryo; En, endosperm; Nu, nucellus

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 115

116 ANNALS OF THE MISSOURI BOTANICAL GARDEN

tapetal layer with binucleate cells (Fig. 67). The sporogenous cells differentiate
themselves into microspore mother cells which undergo meiosis and form first
microspore tetrads and then microspores (Figs. 68-71).

Megasporogenesis and female gametophyte. — The crassinucellate ovules are
orthotropous-bitegmic and borne on laminar placentae (Figs. 3, 4, 19-21). The
micropyle is organized by both the integuments. The outer integument is two
layered; the inner is three layered. A single hypodermal archesporial cell is differ-
entiated in a young ovular primordium. It undergoes a transverse division to form
primary sporogenous and parietal cells (Fig. 72). The primary parietal cell, by
anticlinal and transverse divisions, forms a parietal tissue of about 8 superposed cells
(Figs. 73-78).

The primary sporogenous cell divides meiotically to form a dyad (Fig. 74).
During further development, the lower dyad cell divides and the upper does not
(Figs. 75-76). Thus, the dyad gives rise to a "triad" or a linear row of three
cells (Figs. 76, 77) . Of these three cells, the upper two degenerate and the lower-
most develops into the embryo-sac (Figs. 77-79). The antipodal cells are ephem-
eral and seldom seen in a mature embryo sac. The egg apparatus is situated at the
micropylar end and no case of fertilization was observed.

Endosperm. — The endosperm nucleus divides transversely to form a micropylar
chamber and a chalazal chamber. The chalazal chamber behaves as a haustorium
and the nucleus becomes hypertrophied. The mature seed encloses massive peri-
sperm and a little endosperm surrounding the developing embryo (Figs. 80-82).

Houttuynia cordata. — Microsporogenesis and male gametophyte. — Microsporo-
genesis and male gametophyte development follow the sequence given for Satirurus
cernuus, but with some differences. Division of the microspore mother cell shows
various cytological abnormalities resulting in the formation of sterile pollen.
Frequently, micronuclei are seen in microspore tetrads. Very rarely free micro-
spores are observed (Figs. 83-88).

Megasporogenesis and female gametophyte. — The bitegmic orthotropous ovules
are borne on parietal placentae (Figs. 25, 26, 89-90, 99). The ovular primordium
shows a hypodermal archesporium (Figs. 89, 90). Similar to S. cernuus, the outer
integument is made up of two layers and inner of three layers (Fig. 99). Again,
the micropyle is organized by both the integuments (Fig. 99). An archesporial
cell of hypodermal origin functions directly as the megaspore mother cell (Figs.

si,„v\ e

,, dermal layer, c:

rushed middle layer i

; odermal layer;

, .' .

>ther cell, X 666. Fig. 87.

the movem,

Ovulu

''% A

lkrf. ..agC

,,: ihc

rz*»A \

it division of arches

RAJU MORPHOLOGY AND

SAURURACEAE 117

118 ANNALS OF THE MISSOURI BOTANICAL GARDEN

91-98). The megaspore mother cell divides meiotically to give a linear row of
three cells of which the lowermost is the megaspore that develops into the embryo
sac (Figs. 91-98). Antipodal cells in the mature embryo sac are epheremal but
in some cases these cells are seen in mature embryo sacs (Figs. 96-97). The egg
apparatus in the micropylar end is made up of an egg cell and two hooked synergid
cells (Fig. 96). Here, also, fertilization stages were not observed.

Endosperm.— The endosperm behaves as in S. cernuus (Fig. 98). In mature
seeds often only endosperm and perisperm were observed. The embryo was not

Anemopsis californica. — Microsporogenesis and
sporogenesis and development of male gametophyt<
cernuus and Houttuynia cordata (Figs. 100-105).

Megasporogenesis and female gametophyte. — The orthotropous bitegmic ovules
are found on parietal placentae as in H. cordata (Figs. 6-7, 30-31) . A hypodermal
archesporial cell is differentiated and divides to form a parietal cell and primary
sporogenous cell (Fig. 106). The primary sporogenous cell divides meiotically to
give a row of three cells as in other genera (Fig. 107). The lowermost cell
functions as the megaspore and develops into an embryo sac (Figs. 107-110).
Stages in fertilization have not been observed.

Endosperm. — The earlier stages of endosperm development are like those of
Houttuynia and Saururus cernuus. In the mature seed, massive perisperm, degen-
erated embryo and endosperm were observed.

In Saururus simple pistils organize the fruit. The pericarp becomes dry-fleshy
enclosing a single seed. No case of fruit dehiscence has been observed. Further-
more, there is no stony endocarp developed in the pericarp. On the other hand,
seed coat becomes very hard and often closely abutting the inner wall of pericarp.
The fruit is here interpreted as a berry.

In Houttaynia and Anemopsis the fruits are dehiscent capsules with many
smaller seeds. The dehiscence of fruit occurs at the apical region by the disinte-
gration of fused ventral margins in the middle region of three free styles. In a
dehisced fruit a circular port is seen at the top of the ovary leaving the styles free.

Discussion

Observations on inflorescence of the six species of Saururaceae show an inter-
esting evolutionary trend, probably guided by two processes: 1. reduction, and
2. cohesion and adnation. Of the four genera, Saururus shows the least change by
the above processes.

The "common stalk" of flower and bract shows various degree of reduction.

In Saurt

' is much elongated and shows stages c

tion. Further reduction of the same has resulted in a condition seen in Houttuynia
in possessing a recognizable "common stalk" (Figs. 12, 13, 15). Anemopsis, on
the other hand, shows the final culmination where the pistils appear embedded
in the "inflorescence axis" and the bract appears to be situated at the top of the
ovary (Figs. 16, 49). In all genera this process of reduction is associated with

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 119

concomitant cohesion and adnation of floral parts. By a process of reduction in
the number and cohesion of carpels, the polycarpellate condition seen in Saururtis
has resulted in the three or four carpelled syncarpous condition found in the other
genera. This change has been accompanied by the adnation of stamens to carpels
and is exhibited by Houttuynia, Gymnotheca and Anemopsis. The genus Anemop-
sis has an inferior ovary, a feature that has resulted by their "sinking into" the

120 ANNALS OF THE MISSOURI BOTANICAL GARDEN

The phenomena mentioned above may also be explained in the light of floral
anatomy. Although the bract appears to be a simple structure, it is very difficult
to see where the bract begins and where it ends. It is morphologically intriguing
to consider the trace of the "common stalk" as a "compound" one of bract and
flower pedicel. This condition has been interpreted in Saururus cernuus as a case
of adnation of flower to bract, by Murthy (1959).

Present observations show that only the pedicel of the flower is adnate to the
bract; this feature is evidenced by vascular anatomy of Houttuynia and Gymno-
theca. In Saururus, Houttuynia and Gymnotheca, the bract tends to remain free
from the flower proper. But, in the case of Anemopsis, however, the bract is
adnate to the pistil. This feature is of secondary derivation for the bract trace is
completely free from the flower. This secondary condition is, probably, due to
"development sinking-in" of the pistil. It can also be interpreted as a case of
lateral concrescence of compound pistils in the inflorescence.

In Saururus cernuus, stamen traces are formed before the carpellary traces are
differentiated. But in S. chinensis, vascular anatomy indicates that stamens have
originated subsequently from the dorsal carpellary trace. This specialization has
proceeded in increasing degree in Houttuynia, Gymnotheca and Anemopsis where
the stamens are perigynous or epigynous.

In Saururus ventral traces are as prominent as the dorsal traces; they remain
free from one another throughout their course except at the stylar tip where they
show signs of fusion. Similar behavior of ventral and dorsal traces is seen in
Houttuynia and Anemopsis; their fusion in the stylar tip is seldom observed.
Besides, in these two latter genera, ventral traces of adjacent carpels fuse to form
a placental trace. Anemopsis shows a unique feature in that the ventral trace has
independent connections with the dorsal trace (Fig. 49). The exact significance
of these connections is not known.

To begin with, the flower gets a single trace from the inflorescence axis in
Saururus, Houttuynia and Anemopsis. Such a feature has been reported in many
angiosperms. Puri (1951) in his review paper suggested that highly reduced
flowers received just one trace. Whether this condition is simple by primitiveness
or by reduction remains to be critically examined. As far as the present observa-
tions go, it seems that traces of different organs have undergone fusion to form a
single composite trace.

The Saururaceae show both hypogenous and epigynous conditions. The prob-
lem then arises as to how the epigynous condition has been achieved. The problem
of origin of the inferior ovary has been highly debated for a long time. Douglas
(1944) in her comprehensive review has discussed different views regarding
morphology of the inferior ovary. Of the many views postulated, only two have
remained prominent; they are appendicular and axial views.

It is beyond doubt that in all genera of Saururaceae except Saururus cernuus
stamens have fused with the carpels. In Houttuynia and S. chinensis stamens are
situated half way up on the pistil. Anemopsis and Gymnotheca have their stamens
on the ovary just below the styles. These clearly show that concrescence of certain

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 121

floral parts (here stamens) with the pistils has given rise to epigyny. This is in
support of the views expressed by van Tieghem (see Puri, 1951, 1952; Eames,
1931).

In Anemopsis, besides the fusion of stamens with the ovary, there is develop-
mentally a concrescence of pistils resulting in a massive spike. This condition has
given an impression that pistils have "sunken". Adnation of the bract, with its
free trace, to the pistil is an additional evidence to show that the ovary wall is
made up of tissues other than that of the ovary wall itself.

Prior to syncarpy and epigyny, profound morphological and anatomical changes
have taken place in flowers of the Saururaceae. These changes have been well
preserved in the genus Saururus. The carpels with their respective stamens are
arranged "spirally". The last-formed carpel does not possess a stamen or stamens.
This detail suggests that the flowers of S. cernuus, adnate to bracts, are not actually
simple; they are, probably, derived from a structure similar to an "inflorescence".
Absence of an axial tissue, development of a terminal normal or abortive carpel
with or without a stamen or stamens, and close telescoping of simple pistils suggest
that the flower is of the "sympodial type". Each flower then represented by a
carpel with its own stamen or stamens. A similar analogy can be traced in the
flowers of Chloranthaceae (Cardemoy, 1863; Hooker, 1890). However, this idea
has been contradicted on the basis of the absence of supernumerary structures in the
flowers of Chloranthaceae (Swamy, 1953).

The successive arrangements of simple pistils with their associated stamen or
stamens lead to two problems — nature of epigyny or perigyny in a simple pistil
and same in compound pistil. In the case of StntntTUS cbmensk the stamen or
stamens are perigynous. This can be interpreted as a case of adnation. Syncarpous
gynoecium also shows perigynous and epigynous conditions. It is difficult to say
whether the cohesion of carpels appeared first or adnation of stamens to carpels.
Available evidences, however, show in the Saururaceae that adnation of stamens to
carpels appeared first as exemplified by S. chinensis. So, the flowers of Hauttuynia
and Anemopsis suggest that they are made up of three independent flowers which
have undergone phylogenetic fusion. This is also amply evidenced by floral
anatomy.

The above interpretations further lead to the problem whether the pistil is
appendicular or axial. Vascular anatomy shows that there is nowhere in the ovary
a residual trace or an axis around which carpels are arranged. The vascular
elements are completely used up by the differentiated carpels. It is on this basis
that flowers of Saururaceae have been interpreted as condensed inflorescence of the
"sympodial type" simulating to a considerable extent vegetative branches where
sympodial growth prevails. Anatomical evidences do not, however, give a correct
picture of inferior ovary. Developmental studies are necessary for the interpreta-
tion of inferior ovary in some members of Saururaceae.

The trends of specialization postulated by Bailey and Swamy (1951) for the
ranalian carpel have recently been useful in interpreting some primitive carpels.
Although carpels of Houttuynia and Anemopsis are quite specialized, the condupli-
cate nature is still present in their styles which have flared-out stigmatic crests.

122

Although ontogenetic studies have not been made, available observations indicate
that the ovule-bearing capacity of the carpel has brought several modifications in
its train. The closure of carpels has occurred in the basal or ovule-bearing part
leaving behind an extensive stigmatic crest. The stigmatic crest is supplied by both
dorsal and ventral bundles. This feature is seen in Saurtirus, Houttuynia and
Anemopsis. In some of the carpels, particularly of S. cernuus, retraction of ventral
traces from stigmatic region toward the ovule-bearing part may be noticed. This
retraction leaves a short style between ovary and stigma. In the case of S. chinen-
sis, the stigmatic crest is supplied by both ventral and dorsal traces; it is situated
just above the ovule-bearing part. This progressive phylogenetic specialization
leading to the formation of a style between stigma and ovary is exhibited by the
carpels of Saururus, Houttuynia and Anemopsis.

Superimposed on the evolutionary specialization mentioned above there is con-
comitant lateral concrescence of carpels resulting in a compound pistil. Further-
more, the "spiral" arrangement of carpels becomes less and less clear. Thus, in the
development of pistils, simple or compound, several independent trends of speciali-
zation have occurred simultaneously in the Saururaceae.

Placentation in the Saururaceae is of two types, laminar and parietal. Laminar
placentation is found in Saururus and parietal placentation in Houttuynia and
Anemopsis. Bailey and Swamy (1951) after an extensive survey of carpels in the
Ranales have come to the conclusion that carpels with laminar placentation are
primitive and show various trends of specialization leading toward other types of
placentation, A detailed study of carpels of Saururus has shown not only a
laminar placentation but also different degrees of incipient syncarpous tendencies.
The carpellary margins undergo concrescence or close approximation. This con-
crescence of carpellary margins does not occur at one level since they are arranged
spirally. Axial tissue that is seen at the base of carpels is actually a composite of
different carpellary margins. As a consequence, the pistil appears to be compound
with more than one locule. In fact, this condition probably led several taxonomists
to interpret this placentation as axile (Bailey, 1958; Lawrence, 1951; Benson,
1957).

From this basic laminar placentation, parietal placentation of Houttuynia and
Anemopsis can be derived. The presence of free styles in the pistils of the above
two genera shows that lateral concrescence of adjacent carpels has occurred in the
ovule-bearing parts of carpels. Closely approximated partitions appear in the
locule of the ovary of Houttuynia and Anemopsis. These partitions are the fused
ventral carpellary margins. In the ovule bearing part, ventral parts of carpels have
been retracted considerably apparently to accommodate the ovules developing
inside.

In the organization of syncarpous ovary from simple carpels in Saururaceae, the
so-called classical "torus" is not involved. The evolutionary trend of specializa-
tion from laminar to parietal placentation involves: firstly, the fusion of ventral
margins of adjacent carpels; and secondly, lateral concrescence of the same carpels.
These are further influenced greatly by the ovule bearing capacity of a syncarpous
ovary. In Houttuynia and Anemopsis the ovary becomes "multilocular". The

RAJU MORPHOLOGY AND ANATOMY OF SAURURACEAE 123

retraction of partitions within the ovary in the lower part may be due to the
acquisition of ovule-bearing capacity of ovary. Normally, the ovaries of genera
mentioned above show multilocular nature in the upper part of ovary and uni-
locular in the lower part. Intermediate conditions showing various stages of
phylogenetic retraction of partitions are not infrequent.

Eichler (1875-78) interpreted the flowers of Saururus and Houttuynia as obdi-
plostemonous; flowers which had two whorls of three stamens each and the outer
whorl had aborted. In the present investigation the above-said feature has not
been observed. In Anemopsis and Gymnotheca, each flower is usually associated
with six stamens. Anatomy of flower has shown that traces for stamens arise from
the dorsal bundle of carpels. In addition to the above feature, the first formed
carpels of Saururus usually possess two stamens each, the last formed does not
develop stamens. The absence of correlation between the number of carpels and
stamens, negates the possibility of recognizing two whorls of stamens. A similar
view has also been advanced by Nozeran (1955).

In his comparative survey of floral structures Nozeran (1955) writes that
"the spike of Houttuynia is not entirely an inflorescence and is not still a flower".
Furthermore, he hypothesized that the lowermost bracts transform into perianth
lobes. He also observed in inflorescence several instances of reduction in number
of stamens and carpels. On the basis of above details, he derived the spike of
Houttuynia from a homogeneous inflorescence which was composed probably of
trimerous flowers. In the present investigation, however, all the details given by
Nozeran (1955) were not found. There is no doubt that at the tip of the inflores-
cence, an overy was often borne, but there has not been enough evidence at present
to support Nozeran's hypothesis.

Embryological investigations show great similarities among the three species,
Saururus cernuus, Houttuynia cordata and Anemopsis. Development of both male
and female gametophytes are very similar. During development of female gameto-
phyte the archesporial cell gives rise to parietal cells in Saururus and Anemopsis.
These parietal cells are absent in Houttuynia. The development of the female
gametophyte in all three species studied conforms to the monosporic type (as here
interpreted). Some stages in the development of endosperm were observed in
Houttuynia and Saururus cernuus. In both cases the endosperm is of the cellular
type. The above two genera also show endosperm haustoria. The embryo in
S. cernuus suggests that the first division of the zygote should be transverse. Seeds
enclose a massive perisperm and a little endosperm surrounding the small embryo.
Similar embryological features have been recorded by Johnson (1900) for

Summary

An account of floral anatomy and embryology suggests that the genera of
Saururaceae are closely interrelated. Structure and morphology of inflorescence
and flower are discussed in the light of anatomy.

Anatomical details indicate that the inflorescence-type seen in Anemopsis is
more specialized than the type seen in Saururus or other genera. This has been

124 ANNALS OF THE MISSOURI BOTANICAL GARDEN

achieved, according to the evidences advanced, by processes of reduction, adnation
and cohesion of various floral entities.

Some of the morphological aspects of ovary, carpels and placentation have been
discussed. Embryological details are similar in all the genera investigated here.

I am greatly indebted to Dr. N. H. Nickerson, for his kindness and generosity
and valuable guidance during the course of this investigation. I am also highly
thankful to Dr. R. E. Woodson, Jr., for generously placing some of the material
used here at my disposal and for helpful suggestions.

Literature Cited

1 oi cultivated plants". Revised Edition, McMillan Co., N. Y,
ssification". D. C. Heath and Co., Boston.

Lawrence, G. H.

,(1952). Placentation in Angiosperms. Bot. Rev. 18:603-651.

Quibell, C. H. (1941). Floral anatomy and morphology of Anemopsis calif ornica H. & A.

, B. G. L.

HEPATICITES DEVONICUS
LIVERWORT FROM THE DEVONIAN OF NEW YORK 1 - 2
FRANCIS M. HUEBER

Introduction
The record of the Hepaticae in the Paleozoic is very meager. One may readily
expect this to be the case owing to the small size and the delicate nature of the
members of this group of plants. Up until now the oldest occurrence of undoubted
hepatics was that of four species from the Upper Carboniferous of Shropshire,
England (Walton, 1925, 1928). They are Hepaticites Kidstoni, H. lobatus, H.
Langi (Walton, 1925), and H. Metzgerioides (Walton, 1928). These four fossil
forms have been compared with the living genera Trcubia, Fossombronia, Kiccardia
(Aneura) and Metzgeria respectively. In all cases the remains are of sterile shoots
only and on the basis of their morphology it has been assumed that all of the forms
are of the anacrogynous series of the Jungermanniales (Walton, 1925, 571).
Definite conclusions can not be made until reproductive organs are found.

Materials and Methods

The plant remains which are described here occur as compressions associated
with other plant remains in a fine-grained, dark-gray shale. The specimens were
collected from a small quarry located on the northwest slope of South Mountain,
Schoharie County, New York, 1.1 miles west of the Schoharie-Greene County line
Livingston 7' 30" map at 74° 16' 30" E., 42° 23' 55" N. and at an elevation of
2180 feet. The horizon at which the quarry is located is very close to the base
of the Onteora "red beds" which are the continental equivalent to the lowermost
two- thirds of the marine Finger Lakes stage, Senecan series (Frasnian) of central
and western New York. The beds are therefore lowermost Upper Devonian.

The plants with which the specimens of Hepaticites are associated, that is in
the same beds and at the same quarry, are Psilophyfon princeps Dawson 1871,
Schizopodium Harris 1929, Cladoxylon Unger 1856, Pseudosporochnus Potonie and
Bernard 1904, Eospermatopteris (Dawson) Goldring 1924, Callixylon Zalessky
1911, and two new genera all of which are to be described in future papers.

While preparing transfers of other plant remains, thin fragments of plant
material were seen in the sludge derived from the dissolution of the shale. The
fragments at first were thought to be epidermal layers from plants contained in the
shale. However, after they were washed, dehydrated and mounted, they were

126

found to possess the characteristics of the genus Hepaticites. Additional prepara-
tions were made using the bulk maceration technique (Harris, 1926). Pieces of
rock in which large quantities of plant remains are evident are placed into poly-
ethylene or polystyrene plastic containers. Concentrated hydrofluoric acid (48%)
is poured over the specimens until they are completely immersed in the acid. From
ten to twenty- four hours are generally required for dissolution of the rock; the
time is of course dependent upon the size of the specimen being dissolved. Frequent
observations of the progress of the reaction are necessary in order that any freed
plant material which is either floating on the surface of the acid or lying in the
sludge can be removed. The remains are easily removed by slipping a flat, wooden
rod under the specimen and then carefully raising the rod out of the solution. The
specimen is "floated" onto the rod in the way that one floats specimens of algae
onto paper for the preparation of herbarium sheets. The writer has found that the
wooden labels used for flower pots are very satisfactory for lifting the specimens
from the solution. The plant material is floated off of the rod into a container of
water and washed through three or four changes of water. The wash water is
either decanted or siphoned off. After the specimen is washed, it is floated onto a
glass slide and positioned on the slide by the use of brushes cut down to a single
bristle. The slide is gradually raised from the water, keeping the slide inclined. A
drop or two of liquid detergent added to the water will help reduce surface tension
and facilitate the removal of the specimen. Any water remaining on the slide and
around the specimen is carefully removed with blotting paper. After the specimen
has partially dried, it is saturated with a 1 % aqueous solution of gum arabic which
when dried serves to hold the specimen in place on the surface of the slide. The
gum arabic solution is applied with a fine-tipped brush. When the specimen is
fully dried, a drop of Harleco Synthetic Mounting Medium or Canada balsam
dissolved in xylene is placed on top of it and a coverslip lowered over the surface.
At this point, if the specimen were not fixed to the surface of the slide with gum
arabic, it would migrate to the edge of the coverslip or be broken by the flow of
the viscous mounting medium. The mounting medium is dried and the prepara-
tion is ready for study. Nearly eighty fragments were prepared in this way.

Genus HEPATICITES Walton 1925
Hepaticites devonicus sp. nov. — Description
The thallus is very simply differentiated into two parts, a rhizomatous portion
and a thallose region. The latter is differentiated into two parts, an axial or costal
region composed of elongate parenchymatous cells and marginal wings or lamellae
composed of polygonal parenchymatous cells (figs. 1-5). The axial region aver-
ages 0.56 mm. in width. It is two to several cells thick; the actual number is
impossible to determine because of the complex reticulum formed by the outlines
of the cells. A midrib composed of very elongate cells and averaging 0.22 mm. in
width forms the central portion of the axial region. The midrib is in turn sur-
rounded by a zone of elongate cells which are transitional in size between the
midrib and the cells of the wings. The cells are elongate parallel to the long axis
of the thallus and are 14 to 26 ,i in diameter. Because of the incomplete preserva-

HUEBER HEPATICITES DEVONICUS 127

tion of the end walls, the full length of any one of the cells of the midrib is
unknown. The cells which form the transition zone between the midrib and the
wings average 140 /i in length. No rhizoids are evident along the axial region.
The wings are only one cell thick and vary from 1.3 to 1.75 mm. in width. The
cells are 46.8 to 73.2 /j. in diameter and are polygonal in outline. Scattered through
the tissue of the wings are dark cells (fig. 5). The contents may indicate that
some of them were storage or secretory cells. Some, however, may be darkened due
to the activity of fungi, for fungal hyphae are commonly observed adhering to the
surfaces of the thallus. The margins of the wings are serrate, the teeth are
spaced at intervals of 0.18 mm. (figs. 6, 7), but otherwise the margins show no
well pronounced lobing. Dichotomous branching of the thallus is evidenced in
three specimens, two of which are illustrated in figs. 2 and 3. Fragments of the
wings are still attached to the axial regions beyond the dichotomy. Growth of
the plant was apparently the result of the divisions of a single apical cell borne in
a notch at the apex of the thallus. A fragment of the thallus illustrating what
may be interpreted as an example of the apical notch is shown in fig. 4.

The rhizomatous portion of the plant (fig. 8) varies from 0.58 to 1.7 mm. in
width and is composed wholly of elongate parenchymatous cells. The midrib racial
from 0.15 to 0.28 mm. in width; the cells of which it is composed vary from 14
to 29 fi in diameter and as in the midrib of the thallus portion their length
is unknown. The cells which surround the midrib are approximately the same
diameter as those in the midrib and they vary from 136 to 220 fi in length.
Numerous rhizoids arise from the cells on the surface of the rhizome (figs. 8 and
9). They are nonseptate and average 21 /x in diameter just above the broadened
base. The tips of the rhizoids are broken off and therefore the full length of any
one individual is not known. The largest fragment is 0.56 mm. in length. It
must be stated here that no rhizomatous portion has been found in direct organic
connection with the thallose portion of the plant. Both parts were isolated from
the same matrix and since the part which is here referred to as the rhizome has the
same general form and structure of the axial region of the thallose segment,
the writer feels justified in considering them parts of the same plant.

A comparsion of the illustrations of Hepaticites devonicus with those for the
species H. Metzgerioides (Walton, 1928, Plate XII, Figs. 1-6) will show how
similar the thallose portion of H. devonicus is to one which has already been
included in the genus Hepaticites. The finely serrate margin, the lack of rhizoids
on the thallose portion of the plant, and the differentiation of a rhizome bearing
rhizoids, all characteristics of H. devonicus, are the evident morphological differ-
ences between the two species. The difference in the gross size between the two
(H. devonicus is perhaps four to five times larger than H. Metzgerioides) may be
of specific value, but the similar size of the constituent cells in the two plants
suggests that the difference is probably one of growth form.

As in H. Metzgerioides, H. devonicus shows a clear distinction between midrib
and wing which is also characteristic of the modern genera Pallavicinia Gray, corr.
Carruthers and Metzgeria Raddi. Hepaticites devonicus combines characteristics

128 ANNALS OF THE MISSOURI BOTANICAL GARDEN

of both of the modern genera. It is similar to Pallavicinia in the width of the
thallus (3 to 4 mm.) and may represent a plant like Pallavicinia Zollingeri
(Gottsch.) Schiffn. (See Smith, 1938, especially Fig. 32 and Campbell and
Williams, 1914, especially Fig. 4) in which there is differentiation of a prostrate
portion bearing rhizoids and an erect, branched portion on which rhizoids are not
present. The small marginal teeth of the thallus in H. devonicus is another char-
acter which is shared with Pallavicinia Zollingeri. Hepaticites devomcm differ!
from Pallavicinia in having a less differentiated central strand and larger central
cells in the strand. Hepaticites devonicus and Metzgeria are similar in their slightly
differentiated central strands, in the size of the cells in the central strand, the size
of the cells in the wings, and the possibility that the teeth on H. devonicus are
homologous to the hairs on the margins of the thallus in Metzgeria. They differ
in that Metzgeria is smaller (1 to 2 mm. wide) and bears rhizoids along its midrib.
Ifepaticites devonicus resembles Metzgeria in cellular details but resembles Palla-
vicinia in size and form.

Reproductive structures are not known for H. devonicus but on the basis of
its similarity in habit and the many details of structure which it shares with mem-
bers of the Anacrogynae the writer suggests that it be classed with this group until
such a time as further evidence of its relationship is obtained.

The occurence in Devonian sediments of remains assignable to the genus
Hepaticites extends the range of the Hepaticae even further back in geologic time,
and it emphasizes that this class of plants had evolved to a high order of organiza-
tion far earlier than heretofore recorded.

Hepaticites devonicus Hueber sp. nov.

Diagnosis: Dichotomously branching thallus differentiated into rhizomatous por-
tion bearing rhizoids and thallose portion composed of axial region averaging 0.56
mm. in diameter and wings one cell in thickness ranging from 1.3 to 1.75 mm. in
width. Cells of the axial region elongate, 14 to 26 /x in diameter, length of central
cells undetermined, length of transitional cells averaging 140 fi. Cells of the wings
polygonal, 46.8 to 73.2 p. in diameter. Margins of wings finely serrate, lobing
indefinite. Rhizomatous portion 0.58 to 1.7 mm. in diameter, cells elongate as in
axial region of thallose portion, numerous non septate rhizoids averaging 21 p. in
diameter borne on surface of rhizome.

Holotype number 23, specimen 1579-294, slides 6-12, paleobotanical collections of
the Department of Botany, Cornell University, New York.
Locality: Northwest slope South Mountain, Schoharie County, New York.
Horizon: Base of the Onteora "red beds" lowermost Upper Devonian.

A new species of Hepaticites, H. devonicus, is described from the lowermost
Upper Devonian of New York. The plant is differentiated into two parts, a
rhizomatous portion and a thallose region. The latter is differentiated into two
parts, an axial region composed of elongate parenchymatous cells and marginal
wings which are one cell thick and composed of polygonal parenchymatous cells.

HUEBER HEPATICITES DEVONICUS 129

The margins of the wings are finely serrate and show no lobing. There are no
rhizoids borne on the thallose portion of the plant. The rhizome is composed
wholly of elongate parenchymatous cells and resembles the axial region of the
thallose portion of the plant. Rhizoids occur over the entire surface of the
rhizome. No reproductive structures are known.

The plant is compared with Pallavicinia Zollingeri and Metzgeria and placed
for the time being with the fossil forms of the Anacrogynae.

The report constitutes the earliest record of the Hepaticae in the geologic
column.

The writer wishes to express his sincere appreciation to Dr. Harlan P. Banks for
his supervision and guidance during the course of the research which resulted in
the discovery of a diverse flora in the Devonian of eastern New York. Special
acknowledgement is extended to the National Science Foundation for the support
afforded by graduate fellowships for the years 1957-1958, 1958-1959 and 1959-
1960 during which time the research was carried to completion. Thanks arc
extended to Dr. Rudolph Schuster for his suggestions concerning the specimens of
Hepaticites and to Mrs. J. Douglas Grierson for the photographic prints used to
illustrate the present paper.

Literature Cited

Scria 15:1-44.
Harris, T. M. (1926). Note on a New Method for the Investigation of Fossil Plants. New Phytol.

Smith, G. M. (1938). Cryptgamic Botany, Vol. II: Bryophytes and Pteridophytes. (First Ed.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Slide 8,

Fig. 3. Portion of the thallose region of the plant showing a dicho
The marginal wings are folded over the midrib in upper portions of the s

Fig. 4. Apex of the thallose portion of the plant. Slide 11, X 58.

Fig. 5. Detail of the marginal wing of the thallose portion of the

darkened cells which may have been storage or secretory cells. Slide 10,

A-. Mo. Bo,.C,,k„. V

01. -is. 19*]

r 1

IS! *

1 K

■m *i

If

j ]

* 1

r - /

^'

^ «Xv »11

a*n

* ft

4 ^l^V

HUEBER— HEPATICITE

HUEBER HEPATICITES DEVONICUS

Explanation 01

Plate

PLATE II

Hcpaticites dev

nicus

Jr° s t°°£°

Ltinr 8 SUden 2 7. the

thallose regie

the plant illustra

Detail of

one of

the marginal teeth s

hown on the

speL

men in Fig. 6. S

Fragment

of the

rhizomatous portion

of the plant

N

te the dense cove

! rhizoids. The tips c

INDEX OF ORCHID NAMES— 1960

ROBERT L. DRESSLER

As an aid to ourselves and others studying the orchids, we have started an index
of new orchid names, to be published annually. It is felt that such an index may
be of value to the many botanists and others interested in this large and complex
family. For new taxa we have indicated the country or origin and the location
of type material (using the standard abbreviations), if this is known. For new
combinations, new names or changes in rank, we cite the basionym with the
new name. "We have abbreviated the authors' names less than is customary, as we
feel that a saving of only one or two spaces does not justify the loss of clarity. A
few names which were published in mimeographed news letters are purposely
omitted. Such publication is of dubious validity under the rules of nomenclature,
and certainly should be avoided. We would be very glad to receive corrections or
additions to the index, to be appended to the Index for 1961.

Subfamily Cypripedioideae Garay, Bot. Mus. Leafl. 19:86. Type genus:

Cypripedium L.
Subfamily Epidendroideae Dressier, Taxon 9:213. Based on Kerospbaeroideae

Subfamily Kerosphaeroideae Garay, Bot. Mus. Leafl. 19:87. Type genus:

Epidendrum L. This name is not legitimate under the Rules of Nomenclature.
Subfamily Neottioideae Garay, Bot. Mus. Leafl. 19:86. Type genus: Neottia Sw.
Subfamily Ophrydoideae Garay, Bot. Mus. Leafl. 19:87. Type genus: Orchis L.

This name is not legitimate under the Rules of Nomenclature; this subfamily

must be known as the Orchidoideae.
Aeranthes albidiflora Toilliez-Genoud, Ursch & Bosser, Not. Syst. Paris 16: 205.

Madagascar (TAN).
Aeranthes ambrensis Toilliez-Genoud, Ursch & Bosser, Not. Syst. Paris 16: 209.

Madagascar (TAN).
Aeranthes carnosa Toilliez-Genoud, Ursch & Bosser, Not. Syst. Paris 16: 209.

Madagascar (TAN).
Aeranthes denticulata Toilliez-Genoud, Ursch & Bosser, Not. Syst. Paris 16: 207.

Madagascar (TAN).
Aeranthes neoperrieri Toilliez-Genoud, Ursch & Bosser, Not. Syst. Paris 16: 212.

Madagascar (TAN).
Aeranthes orthopoda Toilliez-Genoud, Ursch & Bosser, Not. Syst. Paris 16: 212:

Madagascar (TAN).
Bletia purpurea var. alba Ariza- Julia & Jimenez, Rhodora 62:256. Dominican

Republic.
Bulbophyllum jacquetii var. rosea Guillaumin, Bull. Mus. Hist. Nat. Paris II 32:

189. Annam (P?).
Bulbophyllum rugosibulbum Summerh., Kew Bull. 14:138. Northern Rhodesia

(K).

ull. Mus. Hist. Nat. Paris II 32:

(133)

134 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Chroniochilus virescens (Ridley) Holttum, Kew Bull. 14:273. Sarcochilus vires-
cens Ridley.

Clehostoma tixieri Guillaumin, Bull. Mus. Hist. Nat. Paris II 32:369. Annam
(P?).

Cypripedium X andrewsii nm. favillianum (J. T. Curtis) Boivin, Naturaliste
Canadien 87:32. Cypripedium X favillianum J. T. Curtis.

Cypripedium X andrewsii nm. landonii (Garay) Boivin, Naturaliste Canadien 87:
32. Cypripedium X landonii Garay.

Cyrtorchh arcuata ssp. leonensis Summerh., Kew Bull. 14:149. Sierra Leone
(K).

Cyrtorchh arcuata ssp. variabilis Summerh., Kew Bull. 14:148. Tanganyika
Terr. (K).

Cyrtorchh arcuata ssp. whytei (Rolfe) Summerh., Kew Bull 14:147. Ustrosta-
chys whytei Rolfe.

Cyrtorchh neglecta Summerh., Kew Bull. 14:149. Tanganyika Terr. (K.)

Dactylorhiza— Sod (Ann. Univ. Scient. Budapest. Eotvos Biol. 3:335-357) gives
a synopsis of Dactylorhiza (Dactylorchh) which includes many apparent new
combinations for species, subspecies and hybrid populations. "We do not list
these here because there is not "a full and direct reference to its author and
original publication" with the basionyms. Further, it is quite unclear just
which names Soo intends to publish as new combinations. It seems unlikely
that one would intentionally publish so many new combinations in Dactylorhiza
while recommending that Dactylorchh be considered for conservation.

Diaphananthe section Rhipidoglossum (Schltr.) Summerh., Kew Bull. 14:141.
Rhipidoglossum Schltr.

Diaphananthe brevifolia (Summerh.) Summerh., Kew Bull. 14:143. Rhipido-
glossum brevifolium Summerh.

Diaphananthe cuneata Summerh., Kew Bull. 14:141. Uganda (K).

Diaphananthe densiflora (Summerh.) Summerh., Kew Bull. 14:143. Rhipido
glossum densiflorum Summerh.

Diaphananthe erecto-calcarata (De Wild.) Summerh., Kew Bull. 14:143
Angraecum erecto-calcaratum De Wild.

Diaphananthe globoso-calcarata (De Wild.) Summerh., Kew Bull. 14:142. An
graecum glohoso-calcaratum De Wild.

Diaphananthe laxiflora (Summerh.) Summerh., Kew Bull. 14:141. Rhipidoglos-
sum laxiflorum Summerh.

Diaphananthe longicalcar (Summerh.) Summerh., Kew Bull. 14:142. Rhipido
glossum longicalcar Summerh.

Diaphananthe microphylla (Summerh.) Summerh., Kew Bull. 14:142. Rhipido-
glossum microphyllum Summerh.

Diaphananthe obanensis (Rendle) Summerh., Kew Bull. 14:142. Angraecum
ohanense Rendle.

Diaphananthe peglerae (Bolus) Summerh., Kew Bull. 14:143. Mystacidium
peglerae Bolus.

Diaphananthe rohrii (Reichb. f.) Summerh., Kew Bull. 14:140. Angraecum
rohrii Reichb. f .

DRESSLER INDEX OF ORCHID NAMES 1960 135

Diaphananthe rutila (Reichb. f.) Summerh., Kew Bull. 14:143. Aeranthus

rutilus Reichb. f.
Diaphananthe xanthopollinia (Reichb. f.) Summerh., Kew Bull. 14:143. Aeran-
thus xanthopollinius Reichb. f.
Dinklageella minor Summerh., Kew Bull. 14:156. Liberia (K).
Epidendrum hioramii Acuiia & Alain, Mem. Soc. Cubana Hist. Nat. 24:110.

Cuba (SV).
Epipactis thunbergii forma subconformis Sakata, Jour. Jap. Bot. 35:224. Japan

(Herb.T.Sakata).
Eulophia laurentii (De Wild.) Summerh., Kew Bull. 14:139. Lissochilus

laurentii De Wild.
Habenaria section Kryptostoma Summerh., Kew Bull. 14:135. Type species:

H. tentaculigera Reichb. f .
Habenaria gonatosiphon Summerh., Kew Bull. 14:134. Tanganyika Terr. (K).
Habenaria magnirostris Summerh., Kew Bull. 14:132. Tanganyika Terr. (K).
Habenaria richardsiae Summerh., Kew Bull. 14:131. Tanganyika Terr. (K).
Holothrix brachycheira Summerh., Kew Bull. 14:127. Northern Rhodesia (K).
Holothrix papillosa Summerh., Kew Bull. 14:128. Tanganyika Terr. (K).
Holothrix pentadactyla (Summerh.) Summerh., Kew Bull. 14:129. Deroemera

pentadactyla Summerh.
Laeliopsis domingensis var. alba Ariza-Julia & Jimenez, Rhodora 62:236. Domini-
can Republic.
Luisia acutilabris Guillaumin, Bull. Mus. Hist. Nat. Paris II 32:188. Laos (P?).
Masdevallia horrida Teuscher & Garay, Amer. Orchid Soc. Bull. 29:23. Costa

Rica (AMES) .
MaxiUaria imbraicata var. carinata (Rodr.) Pabst, Orquidea (Brazil) 22: 134.

Maxillaria carinata Rodr.
MaxiUaria imbricata var. iguapensis (Hoehne & Schltr.) Pabst, Orquidea (Bra-
zil) 22: 134. Maxillaria iguapensis Hoehne & Schltr.
Oberonia attenuata Dockrill, No. Queensland Naturalist 29:4. Australia (BRI).
Oberonia recurva var. Hngmalensis (Blatt. & McC.) Sant. & Kapadia, Jour.

Bombay Nat. Hist. Soc. 57:259. Oberonia Hngmalensis Blatt. & McC.
Oberonia santapaui Kapadia, Jour. Bombay Nat. Hist. Soc. 57:265. Oberonia

lindleyana Wight, 1851, not O. lindleyana Duperr., 1834.
Oberonia trullaelabris Guillaumin, Bull. Mus. Hist. Nat. Paris TI 32:115.

Annam (P>).
Orchis rotnndifolia forma beckettii Boivin, Naturaliste Canadien 87:42. Canada

(DAO).
Peristylus densus (Lindley) Santapau & Kapadia, Jour. Bombay Nat. Hist. Soc.

57: 128. Coeloglossum densum Lindley.
Polys tachya bella Summerh., Kew Bull. 14:137. Kenya Colony (K) .
Polystachya testuana Summerh., Kew Bull. 14:136. Gabon (Herb. Le Testu).
Pteroceras alatum (Holttum) Holttum, Kew Bull. 14:269. Sarcochilus alatum

Holttum.
Pteroceras appendiculatum (Bl.) Holttum, Kew Bull. 14:269. Dendrocolla
appendiculata Bl.

136 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Pteroceras berkeleyi (Reichb. f.) Holttum, Kew Bull. 14:269. Thrixspermum

berkleyi Reichb. f .
Pteroceras biserratum (Ridley) Holttum, Kew Bull. 14:269. Sarcochilus bi-

serratus Ridley.
Pteroceras caligare (Ridley) Holttum, Kew Bull. 14:269. Sarcochilus caligaris

Ridley.
Pteroceras carrii (L. ¥ras.) Holttum, Kew Bull. 14:269. Sarcochilus carrii

L. Vms.
Pteroceras ciliatum (Ridley) Holttum, Kew Bull. 14:269. Dendrocolla ciliata

Ridley.
Pteroceras compressum (Bl.) Holttum, Kew Bull. 14:270. Dendrocolla com-

pressa Bl.
Pteroceras ecalcaratum (Holttum) Holttum, Kew Bull. 14:270. Sarcochilus

ecalcaratus Holttum.
Pteroceras emarginatum (Bl.) Holttum, Kew Bull. 14:270. Dendrocolla emar-

ginata Bl.
Pteroceras fasciculatum (Carr) Holttum, Kew Bull. 14:270. Sarcochilus fas-

ciculatus Carr.
Pteroceras hirsutum (Hooker f.) Holttum, Kew Bull. 14:270. Sarcochilus

hirsutus Hooker f.
Pteroceras johorense (Holttum) Holttum, Kew Bull. 14:270. Sarcochilus

johorensis Holttum.
Pteroceras minutiflorum (Ridley) Holttum, Kew Bull. 14:270. Asochilus

minutiflortis Ridley.
Pteroceras pallidum (Bl.) Holttum, Kew Bull. 14: 270. Dendrocolla pallida Bl.
Pteroceras siamense (Ridley) Holttum, Kew Bull. 14:271. Asochilus tkmentk

Ridley.
Pteroceras stenoglottis (Hooker f.) Holttum, Kew Bull. 14:271. Sarcochilus

stenoglottis Hooker f .
Pteroceras sauveolens (Roxb.) Holttum, Kew Bull. 14:271. Aerides suaveolens

Roxb.
Pteroceras tanyphyllum (Ridley) Holttum, Kew Bull 14:271 Sarcochilus tany-

phyllus Ridley.
Pteroceras teres (Bl.) Holttum, Kew Bull. 14:271. Dendrocolla teres Bl.
Pteroceras teysmannii (Miq.) Holttum, Kew Bull. 14:271. Aerides teystnannii

Miq.
Pteroceras violaceum (Ridley) Holttum, Kew Bull. 14:272. Sarcochilus viola-

ceus Ridley.
Pteroceras viridiflorum (Thw.) Holttum, Kew Bull. 14:272. Aerides viridiflora

Thw.
Pteroceras zollingeri (Reichb. f.) Holttum, Kew Bull. 14:272. Dendrocolla

zollingeri Reichb. f .
Schizochilus lepidus Summerh., Kew Bull. 14:130. Mozambique (Southern Rho-
desia Gov. Herb.).
Schomburgkia X parkinsonian H. G. Jones, Jour. Barbados Mus. Hist. Soc. 28:

4. Barbados (parent species introduced).
Thrixspermum pygmaeum (King & Prantl.) Holttum, Kew Bull. 14:275.

Sarcochilus pygmaeus King & Prantl.

PRELIMINARY STUDIES IN THE GENUS STANHOPEA
(ORCHIDACEAE)

C. H. DODSON and G. P. FRYMIRE

Phylogeny and Pol!

The d tributi n .1 tlu -.r-ivie , J i . - l ' t ,;i„nship of the genus

sidered and it is believed that Stanhopea is derived from plants similar to members of
Skvekhtgh. The development of the advanced and complex species from the simpler

To say that the genus Stanhopea is poorly understood is an unders
Rather it is in a state of taxonomic chaos. Before we can say that a genus is under-
stood it is expected that an individual plant, taken at random from a population of
similar plants, can be fitted into a species concept. That is to say, it should be
possible for a trained botanist to determine whether an individual belongs to a
known species or is new and unknown. At the present time the only way to deal
with most individual plants of this genus is to name them as new species, thus
defeating the whole precept of taxonomy. Throughout its range of distribution
Stanhopea is characterized by populations which are extremely variable. The van-
in which are represented many kinds of Stanhopea hitherto considered as different

Because of the plant to plant variability within a single population, a great
number of species have been described from single specimens growing in European
greenhouses. Often the origin of these plants was not even known. Only the fact
that they were different from the other known specimens was taken into consider-
ation and that difference was considered sufficient reason for a new species descrip-
tion. It has been estimated that at least 500 names have been proposed in Stanhopea
(Ames and Correll, 1953). This estimate is high, actually only slightly over 100
names have been published. Recent works have reduced that number to between
25 and 50 which are considered as even reasonably valid.

One of the problems involved in the classification of this group is that the
flowers are large and fleshy and of intricate structure. They do not preserve well
as herbarium specimens and many of the structures of the lip in particular are
destroyed or altered in drying. The flowers are not long lasting and seldom persist
on the plant for more than three or four days. They are therefore difficult to study

(137)

138 ANNALS OF THE MISSOURI BOTANICAL GARDEN

in the field since not many flowers can be seen at any one time within a population.
Their period of flowering is often extended over a period of from five to seven
months and a plant which is not large will generally flower only once in a given

Since they are difficult to study in the he
only way to straighten out the taxonomic
persistent field work. Our work bears out this suggestion and we do not as yet
feel qualified to publish a revision of the genus even though we have studied a
considerable number of large populations. We have also studied most of the
available herbarium material, but find it hopelessly unrewarding. Such large fleshy
flowers as are found in this group do not, as a general rule, make good specimens
when crushed and dried. Even after boiling, the critical characters are often lost
and determination is difficult. Our experience indicates that dried specimens made
from the same plants on different occasions are so altered in drying that they can
easily be given separate species designations. This situation is particularly true in
certain of the critical species groups which are separated on small differences in
structure of the fleshy parts of the lip.

Perhaps all is not as hopeless as it appears on the surface. A number of workers
have made limited regional studies for orchid floras. A survey of their work
reveals that though they do not all agree, they do not seriously disagree in respect
to the majority of the species. In fact, some species appear to be remarkably
distinct and do not demonstrate significant variation within their populations.
Therefore they are given recognition by each of these workers when they occur
in regions treated by them.

When the species in which the problems occur are separated from the ones
which are distinct we find that the latter group is not a difficult one. Each species
has a distinct distribution and to a certain degree a distinct habitat. Within the
problem group the species vary tremendously yet they are in many respects remark-
ably alike, particularly in morphological characteristics. We feel that this is the
key to the problem. They are not only remarkably alike, they are alike and only
represent variations of a pattern If these complexes were treated as representing
fewer species each of which may vary considerably within a particular population,
then the genus probably could be treated taxonomically and could be understood.
Morphology

Let us consider just what combination of characteristics place a plant in the
genus Stanhopea. The vegetative characters are surprisingly uniform, so much so
that it is nearly impossible to distinguish species without flowers. All are epiphytic
or lithophytic. The pseudobulbs are usually large, ovoid, ribbed and unifoliate.
The leaves are broad, elliptic-lanceolate, contracted into a petiole at the base and are
strongly veined and plicate. The inflorescence consists of a short pendulous raceme
with several broad, papery bracts. The flowers are large, long pedicellate and
shielded by a papery bract, this often brightly colored and generally similar in
aspect to the sepals and petals. Nearly all species produce a strong fragrance, some
being agreeable and others not. The sepals are membranaceous and concave, the
dorsal sepal is free and erect and the lateral sepals broad, reflexed and connate at

DODSON & FRYMIRE THE GENUS STAN HOPE A

the base. The petals in most species are membranaceous, subequal to the dorsal
sepal and recurved. The principal characters which separate the species are to be
found in the labellum. In general, the lip consists of three parts: the hypochile or
globular, concave part nearest the base of the flowers; the mesochile, which is short
and often projected on each side to form horns; and the articulated epichile, which
is variously shaped but usually ovate. Certain of the species, such as S. ecornuta
Lehm. and S. pulla Rchb. f., simply have a saccate, entire lip which would corre-
spond to the hypocile of the more complex forms. These should perhaps be con-
sidered as the more primitive forms and later discussion will attempt to show why.
The column is rather similar in most of the species and is elongate and rather
slender with or without broad lateral wings. The anther is ventral, incumbent and
two-celled. The pollen apparatus consists of a lanceolate viscidium which rests on
an elongate spine-like rostellum, a stipe which is long and narrow, and the two
pollinia which are elongate, quite thin, and recurved. The stigma forms a pocket
under the rostellum.

Distribution
The genus is distributed from central Mexico through Central America and
into southern Brazil The plants are generally
ome shade is afforded and where they will rarely

>rthern South

ANNALS OF THE MISSOURI BOTANICAL GARDEN

become dry. They can withstand prolonged periods of moderate drouth but need
considerable water during their growing season.

Figure 1 shows the distribution of the genus. Stanhopea lewhae Ames & Correll
occurs in Guatemala and Costa Rica. Stanhopea ecornuta Lindl. occurs in Hon-
duras, Guatemala, Nicaragua and Costa Rica and S. cirrhata Lindl. is found in
Nicaragua and Costa Rica. Stanhopea pulla Rchb. f. is found in Costa Rica and
Panama. Stanhopea grandi flora Lindl.* extends from Trinidad through Venezuela,
northern Brazil and the Amazon drainage region of Colombia, Ecuador and Peru.
Stanhopea rodigasiana Hort. and S. reichenbachiana Roezl. ex Rchb. f. are little
known species which occur in Colombia. Stanhopea connata Kl. occurs in eastern
Ecuador and Peru and S. tricornh Lindl. is restricted to coastal Ecuador and
southern Colombia. The remaining species in the genus fall into two major com-
plexes. The first of these groups include S. insignis Frost, S. tigrina Batem., S.
martiana Batem. ex Lindl. and S. saccata Lindl. The second group is composed of
S. oculata Lindl., S. bucephalus Lindl., S. wardii. Lodd. ex Lindl., and a considerable
number of lesser known forms which are quite similar. These two groups occur
throughout the range of the genus from Mexico to southern Brazil; however,
certain forms are restricted to geographic regions.

Phylogeny
The genus Stanhopea was taken by Bentham (1881) as the type of the subtribe
and thus the name Stanhopeinae should be used in place of Gongorinae Schltr. (See
Dressier and Dodson, 1960). As close allies Stanhopea has the following genera:
Sievekingta, Polycycnis, Paphinia, Cirrhea and Gongora and perhaps Peristeria and
Houlletia. All of the species within these genera have a labellum which is more

DODSON & FRYMIRE THE GENUS STANHOPEA

; divided into a hypochile and an epichile. Figure 2 illustrates a possible
phylogenetic series demonstrating the relationships of the genera within the
Stanhopeinae.

We do not feel that a discussion of the phylogenetic relationships of the other
members of the Stanhopeinae would fall within the purpose of this paper and
therefore will only discuss those genera which seem to be directly involved with
Stanhopea in a stepwise progression from the simple to the complex. The genera
placed at the center of the diagram in Figure 2 have a rather simple labellum and
as the list radiates outward toward the more advanced genera the lip becomes
more complicated.

Figure 3 demonstrates our concepts of the origin and development of the
Stanhopeas from the poorly known genus Sievekingia, a genus of about seven
species. In general the plants are very similar to small immature Stanhopeas and
for this reason are often neglected by collectors. Very few specimens of these
plants are in herbaria and in several instances the type is the only specimen in
existence. Members of this genus were not discovered until late in the 19 th
Century and thus did not receive so many generic names as some of the earlier
discovered members of the subtribe. Without doubt certain of the species placed
in Sievekingia show fewer morphological relationships to each other than do the
genera to each other: Acineta, Lycomormium, Teristeria, Coeliopsis,

142 ANNALS OF THE MISSOURI BOTANICAL GARDEN

d d

Neomoorea, Lacaena and Lueddemania. The flowers of the Sievekingias are for the
most part less complex than the Stanhopeas. Stevekingia shepbeardii Rolfe has a
semierect inflorescence with rather simple flowers which have only a slightly lobed,
truncate lip with a simple fimbriate callus in the center. This is certainly the most
primitive known species at least in respect to the labellum. Stevekingia suavis
Rchb. f. and S. peruviana Rolfe have somewhat larger, erect side lobes which form
a concave hypochile at the base of the lip. From this point the close relationships
of the more primitive Stanhopeas are obvious. It is a simple step from Sievekingia
to Stanhopea ecomuta and S. pulla in which the lip is saccate and without the
complicated mesochile and epichile of the more advanced Stanhopeas.

Stanhopea cirrhata (Fig. 8) is probably an advancement from a plant similar
to S. ecomuta (Fig. 6) or S. pulla (Fig. 7) through the development of lobes along
the upper edge of the hypochile. In other respects S. cirrhata is rather similar to
the two primitive species except for the narrow, caudate wings formed along each

Eulaema bomboidt

side of the column. This may be the basic form from which the two major
branches of Stanhopea have developed. On one side, S. reichenbachiana (Fig. 10)
could have been derived from S. cirrhata and could well have been the form from
which Stanhopea grandi flora (Fig. 11) developed. Stanhopea grandi flora has a
fairly complex lip with short horns produced from the hypochile rather than from
the mesochile as in many species.

Stanhopea cirrhata could also be considered as similar to the ancestral form
from which S. lewisae (Fig. 13) developed. Stanhopea lewisae has very small,
slightly developed horns on the mesochile. Most of the advanced species with well
developed horns on the mesochile may well have been derived from S. lewisae or a
similar form.

Stanhopea tricornis (Fig. 9) shows little relationship to the other species in the
genus. The petals are thick and fleshy and are held parallel to the lip and column.
The lip has a hypochile which is only slightly saccate, no true mesochile and a very
thick and fleshy epichile. The horns are produced as an extension of the hypochile
and a sharp spine-like callus forms a third horn between the two lateral horns in
the place of a mesochile. Two possibilities can be postulated for the origin of this
unique species. First, it could have developed through evolution from the other
known species of primitive Stanhopea as an isolated form. No intermediates are
known, however, between this species and the other known Stanhopeas. The
second hypothesis may seem a bit far-fetched, but there is some morphological
evidence for derivation of S. tricornis from a species of Sievekingia, S. trollii
Mansf., recently encountered as sympatric with this Stanhopea. This Sievekingia
has petals similar to S. tricornis. The plant is so similar that it doubtless has been
passed over by previous collectors (including ourselves) as a young Stanhopea.
A colored slide taken of this species, without a scale to indicate size, was at first

144 ANNALS OF THE MISSOURI BOTANICAL GARDEN

determined as a highly floriferous form of S. tricornis. The lip of this Sievekingia
is saccate at the base, has two side lobes which flare outward and an epichile very
similar to S. tricornis. The callus of the lip is in some respects similar to the
center horn of S. tricornis. It seems feasible that S. tricornis may have been
derived by way of this species of Sievekingia. Stanhopea tricornis may possibly
have been derived as a result of hybridization between Sievekingia trollii or a
similar form and some Stanhopea sympatric with it in distribution. Stanhopea
bucephalus, S. tricornis and Sievekingia trollii are sympatric along the western
slopes of the Andes in Ecuador. If chance pollination of Stanhopea bucephalus by
the Sievekingia should occur the resulting intermediate hybrid might well be visited
by a pollinator not utilized by either parental species and thereby be stabilized as
a species. When an attempt is made to draw an intermediate between the morpho-
logical characteristics of the Sievekingia and Stanhopea bucephalus the result is
surprisingly similar to the labellum of S. tricornis.

Stanhopea rodigasiana (Fig. 12) is unique in the genus. The strictly one-
flowered inflorescence, ovoid, smooth pseudobulbs, bifid rostellum, round viscidium
and unique horns on the lip are characters not found in any other Stanhopea. In
some respects it suggests a possible relationship to Paphinia, but it is certainly close
to Stanhopea. It is surprising to us that this species was never placed in a mono-
typic genus. Although it is obviously closely related to Stanhopea it is sufficiently
different from the other species in the genus to allow any taxonomist with a narrow
generic concept to separate it without qualms of conscience. The reason for its
not having been so separated is very likely that it was not discovered until after the

been discovered 50 years earlier when Peristeria, Acineta, Lycomormium, and other
closely allied genera, were being separated, it would surely have been considered as
generically distinct. We do not believe that it warrants special distinction nor do
we believe that so many genera should be recognized in the Peristeria complex. We
hope that through more careful study of the entire subtribe Stanhopeinae we will
be able to develop a useful revision of its members.

The above mentioned species of Stanhopea are for the most part easily recog-
nized, have distinct morphological characters and distributional patterns. They
present no serious problems to the taxonomist. The remaining species within the
genus form the difficult portion. All of them have a more complex labellum than
the species previously discussed. The hypochile is saccate and of various sizes but
it follows a similar pattern in all of the species. The mesochile is short and inserted
on the apex of the hypochile with two lateral, elongate horns of various sizes. The
epichile is articulated to the apex of the mesochile and while it may vary in size
and proportions, they are all similar. Two major groups can be separated and while
we are not yet prepared to present a formal revision of the genus we feel that some
designation is necessary in order to be able to group the excessively large number
of species names in such a way that they can be discussed with relative ease. We
will use the terms "oculata complex" for those plants similar to S. oculata, and
"insignis complex" for the plants allied to S. insignis. The "oculata complex" will
therefore comprise those concepts which have a rather long narrow or rectangular

FRYMIRE THE GENUS STAN HOPE A

hypochile, large mesochile horns and always have an entire, acute or apiculate
epichile. The "oculata complex" comprises such taxa as: S. xvardii Lodd. ex Lindl.,
(Fig. 21), S. oculata (Lodd.) Lindl. (Fig. 19), S. bucephalus Lindl, (Fig. 20),
S. haseloviana Rchb. f., S. platyceras Rchb. f., etc. The "insignis complex" will
consist of those forms which have a saccate or globular hypochile, large mesochile
horns and either entire, acute or three-lobed epichiles. The "insignis complex"
comprises such species as S. tigrina Batem. ex Lindl. (Fig. 16), S. martiana Batem.
ex Lindl. (Fig. 17), S. insignis Frost (Fig. 15), S. saccata Batem. (Fig. 18) and
S. devoniensis Lindl.

Stanhopea quadricornis Lindl., if such a species exists, has a lip which is within
the variation pattern of S. oculata with the exception that the base of the thicken-
ings of the margin of the hypochile are projected into short horns. Apparently
only one specimen was collected and was named by Lindley. It is quite possible
that it was only an aberrant specimen or a hybrid and was not representative of
any population.

It seems apparent that Stanhopea is a rather terminal group in evolution, with
Coryanthes, a genus often said to be closely allied to Stanhopea, having a derivation

146 ANNALS OF THE MISSOURI BOTANICAL GARDEN

from some similar ancestor, but one probably more closely allied to Peristeria than
to Stanhopea. Cirrhea and Gongora appear to be the termination of a separate line.
In general, the Orchidaceae appears to be a rather recent but rapidly evolving
family in which less extinction is evident than in older families. Such a situation
often makes classification difficult due to the intergrading forms, but on the other
hand, it makes the tracing of phylogeny easier since closely related groups are still
present. We do feel that the Stanhopeas were derived from ancestors very much
like some of the species of Sievekingia which are known today. The species of
Sievekingia are not particularly successful, as is evident from their extreme rarity,
but may be considered as relict forms locally preserved in relatively hospitable
habitats. Such forms would very likely become extinct and be removed from the
record if they were members of plant families subjected to more stringent environ-

POLLINATION

This paper is perhaps somewhat premature in one respect, that is, the pollinators
and pollination mechanisms are not as yet known for all the species. However it
may be many years before all the information can be gathered. We feel that the
major worth of the paper lies in its discussion of the problems within the genus.

The pollination mechanisms and pollinators of two species have been well
worked out and photographed. Pollination studies of the other species will be
described when they are known. We have witnessed the pollination of both
Stanhopea tricornis and S. bucephalm in Ecuador.

A future revision of this genus may well be based in large part upon the ideas
gained from observations of pollination. Polymorphic genera are often difficult to
understand and when confusion is compounded by having highly variable popula-
tions, the naming of which has been multiplied on an unrealistic basis, such a genus

FRYMIRE THE

beco

nes chaotic. Occas

standpoint, can provide f

polli

nation mechanisms o

and

learly made obviou

gem

. The first system

s the flower after lar

sionally new techniques or observations from a different
lashes of insight for the investigator. The totally different
oi these kinds of Stanhopeas, when observed, immediately
is certain of the reasons for the polymorphism within the
is typical of most orchids in which the pollinator simply
nding on the expanded labellum. The transfer of pollinia
occurs as a result of proper positioning of the column and its sexual apparatus in
relation to the size and conformity of the particular pollinator. The majority of
the species of Stanhopea are of this type. These species are not extremely variable
and are easily recognized.

The second is a highly complex system in which the flight characteristics of the
bee are used in the pollination of the flower. After landing, the bee is clumsy and
has difficulty in regaining balance in flight. The bee, unable to grasp the slick
surfaces of the pendant labellum, falls through the flower. This characteristic
flight habit of the bee probably has contributed to the evolution of an entirely
different system of pollination. Though this kind of flower is distinct, it apparently
has been unable to preclude occasional hybrids within the related forms but has
simply led to more variation. Color and morphology of the unessential flower parts
appear to have been subject to little selection pressure for uniformity, and may
vary considerably from plant to plant within a given population. The resulting
confusion, caused by attempts to name every variant in cultivation, has not only
increased out of all proportion but has tended to make the genus nearly impossible
taxonomically.

Stanhopea tricornis occurs frequently in old epiphyte covered cacao trees in the
region of Quevedo in central, coastal Ecuador. The species generally succeeds best in
fairly dense shade and on a large plant the pendant two-flowered racemes are pro-
duced throughout the year. Several racemes may be produced by one pseudobulb.

148 ANNALS OF THE MISSOURI BOTANICAL

In late January of 1960 we succeeded in finding a plant which had two open flowers
(both had been pollinated) and two mature buds which would open the following
day. We removed the plant and carefully carried it to our hotel room. Before dawn
of the next morning we found the flowers to be open and emitting a heavy perfume
very similar to "Diorissimo" manufactured by the Christian Dior Company of

position where it could be easily photographed. At about 9 a.m. a large male bee,
Eulaema meriana (Oliv.) (tribe Euglossini, family Apidae, E. dimidiata is a syno-
nym) came to the flowers, which he proceeded to investigate, apparently searching
for the source of the strong fragrance. Meanwhile a second bee of the same species
arrived; neither bee seemed familiar with the flower and each searched for some time
before discovering the tunnel-like opening, formed by the petals, column and lip,
and entering the flower. It might be noted that this is the only species of Stan-
hopea in which the petals are fleshy and held closely appressed to the column. The
bee forced upward past the rostellar spine and entered the inner portion of the
flower to the hypochile (see Fig. 4). The bee did not attempt to extract food.
In backing out the bee encountered the rostellar spine once again. This time the
rostellar spine slid under the posterior projection of the metathorax of the bee, and
his struggles dislodged the lance-shaped viscidium of the pollinia, which immedi-
ately became cemented to the underside of the metathorax. In further backing
out he pulled the pollinia from the anther cap. The bee then flew away. Even if
the bee had entered the other flower, pollination could not have been accomplished
immediately for the stigmatic pocket of Stanhopea does not open sufficiently for
insertion of pollinia until several hours after the pollinia have been removed. The
following morning what appeared to be a different bee but one with pollinia
attached to his metathorax visited the flowers. The pollinia became lodged in the
stigma and were removed from the back of the bee, thus effecting pollination. In
total, five instances of pollination of Stanhopea tricornis were observed with six
bees, all E. meriana, involved.

Other species of Eulaema were noted in the immediate vicinity and were quite
curious about the camera, approaching to examine it, but showed no interest in the
Stanhopea tricornis flowers only a foot or so away. These species were E. tropica
(Linne) and E. cingulata (Fab.)* both of which are responsible for pollinating
numerous other orchids including several species of Catasetum, Cycnoches, Sobralia
violacea Lindl., Maxillaria grandiflora Lindl, Aspasia epidendroides Lindl. and
Pescatoria wallisii Linden & Rchb. f.

From these observations we can glean the following important points:

1. Eulaema meriana appears to be the specific pollinator of S. tricornis.

2. Eulaema tropica and E. cingulata are sympatric but showed no interest in S.
tricornis.

3. Eulaema meriana is attracted to the fragrance emitted by S. tricornis and
without prior experience with the flower must search to find the entrance.

FRYMIRE THE GENUS STANHOPEA 149

Flower color and structure are certainly secondary if evei

the actual attraction of the bee.

Flowers of S. tricornis were carefully examined for nectaries
structures which would offer food for the bee. None was found. The fragrance
is produced in the hypochile; the other parts, when removed from the flower, have
little or no fragrance. The inside of the hypochile is covered with long fleshy
hairs which the bee might eat but in the flowers entered by E. meriana these hairs
did not appear to have been disturbed. It appears then, that the bee is attracted
strictly on the basis of fragrance and does not receive food. He apparently does
not expect to receive food for he does not lower his tongue, which is so long that
it must be lowered and swung forward before a narrow flower is entered. This
pattern has been observed in other members of the Catasetinae and Stanhopeinae
in which male Eulaemas are attracted by the fragrance and the bees do not appear
to receive food of any kind.

In all of these orchids the flowers are generally not long lasting but produce
their strong, heady, spicy fragrance in such quantity that they can be detected by

150 ANNALS OF THE MISSOURI

the human nose, a notably weak olfactory organ, at distances of more than thirty
feet. The bees are usually quite wary when they first come to the flowers but after
being around them for some time seem to become drugged for they fly erratically
and can be approached easily. Only the male bees of Eulaema appear to be
attracted to this type of flower; the females, which are busy with nest building
and food gathering to provision the nests, do not approach these flowers. The male
bees have never been observed taking part in nest building or care of the nest and
seem to leave the nest as soon as possible and live a vagabond life. They secure food
from many of the same flowers as the female bees but also visit Stanhopeas, Cata-
setums, Cycnoches and other flowers which offer only the heady and seemingly
intoxicating fragrances. It is possible that the female bees are not attracted to
these flowers because they cannot sense them. It is known that female bees of
Apis mellifera, the domestic bee, have fewer olfactory pits in their antennae than
do the males. The only orchids in which both male and female bees were noted
were Sobralia violacea and Maxillaria grandiflora. Sobralia is a rather primitive
orchid which does provide nectar for the bee and which is visited by several bees of
different families. Maxillaria grandiflora provides starchy hairs at the base of the
lip. One morphological feature found in the male bees of the Euglossini which is
not present in the females is the presence of chemoreceptive hairs on the tarsi of the
front legs. The male bees brush these pads through the scratched surface of
the labellum of the orchids which they visit.

Stanhopea bucephalus is pollinated in a somewhat different manner. This vari-
able species is distributed along the western slopes of the Andes from elevations of
about 400 meters up to 1200 meters and is to be found on the summits of the
coastal mountains where they rise to these elevations. It is sympatric with S. tri-
cornis in the region of Quevedo and the two species appear to be sympatric for at
least 60 kilometers to the south.

Pollination of this species was observed during the month of March at Olimpo
on the railroad from Guayaquil to Quito. Unlike S. tricornis this species has thin,
narrow petals which became retracted and curl up around the dorsal sepal shortly
after the flower opens. The horns are not produced as an extension of the hypochile
as in S. tricornis but are larger and developed from a definite mesochile. Thus there
is left a large open space between the lip and the column at the level of the hypo-
chile. The lower part of the column is effectively surrounded, to form a tunnel,
by the horns of the mesochile, the epichile and the wings of the column. The
pollinator of this species is also a male Eulaema, E. bomboides Friese, but this one
is only about two thirds as large as E. meriana. Eulaema bomboides is a little larger
than either E. cingulata or E. tropica.

Eulaema bomboides is attracted by the fragrance of this Stanhopea which is
honey-sweet, quite unlike S. tricornis. The bees come to the flower of S. bucephalus
but immediately fly into the space between the column and the hypochile and
attempt to land on the hypochile, or they land on the side of the hypochile and
attempt to crawl around and enter the hypochile to reach the source of the fra-
grance. The surfaces of all parts of the Up are waxy and extremely slick, so
that the bee can not successfully land and hold fast. He falls down and is guided

FRYMIRE T

GENUS STANHOPEA

by the column and its wings at his back, the horns of the mesochile at his sides
and the slick epichile beneath him so that he falls directly through the tunnel. At
the lower end of the tunnel, almost at its exit, the rostellar spine with the viscidum
of the pollinia on its outer surface projects into the path of the falling bee and
is caught under the posterior projection of his thorax as in S. tricornis. The bee
hits the rostellar spine with such force that it gives way and allows the bee to fall
through. In doing so the viscidium of the pollinia is cemented firmly into place
under the thorax (see Fig. 6). The bee is apparently disturbed and tends to fly
away but he repeats the process when he encounters another flower. If the flower
is older with an open stigma the pollinia are this time placed in the stigmatic pocket
during his fall and pollination is effected.

The pollination of S. bucephalus is in many respects quite similar to the pollina-
tion of Gongora maculata Lindl. as reported by Allen (1956), with the exception
that the Gongora is pollinated by species of Euglossa. Euglossa is a genus of bees
which is very closely related to Eulaema and is also responsible for the pollination
of Coryanthes (Allen 1950). In Gongora the bee, hanging upside down loses its
hold on the slick surfaces of the lip, falls and is guided down the pendant column
during his fall with the result that he also makes contact with the viscidium of the
pollinia much in the same manner as Eulaema in S. bucephalus.

Several plants of this Stanhopea were brought back to the garden of the Insti-
tuto Botanico where quantities of Eulaema tropica and E. cingulata are common
and at no time did any of these bees show interest in the flowers. Even when
flowering plants of Catasetum were placed in the immediate proximity the Eulaemas
ignored the Stanhopea flowers completely.

152 ANNALS OF THE MISSOURI BOTANICAL GARDEN

From these observations we can add the following facts:

1. Eulaema botnboides appears to be the specific pollinator of S. bucepbalus in the
region studied.

2. Stanbopea bucepbalus is pollinated on an entirely different principle from S.
tricornh.

3. Eulaema bomboides is attracted by the fragrance of S. bucepbalus which is
different from the fragrance of S. tricornis.

4. If E. bomboides does not visit S. tricornis then S. tricornis and S. bucepbalus
are effectively isolated genetically.

The last point is by far the most important. Eulaema meriana is too large to
fall through S. bucepbalus. If E. bomboides is not attracted by the fragrance of

5. tricornis then the possibilities of the two species of Stanbopea forming hybrids
are effectively nil. This appeared to be the case. Stanbopea tricornis shows very
little variability throughout its populations. The sympatric species S. bucepbalus
is quite variable but shows no characters indicating introgression from S. tricornis.

If S. tricornis evolved in this region with E. meriana as its pollinator and S.
bucepbalus was dependent upon a different insect pollinator and a different pollina-
tion procedure the two would not hybridize. If by accident a bee of E. bomboides
did enter S. tricornis and effect a hybridization between the two species of Stan-
bopea, and it is feasible that this might occur, then the resulting hybrids would
probably find no effective pollinator. If such a pollinator did exist then the chances
might be better for development of a completely different, stable species with its
own characteristics and its own structurally isolated population.

Interestingly enough, perhaps not all Stanhopeas are pollinated by members of
the genus Eulaema. The type locality of Euglossa viridissima Friese is given as "at
Stanbopea tigrina at Cordoba and Orizaba in Mexico". If this report is correct,
and there is no reason to doubt its validity, it indicates that perhaps some of the
species are even more widely separated ecologically than are S. tricornis and S.
bucepbalus. Our observations indicate, at least in Ecuador, that Euglossas and
Eulaemas are not mutually attracted to the same orchid flowers. Where they are
attracted to orchid species of the same genus (i.e., Eulaema cingulata is attracted
to Cycnocbes lebmanii and C. ventricosum while Euglossa viridissima is attracted to
C. egertonianum) the flowers are extremely distinct morphologically. The prin-
cipal morphological difference between S. tigrina and S. bucepbalus which would be
of importance in pollination procedure lies in the extreme closeness of the column
and terminal portions of the lip in the former species whereas in the latter these
structures are rather open. A large bee like a Eulaema would not be able to fall
through S. tigrina due to this constriction of parts. On the other hand a Euglossa
would fall through S. bucepbalus without touching any of the parts and would
be ineffectual in pollination. We feel that the significant factor in this situation
does not lie in the morphological form being a deterrent to effective pollination, but
rather that the fragrance of the flower does not attract a pollinator which would
be ineffective. More likely the morphological structures of the flower have been
modified to accommodate the type of pollinator which was attracted to the flower.

DODSON & FRYMIRE THE GENUS STANHOPEA 153

ANNALS OF THE MISSOURI BOTANICAL GARDEN

~Z& .

Side view of the lip (
in (X Vi). i. Pollinar

Variation

We have pointed out that populations of species in Stanhopea are often ex-
tremely variable. Unfortunately no extensive analysis of specific populations are
yet available; however, we did superficially study two large populations in Ecuador.
The first occurs along the western slopes of the Andes from Macuchi in north-
western Ecuador to Paccha in the south near the Peruvian border. This population
is approximately 400 miles long and ranges from 400 to 1,200 meters in elevation.

Near Macuchi some of the plants correspond to a species named by Rolfe as
S. anfracta. This species is very close to S. wardii and had been considered by
Lindley as S. wardii var. venusta. Schweinfurth (1958) has placed it in synonymy
with S. wardii. Other plants are similar to what has been known as S. bucephalus.
The population was sampled at four locations through its central portion and was
rather uniformly S. bucephalus with only occasional plants occurring which could
be considered as referable to S. wardii. However, at Santa Isabela near the southern
extension of the population the situation changed considerably. Plants were so

DODSON & FRYMIRE— THE GENUS STANHOPEA 155

specimens were found which corresponded precisely to S. anfracta,
S. bucepbalus, S. graveolens, S. oculata, S. ivardii, S. haseloviana, S. guttulata, S.
,'. S. platyceras, and S. peruviana. Many plants also occurred in the
population which were intermediate between these concepts. Most of these species
are actually color forms and many have been relegated to synonymy by other
authors. Some of them have been maintained, largely because no botanist knew
much about them, for they were described from individual specimens cultivated in
Europe and their origins were unknown.

Plants from the population at Paccha, about 20 miles south of Santa Isabela
were sent to Europe by a number of early collectors. Kruth named a plant which
Humboldt and Bonpland had collected here as Epidcndrum grandiflorum. Lindley
later named another plant from the same locality as S. buccphalus and it was
illustrated in the Botanical Register, plate 24, in 1845. Reichenbach changed the
name to S. grandiflora, based on Kunth's name. Another illustration of a plant in
the Botanical Magazine, Plate 5278 given as S. bucephalus is so inaccurate that four

156 ANNALS OF THE MISSOURI BOTANICAL GARDEN

"species" can be identified from the same plate. Plate 8 5 1 7 in the Botanical Maga-
zine is actually a plant much nearer S. wardii. More recently S. bucephalus has
been reduced to synonymy under S. oculata. This confusion has pointed up
the results of describing species from individual specimens without knowledge
of the variation within populations. We feel that this situation is not uncommon
and clearly demonstrates the need for careful and extensive population studies of
the genus.

A second variable population of Stanbopea occurs in the valley formed by the
headwaters of the Zamora river on the eastern slopes of the Andes in southern
Ecuador. The population is extensive and we feel that it demonstrates another
of the kinds of phenomena involved in variation of the genus. The upper end of
the population reaches a point near Km. 45 on the road from Zamora to Loja, at

rains are the rule and temperatures are usually low, around 60° F.

Normally the sun breaks through only for an hour or two each day but occa-
sional days are sunny. When the sun is out for several hours the temperature rises
to as high as 75 ° F. The elevation decreases rapidly to the east and at Zamora, 62
km. from Loja, is about 1000 meters. At Zamora the valley levels out and becomes
broader, the normal daily temperature is around 75° F. and occasionally reaches
85° F. The weather conditions change little during the year but less rainfall
occurs in November and December. The population was studied as far as Yansasa,
about 60 km. northeast of Zamora. Here the elevation is near 800 meters and
weather conditions are similar to Zamora. In all, the portion of the population
studied extended 85 km. and varied from 1 km. in width at the upper end to 3 km.
wide at the lower end. Elevations run from 1800 meters at the upper end to 800
meters at the lower end.

Though they are variable in color at the lower end of the population the flowers
correspond to Stanbopea anfracta Rolfe. The flowers of plants at the upper end
are similar to the species described by Schlechter as S. peruviana and demonstrate
little variation. Both of these species have been reduced to synonymy under
S. wardii. A cline occurs from the upper end of the population to the lower, the
plants in the center having flowers essentially intermediate between the extremes.
Variation increases considerably toward the lower end of the population. Vege-
tatively the plants are indistinguishable throughout the population.

The flowers of the plants at the upper end of the population are morphologically
nearly identical to the flowers at the opposite extremes but are colored quite differ-
ently and are much thicker in substance. The flowers are completely tangerine-
yellow with no other color involved. The sepals and petals have heavy texture and
are not reflexed. The flowers usually last for eight days.

At the lower end of the population the flowers are thin in substance and have
their sepals and petals strongly reflexed. The basal part of the lip, sepals and petals
are yellow -orange flecked and spotted with maroon-red, and the extremities of the
sepals, petals and lip are white flecked with red. The flowers last only one or two
days. Plants from around Zamora had flowers more or less intermediate between
the two extremes in all characters.

From Zamora to Yansasa bees of the genus Eulaema were very com
species in particular was quite common and appeared much like E.
They were of the same size but slightly different in color pattern and probably
represent E. nivofasciata Friese. Though none was actually observed pollinating
Stanhopeas it is probable that they are the pollinators. Eulaema cingulata was also
common and while they visited other flowers placed in proximity they showed no
interest in Stanhopeas and would not enter them.

The bees of this genus are partial to the warmer tropical climates. They nest
in hollow limbs or in holes in embankments and their nests have not been encoun-
tered at elevations above 1,000 meters in Ecuador. From Zamora to Yansasa they
are commonly observed visiting Costus, Inga and Bixa. They are fast flying bees
and are capable of long flights. They are seldom seen in the upper end of the
valley and then only on the rare, sunny, warm days. It would appear that they do
not nest in the upper extensions of the valley. Plants in the lower regions com-
monly produce one or two seed capsules indicating considerable visitation by the in-
sects. In the upper regions, on the other hand, seed capsules are rarely encountered.

158 ANNALS OF THE MISSOURI BOTANICAL GARDEN

This population in some respects appears rather typical of the populations of
Stanhopea in Ecuador. It is not as complex as the population of Stanhopea on the
western slope of the Andes which is extremely variable in the direction of S. wardii
in its northern portion, rather stable in its central portion corresponding to S.
bucephalus and mixed with typical S. oculata in its southern extension. It is our
opinion that this type of variation is typical of S. oculata and allied species through-
out their range. It would appear that the relatively stable, heavy substanced, long
lasting type of flower at the upper extension of the Zamora population may be due
to strong natural selection for that type of flower. In the upper elevations where
pollinators are not abundant it would be of more value to the population to have
long lived flowers which would have a greater possibility of visitation and pollina-
tion. Such flowers may have to wait several days for weather conditions to be
satisfactory for bees to reach them. Short lived flowers would stand little chance
of being visited and therefore would rarely reproduce. Long life in these flowers
depends on thickness of their parts. A selection pressure is therefore exerted for

DODSON & FRYMIRE THE GENUS STANHOPEA 159

flowers with heavy substance. In the lower regions flowers are normally visited
shortly after opening and long lived flowers would offer no selective advantage.

Some botanists would attempt to attribute such extreme population variability
as is found in Stanhopea to hybridization and introgression. Perhaps this is a partial
answer but in many cases the variable populations are so widespread as to preclude
hybridization. One might also suggest that the variation is due to high mutation
rate. Perhaps this is also a partial answer. It is apparent from what has been
pointed out above in the pollination section of this paper that the bees which
accomplish the pollination are attracted by the strong fragrance of the flowers.
The color, and to a degree, the form of the flower appears to be of minor impor-
tance in its pollination. This is also true in the genera related to Stanhopea, such
as Gongora and Coryanthes.

160 ANNALS OF THE MISSOURI BOTANICAL GARDEN

(x :x

If the attraction of the fragrance is so strong, color may be a minor factor in
selection and the bees will still visit the flowers even if color varies widely. Also
form could vary widely in parts which are not essential to the actual mechanisms
of pollination, allowing variants due to mutation to remain within the population
as fully effective breeding members contributing to the gene pool as a whole. In
S. oculata and its allied species the characters which vary the most are the form of
the hypochile of the lip and the color of the lip, sepals and petals. These are char-
acters which are not actually involved in the mechanisms of pollination, and may
not be important in the actual attraction of the pollinator. Selection would not be
strong for these characters and a plant which deviated from the norm to a consid-
erable extent could still remain a breeding member of the population as long as it
retained the character of fragrance to attract the pollinator and the correct propor-

DODSON & FRYMIRE THE GENUS STANHOPEA 161

tions of the tunnel of the lip, formed by the mesochile and column, to properly
place the pollinia for transport by the bee to another flower.

The fact that cross-pollination is nearly obligate in these genera may be of
importance also. Certainly more variation would be produced, maintained and
dispersed in a population of obligate cross-pollinating plants than in a population
where self-pollination is possible and often of regular occurrence. Certain of the
genera in the Catasetinae, i.e., Catasetum and Cycnoches, are characterized by
unisexual flowers which make cross-pollination obligate. These are genera in
which the species are notoriously variable. In many of the other genera self-

162 ANNALS OF THE MISSOURI BOTANICAL GARDEN

incompatibility is the rule. Not a great deal of information is available on this
phenomenon but the authors have made numerous attempts to self-pollinate
Stanhopea bucephalus and Gongora maculata with no success. The resulting seed
pods dropped after a few weeks.

In most of these variable genera the flowers are highly fragrant but extremely
short lived, seldom lasting more than two or three days. The fragrance ceases very
soon after the flower has been pollinated. It would appear that there is a correla-
tion involved between the characteristics of powerful fragrance and short lived
flowers. These species seem to depend on their fragrance to attract pollinators and
are usually very successful, for a considerable portion of the flowers are pollinated.
They have developed in the direction of stronger and more attractive fragrances but
offer no food or nectar to the pollinator. Selection may even favor short-lived
flowers which do not attract bees long after their pollination (i.e., do not compete
with the younger flowers). Here there seems to be a strong selection pressure for
fragrance for pollinator attraction with lower selection pressure for constancy of
morphological characters.

Many of the orchids which have long lasting flowers are either not fragrant or
are weakly fragrant. These orchids seem to depend on numbers of flowers few of
which are ever pollinated. These species for the most part are not as variable as the
fragrant types and have developed other mechanisms for attraction of the pol-
linator. Where populations of these species are extremely variable, it is usually in
a relatively limited geographical area and their variability can often be attributed
to introgressive hybridization.

SUCCESSFULNESS OF THE GENUS

The objective measurement of success in a group of plants is difficult, as too
many factors are involved. Obviously, so called weedy species would rate high on
the scale. However, to attempt to rate other plants as being highly successful or
only fairly successful is scarcely possible. We do feel, however, that we can sub-
jectively rate the success of a group of orchids in relation to other orchids. In such
a subjective sense we can use extent of distribution and density of populations as
measuring tools. Many orchids are highly successful in respect to density of
population but are limited in distribution. Others have wide distributions but are
seldom encountered in quantity in a given area. The majority of the species of
Stanhopea are in one or the other of these categories. The "insignis and oculata
complexes" on this basis are highly successful in that they are commonly found
in quantity (in undisturbed habitats) and have broad ranges. They are to be
encountered in any region of the American tropics which provides the basic needs
of the plants. They are found from sea level to 7000 feet and grow as well in
either location provided there exists sufficient humidity and shade and the temper-

This very success, we feel, contributes to their variability, but on the other
hand, may have suppressed speciation. The extremes of the range may be expected
to produce forms which are unlike due simply to spatial relations. However
isolation is difficult to achieve in a dense and essentially continuous population. In

DODSON & FRYMIRE THE GENUS STANHOPEA

this respect "success" and taxonomic diversity may be mutually incompatible.
This situation is by no means limited to Stanhopea. The hooded types of Cata-
setum reflect the same phenomena, though not quite to the same degree. In Mexico
and middle Central America a form known as C. integerrimum Hook, is common
and variable. Further south in Costa Rica and Panama intergrading populations
known as C. viridiflavum Hook., C. oerstedii Rchb. f. and C maculatum Kunth
occur. From Panama extending through Venezuela, the Guianas and northern
Brazil the form known as C. macrocarpum L. C. Rich, is found. In Ecuador C.
macroglossum Rchb. f. is common. All of these populations are highly variable
and in Ecuador alone, specimens corresponding to all of these "species" may be
found. There is little reason to recognize these species when they are now known
5 of one widely distributed and highly successful popula-

164 ANNALS OF THE MISSOURI BOTANICAL GARDEN

tion. Here, as in Stanbopea, speciation seems to be inhibited by the otherwise
highly successful population structure.

Gongora maculata Lindl. demonstrates a close parallel. Several species of
Cycnoches, Mormodes, Coryanthes, Epidendrum and Oncidium have similar pat-
terns. Many other instances of highly successful, widely distributed orchids can
be given in which the most striking chracteristic is their extreme variability. We
feel that speciation has taken a back seat to success in the sense that development
of distinct taxa is inhibited where gene flow is unrestricted and selection is weak.

Perhaps the most successful method of handling the great variability in Stan-
bopea will be to follow that used by Allen (1952) in his particularly workable and
useful revision of Cycnoches. Cycnoches reflects the same kinds of problems
encountered in Stanbopea and by placing several closely related, intergrading
populations, which had been previously known as species, as subspecies under one
inclusive species, Allen was able to indicate the close relationships of these popula-
tions. In our opinion this is the most practicable manner to handle the problems of
such extreme variability in obviously closely related populations. Certainly, it is
folly to attempt to apply species or even subspecies names to all of the variants
found; the number would be endless. There may be some difference of opinion as
to how many subspecies should be recognized, some authors suggesting more and
others less. When numerous well authenticated population studies are accomplished
in Stanbopea it is quite possible that certain of the dubious species within the
"insignis and oculata complexes" will be found to exist as discrete populations and
will be sufficiently distinct to be recognized, others will probably be reduced.

Recent correspondence with a rather well informed amateur orchid enthusiast
indicates that he has encountered a stable population of a Stanbopea. Specimens
which he has kindly forwarded appear near the type form of S. oculata. He main-
tains that these are different from typical S. oculata, are stable within what he
knows of the population and should therefore be recognized as a new species.
Perhaps with exhaustive study this population might be found to be distinct
throughout its extent and should therefore be considered as a valid species. Our
experience with populations of Stanbopea in Mexico, Guatemala, Costa Rica,
Panama, Ecuador and Peru would indicate that this population, throughout its
entire range, would very doubtfully represent even a valid subspecies of S. oculata.
This case, however, is quite representative of the problems which Stanbopea offers
and indicates the necessity of thorough knowledge of large percentages of over-all
populations before rash action is taken in applying more epithets in the group.

The nomenclature of Stanbopea is badly confused and the efforts made by
modern workers to clarify the situation have not always been particularly successful.
Many of the early descriptions of the species were not accompanied by illustrations
and were at best brief. Several of them are nearly impossible to identify with any
particular population. One of the unfortunate pitfalls the modern taxonomist
encounters is the tendency to make one of these early descriptions fit his concept
of a given species or vice versa. Normally the taxonomist has the type specimen

FRYMIRE THE GENUS STANHOPEA 165

to fall back on and this specimen, when carefully examined, supplies the necessar
information to avoid error. Unfortunately, herbarium specimens of Stanhopeas d«
not always fulfill this function even when they exist. The critical characters whicl
separate species in this genus are for the most part destroyed in the preparatioi
of specimens.

Many times the type specimen does not reflect the actual structure of the popu
lation. An example is to be found in S. wardii in which the type descriptior
mentions two teeth produced at the base of the hypochile by the extension of th<
falcate lateral margins. This character is common in the specimens from southerr
Central America and perhaps to a lesser extent in Guatemala, but actually does no
occur in many of the specimens from Mexico, Nicaragua and northern Soutl
America, which are certainly referable to this species on every other criterion. T<
further complicate the situation, this character does occur occasionally in othei
species such as S. oculata and S. bucephalus. Doubtless, the type specimen oi
S. wardii did have this character, but the presence or absence of this one character

166 ANNALS OF THE MISSOURI BOTANICAL GARDEN

since it is not diagnostic of the total population and is found in other species, should
not be considered as a criterion for determination. Several other characters help
to make this species distinct from S. oculata and S. bucephalus. Allen (1949) has
suggested that Lindley was in error in his description of S. oculata and S. bucephalus.
Allen placed S. bucephalus in synonymy under S. oculata and re-erected S. grave-
olens which had been considered as synonymous with S. bucephalus. We cannot,
however, agree with Allen and feel that the epithet, S. bucephalus, should be used
in place of S. graveolens. Since Lindley's descriptions are extremely brief and the
later illustrations which he used of his species correspond rather well with his
descriptions and with actual populations, we feel that Allen has allowed his splendid
knowledge of the Central American species to overshadow the variation of these
species in other areas. Since the type specimens yield little information it will
always be difficult to resolve this situation.

Another of the confusing situations involved in the taxonomy of Stanhopea is
the use of the name "grandiflora". This name was first used by Kunth when he
described and illustrated Epidendrum grandiflorum in 1808. In 1828 Loddiges
described another plant and named it Ceratochilus grandiflorus but the name Cera-
tochilus had previously been used by Blume for an East Indian orchid. In 1832
Lindley transferred the Loddiges plant to the genus Stanhopea of Frost and
described a similar species as Stanhopea eburnea. Later these two plants were
found to be conspecific but have been known as S. eburnea to the present time.
Lindley's name, S. grandiflora, is valid and must be used in place of S. eburnea. In
1832 Lindley also named a plant which was essentially identical to Kunth's plant as
Stanhopea bucephalus. The type specimen was a Ruiz and Pavon collection which,
as Rolfe (1912) has pointed out, is marked as being from Mexico. It appears
that many of Ruiz and Pavon's specimens were mixed with the Sesse and Mocifio
Mexican collections and a great deal of confusion has been caused by this situation.
The probable origin of the Ruiz and Pavon specimen was Peru, as it was then
called, or what is now Ecuador. It must have come from very near the same
locality as Kunth's specimen since the trails of access to the Andes were very few
during this period.

Reichenbach apparently attempted to clear up the situation by transferring
Kunth's plant to Stanhopea and reducing Lindley's S. bucephalus to synonymy
under S. grandiflora Rchb. f. Unfortunately, Reichenbach's S. grandiflora is
invalid because Lindley's combination precedes it.

Another attempt to clarify the situation in the "insignis complex" has been the
reduction of the name S. tigrina Batem. to synonymy under S. devoniensis Lindl.
by Williams (1951). However, the concept, early descriptions, and illustrations of
S. devoniensis and S. tigrina appear to us to be clearly distinct. The plate of S.
devoniensis in Lindley's Sertunt Orchidacearum has little resemblance to Bateman's
plate of S. tigrina in his Orchids of Mexico and Guatemala and very likely repre-
sents either a rarely encountered species or a hybrid form between S. saccata and
S. tigrina. The figure very definitely resembles S. saccata in the features of the lip
more closely than it does S. tigrina. Williams may have been misled by the color,
and was certainly handicapped by the lack of good type material. Occasionally

DODSON & FRYMIRE THE GENUS STANHOPEA 167

168 ANNALS OF THE MISSOURI BOTANICAL GARDEN

plants with flowers somewhat like the description and plate (discounting the three-
lobed epichile which is lacking in the plate but mentioned in the description) turn
up, but are quite rare in comparison to the number of typical S. tigrina found in
the Mexican populations.

Conclusions
This paper has been written as a discussion of Stanhopea preliminary to a formal
revision of the genus. Our original intention was to include the revision of
Stanhopea with this paper but after examination of herbarium material proved to
be disappointing we have concluded that only with further field study of living
populations can we accomplish a truly meaningful treatment of the genus. We
have also found that to merely revise Stanhopea and leave the other members of the
subtribe Stanhopeinae in the condition which they are at the present time would
be to do a less than meaningful work. The members of the Stanhopeinae are so
closely intermeshed with Stanhopea that revision of the whole group is mandatory.
We hope that in the not too distant future such a revision can be accomplished
using classical methods of taxonomy combined with modern biosystematic

In reality then, this paper can only point up the problems involved in the
group; it brings up to date the information that is available from the literature and
adds our own limited observations of the natural phenomena occurring in the
group. These field observations have pointed out the path, so to speak, that should
be followed in the study of such a complex and difficult genus as Stanhopea. It
may well be true that similar procedures will need to be used in treating other
problem groups in the Orchidaceae. Our experience indicates that future studies
of the difficult subtribe Catasetinae can only be approached through field study.
Probably parts of the Sobraliinae, Epidendrinae and Pleurothallidinae will prove to

Hypochile projected into fleshy lobes on its upper edges; column narrowly winged with
the apex of the wings caudate and extended beyond the anther 3. S. CIRR

DODSON & FRYMIRE THE GENUS STANHOPEA

d. Hypochile subsaccate provided with fleshy lobes on each side of the apex; c

short, hypochile conspicuously s
rell developed, hypochile not a

.ist of Species*

Stanhopea ecornuta Lem. in Fl. des Serres 2: t. 181. 1846. (Fig. 6).
Stanhopea calceolus Hort. ex Rchb. f. Xen. Orch. 1:117. 18 58.
Stanhope astrum ecornutum Rchb. f. Xen. Orch. 1:124. 1858.
Stanhopea pulla Rchb. f. in Gard. Chron. 7:810. 1877. (Fig. 7).
Stanhopea cirrhata Lindl. in Journ. Hort. Soc. 5:37. 1850. (Fig. 8).
Stanhopea tricornis Lindl. in Journ. Hort. Soc. 4:236. 1849. (Fig. 9).
^Stanhopea stenochila Lehm. & Kranzl. in Gard. Chron. 369. 1900.
t Stanhopea langlasseana Cogn. in Gard. Chron. 426. 1901.
Stanhopea convulata Rolfe, in Kew Bull. 366. 1909.
Stanhopea reichenbachiana Roezl. ex Rchb. f. in Gard. Chron. 2:40. 18
(Fig. 10).

Stanhopea amesiana Hort. ex Gard. Chron. 3 52. 1893.
^Stanhopea lowii Rolfe, in Kew Bull. 63. 1893.
Stanhopea suavis Hort. in Ospina, Orquidas Colomb.
Stanhopea grandiflora (Lodd.) Lindl. Gen. & Sp. Orch. 158. 1832. (I
11).

Ceratochilus grandiflorvs Lodd. in Bot. Cab. t. 1 41 4. 1828.
Stanhopea eburnea Lindl. in Bot. Reg. 18: t. 152Q. 1832.
Stanhopea calceolata Hort. ex Rchb. f . 1 : 1 1 7. 1858.
Stanhopea calceolata Drap. Lenq. Hort. Univ. 2:127 & 264.
^Stanhopea Candida Barb. Rodr. Gen. Spec. Orch. Nov. 1:101. 1877.
Stanhopea randii Rolfe, in Kew Bull. 363. 1894.
Stanhopea rodigasiana Claes. ex Cogn. in Gard. Chron. 14. Fig. Q. 1 8!

(Fig. 12).
Stanhopea lewisae Ames & Correll, in Bot. Mus. Leafl. Harv. Univ. 10*:i

PI. 10. 1942. (Fig. 13).
Stanhopea connata Kl. in Otto & Deitz. Alleg. Gartenz. 22:226. 18;
(Fig. 14).

Stanhopea graveolens Kl. ex Rchb. f. Xen. Orch. 1:118. 1 879

ANNALS OF THE MISSOURI BOTANICAL GARDEN

"insignis complex"

Stanhopea insignis Frost ex Hook, in Bot. Mag. /. 1948 & 1949. 1829.

(Fig. 15).

Stanbopea flava Lodd. ex Beer. Prakt. Stud. Fam. Orch. 312. 18 54.

Stanbopea atropurpurea Lodd. ex Planch. Hort. Donat. Orch. 215. 1858.

Stanbopea odoratissima Hort. ex Planch. Hort. Donat. Orch. 216. 18 58.

Stanbopea macrochila Lem. in Illustr. Hort. 4: Misc. 71. 1859.
Stanhopea tigrin a Batem. ex Lindl. Sert. Orch. 1.1838. ( Fig. 16).

Stanbopea cavendisbii Lindl. ex Batem. Lond. Hort. Brit. Suppl. 3:643.

Stanhopea maculosa Knowles & Westc. Flor. Cab. 3: /. 1 21. 1839.

Stanhopea nigroviolacea Morren. ex Beer. Prakt. Orch. 313. 18 54.

Stanhopea expansa P. N. Don. Hort. Cantab. 13:721. 1845.

Stanhopea lyncea P. N. Don. Hort. Cantab. 13:608. 1845.
Stanhopea devoniensis Lindl. Sert. Orch. t. I. 1838.
Stanhopea martiana Batem. ex Lindl. in Bot. Reg. 26: Misc. 50. 1840.

(Fig. 17).

Stanhopea velata Morren, in Ann. Soc. Gand. 3. t. 153. 1847.
Stanhopea saccata Batem. Orch. Mex. & Guatem. t. 15. 1839.

Stanhopea implicata Westc. ex Lindl. Fol. Orch. Stan. 6. 1852.

Stanhopea radiosa Lem. Illustr. Hort. 4: Misc. 72. 1859.

Stanhopea marshii Rchb. f. Xen. Orch. 1:120. 1879.

Stanhopea elegantula Rolfe, in Kew Bull. 161. 1910.
Stanhopea intermedia Kl. in Act. Hort. Petrop. 17:142. 1898.
"oculata complex"
Stanhopea oculata (Lodd.) Lindl. Gen. & Sp. Orch. PI. 158. 1832. (Fig.

19).

Ceratochilus oculatus Lodd. in Bot. Cab. t. 1764. 1832.

Stanhopea lindleyi Zuccar, Abh. Akad. Muenich 2:320. 1831-33.

Stanhopea guttulata Lindl. in Bot. Reg. 29: Misc. 75. 1843.

Stanhopea graveolens Morren in Ann. de Gard. 2:55. /. 54. 1846.

Stanhopea guttata Kich. Berl. Allg. Gartenz. 364. 1858.

Stanhopea ornatissima Lem. in Illustr. Hort. 9:325. 1862.

Stanhopea cymbiformis Rchb. f. Xen. Orch. 2:84. t. 124. 1865.

Stanhopea minor Schltr. Notizbl. Bot. Gart. Mus. Berlin 6:483. 1917.
Stanhopea Bucephalus Lindl. Gen. & Sp. Orch. Pi. 157. 1832. (Fig. 20).

Stanhopea graveolens Lindl. in Bot. Reg. 3: Misc. 59. 1840.

Stanhopea aurantia Lodd. ex P. N. Don in Hort. Cantab. 608. 1845.

Stanhopea jenisbiana Kramer ex Rchb. f. Bot. Zeit. 10:934. 1852.

Stanhopea warsceivicziana Kl. in Allg. Gartenz. 20:214. 1852.

Stanhopea guttata Beer. Prakt. Orch. 312. 1854.

Stanhopea inodora Rchb. f. Xen. Orch. 2:157. 18 58.

Stanhopea aurata Hort. ex Planch. Hort. Donat. Orch. 104. 1858.

Stanhopea aurata Beer, P HO. 1858.

Stanhopea costaricensis Rchb. f. in Hamb. Gartenz. 16:424. 1860.

Stanhopea grandiflora Rchb. f. Walp. Ann. 6:587. 1863.

Stanhopea (H i. in Acta Hort. Petrop. 17:15. 1898.

Stanhopea lietzei Schltr. in Fedde Rep. Sp. Nov. 16:248. 1919.

Stanhopea remota Hort.
Stanhopea wardh Lodd. ex Lindl. Sert. Orch. t. 20. 1838. (Fig. 21).

Stanhopea aurea Lodd. ex Lindl. in Bot. Reg. 4: Misc. 11. 1841.

Stanhopea venusta Lindl. in Bot. Reg. 4: Misc. 11. 1841.

Stanhopea inodora Lodd. ex Lindl. in Bot. Reg. /. 65. 1845.

Stanhopea amoena Kl. in Allg. Gartenz. 20:273. 1852.

Stanhopea anfracta Rolfe, in Orch. Rev. 12:357. 1904.

Stanhopea pur pus ii Schltr. in Orchis 10:186. 1916.

Stanhopea peruviana Rolfe, in Bot. Mag. t. 8417. 1912.

DODSON & FRYMIRE THE GENUS STANHOPEA

Stanhopea annulata Mansf. in Orchis 16:19. 1938.

Stanhopea bicolor C. Koch. Berl. Allg. Gartenz. 209. 1857.

Stanhopea deltoidea Lem. in Illustr. Hort. 9:340. 1862.

Stanhopea Florida Rchb. f. in Gard. Chron. 615. 1879.

Stanhopea fregeana Rchb. f. in Otto & Dietr. Allg. Gartenz. 23:313. 1855.

Stanhopea fuerstenbergiae Kranzl. in Gard. Chron. 161. 1899.

Stanhopea gibbosa Rchb. f. in Gard. Chron. 1254. 1869.

Stanhopea haseloviana Rchb. f. in Otto & Dietr. Alg. Gartenz. 23:322.

1855.
Stanhopea hoppii Schltr. in Fedde Rep. Sp. Nov. 27:82. 1924.
Stanhopea impressa Rolfe, in Kew Bull. 196. 1898.
Stanhopea madouxiana Cogn. in Gard. Chron. 134. 1898.
Stanhopea moliana Rolfe, in Lindenia 7:89. /. 331. 1891.
Stanhopea nigripes Rolfe, in Kew Bull. 364. 1894.
Stanhopea platyceras Rchb. f. in Gard. Chron. 27. 1868.
Stanhopea ruckeri Lindl. in Bot. Reg. subt. 44. 1843.
Stanhopea schiixeriana Rchb. f. Xen. Orch. 2:158. 1858.
Stanhopea shuttleworthii Rchb. f. in Gard. Chron. 795. 1876.
Stanhopea uncinata Drap. Hort. Univ. 4:65. 1843.
Stanhopea xytriophora Rchb. f. in Gard. Chron. 842. 1868.

Dubious species
Stanhopea quadricornis Lindl. in Bot. Reg. 24: t. 5. 1838.

Stanhopea russelliana Lodd. ex P. N. Don in Donn. Hort. Cantab. 13:721. 1 845.
Stanhopea violacea Hort. ex Beer. Beitr. Morph. u Biol. Orch. /. 4 & 8. 1863.
Stanhopea wallisii Rchb. f . in Linnaea 41 : 1 09. 1 877.
Stanhopea hernandezii (Kunth) Schlechter, Beih. Bot. Centr. 2:490. 1918.

LrrERATURE CrTED

Allen, P. A. 1949. Flora of Panama (Orchidaecae). Ann. Mo. Bot. Gard. 36:53-59.

. 1951. Pollination in Coryanthes speciosa. Amer. Orch. Soc. Bull. 19:528.

. 1952. The Swan Orchids, A Revision of the Genus Cycnoches. Orch. Journ. l 4 ~ s :226.

. 1956. Pollination in Gongora maculata. Amer. Orch. Soc. Bull. 24:230.

Ames, O. & D. S. Correll. 1952. Flora of Guatemala (Orchidaceae). Fieldiana: Bot. 26 2 :528-536.

Dressler.'R. L. & C. H. Dodson. 1960. Classification and Phylogeny in the Orchidaceae. Ann. Mo.

Rolfe, R. A. 1912. Stanhopea grandiflora and S. bucephalus. Orch. Rev. 20:172.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

A. Euiaema bomboides attempting to land on the hypochile of Slanhopea bucephalus.
B. Bee falling through the flower and engaging the spine of the rostellum under his meta-
thorax. C. Pollinia being inserted into the stigmatic cleft while the bee is falling through
the flower. D. Euiaema moriana approaching the flower of S. tricornis.

DODSON & FRYMIRE— THE GENUS ST,

MISSOURI BOTANICAL GARDEN

STAFF

Direc

Frits W.

Went

DERSON,

John D. Dwyer,

Henry N. Andrews,

Norton H. Nickerson,

Bunting,

Trifon von Schrenk,

Hugh C. Cutler,

Owen J. Sexton,

Carroix W. Dodge,

George B

Robert L.

TaX t°h n e° A&E? Edit ° r °

Robert E. Woodson, Jr

BOARD OF TRUSTEES

President

Robert Brookings Smith

Vice-President

Leicester B. Faust

Howard F. Ba

Volume XL VIII

Annals

of the

Missouri Botanical
Garden

SEPTEMBER, 1961

Preliminary Revision of the Genus Dryrnaria . James A. Duke 173-268
Sanango : New Amazonian Genus of Loganiaceae .....

George S. Bunting and James A. Duke 269-274

pub:

Annals

of the

Missouri Botanical Garden

A quarterly journal containing scientific contributions fi
Missouri Botanical Garden and the Henry Shaw School of Be
Washington University in affiliation with the Missouri B

Beginning with Volume 45, 1958:

Subscription Price._ _^ $12.00 per volume

Single Numbers 3.00 each

Contents of previous issues of the Annajls of the Missouri Botanical
kbdes are listed in the Agricultural Index, published by the H. W. Wilson

Annals

of the

Missouri Botanical Garden

PRELIMINARY REVISION OF
THE GENUS DRYMARIA

JAMES A. DUKE
le days of DeCandolle, a revision of all the known species of Dnnunu is

The genus Drymaria comprised four species at the time of its establishment by
Willdenow (in Roem. & Schult. Syst. Veg. 5:406. 1819). Less than 150 years
later the number of names has multiplied more than thirty-fold, with most of the
novelties being described before the turn of the present century. No complete
revision has been attempted since this prodigious multiplication of epithets.
Wiggins (in Proc. Cal. Acad. 4 35 :189. 1944) has completed an able revision of
the genus as represented in the vicinity of the Sonoran Desert. His conclusions
were quite sound, and the only consequential differences expressed in the present

In addition to Wiggins, the most important recent American studies on Dry-
maria have been pursued by Fosberg, who has clarified many of the problems created
by M. E. Jones, proposed a few new taxa, and made several appropriate reductions,
both in the literature and in various herbaria. Macbride's account in the Flora of
Peru, although promulgating a few errors, has been a forward step toward under-
standing the genus in an unusually complicated region, where speciation in Dry-
maria has run rampant. I. M. Johnston's studies in the equally perplexing Coahuilan
Desert have resulted in the description of some of the most bizarre species of
Drymaria. The major contributor from Europe has been the late Dr. Mattfeld
whose discerning work has been responsible for the recognition of several new
South American species, all of them herein maintained. He has also clarified some
of the errors stemming from the work of his predecessor Briquet, who described
several new Mexican species, two of which were probably based upon Pavon's

(173)

174 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Gross Morphology

An amazing array of characters is exhibited in the genus, and will doubtless
tempt the description of minor variants. The petals, for example, are nearly as
variable as those of Silene. Almost every portion of the plant has been utilized as
a taxonomic tool at one place or another in the genus.

Various habits are assumed by members of the genus. Perhaps most common
is the clambering and sprawling habit exhibited by D. cordata, which occasionally
roots at the nodes. Erect suffrutescent perennials are not uncommon, and may
raise an eyebrow on those not intimately familiar with the genus. Depressed sub-
cespitose forms such as occur in some varieties of D. effusa recall some species of
Arenaria, while delicate subvirigate annuals such as D. lepiophylla may suggest
Stipulicida. Too often, the duration has been used as a key character, but the
duration of a plant is frequently a function of the environment. Plants grown in
more xeric habitats often exhibit more attributes of a perennial.

Although the leaves of most species are of an ovate type and often cordate at
the base, every extreme, from linear and subsucculent to reniform and membrana-
ceous, is found in one species or another. The leaves are usually entire, but undula-
tions occur in the leaf margins of some South American species. The opposite
arrangement prevails, but in a few series pseudoverticillate leaves are common.
Venation of the leaves is usually obscure and rarely of taxonomic significance, but
the texture and pigmentation occasionally aid in identification. The petioles may
be as long as the blades or completely obsolete. Most species have stipules which
may be entire, bifid, or irregularly lacerate, and the nature of the division is of
occasional importance as a taxonomic criterion.

Indumental variations are often called upon for specific delimitations. Septate
or moniliform hairs occur in several species. Forms of D. villosa may be hirsute
or villose in nature; forms of D. ovata are best described as lanose; stipitate or
sessile glands are present in many species. In D. cordata the glands commonly take
the form of a conspicuous farinose girdle on the pedicels. The indument is usually
more or less translucent, but is not uncommonly white, or even reddish or yellowish.
Completely glabrous species occur, and some of the xeromorphic and halophytic
species are decidedly glaucous.

Predominating in the genus is a terminal cymose inflorescence, but here again,
numerous departures from the norm exist. In most species with the leaves pseudo-
verticillate, the flowers reflect this arrangement. Rarely the flowers may be solitary
or clustered in the axils of barely modified leaves, but more commonly each flower
is subtended by more or less scariose bracts. Many inflorescences show racemose
tendencies; in D. anomala, the first and sometimes the second branches of the
inflorescence are dichasial, but subsequent branching tends to be monopodial. Pro-
portionate lengths of bracts, pedicels, peduncles and flowers, as well as relative
variations in their induments, have been used in distinguishing species.

Almost invariably there are five sepals in a flower, but aberrant flowers are
known with four or six sepals. The sepals are most commonly lanceolate or ovate
but in some they are nearly orbicular and in others nearly deltoid. One to nine
nerves may traverse each sepal, the 3 -nerved sepal being most prevalent. In some

DUKE REVISION OF DRYMARIA 175

few species the venation is more or less dendritic and in others so obscure as to be
difficult of interpretation. Occasionally the sepals are conspicuously carinate or
cucullate. Shape, apex, nervation, indument and comparison of the sepals are

Petals are usually five in number, but three is characteristic of certain species,
and the petals are absent or drastically reduced in D. apetala and in forms of
D. villosa and D. xerophylla. Typically the petal is bifid for about half its length,
but this basic construction is hard to visualize in some of the members of the first
few series whose ornate petals are so bizarre as to defy terse terminology to describe
them. The lobes in some more easily described forms are merely basally auriculate,

5 nature of the i

lobes. Most of the bilobed petals are tapered directly to a narrow claw,
some members (e.g. in Series fasciculatae) there is a longer or shorter trunk
interpolated between the lobes and the claw. In some forms (e.g. Series holo-
steoides) the trunk never narrows to a claw; such forms are herein referred to as
exunguiculate. The lobes of the bilobed petals are 1 -many-nerved, the 1 -nerved
condition being characteristic of the Series cordaTae, and certain species in other

Less constant in number than the petals are the stamens, which are occasionally
reduced to two in number. In species with the reduced stamen number, the anthers
are often abortive, and tend to be more orbicular than in those species possessing
the higher number of stamens; in pentandrous flowers the anthers are usually
oblong. Not infrequently the filaments are shallowly connate into a cup, and

such a cup. Rarely, as in D. stipitata, there are conspicuous saccate staminodia. In
the Series frutescentes occasionally floral dimorphism suggests that monoecism
is in the early stages of evolution.

Offering little in the way of systematic clues is the gynoecium. In a few species
the ovary is borne at the summit of a rather prominent stipe, the presence of which
does not seem to be necessarily constant. Some species, which are also characterized
by the development of staminodia, have a rather prominent stylopodium atop the
ovary. In most species, there are three styles, but occasionally there may be two
or four. Such numerical variations have been observed in a single inflorescence.
The degree of fusion of the styles on the other hand seems to be a more reliable
taxonomic criterion, the styles being fused to a greater or lesser degree in most
species. Only with hesitancy do I use the word fused; it would seem to be just
as well to speak of one style with three branches as to speak of three styles partially
fused. For the sake of tersity, I have often described the style as being bifid or
trifid half its length, when it might have been more correct morphologically to say
that there were two or three styles fused for half their length. Interestingly the
two or three styles are almost completely free in D. cordata, a species which has
the most reduced androecium, and a species which I would judge to be one of the
most advanced in the genus.

The capsule may be three-valved or two- or four-valved, correlated with the
number of styles. Rarely the size of the capsule may be diagnostic. Inside the

176 ANNALS OF THE MISSOURI BOTANICAL GARDEN

capsule, 1-many of the campy lotropous ovules may mature into seeds, and the
seeds have proven very valuable in the taxonomy of the group. The number, shape,
color and sculpture of the seeds all play some part in species determination. Most
species are characterized by many-seeded capsules, but in the Series cordatae, some
occasionally mature only one seed per capsule. The seeds may assume any of several
shapes. Perhaps most common is the cochleate, or snail-shaped seed, characteristic
of Series cordatae. Foetiform seeds, shaped much like the human foetus, are found
in the Series holosteoides, and the vestiges of this shape are retained in some
subsequent groups. Ampulliform, or retort-shaped, seeds occur in D. debilh
and this shape is approached by a few species of other sections. The seeds of D.
viscosa are unusual not only in being lacrimiform, but also in being smooth, dull
and yellowish. Many members of the earlier series have hippocrepiform seeds and
species of the Series lyropetala have hairs on the hump of the horseshoe. Seeds
of most species are characterized by various kinds of sculpture. In the earlier
series the seeds are merely granular, if we exclude the hairs of Series lyropetala
from the realm of sculpture. Later sections, especially in the Series villosae are
endowed with species whose seeds are ornate indeed. For example the tubercles in
D, villosa are stellate in outline, those in its subsp. palustris are often capitate, those
in D. divaricata var. stricta are secondarily tuberculate. On the other hand,
forms of D. ovata have smooth and lustrous black seeds. Usually the seeds are
reddish-brown, but in D. viscosa they are yellow; occasionally seeds of some species

Seeds have turned out to be rather critical indicators of a species' systematic position
in Drymaria. It should be pointed out however, that in working with Spergula,
J. K. New (in Ann. Bot. 23:23. 1960) has found variations in the number of
papillae on the seed coats to be environmentally induced. Furthermore she has
found no genetic correlation between the variations in seed-coats and in hairiness.
Species of Drymaria, whose seeds are highly viable a year after collection and
germinate rapidly in the greenhouse, should prove interesting subjects for similar

Geography

Departing somewhat from the "typical" picture of the Caryophyllaceae, Dry-
maria is largely a subtropical genus. Of the forty-eight species herein maintained,
all but two are strictly American, and only one subspecies does not occur in
America. As a whole in America the genus ranges from the western United States
through Central and South America to Argentina and Chile, the large majority of
the species occuring on the western halves of the continents. Florida, Surinam
and Brazil, on the eastern shore, seem to possess only the pantropical D. cordata.
For a geographical tabulation of species, the reader is referred to the citation of
specimens which terminates this paper.

Of particularly geographical interest is D. cordata, whose varietal potential and
geographical distribution has been studied in some detail by Mizushima (in Jour. Jap.
Bot. 32:69. 1957). Almost all material from continental Asia belongs to the
subspecies diandra (maintained as a species by Mizushima), a subspecies completely

DUKE REVISION OF DRYMARIA 177

lacking in America (excl. Hawaii).

rather distinct variety pacifica occu

large. In Africa however, there are about as many intermediates as there are pure

representatives of the two subspecies.

A second species occurs in Malaysia and, to judge from the literature, has been
introduced there only within the present century. Of course there is the more
remote possibility that it has merely been overlooked until this century. The
species, D. villosa, is so distinct that I prefer to believe that it is a recent introduc-
tion from the Americas.

Several interesting disjunctions are apparent in the genus. Most striking is that
illustrated by D. ladewii, known only from two collections in Bolivia, but repre-
sented by bountiful material from Guatemala and southern Mexico. More easy of
interpretation are several vicarious disjunctions between Baja California and main-
land Mexico, e.g. D. holostcoidcs & D. pacbyphylla; D. arcnarioides ssp. peninsidaris
& D. arenarioides ssp. arenarioides; D. gracilis ssp. carina ta & D. gracilis ssp. gracilis.

Economics
Often a conspicuous element in the herbaceous vegetation, Drymaria probably
plays in the subtropics the insignificant role played by the chickweed in temperate
North America, fairly well known, but hardly a topic for conversation. Drymaria
cordata, and perhaps others, which spread radially and root at the nodes forming
rather lush mats, have been planted as ground covers. In Ecuador various species
are used as curatives for liver and kidney ailments. In Mexico, species are reputedly
used to cure "yaza" and in the Orient, infusions are purportedly used to alleviate
headaches. One should be careful of the plant, especially as an oral administration.
Drymaria pacbyphylla (and probably related species) is seriously toxic to livestock,
and all measures should be taken to prevent its spread in the arid West, where
palatable fodder is at a minimum. Fortunately the seeds alone of D. pacbyphylla
are so distinctive (see Fig. 1C) as to separate it from any other species.

In the course of this study I have been fortunate enough to study all or most
of the material of the Gray Herbarium, the University of California, the New York
Botanical Garden, the Field Museum, the Pomona College Herbarium, the Missouri
Botanical Garden, the Rijksherbarium, the Leningrad Herbarium, the Herbarium
of the University of Glasgow, the Botanisches Museum of Berlin, the Botanische
Staatssammlung at Munich, the Museum National d'Histoire Naturelle at Paris,
and the University of Michigan. Critical material has also been loaned by the
curators of the herbaria at Kew, Prague, Stockholm, Lija, Lund, Geneva, Gottingen,
the University of Brazil, and the California Academy of Sciences. Dr. Ramon
Ferreyra has forwarded many interesting Peruvian specimens for determination.
To him and to the curators and my correspondents at all the aforementioned
, I express my sincere gratitude for making their time and efforts avail-
To my colleagues here at the Missouri Botanical Garden, I am indebted

178 ANNALS OF THE MISSOURI BOTANICAL GARDEN

for their suggestions and comments. * Special thanks are extended to my wife,
Peggy, who has been kind enough to prepare the illustrations.

DRYMARIA Willd.
Drymaria Willd. ex Roem. & Schult. Syst. 5:31. 1819.
Pinosia Urban, in Arkiv. Bot. 23A 5 :70. pi. 2. 1930.
Mollugopbytum M. E. Jones, in Extr. Contr. West. Bot. 18:3 5. 1933.

Annual or perennial, glabrous or pubescent herbs, occasionally subligneous
below, prostrate, spreading or erect. Leaves opposite or pseudoverticillate, glabrous
to villose or glandular, sessile to long-petiolate, usually with persistent or fugaceous
small stipules. Flowers few to many in dichasial cymes, rarely in pseudoverticils,
racemes, or solitary or clustered in the leaf axils. Sepals 5, not connate. Petals
(0-) 3-5, white, usually 2-cleft, occasionally appendiculate in the sinus, often
auriculate. Stamens 2-5, the anthers versatile, 2-celled, the flattened filaments
slightly connate at the base, rarely alternating with prominent staminodia. Ovary
superior, slightly stipitate; carpels mostly 3, the 3 styles more or less united below;
ovules few to many, campylotropous on free central placentae. Capsule ovoid to
spheroid, usually dehiscing into 3 entire valves; seeds 1 to many, cochleate, foeti-
form, hippocrepiform, or ampulliform, usually tuberculate, the embryo curved
about the perisperm.

As here conceived the genus consists of 48 species, all but two of them exclu-
sively American. Since the genus has never been monographed in its entirety, and
no supraspecific classification has been elaborated, I would like to propose the
following informal series, most of which seem to be rather homogeneous. Follow-
ing a tabular presentation of the series is an artificial analytic key to all species.
Following this is a systematic key to the series. The latter key, although not
assuredly phylogenetic, is so set up that species with the more elaborate floral
characters, often correlated with certain vegetative factors, come first, and the
species with simpler floral construction come later. I am not prepared to say
whether evolution in the genus has been a matter of reduction or amplification

Supraspecific Classification

A. Series holosteoides

D. Series arenarioidi

1. D. holosteoides

2. D. pachyphylla

8. D. molluginea

9. D. arenarioides

B. Series lyropetala

4. D. subumbellata

5. D. suffruticosa

6. D. lyropetala

U. D.barkleyi

12. D. poly car poide

13. D.viscosa

C. Series stipitatae

F. Series ortecioides

7. D. stipitata

14. D. ortegioides

could not be used to full advantage.

cindly rev,
Dr. Fosber

iewed an early draft of this pa

DUKE REVISION OF DRYMARIA

N. Series villosae

D. hypericifolia
D. longepedunc

L. Series debiles

29. D. debilis
M. Series ]

30. D. laxiflora

', D. glaberrh
, D. monticol

Q. Series

44. D. gn

45. D. «/,

Artificial Key ■

4. Sepals 7.5-9.0 mm. long; Pe.
2. Sepals obtuse

r polylacerate apically:

ANNALS OF THE

BOTANICAL GARDEN

tracts clearly different

Manes glandular; petals
. Sepals and petals 4-7
. Sepals and petals 2.8-

' with 4 lobes; sepals

Sepals withont stipitate glands; continental No

21. Leaves ell.pnc

lepals acute or strongly 3 -ribbed, if obtuse, the fl

.Sepals obtuse (D.effusa):

23. Bract* mostly shorter than the pedicels:

s mostly longer than the pedicels; plat
ides stipitate-glandular; pedicels (i

Plants rarely suffruticose; No
34. Sepals all obtuse, often c

DUKE REVISION OF DRYMARIA

Peduncles and pedicels stipitate-glandular 18:

. Cyanic glaucous perennials of the Coahuilan Des

Pedicels mostly longer than t

49. Sepals acuminate, glabrous.
. Petals with a definite claw, the t

Leaves broadest below
S3. Sepals 4.5 mm. long

182 ANNALS OF THE MISSOURI BOTANICAL GARDEN

; seeds with simple or c

. Seeds with ornate tubercles; sepals no more t
61. Sepals glabrous, obtuse

62. Sepals obtuse; Venezuela and Colombi

. Sepals oblong to orbicular, ob
ciliate auricles; tubercles of the st

entire; Chiapas, Guatemala; Bolivia...,

I mm. long, the tubercles rectanguloid,

DUKE REVISION OF DRYMARIA

. Sepals glandular or villosulous:
78. Petals clearly exerted, the

r rarely carinate, the i,

Sepals rarely carinat

viscid-glandular;

184 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Sepals and herbage subglabrous to 1
flowers at anthesis broader, terete; lc

43d. D. DIVARICA!

Leaves long-petiolate; sepals basally saccate

>7. Sepals strongly nerved, cannate; leaves
often crenulate

Key to the Series of Drymaria

or hippocrepiform, not dorsally hispid Series A: I

pseudoverticillate, stipulate; seeds, if foetiform, dorsally hispid:

inear; plants upright perennials; petals with many distal lobes; seeds foetiform,

t foetiform, not dorsally hispid:

■i-r linear, <- c : - • • ,- . ■.

own, facially transparent, revealing the yellowish embryo. Baja California

e or glandular; seeds tuber-

DURE REVISION OF DRYMARIA

imbricate. Cuba.

, Large-flowered suffrv
. Small-flowered delica

. Glabrous to villose plants wit!,
the lobes sometime crrni rinate or auriculate; stamens 2-5, if unequal, the

' exauri.-ulaic/ Western U. S. to Central America. .. ..." .....'....'.

m. Trunk of the petals shorter than the claw (except D. glaberrima), the

seeds stellate, cylindric or capitate; North an'd S
o. Petals exaur .1 - i not ciliatc: tul cycles of the seeds

or mammiliform auricles; leaves short- to long-

membranaceous, net densely villose Series P:

. Lobes of the petals 1 -nerved (except D. glandttlosa \

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Leaves elliptic, pseudoverticillate, estipulate, often somewhat glaucous and
succulent. Flowers in pseudoverticils. Sepals obtuse, often somewhat cucullate,
the venation subdendritic. Seeds foetiform or hippocrepiform, merely granular,
occasionally gibbose and umbonate. Two halophytic species of western United
States and Mexico, both probably toxic to cattle. Fig. 1.

Leaves and stems often stipitate-glandular;
exceeding the sepals; seeds often strongly v

icous, rarely stipitate-glandular, often tinted with s
sepals usually longer; seeds rarely gibbose

HOLOSTEOIDES

Drymaria holosteoides Benth. Bot. Voy. Sulph. 16. 1844. (hc

at Magdalena Bay; Kl; isotypes at GH & LE.)
Drymaria veatchii Curran, in Proc. Cal. Acad. 2 1 :227. 188

1888 on Cedros Island, UC!)
Mollugophytum holosteoides M. E. Jones, Extr. Contr. West. Bot. 18:3 5. 1933.

Radially spreading herbaceous to lignescent annuals to as much as 40 cm. long,
branching at nearly every node, stipitate-glandular to glabrescent, the internodes
mostly much longer than the leaves, with a caducous rosette of spatulate leaves.
Cauline leaves pseudoverticillate, the blades glabrous or scantily glandular, occasion-
ally somewhat glaucous, narrowly to broadly elliptic, apically rounded or acutish,
basally acute, 4-12 mm. long, 2-7 mm. broad, the petioles glabrous to stipitate
glandular, 2-8 mm. long; stipules absent. Inflorescences of terminal and axillary
contracted umbelloid cymes, subtended by both normal foliage leaves and scariose
ovate bracts 1-2 mm. long; pedicels 2-6 mm. long, glandular or stipitate-glandular.
Sepals 5, glandular or stipitate-glandular, broadly oblong, obovate or suborbicular,
apically obtuse and often cucullate, 3-4 mm. long, 1.6-3.0 mm. broad, the venation
obscure, dendritic; petals 5, 2.8-3.5 mm. long, bifid for about half their length,
with 4-6 laciniae in the cleft, these about half as long as the oblong lobes, the
trunk suborbicular, denticulate or crenulate, cordate or truncate to the minute
claw; stamens 5, ca. 3 mm. long, the oblong anthers ca. 0.6 mm. long; ovary at
anthesis globose, the styles attaining the anthers, bifid or trifid less than a third of
its length. Capsule subglobose, 1.5-3.0 mm. long, 6-20-seeded, the seeds foeti-
form, basally umbonate, laterally gibbose and often dorsally flanged, granular, 0.5-

. D. holosteoides var. holosteoide:
>X. B. D. bolosteoides var. rr«
pachyphylla (holotype) petal, 10!

1.0 mm. long, nearly or quite as broad as long. Fig. 1A (Holotype of D. veatchii).

Sandy soils of Baja California, Mexico, mostly north of Cape San Lucas.

The isotype at Leningrad has unusually long styles to 2 mm. long which are
divided less than one-third their length, a character more frequent in the var.
crassi folia. Indeed the type is in some other respects (as well as aspect) inter-
mediate between these closely related varieties but by far the majority of its
characteristics are those of the more northern populations. Among other inter-
mediate specimens which forced me to reduce the following species to a variety
may be cited Constance 3185, which in its strongly gibbose seeds and stipitate-
glandular stems approaches the typical variety. However the specimen seems to
possess more characteristics of var. crassifolia, e. g. the petals are much shorter
than the sepals, the internodes are relatively short and the plant is obviously a

, crassifolia (Benth.) J. Duke comb. &

Drymaria crassifolia Benth. Bot. Voy. Sulph. 16. 1844. (ho

" Cape San Lucas; K !; isotypes at GH, LE !)
Mollugophytum crassifolium M. E. Jones, Extr. Contr. West. ]

188 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Subcespitose herbaceous to lignescent perennials verticillately branching but
rarely branching at every node, glabrous or very rarely stipitate-glandular, to as
much as 20 cm. long, some of the internodes occasionally shorter than the leaves,
with a caducous rosette of spatulate leaves, the taproot to 8 mm. broad. Cauline
leaves pseudoverticillate, the blades glabrous, carnose, glaucous, often cyanic,
narrowly to broadly elliptic, 4-12 mm. long, 3-8 mm. broad, the petioles 4-12 mm.
long, glabrous or glandular; stipules absent. Inflorescences of terminal and axillary
contracted umbelloid verticils subtended by normal foliage leaves and by bracts
1-2 mm. long; pedicels 5-12 mm. long, glabrous to stipitate-glandular. Sepals 5,
glabrous to scantily sessile-glandular, broadly oblong to elliptic, apically obtuse and
often cucullate, 3.0-4.5 mm. long, 1.5-2.5 mm. broad, the venation obscure,
dendritic; petals 5, 2.0-3.5 mm. long, bifid for about half their length, with 4-6
laciniae in the cleft, these about half as long as the main lobes; trunks ovate,
denticulate, cordate or truncate to the minute claw; stamens 5, ca. 3 mm. long,
the oblong anthers ca. 0.6 mm. long; ovary at an thesis subglobose, the styles attain-
ing the anthers, bifid or trifid less than a third of their length. Capsule subglobose
to ovoid or ellipsoid, 3-5 mm. long, 7-20-seeded, the seeds hippocrepiform, granular,
scarcely umbonate or gibbose, 0.8-1.0 mm. broad, nearly or quite as long as broad.
Fig. IB. (Holotype).

Saline soils around Cape San Lucas, Baja California, Mexico.

Wooton & Standi, in Contr. U. S. Nat Herb. 16:
Wooton 405, dry plains south of the White Sands,

N. Mex., US; isotypes at F, GH, MO, UC, etc.!)

Glaucous subsucculent annuals, the vegetative branching largely confined to
radially diverging branches from a slender yellowish rootstock or to terminal
pseudoverticillate branch systems, the elongate spreading internodes much exceed-
ing the terminally crowded leaves. Leaves pseudoverticillate, glaucous, subsuccu-
lent, wrinkling in drying, broadly elliptic to suborbicular, apically obtuse or acutish,
basally tapering to the petiole, 4-14 mm. long, 4-12 mm. broad, the glabrous
petioles 4-8 mm. long; stipules absent, the petioles clasping. Inflorescences of
terminal and axillary contracted umbelloid verticils subtended by normal foliage
leaves and bracts 0.5-1.5 mm. long, the pedicels 1-5 mm. long, the bracts ovate,
obtuse, scariose and almost nerveless. Sepals 5, subequal, glabrous, glaucous,
ellipsoid, obtuse, 2.5—3.3 mm. long, 1.5-2.0 mm. broad, obscurely 3— 5 -nerved, the
central portion green, the margins scariose; petals 5, 2.5-3.0 mm. long, bifid about
half their length, with 2 oblong laciniae in the cleft, the trunk serrulate, tapered
to the base, the claw not clearly delineated; stamens 5, 1.0-1.5 mm. long, the
oblong anthers ca. 0.5 mm. long; ovary at anthesis subglobose, the short style bifid
or trifid more than half its length, slightly exceeding the anthers. Capsule sub-
globose, 3-4 mm. long, mostly exceeding the sepals, 15-25-seeded, the seeds
vermiculiform, facially gibbose, more or less tessellate, 0.8-1.8 mm. long, ca. twice
as long as broad. Fig. 1C (Holotype).

Heavy saline soils in denudated or ruderal sites, often as a pioneer on bare
areas, western U. S. A. and northern Mexico.

DUKE REVISION OF DRYMARIA 189

Although certain authorities have regarded this species as synonymous with
D. holosteoides, there seems to be no justification for such a conclusion, as the
mainland species differs constantly in many respects from the preceding species of
Baja California. The seed alone would suffice to distinguish this species from all
other species of Drymarra.

Little (in Ecology 18:416. 1937), who regards D. pachyphylla as a synonym
of D. holosteoides, presents some very interesting information about the ecology
and toxicity of this serious weed. All the aerial portions of the plant are
toxic to cattle and sheep. To complicate matters, the xeromorphic nature of the
leaves is responsible for a degree of resistance to drought and to chemical sprays.
Campbell (in Journ. Agr. Res. 43:1027. 1931) has shown the species to be a
pioneer in plant succession on clay soils such as the adobe soils of portions of our
western ranges.

B. Series lyropetalae
Leaves linear, pseudoverticillate, estipulate or with entire caducous stipules.
Flowers in pseudoverticils or in cymes. Sepals obtuse or acute, the venation sub-
dendritic. Petals apically 6-14-lobed, laterally denticulate, minutely unguiculate.
Seeds hippocrepiform or foetiform, dorsally hispidulous, etuberculate. Four gypso-
phytic species of the Coahuilan Desert of northeastern Mexico. Fig. 2.

a. Inflorescence pseudoverticillate, compact, the bracts mostly contiguous; petals mostly 6-8-
lobed, the lobes lance-deltoid, attenuate, about half as long as the trunk, the trunk occa-

. Sepals 3.0-4.5 mm. long, often glandular
long, the ligular staminodia 0.2-0.5 mm.

Sepals 5-6 mm. long, glabrous or bull;

dSinlilir/bra^ts °Q*".5 mm/long" plan" glLZlll^"—'..^!... T!"Z. dTyrotetala

3. Drymaria elata I. M. Johnston, in Jour. Arn. Arb. 21:68. 1940. (holo-
type: I. M. Johnston 7823-, 10 km. s. of Laguna del Rey, locally abundant in
gypsum silt; Coahuila. GH!)

Upright lignescent perennials, the branches opposite or pseudoverticillate,
glabrous or glaucous to stipitate-glandular, the internodes mostly longer than the
leaves, the tap root to 1 cm. thick. Leaves opposite or pseudoverticillate, glabrous
or glaucescent, succulent, obscurely veined, linear, apically obtuse, 1-7 cm. long,
0.5-1.5 mm. broad. Inflorescences of terminal contracted 5-13-flowered umbelloid
cymes, the peduncles mostly 5-8 cm. long, glabrous; bracts 0.5-1.0 mm. long,

190 ANNALS OF THE MISSOURI BOTANICAL GARDEN

DUKE REVISION OF DRYMARIA 191

, the mature pedicels 4-7 mm. long, puberulent or stipitate-glandular,
the outer ones spreading or deflexed. Sepals 5, 3.0-4.5 mm. long, 2-3 mm. broad,
stipitate-glandular or glabrate, broadly ellipsoid to suborbicular, apically obtuse,
the venation obscure, with 7-9 veins arising from the base, all but the midrib
dichotomizing; petals 5, 2.5-3.5 mm. long, with about 6 subequal apical laciniae,
these lance-deltoid, falcate, attenuate, and denticulate, the trunk longer than the
laciniae, apically bilinguate, laterally lacerate, ca. 1.5 mm. broad, basally cordate,
the claw binute; stamens 5, 2-3 mm. long, the oblong anthers 0.5-0.8 mm. long;
staminodia conspicuous, flap-shaped, 0.4 mm. high, 0.8 mm. broad; ovary at
anthesis ovoid, the style slightly exceeding the anthers, trifid less than one-fourth
of its length. Capsule subglobose, exceeding the sepals, 3.0-4.5 mm. long, 10-
20-seeded, the seeds hippocrepiform, 1.0-1.5 mm. long, densely hispidulous dorsally,
ventrally slightly umbonate, the facies merely puncticulatc. Fig. 2A. (Holotype).

Known only from saline soils in the vicinity of Sierra del Rey and Laguna del
Rey in Coahuila, Mexico.

The petals are completely unlike those of any other species and suggest the
dorsal profile of a beetle. This peculiar aspect is approached, but only slightly, by
the petals of the following species.

4. Drymaria subumbellata I. M. Johnston, in Jour. Arn. Arb. 31:188. 1950.
(holotype: 7. M. Johnston 84.8Q, s. end of Canada Oscuro near Tanque la
Luz, confined to gypsum beds on the escarpment, western Coahuila; GH!)
Upright or ascending lignescent perennials to as much as 25 cm. high, the
branches mostly opposite, the internodes longer or shorter than the leaves, glabrous
or glaucescent, the taproot to 1 cm. thick. Leaves opposite or pseudoverticillate,
the blades 1.5-4.0 cm. long, 0.5-1.0 mm. broad, glabrous or glaucescent, succulent,
linear, apically obtuse, attenuate to the sessile base; true. stipules absent. Inflores-
cences of terminal subumbelloid 3-9-flowered cymes, the peduncles 4-8 cm. long,
the bracts mostly imbricate, lance-ovate, 2.5-4.0 mm. long; mature pedicels 3-7
mm. long, glabrous, the outer ones spreading. Sepals 5, 5-6 mm. long, 2-3 mm.
broad, glabrous or bullate, broadly ovate, apically obtuse or acutish, the margins
involute, the venation obscure, with 7-9 veins arising from the base, all but the
midrib dichotomizing; petals 5, 4.5-6.0 mm. long, with 6-8 subequal apical
laciniae, these lance-deltoid, attenuate, falcate, denticulate, the trunk longer than
the laciniae, laterally lacerate, basally truncate to rounded, 1.0-2.0 mm. broad;
stamens 5, 3.5-4.0 mm. long, the oblong anthers 0.8-1.0 mm. long; staminodia
obsolescent; ovary at anthesis ovoid, the style slightly exceeding the anthers, bifid
or trifid less than one-fourth its length, nearly as long as the ovary. Capsule
subglobose, 5-6 mm. long, exceeding the sepals, 10-20 -seeded, the seeds hippo-
crepiform, 1.2-1.5 mm. long, dorsally provided with dense white hairs to 0.5
mm. long, ventrally umbonulate, the facies smooth or puncticulate. Fig. 2B.
(Holotype).

Known only from the type collection.

In the nature of the inflorescence and the petals, this bizarre species is inter-
mediate between D. elata and the other two species in this homogeneous series.

192 ANNALS OF THE MISSOURI BOTANICAL GARDEN

5. Drymaria suffruticosa A. Gray ex S. Watson, in Proc. Am. Acad. 17:328.

1882. (holotype: Palmer 74; San Lorenzo de Laguna, Coahuila; GH!;

isotypesatUS,MO,UC,etc!)

Upright suffrutescent glabrous perennials to as much as 30 cm. tall, the leaves
often fascicled at the geniculate nodes, the taproot to as much as 8 mm. thick.
Leaves pseudoverticillate, glabrous or glaucescent, somewhat succulent, linear,
obtuse, 10-25 mm. long, ca. 1 mm. broad, basally tapering to the stem; stipules
apparently lacking. Flowers globose in few-flowered terminal cymes, the peduncles
and pedicels glabrous; bracts lanceolate, 3-5 mm. long, about equaling the pedicels.
Sepals 5, 5-6 mm. long, 3-4 mm. broad, broadly ellipsoid, apically rounded but with
a minute acumen, strongly venose throughout, the small scariose margins involute;
petals 5, 4.5-5.5 mm. long, ca. 2 mm. broad, with 10-14 subequal apical laciniae,
laterally lacerate, basally truncate or tapered to the minute claw; stamens 5, 4-5
mm. long, the oblong anthers ca. 1 mm. long; staminodia cupular, minute; ovary
at anthesis renoid, the minute bifid style slightly exceeded by the anthers. Capsule
broadly ellipsoid, ca. 6 mm. long, often capped by a stylopodium, many-seeded, the
seeds hippocrepiform, ca. 0.9 mm. broad, dorsally provided with a row of stiff
hairs to 0.5 mm. long. Fig. 2C. (Holotype).

Known only from the Coahuilan Desert of Mexico.

6. Drymaria lyropetala I. M. Johnston, in Journ. Arn. Arb. 21:68. 1940.
(holotype: 7. M. Johnston 7594.; 3.5 km. s. of Cedral, gypsum plain, locally
common, GH!)

Drymaria lyropetala var. coahuilana I. M. Johnston, in Jour. Arn. Arb. 31:189. 1950.
(holotype: R. N, Stewart 567 [cited as Johnston 567 in original description]; 2 km.
s. of Santa Elena, Sierra de las Cruces, Gypsum Ridge, Coahuila, GH!)
Upright or ascending glandular or glabrous perennials to as much as 30 cm.
high, branching mostly at the base, the internodes shorter to longer than the
leaves, apparently elongating belatedly, the taproot to 1 cm. thick. Leaves
opposite or pseudoverticillate, 3-15 mm. long, 0.5-1.2 mm. broad, glabrous or
glandular, glaucescent, succulent, linear, apically obtuse, attenuate to the sessile
base, the upper leaves minutely stipulate. Inflorescences of terminal lax 3-
9-flowered cymes, the peduncles 4-60 mm. long, glabrous to stipitate-giandular,
the bracts 0.5-3.5 mm. long, remote, the mature pedicels 3-6 mm. long, glabrous
or glandular, not recurved. Sepals 5, 3-5 mm. long, 1.5-3.0 mm. broad, glabrous
to stipitate-giandular, ovate, apically obtuse, the venation obscure, with 7-9 veins
arising from the base, all but the midrib dichotomizing; petals 5, 3.2-5.2 mm.
long, with 10-12 unequal apical laciniae, these linear, obtuse, crenulate, the outer-
most as long as or longer than the trunk, the trunk laterally lacerate, basally
cordate to truncate, 1-2 mm. broad; stamens 5, 3-6 mm. long, the oblong anthers
0.7-1.0 mm. long; staminodia cupular, to 0.4 mm. high; ovary at anthesis sub-
cylindric, the style about as long, 1.5-2.5 mm. long, trifid about one-fourth its
length; stylopodium well-developed. Capsule globose, 3.0-3.5 mm. long, 3-
24-seeded, the seeds hippocrepiform, 1.0-1.5 mm. broad, ventrally umbonulate,

DUKE REVISION OF DRYMARIA

dorsaliy hispid, the hairs 0.2-0.3 (-0.5) mm. long, scanty and spreading. Fig.
(Holotype of var. lyropetala) ; Fig. IE (Holotype of var. coabuilana).

Local on gypsum soils of the Coahuilan Desert of Mexico.

The two varieties of this species may be separated by the following key:
a. Bracts 2-3.3 mm. long; leaves 5-15 mm. long, shorter than the internodes; plants glabrous

Johnston erred very slightly in his description of the more southern var. lyropetala,
stating that it was glabrous, the glandulosity occurring only in the northern
variety. The type of var. lyropetala has a few scattered glands. It should be noted
here that the UC isotype has leaves which are mostly longer than the internodes
while the converse is true of the holotype, which incidentally is less conspicuously

C Seri
7. Drymaria stipitata Fosberg, in Lloydia 4:281. 1941. (holotype: C. H.

Muller 3301; Mun. de Sierra Mojada, Sierra Mojada, Canon de San Salvador;

common in high oak chaparral; Coahuila; US!)

Upright suffrutescent perennials to as much as 20 cm. high, glabrous below,
becoming stipitate-glandular above. Leaves opposite, the blades 5-15 mm. long,
1.5-5.5 mm. broad, glabrous, somewhat carnose, almost veinless, narrowly ovate to
elliptic, apically acute, basally acutely tapering to the petiole, the petiole 1-2 mm.
long; stipules apparently lacking. Inflorescences of few-flowered terminal cymes,
the axes glandular-puberulent, locally densely so; bracts lance-ovate, 1.5-2.5 mm.
long, apically acuminate; pedicels mostly 3-5 mm. long. Sepals 5, subequal,
lanceolate to ovate, apically acuminate, basally umbonate, 5-6 mm. long, 1.5-2.5
mm. broad, transparent save for the 3 ribs; petals 5, 5.0-5.5 mm. long, apically and
laterally lacerate, the laciniae linear, often dichotomous, some of the lateral laciniae
directed downwardly; stamens 5, 5.0-5.5 mm. long, the broadly oblong anthers
ca. 0.8 mm. long, alternating with 5 clavate staminodia to as much as 2.2 mm.
long; ovary at anthesis ellipsoid, much shorter than the style, the style trifid for ca.
one-fifth its length, exceeding the anthers. Capsule ca. 4 mm. long, 6-8 -seeded,
the seeds 0.8-1.0 mm. long, symmetrically hippocrepiform, the surfaces irregularly
corrugated. Fig. 3 A (Holotype).

Coahuila, Tamaulipas and probably Nuevo Leon, Mexico.

D. Series arenarioides
Leaves linear to narrowly ovate, opposite or pseudoverticillate, subsessile, stipu-
late. Flowers in cymes, racemes or solitary in the axils of slightly reduced foliage
leaves. Sepals mostly obtuse and weakly nerved. Petals 2-8-lobed, the trunk
much longer than the claw, entire or denticulate. Seeds hippocrepiform and
dorsaliy sulcate or lacrimiform, nearly smooth or granular. Five species in western
U. S., northern Mexico and Baja California. Figs. 3B, 3C, 4.

194 ANNALS OF THE MISSOURI BOTANICAL GARDEN

(Lag.) Didr. in Linnaea 29:738. 1859. (hold-
type: "Lagasca 1806" annotated by De Candolle in the De Candolle Her-
barium; G. n. v.; photograph and drawing of type at GH!)

Alsine molluginea Lag. Gen. & Sp. 13. 1815.

Arenaria ? molluginea Ser. ex DC. Prodr. 1:400. 1824.

Drymaria sperguloides A. Gray, PI. Fendler 11. 1849. (holotype: Fendler 55; New

Mexico, GH!; isotypes at US, UC, F, Mo, etc.)
Lepigonum molluginea Fries, Ind. Hort. Sem. 1856.
Mollugophytum sperguloides (A. Gray) M. E. Jones, Extr. Contr. West. Bot. 18:3 5. 1933.

Slender erect virgate or dichotomously branched annuals to as much as 20 cm.
high, the internodes shorter to longer than the leaves, glabrous or with sessile
glands, the tap root ca. 1 mm. thick. Leaves mostly pseudoverticillate, glabrous
or scantily glandular, 9-25 mm. long, 0.5-2.0 mm. broad, linear, apically obtuse,
attenuate to the sessile base, the stipules lance-deltoid, 1-2 mm. long. Inflores-
cences of terminal cymose racemes, only the first 1 or 2 branches cymose, subse-
quent branching tending to be racemose, the peduncle 8-30 mm. long; bracts
1.0-2.5 mm. long, the pedicels 2-5 mm. long, glabrous to minutely stipitate-
glandular. Sepals 5, 2.5-3.5 mm. long, glabrous or minutely glandular, oblong,
apically obtuse and often cucullate, only the midrib prominent; petals 5, 1.7-2.5
mm. long, 0.4-0.7 mm. broad, apically provided with usually 4 lacinae, the outer
one- third to one-half as long as the trunk, the inner occasionally absent or reduced
to mere dentations, the trunk subdeltoid, laterally denticulate, truncate or tapered
to the minute claw; stamens 5, 1.5-2.0 mm. long, the oblong to suborbicular
anthers 0.4-0.5 mm. long; staminodia absent; ovary at anthesis globose, ca. 2 mm.
long, the style 2-3 -cleft nearly to the base, less than 1 mm. long; stylopodium
absent. Capsule subglobose, 2-3 mm. long, exceeding the sepals, 3-17-seeded, the
seeds 0.9-1.4 mm. broad, dark brown or purplish in age, hippocrepiform, dorsally
flat, scarcely umbonate ventrally, minutely corrugated or tuberculate, the tubercles
broader than long. Fig. 3B (Holotype of D. sperguloides) ; Fig. 3C ("D. molluginea
Fenzl"atLE).

-REVISION OF DRYMARIA

195

Sandy soils, Arizona to western Texas, U. S. A.; south to Puebla, Mexico.

The specimen from Leningrad (Fig. 3C) apparently from the Ledebour her-
barium, and labeled "Drymaria violluginca Fenzl" lacks the inner laciniae of the
petals. I doubt that this feature has any taxonomic significance. A close parallel
is exhibited by D. arenarioides ssp. peninsularis where the form which lacks the
two central laciniae has been described as D. johnstonii. In both D. mollugitica
and D. arenarioides ssp. peninsularis specimens have been examined in which 1 or 2
of many flowers dissected had the inner laciniae reduced to mere dentations or
apparently completely lacking.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

9a. Drymaria arenarioides s

Drymaria arenarioides Humb. & Bonpl.

atP!)
Drymaria frankenioides HBK. Nov. Gen. & Sp. 6:21. pi. 515. 1823.

Prostrate radially spreading herbaceous or lignescent perennials to as much as
20 cm. long, branching mostly at the base, the internodes shorter to longer than the
leaves, stipitate-glandular, the taproot to as much as 5 mm. thick. Leaves opposite
or pseudoverticillate, stipitate-glandular, the blades linear-oblong to narrowly
elliptic, apically obtuse, obscurely 1 -nerved, 5-15 mm. long, 1-3 mm. broad,
basally tapered to the petiole, the petiole ca. 1 mm. long; stipules lance-deltoid,
1.0-2.5 mm. long, entire. Flowers, except the first formed, solitary in the axils of
slightly and gradually reduced foliage leaves, the mature pedicels 3—15 mm. long,
stipitate-glandular, the subtending leaves stipulate. Sepals 5, 4-7 mm. long, 1-3
mm. broad, stipitate-glandular, oblong to broadly elliptic, the venation obscure,
subdendritic, the inner sepals shorter, broader and less glandular; petals 5, 4.0-6.5
mm. long, apically (4-) 6-8-lobate, the outer lobes slightly longer and broader,
about as long as the trunk, the inner lobes linear, equal, the trunk laterally denticu-
late, 0.8-1.5 mm. broad; stamens 5, ultimately 3-5 mm. long, the oblong anthers
0.8-1.2 mm. long; ovary at anthesis ellipsoid, the style bifid or trifid less than half
its length, slightly exceeding the stamens; stylopodium often conspicuous. Capsule
ovoid, 3-5 mm. long, 15-25-seeded, the seeds hippocrepiform, dorsally flattened or
sulcate, exumbonate, minutely tuberculate, 0.8-1.2 mm. broad. Fig. 4A.

Sonora, Chihuahua, Zacatecas, San Luis Potosi, Guanajuato and Hidalgo, Mexico,
usually in sandy soils.

It would appear that Humboldt and Bonpland or Kunth had some reason for
changing the name D. arenarioides to D. frankenioides as Roemer and Schultes cited
the name as follows: "D. arenarioides Humb. et Bonpl. . . . Reliqu. Willd. MS".
I would judge that Willdenow had written the name "D. arenarioides Humb. &
Bonpl." on his herbarium specimen and that Roemer and Schultes were the first to
publish the name, probably furnished earlier by Humboldt and Bonpland. Two
other names in Drymaria seem to have the same history, i.e. D. ovata and D. stel-
larioides. The question then arises whether the name should bear the authority
"Humb. & Bonpl. in Roem. & Schult.", "Humb. & Bonpl.: Willd. ex Roem. &
Schult." or "Willd. ex Roem. & Schult." Most publications appear to make use
of the latter, but I feel it more proper to credit the name to Humboldt and Bon-
pland as did Willdenow in his herbarium. At any rate, D. arenarioides and D.
frankenioides are names based on the same specimen (typonyms).

In its diminished leaves, sepals and petals, Parry # Palmer 49 from San Luis
Potosi, approaches D. arenarioides subsp. peninsularis, but the petals, which seem
to possess more constant characters, are like those of typical arenarioides, i.e. two
larger outer lobes and two basally dichotomous inner lobes, the linear divisions of
the inner lobes being nearly as long as the oblong outer lobes. The very long style
of this specimen further corroborates its determination as D. arenarioides ssp.
arenarioides.

DUKE REVISION OF DRYMARIA

198 ANNALS OF THE MISSOURI BOTANICAL GARDEN

9b. Drymaria arenarioides ssp. peninsularis (S. F. Blake) J. Duke comb. &

F. Blake, in Jour. Wash. Acad. 14:285. 1924. (H<
Purpus 423; Cape Region, Baja California, US!; isotypes at UC, F, GH, MO)

Jobm

Radially spreading or ascending annuals or suffrutescent perennials to as much
as 25 cm. long, branching mostly at the base, the internodes shorter to longer than
the leaves, stipitate-glandular, the taproot to as much as 1 cm. thick. Leaves
opposite or pseudoverticillate, stipitate-glandular, linear, obtuse, obscurely veined,
4-25 mm. long, 0.5-1.0 mm. broad, attenuate to the sessile base, the stipules lance-
deltoid, entire, 0.5-1.5 mm. long. Flowers, except the first formed, solitary in the
axils of slightly and gradually reduced foliage leaves, the mature pedicels 4-12 mm.
long, stipitate-glandular, the subtending leaves stipulate. Sepals 5, 2.8-4.5
mm. long, 1.0-2.2 mm. broad, stipitate-glandular, oblong to broadly elliptic,
obtuse, the venation obscure, subdendritic, the inner sepals shorter, broader and
less glandular; petals 5, 2.8-4.0 mm. long, apically 4(-8)-lobate, the outer lobes
longer, about as long as the trunk, the inner about half as long as the outer, the
trunk laterally denticulate, basally truncate or tapered to the minute claw; stamens
5, ultimately 2-3 mm. long, the oblong anthers 0.5-0.8 mm. long; ovary at
anthesis obovoid, stipitate, the minute style slightly exceeded by the anthers.
Capsule globose, 2.5-4.0 mm. long, the style less than 1 mm. long, trifid about half
its length, the seeds 10-20, hippocrepiform, dorsally flattened or sulcate, scarcely
umbonate ventrally, reticulate or minutely tuberculate, 0.6-1.0 mm. long. Fig
4B (Holotype of D. peninsularis) 5 Fig. 4D (Holotype of D. johnstonii).

Sandy soils; Cape Region of Baja California, Mexico.

Unquestionably this subspecies is intimately related to ssp. arenarioides. The
two subspecies are nicely separated by the Gulf of California in addition to the
diagnostic characters listed in the key. I have seen many specimens of both
subspecies, and intermediates are few indeed. Craig 738 and Hoivell 10576 appear
to represent intermediates with many of the petals similar to those of the typical
subspecies. Other features, coupled with the geography, seem to justify their
relegation to ssp. peninsularis.

10. Drymaria axillaris Brandegee, in Univ. Cal. Publ. Bot. 4:178. 1911.
(holotype: Purpjis 4526; Sierra del Rey, Coahuila, UC!; isotypes at GH, US,
F, MO, etc.)

Upright glandular-pubescent, cyanic glaucous perennials to 10 cm. high with
many suffrutescent branches arising near the base, the taproot to 6 mm. in diameter.
Leaves opposite, cyanic, glaucous, densely glandular-pubescent with capitate hairs
to 0.7 mm. long, broadly ellipsoid, apically obtuse to acutish, basally subcordate or
tapered to the petiole, 3-12 mm. long, 2.5-8.0 mm. broad; veins not apparent;
petioles 0.5-2.0 mm. long, the stipules entire, broadly deltoid, 0.5-1.0 mm. long.
Flowers solitary in the axils of reduced leaves, the glandular-pubescent pedicels

DUKE REVISION OF DRYMARIA 199

ebracteate, 1.5-3.0 mm. long. Sepals 5, subequal, subsucculent, glandular-
pubescent, broadly lanceolate to ovate, slightly concavo-convex, obtuse to acutish,
4-5 mm. long, 1.5-2.0 mm. broad; venation obscure, the margins scariose; petals
5, 3.5-4.0 mm. long, bifid about half their length, the lobes obtuse, ca. 0.6 mm.
broad, exappendiculate, the claw 0.5-1.0 mm. broad; stamens 5, 3.0-3.5 mm.
long, the oblong anthers ca. 0.8 mm. long; ovary at anthesis broadly ellipsoid, the
style bifid about one-third its length, slightly exceeding the anthers. Capsule
bivalved, ca. 2.5 mm. long, many-seeded, the seeds hippocrepiform, merely reticu-
late or puncticulate, ca. 0.7 mm. broad. Fig. 4C (Isotype).
Known only from the type locality in Coahuila, Mexico.

11. Drymaria barkleyi J. Duke and Steyermark sp. nov.

Plantae annuae graciles ramis paucis internodis glaucis quam foliis longioribus
radice non viso. Folia opposita laminis glabris glaucis bullatisque 4-10 mm. longis,
1.0-2.5 mm. latis, ellipticis apice obtusis subsessilibus stipulis minutis deltoideis
caducis apice saepe bifidis. Flores solitarii in axillis foliorum vix reductorum pedi-
cellis glabris 4-8 mm. longis. Sepala (4-) 5, 3.0-4.5 mm. longa ca. 1 mm. lata
glabra oblongo-elliptica obscuriter venosa; petala 5, 2.5-3.5 mm. longa ultra
medium bifida vel rariter trifida lobis oblongis obtusis emarginatisve denticulatis
basi acutis unguibus minutis; stamina (4-) 5, 3 mm. longa antheris oblongis 0.8-
1.0 mm. longis; ovarium oblongum, 1-2 mm. longum saepe 10-ovulatum stylis ca.
1.5 mm. longis (3-) 4-lobis minutis. Capsulae non visae. Fig. 4E (Holotype
[except seed] ) .

Coahuila: low pointed hill, almost destitute of soil, apparently a limestone shale, 25
mi. sw. of Monterrey, Warnock & Barkley 148 26 M. (holotype: F, isotype at US) .

Fortunately another specimen with mature capsules is referable to the new
species. Hernandez, Kowell tf Barkley 16M531, collected on talc-like limy soil and
shaly limestone, (in Nuevo Leon) , 1 1 mi. w. of Santa Catarina differs only in having
a few sessile glands on the herbage. The capsule is yellow, 3-4-toothed, 2.5-3.0
mm. long, 4—8 seeded. The seeds are lacrimiform, ca. 0.7 mm. in diameter,
minutely and remotely tuberculate. Apparently the closest related species is the
preceding with which the new species shares the cyanic tinting, solitary axillary
flowers, and minute deltoid stipules. The differences far outnumber the strictly
superficial similarities however. The sharply pointed long-beaked seed is slightly
reminiscent of D. debilis.

12. Drymaria polycarpoides A. Gray, PI. Fendl. 12. 1849. (holotype:
Gregg s. *.; valley of Bolson de Mapimi, Mexico; GH!)

Prostrate or ascending profusely branching perennials to as much as 10 cm.
long, branching mostly from the base, the nodes geniculate, fragile, the internodes
mostly shorter than the leaves, glabrous or quite glaucous, the taproot to as much
as 1 cm. thick. Leaves opposite, carnose, glaucous, often cyanic, elliptic, apically
acute to obtuse, basally tapered, 5-12 mm. long, 2-5 mm. broad, weakly if at all
nerved; petioles 0.5-1.5 mm. long, exceeding the minute, entire, lance-deltoid

200 ANNALS OF THE MISSOURI BOTANICAL GARDEN

stipules. Flowers in few-flowered strongly dichotomous cymes, the pedicels 1-4
mm. long, the bracts deltoid, obtuse, not scariose, 0.8-1.5 mm. long. Sepals 5,
3.5-5.0 mm. long, 1.5-2.0 mm. broad, glaucous, oblong to ovate, apically obtuse
and often cucullate, the scariose margin as broad as or broader than the central
3 -nerved portion; petals 5, 3.5-4.5 mm. long, bifid little more than half their
length, the lobes broadly oblong, apically obtuse, 1-3 -nerved, merging imper-
ceptibly with the trunk, the trunk 1.0-1.5 mm. long, nearly as broad, exunguicu-
late or subexunguiculate; stamens 5, the anthers 0.8-1.0 mm. long, the filaments
1.5-2.5 mm. long, devoid of staminodia; ovary at anthesis ovoid, equaled or
exceeded by the style, the style trifid for less than one-fourth its length. Capsule
ellipsoid to oblongoid, 2.0-2.5 mm. long, few-seeded, the seeds lacrimiform, not
circinate, umbonulate ventrally, smooth, yellowish- brown. Fig. 4F (Holotype
[except seed]).

Known only from the Coahuilan Desert of Mexico.

E. Series viscosae
13. Drymaria viscosa S. Watson ex Orcutt, in West. Am. Sci. 2:57. 1886.

nomen nudum. S. Watson in Proc. Am. Acad. 22:469. 1887. (holotype:

Orcutt 1330, Socorro ("Socono"), n. Baja California, GH!; isotypes at UC, D,

F, US, MO, NY, etc.)

Prostrate radially spreading or ascending diffusely branched viscid annuals, the
branching below the inflorescence mostly dichotomous, the taproot to as much as
5 mm. thick. Leaves pseudoverticillate or opposite, glandular-puberulent, linear,
apically obtuse, basally tapering to the clasping petiole, 3-15 mm. long, 0.5-1.5
mm. broad, the petiole 1-3 mm. long; stipules acicular, caducous, 1.0-1.5 mm.
long. Flowers campanula te in 1-5 -flowered cymes, the peduncles mostly glabrous;
bracts deltoid, marginally ciliolate, 0.5-1.0 mm. long, the pedicels 0-3 mm. long,
all but the central equaled or exceeded by the bracts. Sepals 5, 2.3-3.0 mm. long,
1-1.5 mm. broad, glandular-puberulent, oblong, obtuse, obscurely 3 -nerved, the
midrib subapically excurrent; petals 5, 1.8-2.2 mm. long, bifid about two-thirds
their length, the lobes oblong, obtuse, tapered to the unguiculate base, exappendicu-
late; stamens 5, ca. 2 mm. long, the broadly oblong anthers ca. 0.4 mm. long;
ovary at anthesis ellipsoid, about equaled by the style, the style trifid about half
its length, barely exceeding the anthers. Capsule 2.0-2.5 mm. long, equaling or
exceeded by the sepals, 5-1 5 -seeded, the seeds cochleate, nearly smooth, 0.5-0.6 mm.
broad, dorsally brown, facially transparent and colorless, the yellowish embryo
visible. Fig. 5 A (Isotype).

Sandy soils of Baja California, Mexico.

In habit, this plant is not unlike D. molluginea as pictured in Fig. 3D.

14. Drymaria ortegioides Griseb. Cat. PL Cub. 21. 1866. (holotype:

2019; Cuba occ, GOET!; isotypes at F, GH, MO, US, etc.)
Pinosia ortegioides (Griseb.) Urban, in Ark. Bot. 23A5:71. 1930.

DUKE REVISION OF DRYMARIA 201

Upright sparsely dichotomizing glabrous or glandular perennials to as much as
20 cm. high, the internodes much longer than the leaves, the tap root to 5 mm.
thick. Leaves opposite, glabrous, linear to elliptic, apically acute to obtuse, attenu-
ate to the subsessile base, 3-12 mm. long, 1.0-3.5 mm. broad, the stipules lance-
deltoid, 0.3-0.8 mm. long, entire. Inflorescences of terminal condensed 5-21-
flowered dichotomous cymes; peduncles 1.5-3.5 cm. long, the bracts 1-2 mm. long,
imbricate or subimbricate, the glabrous pedicels 1.5-4.0 mm. long. Sepals 5, 2.5-
3.5 mm. long, glabrous, ovate to obovate, acute to obtuse, the venation obscure,
dichotomous and reticulate; petals 3, 2.3-3.2 mm. long, twice dichotomous, the
lobes oblong, obtuse, nearly as long as the trunk, exappendiculate, the trunk
merging gradually with the claw; stamens 3, the oblong anthers ca. 0.5 mm. long;
ovary subglobose, the styles 0.5-1 mm. long. Capsule globose, 1.5-2.0 mm. long,
shorter than the sepals, 1-4-seeded, the seeds hippocrepiform, scarcely if at all
umbonulate, closely tuberculate, 1.0-1.5 mm. long, dark brown. Fig. 5B, 5C.

Sandy savannas of Cuba.

Leaves deltoid-ovate to reniform, opposite, stipulate. Flowers long-pedicel'
in lax cymes. Sepals acute or obtuse, trinerved. Petals bifid, the lobes thcmsel
dichotomizing, ciliate-auriculate at their bases. Seeds cochleate, tuberculate,
tubercles domical or substellate. 3 species in Mexico. Fig 5 D-F.

15. Drymaria hypericifolia Briq. in Ann. Cons. & Jard. Geneva 14:369. 1911.

(holotype: Jurgenson 38, de Lecambre de Yolotepeque a Juquilla, au sud-

ouest d'Oaxaca; Mexicum G!; isotype at K)

Erect or ascending perennials (?) to as much as 25 cm. high, the internodes
about as long as the leaves, stipitate-glandular. Leaves opposite, 8-22 mm. long,
5-14 mm. broad, narrowly deltoid-ovate, glabrous or minutely glandular-puberulent
below, apically acute, marginally somewhat undulate, basally obtuse to acute,
plinerved; petioles 1-2 mm. long, the stipules entire or bifid, the divisions lance-
deltoid, 0.5-1.2 mm. long. Inflorescences of terminal 5-1 1-flowered cymes, the
peduncles 2-3 cm. long, the bracts lanceolate, 1.5-2.5 mm. long, the pedicels
stipitate-glandular, 4-6 mm. long. Sepals 5, 4.5-5.5 mm. long, lanceolate to
narrowly ovate, apically acute, the outer sepals scantily stipitate-glandular, tri-
nerved; petals 5, ca. 6 mm. long, four times dichotomous, the ultimate segments
emarginate, the trunk laterally provided with deciduous cilia te auricles; stamens 5,
ca. 4 mm. long, the anthers oblong, ca. 0.7 mm. long; ovary at anthesis ellipsoid,
about equaled by the partially trifid style, the style slightly exceeded by the anthers.

202 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Capsule ellipsoid, ca. 5 mm. long, 3-valved, apically constricted, many-seeded, the
seeds cochleate, 0.7-1.0 mm. long, irregularly corrugated or tuberculate; seeds
immature. Fig. 5D (Holotype).

Known only from two collections near Juquila in Oaxaca, Mexico.

North American authors have confused Drymaria ladcwii with this species,
probaWy because of certain vegetative resemblances. The flowers of the two
species however contrast vividly.

16. Drymaria excisa Standi, in Field Mus. Bot. 8:11. 1930. (holotype:
Mexia 174.8', Real Alto, trail to El Tajo de Santiago, 2500 m., Jalisco, F!; iso-
typesatUC, GH, etc.)
Drymaria grandis Bullock, in Kew Bull. 1936:389. 1936. (holotype: Hinton 5427, Los
Hornos, Temascaltepec, 2500 m., Mexico, K!; isotypes at GH, MO, etc.)
Robust spreading perennials, the procumbent or erect branches to 3 m. long
(fide Bullock), the internodes mostly longer than the leaves, glandular-puberulent,
glabrescent. Leaves opposite, 5-25 mm. long, 5-30 mm. broad, glabrous or
minutely glandular, the blades orbicular to obcordate, apically emarginate or
apiculate, marginally entire, basally obtuse to truncate, 3-pli-nerved; petioles 5-25
mm. long, the stipules entire or apically lacerate, lanceolate, 2—4 mm. long.
Inflorescences of terminal lax 3 -many- flowered cymes, the peduncles 3-6 (-15)
cm. long; bracts lanceolate to ovate, 3.0-4.5 mm. long, the pedicels 3-15 mm.
long, stipitate-glandular, glabrescent. Sepals 5, 5.0-10.0 mm. long, 2.0-2.8
mm. broad, ovate, apically acute, weakly 3-ribbed; petals 5, 6-12 mm. long,
2— 3 -times dichotomous, the ultimate segments entire or emarginate, the trunk
laterally provided with several deciduous ciliate auricles; stamens 5, 5.0-6.5 mm.
long, the oblong anthers ca. 1 mm. long, the staminodia minute, semicircular;
ovary at anthesis ovoid, about equaling the style, the style 1.5-3.0 mm. long, trifid
about one-third its length. Capsule ellipsoid, apically constricted, 3.0-5.5 mm.
long, many-seeded, the seeds tightly cochleate, 0.9-1.2 mm. broad, evenly tuber-
culate, the dorsal tubercles somewhat longer than broad, the facial tubercles
polygonal (rarely substellate) in outline. Fig 5E (Holotype).

Jalisco, Michoacan, Mexico, and Morelos, Mexico, usually above 2000 m.
It is interesting to note that in none of the original descriptions of taxa refer-
able to this species (or series) was any mention made of the poly dichotomous
nature of the petals. This may be due to the fact that the foliage is so outstanding
that the species can usually be determined without resorting to the tedium of
dissection.

17. Drymaria

1890. (b

jara, GH!; isotypes at UC, US, etc.)

Laxly spreading or rarely erect annuals, the internodes mostly longer than the
leaves, villose with spreading jointed hairs or stipitate-glandular. Leaves opposite,
5-14 mm. long, 5-20 mm. broad, villose to glabrous, reniform to deltoid-ovate,
apically rounded to acute and apiculate, marginally entire, basally truncate or

204 ANNALS OF THE MISSOURI BOTANICAL GARDEN

cordate, 3-7-pli-nerved; petioles 1-6 mm. long, stipules entire or bifid, the divisions
filiform to narrowly lanceolate, 1.0-1.5 mm. long. Inflorescences of lax terminal
3-many-flowered cymes, the peduncles 1-10 cm. long; bracts lanceolate, 1.5-2.5
mm. long, the pedicels 5-25 mm. long, sparingly villose with glandular hairs
occasionally admixed. Sepals 5, 4.2-5.5 mm. long, 1.7-2.7 mm. wide, glabrous
to villose, narrowly to broadly ovate, the outer acute, villose, and weakly 3 -nerved,
the inner rounded and glabrous; petals 5, 5.6-7.0 mm. long, once or twice
dichotomous, the ultimate segments emarginate, the trunk laterally provided with
several ciliate auricles; stamens 5, 4-5 mm. long, the oblong anthers 0.8-1.0 mm.
long; ovary at anthesis subglobose, 1-2 mm. long, the style 2.0-2.5 mm. long,
trifid about one-third its length, exceeded by the anthers. Capsule 3.5-5.0 mm.
long, many-seeded, the seeds 0.7-0.9 mm. long, cochleate, evenly tuberculate, the
dorsal tubercles conical, the facial tubercles elongate-stellate. Fig. 5F (Holotype).

Jalisco, Michoacan, and Mexico, Mexico.

Here the tendency toward a second dichotomy in the petals is often suppressed
to the point that the petal is merely bilobed, but the lobes are emarginate, and the
twice-dichotomous nature is still reflected in the venation. This species affords a
good transition to the series villosae. McVaugh & Koeh 897 from Jalisco makes
a marked approach toward D. multiflora, and may represent a hybrid. Indicative
of hybrid tendencies are the broadly reniform leaves, the absence of villosity, and
the weakly emarginate petals. The floral parts in general are a bit small for
D. longepeduncnlata and come closer to the range of dimension exhibited by the
flowers of D. multiflora.

H. Seri

Leaves linear, oblong or narrowly spatulate, subsessile, stipulate, the stipules
usually entire. Flowers in bracteate cymes, the pedicels longer or shorter than the
bracts. Sepals obtuse to acuminate, trinerved. Petals bilobed, unguiculate,
exauriculate, the lobes 1 (-3) -nerved. Seeds cochleate, tuberculate, the tubercles
not ornate. Two polymorphic intergrading delicate annual (or rarely perennial)
species; western U. S. A. to Guatemala. Fig. 6.

inflorescence; bracts mostly exceeding the subtended pedic<
1. Some or all of the sepals acute to acuminate; seeds tightly co

DUKE REVISION OF DRYMARIA

Peduncles and often the pedicels and sepals stipitate-glandular, the pedicels mostly
>y or exerted from the calyx:

18a. Drymaria effusa var. effusa

HI)

Erect, delicate, sparsely branched annuals to as much as 25 cm. high, the inter-
nodes mostly longer than the leaves, minutely glandular-puberulent or glabrate,
the roots to ca. 1 mm. thick. Basal leaves pseudoverticillate, spatulate to orbicular;
cauline leaves opposite, rarely pseudoverticillate, 5-25 mm. long, 0.5-1.2 mm.
broad, minutely puberulent or glabrous, linear, apically obtuse or apiculate, sessile,
the stipules entire, acicular, caducous, 0.5-1.0 mm. long. Inflorescences of terminal
multifloriferous cymes, the ultimate branches showing tendencies to become
monopodial; peduncles 2-5 cm. long, stipitate-glandular; bracts ovate, 0.5-1.8 mm.
long, the mature pedicels 1-5 mm. long, stipitate-glandular, usually exceeding the
bracts. Sepals 5, 1.4—2.2 mm. long, the outer narrower, shorter, more obtuse and
often cucullate, glandular, the inner more ovate, often acutish, all trinerved, the
veins subapically confluent; petals 5, 2-4 mm. long, bifid half their length or more,
the lobes linear to oblong, 1 -nerved, obtuse, tapering gradually to the claw, exap-
pendiculate; stamens 5, 1.2-1.6 mm. long, the oblong anthers ca. 0.3 mm. long;
ovary at anthesis ellipsoid, the style rather elongate, trifid about half its length,
about attaining the anthers. Capsule ellipsoid, nearly equaling the sepals, the
stigmata exerted; seeds 2-9, cochleate to lacrimiform, coarsely tuberculate, 0.5-
0.8 mm. broad. Fig. 6B (Holotype).

Arizona, U.S.A.; Sonora and Sinaloa, Mexico, the Sinaloa specimens approach-
ing D. leptophylla.

18b. Drymaria effusa var. confusa (J. N. Rose) J. Duke comb. & stat. nov.
tr. U. S. Nat. Herb. 5:133. 1897. (holotype:
, US!, isotypes at GH, MO, etc.)
Delicate sparsely branching glabrous or subglabrous annuals to as much as 7.5
cm. high, scarcely branching at the base. Basal leaves glabrous, orbicular to
broadly elliptic, 2.5-3.5 mm. long, 2.5-3.0 mm. broad, the petioles 1.5-2.5 mm.
long; cauline leaves linear to oblong, obtuse, slightly narrowed to the clasping base,
10-20 mm. long, ca. 1 mm. broad, the stipules entire, acicular, 0.5-1.0 mm. long.
Inflorescences of terminal diffuse several-flowered cymes, the glabrous peduncles
10-25 mm. long; bracts ovate to narrowly lance-deltoid, 0.5-0.8 mm. long,
scariose and transparent save for the darkened midrib; pedicels glabrous, 2-8 mm.
long, at maturity usually longer than the flowers. Sepals 5, 1.5-2.2 mm. long, 0.5-
0.7 mm. broad, minutely glandular, broadly oblong, apically obtuse, transparent
save for the 3 ribs; petals 5, 2.0-2.5 mm. long, bifid about half their length, the
lobes spatulate to narrowly oblong, trinerved, exappendiculate, merging imper-

206 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ceptibly with the narrower claw; stamens (3?-) 5, ca. 2 mm. long, the oblong
anthers ca. 0.4 mm. long; ovary at anthesis ellipsoid, slightly exceeded by the
anthers, the slightly trind style exceeding the anthers. Capsule ca. 1.8 mm. long,
few seeded, the seeds cochleate, minutely tuberculate, ca. 0.6 mm. long. Fig. 6C
(Holotype).

Chihuahua and Durango, Mexico.

Nelson 4813 is determined as this variety only with some misgivings. All the
sepals are obtuse and cucullat^ and many of the pedicels are exceeded by the sub-
tending bracts, in these respects approaching var. depressa. Although suggestive
of var. depressa, this plant's upright habit and obovate 3 -nerved petal-lobes deny
this possibility.

18c. Drymaria effusa var. depressa (Greene) J. Duke comb. & stat. nov.

Drymaria depressa Greene, Leafl. Bot. Obs. 1:153. 1905. (holotype: Metcalf 143O;

open glades of the Black Range, alt. 9,500 ft., N. Mex., ND; isotypes at F, GH, MO,

UC,US,etc!)
Drymaria minuscula Standi. & Steyerm. in Field Mus. Bot. 23:52. 1944. (holotype:

Steyermark 50243; on rocky limestone outcrops under Juniperus standleyi, alt. 3700

Guatemala, F!)

Minute, subcespitose, glabrous to minutely puberulent annuals rarely to 5 cm.
high, the branching largely confined to the base and the inflorescence. Rosette
leaves orbicular to spatulate, 1.5-10.0 mm. long, 1-3 mm. broad; cauline leaves
when present opposite, the blades oblong, 3-10 mm. long, 1-4 mm. broad, apically
obtuse, basally tapered, subsessile, the stipules entire, 0.5-1.2 mm. long, acicular,
caducous. Inflorescences of terminal lax (rarely condensed) several-flowered cymes,
the bracts 1.5-2.0 mm. long, mostly longer than the pedicels, the pedicels and
peduncles glabrous or glandular. Sepals 5, 1.5-2.5 mm. long, ca. 1 mm. broad,
glabrous or glandular, oblong to ovate, apically obtuse, cucullate, 3 -ribbed, the
ribs subapically confluent; petals 5, 1.5-2.8 mm. long, bifid about half their length,
the lobes linear to narrowly oblong, apically obtuse, usually 1 -nerved, gradually
tapering to the claw; stamens 5, 0.7-1.7 mm. long, the oblong anthers 0.2-0.3 mm.
long; ovary at anthesis globose to obturbinate, the styles 2-3 -fid about half their
length, about equaling the anthers. Capsule ellipsoid to subglobose, about equaling
the sepals, 6-1 2 -seeded, the seed cochleate to lacrimiform, 0.5-0.7 mm. long,
dorsally sulcate, minutely tuberculate. Fig 6A. (Isotype).

High altitudes, New Mexico, Colorado and Arizona, U. S. A., south to Guate-

As a rule this is quite a distinctive variety but it is probable that it hybridizes
with other members of the series. Metcalfe 1428 strongly suggests that hybridiza-
tion has occurred with D. leptophylla or one of its varieties. Except rarely, D.
leptophylla and D. effusa var. depressa are quite distinct, the former being char-
acterized by an elongate subvirgate habit, the capsules shorter than the inner acute
sepals, and the deeply cleft style, the latter being differentiated by the dwarf
subcespitose habit, the capsules about equaling the obtuse, cucullate inner sepals,
and the shallowly cleft style.

DUKE REVISION OF DRYMARIA 207

208 ANNALS OF THE MISSOURI BOTANICAL GARDEN

In correspondence with Dr. Steyermark, I have suggested to him that D. minus-
cala probably represents an alpine reduction of what I here reduce to the var.
depressa, and he tends to agree with this disposition of his diminutive species. The
type from Guatemala has somewhat reduced nodes, leaves, sepals, etc., but such
reductions also occur at the other end of the range of the variety, e. g. Colorado,
and it seems doubtful that such diminutions are worthy of formal status.

19a. Drymaria leptophylla var. leptophylla

Drymaria leptophylla (Cham. & Schlecht.) Fenzl ex Rohrb. in Linnaea 37:195. 1871.

(holotype: Schiede 6 Deppe 511; ad radices montis Orizabae, B; n.v.)
Arenaria Uptopby in Linnaea 5:233. 1830.

Drymaria tenella A. Gray, Pi. Fendl. 12. 1849. (holotype: Fendler 56; eight miles west

of las Vegas, New Mexico; GH!)
Drymaria nodosa var. ? gracillima Hemsl. Diag. PI. Nov. 2:22. 1879. (holotype: Parry

& Palmer 6o\ in regione San Luis Potosi, alt. 6000-8000 ped., Mexico, K n. v.; isotypes

at MO, GH, etc.!)
Drymaria gracillima (Hemsl.) J. N. Rose, in Contr. U. S. Nat. Herb. 5:132. 1897.

Erect delicate, usually sparsely branched annuals to as much as 20 cm. high,
the internodes much longer than the leaves, glabrous to minutely glandular, the
taproot ca. 1 mm. thick. Leaves opposite, rarely pseudoverticillate, glabrous,
linear to narrowly oblong, 5-25 mm. long, 0.5-1.0 mm. broad, often involute and
circinate, apically obtuse or acute, tapered to the subsessile base; stipules entire,
0.2-0.8 mm. long, acicular, caducous. Inflorescences of terminal many-flowered
dichasial cymes, the peduncles 1-5 cm. long, glabrous to minutely glandular;
bracts ovate, 0.5-1.0 (-1.5) mm. long, equaling or exceeding the pedicels. Sepals
5, ovate, 1.5-3.5 mm. long, narrowly ovate, apically acute and often somewhat
reflexed (the outer sepals obtuse and cucullate in some northern material) , 3 -ribbed,
glabrous or with a few sessile glands; petals 5, 1.2-2.4 mm. long, bifid about half
their length, the lobes linear, obtuse or acute; stamens 5, 1.0-1.5 mm. long, the
suborbicular anthers ca. 0.2 mm. long, the filaments basally and briefly connate;
ovary at anthesis subglobose, the style about attaining the anthers, divided nearly
to the base. Capsule 1.5-2.0 mm. long, 5-20-seeded, the seeds cochleate, 0.5-
0.7 mm. long, evenly tuberculate with domical tubercles. Fig 6G (Isotype of
D. gracillima) .

Arizona to Colorado, U.S.A., south to southern Mexico, and in Baja California.

Toward the northern end of its range, the specimens belonging to the typical
variety show frequent, but inconsistent, tendencies toward more obtuse, often
cucullate, outer sepals and more pronouncedly coiled and contorted leaves. In the
northern material the outer sepals are more frequently shorter than the inner, while
the converse condition is more frequent toward the south. Although there is not
enough material to justify any rigid hypothesis, I would guess that clinal variation
is at work in this complex variety.

19b. Drymaria leptophylla var. cognata (S. F. Blake) J. Duke comb. & stat.

DUKE REVISION OF DRYMARIA 209

Upright, diffusely branching, locally glandular-puberulent annual to as much
as 15 cm. high. Leaves opposite, the blades 6-20 mm. long, 0.5-1.5 mm. broad,
glabrous, lance-linear, apically acute, gradually tapering to the subsessile somewhat
clasping base, veinless or obscurely 3 -nerved; stipules mostly entire, acicular, 0.5-
0.8 mm. long. Inflorescences of several-flowered, diffuse, trichotomous cymes, the
axes locally glandular-puberulent; bracts lance-ovate to ovate, aristate, scariose save
for the green excurrent midrib, 1-2 mm. long, usually shorter than the subtended
pedicels, the pedicels 0.5-4.0 mm. long. Sepals 5, 3.3-4.5 mm. long, 0.8-1.3 mm.
broad, glabrous or very sparsely glandular, lance-ovate, apically attenuate, 3 -ribbed,
transparent save for the green ribbed area; petals 5, 3.0-3.5 mm. long, bifid
slightly more than half their length, the sinus broad, the lobes narrowly oblong,
obtuse, the trunk flaring before tapering rather abruptly to the claw; stamens 5,
2.0-2.5 mm. long, the oblong anthers ca. 0.5 mm. long; ovary at anthesis sub-
globose, the style slightly shorter, trifid about half its length, slightly exceeding
the anthers. Capsule ellipsoid, 2-3 mm. long, ca. 5-seeded, the seeds cochleate, ca.
0.8 mm. broad, minutely and regularly tuberculate. Fig. 6D (Holotype).

Durango and Chihuahua, Mexico.

19c. Drymaria leptophylla var. nodosa (Engelm.) J. Duke comb. nov.

Drymaria nodosa Engelm. in A. Gray, PI. Fer

from seed collected by Wislizenius in the m
Drymaria tcuclla var. nodosa (Engelm.) "Wiggir
Drymaria gentryi Fosberg, in Proc. Biol. Soc. Wash. 62:147. 1949. (HOLOTYPE: Gentry

266g; transition pine slope, los Cascarones, Rio Mayo; Chihuahua; US!; isotypes at MO,

UC, etc.)

Erect delicate, sparsely branched annuals (or perennials?) to as much as 25 cm.
high, the internodes mostly longer than the leaves, usually glandular-puberulent,
the taproot to 2 mm. in diameter. Basal leaves spatulate to orbicular, opposite or
pseudoverticillate, caducous; cauline leaves opposite, glabrous or minutely glandu-
lar, 5-30 mm. long, 0.5-2.5 mm. broad, 1 -nerved, linear to narrowly elliptic,
apically obtuse to apiculate, sessile; stipules entire, acicular, caducous, 0.5-1.5 mm.
long. Inflorescences of terminal many-flowered dichasial cymes, the peduncles
2-5 cm. long, stipitate-glandular; bracts lance-deltoid to ovate, acuminate, 0.7-2.5
mm. long, mostly shorter than the pedicels, the pedicels stipitate-glandular, 0.5-
5.0 mm. long. Sepals 5, subequal, 2.3-3.5 mm. long, 0.5-1.5 mm. broad, lanceolate
to ovate, acuminate, basally or entirely stipitate-glandular, 3 -ribbed, the ribs
distally confluent; petals 5, 2.0-3.5 mm. long, bifid slightly more than half their
length, the lobes linear to oblong or narrowly spatulate, obtuse, mostly 1 -nerved,
exappendiculate; stamens 5, 1.5-2.5 mm. long, the oblong anthers ca. 0.4 mm.
long, the filaments basally connate, often with inconspicuous staminodial flaps;
ovary at anthesis ellipsoid to subglobose, the styles elongate, nearly as long as the
ovary, bifid or trifid less than half their length, equaling or exceeding the anthers.
Capsule ellipsoid, 1.5-2.0 mm. long, shorter than the sepals, 1-9-seeded, the seeds
cochleate, the anterior end strongly recurved, minutely and regularly reticulate,
0.6-1.0 mm. broad. Fig. 6E (Holotype of D. nodosa), Fig. 6F (Holotype of
D. gentryi) .

210 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Chihuahua, Sonora, Durango, San Luis Potosi and Hidalgo, Mexico.

Gentry 2669, type of D. gentryi, stands out with its rather broad leaves, but
examination of other specimens e.g. LeSueur 623, also from Chihuahua, shows that
Fosberg's species cannot stand as distinct from D. leptophylla var. nodosa.

Equally aberrant are S. S. White 418$ and 4776 from Sonora. In these speci-
mens, the inflorescence is strikingly congested and the petal-lobes are broader than
those of most specimens of the variety nodosa. In spite of the bizarre inflorescence,
it seems best to retain them within the variety, with which they coincide in other
fundamental characteristics.

Leaves opposite, lanceolate, petiolate, stipulate, the stipules mostly lacerate.
Flowers in monopodial or dichasial cymes, the pedicels longer or shorter than the
bracts. Sepals acute, 1-3-nerved. Petals bifid, the lobes 1-5-nerved, unguiculate,
exauriculate. Stamens 2-5. Seeds cochlea te, tuberculate, the tubercles not ornate.
Two diffusely spreading annual or perennial Mexican species. Fig. 7.

20. Drymaria anomala S. Watson, in Proc. Am. Acad. 25:143. 1889. (holo-
type: Pringle 2847, Carneros Pass, Coahuila, US!; isotype at GH)

Upright or spreading annuals or perennials, much branched from below, the
internodes mostly longer than the leaves, stipitate-glandular, the taproot to 5 mm.
thick. Cauline leaves opposite, 3-8 mm. long, 2-4 mm broad, glabrous to glandu-
lar or puberulent, the blades lance elliptic, obscurely veined, apically acute, basally
cuneate to acute, the petioles mostly less than 1 mm. long; stipules entire or bifid,
lance-deltoid to acicular, persistent, 0.5-1.5 mm. long. Inflorescences of terminal
many-flowered cymes, the first branches dichasial, subsequent branches becoming
monopodial; peduncles 2-5 cm. long, glabrous to stipitate-glandular; bracts 0.5-
2.0 mm. long, exceeding the pedicels, the pedicels stipitate-glandular or glabrescent.
Sepals 5, 2-3 mm. long, 1.0-1.5 mm. broad, stipitate-glandular, broadly elliptic,
apically acute, with 3 prominent veins, the outer submarginal, there being little or
no scariose margin; petals 5, 1.5-2.0 mm. long, bifid a little over half their length,
the lobes oblong, obtuse; stamens (2-) 3-5, 1.0-1.4 mm. long, the anthers oblong,
ca. 0.3 mm. long; ovary at anthesis obturbinate, the style trifid about half its
length, much shorter than the ovary, about attaining the anthers. Capsule about
equaling the sepals, 5-9-seeded, the seeds cochleate, closely tuberculate, 0.6-0.9
mm. broad. Fig. 7 A (details from Holotype).

Coahuila, Zacatecas, San Luis Potosi and Hidalgo, Mexico.

21. Drymaria tenuis S. Watson, in Proc. Am. Acad. 25:142. 1889. (holo-
type: Pringle 2120; under ledges of the barranca near Guadalajara, Jalisco,
US!; isotype at GH)

DUKE REVISION OF DRYMARIA

in Ann. Cons. & Jard. Bot. Gen. 13:374. 1911.
Drymaria tenuis P var. jaliscana Briq. loc. cit. 374. 1911. (holotype: Pringle 4536;
Civitas Jalisco, in declivibus altis prope Guadalajara, G; n. v.; isotypes at MO, US,

Erect or clambering delicate annuals or perennials to as much as 50 cm. long,
the internodes much longer than the leaves, glabrous or pilosulous, the root some-
times lignescent, to as much as 6 mm. thick. Leaves opposite, glabrous, glandular
or pilosulous, the blades lance-elliptic to ovate, apically and basally acute, 2.5-10.0
mm. long, 1-7 mm. broad, the petioles 0.5-5.0 mm. long; stipules lacerate, the

the peduncles very elongate, glal

212 ANNALS OF THE MISSOURI BOTANICAL GARDEN

pedicels 2-7 mm. long, glabrous, usually much longer than the subtending bracts.
Sepals 5, the outer ovate, apically obtuse, 2.0-2.5 mm. long, ca. 1 mm. broad, the

mm. long; petals 5, 1-2 mm. long, bifid about two-thirds their length, the lobes
oblong, apically obtuse or acutish, exappendiculate, gradually tapering to the linear
claw; stamens 5, 1.0-1.3 mm. long, the oblong anthers ca. 0.2 mm. long, the
filaments basally connate; ovary at anthesis subglobose, ca. 0.8 mm. long; style
bifid or trifid about half its length, about attaining the anthers. Capsule ovoid,
about equaling the sepals, 5-15-seeded, the seeds cochleate, 0.5-0.7 mm. broad,
evenly tuberculate, the dorsal tubercles mostly longer than broad. Fig. 7B
(Holotype of D. filiformis Robinson & Seaton) ; Fig. 7C (Holotype of D. tenuis).

Michoacan, Jalisco, Guanajuato, Zacatecas and Veracruz, Mexico.

The type of D. filiformis from Mt. Orizaba differs only in having shorter
petioles and smaller, more obtuse petal lobes, differences possibly resulting from the
alpine environment and probably of no taxonomic consequence. Unfortunately
Briquet relied too heavily on Watson's description of D. tenuis as glabrous, and
published the variety jaliscana, believing it to differ in being pubescent. Examina-
tion of the holotype of D. tenuis reveals that Watson's description was erroneous;
the type is locally villosulous with hairs up to 0.5 mm. long. Briquet's pubescent
variety, collected very near the type locality of the typical variety, is probably no
more pubescent than the typical variety.

Leaves linear to lanceolate, opposite, occasionally pergameneous, sessile or briefly
petiolate, stipulate or estipulate. Flowers, showing tendencies toward monoecism,
in dichasial bracteate cymes. Sepals acute, 3-5-ribbed. Petals bifid, the lobes 3-
9-nerved, exauriculate, unguiculate or exunguiculate, the trunk entire or denticu-
late. Sepals cochleate, tuberculate, the tubercles acute or obtuse. Four more or
less erect suffrutescent perennial species of Peru and Ecuador. Fig. 8.

'' d>" V lob t 'ir'S' ; ci;» 'horur' Sl»

the

! trunk; seedi

rv-u

fafE

"aftuberde!!

- ^:r^:t'\z?<L

l dJ

- ] :r:i

!:::<,

■BSruasjss

Pedicels and sepals glabrous, the mi

,i the 5 ep,ls

excu

with domical

:tals tapered to the claw. Peru

nd sep.h ,hb

reus,

>stly 5-nerved;

22. Drymaria stellarioides Humb. & Bonpl. ex Roem. & Schult. Syst. Veg. 5:
406. 1819. (holotype: Humb. # Bonpl. s.n.; crescit prope Hambato, alt.
1380 hex. [Regno Quitensi]; probably destroyed; fragment and photograph at
F! illustr. HBK. Nov. Gen. & Sp. 6; pi. 516. 1823.)

214

Upright or ascending dichotomous lignescent perennials from a stout tap-
root to 15 mm. thick, the internodes mostly longer than the leaves except on
condensed lateral branches, glabrous to glandular-pubescent. Leaves opposite,
glabrous or rarely glandular-puberulent, lanceolate to oblanceolate, apically acute,
marginally entire, 3 -ribbed, 5-20 mm. long, 2.5-7.5 mm. broad, tapered to the
sessile, clasping base; stipules entire, apparently fused or occasionally absent, 0.0-
2.5 mm. long. Inflorescences of lax to crowded 5-many-flowered cymes, the
peduncles 1-8 cm. long; bracts lanceolate, 3-5 mm. long, intergrading with the
foliage leaves; pedicels glabrous to densely glandular-puberulent, all but the central
pedicel usually exceeded by the bracts. Sepals 5, narrowly to broadly lanceolate,
glabrous to densely glandular-puberulent, strongly 3 -ribbed, apically acute, (4-)
5-9 mm. long, 2-3 mm. broad; petals 5, 5-10 mm. long, bifid about half their
length, the lobes oblong, obtuse, 4-8 -nerved, exauriculate, constricted slightly
toward the trunk, the trunk rhombic to rectangular, truncate to the short claw;
stamens 5, 3-6 mm. long, those of fertile flowers (destined to mature seeds)
usually much shorter than those of infertile flowers; anthers 0.7-1.1 mm. long;
filaments basally connate into a cup ca. 1 mm. deep; ovary (occasionally absent in
flowers with long stamens) ellipsoid, multiovulate, those of fertile flowers about
equaling the anthers; styles 1.5-3.0 mm. long, trifid for about one-fourth their
length. Capsule ellipsoid, substylopodiate, 3-5 mm. long, 2-many-seeded, the seeds
laxly circinnate, 0.8-1.2 mm. broad, tuberculate, the dorsal tubercles subspinulose,
the facial tubercles tending to be polygonal. Fig. 8A.

Pachano Q5 (Fig. 8A) and the isotype seem to differ from other representatives
of this species in that the sepals and petals are unusually long and the petal lobes
have more numerous veins. Seeds of this species, as illustrated in HBK. Nov. Gen.
& Sp. 6: pi. 516. 1823, appear to be constantly subspinulose-tuberculate. There
is no indication in the aforementioned illustration however of the polygamous
tendency which I believe characterizes this species. Although the material available
for dissection has not been too copious, I feel that certain correlates exist in differ-
ent types of flowers. Many of the ovaries bear aberrant ovules and in some
flowers the ovary is completely lacking. In flowers which mature seed, the anthers
rarely surpass the body of the capsule, while in those flowers with aberrant ovules,
the filaments seem to have elongated so that the anthers equal or surpass the style
branches. In some unnumbered specimens collected by Jameson near Cuenca and
Loxa, there are flowers with long glandular sepals and others with shorter glabrous
sepals; those with the shorter glabrous sepals proved to be fertile while some of
those with longer glandular sepals were completely devoid of ovaries.

23. Drymaria stereophylla Mattf. in Notizbl. Bot. Gart. Berl. 13:436. 1936.
(holotype: Kaimondi 2$02\ Peru. Dep. Junin: Prov. Jauja; Abhang am Fusse
des Chacapalpa zwischen Huari und Jauja, B, n. v.; probably destroyed)

OF DRYMARIA 215

Drymaria stereophylla <ru. exstipmUt* Mattf. in Notizbl. Bot. Gart. Berl. 13:437. 1936.
r. 12° 5' S., an Felsen, 3600-3700 m ii. M.;

Laxly spreading or ascending lignescent perennials to as much as 3 5 cm. tall, the
internodes shorter to longer than the leaves, glandular-puberulent, the taproot to
as much as 1 cm. thick. Leaves opposite, 4-12 mm. long, 2-6 mm. broad,
glandular-puberulent, broadly lanceolate to narrowly ovate, apically acute to
aristately acuminate, subsessile, the brief petioles clasping; stipules entire, lance-
linear, caducous, 0.5-1.0 mm. long, absent in the variety. Inflorescences of lax
occasionally congested terminal 3-many-flowered cymes, the peduncles 1-6 cm.
long; bracts lanceolate to lance-ovate, 3-5 mm. long, the pedicels 2-10 mm. long,
stipitate-glandular. Sepals 5, 5.0-7.5 mm. long, 2.0-4.5 mm. broad, glandular,
ovate, 3-4-nerved, the midrib not excurrent, the outer acute, the inner obtuse;
petals 5, 5-9 mm. long, deeply bifid, the lobes oblong, obtuse, 5-8 -nerved, exap-
pendiculate, not constricted at the junction with the trunk, the trunk truncate to
the claw; the claw minute, or, in the variety, as long as the trunk; stamens 5, 4-6
mm. long, the oblong anthers ca. 1 mm. long, the filaments basally connate forming
a cup about 1 mm. deep; ovary at anthesis turbinate to ellipsoid, the style 1.0-2.5
mm. long, bifid or trifid about one-third its length. Capsule ovoid, 2.5-3.5 mm.
long, 4-27-seeded, the seeds loosely circinnate, 0.9-1.3 mm. broad, tuberculatc, the
tubercles low and contiguous. Fig. 8B (Isotype of var. stereophylla) ; Fig. 8C
(Isotype of var. exstipulata).

Departments Junin, Cuzco and Lima, Peru.

Mattfield commented on the homogeneity of the specimens which he cited in
the original description of D. stereophylla var. exstipulata. Material collected sub-
sequently has further substantiated the differences in the varieties and other differ-
ences have been uncovered. The shape of the petals, frequent telescoping of the
internodes, and the absence of stipules in the var. exstipulata are characteristics
which point toward the bizarre D. frutescens. The two varieties of D. stereophylla
may be keyed as follows:

a. Stipules present, the leaves rigid, glabrous or with a few sessile glands; trunk of the petals

24. Drymaria auriculipetala Mattf. in Notizbl. Bot. Gard. Berl. 13:439.
1936. (holotype: MacBride & Featherstone 2264; Peru; Llata, von Fels-
bandern herabhangend, ca. 2100 m ii. M.; F!; isotypes at GH, MO, US)
Upright or ascending glabrous suffruticose annuals or perennials to as much as
60 cm. long. Leaves opposite, glabrous, 5-15 mm. long, 1.5-3.0 mm. broad, the
blades narrowly lanceolate, slightly falcate, apically attenuate, aristate, basally
clasping, the stipules entire, acicular, tardily deciduous, 1.0-1.5 mm. long. Inflores-
cences of lax terminal 1-7-flowered cymes, the glabrous peduncles 5-40 mm. long;
bracts broadly ovate, strongly 1-ribbed, 2.5-3.5 mm. long, exceeding the pedicels.
Sepals 5, subequal, 7-8 mm. long, 2-3 mm. broad, glabrous or minutely glandular,

216 ANNALS OF THE MISSOURI

lanceolate, apically attenuate and aristate, basally unguiculate or umbonulate, only
the midrib conspicuous; petals 5, 6.5-7.5 mm. long, bifid about half their length,
the lobes broadly oblong, apically rounded, 1.0-1.3 mm. broad, 4-6-nerved, ex-
apendiculate, hardly constricted at the junction with the trunk, the trunk 1.2-1.6
mm. broad, the claw minute; stamens 5, ca. 5.5 mm. long, the oblong anthers ca.
0.8 mm. long, often somewhat twisted; ovary at anthesis obovoid, slightly exceeded
by the anthers, the style trifid about half its length, about equaling the anthers.
Capsule ca. 4 mm. long, many-seeded, the seeds cochleate, ca. 1 mm. broad, dorsally
flattened or sulcate, tuberculate, the tubercles low and contiguous. Fig. 8D
(Holotype).

Known only from the type locality; Llata, Peru.

tf. in Notizbl. Bot. Gart. Berl. 13:439. 1936.
Weber bauer 7203; Peru; Dep. Libertad, Prov. Santiago de Chuco,
iiber der Hacienda Angasmarca; Grasssteppe mit zahlreichen eingestreuten
Strauchern, an felsigen Abhangen, 3650 m ii. M.; B, «. v.; isotypes at F, US!)
Ascending frutescent glabrous perennials to 30 cm. tall, the leaves closely
appressed to the stems. Leaves 2-6 mm. long, 1.0-1.5 mm. broad, imbricate in two
ranks, lance-ovate, apically aristate-acuminate, basally clasping, 3 -ribbed, margin-
ally transparent; stipules entire, narrowly lance-deltoid, ca. 0.5 mm. long. Flowers
solitary at the ends of the branches, the bracts similar to the foliage leaves, the
glabrous pedicels 2-3 mm. long. Sepals 5, subequal, 5.5-7.0 mm. long, 1.5-2.0
mm. broad, glabrous concavo-convex, ovate, apically acute or shortly acuminate,
transparent save for the 7 nerves; petals 5, 5-7 mm. long, bifid for about two-
thirds their length, the lobes oblong to spatulate, obtuse or emarginate, exappen-
diculate, the trunk and claw merging almost imperceptibly; stamens 5, 2.5-4.5
mm. long, the oblong anthers ca. 0.8 mm. long, often spirally contorted in drying;
ovary at anthesis ellipsoid, exceeded by the anthers, the style trifid about half its
length, about equaling the anthers. Capsule 2.5-3.0 mm. long; mature seeds not
seen. Fig. 8E (Isotype).

Known only from the type locality in Dept. Libertad, Peru.

K. Seri

Leaves opposite, linear-elliptic to narrowly ovate, subsessile, stipulate, the
stipules entire or bifid. Flowers in rather dense bracteate cymes, the bracts often
stipulate. Sepals acute, 1-3-nerved. Petals bifid, the lobes 1-3-nerved, exauricu-
late, the claw minute or completely absent (except in D. praecox) , the trunk as
wide as the combined width of the lobes. Stamens 3-5. Seeds loosely or tightly
cochleate, tuberculate. Three species in Peru, two rather delicate and minute
annuals and the more or less suffrutescent D. fasciculata. Fig. 9.

a. Plants upright, often suffrutescent; sepals 5-8 mm. long; leaves mostly larger than 6 mm.

long and 4 mm. broad; bracts not stipulate 26. D. fasciculata

a. Minute depressed or ascending annuals; sepals 2.0-4.5 mm. long, glabrous, glandular or

villosulous; leaves mostly less than 6 mm. long and 4 mm. broad; bracts often stipulate:

DUKE REVISION OF DRYMARIA

. Sepals acuminate, glabrous, apically reflexed in age; herbage occasionally glandular
not villose. Dept. Li . . •: Peru 27b. D. encleman/

. Sepals acute, stipitate-glandular, occasionally apically inflexed; herbage villos*
villosulous. Dept. Ancash and Cuzco; Peru:

26. Drymaria fasciculata A. Gray, Bot. U. S. Expl. Exped. 125. 1854. (holo-
type: Wilkes exped. s. «.; Andes of Peru, near Obrajillo, US!)
Upright dichotomous, paucifoliate annuals to as much as 25 cm. high, sub-
glabrous except for the glandular axes of the inflorescence, the taproot to as much
as 3 mm. broad. Leaves opposite, the blades 8-18 mm. long, 4-10 mm. broad,
ovate (the lowermost occasionally spatulate), apically acute or acuminate, margin-
ally somewhat undulate, basally clasping, only the midrib obvious, the stipules
entire, linear-lanceolate, caducous, ca. 1 mm. long. Inflorescences of terminal
fasciculate, strongly dichotomous, 3-many-flowered cymes, the glabrous to densely
glandular peduncles 2-6 cm. long, the bracts lanceolate to ovate, acute or
mucronate, transparent save for the midrib, 4-6 mm. long, 2-4 mm. broad,
obscuring the short pedicels. Sepals 5, subequal, 4-7 mm. long, 1.0-1.8 mm. broad,
lance-oblong to narrowly elliptic, acuminate, rather strongly 3-ribbed, often with
a scariose margin; petals 5, 2.5-7.0 mm. long, bifid about half their length, the
lobes oblong, apically rounded to emarginate, 2-3-nerved, their outer margins
forming straight lines with the margins of the trunk, exappendiculate, the trunk
exunguiculate or subexunguiculate; stamens 5, 2-6 mm. long, the oblong anthers
0.4-1.2 mm. long; ovary at anthesis subcylindric, slightly exceeded by the anthers,
the barely trifid style slightly exceeding the anthers. Capsule narrowly ellipsoid,
included, 3-4 mm. long, several-seeded, the seeds laxly circinnate, minutely and
rather evenly tuberculate, the anterior end usually long and pointed. Fig. 9A
(details from Holotype).
Andes of Peru.

27a. Drymaria engleriana var. engleriana
Polycarpan englerianum Muschler, in Bot. Jahrb. 45:452. 1911.

(holotype: Weberbaver 3WI; Peru; Ancash; Huaraz, 4300 m.; probably destroyed

atB; iso types at F,G!)
Drymaria engleriana (Muschler) Baehni & MacBride, in Field Mus. Bot. 13 2 :621. 1937.

Minute prostrate annuals to 5 cm. high, the internodes mostly longer than the
leaves, hirsutulous. Leaves opposite, the blades 2-8 mm. long, 1.0-4.5 mm. broad,
ovate, apically acute, marginally entire, basally attenuate to the alate clasping
petioles, weakly 3 -nerved, scantily hirsutulous, the stipules entire or bifid, the
divisions filiform to lance-deltoid. Inflorescences of terminal, rather lax, several-
flowered cymes, the axes hirsutulous, the bracts 1.5-3.0 mm. mostly longer than the
pedicels. Sepals 5, 2.0-3.6 mm. long, 0.6-1.2 mm. broad, narrowly ovate, apically
acute or briefly acuminate, often with sessile or stipitate glands, 3-ribbed, only the
midrib prominent, the apex not reflexed, often inflexed; petals 5, 1.0-2.4 mm. long,

218 ANNALS OF THE MISSOURI BOTANICAL GARDEN

praecox (holotype) ;

DUKE REVISION OF DRYMARIA 219

bifid about half their length, the lobes obtuse to acute, oblong, 1 -nerved, the outer
margin forming more or less a straight line with the margin of the trunk, the claw
not apparent; stamens 3-5, 1.0-1.5 mm. long, the suborbicular to oblong anthers
ca. 0.2 mm. long, the filaments basally connate into a minute carnose column;
ovary renoid, the style shorter than the ovary, ca. 0.5 mm. long, trifid about half
its length. Capsule globose to ellipsoid, included, 1.5-2.0 mm. long, 2-8-seeded,
the seeds laxly circinnate, minutely and evenly tuberculate, 0.7-0.9 mm. broad.
Fig.9C (Isotype).

High altitudes, Dept. Ancash, Peru.

Although Baehni and MacBride (in Field Mus. Bot. 13 2 :621. 1937) correctly
transferred this species to Drymaria from Poly car pott, MacBride (p. 627) also
retains it as a species of Polycarpon, curiously stating that it is "probably only a
densely cespitose form of P. apurense."

27b. Drymaria engleriana var. devia (Baehni & MacBride) J. Duke comb. &

Drymaria devia Baehni & MacBride, in Field Mus. Bot. 13 2 :620. 1937. (holotype: Vcn-

nell 14655; Lima; open rocky slopes near Canta, 4000 m., F!)

Delicate prostrate or ascending subglabrous annuals usually less than 5 cm. tall,
branching mostly from the base. Leaves opposite, the blades 2.5-5.0 mm. long, 2-3
mm. broad, subcoriaceous, ovate, occasionally glandular, apically acute or attenuate
and apiculate, basally clasping, the uppermost rather strongly 3 -ribbed; stipules
entire or bifid, the divisions narrowly lance-deltoid, ca. 1 mm. long. Inflorescences
of dense terminal 1-1 1 -flowered cymes, the glabrous peduncles 4-8 mm. long, the
bracts 2-3 mm. long, ca. 1 mm. broad, ovate, aristate-acuminate, the midrib often
cyanic; pedicels mostly less than 1 mm. long, concealed by the bracts. Sepals 5,
4.0-4.5 mm. long, 1.5-2.0 mm. broad, narrowly ovate, aristate-acuminate, strongly
3 -ribbed, glabrous to minutely glandular, transparent save for the ribs; petals 5,
2.5-3.0 mm. long, bifid about half their length, the lobes oblong, obtuse, 1-nerved,
curving outwardly to the inflated trunk, the trunk ca. 1.2 mm. broad, exappen-
diculate, the claw not apparent; stamens 5, ca. 2.2 mm. long, the oblong anthers ca.
0.4 mm. long; ovary at anthesis slightly exceeding the anthers, the style trifid about
half its length. Capsule 2.5-3.0 mm. long, ca. 8-seeded, the seeds dorsally flattened,
minutely tuberculate, ca. 0.6 mm. broad, dark brown. Fig. 9B (Holotype).

At high altitudes in Depts. Lima, Puno and Huancavelica, Peru.

This diminutive variety is closely related to the preceding and seems sufficiently

ribbed sepals and the lack of jointed hairs seem to characterize the few specimens
of the variety examined.

28. Drymaria praecox Baehni & MacBride, in Field Mus. Bot. 13 2 :625. 1937.
(holotype: Weber bauer 6gi6; between Pisac & Paucartambo, prov. Paucar-
tambo, Dept. Cuzco, 4100 m., Peru, F!; isotypes at MO, US, etc.)
Delicate, virgate, glandular-puberulent annuals to as much as 5 cm. high, with

few nodes below the inflorescence, the basal rosette often persistent. Basal leaves

220 ANNALS OF THE MISSOURI BOTANICAL GARDEN

spatulate; cauline leaves opposite, the blades 2—4 mm. ling, 1.5-3.0 mm. broad;
villosulous, ovate, apically acute, basally clasping, obscurely trinerved, the stipules
entire, acicular, 0.5-0.8 mm. long. Inflorescences of diffuse 1-9-flowered cymes,
the axes densely glandular-puberulent; bracts ovate to lanceolate, stipitate, 1 -ribbed,
the rib often excurrent, 1.0-1.5 mm. long, the pedicels usually much longer.
Sepals 5, 2.5-3.5 mm. long, 1.0-1.4 mm. broad, glandular-pubescent, narrowly
ovate, apically acute or obtuse, often inflexed, 3 -ribbed, transparent save for the
ribs; petals 5, ca. 3 mm. long, bifid half their length or more, the lobes oblong,
obtuse, exappendiculate, 3 -nerved, gradually tapering to the base, the trunk and
claw not discernibly differentiated; stamens 5, 1.2-1.5 mm. long, the suborbicular
anthers ca. 0.3 mm. long; ovary at anthesis compressed-globose, exceeded by the
anthers, about as long as the style; style trifid about half its length, slightly exceed-
ing the anthers. Capsule included, ca. 2.2 mm. long, ca. 5-seeded, the seeds
cochleate, ca. 0.7 mm. broad, tuberculate, the tubercles rather prominent, nearly
contiguous. Fig. 9D (Holotype).

Known only from the type collection from Cuzco, Peru.

Perhaps a misfit in this series, this species may be of hybrid origin, and in some
respects it recalls D. glandulosa var. galeottiana, perhaps also of hybrid origin.

L. Series debiles

29. Drymaria debilis Brandegee, in Proc. Cal. Acad. 2 2 :131. 1889. (holotype:

Brandegees. «.; Purisima, Feb. 13, 18 89; UC!)
Drymaria polystachya Brandegee, in Zoe 2:70. 1891. (holotype: Brandegee J5; cliffs

near San Jose del Cabo, Baja California; UC!; isotypes at D, US, etc.)
Drymaria diffusa J. N. Rose, in Contr. U. S. Nat. Herb. 1:13 0. />/. 12. 1892. (holotype:

Palmer 8lQ; Carmen Island; US!; isotypes at F, MO, US, NY, etc.)
Drymaria polystachya var. diffusa (Rose) Wiggins, in Proc. Cal. Acad. 4 25 :198. 1944.

Ascending or clambering, occasionally subscandent annuals or perennials, the
tenuous, villose or glandular internodes shorter to longer than the leaves, the tap-
root to as much as 15 mm. thick. Leaves opposite, the blades 3-20 mm. long, 3-20
mm. broad, villose or glandular, broadly ovate to reniform, apically acutish and
apiculate, basally cordate, often drying with a yellowish tinge, the petioles nearly
as long as the blades, the stipules mostly entire, filiform, 1-3 mm. long. Inflores-
cences of terminal lax 3 -many-flowered cymes, the axes villose or glandular
puberulent, the peduncles 1-10 mm. long; bracts narrowly to broadly ovate,
scariose, 1.5-2.5 mm. long, much exceeded by the pedicels, the villose or glandular
pedicels to as much as 25 mm. long. Sepals 5, 2.5-5.5 mm. long, 1.5-2.4 mm.
broad, glandular to villose, oblong to elliptic or ovate, apically obtuse, emarginate
or with the midrib occasionally excurrent, the venation obscurely dichotomous and
reticulate; petals 5, 3.0-6.5 mm. long, bifid about half their length, the lobes
oblong, rounded to truncate at the apex, 1-2-nerved, gradually tapered to the
claw; stamens 5, 1.0-3.5 mm. long, the filaments usually dimorphic, two or three
being distinctly longer than the others, the anthers oblong 0.5-1.0 mm. long; ovary
at anthesis ellipsoid, often stipitate, exceeded by the anthers, the style 1-2 mm.
long, trifid to half its length. Capsule ellipsoid to ovoid, 1.0-3.5 mm. long, 6-

DUKE REVISION OF DRYMARIA

222 ANNALS OF THE MISSOURI BOTANICAL GARDEN

15-seeded, the seeds ampulliform, 0.6-1.0 mm. broad, the pointed anterior ends not
strongly recoiled, the tubercles low, polygonal, rather remote. Fig. 10A (Holotype
of D. polystachya) ; Fig. 10B (Holotype of D. diffusa); Fig. IOC (Holotype of
D.debilis).

Baja California, Mexico.

This polymorphic species has generally been regarded as three separate species,
the more or less sympatric differences purported to reside in the duration, pubes-
cence, and the length and shape of the sepals and petals. I am more impressed by
the homogeneity of certain other characters (e. g. the peculiar retort-shaped seeds,
the inequality of the stamens, the venation of the sepals) than by the differences
supposed to separate the three species. Particularly suspicious to me are keys in
which the primary dichotomy is annual vs. perennial. Many Drymaria species,
originally described as annual, have turned out to have perennial representatives.
I prefer to think of D. debilis as a species embracing many contrasting character-
istics, but with few, if any, correlated contrasts.

M. Series laxiflores

30. Drymaria laxiflora Benth. PL Hartw. 73. 1839. (holotype: Hartweg
523; Guatemala; in rupibus Sunilo, BM? n. v.; isotypes at GL, F, LE, P, etc!)

in Ann. Cons. & Jard. Bot. Gen. 13: 370. 1911. (holo-
type: Pringle 331 ; M« us umbrosis prope Chihua-
hua; G, n. v.; isotypes at F, GH, MO, NY, P, US, etc.!)

Glabrous to densely stipitate-glandular diffuse to subcespitose perennials to as
much as 30 cm. high, the internodes mostly longer than the leaf-blades, the taproot
to as much as 6 mm. thick. Leaves opposite, the blades glabrous to densely
stipitate-glandular, ovate to reniform, 4-12 mm. long, 3-14 mm. broad, trinerved,
apically obtuse to acute and mucronulate, marginally entire, basally acute to
subcordate, the petioles 1-5 mm. long, the stipules bifid or trifid, rarely entire,
the divisions setaceous, 0.5-3.5 mm. long. Inflorescences of lax few-flowered
cymes, the peduncles 5-40 mm. long; bracts ovate, strongly 1 -ribbed, scariose,
apiculate, 1.5-3.0 mm. long, transparent save for the midrib, mostly shorter than
the pedicels, the pedicels subglabrous to densely stipitate-glandular, 2-6 mm. long.
Sepals 5, 2.5-6.0 mm. long, 1.5-2.2 mm. broad, glabrous to densely stipitate-
glandular, lanceolate to narrowly ovate, trinerved, acute, the midrib occasionally
excurrent, the laterals often subapically dissipating; petals 5, 2.5-6.0 mm. long,
bifid one-half to two-thirds their length, the lobes oblong, apically emarginate, with
one dichotomous vein, exappendiculate, continuous with the trunk, the trunk as
long as or longer than the claw; stamens 4-5, 2-4 mm. long, the oblong anthers
0.3-0.8 mm. long; ovary at anthesis ellipsoid, exceeded by the anthers, the elongate
style trifid one-third to one-half its length, slightly exceeding the anthers. Capsule
3-4-valved, 2-4 mm. long, many-seeded, the seeds 0.5-0.7 mm. broad, evenly
tuberculate, the dorsal tubercles conical, the facial tubercles substellate. Fig 10D
(Isotype of D. chihuahuensis) ; Fig. 10E (Isotype of D. laxiflora).

Texas through continental Mexico to Guatemala, to the south becoming less
glandular and tending to have larger floral parts.

DUKE REVISION OF DRYMARIA 223

Examination of several isotypes of Hartweg 523 from Guatemala reveals that
the plant, although described as glabrous, possesses a few scattered stipitate glands.
Bernoulli 6 Cario 3255 at Leningrad, collected in Quezaltenango, matches the types
of D. laxiflora in all respects except that it is more glandular. Toward the
northern end of the range, the species has some very glandular representatives whose
seeds are nearly black, pleiochroistic and very ornately sculptured. In spite of the
extreme variation, some of which appears to be along a north-south dine, only one
species seems to be involved. Bourgeau 2Q46 and Rose & Hay 5676 from Mt.
Orizaba have petals, the lobes of which are so deeply emarginate as to recall the
twice dichotomous petals of D. longepedunculata, and it is here suggested that
the Series laxiflores is genetically closely related to the Series excisae.

Leaves opposite, stipulate, mostly long-petiolate, and ovate to orbicular, often
villose. Flowers in few-many-flowered cymes, rarely subsolitary in the axils.
Sepals acute or obtuse, 1-3-nerved. Petals (occasionally absent) merely bifid, the
lobes usually ciliate-auriculate, 1-7-nerved, unguiculate. Stamens 3-5. Seeds
cochleate, tuberculate, the tubercles stellate, cylindric or capitate. Four species of
continental Mexico, Central and South America, one introduced in Malaysia.
Figs. 11 and 12.

224 ANNALS OF THE MISSOURI BOTANICAL GARDEN

31. Drymaria multiflora Brandegee, in Zoe 5:232. 1906. (holotype: Purpus
1653; dry hills, Salto de Agua, Mexico, UC!; isotypes at F, GH, MO, US, etc.)
Erect or ascending perennials to as much as 50 cm. tall, much branched, the
internodes mostly longer than the leaves, glabrous to stipitate- glandular, rarely
villosulous. Leaves opposite, the blades glabrous, 4-22 mm. long, 4-25 mm. broad,
broadly deltoid-ovate, orbicular or reniform, apically rounded to deeply emarginate,
marginally entire, basally truncate to acute, not cordate, weakly 3-5 -nerved; peti-
oles 2-20 mm. long, the stipules entire, lanceolate, 1.0-2.5 mm. long. Inflores-
cences of terminal 3 -many-flowered cymes, the peduncles to as much as 10 cm.
long, often stipitate- glandular; bracts ovate, 2.5-5.0 mm. long, the pedicels mostly
5-22 mm. long, glabrous or stipitate-glandular. Sepals 5, glabrous, unctuous,
rarely with sessile glands, lance-deltoid to deltoid-ovate, apically acute, at least
by extension of the midrib, 3-6 mm. long, 1.2-2.0 mm. broad, often saccate
basally; petals 5, 2.5-7.0 mm. long, usually equaling or exceeding the sepals, bifid
about two-thirds their length, the lobes apically rounded, oblong 2-5-nerved,
basally provided with ca. 3 ciliate auricles; stamens 3-5, 2.0-3.5 mm. long, the
oblong anthers 0.3-0.5 mm. long; ovary at anthesis globose to ovoid, the style 1-2
mm. long, bifid or trifid about half its length. Capsule 2.0-3.5 mm. long, ovoid,
4-26-seeded, the seeds ca. 1 mm. broad, the dorsal tubercles longer than broad, the
facial tubercles substellate. Fig. 11A (Isotype).
Sinaloa, Mexico, to Costa Rica.

Standley 834.78 from Quezaltenango is highly aberrant. Many flowers are
apetalous; others have entire liguliform petals; two flowers had twice-dichotomous
petals suggestive of D. longepedunculata.

Drymaria conzattii J. Duke

Plantae annuae erectae vel diffusae glanduloso-villosae usque ad 45 cm. altae
internodis saepe quam foliis longioribus. Folia opposita glanduloso-villosa reni-
forma apice vix acuminata basi cordata leviter 5-7 nervata, 5-20 mm. longa, 5-30
mm. lata, petiolis brevibus ad 2.5 mm. longis. Stipulae integrae vel laceratae lobis
filiformibus quam petioli longioribus. Inflorescentiae terminales axillaresque in
confertas 3-1 1-floriferes cymas dispositae rhachide glanduloso-villoso bracteis
ovatis 3.0-4.5 mm. longis costis excurrentibus quam pedicellis longioribus. Sepala
5 externa minuta glanduloso-puberula angusto-ovata acuta conspicue 3-costata,
5.0-5.6 mm. longa, 1.8-2.0 mm. lata, interna paulo breviora paene glabrescentia;
petala 5, 4.0-4.5 mm. longa 3/4-bifida lobis oblongis obtusis gradatim ad angusto-

oblongis ca. 0.6 mm. longis anthesi subglobosis ca. 0.6 mm. longis; ovarium anthesi
subglobosum ca. 1 mm. longum stylo subaequale, stylo vix divisio quam antheris
breviore. Capsulae maturae non visae. Fig. 1 IB (Holotype).

Mexico: Oaxaca: de Almoloya a Sta. Catarina, 1000 m., C. Conzatti 1688. holo-
type MO; isotype US.

Known only from the type collection.

DUKE REVISION OF DRYMARIA

225

33. Drymaria malachioides Briq. in Ann. Cons. & Jard. Bot. Gen. 13:372.

1911. (holotype: Galeotti 4415; Mexicum: Cordillere d'Ario, 4000'; G!;

isotypeatP!)

Clambering or ascending herbaceous annuals to as much as 45 cm. high, the
internodes longer than the leaves, villose with jointed hairs to as much as 1 mm.
long. Leaves opposite, the blades reniform to deltoid-ovate, 1-3 cm. long, 1-3 cm.
broad, apically rounded, marginally entire, obtuse to cordate basally, weakly

iotype); petal, 7^X; sepal. 7 ' , V. :

226 ANNALS OF THE MISSOURI

5 -nerved, scantily villose; petioles 1-20 mm. long, the stipules entire or bifid, the
divisions lanceolate to filiform, 1-2 mm. long. Inflorescences of terminal 3-many-
flowered cymes, ultimately lax, the peduncles 1-6 cm. long; bracts 1.5-2.5 mm.
long, the villose pedicels to as much as 12 mm. long. Sepals 5, 4.5-5.0 mm. long,
1.5-2.0 mm. broad, elliptic, apically acute, obscurely 3-nerved, with a few sessile
glands; petals 5, 5-6 mm. long, bifid about two- thirds their length, the lobes
spatulate, apically rounded to apiculate, several-nerved, basally provided with 4-6
ciliate auricles; stamens 5, 2-4 mm. long, the oblong anthers 0.5-0.8 mm. long;
ovary at anthesis ovoid, the styles longer than the ovary, trifid about half their
length. Capsule ellipsoid, ca. 4 mm. long, several-seeded, the seeds ca. 0.6 mm.
broad, cochleate, tuberculate, the dorsal tubercles filiform, the facial substellate.
Fig. 11C (Holotype).

Known only from the type collection, probably in Michoacan, and a few other
collections near Tancitaro, Michoacan.

Certain characteristics of the plant suggest that it might have been derived
from D. villosa, perhaps through polyploidy, since most of the plant parts are
proportionately larger than their equivalents in D. villosa.

34a. Drymaria villosa subsp. villosa

Drymaria villosa Cham. & Schlecht. in Linnaea 5:232. 1830. (holotype: Schiede &
Deppe 505; in aquosis prope Jalapam, B, probably destroyed; isotype at LE!)

birsuta Bartl. in Presl, Rel. Haenk. 2:8. 1831. (holotype: Haenke s.n.
Habitat in Peruviae montanis huanoccensibus, PR!)
Drymaria cubensis Regel, in Otto & Dietr. Allg. Garten. 8:298. 1840. (No type indi-
cated; horticultural)
Drymori* ciUaris Hort. Berol. ex C. A. Mey. in Ind. Sem. Hort. Petrop. 9:71. 1843.

(Hort.; no type indicated)
Drymaria ciliata Hort. Berol. ex C. A. Mey. loc. cit. 9:71. 1843. (Hort.; no type

indicated)
Drymaria cordata var. pilosa Schlecht. in Linnaea 26:374. 1853. (no type indicated)
Drymaria cordata var. 5 villosa (Cham. & Schlecht.) Rohrb. in Mart. Fl. Bras. 14 2 :260.

1872.
Drymaria stylosa Backer, in Bull. Jard. Bot. Buitenz. 2 12 :15. 1913. (no type indicated)
Drymaria tepicana M. E. Jones, in Contr. West. Bot. 15:124. 1929. (holotype: M. E.

Jones 22847; Tepic, Nayarit, POM, !)
Drymaria barrancae M. E. Jones, loc. cit. 18:65. 1931. (holotype: M. E. Jones 27051a;
La Barranca, Guadalajara, POM!)

Prostrate or ascending annuals to as much as 45 cm. long, the internodes mostly
longer than the leaves, villose to hirsute with septate hairs to as much as 2 mm.
long. Leaves opposite, the blades scantily to densely villose or hirsute with
cinereous or ochraceous hairs, orbicular to reniform, apically rounded to acute and
apiculate, basally cordate to truncate, weakly 3 -7- veined, 5-15 mm. long, 5-15
mm. broad; petioles 1-10 mm. long, the stipules mostly entire, 0.5-1.5 mm. long,
scarcely distinguishable from the indument. Inflorescences of terminal and axillary
5-many-flowered cymes, the ultimate branches often tending to be racemose;
peduncles 1-5 cm. long; bracts 0.5-1.5 mm. long, the pedicels 2-20 mm. long,
villose. Sepals 5, 2.0-3.6 mm. long, 1-2 mm. broad, narrowly to broadly ovate
or elliptic, apically acute to obtuse, villose, occasionally glabrescent, weakly

DUKE REVISION OF DRYMARIA 227

cally reduced in
IT.], 2.0-3.6 mm.

long, bifid for half their length or more, the lobes apically acute to deeply
emarginate, 4-nerved, basally provided with caducous filiform auricles, these
variable in number and orientation; stamens usually 5, 2.0-3.5 mm. long, the
oblong anthers 0.3-0.5 mm. long, the filaments shallowly connate, devoid of
staminodia; ovary at anthesis ovoid to globose, the style 1.0-1.5 mm. long, trifid
to as much as half its length. Capsules ovoid to ellipsoid, 2.0-3.5 mm. long,
equaling or exceeding the sepals, many-seeded (rarely as few as two in D. bar-
rancae), the seeds cochleate, 0.5-0.9 mm. broad, tuberculate, the dorsal tubercles
cylindric to capitate, the facial tubercles stellate. Fig. 11D (Holotypc of D.
barrancae) ; Fig. 1 IE (Isotype of D. hirsuta) .

Central Mexico through Central America along the western coast of South
America to Peru; apparently introduced and rather widespread in the East Indies.

Examination of adequate material from Mexico, type locality of D. lillosa, from
Peru, type locality of D. hirsuta, and from Java, type locality of D. stylosa, clearly
shows that these are conspecific. Indonesian material is marked although incon-
stantly, by glabrous sepals and hirsute indument, in contrast to Mexican material,
which is predominantly villose.

I have seen no specimen representing this species from the West Indies, and
believe very strongly that the epithet cubensis results from error or misconception.
The contention seems to be supported by the following quote (from Ind. Sem.
Hort. Petrop. 9:71. 1843): "1093. Drymaria ciliata. Eadcm planta, jam diu
mis et D. ciliatae H. Berol. nomine culta, nunc sub nomine D. cubensis
H. Berol. in hort. occurit. Num sit D. cubensis species a D. ciliata distincta?
Ignoramus. — Nostra planta D. villosae est proxima et vix nisi pilis in caule et
foliis rarioribus ad ilia differt. M." I have seen numerous horticultural specimens
under these names from various European herbaria and they were all referable to

(Cham. & Schlecht.) J. Duke comb.

Drymaria palustris Chan

Deppe 504; in palu

tfDeppe 405, LE\)
Drymaria pauciflora Bartl. in Presl, Rel. Haenk. 2

Drymaria cordata var. /3 palustris (Cham. & Schlecht.) Rohrb. in Mart. Fl. Bras. 14-:260.
1872.

Drymaria townsendii Robinson, in Bot. Gaz. 30:58. 1900. (holotype: Towmend f$
Barber 231 ; on the Sierra Madre 8 km. southeast of Colonia Garcia, Chihuahua, GUI;
isotypes at F, MO, US, etc.)

Drymaria nummularia Briq. in Ann. Conserv. & Jard. Bot. Gen. 13:371. 1911. (holo-
type: Galeotti 4416; Mexicum: Cordillere d'Ario, 4000', G!; isotype at K).

Drymaria mbm UsVL T jonei in Contr. West. Bot. 15:125. 1929. (holotype: M. E.

Drymaria sphagnophila Baehni & MacBride, in Field Mus. Bot. 13 2 :625. 1937. (holo-
type: MacBride 1542; sunny sphagnum bog, Mito, Huanuco, Peru; F!)

228 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Prostrate annuals or erect perennials (in var. perennis J. Duke var. nov.*)
shorter to longer than the leaves, glabrous, glandular, or rarely villose and glabres-
cent, the prostrate forms often rooting at the nodes, the erect forms tending to
become suffrutescent. Leaves opposite, the blades 2-15 mm. long, 2-15 mm. broad,
deltoid-ovate to reniform, apically rounded to apiculate, marginally entire, basally
obtuse to cordate, glabrous, glandular, or villose and usually glabrescent, weakly
3-7-nerved; petioles 0.5-6.0 mm. long, the stipules entire, (rarely lacerate in
var. perennis) filiform to lance-deltoid, 0.5-2.0 mm. long. Inflorescences of
terminal and axillary 1-many-flowered cymes, the peduncles 0-10 cm. long; bracts
ovate, to elliptic, acute to obtuse, 0.5-2.0 mm. long, the pedicels glabrous, rarely
glandular or villosulous, 1-15 mm. long. Sepals 5, 1.0-3.0 mm. long, 1-2 mm.
broad, glabrous or with sessile glands, oblong to orbicular, apically obtuse and
often cucullate, very obscurely trinerved, the midrib subapically dissipating, often
distally saccate; petals 5 [absent in subsp. palustris forma townsendii (Robinson)
J. Duke comb. & stat. nov.], 1-4 mm. long, bifid from half to nearly
all their length, the lobes oblanceolate to oblong, apically acute to emarginate, 1-
2-nerved, the shallowly cleft petals exauriculate or with mere dentations, the deeply
cleft petals usually with ciliate, downwardly directed auricles; stamens 2-5, 1.0-2.5
mm. long, the oblong to suborbicular anthers 0.2-0.4 mm. long; ovary at anthesis
ovoid to globose, the styles 1.0-1.5 mm. long, bifid or trifid half their length or
more, usually at least as long as the ovary. Capsule 2.0-3.5 mm. long, usually
exceeding the sepals, occasionally twice as long, 5-24-seeded, the seeds cochleate,
0.6-1.1 mm. broad, densely tuberculate, the dorsal tubercles longer than broad,
occasionally capitate, the facial tubercles usually stellate. Fig. 12A (Holotype of
D. villosa ssp. palustris var. perennis) ; Fig. 12B (Isotype of D. nummularia) ; Fig.
12C (Isotype of D. pauciflora) ; Fig. 12D (Holotype of D. sphagnophila) ; Fig. 12E
(Holotype of D. townsendii) ; Fig. 12F (Holotype of D. subsessilis).

Mexico through Central America at high altitudes down western South Amer-
ica to Peru. Bolivia.

Three rather easily discernible elements may be distinguished in this subspecies.
Firstly, there is typical ssp. palustris, which (if I am correct in assuming that
Schiede S? Deppe 405 at Leningrad is the same as Schiede & Deppe 504, the
holotype) has large petals cleft nearly to the base and provided with ciliate
auricles. This is an annual plant, usually prostrate and often rooting at the nodes.
I have seen no specimens from outside Mexico. Among Mexican specimens of this
element are M. E. Jones 22848, Bell & Duke 16624, Rose & Hay 4800, H. E. Moore
2801, Hinton et al. 12233, and /. H. Maysilles 7782.

By far the most common element is that prostrate form whose petals are not
so deeply cleft and not so ornate, i.e. there are no ciliate auricles. In habit this
is similar to the typical form and no other characters are discernably correlated.

Potosi, Mexico, MO.

DUKE REVISION OF DRYMARIA

230 ANNALS OF THE MISSOURI BOTANICAL GARDEN

To this more prevalent portion of the population, the following names have been
applied: D. paucifiora, D. ..gnopbila.

The third element seems to have received no name previously. To this belong
erect perennial relatives which have little else in the way of differentia, except a
more pronounced tendency to have subacute, somewhat cucullate sepals and
glandular herbage. Representing specimens of the var. percnnis, which may have
resulted from gene exchange with D. laxiflora or one of its relatives, are Palmer 356,
and 187, Pringle 3265, I. M. Johnston 8965, M. T. Edwards 804B, H. E. Moore
1669, and Salinas, Kowell 8 Barkley 16M373.

In Ecuador where the plants are considered medicinal, this, the "Drimaria
llamba" is reputed to be the most potently therapeutic of the local species.

34c. Drymaria villosa subsp. paramorum (S. F. Blake) J. Duke comb. & stat.

Drymaria paramorum S. F. Blake, in Contr. U. S. Nat. Herb. 20:521. 1924. (holotype:
A. Jabn III; Paramo de la Cristalina, Trujillo, Venez.; alt. 2900 m.; US!)
Prostrate or ascending annuals to as much as 20 cm. long, the internodes
mostly longer than the leaves, sparsely to densely villose with caducous septate
hairs. Leaves opposite, the blades 5-10 mm. long, 4-8 mm. broad, broadly ovate,
apically acutish, basally rounded to cordate, trinerved, sparsely to densely villose;
petioles 0.5-3.0 mm. long, the stipules entire, scariose, lance-deltoid, ca. 1 mm.
long. Inflorescences of lax 3-many-flowered cymes, the glabrous to villose
peduncles to 4 cm. long; bracts scariose, ovate, acute, mucronulate, 1.0-2.5 mm.
long, the glabrous to villose pedicels 2-6 mm. long. Sepals 4-5, subequal, 3.5-4.2
mm. long, 1.2-2.0 mm. broad, the outer villose, broadly oblong, obtuse, weakly
trinerved, marginally scariose, the inner glabrous, almost nerveless, broader; petals
4-5, 3.5-5.0, bifid about half their length, the lobes oblong, 3-7-nerved, obtuse,
exauriculate or with minute deltoid auricles at the base, tapered abruptly to the
linear claw; stamens 4-5, 2-3 mm. long, the oblong anthers ca. 0.5 mm. long;
ovary at anthesis subglobose, the style trifid about half its length, slightly exceeded
by the anthers. Capsule equaling or exceeding the sepals, ca. 3.5 mm. long, 15-
34-seeded, the seeds 0.8-1.2 mm. broad, tuberculate, the tubercles stellate. Fig.
12G (Holotype; [seed from Vittier 13170]).

Trujillo and Merida, Venezuela; also from the Cordillera Oriental in Colombia.

Certain specimens of D. ovata from Ecuador (e. g. Fosberg & Giles 23225 and

Rose 8 Rose 23225) seem to share some of the characteristics of this subspecies.

O. Series grandiflores
Leaves opposite, ovate to reniform, sessile or petiolate, entire or crenulate,
stipulate, the stipules entire or lacerate. Flowers in lax or subcapitate terminal or
axillary many-flowered cymes. Sepals acute or obtuse, weakly to strongly 3-7-
nerved, the midrib often carinate. Petals bifid (absent in D. apetala), unguiculate
(subexunguiculate in D. glaberimma) , exauriculate or with deltoid auricles, the
lobes 2-15-nerved. Stamens 5. Seeds cochleate, tuberculate (smooth in forms of
D. ovata) , the tubercles domical or conical, not stellate nor capitate nor cylindric.

DUKE REVISION OF DRYMARIA 231

8 species from Colombia to Argentina, along western South America. Figs. 13

glabrous; inflorescences often long-pedunculate in
iubexunguiculate, the lobes oblong.
of the seeds secondarily tuberculate; trunk of the p

ovate, glabrous to

.

g. Sepals more or

less obtuse, sho

g. Sepals acute, a

f. Petals less than 6

to the claw; lea

es ovate, glabr

i. Petal-lobes oblong, tapered to the claw, slightly if at all exserted; inflore

Sepals obtuse to acute, if carinate, then obtuse, 2.5-5.0 mm. long, glabrc
slightly glandular; Peru and the Galapagos Islands, Ecuador and Bolivia:
j. Sepals ellipsoid to orbicular, very obtuse, occasionally slightly glandular....

35. Drymaria firmula Steyermark, in Field Mus. Bot. 28 1 :227. 1951. (holo-
type: Steyermark 56536; Venezuela; Merida; Paramo de los Colorados, between
El Molino & San Isidro Alto, 2745-2955 m. F!)

Ascending or erect herbs to as much as 25 cm. high, the internodes mostly
longer than the leaves, often cyanic. Leaves opposite, the blades glabrous, sub-
pergameneous, 3.0-8.5 mm. long, 2.5-7.0 mm., apically acute, marginally involute,
basally cordate, subsessile or on petioles to 1.5 mm. long, the stipules lacerate, the
divisions linear, caducous, equaling or exceeding the petioles. Inflorescences of lax
terminal 3-7-flowered cymes, the axes glabrous to minutely glandular; bracts
ovate, 1 -nerved, scariose, 1.5-2.5 mm. long, mostly equaling or exceeding the
subtended pedicels. Sepals 5, subequal, glabrous, 3.8-4.8 mm. long, 1.4-2.0 mm.,

232 ANNALS OF THE MISSOURI BOTANICAL GARDEN

DUKE REVISION OF DRYMARIA 233

narrowly ovate to oblong, obtuse, trinerved, the midrib of the outer, somewhat
narrower, sepals often excurrent; petals 5, 4.5-5.0 mm. long, bifid slightly more
than half their length, the lobes broadly oblong, obtuse, 1-4-nerved, ca. 1 mm.
broad, gradually tapered to the claw, exappendiculate or unidentate; stamens 5,
3.0-3.5 mm. long, the oblong anthers ca. 0.6 mm. long; ovary at anthesis ellipsoid,
the long style trifid for about one-third its length, about attaining the apex of the
sepals. Capsule ovoid, 3-valved, 2 mm. long, 2-seeded, the seed cochleate, 1.2-1.5
mm. broad, dorsally tuberculate, the tubercles low and domical, facially merely
reticulate, sublustrous. Fig. 13A (Holotype).

Known only from the type collection.

The lacerate stipules, small domical tubercles on the large seeds and the occa-
sionally 1 -nerved petal-lobes suggest a close affinity with D. cordata. On the other
hand, the size and shape of the petals, coupled with the subsessile, lubpergtmeneoui
leaves, suggest a closer alliance with D. ovata, D. monticola and D. glaherrima,
and these four poorly known species seem to form a closely knit group.

hex. n.v.)

Prostrate to rigidly ascending annuals or perennials to as much as 50 cm. high,
the internodes mostly longer than the leaves, glabrous, farinose, glandular-
puberulent or densely villose. Leaves opposite, the blades glabrous to villose,
membranaceous to subpergameneous, deltoid-ovate to reniform, apically acute and
apiculate to rounded, basally truncate to subcordate, 5-40 mm. long, 5-30 mm.
broad, subsessile or the petioles to as much as 8 mm. long, the stipules mostly
lacerate, the divisions filiform to lanceolate. Flowers broadly obconic, 1-many in
lax terminal, rarely axillary, cymes, the peduncles to as much as 12 cm. long;
bracts lanceolate, 2-5 mm. long, usually shorter than the glabrous to villose
pedicels. Sepals 5, glabrous to villose, elliptic to ovate, 4.0-6.5 mm. long, apically
acute to obtuse, marginally entire, 3-8-nerved, the inner broader, more obtuse and
glabrous; petals 5, 4-7 mm. long, bifid one-half to three-fourths their length, the
lobes narrowly oblong, 2-5-nerved, apically obtuse or emarginate, basally truncate
or rarely tapered to the narrow claw, usually exappendiculate; stamens 5 (-6), the
oblong anthers 0.4-1.0 mm., the filaments 2-5 mm. long; ovary at anthesis ovoid,
the style 1-2 mm. long, bifid or trifid about half its length. Capsules 2.0-4.5 mm.
long, 2-15-seeded, the seeds tightly circinnate, subdiscoid, 0.9-1.5 mm. broad,
smooth or provided with low domical tubercles, often nearly black, lustrous. Fig.
13B (Lebmann6ll2).

Along the Andes from Venezuela to northern Argentina.

Exhibiting a confusing array of variation, this polymorphic species is difficult
of definition, but a rather constant correlation of flower- and petal-shape seems to
prevail. The lustre of the seeds, equaled nowhere else in the genus, is usually
evident, even in those forms which have tuberculate seeds. The following con-
trasting characteristics may be found in representative specimens: ovate vs. reni-

234 ANNALS OF THE MISSOURI BOTANICAL GARDEN

form leaves, sessile vs. petiolate leaves, membranaceou:
petals tapered vs. truncate to the claw, and smooth
this number of variables, and undoubtedly more exist, many combinations are
possible, and I have seen many of them, but am unable to detect any correlations
among these characteristics except that, south of Peru, all specimens have had
smooth seeds and large leaves. Weberbajier 654$ from Peru differs in having
subpergameneous leaves, 1-2-nerved petal-lobes, and a farinose girdle on the
pedicels, characteristics suggesting that it might have resulted from hybridization
with D. cordata. From Huanuco, Macbride & Featherstone 1714 has most of the
features of D. cordata but the seeds are nearly smooth and lustrous, and the veins
in the petal-lobes have several branches.

Collections made by Prieto for W. W. Camp in Prov. Cariar and Azuay of
Ecuador possessed the colloquial names "Drimaria del cerco" and "Drimaria macho"
and the plants are utilized as treatments for liver and kidney ailments.

37. Drymaria apetala Bartl. in Presl, Rel. Haenk. 2:7. 1831. (h
Haenke s. n. Habitat in Chile; PR! )

Stellaria virgata Seringe, in DC. Prodr. 1:396. 1824. not D. virgata Briq.,

type: patr. ign. "Alsine virgata Deless. herb.": G, n.v.)
Drymaria nmcrautba A. Gray, Bot. 1

exped.s.n.- Obrajillo; US!)
Drymaria virgata Briq. in Ann. Cons. & Jard. Bot. 13 :370. 1911.

Virgate or sparingly branched suffrutescent perennials (?) to as much as 30
cm. high, the internodes mostly longer than the leaves, locally farinose. Leaves
opposite, the blades glabrous to minutely glandular, narrowly to broadly ovate,
apically acute, basally cordate and somewhat clasping, 5-15 mm. long, 3-9 mm.
broad; stipules mostly entire, deltoid, acuminate, caducous, 0.5-1.5 mm. long.
Inflorescences of rather dense few-flowered cymes, the axes often densely glandular;
bracts lanceolate, acuminate, strongly 1-ribbed, 2.5-4.0 mm. long; pedicels 2-10
mm. long. Sepals 5, subequal, rather densely glandular, 7.5-9.0 mm. long,
narrowly ovate, apically acute with the midrib often excurrent as an apiculum,
basally rather strongly trinerved; petals 5, reduced to mere vestiges or completely
absent; stamens 5, 5.5-6.5 mm. long, the oblong anthers 1.8-2.0 mm. long; ovary
at anthesis narrowly ellipsoid, the trifid style exceeded by the anthers. Capsule
included, 5.5-6.5 mm. long, many-seeded, the submature seeds cochleate, rather
sharply pointed at the hypocotylar end, dorsally sulcate, weakly tuberculate, the
tubercles low and subhemispherical. Fig. 13C.

Although attributed to Mexico, erroneously, and questionably to Chile, I have
seen no specimens that were collected outside of Peru. The Pavon specimens from
"Mexique" and from Peru, probably are duplicates of one collection.

38. Drymaria glaberrima Bart
Haenke s. «.; habitat in Peru
GOET)
Procumbent or ascending lignescent annuals or perennials to

high, the internodes mostly longer than the leaves. Leaves <

DUKE REVISION OF DRYMARIA 235

glabrous, narrowly to broadly ovate or elliptic, apically acute, marginally entire or
undulate, basally obtuse to truncate, 3-5-nerved, subsessile, the stipules filiform,
caducous, usually exceeding the petioles. Inflorescences of mostly axillary 3-11
flowered cymes, the peduncles 1-8 cm. long; bracts lanceolate to ovate, 1.5-2.5
mm., all except those of the central flowers equaling or exceeding the glabrous
subtended pedicels. Sepals 5, glabrous, narrowly ovate, apically acute to obtuse,
weakly 3-5-nerved, 4.0-5.0 mm. long, the inner obtuse, with a broad hyaline
margin, ca. 2 mm. broad; petals 5, 3.0-4.5 mm. long, bifid halfway or slightly
more, the lobes oblong, obtuse, 3-7-nerved, ca. 0.5 mm. broad, the trunk ca. 1 mm.
broad, truncate to a minute claw, or completely exunguiculate; stamens 5, ca.
3 mm. long, the oblong anthers ca. 0.6 mm. long, the filaments basally connate
into a very shallow cup; ovary broadly oblongoid, the style ca. 1.5 mm. long,
trifid ca. half its length. Capsule oblongoid, 2-10-seeded, the seeds tightly circin-
nate, tuberculate, the dorsal tubercles longer than broad, often themselves second-
arily tubercululate, the facial tubercles low and domical or slightly elongated.
Fig. 13D (Isotype).

Known from several stations, all in Peru.

The material that I am citing for this species is heterogeneous indeed and it is
possible that more than one taxon is involved. Intensive collecting and field
studies will be needed to clarify this whole series, which is still full of problems.
39 Drymaria monticola Howell in Proc. Cal. Acad. 421:329. 1935. (holo-

type: /. T. Hoivell 9243; Mt. Crocker, Indefatigable Island; CAS!)

Decumbent or ascending perennials, to as much as 40 cm. high, the glabrous
internodes exceeding the leaves, the roots to 4 mm. thick. Leaves opposite, the
blades glabrous, 3-5-nerved, ovate, acute, marginally undulate, basally truncate to
subcordate, subsessile; the stipules mostly bifid, 1-2 mm. long, exceeding the
petioles. Flowers in rather dense 5-9-flowered cymes, the peduncles 2-7 mm. long,
glabrous, the bracts ovate, 1.0-1.5 mm. long, shorter or slightly longer than the
glabrous pedicels. Sepals 4-5, glabrous, narrowly to broadly ovate, weakly 3 -ribbed
4.5-6.0 mm. long, 1.5-2.5 mm. broad, apically acute, the midrib subapically
dissipating, the lateral ribs weak, with 2 intercalary veins; petals 4-5, 4-5 mm.
long, bifid about half their length, the lobes oblong, obtuse, 2-4-nerved, often
basally auriculate, tapered to the linear claw; stamens 4-5, 3.5-4.5 mm. long, the
oblong anthers 0.7-0.9 mm. long, the filaments basally connate into a cup to as
much as 0.7 mm. deep; ovary globose to ovoid, stipitate, the styles 1.5-2.0 mm.
long, trifid a little more than half their length. Capsule 2.0-3.5 mm. long, 8-
16-seeded, the seeds tightly cochleate, ca. 1 mm. broad, tuberculate, the dorsal
tubercles domical, slightly higher than broad, grading into the polygonal facial
tubercles. Fig. 13E (Holotype).

Known only from the Galapagos Islands.

This species was likened by its author to D. apetala (D. macrantha) and in its
sessile leaves it does somewhat resemble that species. A closer relative appears to
be D. firviula from Venezuela which differs somewhat in its strongly three-ribbed
sepals and exappendiculate petals. D. glaberrima also belongs in this relationship
but differs in its tuberculate tubercles and obtuse inner sepals.

236 ANNALS OF THE MISSOURI BOTANICAL GARDEN

40. Drymaria grandiflora Bartl. in Presl, Rel. Haenk.

Haenke s. n. Habitat in Chile, et in Peruviae montanis huanoccensibus; PR!)
Prostrate or clambering rarely ascending perennials, the internodes mostly longer
than the leaves, villose to glabrescent. Leaves opposite, the blades glandular-
villosulous, rarely glabrous, narrowly to broadly ovate, weakly to strongly 3 -nerved,
apically acute, marginally entire, basally rounded to truncate, usually clasping; the
main leaves 10-30 mm. long and 5-20 mm. broad, those of the branches much
smaller; stipules entire to lacerate 1.0-2.5 mm. long, usually longer than the
petioles. Flowers narrowly to broadly campanulate in terminal, lax, few-flowered
cymes, the peduncles 1-15 cm. long; bracts 2-5 mm. long, usually much exceeded
by the pedicels, the pedicels usually densely glandular- villosulous, 3-25 mm. long.
Sepals 5, glandular-puberulent to glabrous, ovate, weakly nerved, 6-8 mm. long,
the outer sometimes carinate and acute, the inner flatter and often obtuse; petals 5,
7—12 mm. long, bifid about two-thirds their length, the lobes linear to oblong,
apically obtuse, many nerved, to as much as 3 mm. broad, basally gradually merging
with the poorly differentiated trunk and claw, exauriculate or rarely unilaterally
unidenticulate; stamens 5, the anthers oblong 1.0-1.5 mm. long, the filaments 4-6
mm. long, connate into a cup as much as 1.5 mm. deep; ovary at anthesis ellipsoid,
usually longer than the style, the style 2-3 mm. long, trifid about one-third its
length. Capsule ellipsoid, 4-5 mm. long, 5-many-seeded, the seeds tightly circin-
nate 1.2-1.6 mm. broad, coarsely tuberculate, the tubercles dorsally conical, some-
what flattened, grading facially into polygonal tubercles. Fig. 13F.

Peru; departments Huanuco, Cajamarca, and Lima; reported from Chile, but

41a. Drymaria
Drymaria paposana Phi

valle Tartal alisque locis ejusdem regionis crescit, SGO? n. v. photograph of isotype

from destroyed Austrian herbarium; MO!)

Upright or ascending annuals to as much as 25 cm. high, the internodes mostly
longer than the leaves, glandular-puberulent, the taproot to as much as 2 mm.
thick. Leaves opposite, the blades narrowly to broadly ovate, apically acute,
basally cuneate to subcordate, glabrous or scantily glandular-puberulent, the upper
leaves sessile, 5-20 mm. long, 4-18 mm. broad, the stipules lacerate, the divisions
filiform, mostly exceeding the petioles. Inflorescences of terminal subcapitate or
monopodial cymes, the peduncles 1-30 mm. long, glandular-puberulent, the bracts
ovate, usually longer than the glandular pedicels. Sepals 5, 3.5-5.0 mm. long,
ellipsoid, glandular to glabrous, apically obtuse, weakly 3 -nerved, the midrib sub-
apically dissipating, the outermost sepals occasionally carinate; petals 5, 3.5-4.5
mm. long, bifid about half their length, the lobes oblong, rarely to 1 mm. broad,
apically obtuse or emarginate, 2-3-nerved, tapered gently to the claw, exauriculate;
stamens 5, 2.7-3.0 mm. long, the oblong anthers ca. 0.6 mm. long; ovary at
anthesis ovoid, the style ca. 1.2 mm. long, trifid about half its length. Capsules
2.6-3.0 mm. long, many-seeded, the seeds cochleate, 0.6-0.8 mm. broad, the
tubercles conical, acute. Fig. 14D.

DUKE REVISION OF DRYMARIA 237

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Northern Chile.

Although I. M. Johnston lumped this with D. cordata, it is quite distinct from
that species in many respects.

41b. Drymaria paposana var. weberbaueri (Muschl.) J. Duke comb. & stat.

Drytnaria weberbaueri Muschl. in Engl. Bot. Jahrb. 45:451. July 25, 1911. (holotype:

s. m.; B, n.v., probably destroyed; as lectotype may be designated Weberbauer 1657

Drymaria fenzliana Briq. in Ann. Cons. & Jard. Bot. Gen. 13:373. May 10, 1911. (holo-

Erect or ascending annuals to as much as 3 cm. tall, the few internodes mostly
longer than the leaves, glabrous to glandular-pubescent. Leaves opposite, the blades
glabrous to glandular-puberulent, narrowly to broadly ovate, 5-20 mm. long, 4-15
mm. broad, apically acute, basally cuneate to truncate or subcordate, usually sub-
sessile, the lacerate stipules mostly longer than the petioles. Inflorescences of
terminal, usually subcapitate, 3-many-flowered cymes, the peduncles 1-4 cm. long,
the pedicels 0.5-3.0 (-5.0) mm. long, glandular-puberulent, usually exceeded by
the ovate, occasionally dentate, bracts. Sepals 5, 3.5-7.0 mm. long, ovate, usually
glandular, the outer acute and often carinate, 1-3 -nerved, the inner often broader,
more obtuse, less glandular; petals 5; 5-7 mm. long, usually exceeding the sepals,
bifid little more than half their length, the lobes broadly oblong to spatulate, to
2 mm. broad, 3-many-nerved, apically obtuse or emarginate, basally auriculate,
often asymmetrically so, more or less truncate to the linear claw; stamens 5, 3.0-
5.0 mm. long, the oblong anthers 0.5-1.0 mm. long, the filaments shallowly
connate; ovary at anthesis ovoid to ellipsoid, the style 1.0-2.5 mm. long, trifid for
about one-third its length. Capsule ovoid, 3-valved, often stipitate, 2.5-4.0 mm.
long, many-seeded, the seeds cochleate, 0.5-1.0 mm. broad, tuberculate, the
tubercles mostly domical and obtuse. Fig. 14B (Holotype of D. fenzliana); Fig.
14C (lectotype of D. weberbaueri) .

Lomas of Peru from the departments of Lima to Arequipa.

With its rather dense clover-like heads, this is perhaps the most handsome
'Drymaria. A nice photograph may be seen in Goodspeed and Stork (in Univ. Cal.
Publ. Bot. 28: pi. 8. 1955). Although Briquet in his description of D. fenzliana
believed it to be a Mexican species, it becomes apparent that the label on the Pavon
specimen had suffered the same fate as those on some of his specimens of D. apetala,
some of which were labeled as being from Mexico and others from Peru, although
obviously belonging to the same collection.

Toward the southern end of its range, this variety intergrades with the var.
paposana and the newly described var. serrulata.

41c. Drymaria paposana var. serrulata J. Duke var. nov.

Plantae perennes erectae vel glabrae sublignosae internodis quam
Iongioribus, radice usque 1 cm. in diametro. Folio opposita laminis glabris anguste
vel late ovatis apice acutis

DUKE REVISION OF DRYMARIA 239

marginibus conspicue crenulatis 5-nervatisque subsessilibus, stipulis laceratis quam
petiolis longioribus. Flores 3 usque multi cymis terminalibus laxis aut vix con-
fertis pedunculis 1-8 cm. longis, pedicellis minus 1 cm. longis rare divaricatis.
Sepala 5, 5-8 mm. longa, 2.5-3.5 mm. lata minute granulosa angusto-ovata vel
elliptica, apice acuta (externissimo carinato cum carinis saepe serrulatis) ; petala 5,
6-8 mm. longa, 2/3 -bifida lobis oblongis 4-6-nervatis exauriculatis vel auriculatis
parvis vix ungue constricto; stamina 5, 4.5-6.5 mm. longa, antheris oblongis ca.
1.5 mm. longis; ovaria ovoidea vel ellipsoidea, stylo ca. 2 mm. longo, 3/4-trifido.
Capsulae inclusae seminibus multis ca. 0.8 mm. latis crasse tuberculatis opacis.
Fig. 14A.

Peru: Dept. Arequipa, Prov. Caraveli; Lomas de Atiquipa (Taimara) ; falda arcillosa-
pedregosa, 500-550 m.; Ramon Ferreyra 13490. (holotype MO).

Not uncommon in the lomas from the departments of Arequipa and Moquegua,
this unusual variety seems largely to replace var. weberbaueri, and although it
resembles that variety in habit, the differences in their flowers are quite striking.
Although the epithet applies to the carina of the outer sepals, the leaves themselves
are often distinctively crenulate, so much so as to appear serrulate. See Fig. 14A.

42a. Drymaria rotundifolia var. rotundifolia

Drymaria rotundifolia A. Gray, Bot. U. S. Expl. Exped. 123. 1854. (holotype: Wilkes

Expedition s. n. Obrajillo, Peru US!)

Upright annuals to as much as 20 cm. high, the internodes mostly longer than
the leaves, glabrous to glandular-puberulent. Leaves opposite, the blades glabrous
to minutely glandular, reniform, apically rounded or apiculate, basally clasping,
4-10 mm. long, 5-15 mm. broad; the stipules lacerate, caducous. Inflorescence of
rather congested 3-1 1-flowered cymes, the peduncles glandular, 1-45 mm. long;
bracts scarious, ovate, acute, 0.5-2.0 mm. long; pedicels mostly 1.5-3.0 mm. long,
those of the central flowers to 8 mm. long; glandular. Sepals 5, glabrous or sessile-
glandular, broadly ellipsoid, obtuse or rounded, subcucullate, weakly 3 -nerved, the
midvein subapically dissipating, often carinate, marginally scarious and transparent,
2.5-3.5 mm. long, 1.2-2.8 mm. broad; petals 5, 3.2^.5 mm. long, bifid about half
their length, the lobes oblong, ca. 0.5 mm. broad, exappendiculate but somewhat
flared before narrowing into the linear claw; stamens 5, about 3 mm. long, the
oblong anthers about 0.6 mm. long; ovary at anthesis turbinate to ellipsoid, the
style trifid about half its length, ca. 1.5 mm. long, exceeded by the anthers.
Mature capsule ellipsoid, 2.5-3.5 mm. long, several-seeded, the seeds cochlea te,
often dorsally sulcate, 0.6-0.8 mm. broad, tuberculate, the tubercles conical, obtuse
or rarely acute. Fig. 14E (Holotype).

Peru.

Stewart 1 512 from the Galapagos Islands might just as well be included here as
in var. nitida. The specimen has some of the characters of both varieties, and
perhaps represents a new entity.

42b. Drymaria rotundifolia var. nitida (J. Ball) J. Duke comb. & stat. nov.
Drymaria nitida J. Ball, in Jour. Linn. Soc. 22:31. 1887. (holotype: Peru: Ex saxosis

Andium Peruviae juxta pagum Chicla, 12-13000' s. m. /. Ball s. n. K, n. v., isotype at

GH!)

240 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Minute to rather large ascending annuals to as much as 1 5 cm. high, the inter-
nodes glabrous, mostly longer than the leaves. Leaves opposite, the blades glabrous,
often somewhat fleshy, narrowly ovate to elliptic, 3-10 mm. long, 2-8 mm. broad,
glabrous, apically acute to cuspidate, marginally hyaline, basally attenuate and
clasping, the venation obscure; stipules entire, very fugaceous, acicular, less than
1 mm. long. Flowers campanulate, in rather dense few-flowered terminal cymes,
the lowermost bracts merging with the leaves, often stipulate, ovate, ca. 1.5 mm.
long; peduncles 4-7 mm. long, the pedicels glabrous, exceeded by or slightly
exceeding the bracts. Sepals 5, glabrous, 2.5-3.0 mm. long, the outer broadly
oblong and weakly carinate, the inner suborbicular, obtuse, to 2 mm. broad,
obscurely nerved, the midnerve often subapically dissipating; petals 5, 3.2-4.0 mm.
long, bifid about two-thirds their length, the lobes narrowly spatulate, apically
obtuse, 2-3-nerved, exauriculate, acutely tapered to the claw; stamens 3—5, 2.5-3.5
mm. long, the anthers oblong; ovary at an thesis cordate, ca. 8 -ovulate, the style
deeply trifid. Capsules subglobose, 3-8-seeded, ca. 2 mm. long, the seeds ca. 1 mm.
broad, tuberculate, the tubercles subdendritic. Fig. 14F (Isotype).

Peru, Ecuador and the Galapagos Islands.

Although the type is a diminutive plant, other specimens that I refer to this
variety attain a good size and in some respects approach D. glaberrima: e.g.
Mandon Q48 from Bolivia. Characters in the flowers however suggest that it is
closer allied to D. rotundifolia and D. paposana, especially its var. weberbaueri.

Leaves opposite, petiolate, elliptic to reniform, stipulate. Flowers in cymes, the
pedicels usually long and divaricate, exceeding the bracts. Sepals trinerved, acute,
occasionally saccate basally. Petals bifid about half their length, with mamilliform
auricles (except D. divaricata forms), the lobes (1-) 2-5-nerved. Stamens 4-5.
Seeds cochleate, tuberculate, the tubercles domical conical, occasionally subdendritic
or mamilliform. Five poorly defined varieties in Peru and Bolivia. Fig. 1 5 .

a. Leaves broadly ovate to reniform, short- to long -petiolate, glabrous to villosulous; pedicels

sepals; seeds with mar

b - Sft?

Zt£X^

: venly keeled,
les of the seeds
iscid-glandular:

not basally saccate

d. Sepals

and herbage subg.abrc

,u, to vilU»u1o

us; pedicels diva™

:at C ;

. Leaves e'ln

ttSArs

rs, glabrous; se

t, glabrous; pedi«
pals glabrous; lorn;

Is strongly auri,
alpine plants....

DUKE REVISION OF DRYMARIA

ANNALS OF THE MISSOURI

Drymaria divaricata HBK. Nov. Gen. & Sp. 6:24. 1832. (holotype: Humb. 6 Bonpl.

s. n. Crescit ad litora Oceani Pacifici, prope Lima Peruvianorum, P n. v.)

Laxly spreading annuals, the internodes much longer than the leaves, to as much
as 10 cm. long, minutely glandular-puberulent. Leaves opposite, the blades
glabrous or minutely puberulent or villosulous, narrowly to broadly deltoid-ovate,
6-12 mm. long, 6-15 mm. broad, apically acute, marginally entire, basally trun-
cate; the distally alate petioles 2-5 mm. long, the stipules filiform, caducous,
mostly entire, 1.0—1.5 mm. long. Inflorescence of very lax terminal cymes, the
flowers many, squarely campanula te; the glandular-puberulent peduncles to 10 cm.
long; bracts oblong to lance-linear, ca. 1 mm. long; the pedicels glandular-
puberulent, to as much as 35 mm. long. Sepals 5, glandular-puberulent to glabrous,
unctuous, triangular-ovate, apically obtuse to acutish, and often cucullate, basally
saccate, 3.0-4.5 mm. long; petals 5, 5-9 mm. long, bifid about two-thirds their
length, the lobes narrowly spatula te, weakly 2-5 -veined, apically obtuse, basally
merging gradually with the linear claw, exauriculate to minutely denticulate;
stamens 5, 3.5-5.0 mm. long, the anthers oblong 0.3-0.6 mm. long. Ovary at
anthesis ovoid, the style trifid one-fourth to one-half its length 1.5-2.5 mm. long.
Capsule ovoid, 2.5-3.5 mm. long, ca. 8-seeded, the seeds tightly circinnate, 0.8-1.5
mm. broad, coarsely tuberculate, the dorsal tubercles longer than broad, sub-
dendritic. Fig. 15B (Pennell 14560).

Province of Lima, Peru.

Since I have not seen the type of D. divaricata, I sent specimens of all the
varieties herein recognized to Dra. Lourteig for comparison with the Humboldt and
Bonpland specimen in the Historical Herbarium at Paris. After comparison with
the various specimens sent, she selected Ferreyra 3460 as the best match for D.
divaricata HBK. Also representing the typical variety is Pennell 14560 which is
illustrated in Fig. 15B. This has quite long petals proportionately much longer
than those of Ferreyra 3460. Both specimens have peculiar dwarf flowers in the
axils of some of the unusually small bracts. Both were collected at altitudes higher
than 2700 meters, which might seem incompatible with the type locality, the
description of which would at least suggest that D. divaricata might be a loma

43b. Drymaria divaricata var. stricta (Rusby) J. Duke comb. & stat. nov.
Drymaria stricta Rusby in Phytologia 1:55. 1934. (holotype: G. H. H. Tate 160.

Pongo de Quime Bolivia, alt. 12,000 ft. NY!)
Drymaria agapatensis Baehni & Macbride, in Field Mus. Bot. 13:619. 1937. (holotype:
Lechler 1947% Pnno, Agapata, G!; isotype at UPS, F, MO, P, etc.)
Upright or ascending lignescent annuals or perennials to as much as 30 cm.
high, the internodes mostly longer than the leaves, minutely but obviously
glandular-puberulent. Leaves opposite, glandular-puberulent, the blades ovate to
reniform, apically acute, marginally entire or undulate, basally truncate to cordate,
5-14 mm. long, 5-20 mm. broad, the petioles often alate, 2-6 mm. long, stipules
lacerate, the laciniae filiform to acicular, 1.5-2.0 mm. long. Inflorescence of

OF DRYMARIA 243

mostly terminal lax, several-flowered cymes, the peduncles 1-6 cm. long, glandular-
puberulent, bracts oblong to lance-deltoid, 1-2 mm. long, mostly exceeded by the
pedicels, these 1-10 mm. long, sepals 5, glandular-puberulent, ovate, basalb nccatt
but not carinate, apically acute to shortly acuminate, subcucullate, 4.5-5.5 mm.
long, weakly 3-4-nerved; petals 5, 4.5-5.5 mm. long, bifid for 2-3 mm., the lobes
linear, obtuse to emarginate, narrowed to the linear claw, 1 -nerved, usually
auriculate; stamens 5, the anthers often aberrant, oblong to orbicular, ca. 0.3 mm.
long; ovary at an thesis ovoid, the style about as long, 1.5-2.5 mm. long, trifid
about half its length. Capsule 2.5-3.5 mm. long 2-8 -seeded, the seeds tightly
circinnate, 0.8-1.0 mm. broad, densely tuberculate, the dorsal tubercles secondarily
tubercululate, the facial irregularly polygonal. Fig. 15C (Isotype of D. agapatensh) .

Peru and Bolivia.

The seeds of this variety are among the most bizarre in Drymaria. Some speci-
mens are referred here only reluctantly as they approach D. glaberrima, a species
which shares the unusual tuberculation of the seeds.

43c. Drymaria divaricata var. viscidula (A. Gray) J. Duke comb. & stat. nov.

Upright annuals to as much as 35 cm., the internodes mostly longer than the
leaves. Leaves opposite, the blades glabrescent, broadly ovate, apically acute and
mucronulate, basally subcordate, 5-8 mm. long, 6-10 mm. broad, the veins very
obscure; petioles 1-2 mm. long, glabrescent; stipules entire or bifid, the divisions
linear, deciduous, 1.0-1.5 mm. long. Inflorescences of 3-9-flowered cymes, the
peduncles rather densely glandular-pubescent; bracts linear, 1.0-2.5 mm. long,
gradually reduced; pedicels 1.0-2.5 mm. long. Sepals 5, subequal, the outer
glandular, broadly lanceolate, acute, strongly 3 -nerved, 3-4 mm. long, ca. 1 mm.
broad, the inner flatter, obtuse, 1.0—1.5 mm. broad, with scariose margins outside
the 3 ribs; petals 5, 3.5-4.0 mm. long, bifid about half their length, the lobes
oblong, ca. 0.5 mm. broad, basally auriculate, the auricles obtusely deltoid, the claw
gradually tapering to the auricle; stamens 5, ca. 3.2 mm. long, the anthers oblong,
ca. 0.5 mm. long; ovary at an thesis subglobose, exceeded by the anthers, the style
trifid more than half its length, slightly exceeding the anthers. Capsule ovoid,
2.8-3.5 mm. long, maturing ca. 6 seeds, the seeds cochleate, ca. 0.7 mm. broad,
dorsally tuberculate with low rounded tubercles broader than long, facially
flattened or sulcate, merely lineolate. Fig. 15D (Holotype).

Peru, at rather high elevations.

Although in some respects (especially the seeds) suggestive of D. glanduhsa, I
think a perusal of the plate depicting the Series divaricatae will convince one that
the var. viscidula is more properly aligned here.

43d. Drymaria divaricata var. divergens J. Duke var. nov.

Herbae annuae graciles diffusae glanduloso-pilosulosae internodis saepe quam
foliis longioribus. Folia opposita laminis reniformibus apice rotundatis vel obtusis

244 ANNALS OF THE MISSOURI BOTANICAL GARDEN

mucronatisque basi truncatis aut rotundatis, 5-12 mm. longis, ea in ramis lateral-
ibus saepe reducta petioli saepe in longitudine laminae foliorum aequales stipulis
lanceolatis frequenter quam petiolis brevioribus. Flores in cymas terminales pleures
dispositi pedicellis saepe divergentibus frequenter 2-4 X in longitudine pluribus
quam floribus bracteis lanceolato-linearibus usque ad 3 mm. longis. Sepala 5,
3.5-4.5 mm. longa ovata acuta 3-nervata vix stipitato-glandulosa; petala 5, sepalis
aequalia ut vix plus, bifida lobis angusto-spatulatis obtusis vel truncatis 3-5-
nervatis basi auricula tis ungue lineari; stamina 5 quam petala brevioria, antheris
oblongis ca. 0.6 mm. latis. Capsulae inclusae seminis cochlea tis. Ca. 0.8 mm. latis
tuberculatis tuberculis dorsalibus latis quam longioribus. Fig. 15E (Holotype).
Peru: Dept. Lima, prov. Pativilca; Lomas de Pacar, alt. 150-700 m. O. V. Nunez

2320 (HOLOTYPE, US).

If I have erred in my interpretation of the typification of D. divaricata HBK,
this variety of the Peruvian lomas would be my second choice, and indeed were it
not for the kind assistance of Dra. Lourteig, I believe this new variety would be
masquerading under the name D. divaricata HBK. The similarities of the two
concepts are strong, but the slight morphological differences plus the ecologic

43e. Drymaria divaricata var. reflexiflora J. Duke var. no v.
Drymaria divaricata Auct. Am. non HBK.

Plantae parvae ascendentes ad 25 cm. altae internodis quam foliis saepe longi-
oribus glabris. Folia opposita glabris, eis in rostellam aggregatis spatulatis longe
petiolatisque gradatim foliis caulis anguste ad late ovatis vel elliptico-ovatis acutis
basi truncatis saepe undulatis (in marginibus) mixtis, eisdem etiam secundum
breves alatos petiolos attenuatis, petiolis saepe quam stipulis brevioribus stipulis
saepe trifidis lobis filiformibus, 1.0-2.5 mm. longis. Flores campanulata in laxas
evidenter trichotomas cymas foliis veris ovatis bracteis conmixtis; bractea 2-4
mm. longae, acutae vel acuminatae, prope aequales vel praestantes divaricatis glabris
pedicellis. Sepala 5, glabra ovata vel elliptica, 3.0-4.5 (-5) mm. longa evidenter
3-costata basim umbonata vix saccata exteriora acuta, interiora obtusa vel acuta;
petala 5, vix exserta 3.0-4.5 mm. longa, 2/3 -bifida lobis angusto-spatulata apice
obtusa 3-7-nervata, basi auriculata vel biauriculata turn in unguem linearem con-
stricta; stamina 5 antheris fertilibus saepe reductione 3 oblongis filamentis 3.0-3.5
mm. longis; ovarium ovoideum, 2.5-3.5 mm. longum, stylo 1.0-1.8 mm. longo,
1/2-fido trilobato. Capsule ovoideae, 15-25-seminiferes semini
circinatis ca. 0.6 mm. latis tuberculis dorsalibus latis quam longioribus sc
superficie tuberculis conicis vel ellipsoideis munitis. Fig. 15A (Holotype).

Peru: Dept. Lima, Prov. Chancay; Lomas de Chancay, 500-600 m.; Ferreyra I.
(holotype, MO).

Although long confused with typical D. divaricata HBK., also reputed 1
a loma plant, there is little doubt that this variety is distinct, morphologically
possibly ecologically. Plants representing this new variety bear little resemb
to var. divaricata, as depicted in a photograph of the type.

OF DRYMARIA
Q. Serie
Leaves opposite, ov

Sepals mostly acute, 1-3-nerved. Petals bifid (occasionally ligular in D. xero-
phylla), unguiculate, exauriculate (poorly defined auricles occasionally present in
D. gracilis), the lobes 1 -nerved (except D. galeottiana) . Stamens 2-5, the anthers
orbicular to oblong. Styles 2-3, free or basally united. Seeds 1-many, cochleate,
tuberculate, the tubercles domical or rectanguloid, contiguous to remote. Five
species, 1 more or less pantropical, 1 ranging from western U. S. A. to Argentina,
the others confined to Mexico and vicinity. Figs. 16, 17, 18.

itico and r

tubercles more or less domical and robe r lacerate:
f. Leaves deltoid-ovate, the stipules entire; secondary veins forming a crinkled inter-
rupted reticulum; stamens 4-5 I tor half their length or more;
southern Mexico, Guatemala and Bolivia 47. D. ladewii

mm. broad; Old World Tropics .' ' 48b. D. cordata ssp. DIANDRA

44a. Drymaria gracilis ssp. gracilis

Drymaria gracilis Cham. & Schlecht. in Linnaea 5:232. 1830. (holotype: Schiedc &

Deppe 503; prope Jalapa. B, n. v. isotype at GOET!)
Drymaria cordata var. 7 gracilis (Cham. & Schlecht.) Rohrb. in Mart. Fl. Bras. 14 2 :260.

1872.

Erect or ascending perennials to as much as 45 cm. tall; the internodes mostly
longer than the leaves, glabrous or minutely stipitate-glandular; leaves opposite, the
blades glabrous, rarely villosulous or punctate, deltoid-ovate (narrower upwards),
5-15 mm. long, 3-15 mm. broad, apically acute to rounded and apiculate, mar-
ginally entire, basally truncate 3-5 -nerved, the petioles 1-8 mm. long; the stipules

246

ANNALS OF THE MISSOURI

GARDEN

bifid or trifid, 1.0-2.5 mm. long, the divisions filiform. Inflorescences of lax
terminal, many-flowered cymes, the peduncles 2-6 cm. long; bracts ovate, carinate,
0.5-1.5 mm. long; pedicels 3-6 mm. long, glabrous, divaricate. Sepals 5, ovate,
glabrous, 2.0-2.7 mm. long, apically acute, 1-nerved, carinate; petals 5, 2.0-2.7
mm. long, bifid at least two-thirds their length, the lobes obtuse to emarginate,
basally entire to auriculate, divaricate, tapered or truncate to the minute claw;
stamens 5, rarely fewer, 1.8-2.2 mm. long, the anthers oblong, 0.4-0.5 mm. long,
briefly connate into a shallow cup; ovary at anthesis globose, the styles elongate,
bifid or trifid less than half their length, exceeding the ovary; capsule globose,
equaling or exceeding the sepals, 1.8-2.5 mm. long, many-seeded, the seeds
cochleate, 0.5-0.6 mm. broad, coarsely tuberculate, the dorsal tubercles domical,
the facial tubercles polygonal, approximate. Fig. 16B.

DUKE REVISION OF DRYMARIA 247

Hidalgo to Vera Cruz, Mexico.

Specimens in the Lund Herbarium, and the herbaria at Leningrad and the Field
Museum, although labeled as grown from seeds of D. divaricata, are clearly refer-
able to D. gracilis. Somewhere a mixup in labels must have occurred. For that
reason, I sent a specimen of this species to Dra. Lourteig for comparison with the
type of D. divaricata. As mentioned in the discussion of D. divaricata previously,
she cited the Peruvian Ferreyra 3460, very different from the Mexican D. gracilis, as
comparable to D. divaricata HBK.

44b. Drymaria gracilis ssp. carinata (Brandegee) J. Duke comb. & stat. nov.

la Laguna, Baja Calif. UC!)
Drymaria carinata var. perennis Wiggins, in Proc. Cal. Acad. Sci. 4 25 :197. 1944. (holo-
type: Gentry 441$; La Laguna, Sierra Laguna UC!)

Prostrate to ascending annuals or perennials to as much as 40 cm. high, the
internodes mostly longer than the leaves, glabrous. Leaves opposite, the blades
glabrous, narrowly to broadly deltoid-ovate, 5-20 mm. long, 2-15 mm. broad,
apically rounded to acute, marginally entire, basally truncate, 3-nerved, petioles
1-10 mm. long, the stipules entire, 1-2 mm. long, filiform. Inflorescences of
terminal lax 3-many-flowered cymes, the peduncles 2-6 cm. long; bracts ovate,
carinate, 1.0-1.5 mm. long, the pedicels 2-12 mm. long, glabrous. Sepals 5, ovate,
glabrous, 1.5-2.5 mm. long, apically acute to acuminate, 1 -nerved and carinate;
petals 5, 1.5-3.0 mm. long, bifid at least two-thirds their length, the lobes apically
obtuse to emarginate, 1 -nerved, divaricate, basally auriculate to entire, abruptly

0.3-0.4 mm. long, shallowly connate into a cup; ovary at anthesis globose, the
style trifid one-half to five-sixths its length. Capsule globose, 1.5-2.0 mm. long,
many-seeded, the seeds cochleate, 0.4-0.6 mm. long, minutely and remotely tuber-
culate. Fig. 16C (Isotype).

Baja California, Mexico.

This subspecies shares with the typia
niality. Probably separated for some ti
subspecies exhibit an interesting disjuncti

45a. Drymaria glandulosa

Drymaria glandulosa Presl, Rel. Haenk. 2:9. 1831. (holotype: Haenke s. «.; habitat in
Mexico, Pr!)

Drymaria rar, nnaea 12:206. 1838. (holotype: Hegewiscb s. n.,

prope urbem Mexico, GOET (?) n. v.)

Drymaria kptoclados Hemsl. Diag. Pi. Nov. 2. 1878. (holotype: Bernoulli 240, Guate-
mala; Camino del Sapote; K!; isotype at NY)

Drymaria fendleri S. Wats, in Proc. Am. Acad. 17:328. 1882. (holotype: Fendler 60,
New Mexico, GH!)

Drymaria leptoclados var. peruviana J. Ball, in Journ. Proc. Linn. Soc. 22:32. 1885.
(holotype: /. Ball s.n.-.i * < ?) at GH!)

Drymaria blasiana M. E. Jones, Contr. West. Bot. 15:125. 1929. (holotype: M. E.
Jones 22845; in open places, San Bias, Sinaloa; POM!)

Drymaria fendleri var. perennis M. E. Jones, Extr. Contr^ West. Bot. 18:65 1933.
E. Jones 27050, Laguna Mtns., Baja Cal. POM!; isotypes at UC, etc.)

248 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Drymaria glandulosa var. fendleri (S. Wats.) Fosberg, in Morton, in Contr. U. S. Nat.

Herb. 29:96. 1945.
Drymaria glandulosa var. perennis (M. E. Jones) Fosberg, loc. cit. 29:96. 1945.

Sparingly to densely stipitate glandular upright or ascending annuals or peren-
nials, the internodes mostly longer than the leaves, the taproot to 1 cm. thick.
Leaves opposite, the blades sparingly to densely glandular-puberulent, ovate to
reniform, 5-12 mm. long, 5-20 mm. broad, apically rounded and apiculate, basally
cordate, marginally entire, the petioles 1-8 mm. long; stipules lacerate, the divisions
lance- linear to filiform. Inflorescences of terminal and rarely axillary lax to dense
5 -many- flowered cymes, the peduncles (0-) 2-8 cm. long (the peduncles often
axillary and abbreviated in var. fendleri); bracts lanceolate-aristate 2.5-5.0 mm.
long, often exceeding the pedicels (especially in var. fendleri), lanceolate, weakly
to strongly 3 -ribbed, the ribs subapically confluent, the mid-rib often excurrent,
the outer sepals frequently conspicuously shorter than the inner and more attenuate
at the apex, 3.0-4.6 mm. long; petals 3-5, 1.2-3.0 mm. long, bifid about half their
length, the lobes linear, apically acute to emarginate, 1 -nerved, basally auriculate or
entire; stamens 2-5, the anthers oblong, 0.2-0.4 mm. long, the filaments 1-2 mm.
long; ovary at anthesis subglobose, longer than the style, the style bifid or trifid
about half its length; capsule 1.5-2.8 mm. long (2-) 4-20-seeded, the seeds
cochleate, 0.5-0.8 mm. broad, minutely but regularly tuberculate, the dorsal tuber-
cles domical, the facial tubercles rectanguloid or lineolate, remote. Fig. 16D
(Holotype D. leptoclados); Fig. 16E (Holotype D. fendleri).

Baja California; Arizona through western South America along the Andes to
Argentina; apparently skipping southern Central America.

Fosberg (in Contr. U. S. Nat. Herb. 29:96. 1945) recognizes three varieties
of D. glandulosa, a variable species indeed. The varieties are separated on basis of
varying vegetative characters, var. fendleri having subglabrous leaves and var.
perennis being perennial. Var. fendleri also tends to have cinereous less cordate
leaves and more dense inflorescences. Although the type of D. glandulosa was
described as being an annual, the root is apparently perennial and nearly 3 mm.
thick. Since the perennial habit occurs in both the subglabrous and the densely
pubescent forms and all possible combinations of the characters can be found
throughout the range of the species I see little reason for their recognition.

Some of the forms are only with difficulty separated from D. cordata, but the
seeds of the two have markedly different sculpture. Drymaria glandulosa is also
marked by longer anthers and proportionately longer inner sepals, and the inflores-
cence is more obviously cymose with a conspicuous central leader. The perennial
habit, frequently encountered in D. glandulosa, is rarely if ever encountered in
D. cordata.

Mandon 950 from the Andes of Bolivia (2600-2700 m.) dept. Larecaja,
referred tentatively by Ball (Kew Bull. 22:32) to D. leptocladus var. peruviana is
obviously D. glandulosa, and in its tendency toward glabrescence it does resemble
D. leptocladus.

Berlandier 999 and 1122 from Mexico are distinctive in that the petals have
pli-nerved veins and subcylindric tubercles thus suggesting intergression with a
member of the Series villosa.

DUKE REVISION OF DRYMARIA 249

45b. Drymaria glandulosa var. galeottiana (Briq.) J. Duke comb. & stat.

Drymaria galeottiana Briq. in Ann. Conserv. & Jard. Bot. Gen. 13:373. 1911. (holotype:
Galeotti 4408; near Oaxaca, Mexico, G!; isotypes at K [sub 4416] and L.)
Rather delicate erect annuals to as much as 20 cm. high, the internodes mostly
longer than the leaves, glandular-puberulent, the roots fusiform. Leaves opposite,
the blades glandular-puberulent and glabrescent, broadly ovate to reniform, apically
obtuse and apiculate, marginally entire, basally truncate to subcordate, subsessile,
the stipules bifid or trifid, 1-2 mm. long, about equaling the petioles. Flowers
obconic, in rather dense terminal cymes, the peduncles 1-2 cm. long, glandular-
all but the central pedicels which may be up to 6 mm. long. Sepals 5, subequal,
glandular-puberulent, 3.0-3.6 mm. long, 1.5-2.8 mm. broad, ovate to broadly
oblong, 3 -nerved, the midrib sometimes prominent and excurrent, the inner sepals
obtuse, the outer acute, transparent save for the nerves; petals 5, 2.5-3.2 mm. long,
bifid ca. half their length, the lobes oblong, obtuse, 2-4-nerved, tapering gradually
to the claw, exappendiculate; stamens 3-5, 1.5-2.2 mm. long, the anthers sub-
orbicular, 0.2-0.3 mm. long; ovary at anthesis subglobose, the style ca. 1 mm.
long, trifid ca. half its length. Mature capsule subglobose, 2.5-3.0 mm. long,
many-seeded, the seeds cochleate, about as long as broad, ca. 0.6 mm. broad, the
dorsal tubercles low and domical, the facial tubercles rectangular, somewhat remote.
Fig. 16A (Holotype).

Known only from few scattered collections in Oaxaca and Hidalgo, Mexico.
At Kew, the specimen labeled 4416 is called D. galeottiana and that labeled 4408 is
called D. nummularia, so apparently the labels have been crossed. This variety is a
misfit in this series because of the petal nervation and it may prove to represent a
mere hybrid. The seeds, sepals, and pubescence suggest D. glandulosa whereas the
nervation of the petals and to some extent their shape and size, recall D. laxiflora.
These two are the most likely candidates for the parents if indeed the plant is
a hybrid.

A. Gray, PL Wright. 2:18. 18 5:
Conifer 722) .

Ascending perennials profusely branched from the base, the internodes longer
or shorter than the leaves, glabrous. Leaves opposite, the blades broadly to nar-
rowly ovate, glabrous, strongly 3-7-nerved, 5-10 mm. long, 3-10 mm. broad,
apically acute and apiculate, marginally entire, basally rounded to obtuse, subsessile;
the stipules lacerate, the divisions lance-linear to filiform, exceeding the petioles.
Inflorescences of dense axillary cymes, or fascicles, the lanceolate acuminate bracts
longer than the glabrous peduncles and pedicels, 1.5-2.5 mm. long. Sepals 5,
oblong, apically acute to obtuse, strongly 3 -nerved with the nerves subapically
confluent, basally umbonate, 3-4 mm. long, glabrous to scantily stipitate-glandular;
petals 3-5, entire and Iigulate or bifid for half their length, the lobes apically

ANNALS OF THE MISSOURI BOTANICAL GARDEN

denticulate or entire, 1-nerved, 1.0-1.8 mm. long; stamens 3, ca. 1 mm.
anthers oblong, the filaments basally connate into a shallow cup; ovary ai
subglobose, the style about as long, bifid or trifid about half its length,
ellipsoid, 3-valved, 1-3-seeded, the seeds cochleate, 0.7-0.8 mm. broad,
tuberculate or merely reticulate on the fades. Fig. 17.
San Luis Potosi to Oaxaca, Mexico.

Rusby, in Phytologia 1:54. 1934.
; Nequejahuira; alt. 8000 ft. NY!)

Drymaria hyperidfolia

Diffusely spreadin

leaves, strongly divar

DUKE REVISION OF DRYMARIA 251

deltoid-ovate, apically rounded to acute and apiculate, marginally entire, basally
more or less truncate; stipules entire, more or less lanceolate; larger leaves to 2.8
cm. long, 2.5 cm. broad, trinerved, with the secondary veins forming an obvious
but interrupted reticulum, the leaves as the rather long petioles strongly reduced
upwardly. Inflorescences of terminal and axillary 5-many-flowered cymes, the
axes glandular-puberulent, the bracts lanceolate, 1.5-3.0 mm. long, exceeded by the
long, divaricate pedicels, the pedicels to 12 mm. long. Sepals 5, lance-oblong, 2-3
mm. long, apically obtuse or acute by extension of the midrib, weakly 1-3 nerved,
the midrib often distally subcyanic; petals 5, about equaling the sepals in length,
bifid about three-fourths their length, the lobes obtuse, exaunculate, subdivancate,
weakly 2-5 -nerved, the claw carnose, subcylindric; stamens 5; 1.5-2.5 mm. long,
the oblong anthers ca. 0.5 mm. long, the filaments basally connate into a more or
less carnose ring; ovary at anthesis turbinate, the style trifid ca. half its length,
about attaining the anthers. Capsule turbinate, 3-valved, included (seeds rarely
mature in the specimens at hand), 1-5-seedcd, the seeds similar to those of
D. cor data. Fig. 18 A.

Southern Mexico, Guatemala, Bolivia.

After scrutiny of the copious Guatemalan material and the scanty Bolivian
material (2 specimens, one from the O. Kuntze herbarium, the other the type;
[the second specimen cited by Rusby, in his description of the species, apparently
is a hybrid with D. cordata or some related species and must be excluded from
D. laden H] ) , I am unable to find the slightest differences between them. Asserting
rather strongly that there is no obvious difference between the Guatemalan and
Bolivian material, I am quite at a loss to explain the peculiar disjunction exhibited

n.v.)
Holosteum cordatum L. Sp. PI. 88. 1753.

Holostcum diandrum Sw. Prodr. 27. 1788. (holotype: habitat in Jamaica, S!)
Drymaria diandra MacFadyen, Fl. Jam. 1:52. 1837. not Blume, 1825.
Drymaria cordata /3 diandra (Sw.) Griseb. Fl. Brit. W. Ind. 56. 1859.
Drymaria cordata p puberula Tr. & Pi. in Ann. Sci. Nat. 4 17 :148. 1862. (holotype

Triana s. «.; Bogota alt. 2600 m., P, n. v.; isotype at NY!)
Drymaria procumbent J. N. Rose, in Contr. U. S. Nat. Herb. 1:304. 1895. (holotype

Palmer 1165, near water about Colima, Mex., US!; isotype at GH)
Drymaria adenophora Urban, in Fedde, Repert. Sp. Nov. 21:213. 1925. (holotype

Ekman 12978, Cuba; prov. Pinar del Rio prope Cortez, B?; isotype at G, S, NY!)
Stellaria adenophora (Urb.) Leon in Leon & Alain, Flora de Cuba 2:154. 1950.
Drymaria cordata var. pacifica Mizushima, in Jour. Jap. Bot. 32:78. 1957 (holotype

Svenson 65; Academy Bay, Indefatigable Island, Galapagos Islands, 100 ft. alt.; GH.)

Glandular-puberulent to glabrate ramifying annuals, prostrate and spreading 01
erect, the internodes mostly longer than the leaves, glabrous to densely glandular
often rooting at the nodes. Leaves glabrous to scantily puberulent, orbicular tc
reniform, apically rounded and occasionally mucronulate, basally rounded t<
cordate, 5-25 mm. long, 5-30 mm. broad, the petioles 2-15 mm. long; stipule

stly lacerate, rather persis

Inflorescences (

252

ANNALS OF THE MISSOURI BOTANICAL GARDEN

axillary few-many-flowered dichasial cymes, (flowers rarely solitary in the axils)
the bracteate pedicels locally girdled with a dense band of glandular pubescence,
rarely subglabrous, 2-15 mm. long, much exceeding the subtending bracts. Sepals
5, lanceolate to ovate, acute, glandular-puberulent to glabrous, obscurely to strongly
3-nerved, 2.5-4.0 (-5.0) mm. long, the borders translucent; petals 5, 2-3 mm.
long, deeply bifid, the lobes linear, acute, rarely obtuse, 1 -nerved, basally exauricu-
late but rarely subdentate; stamens 2-3 (-5), the flattened filaments 2.0-2.5 mm.
long, the anthers suborbicular, 0.2-0.3 mm. long; style 0.5-1.0 mm. long, bifid or

DUKE REVISION OF DRYMARIA 253

trifid from half to its entire length; ovules few to many, campylotropous on free
central placentae. Capsule ovoid, 1.5-2.5 mm. long, the 3 valves entire; seeds
1-many, 1.0-1.5 mm. broad, cochleate, dark reddish brown, tuberculate in lines, all
the tubercles low and domical, closely approximated or contiguous. Fig. 18B
(Holotype of var. pacifica) ; Fig. 18C (Holotype of D. procumbent).

A nearly pantropical species, in America ranging from Florida and Mexico
through the West Indies and Central America to Argentina.

Mizushima (in Jour. Jap. Bot. 32:69. 1957), studying the varietal potentials of
this polymorphic species, described var. pacifica, which differs from typical cordata
in having glabrous sepals, more seeds, and a tendency to have a greater number of
stamens and the styles divided completely to the ovary. This rather striking variety
seems to be lacking in continental North America. Occuring in both var. pacifica
and var. puberula, for those who wish to recognize them, are variants in which the
flowers are subsolitary in the axils (e.g. Hoi ton 733 from Colombia). Surprisingly
and fittingly these have as yet received no taxonomic status. Drymaria adcnophora
Urban represents a more or less viscid form with diminutive leaves and flowers
and apparently is very similar to specimen 109-1 in the Linnaean herbarium. A
similar reduction is exhibited by the type of Holosteum Ji.in.hu>-: Sw. generously
loaned by Stockholm. Mosen 40 5 at Lund from Brazil with the sepals to as much
as 5.2 mm. long represents the opposite extreme of D. adcnophora, which I believe
is typical D. cordata.

Perhaps the strangest aberration is that represented by the flowers of the type
of D. procumbens, here considered to represent only a monstrosity. The petals are
bilobed but one of the lobes is clearly staminal in nature (see Fig. 18C), the other
lobe being normal and petaloid. I have seen a similar aberration in one specimen
of D. paposana var. weberbaueri, perhaps lending weight to Pax's interpretation of
the petals as staminodia.

48b. Drymaria cordata ssp. diandra (Blume) J. Duke comb. & stat. nov.
Drymaria diandra Blume, Bidjr. Fl. Nederl. Ind. 62. 1825. (lectotype: after Mizu-

Drymaria retusa Wallich, ex Wight & Arnott, Prodr. Fl. Ind. Or. 3 59. 1834. as synonym.

(? type?: Wight 152, G, L, !)
Drymaria cordata forma indica Miquel, in Junghuhn, Pi. Jungh. 395. 1855.
Drymaria externa Wallich, ex Edgeworth, & Hook. f. in Hook. f. Fl. Brit. Ind. 1 2 :244.

1874. (?type?: Wallich 647, Gl)
Drymaria gerontogea F. Muller, Descr. Papuan PL 1:87. 1877.

Prostrate or ascending annuals, the internodes longer than the leaves, glabrous
to stipitate-glandular. Leaves opposite, the blades glabrous, rarely minutely
puberulent, deltoid-ovate to orbicular or subreniform, apically rounded and often
apiculate, marginally entire, basally truncate to obtuse, rarely cordate, 5-25 mm.
long, 3-20 mm. broad, 3-7-nerved; the petioles 2-8 mm. long usually exceeding the
stipules; the stipules lacerate, the segments filiform, 1-3 mm. long. Inflorescence
a terminal lax to dense 3-many-flowered cyme, the peduncles 1-12 cm. long,
glabrous or glandular-puberulent; bracts lanceolate, 2-5 mm. long; pedicels evenly
glandular-puberulent; 1-8 mm. long. Sepals 5, narrowly obovate, 2.0-4.5 mm.
long, the outer glandular-puberulent, strongly carinate, strongly three-ribbed, the

254 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ribs protrusive and often connected by cross- veins; petals 3-5, 1.5-3.0 mm. long,
bifid one-half to two-thirds their length, the lobes oblong, 1 -nerved, apically obtuse
to emarginate, basally tapered to the linear claw. Stamens 2-3 (-4), 1.6-2.2 mm.
long, the anthers suborbicular, the filaments basally connate into a shallow cup
(the cup often carnose in African forms). Ovary at anthesis globose, the styles
short, bifid or trifid from half to all their length. Capsule 2-3-valved, 1.5-2.5
mm. long, 1-2 -seeded, the seeds cochleate, 1.4-2.0 mm. broad, rather densely
tuberculate. Fig. 18D (type of D. retusa ?) ; Fig. 18E (lectotype of D. diandra).

Tropical Africa, Australia, and Tropical Asia (China, S. & W.; India, Indo-
china, Malaysia, Formosa, Ryukyu Archipelago), Oceania, Hawaii, Philippines, New
Guinea, Celebes.

Mizushima (in Journ. Jap. Bot. 32:69. 1957), after a detailed study of the
variation of D. cordata, reinstated D. diandra, long lost in synonymy under D.
cordata, to specific status. Intermediates do occur between the two, especially in
Africa. The absence of the taxon from the New World when weighed with the
characters enumerated in the key, none of them invariably constant, suggest that
reduction to a subspecies more truly reflects the hierarchial relationships.

African forms are very perplexing and differ from the typical material in the
large sepals and petals, higher number of stamens, with prominent, often carnose,
staminodial cups. The flowers often lack the pyriform shape which so frequently
differentiates ssp. diandra from ssp. cordata. Description of these variants as formal

Species Excludendae
D. filiformis Benth. = Stellaria filiformis (Benth.) Mattf.
D. idiopoda S. F. Blake = Stellaria sp.
D. mairei Leville rz Cerastium sp.
D. palmeri Hemsl. = Stellaria sp.
D. rotundifolia Harriot = Stellaria rotundifolia Poir. fide Kew Index

Species non Satis. Notae
D. adiantoides Muschl. (type: Sodiro I2QI) ex char. = D. villosa}
D. grandi flora Schlecht. non Bartl. May be Stellaria}
D. megalantha Steud. " "

D. oxalidea Pax (type: Schaffner 827) ex char. veg. = D. excisa}}}
D. peruviana Muschl. (type: Weberbatter 2601) ex char. veg. — D. divaricata?
D. sessili flora Fiori (Descr. non viso) ex geogr. — D. cordata

COMBINATIONES ET TaXA NOVAE

D. holosteoides var. crassi folia (Benth.) J. Duke

D. arenarioides ssp. peninsularis (S. F. Blake) J. Duke

D. barkleyi J. Duke & Steyermark

D.conzatti J. Duke

D. effma var. confusa (J. N. Rose) J. Duke

DUKE REVISION OF DRYMARIA

D. effusa var. depressa (E. L. Greene) J. Duke

D. leptopbylla var. cognata (S. F. Blake) J. Duke

D. leptopbylla var. nodosa (Engelm.) J. Duke

D. engleriana var. devia (Baehni & MacBride) J. Duke

D. villosa ssp. palustris (Cham. & Schlecht.) J. Duke

D. villosa ssp. palustrh forma tounscndii (Robins.) J. Duke

D. villosa ssp. pahistris var. perennis J. Duke

D. i7'//ostf ssp. t'/7/ostf forma tepicana (M. E. Jones) J. Duke

D. villosa ssp. paramorum (S. F. Blake) J. Duke

D. paposana var. serrulata J. Duke

D. paposana var. weberbaueri (Muschl.) J. Duke

D. rotundi folia var. nitida (J. Ball) J. Duke

D. divaricata var. s/r/c/a (Rusby) J. Duke

D. divaricata var. viscidula (A. Gray) J. Duke

D. divaricata var. divergens J. Duke

D. divaricata var. reflexiflora J. Duke

D. gracilis ssp. carinata (Brandegee) J. Duke

D. glandulosa var. galeottiana (Briq.) J. Duke

D. con/*/* ss£. <//Wr«x (Blume) J. Duke

Citation of Specimens

Fredholm 6487

Godfrey, R. K.

McFarlin, J. B.

Reese, W. D. 1!

Schallert, P. O.

7 50

Small, J. K. 127

Small, Britton &

Dc\\ i:

nkeler

Small, Mosier &

l\.A\ii,

keler

Wright 869 (type)

D. effusa var. depres
J. Duke
Gould & Robinson 49
Metcalfe 1428 (aber.
Neally 251

D. glandulosa Presl

Barkley 14NM715
Earle & Earle 460
Eggleston 163 81; 16^

Cory 30590
Hinckley 321, 3 57,
Innes & Moon 1136
Moore & Steyermark
Warnock W812

Cory 30375
Darrow 3 517

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Harrison & Kean
Hinckley 1290
Metcalfe 527 (a

Standley 13932
Wright, C. 868

Harrison, Kearney & Hast

ings6228;

Kearney & Peebles 10042

; 12723

Lemmon 3002

Metcalfe 349; 1424

Mulford 806

Parker & McClintock 75

Purpus 8148

Standley 14212

Wright 867

Woot. & Standi.

Gould 3962
Gould & Haskell *
Greene 12600

,M '

Hershey 3012
Hinckley 3522

Mexico (Continental)

1. ANOMALA S. Wa

tson

D. CORDATA (L.)

Bell & Duke 1658
Johnston & Muller

Pringle 2847 (tyi

622A

•E); 3278

Berlandier 1217
Ervendberg 22!
Hinton 7442 (
Karwinsky 29

Lundell 53 51
Moore, HE. 15
Muller, C. H. 3
Parry & Palmer

Thurber 815
White, S. S. 2
Wiggins 7235

Barkley, Rowell & >

i& Barkley 16M699

). barkleyi J. Duke & Steyermark
Hernandez, Rowell & Barkley 16M53 1
Warnock & Barkley 148 26M (type)

DUKE REVISION OF

D.

\UVa
Mom.

ugh WIS^IOMT? 7

idulosa Presl

: !-v.

ar 1782

e 513; 595; 2295 (a

Ikrh>

idier 677; 841; 1122
eau 283; 552; 592a

Fisher 44293
Johnston 44293

Lloyd 370
Lundell 5095

Pringle 5089; 7122

Rose & Hay' 5511

Rose & Painter 6617; 6775; 708

var. galeottiana) 7755
Rose, Standley & Russell 12697
:Barkley 17M-

itrncr

332b

Sesse, Mocino e

Wynd & Mueller

Fenz

iphylla (Cham.
1. ex Rohrb.

Barlow 30
Coulter 697
Gentry 8474
McVaugh 131

05

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Woot. & Standi.

J. Duke
Gentry 571M; 18

gentryi)
Hartman 13 8
Knobloch 5522; 5i
LeSueur 258; 623
Muller 3426
Pringle 581; 716;

Schaffner 5
Townsend & Barbe
Tucker 2457

D. stipitata Fosberg

Johnston & Muller 241

Marsh 1571

Stewart 567 (type: va

Hinton 15505 (?)

Bourgeau 553

Leavenworth 595 (?)

Kenoyer A-457
McVaugh 16795; 177

>. molluginea (Lag.) Didr.

Pringle 2120 (type);

Arsene 1616

var. jaliscana) 7648

; 11928

Griffiths 6749

Scharlner 363; 416

Mearns 2248

Seaton 267 (type: D

. filiformis)

Pringle 1192; 1571; 6481

Sesse, Mocino et al. 44

8, 466; 467;

5184

Sesse, Mocino et al. 984

D. villosa Cham. & Schlecht

». mui^exora Brandegee

A f 8 To 5129 p ' p,; '

1214; p. p.;

5281;

Arsene 5309; 6710

Bell & Duke 16648

Botteri 350; 944

Hinton 1549; 2596; 6578;

12167; 12489; 15050; 15263

McVaugh 12962; 13441; 13442;

14066; Fischer 2092

16524

Galeotti 1726; 4414

McVaugh &Koelz 1264

Hinton 1384; 5152; 5

Jones, M. E. 22847

(type: D.

Purpus 1653 (type)

: D. barrana

Rose & Painter 6517; 6773 p. p.;

6774; Leavenworth & Hoogstraal 1295

7219

Matuda 4752

Miiller 120; 575a

DUKE REVISION OF DRYMARIA

Parry & Palmer 56
Pringle 1501; 7296
Rose & Hay 4800

Townscnd & Barber 23

Rose, Standley & Russell 14145 & 143 86

townsendii)

(both f. tepicana)

Schiede & Deppe 505 (type)

J. Duke

Sezaya, Paxon & Barkley 16M833

Edwards 804B

Sumichrast 1568; 1945

VILLOSA SS

Kenoyer 2449 (appr. laxiftora)
Palmer 187 (type); 356

(Cham. &' Schlecht.) J. Duke

Arsene 16; 5191; 5391; 5416; 5911

Pringle 3265

Balls B4568;B5 504

Barkley et al. 7489

D. xerophylla A. Gray

Frye & Frye 2606

Arsene 46

Coulter 722 (type)

McVaugh 16793

Jones, M. E. 90 (type: D. svBsessilis)

LeSueur 1322 (f. toivtisendii)

Pringle 6910

Leavenworth 525

Maysilles 7782

Roseetal. 9679

McVaugh 13415; 16159

Schaffner 525

Moore 2801; 2955

Sesse, Mocirio et al. 5003

Baja California

. ARENARIOIDES SSp. PENINSULARIS (S. F.

D. gracilis ssp. carinata (Brandegee) J.

Blake) J. Duke

Duke

Carter et al. 2257

Brandegee 34 (type)

Craig 73 8

Jones, M. E. 24193

Howell 10576

Johnston 3972 (type: D. johnstonii)

Purpus 337; 423 (type); 496

Rose 16357

Bryant 1888 (type: D. veatchii)

. debilis Brandegee

Carter et al. 1935; 2474

Brandegee 3 5 (type: D. polystachya) ;

Carter & Kellogg 3089; 3221

s.n. Feb. 13, 1889 UC (type)

Chambers 837

Constance 3135

Carter & Ferris 3709; 3812; 4093

Carter & Kellogg 2922

Hinds s.n. (type)

ixney & Kempton 157; 211

Johnston 3048; 3949; 4263

Dawson 1095

Jones, M. E. 24076

Gentry 3751

Mason 1908

Palmer 4; 819 (type: D. diffusa)

Palmer 570

Reed 6156; 6255

Rose 16268; 16528

. GLANDULOSA Presl

Shreve 7053; 7095

Carteret al. 2379; 2447

Wiggins 5456; 5531; 7898; 11350

Carter & Ferris 3774; 3907

Jones, M. E. 24149; 24190; 24191;

D. holosteoides var. crassifolia (Benth.)

27050 (type: D. fendleri var. perennis)

Purpus 469

Carter et al. 2255

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Nelson 3706

Stanley 58398; 59961; 61978; 65477;

67942; 68554; 80175; 80683; 83275;

83294; 83823; 85147; 85599; 85827;

17158; 17541; 19236; 20236; 21557;
22502; 22650; 22714; 23493; 24022;
44268; 53336; 66994; 72489; 75828;
76914; 87923; 88760; 89508
Standley & Elias 2542; 2687

). laxiflora Benth.
Bernoulli & Cario 3 255
Hartweg 523 (type)

Steyermark 50243 (

lernoulli 240 (type: D. leptoclados)
tandley 58053; 59111; 59852; 60343;

60837; 60874; 60893; 61235; 62818;

62897; 63790; 64317 p. p.; 64427;

66095; 66170; 67287; 77390; 80085;

83193'; 83247^ 83512^ 83884J 84704;

3639; 3651; 3652

DUKE REVISION OF DRYMARIA 261

Standley 1897; 11000; 11873; 12155; Williams & Molina 10076; 12016

12330; 13284; 13324; 15488; 21136 Woodson, Allen & Seibert 1162; 1771

p. p.; 20257; 21730; 21786; 21827; p.p.

22048; 22194; 22651; 22713; 22715; Woodson & Sc he ry 710

23256; 24551; 25023; 29145; 56318; Yuncker, Dawson & Youse 5856

76602; 78213,' 80809; 8105o| 8224o| Dl vlLLOSA SS P- pa lustris

83258; 91341 (Cham. & Schlecht.) J. Duke

Standley, Allen et al. 919 Muenscher 12428

Standley & Elias 2533 skutch 298

Standley & Jaime 6748; 6926 p. p. Standley 61036; 61054; 61553; 62663;

Standley & Lindelie 7469 62855; 62969- 64316- 64697; 64709;

Standley, Williams et al. 1225 659 i 5; 678 i 6 ; 82677 j 83746; 85203;

Steyermark 29615; 29910; 35044 p. p.; 85816; 87192

50528 Standley & Elias 3170; 3649

Tonduz 749 Standley & Jaime 5503

von Tuerckheim 684 (appr. D. malachi- Steyermark 34454; 3 5044 p. p.; 3 5597;

Jack 6254; 8066
Leon & Clement 64^
Leonard 8347; 9217
Leonard & Leonard
Lloyd 405; 455

Brown 111

Clement 40!

Cowell 776

Duss 1779;

3092

Eggers 436b

} 5 3 s

4972.

Proctor 17087

Rose, Fitch & Rt
Shafer 53; 208;

Stahl 779
Taylor 202
Thompson

Heller 4577
Heller & Heller 175;
Hodge 438; 439; 440
Hodge & Hodge 1519

Wright 24
Wright, Parry &

OF THE MISSOURI BOTANICAL GARDEN

D. firmula Steyermark

irk 56536 (type)

Bonpland 222

St<

Broadway 575

Farenholtz 554; 665

(var. pacific*]

Fa

Pit

D. v:

Pittier 9458; 9827;

Vogl 253; 324

J*:

Pil

Colombia

Cuatrecasas, Schultes & Smith
Fosberg 19276; 19491, 20367
Gutierrez & Barkley 17C529

Lehmann 5997 i
Pennell 1563
Pring 63

Schiefei
Toro 15

Rusby & Pennell 622
). villosa Cham. & Schlecht.
Andre 4289 (appr. ssp. palust
Arbelaez & Cuatrecasas 5260
Killip & Smith 16142 p. p.; IS

Apollinaire 24
). ovata Humb.

Rusby & Pennell 644

). VILLOSA SSp. PARAMOl

(S. F. Blake) J. Du

70 (var. pacijica)
Bonpl. ex Roem.

appr. D. glaberrima)

s Humb. & Bonpl. ex Roem.

22888; 23225 (appr. D.

Rose, Pachano & Rose
Rose & Rose 22227;

:Rose 22952; 23183 f
. & Schlecht.

Camp E-1929; E-2624
Fosberg & Prieto 22689
Rose, Pachano & Rose 2;

OF DRYMARIA

Howell 9158; 9671

Svenson 65 (type:
(var. pacifica)

MacBride & Featherstc

MacBride 391
Nunez, O. V.
Weberbauer 6

(Bachni & Mac-

Bride) J. Duke
MacBride & Fcathcrstone
Pennell 13462; 14655 ("n
Tovar 1277B

Ferreyra 7422
MacBride & Featherstoru
glandulosa) ; 693

Haenke s. n. (type

). DIVARICATA Var. DIVERGENS J. Duke

MacBride 3 506

Asplund, E. 13731

Woytkowski 34228

(???)

SSTaSW*

D. GRANDIELORA Bart

►. DIVARICATA Var. REFLEXIFLORA J. Duke

Ferreyra 6245; 85

B7 (appr. D. glaber

Ferreyra 3855 p. p.; 3958; 8695a; 11522
Nunez 2261; 3579; 3754; 3752; 3755;
3757; 3758; 4198; 4291

Pennell 14724
Vargas 1903

>. divaricata var. STRicTA(Rusby)J. Duke

D * Thu A it.? umb " &

Bonpl. ex Roem. 8

Ferreyra 763 (appr. D. glaberrima) ;

Espinosa 199
MacBride & Feather:

1714

Lechler 1947 (type: D. agapatensis)

Weberbauer 5327 (appr. D. glaberrima)

D. PAPOSANA Phil.

i. divaricata var. viscidula (A. Gray)

J. Duke

Ferreyra 6102

D. paposana var. tb

3ERBAUERI (Muschl.]

MacBride & Featherstone 189?

Pennell 14518, 14516

Stork 11456

8025; 8695;'9580

5; 3855; 6328 p. p.

Wilkes Exped. s. n. (type)

Grant 7441, 7468

ANNALS OF THE MISSOURI

Weberbauer 1657
Worth & Morrison
paposana)

closely resembles Ball ;

Balls, E. K. 7021.B7021
Goodspeed, Stork & Horton I
KilHp& Smith 22132; 22134
MacBride & Featherstone

praecox Baehni & MacBride

Duke
Ball, J. s.n. in 1882 (TO
Balls, E. K. 6789
Cerrate 150; 627; 1376

MacBric
Poeppig 157!
Vargas 2100

lam. & Schlecht.
, (type: D. hirst*
th 23212

Featherstone 154

Bang 233

X glandulosa Presl
Bang 92 (appr. D. cordata) :

Cardenas 609a; 2424
Eyerdam 24941; 24957
Fiebrig 2837
Hammarlund 404

). ovata Humb. & Bonpl.
Schult.
Cardenas 2332; 3379
Herzog 2390

Duke
Mandon 948 (appr. D. ovata)
). villosa Cham. & Schlecht.
Buchtien 183 p. p.

-REVISION OF DRYMARIA

Burchell 1011; 9711

Lorentz & Hieronymu:

Parodi 9080

Venturi 5104; (var.

de Samp
Segadas-

Argentina

Parodi 803!

Bonpl. ex Roem.

). cordata (L.) Willd.

pacifica); 8642 (appr. D. diandra)
Bakhuizen 28a (var. pacifica)
Buchtien 695, 795 (var. pacifica)
Bunnemeyer 11072
Cid 20 (var. pacifica)

> Oceania

Edano 1
Flecker J
Heller 2i

ANNALS OF THE

BOTANICAL GARDEN

Nodi&Idjan 253

Posthumus 3459 (var. pacifica)

Merrill 992; 4341

Purseglove 4322 (var. pacifica)

>56; 1958; 1960; 1961 (var.

Oldham 23

pacifica)

Peekel 90

St. John 9938

Poilane P-M-21; 156C

StJohn&Catto 17850

Pulle 1148

Steiner 34787

Ramos 3 0327

Topping 2999

Robinson 18233

VanRoyen 4312

Womersley & Van Royen 5911 (var.

Rock 6419

pacifica)

Savinierre 283

Womersley, Hoogland & Taylor 4909

Stewart 14437

(var. pacifica)

Tanaka 5498

Thwaites 1091

>. cordata ssp. diandra (Blume) J. Duke

Tsai 62067

Tsang 21658,25780

Tsang, Tang & Fung

Bloembergen 3371

Tsiang 9221

Boeea 8482; 8653; 8921

Zollinger 892
). villosa Cham. & Schlecht.
Backer 5240; 5619; 10665; 12940 (t

Griffith 339

Jacquemont 1523
Kajewski 2176
Kalkman 4526
Kanehira & Hatusim
Kaudern 168

Schiffner 1952; 1957
Van Ooststrom 13 059:
VanSteenis 10916; 11

DUKE REVISION OF DRYMARIA

1236 (appr. ssp.

Systematic Index

xepted, preexisting names; ital

diandra MacFayden

Drymaria adompbora

251

axillaris

harklevi

21 3, 215

var. confusa .....2

!lX™ _ !

fendhrt . ....

fenzliana

conzattii

224 ^25

filiformis Benth.

filiformis Robins. & Seaton

frankenioides ..

var -V vracilh

245 246

frutescens . .

f. indica

j
var. pilosa _
/3 puberula

ANNALS OF THE MISSOURI BOTANICAL GARDEN

grandiflora Bartl.
grandiflora Schlech

coahuilana

232, 23 5 ssp.

"

n> l U " dlf nWda' G " y

-—-237,

spbagnopbila

stellarioides

......212,

stl pu« xsnpul '

.^..213,
193,

K 'E,"3™I::::: : '"':

:z 2 io;

Vi ""p'' pal U „ris "

225,

f. townsendii ...228,

■

Pinosia ortegioides

Polycarpon englerianvm

Stellaria adenopbora

filiformis

SANANGO: NEW AMAZONIAN GENUS OF LOGANIACEAE
GEORGE S. BUNTING and JAMES A. DUKE

.orge S. Bunting.
James A. Duke, Mis

Among recent collections of Felix Woytkowski from J
men of a plant that appears to be intermediate between the families Loganiaceae and
Scrophulariaceae! It is a fine example of the situation which has led to the wide-
spread view that the Loganiaceae are an artificial group constituted primarily of
genera excluded for one reason or another from other well-marked, natural families.
There is still ample cause to argue this case, for there is no obvious "family
resemblance" to guide us, and often the circumscription of the family is inadequate
to positively include or exclude an unknown plant.

A century ago George Bentham (Jour. Linn. Soc. Bot. 1:52-114. 1854) wrote:
"There can be no doubt that Rubiaceae, Apocyneae, Gentianeae, and Scrophulari-
neae are large independent orders indicated in nature, yet those genera now amalga-
mated under the name of Loganiaceae bind them so firmly together, that some of
these genera will be found even more closely allied to certain others of each of the
above orders respectively than they are to each other." And he continues: "The
Loganiaceae lie very near to some part or other of the vast family of Rubiaceae, but
by their free ovary, they are absolutely, and by very few exceptions, clearly
separated. The connecting genera with the Apocyneae, Gentianeae, and Scrophu-
larineae are on the other hand, much fewer, but their union is much closer. With
Scrophularineae in particular, although the general affinity is more remote, the few
intermediate genera and species are intermediate in every respect, in habit as in
technical characters. The main distinction, the presence of stipules in Loganiaceae,
disappears very gradually, and the difficulty of drawing the line is the greater from
there being no habit or family resemblance to unite the several members of the
Loganiaceae."

It is in this paper that Bentham discussed the similarity of the genus BuJdleja
(previously included in Scrophulariaceae) to Logania, and concluded the necessity to
place them together, along with the genus Polypremum, in the family Loganiaceae.

Now we have discovered another plant that seems to stand nearly between the
Loganiaceae and Scrophulariaceae. The fruiting specimen has a distinct aspect of
Buddleja, but several technical features prevent its inclusion in that genus. It is
equally, if remotely, allied to the genus Peltantbera. After thorough study of this
plant, and consultation with N. Y. Sandwith at Kew, and J. J. Wurdack at Wash-
ington, we are convinced that it represents an unnamed genus, and describe it

270 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Sanango Bunting et Duke, gen. nov.

Arbores parvae ligno valde duro, f oliis oppositis petiolatis binis lineis c
Flores in thyrsis bracteolatis, calyce ad medium lobato, segmentis lanceolatis sub-
aequalibus, corolla 5-lobata, ± hypocrateriformi, gibbosa, segmentis rotundatis
subaequalibus, aestivatione cochleari, staminibus 4 (!) sub sinibus prope tubi basim
affixis, filamentis inflatis, antheris subrotundis singulatim 2-thecas ferentibus,
introrsis versatilibus in linea ± hippocrepiformi aperientibus, staminodio parvulo,
ovario globoso leviter 4-lobato biloculari disco cupuliformi circumdato placentis
axillaribus multiovulatis, stylo brevissimo, stigmate capitato bilobato. Capsula
polysperma septicidalis, utroque carpello iterum scisso.

Sanango durum Bunting et Duke, sp. nov.

Arbor 16 m. alta ramulis lenticellatis. Foliorum lamina 10-25 cm. longa 4-7
cm. lata elliptica vel anguste obovata ad apicem acuminata, ad basim cuneata vel
attenuata, parce et inaequaliter denticulata praecipue in dimidio superiore, venis
lateralibus utrinque 5-7 arcuatis ascendentibus, infra pallida venis strigulosis
instructa; petiolus 0.5-1.0 cm. longus, ad basim latus, facie adaxiali velutina
ferruginea. Inflorescentia 5-12 cm. longa terminalis vel axillaris ± secunda,
pedunculo et pedicellis strigulosis, pedicello ca. 1 mm. longo; florum calyces ca.
3.5 mm. longi puberuli lobis 5 acutis 1.5-2 mm. longis; corollae albae aliquantum
carnosae, ca. 1.2 cm. longae extus puberulae et glanduloso-puberulae, intus
glanduloso-puberulae et etiam pilosae in tubi basi, lobis ca. 3-4 mm. longis; fila-
mentorum staminalium partes liberae pilulosae vix 2.5 mm. longae antheris ca. 0.5
mm. latis brevioribus quam latis; staminodium parvulum vix 0.2 mm. longum;
ovarium globosum ca. 1 mm. longum, disco cupuliformi tarn procero atque
ovario; stylus crassus persistens, una cum stigmate magno ca. 1 mm. longus.
Capsula ellipsoidea vel anguste obovoidea ca. 3.5-4.0 mm. lata parietibus osseis;
semina orthotropa ± elongata fusiformia ca. 0.7 mm. longa recta vel sigmoidea.

Type collection: Felix Woytkowski 5619, "tree 12 m. tall; no latex; flowers
white; in forest, elev. 300 m.; Aramango, s. of Nazaret," depto. Amazonas, Peru;
2 April 1960 (holotype MO, isotype K, US).

Distribution: Forests between 300 and 750 m. elevation, west central Ama-
zonas and central Junin, Peru. These localities are separated by nearly 500 air
miles. It seems plausible that the distribution may be more or less continuous
between these points.

Peru: Junin: Satipo, Aug. 1960, Woytkowski 5925 (MO).

Vernacular name: "Sanango" (Amazonas).

The generic name Sanango is the colloquial name furnished by the collector of
this plant. It is a corruption of a Quechuan word, "sanangu", an Indian word
applied also to species of Tabernaemontana, Rautvolfia, and Malpighia. The specific
epithet refers to the hard wood to which Woytkowski called attention in his field
blade when one attempts to cut

272 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Sanango belongs to the subfamily Buddlejoideae. Its nearest relatives appear
to be Buddleja, Nuxia, and Peltantbera. It should be noted that Sanango has been
collected only in localities that lie in the middle of the known range of Peltantbera. 1
Although the affinity is not strong, these latter two genera share the following
characteristics: hippocrepiform dehiscence of the anthers, a disk between the ovary
and corolla, multicellular hairs, and a lack of stipules.

The combination of characters occurring in Sanango is unusual in Loganiaceae,
but most of them occur singly in some other genus of the family. The hippo-
crepiform mode of dehiscence of the stamens occurs also in the African genus
Nuxia.

There is a very well developed annular disk of firm texture in the flower of
Sanango. In Peltantbera, a very short disk is present, conspicuous only in young
flowers. In Nuxia, however, the disk is annular, and though membranous, very
similar in form to that in Sanango. 2 Notwithstanding, the style and stigma of
each of these three genera are extremely different.

An approach to the Sanango-type pistil occurs in Buddleja, and in the herba-
ceous genera Mitreola and Polypremum. In Mitreola, the style is short, the stigma
smaller than in Sanango, and the ovary splits in two at the top as it develops. Even
more similar is the fruit of Polypremum which splits into four segments at
maturity as it does also in Buddleja.

Multicellular hairs similar to the dendroid or branched type in Peltantbera have
not been observed in Sanango, nor have we encountered the glandular, peltate type
similar to those described by Solereder (Pflanzenfam. 4 (2) :23, fig. 12. 1892) and
by Hunziker and Fulvio (op. cit. in footnote). Certain types of multicellular
hairs are, however, common in Sanango on both vegetative and floral parts.

The corolla is 5-lobed in some genera of Loganiaceae and 4-lobed in others.
Indeed both conditions are known to exist in a single species of Buddleja. The
loganiaceous flower typically has an equal number of stamens and corolla lobes.
Sanango is unique 3 in the Loganiaceae in possessing a five-lobed corolla with only
four stamens, the fifth one merely suggested by a very small appendage not apparent
in all flowers. This characteristic is typical of such families as Verbenaceae
and Scrophulariaceae. We feel, nonetheless, that Sanango is more closely
allied to Loganiaceae than to any other. Such an opinion is dictated by the
presence of a stipular line between the leaf bases of each leaf pair, and the strong
similarity of the pistil and the fruits to Buddleja, which clearly is related to
Logania. The more or less irregular nature of the corolla in aestivation, and the
reduction of one stamen in Sanango decidedly suggest Scrophulariaceae. There is,

llent paper on Peltantbera, Hunziker and Fulvio (Bol. i

noted that the genus is known from Costa Rica, Pert

udy we have encountered the collection of A. E. Lawr

depto. Boyaca, Colombia) which clearly represents P. ft

aper. The Colombian specimen agrees in all its floral m<

: half the length of the ovary as is characteristic of the B

f Nuxia ftoribunda Benth. is deeply two-lobed, which,

;estive of Rauwolfia and other Apocynaceae.

in genus Usteria, three of the four stamens regularly aboi

Aestivation of the flowers of this genus is cochlear, i.e., one corolla lobe is
interior, one exterior, and the other three have one edge overlapping and exposed
with the opposite edge hidden by the adjacent overlapping lobe (Fig. if). The
corolla tube is gibbous near the base, and the flower appears to be oriented so that
the interior corolla lobe and the gibbous side of the tube are uppermost (adaxial).
The vestigial fifth stamen arises on the gibbous side, below one of the two sinuses
adjacent to the interior corolla lobe (Fig. lc). The bud is distinctly zygomorphic,
more so than in Peltanthera. It is impossible to ascertain from the dried material
to what degree the open flower may show this character.

We are unaware of an identical type of aestivation among the other Loganiaceae,
though it is not uncommon among the dicots. The usual pattern in five-lobed
corollas of this family is with two lobes external, two internal, and one imbricate.

Vesture

At least four types of multicellular hairs occur in Sanango. The largest are
present on the adaxial face of the basal part of the petiole. These are composed of
3-5 elongate, hyaline cells, the basal one inflated, the terminal one attenuate (Plate
IVa) . Very young stems and leaves are clothed with hairs of a similar type, a few
of which persist on the veins on the lower surface of the mature leaves.

The calyx bears hairs of similar structure, but very much smaller than those
described above. In some, however, it appears that the terminal cell is slightly
inflated (Plate IVb) .

Both surfaces of the corolla are very hairy (Plate IVc). The outside is closely
covered by very small, mostly 3 -celled attenuate hairs similar in form to those oi
the leaves, and intermixed with these are some still shorter gland- tipped hairs (Plate
IVe). The glandular tip cell is opaque, borne upon a greatly inflated, hyaline
basal cell.

The inside of the corolla bears 3-4-celled, gland-tipped hairs about twice as
long as those on the outside of the corolla (Plate IVc). [The glandular tip may be
composed of several cells arranged radially, but the small size of the structure has
prevented us from ascertaining its exact nature. Some lobing is apparent in certain
preparations.] In addition, there are many very long, slender hairs in the throat
of the corolla that appear to have no cross walls dividing them into cells (Plate
TVd).

A pollen sample was sent to Dr. G. Erdtman for study. He reports that "the
grains are small, uninteresting from a morphological point of view, and similar to
those in Peltanthera ..." His diagnosis follows: "Pollen grains small, 3-coIporate
(about 15 X 12.5). Colpi about 12 fx long. Ora about 2.5-3 n high. Exine thin.
Sexine of about the same thickness as nexine. [No pattern or traces of bacula
etc. could be seen even by means of an apochromatic objective with the numerical

274 ANNALS OF THE MISSOURI BOTANICAL GARDEN

aperture 1.40. The pollen grains in Peltanthera costaricensh are equally 3-colporate
and measure about 15 X H /*; tne exme seems to be very minutely reticulate.]"
Anatomical Considerations
Dr. Stern, U. S. National Museum, plans to do a detailed anatomical study of
the wood of this genus. We wish, however, to point out that there appears to be
no included phloem in this species, which is the typical condition in the
Buddlejoideae. This contrasts strongly with the intraxylary phloem present in
members of the Loganioideae (Metcalfe and Chalk 2:931. 1950). The wood does
not appear to differ strikingly from that of Veltanthera.

Explanation of Plate
PLATE IV

Types of trichomes in Sanango
a. from adaxial face of petiole b. on calyx c. on corolla, exterior (left) ;
(right) [a-c all at same scale] d. enlargement of long hair on interior oi
single short gland-tapped one at its base (note the apparent lobing of the glai
e. enlargement of a single gland-tipped hair on exterior of corolla

i

MISSOURI BOTANICAL GARDEN

Calavay H. Dodson,
Robert L. Dressier,

BOARD OF TRUSTEES

Volume XL VIII

Annals

of the

Missouri Botanical
Garden

NOVEMBER, 1961

Two New Me
Portraits of PI

Annals

of the

Missouri Botanical Garden

School of Botany of

Annals

of the

Missouri Botanical Garden

Vol. XL VIII NOVEMBER, 1961

RidgMHiaou. SU '

Introduction

The genus Scolecopteris, founded by Zenker in 1837, is based upon silicified
fructifications from the Lower Permian of Saxony. To date the following species
have been described:
S. elegans Zenker, Linnaea 11:509. 1837. (A more detailed description was given

by Strassburger in 1874.)
S. subelegans Grand'Eury, Acad. Sci. Inst. France, Mem. 24:72-73. 1877.
S. ripageriensis Grand'Eury, ibid., p. 73.
S. minor Hoskins, Bot. Gaz. 82:427-436. 1926.
S. oliveri Scott, Journ. Linnean Soc. London, Botany, 49:1-12. 1932, and pp. 309-

321. 1933.
S. tati folia Graham, Bot. Gaz. 95:456-458. 1934. Emend Mamay, Ann. Missouri

Bot. Gard. 37:425-428. 1950.
S. radforthii Andrews, Ann. Missouri Bot. Gard. 30:435-437. 1947.
S. iowensis Mamay, Ann. Missouri Bot. Gard. 37:429-431. 1950.
S. minor Hoskins, var. parvifolia Mamay, ibid., pp. 431-432.
S. incisifolia Mamay, ibid., pp. 433-436.
S. major Mamay, ibid., pp. 436-439.

Scolecopteris is characterized by pecopterid pinnules bearing sporangia in sori
on the abaxial surfaces. The sporangia composing a sorus are attached to a common
pedicel, but are free above, and lack a common synangial sheath.

The following account deals with two additional species that fit these charac-
teristics and thus have been assigned to the genus Scolecopteris. Further evidence
for placing them in this genus will be recorded under the description of each

276 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Materials

Both species are based on fertile pinnules found in coal balls that were collected
by Dr. Henry N. Andrews from a site two miles northeast of Berryville, Illinois.
The outcrop lies in the upper part of the McLeansboro group of the Pennsylvanian
of Illinois.

The Carboniferous Marattiaceae

Fructifications referable to the Marattiaceae are borne mostly on Pecopteris type
foliage. They were borne on the under (abaxial) surface of the leaves, situated
over a veinlet or between the veinlet tip and the margin, and unlike the living
genera the majority of the fossils have radially symmetrical sori.

Mamay (1950) included the following genera in the Carboniferous Maratti-

1. Cyathotracbus Watson, emend. Mamay: Synangia pedicellate; sporangia
enclosed in a continuous synangial sheath; basal portion of synangium with central
column; apical portion hollow and cup- like. Based on petrifactions.

2. Ptychocarpus Weiss: Synangia pedicellate; sporangia enclosed in a continu-
ous synangial sheath, and fused to a central column extending through the entire
length of the synangium. Based mainly on compressions.

3. Scolecopteris Zenker: Synangia pedicellate; sporangia attached to a central
receptacle but free distally. Based mainly on petrifactions.

4. Sturiella Weiss: Synangia pedicellate; sporangia basally fused but free
above; annul us present. Based mainly on petrifactions.

5. Acitheca Schimper: Synangia sessile; sporangia attached at their bases to a
central column, with distal portions free; sporangial apices long and bristle-like.

6. Aster otheca Presl: Synangia sessile; sporangia attached to a receptacle, but
free distally, their long axes parallel to the plane of the pinnule. Based mainly on

7. Eoangiopteris Mamay: Sori linear; sporangia arranged in two rows along
fleshy receptacle; sori located along lateral veins; sporangia free. Monotypic. Based
on petrifactions.

Thus, the accepted criteria for separating the fossil genera are the following:

1. Sori linear or radial.

2. Synangia pedicellate or sessile.

3. Sporangia free above or enclosed in a continuous synangial sheath.

4. Central column present or absent in synangia. (It is important whether
the central column extends through the full length of the synangium or not, as
in distinguishing Cyathotracbus from Ptychocarpus.)

Since the discovery of the pteridosperms, some paleobotanists have regarded
certain of the supposed marattiaceous fructifications as more correctly referable to
Pteridospermae. Kidston (1925) expressed this view in finding resemblances be-
tween Acitheca and Telangium. However, Mamay (1950) concluded that upon
the basis of the sizes of fructifications and spores, the arrangement of sporangia in
son, and the superficial position of sori, this fossil group more closely resembles the
living Marattiaceae than the pteridosperms.

T TWO NEW MEMBERS OF SCOLECOPTERIS

277

Mamay (1950) considered the coenopterid fructification Chorionopterh to best
fill the qualifications of a progenitor of the Marattiaceae. This evolution supposedly
came about by means of a "phyletic slide" from marginal or terminal sori to
superficial ones. He further proposed that those fossil genera with fused sporangia
are most closely allied to the living genera Danaea and Chris tensenia, while those
with free sporangia are most closely allied to the living genera, Angiopteris, Arch-
angiopteris, and Macro glossum.

Description of New Species
Scolecopteris monothrix R. B. Ewart, sp. nov. (figs. 1-12).

This fructification is distinguished by long slender sporangia each of which
terminates in a single, unbranched, multicellular hair. These were borne on
pecopterid pinnules with strongly inflexed and deeply incised margins in synangia
consisting of four, sometimes five, or rarely three sporangia.

One well-preserved specimen was discovered in coal ball WCB 1203. It con-
sists of an ultimate rachis with about 15 attached pinnules. Several detached
pinnules and synangia were found in the same coal ball. A second intact specimen
was found in WCB 1238 along with more fragments.

Nearly all synangia were found attached to pinnules and seem to have been
preserved in their natural position. The morphology of the ultimate rachis and
attached pinnules may be seen in the reconstruction in text-fig. 1.

The average dimension of the pinnules and penultimate mid-rib are as follows:

The penultimate or rachis midrib (fig. 7) shows no cellular detail except for
the vascular supply which consists of a U-shaped strand with the concavity adaxial.
The basal part is two or three tracheids in width while the arms narrow to one
tracheid. Short multicellular hairs are found on the lower side. Figure 11 (A)
shows some of these hairs on the longitudinally sectioned midrib. Some of them
are as long as the diameter of the midrib. Figure 7 shows that these hairs are
several cells in width and are irregularly spaced in contrast to the hairs found in
the sporangia.

As seen in fig. 11, the pinnules are joined laterally where attached to the rachis
but become separate a short distance out from the rachis. They are of the
pecopterid outline characteristic of the genus. The margins are so strongly flexed
downward that they parallel and finally almost enclose the synangia as may be
seen in fig. 10 (B). The pinnule is devoid of synangia along the distal third of its
length and this portion is strongly flexed downward and inward toward the rachis
midrib almost completely enclosing the synangia and the space directly below them.
This space appears to be loosely filled with sporangial hairs which are oriented in
the direction of the midrib. These hairs may be seen, mostly in t

278 ANNALS OF THE MISSOURI BOTANICAL GARDEN

fig. 10 (A). The pinnule margins are deeply incised and may be compared in
structure to those of S. latifolia and S. incisijolia. Figure 12 (B) shows one of
the marginal lobes in vertical section. A row of these lobes in transverse section-
may be seen in figs. 8 and 1 1 (B) . In fig. 8 the cellular contents of these lobes may
be seen. The lobes join laterally at about the level of the sporangial apices, but
their vascular strands continue as discrete bundles. The only other cellular struc-
ture of the lamina that is preserved is a layer of hypodermal cells below the upper
epidermis; it is one or two cells in thickness and may be seen at the top of fig. 4
and in figs. 10, 11, and 12.

The synangia form a single row of nine or ten sori on either side of the midrib
of each pinnule.

A synangium usually contains four sporangia, but those with five, such as the
one in fig. 9, are also common and one has been found with three. The sporangia
composing a synangium are arranged radially on a common pedicel. Due to the
long slender nature of the synangia, it was not possible to find a good median longi-
tudinal section showing both the pedicellate attachment and the apex bearing a
portion of a sporangial hair. Thus, fig. 4 was prepared from three photomicro-
graphs showing these features in their proper proportions: the upper shows the
pedicellate attachment and the proximal half of two sporangia; the middle shows
the distal half of two sporangia and their apices; and the third shows simply the
basal portion of a hair characteristic of the species.

The sporangia are long, slender and roughly cylindrical but are slightly wider
near the base than at the apex; they are tightly appressed to one another and the
synangia themselves are quite crowded. The apex of each sporangium, prior to
dehiscence, is curved inward toward the center of the synangium and fits closely
against the apices of the other sporangia (figs. 2, 4). Thus a central synangial
cavity is formed which is completely enclosed. Dehiscence was accomplished by a
longitudinal slit along the inner face of the sporangium (fig. 1 ) ; in fig. 8 a mass
of spores may be seen discharging through the slit.

This leads to the question of how the spores escaped from the enclosed synangial
cavity. Figure 6 shows a synangium of S. monothrix at a later stage and is from
a different specimen than that shown in most of the figures. The preservation is
poor but it may be seen here that the sporangia have separated except where joined
to the pedicel. This synangium, along with several empty ones, was attached to a
fragment of a pinnule. In all probability the strongly curved pinnules to which
these fructifications are attached straightened somewhat at maturity, decreasing
the crowding of the synangia, and allowing the component sporangia to separate.

In figs. 3 and 4 it may be seen that a single, long multicellular hair is attached
on the outer periphery of a sporangium near the apex. It is likely that such a hair
grew on most if not all sporangia of S. monothrix, but no sporangium has been
observed to have more than one hair. No branching has been observed in these
hairs nor has any deviation in position or structure been noted. In fig. 3, the
"bulb" at the distal end of the hair is merely the point at which the hair leaves
the plane of the section and is not the termination of the hair; the total length is
much greater than is shown. By following the course of a cluster of hairs it

EWART TWO NEW MEMBERS OF SCOLECOPTERIS

is possible to estimate their length. The hairs in such a cluster all tend
one direction, toward the rachis midrib. In fig. 10 (A), a cluster of ha
seen in transverse section in the space below the sporangia. In fig. 12,
hairs may be seen on either side of and pointing toward the rachis n
where they terminate. It is almost certain that these hairs originate on 1
gia, as no trace of an attachment has been observed elsewhere. It is thus
that the sporangial hairs are 2 or 3 times the length of the sporangia or A

The average dimensions of a synangium are as follows:

:te and bilateral with a length of 12 fi
and a width of 9 /*. The smooth walls appear to be thick, but lack any markings
of the exine, except where obviously due to damage. The only member of the
genus with smaller spores is S. elegans with spores 10 /u. in diameter. Using the
method described by Radforth (1939), the average number of spores per sporan-
gium is estimated at 9,000.

From the above description it may appear that S. monothrix has some rather
striking characteristics for this genus. Some thought was given to the possibility
of establishing a new genus but it was decided that this would not be necessary.
The following are the reasons for placing this species in the genus Scolecopteris:

1. It seems most appropriate to refer it to the Marattiaceae on the basis of the
nonindusial sori borne on the abaxial surface of pecopterid leaves.

2. The pedicellate, radial synangia limit it to Scolecopteris, Ptycbocarpus, or
Cyathotrachus.

280 ANNALS OF THE MISSOURI BOTANICAL I

3. Lack of a continuous synangial sheath and central column limit it to
Scolecopteris.

4. It is borne on pinnules with a morphological structure similar to those
of S. latifolia and S. incisifolia and is believed to be closely related to these species.

5. Although it has a closed synangial cavity at maturity, this opens at a later
stage causing the synangia to assume a more scolecopterid appearance. Other
members of the genus, such as S. incisifolia, also have closely appressed sporangia.

6. Although hairs of the S. monothrix type are unique for the genus, S. major
is known to have multicellular hairs on the sporangia.

7. Although S. monothrix has small spores and large sporangia, they do not
represent the extremes of the genus. As mentioned above, S. elegans has smaller
spores, while S. major has sporangia from 0.3 mm. to 1.1 mm. longer than those
of S. monothrix.

Scolecopteris monothrix may be differentiated from other members of the genus
by the following characteristics:

1. No other member of the genus has a single, long, multicellular hair in a
fixed position on the outer periphery of the apex of the sporangium. The synangial
hairs of S. major are quite different.

2. S. oliveri and S. incisifolia are the only other species with bilateral spores.
Of these two, only S. incisifolia compares to S. monothrix in having a smooth exine
but the spores of the former are 22 X 14 fi while those of the latter are only 12
X9 p.

3. In sporangial length, S. monothri
is 0.3 mm. to 1.1 mm. longer, and S. olive
However, both of these species have a greater sporangial width than S. monothrix.
The ratio of sporangial length to width is greater in S. monthrix than in any other
member of the genus, being about 6:1.

4. The blunt connivent apices of S. monothrix further distinguish it from
S. major, which has very acute apices.

The specific name S. monothrix is proposed in reference to the unique sporangial

Specific diagnosis: Synangia consisting of four, sometimes five, rarely three
sporangia basally attached to a common pedicel. Sporangia 1.8 mm. in length and
0.35 mm. in width; apices blunt, connivent in synangia but separating after
dehiscence; dehiscence by longitudinal clefts along inner sporangial wall. A single,
unbranched, multicellular hair one cell in diameter located on the outer periphery
of the apex of each sporangium, reaching a length two or three times the length of
a sporangium. Pinnules averaging 6 mm. in length and 2.5 mm. in width
(exclusive of inflexed margins), margins strongly inflexed and deeply incised,
ultimate rachis with U-shaped vascular strand and thick multicellular hairs aver-
aging 0.8 5 mm. in length. Spores monolete, averaging 12 yu, long and 9 /x wide,
with thick walls and smooth exines.

Age: McLeansboro Group, Upper Pennsylvanian.

Locality: Berryville, Illinois.

Holotype slides: Slides 2987-2995 (from WCB 1203).

EWART TWO NEW MEMBERS OF SCOLECOPTERIS 281

Scolecopteris illinoensis R. B. Ewart, sp. nov. (figs. 13 and 15-26).

This description is based upon specimens from WCB 1204 and WCB 1239,
particularly the latter from which the type slides were chosen. The fructifications
in these coal balls were found to be in an excellent state of preservation. They
were found attached to pinnules which were in turn attached to pinnae midribs.
By cutting WCB 1239 in two planes, it was possible to obtain peels which revealed
the morphology and something of the anatomy of the specimens.

Five or six sporangia are radially grouped on a short common pedicel to form
a synangium. Of the 14 synangia seen in fig. 16, four contain six sporangia, seven
contain five, while it is difficult to determine the numbers of the other three.
Examination of other synangia has shown that those synangia with six sporangia
are almost as numerous as those with five. As may be seen in fig. 16, the synangia
form a single row on either side of the pinnule midrib. There are generally six
or seven synangia per row. A vascular bundle branching from the pinnule midrib
is found at the base of each synangium as seen in fig. 23 (C).

The outer sporangial walls are thick and curved as seen in fig. 1 6, but are com-
posed of but one layer of cells. The lateral walls are thinner and straight, but do not
touch one another, except in synangia which, like the one at the bottom left of
fig. 16, contain spores and are probably less mature than the others. It is likely
that the sporangia in a synangium separate as they mature, like those in S. mono-
thrix. Dehiscence occurs by means of a longitudinal slit along the inner wall of
each sporangium. Most of the sporangia seen in this study had already dehisced
but a few may be seen to contain spores (fig. 22). The sporangia are roughly
ovate in vertical section with acute apices that curve inward toward the center of
the synangium as seen in fig. 23 (B). The sporangial cavity does not extend into
the distal part of the apex. Figure 1 3 shows a transverse section through the apices

The average dimensions of the fructifications are as follows:

Total length of synangium - 0.66 mm.

Maximum diameter of sporangium 0.31 mm.

Diameter of pedicel 0.16 mm.

The spores are spherical in shape and have a crescent-shaped indentation in the
exine at one point as indicated by the arrow in fig. 18. Examination of well pre-
served spores leads to the conclusion that this may be regarded as a natural and
constant character. The diameter of the spores averages 15 /x.

The penultimate midrib may be seen in transverse section in fig. 23 (A). The
vascular supply consists of a U-shaped line of tracheids, and extending from the
free end on the right hand side is a small bundle of tracheids representing a pinnule
trace just departing.

Figure 24 represents an oblique longitudinal section of the vascular supply in
an ultimate pinnule. The arrow indicates a branch of this trace. Following the
trace distally, it may be seen that this is one of six such branches each of which
passes into the pedicel of a synangium. One of these traces may be seen in trans-
verse section in fig. 23 (C).

282 ANNALS OF THE MISSOURI BOTANICAL GARDEN

The pecopterid pinnules are alternately spaced along the rachis midrib as may
be seen in fig. 15. The pinnule margins are minutely toothed (fig. 17) and
reflexed downward (figs. 19-22).

Following are the average dimensions of the rachis midribs and pinnules in
these specimens:

Pinnule length - 5.5 mm.

Pinnule width _ 2.2 mm.

Rachis midrib diameter 1.0 mm.

Figure 19 shows a row of pinnules, each with a pair of sori, arranged along a
rachis midrib and sectioned transversely. At the extreme left is the rachis midrib
in oblique section. The anatomy of the pinnules is well shown in fig. 20. The upper
epidermis is not preserved, but the underlying, large hypodermal cells (A), form
a layer one or two cells thick. Below this layer is a zone of palisade parenchyma
(B), consisting of one or two layers of narrow, tightly-packed, elongated cells.
Underlying this is the spongy parenchyma (C) of loosely- massed rounded cells, and
numerous air chambers. Bounding this is the lower epidermis (D).

A notable feature of Scolecopteris illinoensis is the abundance of hairs on the
lower surfaces of the pinnule midribs (figs. 24, 26). They are multicellular, un-
branched, and one cell in diameter. Figure 21 shows that they extend beyond the
apices of the sporangia. The hairs in transverse section may be seen between the
two rows of synangia in fig. 16. Similar hairs have been observed attached to
the synangial pedicel. In fig. 22, the arrow indicates a small outgrowth of the
sporangial wall which may be the base of a hair, but these have not been seen on
other sporangia.

S. illinoensis clearly belongs in this genus but differs from other members of the
genus as follows:

1. The only other species in which it is not rare to find six sporangia in a
synangium is S. iowensis. S. minor rarely has six, but a comparison of that species
in fig. 14 with S. illinoensis shows a distinct difference in numbers of sporangia.

2. This species differs from S. iowensis which has spores measuring 65-80 fi in
diameter, four times the diameter of those in S. illinoensis. S. iowensis sometimes
has a double series of synangia on each side of the pinnule midrib, while this has
not been observed in the present species.

3. Although multicellular hairs are sparingly present on the lower side of rachis
and pinnules of S. minor, no species has hairs on the pinnules that are as conspicuous
.i- in S. illinoensis.

Specific diagnosis: Synangia consisting of five or six sporangia basally attached
to a common short pedicel. Sporangia 0.65 mm. in length and 0.31 mm. in width;
apices acute and incurved; dehiscence by longitudinal cleft along the inner sporan-
gial walls. Pinnules averaging 5.5 mm. in length and 2.2 mm. in width, with
inflexed minutely toothed margins. Ultimate rachis with U-shaped vascular strand.
Unbranched, multicellular hairs, one cell in diameter and reaching 0.7 mm. in
length produced on pinnule midribs and synangial pedicels. Spores spherical, 15 ft
c-shaped mark on exines.

EWART TWO NEW MEMBERS OF SCOLECOPTERIS

i Group, Upper Pennsylvanian.
Locality: Berryville, Illinois.
Holotype slides: Slides 2996-3005 (from WCB 1239).

Two new species of Scolecopteris are recorded both of which were found in the
Upper Pennsylvanian of Illinois. S. monothrix is distinguished by its long sporangia
each of which has a distally attached long, multicellular hair. The tips of the com-
ponent sporangia were tightly appressed when immature. The spores are very small
and their number per sporangium high. S. illinoensis is characterized by having
five or six sporangia per synangium, small spores, and conspicuous multicellular
hairs on the underside of rachis and pinnules.

This study was carried out in the laboratories of Washington University under
the guidance of Dr. Henry N. Andrews, to whom the writer wishes to express his
sincere thanks; it is a part of a paleobotanical research program aided largely by
the National Science Foundation. Thanks are also extended to Dr. Tom L. Phillips
for assistance with paleobotanical techniques and for aid in other ways.

Literature Cited

Roy. Soc. Edinburgh, i

ANNALS OF THE MISSOURI BOTANICAL GARDEN

PLATE V.
Scolecopteris monothrix
Fig. 1. Transverse section of a synangium sectioned in the plane 1-1 of fig. 4. Note
spores passing into synangial cavity. Slide 2994. X 8 5.

Fig. 2. Transverse section of apex of synangium sectioned in the plane 2-2 of fig.
4. Note that all four apices fit closely together. Also note the four celled base of a
sporangial hair (arrow). Transverse section of several hairs from other sporangia may be

Fig. 3. Longitudinal section of sporangial apex showing hair. "Bulb" at distal end

langial hair indicating character-

a?

MBF.RS OF SCOLECOPTERIS

O NEW MEMBERS OF SCOl I < up I 1 KIN

EWART TWO NEW MEMBERS I

Explanation of Plate

PLATE VI.
Scolecopteris monothrix

Longitudinal

Fig. 8. Transverse section thr ind inflexed pinnule

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation

EW ART— TWO NEW MEMBERS OF SCOLECOPTERIS

EWART TWO NEW MEMBERS OF SCOLECOPTERIS 287

Explanation

Fig. 16 Single pinnule from fig.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

long hairs on midri

-TWO NEW MEMBERS OF SCOLECOPTERIS 289

PLATE X.

ScoL

copteris illinoem

Fig

D

Longitudinal
3. B. Incurvec
Spore mass in

section of fertile part of p
porangium. Slide 2997. X

Fig

uln!

Lon

sill

l^ "If"

Fig

25.

See

ion showi

ig pin

a m

drib with pinnul

Fig

26.

Ha

rs on pinn

ulc mi

r>b.

Slide 2999. X

PORTRAITS OF PLANTS. A LIMITED STUDY OF THE "ICONES"

ERNA RICE EISENDRATH

Jude some portrayal of the habit of growth of the plant; that they be
botarucally accu u n . Each of the "icones" must fulfill each

.r oft-repeated use, and their exception to the

Four genera of the Cw .. L„ffa) are chosen as limiting

Garden discussed in relation to the criteria previously set up. The studs j 5 concluded with a

of Botany, Washington University, St. Loot* 3 0, Missouri*

Botanical Use of the Word "Icones"
The word "icones" was taken into botanical parlance by Otto Brunfels in
1530. Its spelling in the title of his Herbarum vivae eicones still conveys a hint
of orientalism just strong enough to tie in with its common meaning today; to the
layman, an icon is (as it was, to the publishers of Webster's New International
Dictionary of 1918) "a religious image in the Eastern church". For botanists,
however, it is not the gilt on these distorted saints that glows when they speak
reverently of "icones". Their enthusiasm is evoked by a kind of floral illustration
which they automatically lump as "icones" in special sections of libraries; to which
they readily refer as "icones" in conversation; of which they will confidently name
names as true "icones"; but for which they hesitate to pronounce a definition. So,
though everybody knows the special meaning of the word, there remains the prob-
lem of establishing criteria by which to implement its definition; a later problem
will be to list the books containing illustrations that satisfy the criteria.

To start from an authoritative beginning I give you Benjamin Daydon Jackson,
who in 1900 first published in London his most useful Glossary of Botanic Terms.
To him, "icones" 3 are merely "pictorial representations of plants; botanic figures".
In this sense which includes all botanical illustrations it was certainly used by
Pritzel 4 in his Iconum Botanicarum oi 1855 and is also so used in the Latin title
page of the Index Londinensis.

For the purposes of this paper, however, I prefer basing my criteria on the last
word of the last definition of "icon" given in the Oxford English Dictionary
(Oxford, 1933): "esp. applied to 'figures' of animals, plants, etc. in books of
Natural History. Obs." The use of "icon" to include all natural history illustra-
tions is now obsolete; when botanists speak today of the "icones" it is with a

292 ANNALS OF THE MISSOURI BOTANICAL GARDEN

different connotation, not specifically stated in any source that I can find,
implies more than scientifically accurate illustrations, and more t
plant portrayal, falling some place between the two and partaking of both but,
with a kind of hybrid vigor, better than either. Dr. Hugh Cutler bravely ventured
to make the following brief but telling statement. "The 'icones'," he said, "may
have nothing to do with the church exactly, but a great many botanists worship
them." That is, apparently, about as far as educated angels dare to tread; this fool
will now step in to define "icones" by establishing the criteria which distinguish
them from other illustrations in botanical books.

Criteria Distinguishing "Icones"

After examining every lead to illustrations of the key 5 genera of Cucurbi-
taceae available at the Missouri Botanical Garden library (some 157 volumes), I
have set up four arbitrary criteria for "icones". The first is that the illustrations
be full page 6 in size; the second, that they include portrayal of the habit of growth
of the plant. The third is botanical accuracy, and the fourth artistic presentation.

In many instances my first criterion is necessitated in order to eliminate illustra-
tions used simply as elaborations of a text. According to Nissen (1951), "the
documental botanical illustration should, for the co-investigator, replace the plant
itself, be he distant in time or place". 7 So, too, must true "icones" be visual
portraits of plants, dependent in no sense upon the written word. P. J. F. Turpin,
the great French botanical artist, quite naturally believed that the brush was
greater than the pen as a naturalist's tool, and wrote that he "who possesses only
the latter loses perhaps the more significant" (Turpin, 1820). But the brush used
to depict the plant must do so in considerable detail, and the portrait must be
extremely accurate — so accurate as to compete with the botanist's dried herbarium
specimen as a source material for study. (The possibility of this is recognized in
the official acceptance, though rare, of such representations as type specimens.)
Small figures of plants cannot possibly fulfill all of these requirements; hence the
first criterion for true "icones": that they be full page illustrations.

My second criterion for defining true "icones" is that they represent the habit
of plants, and include both vegetative and reproductive organs. It is possible to
object that this is unnecessary, since one can assume that generic characters are
"given", and that only identifying specific characters need be depicted in plant
portraiture. This objection I liken to a statement that a man's fingerprints
identify him; certainly they will differentiate him from all other men, as certain
characters of a species of plants will also differentiate it. Neither fingerprints nor
selected characters, however, will give us portraits of the individual or the species,
and, in defining today's usage of the word "icones", I insist that portrait is the
proper synonym (if any is truly proper) .

"Icones", then, must portray habit of growth, and here is another example of
the need to pin a word to exact definition. What is "habit"? Webster (New

EISENDRATH, A STUDY OF THE ICONES 293

Intern. Diet., 1918) gives as its eighth definition of the word, "Nat. Hist. Char-
acteristic form or mode of occurrence or growth; as, elms have a spreading habit."
Jackson, deriving "habit" from the Latin, gives, "The general i
plant, whether erect, prostrate, climbing, etc." Sir Joseph Dalton Hooker,
systematist, Director of Kew Gardens, and editor of volumes 91-130 of Curtis'
Botanical Magazine, wrote, "Habit of a plant, of a species, a genus, etc., consists of
such general characters as strike the eye at first sight, such as size, colour, ramifica-
tion, arrangement of the leaves, inflorescence, etc., and are chiefly derived from the
organs of vegetation" (Hooker, 1875-97). To this I append the requirement that
either flower or fruit must be depicted; however, it is not always easy to decide
where "habit" is illustrated and where it is not.

I have, for instance, not listed among the "icones" pertinent plates (see plate
XI) in Tournefort's Ins!: ; (1700), even though they are cited

by Linnaeus (Species Plantarum, Stockholm, 1753, vol. 2, pp. 1010-1011) in his
original description of the genus Cucurbit a. My reasons are that Aubriet illustrated
for Tournefort only parts of plants — fruits (and with dissections, too! ) , seeds, flow-
ers, peduncles, etc., but no idea of the appearance of the plant itself is given. Again,
L. H. Bailey, in his Garden of Gourds (1937) , includes page after page of drawings,
almost all of which show sections of the stem, flowers, fruits and leaves. Since
they also show tendrils where tendrils exist, it is apparent that such plants would
climb if given a chance to do so, and I have therefore listed a number of these
illustrations among the "icones"; but I have eliminated such as that of the "Mock
Orange" variety of Cucurbita Pepo, which depicts a pistillate flower, a fruit with
peduncle attached, a tendril, a stem in cross-section, and enlarged dissections of the
style and stigma, as well as of the anther column; but habit, no! — so it is not listed.

illustrate the root. This is undoubtedly a loss to plant portrayal, and probably
came about when interest in plants shifted to a taxonomic basis from the pragmatic
approach of the rhizotomists and their followers. These predecessors of the herbal-
ists performed magic and cures based on lore often dating from the Greek period,
and generally associated, in popular folk ritual, with properties inherent in plant
roots (Singer, 1928). Also worth noting is the rarity of portrayal of habitat
along with habit. This is so important a factor in the systematics of botany, and
was used so wonderfully by Audubon in his portrayals of birds and animals, that
one wishes it were more often used as backdrop to "icones" of our flora. Nissen's
(1951) statement, "Plant sociology ... is of recent date, and representation of it
is prepared more and more with the camera," 8 very possibly explains this lack.

If the portrait of a plant belongs at all within the field of botany, it must be
scientifically accurate, my third criterion for "icones". The degree of precise detail
will vary in each plant portrait, as it does in human portraiture. There is, however,
a most important difference in the requirements of the two. A portrait of a man
shows all the artist can tell us about a single individual, be he godlike or sickly
earthling; but, to quote from Arber (1938), "the drawing which is ideal from the
standpoint of systematic botany, avoids the accidental peculiarities of any indi-

294 ANNALS OF THE MISSOURI BOTANICAL GARDEN

vidual specimen, seeking rather to portray the characters fully typical for the
species." This distinction is well illustrated in the very book that introduced the
word we are trying to define, Brunfels' Herbarum vivae eicones: here we find that
some of Weiditz' magnificent "living portraits" exhibit such "accidental peculiari-
ties", probably as a result of an overly enthusiastic return to direct observation
from nature. Note figure 1, of Weiditz' Arctium lappa L., in which the plant's
poor wilted leaves add nothing except the information that it had been pulled from
the earth too long before Weiditz got around to drawing it. He apparently did not
itch, as Tournefort felt the true botanical artist should, to get his observations on
to paper. "It frets a man", Tournefort wrote (quoted from Blunt, 1951) "to see
fine Objects and not be able to take Draughts of them; for without this help of
Drawing, 'tis impossible any account thereof would be perfectly intelligible." The
requirements of accurate botanical illustration are that the characters of a species
be accurately represented; fortuitous accident in the life story of the individual,
such as its being attacked by voracious insects, is of no interest to anything or
anybody but the insect and the plant.

It is almost pedantic to add that accuracy is not impaired by the additional
quality of aesthetic satisfaction, the fourth of my criteria. Among the "icones"
the two are found in variable proportions, the balance being struck not only on the
basis of the purpose for which the pictures were made, but also depending upon
the date of their publication. The demands of fashion in taste and interest can be
followed clearly throughout the centuries. Many flower illustrations of great
decorative quality have no botanical value whatsoever, and quite correctly end their
usefulness as lamp shades; this kind of work reached its nadir in the 19th century,
when spinster ladies sentimentally painted charming bouquets on china, as well as
any other available surface. When these illustrations appeared in books, they were
usually accompanied by saccharine verse, under such titles as The Moral of Flowers
(Marquand, 1947). 9 Today's floral illustrations represent the opposite extreme.
We tend to file scientific data in compartments well isolated from all aesthetic
consideration; for this reason we must beware of eyeing with suspicion the accuracy
of figures that are pleasing. It is important to remember, as is subsequently
discussed, that the "icones" include the work of Turpin, Fitch and many other
extremely gifted artists, whose talents were exhibited in a field that demanded far
more of them than simple art for art's sake.

An aesthetic quality, then, is the last of four criteria set up for true "icones";
full page illustration, portrayal of habit and botanical accuracy have all been
discussed above. Extremely few illustrations will pass all these tests; but who

, the"

9th century. Linnaeus hims

elf effused

most colorfully:

.

beds, gloriously a

11 the greater so

sified,

Clasp with fond arms, and

romK

utch, Joseph Wood, The Ga

tetter's World, New York,

E.SENDRATH-A STUDY OP THE .CONES

296 ANNALS OF THE MISSOURI BOTANICAL GARDEN

knows better than a botanist that taxa must be elastic, and that no set of characters
is invariably true for all the units that the taxon must include!

"Icones" may be found in regional floras, de luxe tomes concerned only with the
plants growing in a single garden, learned treatises on classification, serials pub-
lished by scientific institutions or nurserymen's catalogues. All of these fall within
the enormous body of illustrated botanical literature, of which the Index
Londinensis lists well over 3000 titles, and for the purposes of this paper it was
necessary to establish a limited approach. One could set up chronological barriers,
choosing, as did Sit well and Blunt (1956), two centuries as a high point in botan-
ical illustration. One could profitably discuss only the work of one great artist,
such as Turpin, whose name appears several times on these pages. One could trace
any of several aspects of the subject through the pages of Curtis' Botanical
Magazine during its almost two hundred years of publication. Or, one can approach
this loose aggregate of illustrated books with a particular subject as the tool by
which to cut it down to workable proportions. This method has been used, with
four genera of the Cucurbitaceae the limiting factor.

Key Species of the Cucurbitaceae as Limiting Factors in This Study
Various reasons may be cited for pursuing the study of "icones" through illus-
trations of members of the Cucurbitaceae. As is later discussed, the family is
generally distinguished by a fleshy fruit, a structure known as a "pepo"; but,
whereas in many plant groups the fruits may be used to delimit various taxa, it
cannot be said of the cucurbits that "by their fruits ye shall know them". There
are "pepos" and "pepos", and the close relatives, among many cultivated forms, will
produce offspring which exhibit tremendous differences. In size alone, the "pepo"
may be as small, let us say, as a tangerine, or absolutely enormous, as witness the
Hubbard squash or the long-as-a-man-is-tall Lagenaria. It is never possible con-
veniently to mount (and seldom even to preserve) such fruits as herbarium mate-
rial; therefore, since it is for the edible, decorative and useable fruits that the family
is cultivated, illustrations of them are of great botanical importance.

The flowers, too, present a problem in survival. Though they may be showy,
they collapse dismally after one day of glory in the sun, a sad truism for the
taxonomist, since they present most interesting morphological variations. (It may
be due to the same truism that the work of many of the best known floral artists
does not appear among "icones" of the Cucurbitaceae; an illustrator who attempts
to draw them painstakingly from life will find that their lives are too brief to enable
him to create the illusion of prolonging them! ) The choice of the key genera was
made for several reasons: Fevillea because it is the primitive group in the family,
serving as starting point for a number of phylogenetic series; Cucurbita, Lagenaria
and Luff a because their distribution is so wide and varied as to insure their appear-
ance in a variety of publications.

The Cucurbitaceae include about one hundred genera (fide Lawrence, 1951).
Not only the placement of the family in a natural system, but also the delimitation
of generic characters have intrigued and baffled botanists over the years. This
problem is, however, entirely beyond the scope of my study.

EISENDRATH A STUDY OF THE ICONES 297

The Cucurbitaceae are generally tropical plants, with some species hardy enough
to extend into the temperate zones. All, however, are frost-tender herbs, and sur-
vive to fruit in chilly climates only because they grow very rapidly. Client hit a is
native to the American tropics, and Luff a (with the exception of L. operculata,
also New World) to the Old World tropics, whence Lagenaria probably also came,
but this genus has been for a very long time universally distributed throughout the
warm parts of the world.

The same three genera are of interestingly varied economic importance. Cucur-
bita is cultivated for its edible and decorative fruits, the pumpkins and squashes
which traditionally are staples of the American diet, and archeologically are known
to have been important long before the advent of the white man to either Amer-
ican continent. Luffa's fruit has a fibrous interior when mature which, dried,
becomes the "vegetable sponge" so widely used by masseurs; the young fruits
(apparently less abrasive!) are edible. Lagenaria fruits in an incredible variety of
forms; wherever it is grown, it teases the imagination to put the fruits to use as
ladles, vases, pitchers, bird houses or merely decorations. A visitor to Georgia and
Florida at the turn of the century (Odell, 1904-05) found that "hardly a domestic
operation there is complete without the aid of the Gourd in some form". It was
even used as a cradle for negro babies!

Fevillea (tropical American) is neither widespread nor economic, but is included
because it is not cultivated, which fact, along with its many primitive characters,
presents interesting botanical contrasts to the other genera. In Fevillea, for
instance, the corolla is distinctly polypetalous, as it is also in Lagenaria. In Luff a
it is deeply lobed; in Cucurbita the five petals are joined in a floral tube. Again,
in Fevillea five stamens with biloculate anthers of an unspecialized type
stand perfectly free, alternating in position with the petals; in Luff a, as in the
other three genera, there are usually basically only three stamens (because of
the coherence by their filaments of two pairs of the original five), but they are
borne free on the calyx tube. In Lagenaria the anthers of the stamens cohere
lightly, but are not truly connate; in Cucurbita the filaments are connate (except
that sometimes they are free at the very base) and the anthers are twisted so
inextricably together in a column as to be almost indistinguishable. In none of the
last three genera are the anthers primitively biloculate; they dehisce by a single
longitudinal split.

Of the books listed below, the following contain "icones" of Fevillea:
Marcgrave-Piso, Plumier, The Botanical Magazine, Dictionnaire des Sciences Natu-
relles, Descourtilz, Martius, Velloso, Engler and Rendle. Of these, all but the
serial, the dictionary and the systematists concern plants of the Caribbean and
South America, the earliest of them being Marcgrave's first edition in 1648.

Lagenaria is illustrated, among the books listed below, by Fuchs, Dodoens,
L'Obel, Dalechamps, Tabernaemontanus, Gerarde, Parkinson, Bauhin, Rheede,
Rumpf, Oskamp-Zorn, Descourtilz, Velloso, Iinuma, Duthie and Fuller, Bettfreund
and Bailey, indicating an early European knowledge of the plant, as well as its wide
spread in tropical areas of the world.

Luffa, rather strangely, is not pictured by those earlier herbalists who illustrated

298 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Lagenaria, though both genera are thought to have originated in the Old World
tropics. Among the "icones" listed we find Luffa first in Vesling, illustrating
Egyptian plants in 1638. It appears next in Commelin, illustrating the plants
cultivated in the Physic Garden of Amsterdam in 1706; later, it is pictured by
Rheede, Rumpf, Cavanilles, The Botanical Magazine, Velloso, Martius, Wight,
Iinuma, Blanco, Duthie and Fuller, Lecomte, Engler and Bailey. This indicates a
wide distribution indeed, although the plant was apparently not cultivated in
Europe as early as was Lagenaria.

Of the four genera considered in this study, Cucurbita is not only the best
known and most widely cultivated today, but archeologically it is of great interest,
being found in diggings where it can be used to trace the agriculture of early man
in the so-called New World. Because of its tremendously variable fruits, seeds and
other organs, Cucurbita also challenges the systematist and intrigues the geneticist;
because its floral parts are large, it is useful to morphologists and physiologists and,
because the cucurbits are easily grown, their large fruits can be retained for study,
while their indeterminate growth permits flowering from maturity to frost. As a
matter of fact, Cucurbita presents satisfying study material for most botanical
specialists, though geneticists, pleased with the large pollen mother-cells, find diffi-
culties in their miniscule chromosomes (Whitaker and Bohn, 1950). However, it
is probably because of its edible and decorative fruits that we find among the
"icones" more illustrations of Cucurbita than of the other three genera selected.

The earliest "icones" of Cucurbita, the two appearing in the fabulous folio
edition of Fuchs, 1542, are probably also the most beautiful. We find Cucurbita
next in the 1552 Aztec herbal known as the Badianus Manuscript; and at short
intervals from then on, mostly in pictures derived more or less from Fuchs,
cucurbits appear in Matthiolus, Dodoens, L'Obel, Dalechamps, Tabernaemontanus,
Gerarde, Parkinson and Bauhin. Original figures are found in Rheede, Rumpf,
Buch'oz, Schkuhr, Vietz, Descourtilz, Chaumeton, Wright, Iinuma, Blanco, Duthie
and Fuller, Bettfreund and Bailey. The list evidences a rapid and wide spread of
knowledge about and interest in the plants.

Earliest Illustrations of Key Species
None of the titles in the list of "icones" of the selected genera predates 1542,
the year of publication of Fuchs' De historia stirpium. This fact is significant
because the historical background to a discussion of plant illustration is strangely
poor. Though all of man's existence (along with that of other animals!) must of
necessity have been ultimately dependent upon the flora of his world, relatively very
few early examples of plant delineation are known, in comparison to those of fauna.
Among the key genera, the earliest I find are chance reprints in modern books, all
of Lagenaria. The first is printed by Behling (1957) , and is taken from the manu-
script Tacuinum Sanitatis in the Austrian National Library. It is described as a
page from the

"Hausbuch der Cerutti, of upper-Italian origin, from the turn of the 14th to the 15th
century. ... One sees the flask-shaped gourd (O, an excellent

EISENDRATH A STUDY OF THE ICONES 299

The second reproduction is Arber's plate iii (1938). This is from Konrad von
Megenberg's Das Puch der Natur, Augsburg, 1475, which contains the first printed
plant illustrations. It shows, in the center of a group of plants, an unmistakeable
Lagenaria fruit, pendulous on a terminal branch, the whole plant arising from a
basal rosette of heart-shaped leaves. The third is in a 1940 facsimile of an Aztec
herbal of 1552 (Codex Barberini, Latin 241, in the Vatican Library). Familiarly
known, from the name of its translator, as the Badianus Manuscript, it was put
together by Martin de la Cruz, described in the text as an "Indian physician . . .
who is not theoretically learned, but is taught only by experience" (Arber, 1938).
The colored illustration shows a Cucurbit* foetidissima, so identified as the leaves
and root seem to belong to a perennial plant. In the herbal it is called the "Ayonel-
huatl", translated as "calabash root", and is described as an ingredient of a complex
concoction recommended to relieve the pain of childbirth.

Fuchs' De historia stirpium

Fuchs' beautiful Dc historu stirptum wi$ published in Basle in 1542, twelve

years after Brunfels introduced "icones" as a botanical word. Like Brunfels', Fuchs'

book is illustrated with portraits of plants made by artists who had actually looked

Plant illustration in the herbals, prior to Brunfels and Fuchs, consisted of constantly
recopied figures, derived from ancient originals, which in each reproduction lost
more of their freshness and validity. Strangely, the invention (circa 1450) of
movable-type printing, with its tremendous impact on the dissemination of all
kinds of knowledge, did not immediately influence publishers to have their artists
take a fresh look at plants. And this, despite the fact that illuminated manu-
scripts of the period often include charmingly naturalistic flower paintings, and
that such artists as the Van Eycks, Van der Goes, Jacopo Bellini and Botticelli
painted flowers beautifully even before a magnificent high of superb flower por-
trayal was reached in the work of Leonardo 10 (1452-1519) and Diirer (1471-
1528).

Brunfels and Fuchs let fresh air into the musty chambers of illustration of the
herbals, Fuchs leaving for us fine woodcuts of both Cucurbit a and Lagenaria; but
it is from them (see below) that subsequent generations of herbalists were satisfied
to derive their illustrations, thus slamming down the window once again.

Fuchs shared with Brunfels the realization of the importance of drawing plants
from life, but his artists, unlike Weiditz, (whose "living portraits", you will
remember, were sometimes limited portrayals of individual plants) also understood
the necessity of depicting plants typical of a species. From his own statement we
know that Fuchs worked closely with the artists.

ANNALS OF THE MISSOURI BOTANICAL GARDEN

shadows, and other less necessary things, by which the delineators sometimes try to wm
artistic glory." (Quoted from Arber, 193 8.)

Fuchs was himself a physician, famed throughout Europe for his success:
:atment of an epidemic disease that hit Germany in 1529, but his love of plai
is more than a professional interest. Again in his preface, he wrote,

Fuchs' great work represents the high point of the Renaissance herbal (Singer,
1928) ; the text has the botanical limitations of all herbals, but the illustrations are
superb. The figures, fitted gracefully into the frame of the full folio page, do not
thereby lose their naturalism. The lack of shading that Fuchs mentions sometimes
makes the outlines appear thin, but this is much relieved by the presence of color,
which the artists probably intended should be applied by hand, after printing
(Blunt, 1951).

The large folio plates of the key genera include two of Cucurbita Pepo, called
by Fuchs "Cucumis Turcicus" and "Cucumer Marinus", and three of Lagenaria
vulgaris, labeled respectively "Cucurbita Maior", "Cucurbita Minor" and "Cu-
curbita Oblonga".

In 1545 an octavo edition of Fuchs appeared which contained the same illus-
trations of the plants under discussion, with exactly the same names given them,
but in very much reduced size and mirror images of the larger cuts. 11 They were
to have a long life indeed.

These small-scale cuts from Fuchs, along with some from other sources, were
used for over a hundred years by subsequent generations of herbalists. It would

criteria set up for "icones"; but I feel they must nonetheless be included in this
study, and I claim justification from the excellent authority of Dr. William Trelease,
first director of the Missouri Botanical Garden. Dr. Trelease wrote,

er of the plant figures of the herbalists are far frc
: shall also see that others which my definition wot
swish of the skirt, play important roles in Linnae

EISENDRATH A STUDY OF THE ICONES 301

With Fuchs' "icones" well in mind, we now embark upon an unbelievably com-
plex task, tracing from his and other true portraits of plants the often inadequate
thumb-nail sketches used and reused by successive herbalists.
Woodcuts Derived from One Source and Used by Several Authors

Bailey (1929) identified Fuchs' "Cucumis Turcicus" as "undoubtedly a
Cucurbita Pepo of the vegetable marrow kind", and Whitaker (1947) agrees. In
1563, in his Cruydeboeck, Dodoens reproduced the smaller cut as "Pepones magni",
using it again in 1616, in his Stir pin »i historic pern p fades, as "Pepo maior ob-
longus". In 1587, in his Historia generalis plantarum, Dalechamps, possibly using
the same cut, called it "Cucurbita Indica longa", and in 1636 Gerarde, in The
herball, described it as "Pepo maximus oblongus". The picture ends its career, as
far as I know, as "Cucurbita foliis asperis sive Zucha, flore luteo, Ic. 11" in Johannes
Bauhin's Historia plant arum of 1650. This last derivative from a magnificent
original has been cut down to diminutive measurements (8/ 2 X 4/ 2 cms.); the
tendrils of the plant are so stylized as to be almost indistinguishable organs, the
character of the peduncle at the point of attachment is entirely lost, and the
illustration becomes as far removed as possible from a portrait of a living plant.

This series presents a good example, however, of the botanical significance of
many of the small derived wood-cut figures of the herbals. Linnaeus in his defini-
tion of Cucurbita (Species plantarum, 1753, vol. II, p. 1010) cited a variety /3
with Caspar Bauhin's Pinax as reference. Caspar, the son of John, did not illustrate
his book, but referred to the pictures mentioned above, from Dodoens' work
of 1616 and Gerarde's 1636 Herball, so Linnaeus' description is indirectly des-
cended from a picture which may be described as a foster child of Fuchs'.

Again, Fuchs' "Cucumer marinus" had a long and mostly unhappy history
(see Figures 2, 3 and 4) . Matthiolus (Comment arii in libros sex Pedacii Dioscoridis
de materia medica) in 15 58 printed as "Zucche Indiane" and in 1560 as "Cucurbita
Indica" a plate which is certainly not a direct copy of Fuchs', but which almost as
certainly was copied from his. Dalechamps in 1587 used Matthiolus' cut, in
mirror image, as "Cucurbita Indica rotunda", and Linnaeus (Species plant arum,
1753, vol. II, p. 1010) cites Dalechamps as a reference for his Cucurbita Pepo.
Bailey (1929) says of this cut that it represents

In 1563 Dodoens printed a cut more exactly derived from Fuchs' octavo "Cu-
cumer marinus", calling it "Pepones rotundi", and in 1616 Dodoens reused it, this
time as "Pepo rotundus minor". By 1636, when Gerarde called it "Pepo minor
sylvestris", the detail was considerably coarsened, but it is very possible that the
same plate was used. In 1650 Bauhin printed another poor, small, derivative cut,
this one "Cucurbita foliis asperis sive Zucha flore luteo, Ic. V."

Fuchs' illustrations of Lagenaria, too, produced a motly swarm of progeny. His

rbita oblonga" is in Dodoens, 1563, as "Cucurbita anguina ,
as "Cucurbita longa, folio molli, flore albo". Fuchs' "Ci

. B.u;hi:

302 ANNALS OF THE MISSOURI BOTANICAL GARDEN

EISENDRATH— A STUDY OF THE ICONES

Peponesrotundi. ftonfc $<pM<tl.

Peponeslati. %>X(t><¥>(potn(tt.

304 ANNALS OF THE MISSOURI BOTANICAL GARDEN

reappears in Dodoens in 1563 as "Cucurbita cameraria maior", which Dalechamps
also called it in 15 87. (His plate, however, is a reverse of Dodoens', with slightly
different measurements.) Gerarde (1636) seems to have used Dodoens 5 block,
calling it "Cucurbita lagenaria", and in 1650 Bauhin brings it to a small, inglorious
end, again, as "Cucurbita latior folio molli, flore albo". Fuchs' third and last
Lagenaria was called "Cucurbita minor", and Dodoens (1563) called it by the same
name. In 1587 it became "Cucurbita cameraria minor" in Dalechamps, and in
1650 "Cucurbita lagenaria" in Bauhin. Obviously, there was some confusion in
nomenclature! All of the Dodoens illustrations are approximately the same size as
the cuts of Fuchs' 1545 edition, and possibly from the same blocks. Dalechamps'
cuts are slightly smaller and all but the "Cucumer marinus" derivative in reverse
of the above, so he cannot have used the originals; Gerarde's print is clearer than
Dodoens', but may be from the same cuts. The cuts used by Bauhin are all smaller
and far coarser than any of the others.

However, not all the wood cuts used in the herbals of the period after Fuchs

by Dodoens is used again by Gerarde in 1636 as "Pepo Indicus angulosus", and
identified by Whitaker (1947) as "C. Pepo, possibly var. White Bush Scallop".
Again, Dalechamps in 15 87 published a charming decorative print of "Cucurbita
verrucosa" with a good deal of fine, accurate detail, identified by Whitaker as
"evidently a warted variety of C. Pepo". The same material, cut down in size, in
mirror image, with far less grace and detail, was published by Bauhin (1650) as
"Cucurbita verrucosa" and cited by Linnaeus, in 1753, in Species plantarum in
his definition of "Cucurbita verrucosa". (But Linnaeus added, in later editions,
that the plant bore marked resemblances, in flowers and seeds especially, to C.
Pepo.) This is another example of the enormous botanical value of even such a
coarse, much-worried hand-me-down as Bauhin's figure, especially as, according to
Bailey (1929) the Linnaean herbarium contains no specimen named "Cucurbita
verrucosa" by Linnaeus. 12

In 1576 (Plantarum seu stirpium hhtoria) and 1591 (Icones stirpium) L'Obel
published as "Pepo maximus Indicus compressus" a cut identified by Whitaker
(1947) as the first illustration of Cucurbita maxima. In 1587 Dalechamps used
the picture in mirror image, as "Pepo maximus Indicus, L'Obel"; his cut measures
12X7 cms., a bit smaller than the one used by L'Obel, by Dodoens as "Pepo
rotundus maior" in 1616, and again in 1636 by Gerarde, as "Pepo maximus com-
pressus". This cut (referred to L'Obel) is cited by Lamarck in his Encycfopedie,
ii, 151 (1786). Bauhin is also cited by Lamarck, for his portrait of "Cucurbita
aspera, folio non fisso, fructu maximo, albo sessili".

In 1591 Tabernaemontanus published in his Neuw Kreuterbuch as "Pepo
Indicus minor angulosus" a cut used by Gerarde in 1597 as "Pepo Indicus angulo-
sus". The print is poor in Gerarde, but the detail in the cut is excellent and is
identified by Whitaker (1947) as "probably C. Pepo, var. 'Table Queen'". As
before, Bauhin has a mirror image of the same cut, in smaller size, and names it as

EISENDRATH A STUDY OF THE ICONES

3 OS

did Tabernaemontanus. The cut is un-
usual among those of the period in that
it does not show a root.

The same authors use another cut,
originally described by Tabernaemontanus
as "Pepo Indicus minor clypeatus" and by
Gerarde (1597) as "Pepo Indicus fungi-
formis". Again Bauhin's use of it is in
mirror image on a smaller scale; he calls it
"Cucurbita clypeiformis cortice molli et
ramosa" and Bailey (1929) says it is "un-
doubtedly one of the forms of Cucurbita
Pepo var. Melopepo" although tendrils are
a conspicuous part of this figure, and
Bailey himself (1954) describes var.
Melopepo as "Plant covering small space,
compact, not running or tendril-bearing".
However, this last, negative character is
not cited by all authorities.

Again, Tabernaemontanus' "Cucur-
bita capitata" appears in Gerarde (1597
and 1636) as "Cucurbita sylvestris fungi-
formis" and in Bauhin (smaller and
reversed) as "Cucurbita capitata Taber-
nomontani sive clypeiformis". This, which
bears no tendrils, Bailey (1929) again iden-
tifies as Cucurbita Pepo var. Melopepo.

Tabernaemontanus' "Melopepo clype- Fig Ure 4 "Cucurbita foliis asperis sive

atus" becomes "Pepo maximus clypeatus" Zuch

in Gerarde (1597) and (diminished in jjjjj^ ZJS^T^£wSil "**"'*
size and reversed in image) "Cucurbita

clypeiformis, sive Siciliana Melopepon latus a nonnullis vocata" in Bauhin. This
cut is again without roots; it does, however, show exceptionally good detail of
flower and fruit structure and is without tendrils. It is particularly important as
it represents another definite citation by Linnaeus in his description of Cucurbita,
this time "C. Melopepo". Bailey (1929) says "the plant is undoubtedly what we
know as Bush Scallop squash".

Bauhin used Tabernaemontanus' "Cucurbita longa" as his "Pepo Indicus minor,
Tabern." and Gerarde in 1597 used Tabernaemontanus' "Melopepo compressus" as
his "Pepo maximus compressus"; Tabernaemontanus' "Melopepoteres" as his "Pepo
maximus rotundus"; and Tabernaemontanus' "Pepo maximus oblongus", as (for
once not complicating the problem) "Pepo maximus oblongus".

Gerarde, in 1636, had access to cuts used by L'Obel in 1576 and 1591. L'Obel's
"Pepo oblongus vulgatissimus" (1576) became "Pepo oblongus" in 1591, but

Icon

306 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Gerarde called it "Pepo maior sylvestris". L'Obel's "Pepo rotui
Melonis effigie" retains its name in both his editions, but Gerarde calls it "Pepo
Indicus minor rotundus", and Parkinson (Theatrum botankum) in a very coarse
1640 reprint, "Melo Indicus parvus".

The last of the cuts of Cucurbit a, in Bauhin, makes a very poor print indeed,
and must, like the others, have been cut down from another wood block, although
I have not found it printed elsewhere. However, it is another important, though
indirect, Linnaean reference. Bauhin called it "Cucurbita foliis asperis sive Zucha
flore luteo, Ic. I." Linnaeus refers in his Hortus Cliffortianus™ and his Hortus
Upsaliensis to Ray's Historia plantarum 14 which in turn refers to this cut in Bauhin.
Bailey (1929) says, "This figure is again C Pepo as we know it".

Gerarde in 1597 used three cuts of Lagenaria which had been used previously
by Tabernaemontanus. The latter author's "Cucurbita Indica minor" became
"Cucurbita anguina" in Gerarde; his "Cucurbita lagenaria minor", "Cucurbita
lagenaria sylvestris". In 1636 Gerarde used cuts as "Cucurbita anguina" and "Cu-
curbita lagenaria sylvestris" which had also been used before and were to be used
again. L'Obel (1576 and 1591) had used the first as "Cucurbita sive Zuccha
omnium maxima anguina" and Dodoens as "Cucurbita longior". Parkinson appar-
ently used the same plate, in 1640, as "Cucurbita longa." L'Obel also in both his
editions used the second of these as, simply, "Cucurbita lagenaria" and Dodoens as
"Cucurbita prior". Again, Parkinson apparently used the same plate, this time as
"Cucurbita lagenaria maior".

The constant and confusing repetition of these figures is largely due to "pools"
of botanical cuts made by the publishers of botanical books. Of these the best
known is Christopher Plantin, the great printer of Antwerp. Plantin's firm con-
tinued publishing for three hundred years, his place of business becoming a museum
in 1876, when it was purchased by the city of Antwerp from the eighth generation
of the family Moretus, direct descendants of Plantin's son-in-law, who was also
his chief assistant.

Plantin published L'Obel's work as well as the later writing of Dodoens. These
two men, along with Charles de L'Ecluse (who apparently used no cuts of the key
genera) were friends who worked so closely together that it is not possible to tell
which or who was originally responsible for material or figures in their publications.

In 1576 Plantin published Plantarum seu stirpium icones using L'Obel's name
although many of the cuts had also been used to illustrate the herbals of Dodoens
and de L'Ecluse. It is thought that it was this selection of woodcuts which
Johnson bought to use in his editions of Gerarde's herbal in 1633 and 1636 (Arber,
1938). Arber (1938) believes this to be the last time the Plantin collection of
blocks was used.

Jacques Dalechamps, whose Historia generalis plantarum appeared at Lyons in

EISENDRATH A STUDY OF THE ICONES 307

1587, and Jakob Tabernaemontanus, whose Ncnw Krcuterbuch was published in
Frankfurt a few years later, both seem somehow to have had access to the Plantin
"pool" of cuts, as is evident from the above tracing of relationships to the same
figures found in Dodoens and L'Obel; or they used, as did so many others, the
octavo blocks from Fuchs' 1545 edition. Arber (1938) says that Tabernae-
montanus made a collection of blocks for himself, and that they were acquired from
him by John Norton to be used in his 1597 edition (the first) of Gerarde's Herball.
According to Bartlett (1949) this was based upon a translation of Dodoens' second
herbal, begun by a Dr. Priest; Gerarde took over the work for Norton, the publisher,
after the death of Dr. Priest, so the material is hardly original. Gerarde is particu-
larly remembered as a horticulturalist, having supervised the care of several famous
English gardens and having himself cultivated for twenty years a fine garden of
his own in Holborn.

Although John Parkinson also had a famous London garden, and was known
as "Herbarist to Charles I", the examples of illustrations of my key genera found in
his T heat rum botanicum leave much to be desired.

Other Woodcuts in Herbals of Sixteenth and Seventeenth Centuries

The search for "icones" of the key genera takes us now to Padua, where in 163 8,
appeared a little book on the plants of Egypt. The woodcuts of Luffa used by
Vesling (De Plantis Aegyptiis, 163 8) are highly stylized and poorly printed, and
are not included as illustrations in Alpino's original (1592) edition of the book to
which Vesling is supposedly only adding "observations et notae". However, Sims,
who edited the 1814 volume of the Botanical Magazine, thought rather highly of
Vesling's cut. "Professor Cavanilles", Sims wrote in connection with Plate 1638,
"supposes that 'Momordica Luffa' of Linnaeus may probably belong to the same
genus (Luffa) . Indeed these two plants appear to be extremely similar, as we judge
from the excellent figures of Veslingius [sic], in his observations on Prosper Alpino".

Most of the seventeenth century botanical books, however, differ from Vesling
in that they reflect the stimulus which all the natural sciences received from the
discovery of the New World. The earliest of these that falls within the limits of
this paper appeared in Amsterdam in 1648, the Historia Nafuralis Brasiliae. The
author of the section entitled Historia Rerum Naturalium was Georg Marcgrave,
who had accompanied a Dutch expedition which conquered Brazil and took it
briefly from the Portuguese. According to Wm. Swainson (1834), "Marcgrave 's
work abounds with a vastness of new and original information, very different from
what was to be found in the crude and verbose compilations of this period." He
has left for us a small wood-cut of FeiiUea, and several other cuts which are
apparently of members of the Cucurbitaceae, but are too crude and rough to be
recognizable. The cut of Fevillea most certainly does not fit my definition of
"icones", but like those of many of the herbalists it has value because of the time
and place it was made. The same cut is found again in Piso's De Indiae ittrimquc,
Amsterdam, 1658, actually the second edition of the volume listed above. The
work of Marcgrave and Piso is so inextricably entangled that it is impossible to

308 ANNALS OF THE MISSOURI BOTANICAL GARDEN

know who is responsible for what. The editor of the 1648 edition gave credit to
both (after Marcgrave's death), but Piso took all the credit for the second edition
unto himself. The credit for the illustrations, however, is of dubious value, as
Nissen (1951) writes, in his inimitable Teutonic verbosity, that "the cuts are in
no way suitable to verify a doubt-free determination of a specific object".

The decadence into which the botanical woodcut sank after reaching a high in
Fuchs' herbal has been traced through a devious history; with the introduction of
metal plate incision comes another breath of good fresh air. This introduction
presents a second of the strange anachronisms in the history of botanical illustra-
tion, comparable to the generally poor plant portrayals in the printed herbals during
a period when beautiful flower paintings were being made.

The technique of engraving on metal stems from the great goldsmith tradition
of the artisans and was translated to the graphic arts sufficiently early to have
reached a high point in the works of Martin Schoengauer, 1445-91. However, the
earliest strictly botanical book com i anted from metal plates

did not appear until a century after Schoengauer's death (Van Schaack, 1959).
This was Fabius Columna's Phytobasanos; it contained no pictures of my selected
Cucurbitaceae, nor have I found such pictures in books printed until a hundred
years later still. This fact, however, gives a falsely exaggerated impression of
delayed adaptation of the new technique to botanical illustration. After the
publication of Columna's etchings, the makers of botanical prints finally recognized
that the incision of metal plates permits a far more delicate line than does wood-
cutting, and the genus flourished through the seventeenth century.

Between 1678 and 1703 appeared the 12 folio volumes of Hortm mdictu jhuLi-
baricus, copiously illustrated with double page copper engravings, by Heinrich
Adrian van Rheede tot Draakestein, and containing "icones" of Cucurbita mos-
chata, Lagenaria vulgaris (see plate XIV), Luff a acutangula and L. cyliudnca.
Malabar lies on the west coast of southern India, and Rheede was governor of the
province. Blunt (1951) feels that the author may himself have had a part in
making the drawings of plants, though the plates were engraved by an Italian
monk, Father Mathieu. The illustrations are sometimes of immature plants and
sometimes (like those of Weiditz) picture in close detail the flaws of individual
specimens, but their large size and strong calligraphic quality give them a vitality
which many other plant illustrations lack. This is one of the best known of the
pre-Linnean botanical books which reflect the stimulus of the introduction into
Europe of new plants found during voyages of exploration or scientific inquiry.
The works of Commelin, Rumpf and Plumier (see below) fit into this same
category, Rheede, Commelin and Rumpf particularly pointing up, as Blunt (1951)
reminds us, "the interest shown by the Dutch in the flora of their colonial

Commelin's book, Horti medici Amstelaedamensis, con
plants in this Physic Garden, with illustrations drawn frc
Luff a operculata, typical of the stiff engravings taken f

EISENDRATH A STUDY OF THE ICONES 309

and Maria Moninckx, to which originals Blunt (1951) claims that justice is not
done. Caspar Commelin, the author, was a professor of botany and the nephew of
Jan Commelin, who had been director of the Physic Garden.

The next illustrations of the key genera appear in the 6 volumes of Georg
Eberhard Rumpf 's Herbarium Amboinense published, again in Amsterdam, between
1741 and 1750. The story of the author's life is a series of mishaps so catastrophic
that one can only wonder that his work was ever published. Rumpf became an
employee of the Dutch East India Company in 1652, and settled at Amboina 1 "'
where he indulged his love of nature until he died in 1702. However, fire, ship-
wreck and blindness all tested Rumpf's patience and courage, so that one cringes
from criticism of the illustrations of Cucurbita (C. Pepo and C. moschata),
Lagenaria and Luff a (L. acutangula and L. cylindrica) in his book. Let me only
say that their quality is extremely variable, due to the fact that Rumpf's drawings
from which his descriptions were made, were burned and he became blind before
he could make more. Hence several people, including his son, made the new set,
in many of which the botanical detail is poor. However, the books remain im-
portant because later authors have used his descriptions and figures as the types of
many binomials of plants of the Malay flora (Merrill, 1917).

Plantarn m americanarum listed hereafter as containing "icones" was published
in Amsterdam between 1755 and 1760, and was edited by Burmann, a well-known
professor of botany in that city. Burmann's "last labor" {General Biographical
Dictionary, 1813) was to procure "engravings to be executed for the drawings of
American plants left by Plumier, to which he added descriptions". Plumier, a
member of a French religious order, had been sent to explore the French settlements
in the West Indies, and succeeded so well that he was subsequently appointed
botanist to the king {General Biographical Dictionary, 1813). He published many
of his drawings in Paris in 1693, but these did not include the Fevilleas that give
him entree to the list. His plant figures were drawn to large scale in bare, cold
outline, and have little artistic merit; but they were often referred to by Linnaeus,
who had access to a set of copies at the University of Groningen (Sitwell and
Blunt, 1956).

The next book {Histoire universale du regne vegetal by Pierre Joseph Buch'oz,
1775-80) containing an illustration {Cucurbita Pepo) that falls within the limits
of this paper is the work of a botanical artist of mixed repute. Buch'oz' works
are listed by Sitwell and Blunt, by Dunthorne and in most of the catalogues of
exhibits of fine flower books, but the opinion in which he was held by contemporary
botanists needs no elaboration beyond the specific epithet "foetida" which L'Heri-
tier applied to a plant of the genus Buchozia. Nissen (1951) quotes Pritzel in a
brief statement which shows no evidence of Buch'oz' reputation having soared
posthumously. "Miserimus compilator", Pritzel called him, "fraude ac ignorantia
aeque emineus". Pierre Joseph Buch'oz is credited with having written on all
branches of natural science though understanding none, and with compiling some
300 volumes though remaining undistinguished himself! But his illustrations are very
attractive indeed and Blunt (1951) grants him a few kind words because he intro-

310 ANNALS OF THE MISSOURI BOTANICAL GARDEN

duced into floral illustration "the decorative qualities which we always associate
with Far Eastern art".

Antonio Jose Cavanilles is quite something else again. Of great botanical inter-
est, his 6 volume Iconcs et urn, concerning the plants that grew
in Spain, either "spontaneously or cultivated in gardens", during his time, contain
rather stiff engravings with no oriental charm whatsoever. These books were
published at Madrid between 1791 and 1801, and among them is a clean cut of a
Luff a acutangula.

The engravings of Rheede, Rumpf, Plumier, Buch'oz and Cavanilles were all
printed and published in fine, folio volumes, representing "the apogee of the art of
botanical illustration, reached in the 18th century" (Mongan, 1952) . Next on my
list is a group of smaller volumes, all published as handbooks to mec

Influence of the Linnaean Sys*

Christian Schkuhr published at Wittenberg, in 1791, 6 octavo volumes of his
Botanisches Handbuch, the last 3 volumes consisting of an atlas of colored (in the
Missouri Bot. Gard. copy) engravings. Nobody but Nissen and I seem to consider
him worth mentioning, but I find the engravings both charming and carefully done,
and Nissen gives Schkuhr credit for both the drawings and the plates.

Schkuhr's Handbuch, post-Linnean, is the first of my list to exemplify the new
spirit which entered the study of botany after the publication of the Species
Plantarum in 1753. Linnaeus' binomial system stimulated revision of older means
of plant description and classification, and "called for a new and different type of
illustration which emphasized scientific accuracy in the delineation of flower, fruit,
leaf and stalk" (Dunthorne, 1938). The wealth of diagnostic detail included in
Schkuhr's plate of Cucurbita Pepo is an excellent example of this new and different
type, and also illustrates Dunthorne's point that "the new scientific artists first
appeared in Germany". The particular care taken in dissections of the androecium
and gynoecium were of course important, because Linnaeus' 16 sexual system was
based on the number of parts included in each of these floral organs.

Mr. Nissen and I are joined by Sitwell and Blunt, and Dunthorne in listing the
next entry, Johannes Zorn's 1 cones Pi ■;/, which also appears in

Dutch translation by Dheoderich Leonard Oskamp as Afbeeldingin der Artseny-
geuassen. The colored engravings of Lagenaria are identical in the two editions,
and are quite charming.

The next series chronologically on my list is mentioned by Nissen and Sitwell

16 Linnaeus' sexual system was introduced in his Systema Naturae of 1735; the genera were grouped

Pistils. Cucurbita comes under Class 21, Monoecia, dcscr . fcmirw in

eadem domo, sed diverso thalamo. Flores masculi et feminei in eadem planta sunt." Besides being

. ncuerunt. Sta-

EISENDRATH A STUDY OF THE ICONES 311

and Blunt. Published by Ferdinand Bernhard Vietz, 10 quarto volumes including
1086 hand-colored engraved plates, Icones plantarum medico-occonotnico-techno-
logicarum, appeared at Vienna between 1800 and 1820. The plate of Cucurbita
Pepo is large, printed on a folded double page, and has a great deal of charm. A
Lagenaria plate is also ascribed to Vietz, but it is not included among the volumes
(1-5) in the Missouri Botanical Garden library.

I bring in Curtis' Botanical Magazine now because its earliest plate of one of
the key genera fits chronologically into the list of "icones" at this point. Actually,

ruptedly until the present day, practically all of its plates being hand-colored until
as late as 1948!

William Curtis, Praefectus Horti and Demonstrator to the Society of Apothe-
caries at Chelsea, founded the magazine to describe and illustrate "the most
Ornamental FOREIGN PLANTS", the illustrations to be "always drawn from the
living plant". When Curtis died the editorship was taken over by his friend,
J. Sims. Sims' plate 1638, published in 1814, gives (as mentioned above)
Gavanilles as a reference for his Luff a acutangula, a lovely copper engraving, prob-
ably by Sansom, from a drawing by Sydenham Edwards. Edwards is described by
Blunt (1951) as "Curtis' own creation", as it was Curtis who recognized talent
in the son of a Welsh schoolmaster, and brought him to London for instruction.
Edwards remained practically the sole illustrator of the Botanical Magazine from
the publication of the second volume in 1788 until 1815, some years after Curtis'
death; he is credited with about 1650 plates!

The only other of the key genera illustrated among the nearly 10,000 plates in
the Botanical Magazine is a Fevillea cordifolia, plate 63 56, published in 1878. It
presents an interesting contrast to the Luffa, as this was a period of low ebb in the
printing arts; it is lithographed on poor quality paper, and has none of the charm
or calligraphic quality of either earlier or later plates. At this time J. D. Hooker
was editor, and it was under him that "the last traces of the Linnaean System
vanished from the classification of plants in the Botanical Magazine" (Blunt,
1951). Sir Joseph Hooker is best known as an early champion of Darwin and for
his botanical explorations, which contributed greatly to the knowledge of plant
geography.

W. Fitch, artist of the Fevillea plate, was "discovered" by William Hooker,
father of Sir Joseph, while the to-be-Director of Kew Gardens was a professor of
botany and the to-be-artist an apprentice designer of calico, both in Glasgow
(Blunt, 1951). Sir Joseph Hooker, after inheriting this enormously prolific drafts-
man, described him as an "incomparable botanical artist", with "unrivaled skill in
seizing the natural character of a plant". Blunt (1951) finds his work too facile
and insensitive, but adds that Fitch "remains the most outstanding botanical artist
of his day in Europe", despite the fact that his drawings were made from dried
plants. Blunt gives him additional credit as the first artist to be able to do this
satisfactorily. Otto Stapf, editor of vols. 148-156 of the Botanical Magazine, has
given a satisfying explanation of why this was possible:

312 ANNALS OF THE MISSOURI BOTANICAL GARDEN

certain generalization' in which the type of the species came to life and took the place of a

This statement concerning Fitch's method 17 is of interest in relation to my
earlier discussion of "icones" as essentially portraits of species rather than of
individuals as were the woodcuts of Weiditz.

Francois Pierre Chaumeton, "docteur en medecine", well understood the value
of fine botanical illustration. "There are few sciences", he wrote in the preface to
his Flore Medicale, 1814-20, "which demand more imperiously than botany, the
help of painting. One would have to try very hard indeed to replace it with the
most precise description". A doctor with such an opinion would also try very hard
indeed to illustrate his book with the most precise drawings, and Chaumeton did
just that. His two pictures of Cucurbita Pepo were made by P. J. F. Turpin, one
of the great botanical draughtsmen of the day. Turpin worked with botanists of
the highest caliber, Poiteau and de Candolle, as well as illustrating many of the
American plants pictured in Humboldt, Bonpland and Kunth's publication con-
cerning the expedition to the "regions equinoxiales du Nouveau Continent, 1799-
1804". 18

Turpin is described as being largely self-taught, but Blunt (1951) sees in his
work "great indebtedness to the tradition of van Spaendonck and Redoute", both
artists from whom no follower need feel shame to borrow. Elizabeth Mongan
(1952) gives an excellent explanation for this:

Two years after Chaumeton began publishing the various editions of his work
(folio, quarto and octavo), publication began, under the general direction of
Cuvier, 20 of the Dictionnaire des Sciences Naturelles. This was a most ambitious
work, appearing in Paris between 1816-30 in 60 volumes of text, with 14 of atlas,
and planned (according to the title page) to "treat methodically the various
natural beings, from the point of view of our knowledge of them and also relative

, D S b S' P t

EISENDRATH A STUDY OF THE ICONES 313

to their use in medicine, agriculture, business and the arts". Dejussieu 21 was
originally in charge of the botanical section; Turpin of the execution of the draw-
ings and the general direction of the engraving. Unfortunately Turpin made only
two plates among the key genera, both of Fevillea cordi folia (see plates XII and
XIII). Turpin's drawings of botanical detail are generally acknowledged to be
exceptionally fine, and these plates well illustrate the fact. One of the plates, for

it from below: the androecium; a single stamen; and a view of // from the rear.
Equally detailed dissections are made of a "fertile" (female) flower; and, need I
add, the habit of the plant is also shown, all in one beautiful, colored, octavo page.
In 1821, M. E. Descourtilz published in Paris his 8 volumes on the Flore Uiii-
cale des Antilles, an assemblage made while he, a doctor, served in official capacity
at Santo Domingo. Descourtilz used illustrations of several of the key genera
(Cucurbita moschata, C. Pepo— both of which are pictured twice under different
names; Fevillea cordifolia and Lagenaria vulgaris) ; these pictures have a good deal
of charm, but are poor specimens of botanical illustration, despite the fact that the
author claims to have drawn them from nature and to have watched his son cut
them down in size before painting them with the greatest care. Though the father
claims that his son Theodore was a pupil of the "celebre Wanspaendonck", the
slipshod attempt at portrayal of detail is only a sort of casual acknowledgment of
its desirability. In contrast to the works of Turpin, these are sad indeed.

Lithographs

The eleven large folio volumes of Velloso's Florae fluminensis icones, published
in Paris in 1827-32, illustrate the flora of the region around Rio de Janeiro. Vel-
loso was a native Brazilian who became one of the most zealous students of
Domenico Vandelli, the Italian botanical explorer, and is thought to have supplied
most of the plants described in Vandelli's works (Nissen, 1951).

Velloso must have been an ardent botanist to collect so many plants, but the
art work in his book leaves much to be desired. The de Candolles (1878) wrote
that "the figures [of Florae fluminensis] like those of Plumier, are very bad. The
safest thing is to class the species of these authors ... in a very doubtful category,
which one can call inextricables". The figures do, however, play some part in the
history of the graphic arts as they were lithographed by Senefelder, who discovered
the process "almost accidentally" (Blunt, 1951) in 1797. The advantage of
lithography, the technique employed in all the remaining "icones" on my list except
those of Iinuma and Bailey, is that it permits subtle gradations of line; as is readily
seen in his Fevillea trilobata, Luffa cylindrica and two Lagenarias, however, Velloso
but drew in bold outline only.
;o illustrate Brazilian plants. Karl Friederich
Philipp von Martius, a German botanist and traveler, was sent to Brazil in 1817 by
the king of Bavaria. He returned to positions at the botanical garden and the
university in Munich, where he began work on his Flora Brasiliensis in 1840.

314 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Fifteen volumes in 40 folio parts were published between that date and 1906 in
Munich, Vienna and Leipzig, the later volumes, of course, edited by younger men.
The 3000 lithographs in the series were made by a number of artists, and include
a Fevillea trilobata and a Luff a operculata, both of which are drawn in unusually
fine detail.

Robert Wight's Icones plantarum Indiae orientalis, 6 quarto volumes of litho-
graphs published in Madras between 1840 and 1856, contains some of the botan-
ically finest "icones" on my list. The two-page spread of Cucurbita maxima
includes an extraordinarily complete floral analysis; the Luff a cylindrica is done in
such detail that one sees even the fine webbing at the edge of the leaf. The two
plates, drawn by the Indian artist Rungiah and lithographed by Dumphy, are
typical of the work of several generations of talented native artists who worked in
India at about this time (Nissen, 1951). Wight, however, was not too happy
about the way they were printed, especially in the early volumes. He apologizes in
his preface with the explanation that lithography, still in its infancy in Europe, was
even more of a fledgling in Madras, where inexperienced printers had not yet learned
to compensate for the atmospheric difficulties which affected the success or failure
of the plates. Hooker and Thomson (1855) in their Introductory Essay to the
Flora Indica call Wight's works "the most important contributions not only to
Botany, but to Natural Science, which have ever been published in India".

Later Work in Various Techniques

When discussing the botanical woodcuts of the 16th and 17th centuries, I
expressed admiration for the manner in which Brunfels' and Fuchs' artists handled
them, but welcomed warmly the increased grace and delicacy made possible by the
technique of engraving. Now, in the last quarter of the 19th century, we find
a series of woodcuts so beautiful and finely made that one is forced to realize (as
one always is!) that a great artist in any field surmounts the limitations of his
medium. These effusions are aroused by Yokusai Iinuma's Somoku-dzusetsu or
"Illustrated Flora of Japan". The set at the Missouri Botanical Garden is the
second edition, printed at Ogaki in 1874 on rice paper of the highest quality. The
set consists of 20 quarto volumes in which the illustrations (Cucurbita Pepo, 2
Lagenaria vulgaris and Luff a cylindrica) of my key genera extend across two pages.

In discussing the beautiful woodcuts of Fuchs, I pointed out that, in black and
white, they sometimes appeared to be drawn in too thin outline, having obviously
been designed with the intent that they be colored by hand. In these Japanese cuts
the opposite is often true; the body of the figure is inked and the line, remaining
white, is so fine as to seem almost impossible of achievement. In several of the
plates I cite, the enlargements of the anther column and gynoecium are printed in
color. Nissen, the verbose and usually all-seeing, only notes the Somoku-dzusetsu
in a list for which he expresses gratitude to others; and, to my surprise, Blunt
(1951) says only that it is illustrated with "effective woodcuts".

Manuel Blanco, in the third edition of whose Flora de Filipinas we find a
Cucurbita Pepo and two Luff as (L. acutangula and L. cylindrica) , was an Augus-
tinian monk with no training whatsoever in botany, and a botanical library con-

EISENDRATH A STUDY OF THE ICONES 315

sisting only of one volume of Linnaeus. Hooker and Thomson (1855) in their
Introductory Essay call his first edition of 1837 a "botanical curiosity", a state-
ment which covers its reception in Europe, though it met with great local acclaim.
Neither the first nor the second edition was illustrated. The third is the work of
Fathers Celestino Fernandez Villar and Andres Naves, whose only contribution,
according to Merrill (1918), is their Latin translations which made Blanco's Spanish
descriptions more generally accessible. The identifications of the plates (drawn
by F. Domingo and lithographed by C. Verdaguer) show many errors, a large
number of them referring to species which Blanco had never seen. None of the
plates of the key genera falls within this category, however. Nissen (1951) goes
so far as to damn the illustrations of this deluxe edition by praise so faint it hardly
fits the definition. Neither the drawings nor their lithographic reproductions, he
felt, compare with the quality of "the old work", but must be compared with their
contemporaries, all, by inference, very bad indeed.

Field and Garden Crops of Duthie and Fuller, published in Roorkee in 1882-93,
is not mentioned by Blunt or Nissen, nor do its illustrations appear in any of the
exhibition catalogues, but the lithographs of Cucurbita moschata, Lagenaria
vulgaris and Luff a cylindrica, albeit utterly lacking in artistic pretension or value,
merit place in my list through their other qualifications. The artist of the plates,
H. Hormusji, a Parsee imported from the Bombay School of Art, drew the botanical
details with great care, and, though they are not pretty, his pictures are surely
portraits of living plants. J. F. Duthie was Superintendent of the Saharanpur
Botanical Gardens; J. B. Fuller, Director of Agriculture and Commerce for the
Northwestern Provinces and Oudh.

Carlos Bettfreund's 3 octavo volumes, Flora Argentina, published in Buenos
Aires between 1898 and 1901, deal with the "plantas vivas" of Argentina, and are
illustrated in the manner of the period, with poorly colored lithographs of no
artistic, but presumably some botanical value. These include a Cucurbita Pepo
and a Lagenaria vulgaris. The same criticism may be leveled at F. Gagnepain's
section on the Cucurbitaceae in Lecomte's Flore Generale de I'lndo-Chine, published
in Paris in 1921. This contains a Luff a cylindrica drawn, according to the plate,
by Miles. Vesque. I find no mention of this name elsewhere, presumably because
the quality of work does not redound to the glory of the artists.

My criteria for "icones" have had to be as much contorted to include the last
two publications, as they were also to stretch across the post-Fuchsian herbals.
Now they can return to their original shape, though they must, perhaps, still retain
a fair amount of elasticity. Engler's Das Fffanzenreich would not be found on
library shelves under the heading of "icones", nor would its many volumes find
their way into the collection of a bibliophile. They do contain, however, besides
the most universally accepted system of plant classification, a number of fine draw-
ings of plants. Of the key genera, only a Fevillea trilobata appears on a full page
and includes the habit of the plant, but I am grateful for it, because it is the best
of its type of illustration, apparently the closest we will come, in our time, to
fulfilling the criteria for "icones". The Fevillea plate contains precise and accurate
botanical detail and, though not showy in the slightest lampshadish way, gives a

316 ANNALS OF THE MISSOURI BOTANICAL GARDEN

very pleasant impression of being a portrait of a living plant. This opinion is
blessed by Blunt (1951), who points out that many of the illustrations used by
Engler, the "greatest of German taxonomists" were unsigned works of Joseph Pohl,
whose drawings were "conscientious, accurate, useful and in great quantity, but
not on a high artistic level". Of even more significance in these, our troubled
times, is the fact that they were made from specimens in the Berlin Herbarium,
subsequently destroyed during World War II.

A full-page illustration of Fevillea cordifolia in Rendle's Classification of Flow-
ering Plants squeezes in among the "icones" by the same reasoning as does Engler's
F. trilobata. The picture is in Rendle's second volume, which did not appear until
1925, although the first was published in 1904, at which time the material for the
second was "almost finished" (Burkill, 1938). It is impossible, therefore, to know
if I have placed the book in proper chronological sequence. For our purposes,
however, it is only important that we have here again a carefully drawn plant
portrait, containing details acceptable to a distinguished taxonomist. Rendle joined
the Department of Botany at the British Museum at the age of twenty-three, in
1888, and remained there for fifty years, pursuing, according to his "keen concep-
tion of his duty", his work in systematics. Burkill (1938) further describes him as
one of those "ships which, as they leave port, have their course set for the whole
voyage". He held many botanical posts of honor and responsibility, but his interest
was always directed toward taxonomy.

L. H. Bailey's Garden of Gourds, a quarto volume issued in New York in 1937,
is illustrated with a number of zincograph black and white plates mechanically
reproduced from pen and ink drawings. Again, they are of no artistic value, but
are pleasing and contain fine botanical detail acceptable to the author, an acknowl-
edged authority on cultivated plants. Of the many illustrations of key genera
printed in Bailey's book, I have listed as "icones" only those which best give an
impression of habit: "Wild Texas Gourd", "Egg Gourd", "Apple Gourd", "Big
Bell Gourd" and "White Pear Gourd", all belonging to the species Cucurbita Pepo,
var. ovifera; Lagenaria vulgaris or Luff a cylindrica.

Conclusion

Bailey's book contains the last of my "icones", representing well a type of
illustration with which it seems that we must learn in these latter days to satisfy
ourselves. Botanically they are, of course, authoritative, but they grade very low
on the artistic scale, as do all the last group listed. Actually there have been very
few on my list that rate high both artistically and botanically, but it must be
remembered that this is probably because the key plants do not appeal as floral
decor, and have not posed for most of the best-known botanical artists.

Of the publications listed, only Fuchs in the 16th century and Rheede in the
17th contain "icones" that completely fulfill the requirements of my definition —
size, habit, botanical accuracy and artistic value; Rumpf, Plumier, Buch'oz and
Cavanilles in the 18 th century meet the criteria in varying degree, all being at the
very least large, fine books from the bibliophile's point of view.

Schkuhr, Zorn, Vietz, Curtis' Botanical Magazine, Chaumeton, the Diction-

EISENDRATH A STUDY OF THE ICONES 317

naire des Sciences Naturelles and Descourtilz are less pretentious publications of the
late 18th and early 19th centuries, with the happy artistic addition of color, again
complying with my requirements to some degree.

Velloso and Martius return to the grand scale, but are more pretentious than
beautiful, and poor Velloso is also not held in very high botanical esteem. From
these two on, except for the Japanese work of Iinuma, the artistic measure must be
dropped entirely, or my list would have to end. I cannot describe Bailey's book
(more attractive at least than the others) by even so slight a word of praise as
charming, and there is no more temptation to use his illustrations than Engler's or
Rendle's on my lamps. This may be a good thing; if the paper is of decent quality,
the bookworms stay away, and if anybody supports our libraries, the books may be
preserved, as books, for posterity. Though beauty is no longer important in these
publications, botanical accuracy is a sine qua non; for this reason and with the
message of Dr. Trelease still in mind, we can feel sure that they will serve the
purposes of mankind for a very long time to come.

For the sake of consistency in the following indices I have employed the spelling
of authors' names used by Nissen (1951), because his bibliography is the most
complete I have found. This has led to one or two unfamiliar spellings, as "Karl
von Linne" and "Rumpf". Consequently, I have used the name "Linnaeus" in the
text, as it is the familiar form in American botany.

Literature Cited

v. ed. New York, 1949.
Bartlett, H. H irom the botanical library of Mrs. Roy i

(Clements Library, Bulletin LVII) Ann A

plates) (Nissen 170) Colored lithographs.

ich'oz, Pierre Joseph. Histoire universale du regne vegetal. Paris, 1 77 5-1
12 v. of plates (v. 1-9, 11, 12 of plates only at Missouri Bot. Gard.)
287; Dunthorne 59) Engravings.

indolle,' Alphonse de, and Candolle, Casimir de, editors. Monographiae ph

318 ANNALS OF THE MISSOURI BOTANICAL GARDEN

. II. Leipzig, 1924. Lithographs.

ae in Engler, A., Pflanzenreich IV. :
e-Cucumerinae in Engler, A., Pflanzenre

60 v. of text and 15 v. of atlas. (Nisscn 2239) Colo
:werp, 1563. (Pritzel 2345; Nissen 512) Small woodcuts,

Encyclopedic methodique: Botanique. By J
J. L. M. Poiret. Paris, 1789-1823. 13 i
Engler, A. See under Cogniaux, A.

[ndo-Chine. Edit.

element 1938-5U.

irarde, John. The herball, or generall historie of plantes. London, 1597. (Prit:

eenough, J. B. (and others). Allen and Greenough's New Latin grammar. Bosto
x>ker, J. D. The flora of British India. London, 1875-97. 7 v.

mma, Yokusai. Some J flora of Japan"). 2d ed. Ogali

i for the Royal Horticultura

M. Taxonomy of vase

See under Flore genera!

Linne, Karl von. Hortus Cliffortianus. Amsterdam, 17:

appendix by W. T. Steam 'and J. L. Heller. Londo.

•L'Obd, Matthu Antwerp, 1591.

correctly attributed to (

1218) Small woodcuts.
MacCurdy, E. The notebooks of Leonardo da Vinci. N<

EISENDRATH A STUDY OF THE ICONES

ttioli, Pier Andrea. Commentarii secundo aucti in libros sex Pedacii Dioscoi

Odell.J. W. Gourds and.
Oskamp.D. L. See under
Oxford English dictionary

*Piso, Willem and ]

ndle, A. B. The classification of flowering plants. Cub, J,..
eede tot Draakestein, Hendrik Adrian van. Hortus Indicus

mpf, Georg Eberhard. Herbarn

C flower books, 1700-1900. Londt

Tamaulipas. 'Amer.' Antiq. 22:352-358. 1957.
Wight, Robert. Icones plantarum Indiae orientalis. Madras, 1840-53. 6 v. (Pritzel 10246; Nissen

[Zorn, Johannes]. Afbeeldingen der artseny-gewassen. Translated into Dutch by D. L. Oskamp.
Amsterdam, 1796-1S00. 6 v. (Nissen 2203) Colored engravings.

320 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Chronological Index of "Icones" of Key Species

Cucurbita foetidissima:

^ucurbita maxima:

1576, L'Obd: "Pepo maximus Indicus compressus",j

159l' L'Obel: "Pepo maximus Indicus compressus", ]
1616, Dodoens: "Pepo rotundus maior", p. 666.

1840, Wight: C. maxima, vol. II, pi. 507 (double pa

nd "Pepo rotundus compi
", vol. I, p. 616; "Cucurb

n, Bettfreund: "Cue

EISENDRATH A STUDY OF THE ICONES

Melopepon lati ,,1. II, p . 224.

Feviltea cordifolia:

1755, Plumier: "Fevillea foliis cordatis", p. 203, pi. 209.

1816, Dictionnaire des Sciences Naturelles: "Fevillea hederacea, Poir.", vol. L

1925, RendU ,!. H, p. 222.

Jillea trllobata:

1658, Piso: "Nhandiroba", Cap. LXVI, ;

. vol. IV, .

Dodoens: "Cucurbita prior", p. 668; "Cucurbita longior" a

» lagen;
-Schora", vol. VIII, pi. 1; "C ~
vol. VIII, pi. 5.
U, Rumpf: "Cucurbita lagenaria", vol. V, p. 398, pi.

7, Velloso: "Cucurbita lagenaria", vol. X, pi. 98; "Cu
4, Iinuma: "Lagenaria Dasistemon Miq.", vol. XX, f
XX, pi. 45.

Duthie'and Fuller: I. acutan^. -...

322 ANNALS OF THE MISSOURI BOTANICAL

., Ser.", vol. XX, pi. 36.
", atlas II, pi. 334.
.uffa Aegyptiaca", vol. 1

, Lecomte: L. n

, linger: /. cylinJricu. vol. IV. .
. Balicv. /.. rylmdrict, p. 105.

Index, by Key Species, of Woodcuts Derived From One
Source and Used by Several Authors

Cucurbita maxima:
I. 1576, L'Obel: "Pepo maximus Indicus compressus"; 7X13 cms.

jchamps: "Cucurbita Indica longa"; 12 X 6 \' z cm,.

eris sive Zucha flore luteo, Ic. II"; V/ 2 X^i
X 7Yi cms.
7Vi cms.

L'Obel: "Pepo oblongus
L'Obel: "Pepo rotundus

, Dalechamps: "<

EISENDRATH A STUDY OF THE ICONES

! Bauhin: "Pepo Indicus minor, Tabern."; 8»/ a X 4J4 cms.

, Tabernaemontanus: "Melopepoteres"; 12X7 cms.
, Gerarde: "Pepo maximus rotundus"; 12X7 cms.
, Tabernaemontanus: "Pepo maximus oblongus"; 11 1 / 2 X?Vi cms.

, Fuchs: "Cucurbita maior"; fo

lio page.

. Fuchs: 'Cucurbita maior"; oc

, Dodoens: "Cucurbita camerari

, Bauhin: "Cucurbita latior folio molli, flo

, Fuchs: "Cucurbita minor"; fo

, Dodoens: "Cucn.

L'Obel: "Cucurbita sive zucch

, Gerarde: "Cucu:

, Parkinson: "Cucurbita longa";

' L'Obel: as above.

Dodoens: "Cue

Parkinson: "Cucurbita lagenar

Tabernaemontanus: "Cucurbit

Gerarde: "Cucurbita anguina"

; 12 X7«:

Tabernaemontanus: "Cucurbit,

Gerarde: "Cucurbita lagenaria

sylvestris"

EISENDRATH-

ANNALS OF THE MISSOURI BOTANICAL GARDEN

BOTWIOl K

NANDTROBB a iouilles <le li<

„crvnRATH-A STUDY OF THE ICON]

EISENDRATH A STUDY OF THE ICONES 327

PLATE XIV
Hs), vol. VIII, pi. 5, Hart

SYNOPSIS OF POINSETTIA (EUPHORBIACEAE)

ROBERT L. DRESSLER

The name P. heterophylla is found to be applicable to a weedy tropical plant,
: only n = 14. The geographic variation of the North American' annual s

Grove Ave., St. Louis 10, Missouri.

There are not many species in the genus Poinsettia, but their bewildering foliar
polymorphism has caused a great deal of confusion. While studying the genus
Pedilanthus (Dressier 1957), some observations were made on other genera of the
tribe Euphorbieae. In the summer of 1957 I was able to observe most of the North
American species of Poinsettia in Mexico and the southwestern United States. •
Many collections made at that time have been grown and studied at the Missouri
Botanical Garden. These observations have permitted considerable clarification of
the annual species, especially. While I am no longer actively studying the Euphor-
bieae, it is felt that these notes on Poinsettia may be of some value.

Generic Concept
The Linnaean genus Euphorbia, as it is frequently used, includes most of the
tribe Euphorbieae, and is surely the most broadly inclusive generic concept in
modern usage (among flowering plants). I much prefer to recognize about four-
teen genera in the tribe Euphorbieae, six of these being included in the genus
Euphorbia by many authors. I have no quarrel with those who prefer the inclusive
concept of Euphorbia, but feel that the same arguments could be used to reduce
most Compositae to Aster, with about equal profit. In the narrower sense, Eu-
phorbia is a distinctive group of about 80 Old World species which are woody or
succulent, and usually have two spines (rarely 1) near each leaf base. These spines
are frequently considered to be stipular, but true stipules appear to be present in
addition to the spines. The large group of south African succulents which are
spineless or have spines formed from axillary branches are not directly related to
the above group. The earliest generic name for the south African group is
Medusea Haw. Neither of these groups is represented in the New World, the
American succulents being members of Agaloma or Tithymalus. The largest and
most difficult of the segregate genera is Tithymalus, which is largely an Old World
group, but includes three major American groups: ( 1 ) Primitive woody species,
such as Euphorbia (Tithymalus) fulva Stapf and Tithymalus calyculatus (HBK.)
Kl. & Gke., of the West Indies, Central America and the Andes. (2) Herbaceous

• I am especially grateful to the Gray Herbarium and the National Science Foundation for their

330 ANNALS OF THE MISSOURI BOTANICAL GARDEN

perennials related to Tithymalus chrysophyllus Kl. & Gke. and Euphorbia (Tithy-
malus) ftoridana Chapm. This group includes two or three species of the south-
eastern United States and a number of southern South America, all of them with
a conspicuous pistillate calyx. (3) Herbaceous species which are closely related
to Old World sorts and are surely geologically late invaders of the New World.
These include Tithymalus commutatus (Engelm.) Kl. & Gke. and T. campester
(Cham. & Schlecht.) Kl. & Gke., and are largely montane forms. There are also
a few anomolous species which do not fall into these groups, such as Tithymalus
trichotomus (HBK.) Kl. & Gke. and Euphorbia (Tithymalus) pteroneura Berger.
Most New World Euphorbieae belong to the distinctive group made up of
Agaloma, Poinsettia and Chamaesyce. Of the three, Chamaesyce is most frequently
treated as a distinct genus, while Agaloma generally has been kept in Euphorbia.
These two groups are closely related, but can be distinguished by habit and by the
distinctive leaf structure of Chamaesyce. While Agaloma can almost always be
distinguished from Tithymalus by the appendaged glands, this feature is not in-
fallible. The diminutive Euphorbia misella Watson, of alpine meadows in central
and southern Mexico appears to lack gland appendages, but close scrutiny shows
rudimentary appendages beneath the glands, indicating that it is a reduced member
of Agaloma. Similarly, Euphorbia heterophylla var. eriocarpa Millsp. may belong
in Agaloma, but lacks the gland appendages characteristic of that group (see
excluded species, p. 340). Agaloma is greatly in need of monographic study, and
there would be very little profit in wholesale new combinations made without such
study. For this reason, I propose to cite the species of Agaloma as, for example,
"Euphorbia (Agaloma) corollata L.," with the understanding that this group is
quite distinct from true Euphorbia. The same system, of course, may be used for
Medusea, or even Tithymalus. A tentative key to the New World genera of
Euphorbieae is given below.

cMo^h^Tshea

jeast the

base of the plant monopodia!;

leaf ■

,e,ns without distinct

2

2. Glands of the ii

1V a 0l p U end e a C r ate " "** ""* ^ °"

o brer

,1 horns, but without

3. Glands of the

"

TlTHWMLsTrcW

4. The involu.

,:<nd.

a lateral

4. The involuvr-1 l.-K-s

"■^t:z,

l Xz<

sf^oSsrs

1 lobes eqi

,;;.:'

5. The involucra

1 ,inBle h00d " h A G s r r ;

ia Rat.

Generic Relationships
As noted above, Poinsettia is closely allied to Agaloma, though readily distin-
guished by the lack of petaloid appendages on the glands. Euphorbia (Agaloma)
lancifolia Schlecht. is especially similar to Poinsettia, and resembles that genus in

SYNOPSIS OF :

habit, in the tendency toward reduced gland number, in the form of the glands,
in the reduced petaloid appendages and in the thick-walled, coarsely reticulate pollen
grains. The seed of E. (A.) lancifolia is very different from that of Poinsettia,
which serves to reinforce the generic distinction. Several other species of Agaloma
have been confused with Poinscftia, though they are less closely related (see ex-
cluded species, p. 340) . Poinsettia is evidently derived from Agaloma by the loss of
petaloid appendages and reduction in number of glands. While the involucre
of Poinsettia generally has only a single gland, the central involucre of a pleio-
chasium, especially, is likely to have a full complement of five glands.

The Identity of Euphorbia heterophylla L.
Much of the confusion concerning the annual poinsettias has centered about the
identity of Linnaeus' Euphorbia heterophylla. The epithet seems very descriptive
of the familiar North American plant with red and green floral bracts, and most
botanists seem to have assumed that it applied to this plant. One finds, though,
that Linnaeus' epithet refers to the shape of the leaves, rather than to their color.

332 ANNALS OF THE MISSOURI BOTANICAL GARDEN

There is no specimen of E. heterophylla in the Linnaean herbarium, so one must turn
to Tithymalus curassavicus, salicis 2f atriplicis foliis varius, caulibus viridantibm
of Plukenet's Almagestum Botanicum (1696). Plukenet's figure shows a plant
with both lanceolate and pandurate leaves, quite characteristic of the weedy tropical
species with green floral bracts. Leaf shape is none too dependable in this group,
but the entire lanceolate form shown is rarely encountered in the North American
P. cyathophora. It is conceivable, of course, that the North American ornamental
had been introduced into Curacao by that date, but it is hardly likely that Plukenet
would have failed to mention the red floral bracts of that species.

It should not be concluded that Linnaeus' epithet has been consistently and
discerningly misapplied to the North American species. Rather, most specimens of
either species from North and Central America have been determined as E. hetero-
phylla, while South American specimens of true E. heterophylla have more often
been labeled as E. geniculata or E. elliptica. The consistent confusion of these two
species is based on very superficial observation, for the two differ from each other
much more than either differs from P. dentata, which is generally recognized as
distinct. A tabular comparison of the two species is given herewith.

Poinsett:

a cyathophora

Poinsettia heterophylla

Stems and petioles glabrous or
sparsely pilose.

Stems and petioles often markedly

Leaves glossy green.

Leaves dull green.

Floral bracts usu

lly basally red.

Floral bracts green or basally pale,
never basally red (often purple

Gland broad, more or less bilabiate,
the opening narrowly oblong.

opelgcir'cTa'r. "*"****> ^

Seed finely and si

arply tuberculate,

Seed coarsely and bluntly tubercu-

Chromosome Number

A number of chromosome counts have been made from material cultivated at
souri Botanical Garden. These are given below, with the counts previously
lished by Perry (1943). Voucher specimens are available for all those which

Poinsettia cyathophora

Dressier 1821 (Tamaulipas)
Dressier 2271 (Tamaulipas)
Dressier 2430 (Nuevo Leon)
s. n. "Washington, Missouri
s. n. Grey Summit, Missouri
Dressier 2373 (Tamaulipas)
Cultivated (Rex Pearce)
Cultivated (Parke)

s Euphorbia heterophylla)

DRESSLER A SYNOPSIS OF POINSETTIA

Dressier 2198 (Texas)

s. n. Empalme Escobedo, Guanajuato

28

Perry, Clarke Co., Virginia
Davis, Oklahoma
Pleasant Hill, Missouri

14, 28
28
56

Poinsettia heterophylla

Dressier 2136 (Tamaulipas)
Dressier 2194 (Arizona)
Dressier 2243 (Vera Cruz)
Howard & Proctor 13 883 (Jamaica)
Manning & Manning 531217 (Jalisco)
s. n. Soledad, Cuba

Gametic

Semitic
No.

Moyer (cited by Perry) as E. geniculata

28

Poinsettia pulcherrima

No.
28

From the above counts, one could consider 14 to
genus, if it were not for Perry's record of a somatic r
of P. dentata. It is at least conceivable that the pla
chance haploid, but more counts are needed from the e

je the base
umber of
nt studied
astern Uni

number for the
14 for one plant

ed States, before

this can be evaluated properly. There is no indication of polyploidy in P. hetero-
phylla, but P. cyathophora and P. dentata both contain polyploids. These are
discussed further in the next section.

Patterns of Variation

The most outstanding feature of the annual species of Poinsettia is their extreme
polymorphism. Linear-lanceolate, ovate and deeply pandurate leaves may be found
not only in the same population, but on the same individual. More frequently,
however, extremely different leaf types are found on different individuals or on
different stages of one individual. It would appear that there is an interaction of
genetic and environmental control in such cases. The perennial species are less
markedly polymorphic, but all of the adequately sampled species show some degree
of foliar polymorphism. Even seed structure shows a good deal of variation in
Poinsettia, though this feature is usually dependably stable in the Euphorbieae.

In spite of the extreme polymorphism, some patterns of geographic or genetic
variation are evident, and merit more detailed study.

1. Poinsettia cyathophora. This species is perhaps the most highly variable of
the genus. The cultivated polyploids (n = 28) are coarse, somewhat weedy plants
that show little polymorphism, and have somewhat coriaceous leaves. When col-
lected in a dry, rocky arroyo, the plants of Dressier 2373 were relatively small, but
when cultivated in St. Louis, they became large and coarse, very closely resembling
the cultivated plants of the same chromosome number. Many of the more delicate

334 ANNALS OF THE MISSOURI BOTANICAL GARDEN

plants of lower chromosome number would seem to have greater ornamental
potential than the polyploids now available.

Collections from Oaxaca and Nuevo Leon have behaved as distinctly short-day
plants, while one collection from Tamaulipas (Rancho del Cielo) has proven to be
relatively day-neutral. The characteristic red bracts of this species are a surpris-
ingly dependable key feature, but some plants of the Missouri Ozarks lack the red
coloring.

2. Poinsettia dent at a. This species shows a good deal of geographic variation,
which is complicated by polyploidy. The plants with n = 14 of southern Mexico
are much branched and often decumbent, strongly pilose plants with very con-
densed, nearly capitate inflorescence and cream or yellowish floral bracts. This is
apparently the form described as Poinsettia schiedeana. These features become less
marked to the north, suggesting a clinal pattern. One also finds erect, rather weedy
plants with little or no bract coloration in Mexico. These are frequent in disturbed
areas of central Mexico, and closely resemble the forms of this species which grow
in Arizona and western Texas. It is probable that these weedy forms have entered
Mexico from the north and are genetically isolated from the native form by their
higher chromosome number.

Some of the semi-desert forms from northern Mexico are striking because of
their small, coriaceous leaves. The type of Engelmann's var. rigidia is not the

3. Poinsettia heterophylla. This tropical species shows less variation than its
temperate counterparts. Occasional populations with whitish floral bracts, strongly
dentate leaves or very pilose stems suggest introgression from P. dentata, and I
suspect that these two do hybridize to a certain degree where P. heterophylla grows
with a population of P. dentata having the same chromosome number.

Taxonomy
Poinsettia Graham, Edinb. New Philos. Jour. 20:412. 1836. Type: Euphorbia
Pulcherrima Willd.

Pleuradena Raf., Atl. Jour. 1(6):182. 1833, not Pleuradenia Raf., 1825. Type: Pleura-

dena coccinea Raf.
Cyathophora Raf., Fl. Tell. 4:117. 1838, not Cyathophora S. F. Gray, 1821. Type: Eu-
phorbia heterophylla L.

Woody or herbaceous, often perennial from a fleshy storage root; leaves alter-
nate or opposite, stipules inconspicuous or absent, appearing glandular; inflorescence
a condensed pleiochasium or dichasium, the branches becoming monochasial;
involucral glands more or less cup-like, usually reduced in number (commonly 1) ;
pollen grains thick- walled and coarsely reticulate; pistillate flower naked.

SYNOPSIS OF POINSETTIA

m. tall, flowering

rt leafless stems, the vegetative ste

~ ^eloping

'!•...'

vkh the outer rim flaring and

::,:;

wlut rct-.l...J

,-T " ' '

green

glands red, yellow or green,

the

s

. All leaves usually o
II. Seeds finely and I

1. Poinsettia pulcherrima (Willd.) Grah., Edinb. New Philos. Jour. 20:412.

Euphorbia pulcherrima Willd. ex Klotzsch, Allg. Gartenz. 2:27. 1834.

Pleuradena coccinea Raf., Atl. Jour. 1 (6) :182. 1833, not Poinsettia coccinea Dressier

Euphorbia erythrophylla Bertol., Nov. Comment. Acad. Scient. Inst. Bonon. 4:419, t. 41.

1840.
Euphorbia fastuosa Sesse & Mocifio, Pi. Nov. Hisp. 81. 1888.

Distribution. The familiar cultivated poinsettia occurs as a wild plant in the
rocky canyons of western Mexico, in Nayarit, Jalisco and Colima. Dr. Rzedowski
suggests (personal communication) that it may also occur further south, in
Michoacan and Guerrero. The habit of the plant, with a fleshy storage root
especially well developed in the seedling stage, is characteristic of the tropical
deciduous forest of this area.

2. Poinsettia radians (Benth.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253.

Euphorbia stormiae Croiz., Rev. Sudam. Bot. 6:13. 1939 [Type: Storm s. n. AH!].
Distribution. Arizona through central Mexico to Oaxaca.

z. t?$l7 (GH), Ctstmrn 1851 (GH), Meyer 3647 (GH)

336 ANNALS OF THE MISSO

3. Poinsettia restiacea (Benth.) comb. nov.
Euphorbia restiacea Benth., Bot. Voy. Sulph. 162. 1844.

Distribution. Western Mexico (Nayarit and Durango) .

Superficially the plants of P. restiacea resemble some species of Agaloma, such
as Euphorbia sphaerorbiza Benth., because of the slender stems, narrow white
bracts and conspicuous white glands. Though the outer rim of the gland is
flattened and somewhat petaloid, there is no sharp distinction between gland and
appendage, such as characterizes the members of Agaloma. The type specimen (K)
is very scrappy, but a photograph (MICH) indicates that it is probably the same
species as the excellent material which Dr. McVaugh has collected in Nayarit.

4. Poinsettia strigosa (Hook. & Arn.) Arthur, Torreya 11:260. 1912.

Euphorbia strigosa Hook. & Arn., Bot. Voy. Beech. 310. 1840.
Poinsettia pedunculata Klotzsch, Seem. Bot. Voy. Herald 277. 1856.
Distribution. Western Mexico (Nayarit and Jalisco).

5. Poinsettia colorata (Engelm.) comb. nov.

Euphorbia colorata Engelm., Rep. U. S. & Mex. Bound. Surv. 2:190. 1859 [Type: Thurber

rt&GHl].
Euphorbia tuberosa Rose, Contr. U. S. Nat. Herb. 1:111. 1891, not E. tuberosa L., 1753.

Distribution. Northwestern Mexico (Sinaloa, Sonora and Zacatecas) .

6. Poinsettia coccinea sp. nov. Figure 2

Perennis, erecta, 40-70 cm. alta, radice tuberosa oriens, caulibus basi lignes-
centibus, glabris; folia alterna, linearia vel anguste lanceolata, 1.5-6 mm. lata, 40-
120 mm. longa, subtus hispidula, ad basin barbellata, apice attenuata, basi attenuata
vel breviter petiolata, marginibus serratis vel subintegris; cymarum bracteae
coloratae, lineari-lanceolatae vel elliptico-lanceolatae, 3.5-7 mm. latae, 30-80 mm.
longae; inflorescentiae compactae; involucra anguste campanulata, 3-3.5 mm. lata,
3.5-4 mm. longa, glandulis involucrum excedentibus, 1.5-2 mm. latis, marginibus
externis crassioribus altioribusque; ovarium glabrum; styli ca. 1 mm. longi, basi
breve connati, fere ad medium bipartiti; semina oblongo-ovoidea, ecarunculata,
tuberculata, ca. 2.5 mm. longa.

es of thickened
Other collections from the
laparral), peak of Cerro de Las Yucas, July 20,

, but apparently this

This species is related to P. colorata, but may be distinguished by the usually
serrate leaves which are not markedly paler beneath and by the smaller green
(rather than red) cyathia; the gland also is smaller and differently shaped. Poin-
settia coccinea will more often be confused with the annual P. cyathophora. Under
favorable conditions P. cyathophora may survive for two or more seasons, but it is

DRESSLER A SYNOPSIS OF POINSETTIA

338 ANNALS OF THE MISSOURI BOTANICAL GARDEN

not known to form the fleshy "sausage-string" roots of P. coccinea. Crutchfield
tf Johnston 5611A, from near Jaumave, Tamaulipas, is definitely a perennial, but
appears to be P. cyatbophora (no roots were present on the specimen seen).

Poinsettia coccinea is frequent in rocky oak scrub and open oak forest in the
region of the type locality. In this area it is quite distinct from P. cyatbophora,
which was found only in the polyploid form (n = 28). To the north of the
Sierra de Tamaulipas other forms of P. cyatbophora occur (Crutchfield, Johnston
Gf McMillan 6082, Graham & Johnston 4695, both near Padilla, Tamaulipas), and
it is quite possible that introgression may occur between the annual and perennial
species in this region. No satisfactory chromosome counts have been obtained for
this species, but the slides which were made indicated approximately 12 to 15
1 this probably has n = 14.

7. Poinsettia cyathophora (Murr.) Kl. & Gke., Monatsb. Akad. Berlin 1859:
253. 1859.

Euphorbia cyathophora Murr., Comm. Gotting

(Murr.) Moench, Meth. 667. 1794. Euph

Boiss., DC. Prod. 15(2) :72. 1862. Euphorbia heterophylla forma cyathophora

(Murr.) Voss, Vilmorin, Blumengartn. (ed. 3) 1:898. 1895.
Euphorbia graminifolia Michx., Fl. Bor.-Am. 2:210. 1803. Poinsettia graminifolia

(Michx.) Millsp., Field Mus. Pub. Bot. 2:3 04. 1909.
Cyathophora picts. R.t., Fl. Tel!. 4:117. 183 8.
Cyathophora ciliata Raf., Fl. Tell. 4:117. 183 8.
Euphorbia barbellata Engelm., Rep. U. S. & Mex. Bound. Surv. 2:190. 18 59 [Type:

Wright s.n., GH!]. Euphorbia heterophylla var. barbellata (Engelm.) Holzinger,

Contrib. U. S. Nat. Herb. 1:216. 1892. Poinsettia barbellata (Engelm.) Small, Fl.

Southeastern U. S. 722, 1334. 1903.
Euphorbia be* ■ lia Engelm., Rep. U. S. & Mex. Bound. Surv. 2:190.

1859.

1859:104. 1860.

Distribution. The eastern United States south through eastern Mexico to the
Isthmus of Tehuantepec; west to Colima in Mexico. Possibly native in the
Greater Antilles. Widely cultivated and often naturalized.

There is a specimen from the Gottingen Botanical Garden (1794) in the
Missouri Botanical Garden herbarium. While this was prepared after the descrip-
tion of Euphorbia cyathophora, it probably represents the population described by

Though this species is usually an annual, clonal material is easy to maintain by
cuttings, which should facilitate the experimental study of its polymorphism.
Under favorable conditions it may be a facultative perennial. I have seen specimens
from northeastern Mexico and southern Florida which had over-wintered at least
one season before collection.

8. Poinsettia dentata (Michx.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253.
1859.

Euphorbia dentata Michx., Fl. Bor.-Am. 2:211. 1803. Anisophyllum dentatum (Michx.)

Haw., Syn. Pi. Succ. 162. 1812.
Euphorbia herronii Riddell, Syn. Fl. W. States 32. 1835.

DRESSLER A SYNOPSIS OF POINSETTIA 339

Euphorbia dentata p rigidia Engelm., Rep. U. S. & Mex. Bound. Surv. 2:190. 18 59 [Type-

Wright 1837, MO!].
Euphorbia dentata 7 cuphosperma Engelm., Rep. U. S. & Mex. Bound. Surv. 2:190. 1859

[Type: Wright 1834, MO!]. Euphorbia cuphosperma (Engelm.) Boiss., DC. Prod.

15(2) :73. 1862. Poinsettia cuphosperma (Engelm.) Small, Fl. Southeastern U. S.

721, 13 34. 1903. Euphorbia dentata forma cuphosperma (Engelm.) Fern., Rhodora

50:148. 1948.

biedeanaKl. & Gke., Abh. Akad. Berlin Phys. 1859:102. 1860.
Euphorbia dentata p Jasiocarpa Boiss., DC. Prod. 15(2) :72. 1862.
Euphorbia dental. -, '';, ,'.> I a^lni. a. li.mvJK. Pr.J. \^\Z\:71. 1862.
Euphorbia dentata var. gracillima Millsp., Pittonia 2:90. 1890.
Euphorbia dentata var. lancifolia Farwell, Am. Midi. Nat. 8:273. 1923.

9. Poinsettia pentadactyla (Griseb.) comb. nov.

Euphorbia pentadactyla Griseb., Gott. Abh. 24:63. 1879.

Distribution. Paraguay and northern Argentina.

Annua, ramosa, 8-14 cm. alta, ramisque procumbens ad 12 cm. longa, caulibus
ia vel ea basalia opposita, ovata vel elliptica, 8-16 mm.
subtus sparse hispidula, apice acuta, basi cuneata, mar-
ginibus leviter serratis et hispidulis, petiolis 2-6 mm. longis; cymae dichotomae,
involucris late campanulatis, ca. 2 mm. longis, 1.2-2 mm. latis, glandulis 0.6-0.8
mm. latis, ostio oblongo; ovarium glabrum styli ca. 1 mm. longi, basi breve connati,
fere ad basin bipartiti; semina quadrato-ovoidea, minute carunculata, sub-laevia,

Peru. Dept. & Prov. Lima, Dist. Pachacamac; Atacongo; among rocks in arid valley,
alt. 195 m., common, scattered, October 14, 193 5, Ynes Mexia 04044 (Type: MO, isotype
GH, presumably also at UC).

This is apparently a deep-rooted, spreading annual. The plant would pass
superficially for a small delicate P. heterophylla, but the gland is quite different in
shape, and the nearly smooth seed with an equatorial row of small, blunt tubercles,
and fainter basal and distal rows, is unlike that of any other species of Poinsettia.

11. Poinsettia heterophylla (L.) Kl. & Gke., Monatsb. Akad. Berlin 1859:

253. 1859.
Euphorbia heterophylla L., Sp. PI. 453. 1753. Tit • (L.) Haw.,

Syn. PL Succ. 141. 1812. Cyathophora heterop> 4:117 1838.

^Euphorbia elliptica Lam., Encyc. Meth. Bot. 2:425. 1788. Euphorbia heterophylla var.

elliptica (Lam.) O. Ktze., Rev. Gen. 2:605. 1891.

U, Symb. Bot. 2:53. 1791. Euphorbia heterophylla var. hmfolta

(Vahl) O. Ktze., Rev. Gen. 2:605. 1891.
Euphorbia geniculata Ortega, Hort. Mat. Dec.

& Gke., Monatsb. Akad. Berlin 1859:253.

(Ort.) Gomez, Anal. Hist. Nat. Madrid 23
Euphorbia prunifolia Jacq., Hort. Schoenb. 3:

(Jacq.) Haw., Syn. Pi. Succ. 143. 1812.

Monatsb. Akad. Berlin 1859:253. 1859.

340 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Euphorbia frangulaefolia HBK., Nov. Gen. et Sp. 2:62. 1817. Poimettia frangulae folia

(HBK.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253. 1859.
Euphorbia trachyphylla A. Rich., La Sagra, Hist. Cuba 9:198. 1850.

Euphorbia morisoniana Kl., Seem. Bot. Voy. Herald 100. 1853. Poimettia morisontana
(KI.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253. 1859.

■• kad. Berlin Phys. 18 59:102. 1860.
, DC. Prod. 15(2) :73. 1862.

a Mull. Arg., Mart. Fl. Bras. 11(2):694. 1874.
Euphorbia frvn . Mart. Fl. Bras. 11(2) :695. 1874.

Euphorbia zonosperma Mull. Arg., Mart. Fl. Bras. 11(2) -.696. 1874.

Euphorbia heterophylla Pelt: O. Ktze., Rev. Gen. 3 (2) :286. 1898.

Poimettia havanemis Small, Fl. Southeastern U. S. 722, 1334. 1903.
Euphorbia aureocincta Croiz., Jour. Arnold Arb. 24:181. 1943 [Type: Rojas 3379, AH!].

Distribution. From Arizona and Tamaulipas southward throughout the Amer-
ican tropics. As a weed throughout the tropics. Naturalized in Louisiana and
Texas and to be expected elsewhere.

Excluded Species

Poimettia eriantha (Benth.) Rose & Standley, Contrib. U. S. Nat. Herb. 16:13.
1912. = Euphorbia (Agaloma) eriantha Benth. This species, E. (A.) exstipulata
Engelm., £. (A.) jaliscensis Rob. & Greenm., and E. (A.) lacera Boiss. are all
closely related. All of these species have small but distinct petaloid appendages
on the glands and have seeds which are unlike those of Poimettia. In E. (A.)
eriantha, at least, the pollen grains are very unlike those of Poimettia.

Euphorbia heterophylla var. eriocarpa Millsp., Proc. Cal. Acad. II 2:230. 1889.
The type (F!) is not a Poinsettia, but appears to be an Agaloma with reduced
gland appendages, or possibly a Tithymalus. It is probably unnamed as a

Poimettia imulana (Veil.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253. 1859.

= Euphorbia (Agaloma) imulana Veil. This and the following two species

are closely related, if not conspecific.
Poimettia land folia (Schlecht.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253.

1859. = Euphorbia (Agaloma) lancifolia (Schlecht.)
Poinsettia oerstediana Kl. & Gke., Abh. Akad. Berlin Phys. 1859:103. 1860. =

Euphorbia (Agaloma) oerstediana (Kl. &Gke.) Boiss.
Poinsettia punicea (Sw.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253. 1859. =

Tithymalus puniceus (Sw.) Haw.
Poinsettia xalapensis (HBK.) Kl. & Gke., Monatsb. Akad. Berlin 1859:253. 1859.

= Euphorbia (Agaloma) xaUpensis HBK.

References

dorsal view of seed )

I of pis
in bold face type; synonyms a
italics; and all other matter, in

GENERAL INDEX TO VOLUME XLVIII

w combi,

Adenogramma oppositifolia,
Adenonema, 97
Agathophytum, 2
Aizoaceae of Panama, 80
Albersia, 13; gracilis, 14
Allaganthera forskalli, 39
Allochlamys, 9

Uternanthera, 3 8; achyranth., 41; amoena,

crucis, 42; denticulata, 39; ficoidea, 41;
ficoidea, 40, var. flavogrisea, 42, 7 halimi-
folia, 42; flavogrisea, 42; halimifolia, 42;
laguroides, 44; lehmannii, 45; linearis, 33;
mexicana, 45, 46; nodi flora, 39; panamen-
sis, 45, 46; paniculata, 34; paronychioides,
40; parviflora,
polygonoides, 4

:< Lim

iC

Amblogyna,
Ambrina, 2; ambrosioides, 5; andicola, 5;
anthelmintica, 5; chilensis, denudata, 5

Anemopsis calif ornica, 107, 709, I/O, 7/J,

Anisopbylhim dcutat

An thy His, 104
Aphanathe, 73
Arctium lappa, 295

ItfS, 92;

./,■-,■

-

lanuginosa subsp. guatemalensis, 93,
94, subsp. lanuginosa, 92, subsp. saxosa,
93, var. longipedunculata, 92, [var.]
diffusa, 92, p diffusa, 92, a genuina, 92;
leptophylla, 208; mearnsii, 93; wo««-
g/wea, 194; nemorosa, 92, £ novograna-

17, C /J««o

...

6, var. hybridus, 16;
ffatt, 18; dubius, 16,
D leptostachys, 16,

:^c«j, 16; hybridus,
15, var. hypocondri-

ANNALS OF THE

BOTANICAL GARDEN

var. paniculata, 54, var. viscosa, 54; dis- hides, 3

color, 52; erecta thornberi, 52; Ixodes, cola var

54; laxa, 54; paniculata, 54, /3 subacuta, phylloid*

52; polymorph*, 54; sonorae, 54; virgata, 5; *<»»*«■
i2; tiscosa, 54, [var.] apiculaU

52; caribaea, 54; diandra, 5'
53, 54; */o»g«/*, 52; erecta, 5
M/a, 54; ramulosa, 54; squamat

fi. ,w7;h,

WW, 2; leptophyllum var. fe£*o-
s, 3; murale, 2, /3 albescens, 2,
\enense, 2, [var.] spissidentatum,
turn, 5; petiolare var. leptophyl-
pratericola, 3, 4, subsp. prateri-

L-ht»,

Ckytoniap

;:7; Mr;«,

Cucurbit:

spathulatum, 5; s«#r«-

• 302, 305, 305, 324
fj/rtl, 33 8; heterophylla,

Bunting, George S., & J

achyranthoides, 27, 7 </«*«-
3 glabrescens, 27; gcnicitlata,
ta, 2 5, 2(5, j3 achyranthoides,

Desmochaeta achyranthoides, 27

Dimeiand'ra, 12

Dimeianthus, 12

Diplochonium, 84

Dodson, C. H., & G. P. Frymir

I grandiflorus, 169; oculatus,

,71

19; altissima, 19, 20, subsp. a/£o-

te, 21, var. rwMfc, 21, 7 <to-
j3 laxi flora, 21; celosioides, 22;

:henopodium,
carthagenenu

artificial key to species,
5 morphology, 174; key to s.

177

, 24V

excisae, 179, 201; ser. fasciculatae, 179,

grandiflores, 179, 23 0; ser. holosteoides,
178, 186; ser. laxiflores, 179, 222; ser.
leptophylla, 179, 204; ser. lyropetalae,
178, 189; ser. ortegioides, 178, 200; ser.
stipitatae, 178, 193; ser. tenues, 179, 210;
ser. viscosae, 178, 200; ser. villosae, 179,
223; adenophora, 103, 251; adiantoides,

index 345

254; agapatensis, 242; anomala, 210 277; carpoides, 707, 199; polystachya, 220,

apetala, 232, 234; arenarioides subsp. are- var. diffusa, 220; praecox, 218, 219; pro-

narioides, 196, 197, subsp. peninsularis, cumbem, 103, 251; ramosmnr.a, 247;

197, 198; auriculipetala, 21 j, 215; axil- retina, 2-~>>; rotuinlirolia var. nitida, 237,

rancae, 102, 226; blasiana, 247; carinata, folia, 254; •.■■.■/,■.'•..;. 254; sperguloides,

247, var. pereimis, 24"; cbibuabuensis, 194; spbagnopbiL:

222; aliarh, 226; ciliata, 226; cognata, 213; stereophylla, 214, var. exsnpul.ua,

208; confusa, 205; conzattii, 224, 225; 27J, 215, var. stereophylla, 213, 215;

cordata, 103, 251, 252, subsp. cordata, stipitata, 193, Jpjj rfric/*,

251, 252, subsp. diandra, 252, 253, var. 102, 226; subsessilis, 227; subumbellata,

7 gracilis, 245, var. pacifica, 103, 251, 190, 707; suffruticpsa, 190, 192; rew//a,

var. j3 palustris, 227, var. #/om, 102, 226, 208, var. nodosa, 209; tenuis, 210, 211,

var. 5 villosa, 102, 226, (3 Diandra, 103, a var. genuina, 211, ^ var. ?«&

251, y3 puberula, 103, 251, f. *«</«:*, 253; tepicana, 102, 226; townsendii, 227;

.. 187; cubensis, 101, 226; de- veatchii, 186; villosa, 707, subsp. palus-

\; diffusa, 220; di- f. townsendii,' 228, 22?, subsp. 'para-

varicata, 244, var. divaricata, 2^7, 242, morum, 229, 230, subsp. villosa, 225,

var. divergens, 24/, 243, var. reflexiflora, 226, f. tepicana, 227; virgata, 234; r/.s-

24/, 244, var. stricta, 241, 242, var. vis- cidula, 243; viscosa, 200, 203; weber-

cidula, 24/, 243; elata, 198, 190; effusa baueri, 238; xerophylla, 249,250

var. confusa, 205, 207, var. depressa, Dufourea, 91

206, 207, var. effusa, 205, 207; engleriana Duke, James A.: Amaranthaceae of Pana-
var. devia, 218, 219, var. engleriana, 217, ma, 6; Caryophyllaceae of Panama, 90;
2l8; excisa, 202, 203; externa, 253; fas- Chenopodiaceae of Panama, 1; Prelimi-
ciculata, 217, 27<?; fendleri, 247, var. nary Revision of the Genus Drymaria,

perennis, : formh, 173; & George S. Bunting: Sanan-

211, 254: firmula, 231, 232; frankeni- go: New Amazonian Genus of Logani-

oides, 196; frutescens, 213, 216; galeotti- aceae, 269

ana, 249; gentryi, 209; gerontogea, 253;

glaberrima, 232, 234; glanulosa var. E

var. glandulo'sa, 246, 247, var. perennis, Eisendrath, Erna Rice: Portraits of Plants.

208;' Eremogone, 91

grandiflora, 232, 236; grandiflora, 254; Esmarchia, 95

grandis, 202; &M*/4, 101, 226; holoste- Euphorbia aureocincta, 340; barbellata, 338;

oides var. crassifolia, 7^7, var. holo- color ata, 336; cuphosperma, 3 39; cyatho-

steoides, 186, 7(?7; hypercifolia, 201, 203; phora, 338; dentata, 338, var.

hypercifolia, 250; idiopoda, 254; jo£«- 339, var. lancifolia, 339, - cuphosperma,

stonii, 198; ladewii, 250, 252; laxiflora, 339, lasiocarpa, 339, 7 /*»*«*, 339,

2 J/. 2 22; hptoclados, 247, var. peruviana, rigidia, 339, f. cuphosperma, 339; e//*>

247; leptophylla var. cognata, 207, 208, ««, 339; eriantha, 340; fryJ

var. leptophylla, 207, 208, var. nodosa, 33 5; exstipulata, 340; fastuosa, 335;

207, 209; longepedunculata, 202, 203; frangulae folia, 340; geniculata, 339;
lyropetala, 192, var. coahuilana, 700, 193, gramini folia, 338; hermmi, 338; hetero-

234; mairei, 254; malachioides, 225; 3 3 8, var. brasiliensis, 340, var. cil:pi:c.i.

megalantha, 254; mmuscula, 206; mol- 339, var. eriocarpa, 340, var. gramini-

luginea, 194, 795; monticola, 2J2, 235; /o/w, 338, var. linifolia, 339, ry^tf-

multiflora, 224, 225; «*/«/*, 239; nodW, /or*, 338, *//#/«•* f. hirticaulis, 340,

209, var. gracillima, 208; nummularia, t geniculata, 3 39, 8 ™/«or, 340, f. ry-

227; ortegioides, 200, 203; ovata, 2J2, athophora, 338; insulana, 340:

233; oxalidea, 254; pachyphylla, 7#7, 340; lacera, 340; lancifolia, 340; lini-

188; palmeri, 254; / 1//, /ri . 227; papo- /oto, 339; morisoniana, 340; oerstediana,

sana var. paposana, 236, 237, var. serru- 340; pentadactyla, 339; prunifolia, 339,

lata, 2J7, 23 8, var. weberbaueri, 237, var 7 angusti folia, 340, var. a n-pW*,

230; pauciflora, 227; 340; pulcherrima, ' " "

25A;

3 36:

ANNALS OF THE MISSOURI BOTANICAL GARDEN

bieae: key to New

World

genera,

Honkenya, 90

Houttuynia cordata, 107, lOp, HO, 112,
117

1 gracih

fr, 14, 7

Hueber, Francis M.: Hepaticites devonicus,
A New Fossil Liverwort from the De-

R. Bradley: Two N<

:w Members of

vonian of New York, 125

jenus Scolecopteris, 2;
F

Hyala, 104
Hylebia, 97

I

i cordifolia, 32$, 326

of the word, 291; chronological index of

e, G. P., & C. H. Dodson:

Prelimi-

icones of key species, 320; distinguishing

of the Cucurbitaceae as \h
:chniques, 314; lithographs,

Glomeraria, \2

llleccbrun ad.

Gomphrena, 47; aggregata, 36; crassifolia,

36; decumbens f. roseiflora, 47; dispersa,

polygonoide.

48; ficoidea, 41; decumbens, 47, 48, var.

carinata, 48, var. genuina, 48, var. grandi-

Index of Orch

folia, 48, f. albiflora, 47; globosa, 49, var.

7rp;/m, 3 1

-19; ixnimcnm, 48; pauiculata, 29, var. 7

angustifolia,

hookeriana, 28; perennis subsp. pseud 0-

34, var. eri,

decumbens, 48, f. simplex, 48; polygo-

36

eriophora, 3

34; gossypi

Gossypianthus decipiens, 28, hookerianus, 28

hassleriana,

-iophylla,

phylla, 3-

34; gra

ana, 31, J2; havanen

Phylla, 31;

Gymnotheca <

107;

involuc

107, no

Gypsophytum,

H

Hagaea, 104

hEZ,91

Hebanthe, 27;

, 2S

28; panicula,

ta, 29;vin

29

Lepto, •

-

Lepyrodiclis,9\
Lestibudesia, 11
Leucodanium, 95
Leucostemma, 97
Lipandra, 2
Litanum, 86

■mculihi, 3 6
Loganiaceae: New Genus, 2
Lopboxera, 11

M

P rbombifolia%\T(n>ata, 74 °

Mitscherlichia, 57

Mohringia, 90

Moenchia, 95

Mo//«, 104; spadicea, 105

Mollugo, 80; arenaria, 81; ax/'
dichotoma, 81; ,/#»», 81;
81; hoffmannseggiana, 81; ja
schrankii, 81; spergulaefolia,
phylla, 81; verticellata, <?/

\l"Uii^op!.ytum, 100; crassifoli
holosteoides, 186; sperguloides,

Morocarpus, 1

Morphology and Anatomy of t]

Paputaria, 82

Paronychia .

frutescens, 42

Pentaple, 95

;;pepo,- J24

." JOJ

Petiveria, 66; al

acea, 67

,/n/,,,

</rw, 68; or£

8; octandr

paragvaycnsis,

Pfaffia, 27; m«

n/Atf, 29

grandiflor

, 28,

var. hookeria

M, 28;

JO

Pharnjccum an

'*""'".

1; f.Trw»

, 81;

72; icosandra, 71, 72, var. <ro-
/*, 71, var. sessiMora, 71,
2; lovgespica, 71; macrostachya,

i-hispania, 71; polystigm
i, 72; purpurascens, 71
rugosa, <5p, 70; sessili flora

Myosotis, 95
M^05o/o«, 96

Neea, 57; amplifolia, 60; delicai
laetevirens, 58, 59; orosiana, 60
60; psychotrioides, 58, 60; pycnantha,
60; urophylla, 60; xanthina, 58

NTevlin •, I.orin I., Jr.: Aizoaceae of Pana-
ma, 80; Portulacaceae of Panama, 85
e of Panama, 51

\NNALS OF THE MISSOURI BOTANICAL GARDEN

3 39; cyathophora, 331, 332, 333, 338

Embryology, 107

dentata, 331, 334, 338; edwardsii, 33 8,

eriantha, 340; frangulaefolia, 340; genic-

Rbodalsine, 91

ulate 3 39; graminifolia, 338; havanensis

340; heterophylla, 331, 332, 334, 339

76; dodecandra, 77; ehrenbergiana, 77;

gracilis, 76; humilis, 75, 76, [var.] /a«/«,

340; pedunculata, 3 36; pentadactyla

76, var. orientalis, 76, a canescens, 76, p

3 39; pinetorum, 33 8; pruni folia, 3 39

canescens, 76, 5 glabra, 76, 7 plumbagini-

folia, 76, a puberula, 76, P scandens, 77;

laevis, 76, var. pubescens, 76, p acumi-

H-hkdcana, 3 39; strigosa, 3 36; xalapensis

nata, 76; lanceolata, 76; moritziana, 77

Folia, 104

P obtusifolia, 77; orientalis, 76; pallida

76; Paraguay ensis, 76; polyandra, 79

Paly car pa bra if olid, 105; corymbosa, 105

portulaccoides, 76; procumbens, 76; pu-

Polycarpaea, 104; atherophora, 105; £r<wi/

berula, 76; scandens, 77; tetrandra, 76

/o/w, 105; corymbosa, /05, var. &re

Rivinia, 74

1 /•/«//<;. 105, var. /y/>;Va, 105; densiflora

105; jifc/o&t, 105; Jtt<//oz, 105; sjW/cea

Poly car pea, 104

Poly car pia, 104

Rosea, 3 1 ; elatior, 3 3

k Limited Study of the

the Genus Stanhop<

18; embryology, 114
rurus chinensis, 107,

Ptychocarpus, 27
Pupalia densiflora
Py xidium, 13

Racoma, 82

Raeder, Ka.

ma, 66

Scleropus, 12

Scolecoptens: Two new m

Genus, 275
Scolecopteris, 275, 276; elegs

ensis, m, 287, 288, 289;

, 275;

2S7,

le: Phytolaccaceae of Pana-
Morphology and Anatomy

parvifolia, 27
, 279, 284, 285, 286; o
veri, 275; radforthii, 275; ripageriens
275; subelegans, 275
erturnera, 17
esuvium, 84; ortegae, 84; pedunculatu

Stacbyarpagopbora, 22; aspersa, 24
Stanhopea, Preliminary Studies, 137
Stanhopea: distribution, 7Jp; key, 168;

morphology, 138; phylogeny, 140, 141;

pollination, 146, 1 72; successfulness of

'•'.'! in ubachii, 97, \
>8; reichenbachii,

nfracta, 170;

aurea, 170; bicolor, 171; bucephalus, 143,
165, 170; calceolata, 169; calceolus, 169;
Candida, 169; cavendisbii, 170; cirrhata,
74<5, 169; connata, 757, 169; conioluta.
169; costaricensis, 170; cymbiformis, 170,

169; ecornuta, 745, 169; elegantula, 170,
expansa, 170; /fcv<*, 170; florida, 171;
fregeana, 171; fuerstenbergiae, 171; gib-

W.

;•■/.■-

171;

170; impressa, 171; tnodora, 1
nis, 75<?, 170; intermedia, 170;
170; langlasseana, 169; lewisae,
fo/zri, 170; /W/ 7, 170; fc
fywo'tf, 170; macrochila, 170;

42; flavogrisea,
TUtifolia, 42; £d

45; microcephi

albiflora, 76, var. gW«, 76
Tit by wain < cyatbnpborus, 3 3 8; heterophyl-

Torrubia, 61; costaricana, 62, 63; panamcn-

, 82, £3; procumbent,

>ua, 17-..;

. 169;

169;

'5-A

169; rwr&eri, 171; russelliana, 171;
cata, 7(5/, 170; schilleriana, 171; shuttle-
worthii, 171; stenochila, 169; ttutvis,
169; tricornis, 742, 7^p, 169; tigrina,
75p, 170; uncinata, 171; *•*/***, 170;
violacea, 171; W/««,

/*,-.

170; xytriophc
Stellaria, 96; adenophor

baJdwini, !
pidata, 97; cuspidata, '.
filiformis, 254; f ranch,
>, 99, 100; lanuginosa, '.

; octandra, 77; polyandra, )

MISSOURI BOTANICAL GARDEN

Frits W. Went

U R,

. . Dodge,
Robert I

BOARD OF TRUS