JOURNAL

OF THE
ARNOLD ARBORETUM

HARVARD UNIVERSITY

EpIToRIAL BOARD

A. C. SMITH, Editor P. C. MANGELSDORF
I. W. BAILEY E. D. MERRILL

J. H. FAULL H. M. RAUP

I. M. JOHNSTON A. REHDER

C. E. KOBUSKI K. SAX

VOLUME XXIV

JAMAICA PLAIN, MASS.
1943

Reprinted with the permission of the
Arnold Arboretum of Harvard University

KRAUS REPRINT CORPORATION
N ork

1968

No.
No.
No.
No.

Oo 0

DATES OF ISSUE

(pp. 1-117, 1 pl.) issued January 15, 1943.
(pp. 119-242, 6 pl.) issued April 15, 1943.
(pp. 243-374, 12 pl.) issued July 15, 1943.
(pp. 375-518, 6 pl.) issued October 15, 1943.

Printed in U.S.A.

TABLE OF CONTENTS

THE AMERICAN SPECIES OF Drimys.. By A. C. Smith ............4..

PLANTAE PAPUANAE ARCHBOLDIANAE, XI. By EF. D. Merrill and L. M.
PBEM a eg SEO Aen eee Cae OC KEN dh + i eee Upeteeeeeameeen ees

STUDIES IN THE THEACEAE, XIV. NoTES ON THE WEST INDIAN SPECIES
OF TERNSTROEMIA. By Clarence E. Kobuskt 1... 0... cccvcccececees

PAPUAN GRASSES COLLECTED BY L. J. Brass, III. By Agnes Chase ....
NEW PHANEROGAMS FROM Mexico, V. By Ivan M. Johnston ..........
SCHIZOMUSSAENDA, A NEW GENUS OF THE RUBIACEAE. By Hut-Lin Li...

SALIX FLORIDANA CHAPMAN, A VALID SpEciES, With one plate. By
PEON TS TS xt icais winder a4 Sy ae RGD RA da OM eae e's

KarELIN (1801-1872) anv Krrirov (1821-1842), Explorers oF
SIBERIA AND MippLe Asia. By Vladimir C. Asmous ..........0 0005

Taxonomic NoTEs ON THE OLD Wor-Lp SPECIES OF WINTERACEAE. By

oe Os a ee ee ee rr ee eee ee re eran
NovELTIES IN AMERICAN EUPHORBIACEAE. By Leon Croizat ........

THE FAMILY HIMANTANDRACEAE. With six plates. By J. W. Bailey,
CHGVIOHE TE. NGS, G00 Ai CS Ee cro OS mons Vaden pee oe
PLANTAE PAPUANAE ARCHBOLDIANAE, XII. By E. D. Merrill and L.
NE PRES. ne Pas 4 a keen ed ois Cid ETE EO eS

New AND NoTEworRTHY POLYPETALAE FROM BrITISH GUIANA. By N.
PO GROUU 50 6.55.5 bend sexe oe Pk s woke eae OR REDE ORs

NOTEWORTHY SPECIES FROM MEXICO AND ADJACENT UNITED STATES,
2

By fvdn Ml. Jonson’ soa ec cek cee dae ss. 0s dae Dees ee 6

PUBLICATION DATES FOR THE BOTANICAL PARTS OF THE PACIFIC RAIL-
ROAD Reports, By Jan: JORMSION @ isa os 5s ce tko ke cere ge wowed
A SECOND SUMMARY OF THE SCROPHULARIACEAE OF NEW GUINEA.
With five plates. By Francis W. Pennell .........0. ccc cece cnces
A Stupy oF CYTOLOGY AND SPECIATION IN THE GENUS PoPputus L.
With four plates: By 2. Chalmers SW esa eh 6654s 04% Cede ee ses
PLANTS OF COAHUILA, EASTERN CHIHUAHUA, AND ADJOINING
ZACATECAS AND Duranoo, I. By Ivan M. Johnston ..............
THE COMPARATIVE MORPHOLOGY OF THE WINTERACEAE, I. POLLEN AND
STAMENS. With three plates. By J. W. Bailey and Charlotte G. Nast

STUDIES OF PaciFic ISLAND PLants, II. NoTES ON THE PACIFIC SPECIES
Ore teen. By ACs ae ee ee ek eB ae ee eae eee

103

107

119
165

190

207

218

rey

243

ra bs

306

iv TABLE OF CONTENTS

Notes ON THE Fiora oF INpo-Cu1naA. By Hui-Lin Li ........ ccueee: Be

PLANTS OF COAHUILA, EASTERN CHIHUAHUA, AND ADJOINING ZACA-
TECAS AND Duranoo, II. By Ivan M. Johnston ......... 000s eee eee

PLANTAE PAPUANAE ARCHBOLDIANAE, XIII. By E. D. Merrill and
L

oes @ 6 6 ee Ole SE ee Oe EO. 2 ae Re Re ES DD. Oo OE SR Be SS 4 41o Oo 88 SB

Nores ON THE FLoRA OF Kwancsi Province, Cu1na. By Hui-Lin Li.. 444

SrupIEs OF SoUTH AMERICAN PLANTS, X. NoTewortTHy MyrisTIca-
CEAE AND VACCINIACEAE. By 4. C. Smith ich eeebeeaeaeaeRrense ses 460

Tue CoMPARATIVE MORPHOLOGY OF THE WINTERACEAE, II, CARPELS.
With six plates. By J. W. Bailey and Charlotte G. Nast ..........-.

ForsytrHiA VAHL, NOMEN GENERICUM CONSERVANDUM. By Alfred
AMIE ois oh os 8 Eee ae ROR bd oO EEE a EKS ERED SERS

Roy.e’s “ILLUSTRATIONS OF THE BOTANY OF THE HIMALAYAN Moun-
CAINS. YT Ca 1. SPCOPR aos vata nd vane enn es eonened Ves ss ,

Tue ARNOLD ARBORETUM DURING THE FiscaL YEAR ENpDED JUNE 30,
1943

Oboe eRe ee oe eee ee ROS Ae Sy OSS O42 ORO Oe 6. 9- 2k ® OSE: DODO. 0 SG OR SS

BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STU-

ENTS JOLY 1, VOTE — TONE Dy 1948 6 oc 5 cece in sh aides ea piewees 498
STAFF OF THE ARNOLD ARBORETUM, 1942-43 2.0.0.0... 000s cece eee 501
FRE ook oe RNS ee Ce SEER CAD EREROEDEE TET OES 503

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXIV JANUARY, 1943 NumBeErR 1

THE AMERICAN SPECIES OF DRIMYS
A. C. SMITH
With three text-figures

Tue family Winteraceae is composed of six genera, of which only one,
Drimys J. R. & G. Forst., is found in both hemispheres. The family is a
very homogeneous one, the various genera having in common a distinct
type of wood structure, leaf vascularization, staminal structure and vascu-
larization, and pollen grain. Prof. I. W. Bailey and the writer (1, 10)
have already expressed agreement with the prevalent opinion that the
Winteraceae is entirely distinct from the Magnoliaceae, and a study of the
family as a whole is in preparation. The scope of the present paper is
merely a taxonomic revision of the American species, which belong entirely
to the genus Drimys. A large part of my discussion is based upon the
careful analyses made by Prof. Bailey, who has prepared material in sub-
stantial quantities for anatomical study. An attempt to correlate the
anatomical characters with the gross morphology and the geographic dis-
tribution has been made. I am also indebted to Dr. I. M. Johnston for
assistance during the preparation of the manuscript.

Often cited as an illustration of a genus with a bihemispheric-Antarctic
distribution, Drimys also illustrates the varying degree of differentiation
which may be found in two sections of a widespread genus. The repre-
sentatives in the New and the Old Worlds have a great deal in common,
and attempts to break up the complex into two genera — Drimys in
America and Tasmannia in the Old World — are not generally accepted,
nor does this division seem merited. However, it is obvious that two good
sections, or perhaps subgenera, are recognizable. The American repre-
sentatives of Drimys are hermaphrodite trees or shrubs, with always per-
fect flowers, and the stigmatic surface of the carpels is limited to the apical
or subapical region, the small stigma being usually short-stipitate. The
Old World species, on the other hand, are dioecious or polygamo-dioecious
(the staminate flowers usually bearing sterile carpels, the pistillate flowers
being either with or without functional stamens), and the stigmatic surface
is extended along the entire ventral suture of the carpel.

2 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

The fact that the above-mentioned characters are so firmly fixed in the
western and eastern hemispheres respectively seems to indicate that the
original division of the genus into these two populations was ancient. In
its Old World area (Australia, New Guinea, Borneo, the Philippines, and
perhaps a few adjacent islands), Drimys shows great polymorphy and about
36 well marked species are recognizable. This variation extends to great
diversity in foliage and floral characters. In America, on the other hand,
no such extensive variability is evident, and a smaller degree of specific
stability has been evolved. It is obvious that the genetic composition of
the genus is much more diverse in the Old World than in the New. To this
extent, Drimys is an excellent illustration of the degree to which the poten-
tial polymorphy of a group may vary in different parts of its area.

Many students have taken Drimys to include some very diverse elements
in the Old World, but, as already expressed (10), I find it advisable to
follow van Tieghem (11) and Hutchinson (7) in segregating the genera
Bubbia, Belliolum, Exospermum, Zygogynum, and the New Zealand species
(neferred by van Tieghem to “Wintera Forst.”). With these elements re-
moved, Drimys is sharply distinguished from other members of the
Winteraceae by having a calyx which completely encloses the bud, com-
posed of two (or rarely three) sepals which are papyraceous to mem-
branaceous in texture. Occasionally these sepals are calyptrate and soon
deciduous, but in most American specimens they are more or less per-
sistent, often remaining attached to one another at their basal margins while
losing their connection with the torus; on mature plants the sepals are
thus often found as a coherent unit encircling the pedicel but free from it.
Many of the Old World species customarily lose the sepals entirely before
anthesis. The remaining genera of Winteraceae have a more ordinary
calyx, composed of two to several more or less deltoid or irregular lobes
which persist through anthesis and which do not enclose the bud, or at
least not after the very earliest stages of ontogeny. The stamens of Drimys
are composed of a carnose essentially terete filament, to the apex of which
the two subvertical locules are laterally adnate, quite free from one another
and dehiscing laterally or subextrorsely. In other members of the Wintera-
ceae the stamens show various and often conspicuous modifications. The
combination of calyx- and stamen-characters here discussed serves to set
Drimys apart from the rest of the family in such a sharp manner that there
can be no serious question of merging other genera with it.

Drimys was founded by J. R. and G. Forster upon two species, D.
Winteri from the Straits of Magellan and D. axillaris from New Zealand.
Since these two species are not congeneric, the genotype of Drimys must be
designated. Van Tieghem (11), who first broke up the older generic con-
cept, took D. Winteri to be the genotype, referring D. axillaris to “Wintera
Forst.” (not Wintera Murr.). In this latter decision he is thwarted by the
International Rules, since Wintera Murr. was proposed specifically to re-
place Drimys J. R. & G. Forst., and “Wintera Forst.” is merely an in-
correct use of Murray’s generic name. The typification of Drimys has
been discussed by Dandy (in Jour. Bot. 71: 119-122. 1933), who has

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 3

proposed the generic name Pseudowintera for D. axillaris and its allies in
New Zealand.

There has been a vast difference of opinion among taxonomists as to the
nomenclatural subdivision of the American population of Drimys, which
ranges from Cape Horn to southern Mexico, with representatives in Juan
Fernandez and in eastern Brazil, adjacent Paraguay and Argentine, and
on Mt. Roraima. No specimens have yet been reported from Bolivia or
Ecuador, but the discovery of the genus in these countries seems possible,
in view of the fact that a specimen from the Andes of northern Peru has
recently been collected. I have seen no specimens from Honduras, El
Salvador, or Nicaragua, but this may be due to our poor material from
these regions. In general, the genus may be said to occur in mountains
throughout the western hemisphere south of Mexico; toward the south it
occurs at or near sea-level. In certain regions it is very common; such
areas are the Magellanic region, central Chile, Minas Geraes, Colombia,
and Costa Rica. Elsewhere in its range the genus occurs sporadically, but
future collections will show whether the above-mentioned centers are
actual or whether they merely represent the best-known parts of the generic
range.

Among the early students who considered the genus, some, such as
De Candolle (2, 3), were inclined to recognize several species with distinct
geographic areas; the number of specimens available to such workers was
very limited and naturally they had an incomplete idea of the amount of
variability within any given portion of the population. Then followed a
period in which it was customary to refer all American members of the
genus to D. Winteri, as varieties, forms, or merely unstable variants. This
trend, given weight by Hooker (6), persisted until Miers (8, 9) considered
the genus as a whole. With his customary narrow specific delimitation,
Miers arranged the population into nine distinct species, separating them
into four groups on the basis of the position and branching of the inflo-
rescence (one division including the New Zealand D. axillaris, which must
be removed from the genus). An examination of the material now avail-
able shows definitely that no satisfactory specific lines can be based upon
the position of the inflorescence. Miers utilized the terminal versus the
axillary position of the inflorescence, but it appears that the inflorescence is
always terminal at its inception, the umbels (or single flowers) being borne
about the growing point of the branchlets. As the plant develops, this grow-
ing point protrudes through the cluster of inflorescences, leaving them
lateral and often pseudoaxillary. Sometimes the inflorescences thus appear
in two or more whorls toward the apices of branchlets, and whether they
are terminal or lateral appears to depend merely upon the stage of develop-
ment. It seems likely that in certain parts of the population (e.g. the
Magellanic area, etc.) growth is sharply seasonal, and in these parts the
inflorescences are most likely to appear terminal. In other areas (e.g.
Colombia, etc.) growth is more or less continuous, and the inflorescences
develop with a more loose arrangement rather than all at one time at the
apex of the branchlets. Therefore, the apical or lateral position of the

I JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

flowers can be given little weight in a taxonomic consideration. Whether
the flowers are single or aggregated into umbels or fascicles is also suscep-
tible to variation in the same individual and therefore cannot be too rigidly
interpreted, although to be sure there are definite trends in this respect in
different geographic areas. Thus, practically all the Magellanic specimens
have the flowers single, while those in the higher Andes of Chile and
Argentine have a strong tendency in this direction. In the rest of the
population the flowers are predominantly umbellate, but the single-flowered
character is shown here and there throughout.

Miers’ treatment has not been accepted by other workers. Eichler (4),
taking up the whole American population for his treatment in Flora
Brasiliensis, referred it all to D. Winteri, recognizing five forms. The most
recent regional treatment of importance is that of Hauman (5), who, dis-
cussing the Argentine plants, recognized two species, D. Winteri and D.
brasiliensis, the former with two varieties and one form, the latter with
two varieties. The various treatments of other workers need not be
examined in detail; the current tendency in herbaria is to follow the line
of least resistance and refer all the material to D. Winteri. From a casual
examination of the material in any single herbarium, one would indeed have
difficulty in following any other course, as the population appears remark-
ably homogeneous. And yet, any two specimens are likely to be quite
different in such details as number and size of floral parts, but to the casual
observer these differences do not seem concomitant with geographic distri-
bution and thus they are commonly disregarded. Differentiation in the
various geographic areas, to be sure, is incomplete, and one searches in vain
for definite and rigid characters upon which a classification can be based.
In the following paragraphs I shall consider the various parts of the plant
in turn, pointing out the degree of variability, if any, which occurs in
different parts of the population.

Hasit. The American representatives of Drimys are shrubs or low trees,
most often occurring in cool moist temperate forest. Individuals growing
in very exposed situations apparently trend to be gnarled and often to
have the leaves closely crowded toward the apices of branchlets. Other
individuals are more symmetrical and have the leaves scattered along the
branchlets. This character is very likely a reflection of environment and
is not dependable. In general, the more southern members, as_ the
Magellanic population, have a more compact habit, with stouter branchlets
and petioles; the Colombian specimens, which are comparatively large, also
have uniformly thick branchlets. The branchlets throughout are brownish
to cinereous, sometimes glaucous when young, subterete or essentially so,
and usually longitudinally striate or rugulose.

Fot1ace. The leaves are alternate and often irregularly crowded, some-
times appearing essentially whorled about the growing points of the
branchlets. The length and thickness of the petiole varies substantially in
any geographic area or even on an individual. In general, the most south-
ern part of the population has the strongest petioles, but the proportion of

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 5

length to diameter of this organ is too variable to receive systematic
attention.

The leaf-blades are coriaceous in varying degrees, variously shaped but
most often oblong or elliptic to obovate, with a stout costa and immersed
or faintly prominulous secondary nerves. The degree to which the sec-
ondaries are immersed bears a vague correlation to geographic distribution,
but this character is quite undependable and cannot be considered any-
thing more than a trend, subject to local fluctuation. The direction of the
secondaries has a certain stability in each region. In color, the leaf-blades
vary on the upper surface from pale green to dark brown, usually being
glaucous beneath when youn

The lower surface of the leaf- blades has a slightly different aspect in the
different parts of the population, and an analysis of this character is found
to be of use. Although it is an obscure character and furthermore is not
rigid and entirely dependable, nevertheless it may serve as a subsidiary
criterion. In using this character, the lower surface of the leaf-blades
should be examined under a magnification of about 50 diameters. The
stomata are always depressed, and the depression is filled with wax-like
and very finely granular material. It is due to these stomatic areas that
the lower leaf-surfaces, not only in Drimys but throughout the family,
appear to be “punctate” with white or pale dots when dry. In Drimys
Sect. Wintera, the stomatic areas appear to vary in size from about 0.02 to
0.05 mm. in diameter

The lower leaf-surface between the stomata may be essentially free of
wax-like granular material, or this material may cover the entire surface
in a more or less uniform layer. In the latter case the entire surface appears
to be whitish or glaucous. It should be noted that the glaucous appearance
is frequently lost in herbarium specimens, due to variations in methods of
drying. In general, the stomatic areas are more conspicuous in the southern
populations (D. Winteri and D. confertifolia), while in D. granadensis and
D. brasiliensis they are often obscure.

In most specimens of D. brasiliensis the lower epidermis is papillate, the
papillae being club-shaped or knob-like protuberances arising from epi-
dermal cells and distributed between the stomatic areas. The papillae may
be covered with wax-like granules, like the plane surfaces mentioned above.

The above-described characters of the lower leaf-surface are not suffi-
ciently well fixed in each part of the population to be considered more than
trends, although the extreme forms are readily recognized. I am indebted
to Prof. Bailey for the above analysis.

INFLORESCENCES. I have already remarked on the inconstant nature of
the position and type of the inflorescence — whether terminal or lateral,
umbellate or single-flowered. These differences may be taxonomically used
only with great caution and at best they demonstrate somewhat inconse-
quential trends rather than fixed characters. The proportionate length of
the peduncle and pedicels varies greatly on the same individual. The in-
florescences are usually subtended by imbricate papyraceous bracts, which

6 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

are generally oblong and obtuse, sometimes up to 15 mm. long and 7 mm.
broad. When the inflorescences are umbellate, the pedicels are subtended
by a whorl of bracteoles similar to the bracts but smaller, not exceeding
10 by 4mm. Both bracts and bracteoles are very early caducous and are
seldom seen on herbarium material; they appear to offer no points of differ-
ence in the various populations, although they may be a little more per-
sistent toward the south.

Sepats. The sepals are usually two, rarely three, and vary in texture
from membranaceous to papyraceous; they may be essentially opaque and
apparently eglandular or with conspicuous glands. In shape they are usu-
ally suborbicular-deltoid, the variations in size being fairly constant in
different geographic areas. In general, the specimens from the northern
part of the range have thicker and larger sepals than those from the south,
but there are many individual exceptions to this generality.

Perats. The petals are uniformly white and are extremely variable in
number, ranging from 4 to 17 and occasionally being as many as 22 or 25
(in forms from Peru and Panama). They are whorled on the torus and
from 1- to 3-seriate. Although the number of petals is inconstant within
any given geographic population, there are certain broad tendencies which
may be noted. For instance, specimens from Chile, Patagonia, and Juan
Fernandez have the petals 4-14 in number, while in other parts of the
range, according to my observations, the petals are 8-17, and rarely as
many as 25. The matter of petal-size is too variable in each geographic
region to make any generalization possible, although it may be noted that
the largest petals have been found in the Colombian population. In texture,
it may be observed that the southernmost specimens have the petals usually
pellucid-glandular or apparently eglandular, while specimens from the
north and from Brazil have the glands usually opaque and more obvious.

STaMENS. The torus is uniformly subglobose or short-cylindric, the
stamens occurring in several (2-4 or rarely 5) whorls, being 15-50 (rarely
to 65) innumber. There appears to be an incipient tendency toward more
numerous stamens in the northern part of the range. The southern speci-
mens have 15-40 stamens, the northern and Brazilian specimens 18-50
(sometimes as many as 65). The filaments are carnose and essentially
terete, more or less glandular with immersed and inconspicuous glands.
The distal portion of the filament, to which the anther-locules are attached
and to which I shall refer as the connective, offers a more or less dependable
character in its glandular marking. Specimens from Chile, Patagonia, and
Juan Fernandez have the connective eglandular or with very inconspicuous
and essentially colorless glands, which are scarcely apparent, even under
considerable magnification. The northern and Brazilian specimens, on the
other hand, have the connective with numerous immersed glands and
toward the apex usually bear a few conspicuous yellowish superficial glands.
These are often very obvious under small magnification and only rarely
lacking. A specimen from Mt. Roraima has the connective produced into
a small apiculum, but otherwise this organ is essentially truncate. The

1943] SMITH, AMERICAN SPECIES OF DRIMYS 7

anther-locules are elliptic to oblong, variable in size within rather narrow
limits which appear to have only inconsequential geographic significance.
The stamens are often noted as yellow by collectors.

Carpets. The carpels are usually arranged in a single whorl around the
blunt apex of the torus, being from 3 to 12 in number (rarely 2-24). The
only geographic significance which can be attached to the number of carpels
is found in the more common occurrence of high numbers toward the north
(up to 24 in a single Panama specimen). However, the great variation in
number of carpels which is found in any geographic area indicates that this
character is nearly useless for taxonomic purposes. The shape of the
carpels is essentially uniform throughout, and their size is similarly quite
uniform. The stigma is either lateral toward the ventral apex or sub-
terminal. Sometimes it is sessile, sometimes short-stipitate, sometimes
obviously exceeding the body of the carpel and sometimes exceeded by that
blunt surface. In general, the southern specimens have the stigma lateral
or at least exceeded by the body of the carpel, while the northern and
Brazilian specimens have the stigma obviously stipitate, subterminal, and
exceeding the body of the carpel. Carpellary characters are not firmly
fixed in geographic regions, and the above-mentioned trends are at best
only incipient points of differentiation. The ovules are biseriate on the
two ventral placentas, varying from 6 to 26 in number. As regards the
number of ovules, a distinct geographic differentiation is discernible. The
Juan Fernandez specimens have uniformly (16—)18-26 ovules, while in
the remainder of the population the ovules are less than 16 (except for one
or two Chilean plants which have 18 ovules). In Chile and Patagonia the
ovules are 9-16(—18); in the northern mountains and in Brazil the ovules
are 6-12.

Fruits are very uniform throughout the range of the genus in America
and offer no stable characters for specific identification. Usually only a
few carpels per flower develop, but sometimes all reach maturity. The
carpels become obovoid or ellipsoid and often slightly falcate berries, 6-15
mm. long and 4-8 mm. broad at maturity. They are dark purple or
reddish black, at length becoming deep black, and as a rule they are short-
stipitate to obtuse at base and rounded to obtuse at apex. The pericarp
becomes soft and subcarnose at maturity and is often conspicuously yellow-
glandular. The seeds are usually fewer than the ovules (often reduced to
2 or 3 in number) and are imbedded in a thin evanescent pulp; they are
black or castaneous, polished, inequilaterally obovoid and usually strongly
falcate, 3-5 mm. long, 1.5-2 mm. broad, acute or subacute at base, rounded
at apex, and with a thin and brittle testa. Miers (9: 127) has described
in detail the structure of the seed, although his terminology may not be
accepted by morphologists. The characters discussed in the above para-
graph are so uniform throughout the American species that they are not
repeated in the specific descriptions given below.

ConcLusIoNns. From the above notes it is perceived that there are no
clear-cut characters in the American population which can be associated

8 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

with geographic areas. It seems likely that the various geographic popula-
tions of Drimys in America have slightly different genetic compositions, as
indicated by the diverse trends which I have pointed out above. None of
these trends has been carried far enough to be recognized as an absolute
character, but nevertheless they can be used in classification, provided that
they are recognized as merely trends or tendencies and are not taken as
definite rigid characters. Single specimens, taken at random from the
range of the genus in America, often defy classification, but when numerous
specimens from each area are examined one perceives the direction of
morphological trends. It might almost be assumed that, given enough time
and continuation of the present-day geographic isolation, the various geo-
graphic populations of Drimys will become narrower in their respective
tendencies until even casually selected individuals will be rigidly char-
acterized.

The question which faces the taxonomist is whether species or even
subspecific units should be established on characters which, instead of being
firmly established, are merely indicated by broad general trends. The
conclusion at which I have arrived is somewhat intermediate between the
despairing attitude of Hooker and most recent herbarium workers and the
optimistic viewpoint of Miers that several good species are discernible. I
am able to recognize four major groups which, in my opinion, are sufficiently
stable and recognizable to be designated as species. ‘These groups are
(1) the Juan Fernandez population, (2) the population from Chile and
southern Argentine, (3) the population from Peru to western Venezuela and
Mexico, and (4) the Brazilian, Paraguayan, and northern Argentinian popu-
lation, this latter including specimens from Mt. Roraima. From the lack
of definite morphological barriers it may be assumed that these populations
would be interfertile, but the fact remains that they are not interfertile in
nature simply because of their present-day geographic isolation from one
another. The morphological trends in each population, although often
trivial and obscure, are nevertheless quite apparent.

Within each of the above-mentioned populations (except the small and
compact group from Juan Fernandez) I have been able to recognize
several minor groups which I designate as varieties. The morphological
tendencies within each of these varieties are not emphatic, but each variety
is geographically restricted and shows a certain amount of incipient differ-
entiation.

In citations of literature, I have taken the liberty of correcting the spell-
ing of the generic name, when necessary, to Drimys (the original spelling)
from Drymis or Drymys. I have examined most of the illustrations por-
traying the American representatives of the genus and have referred these
to the proper subdivision as treated in the present paper. A few plates are
not available to me, while many others are so inadequate that they do not
permit positive identification; these latter are not cited. The great bulk
of the illustrations has been referred to Drimys Winteri, but it is obvious
that many do not portray the Magellanic variety upon which this name
is based.

1943] SMITH, AMERICAN SPECIES OF DRIMYS 9

Citations of specimens are reasonably complete for the larger American
herbaria, but of course additional collections of most entities will be found
in European institutions. I am greatly indebted to the directors and
curators of the following institutions for the loan of material: Arnold
Arboretum (A); Field Museum of Natural History (F); Gray Herbarium
(GH); Missouri Botanical Garden (M); New York Botanical Garden
(NY); University of California (UC); U. S. National Herbarium (US).
The parenthetical letters indicate the place of deposit of the cited specimens.

ee See

;

confertifolia ;

interi var. andina; 4. sc Winteri var. chilensis ;

; 7. D. granadensis var

mexicana; 8. D. granadensis var. chiriquiensis; 9. D. pronadenss var. peruviana;

10. D. brasiliensis var. campestris; 11. D. brasiliensis var. retor 12. D. brasiliensis
var. angustifolia; 13. D. brasiliensis var. roraimensis. From oan $ series of base maps,
201HCW.

10 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

TAXONOMIC TREATMENT
eT - R. & G. Forst. Char. Gen. 83. 1776; Forst. f. in Nova Acta Reg. Soc. Sci.

1825; Lindl. Nat. Syst. Bot. ed. 2. 17. 1836; Meisn. Pl. Vasc. Gen. 3 (pars alt. 5)
1836; Spach, Hist. Nat. Veg. 7: 436. 1839; Endl. Gen. Pl. 839. 1839, Enchir. Bot.
428. 1841; Gay, Fl. Chil. 1: 60. 1845; Lindl. Veg. Kingd. ed. 2. 419. 1847; Miers in
Ann. Mag. Nat. Hist. III. 2: 37. 1858, Contrib. Bot. 1: 126. 1861; Benth. &
Hook. f. Gen. Pl. 1: 17. 1862; Eichl. in Mart. Fl. Bras. 13(1): 133. 1864; Baill.
Hist. Pl. 1: 156, 190. 1867-69; Prantl in E. & P. Nat. Pfl. III. 2:19. 1891; ee
in Bull. Sci. Fr. & Belg. 27: 222 seq., pro parte. 1896; v. Tiegh. in Jour. de Bot.
14: 280-290. 1900; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108. 1906; Lsihfeabh
in oath — 1921: 190. 1921; Hauman in Comun. Mus. Nac. Buenos Aires
2: 45,

Wintera a. Syst. Veg. ed. 14. 507. 1784; Pers. Syn. Pl. 2: 84. 1806; Humb. &
Bonpl. Pl. Aequin. 1: 205. 1808; Lindl. Introd. Nat. Syst. Bot. 26. 1830.

In the above citations I have listed the most important references to the
genus as a whole, omitting those in which only the Old World species are
considered. The genus is readily divided into two sections as follows
Plants hermaphrodite, the flowers always perfect; carpels with the stigmatic surface

limited to the apical or subapical region, the stigma small, subpeltate, often short-
stipitate; flowers solitary or fasciculate or arranged in umbels; ~~. species.
gi Sstai Aba pice is inte Pasce seep larwheay o.n/ Sia) 9.9e9 igus lara vo:6 ial dca. 8diw al guerarp Gum seravane’alardivear end Section Wi wien,
Plants dioecious or polygamo-dioecious, the staminate flowers usually oe sterile
carpels, the pistillate flowers either with or without functional stamens; carpels
with the stigmatic surface extended along the ventral suture, the stigma not stipi-
tate; Howers solitary or fasciculate, never in umbels; Old World species..........
SESW GEES cack bee ae cess Seen Haas seaGis SHAG MOAR AS arewenk Section Tasmannia.

The present treatment is concerned in detail only with the Section
Wintera. A sectional name for the Old World representatives was appar-
ently first proposed by F. v. Mueller (Pl. Indig. Col. Vict. 1: 20. 1860) as
Drimys Sect. Tasmannia, based on Tasmannia R. Br. ex DC. Reg. Veg.
Syst. Nat. 1: 445. 1817. The more detailed synonymy of the Section
Wintera follows:

— Sect. Wintera (Murr.) DC. Reg. Veg. Syst. Nat. 1: 443. 1817, Prodr. 1: 78.

Pan ee Commers. ex Lam. Encycl. 2: 330, as synonym. 1786.

Magallana Commers. ex DC. Reg. Veg. Syst. Nat. 1: 443, as synonym. 1817; Endl.
Enchir. Bot. 428, as synonym. 1841; non Cav.

Winterana Sol. ex Endl. Enchir. Bot. 428, as synonym. 1841; non L.

Drimys Sect. feng ee Vv. grees in a ae Bot. 14: 289. 1900; Pilger in E. & P.
Nat. Pfl. Nachtr. 2:

Drimys Sect. Monoplewr Vv. “Tiegh in Jour. de Bot. 14: 289. 1900; Pilger in E. & P.
Nat. Pfl. Nachtr. 2:

Drimys Sect. esa mens Vv, Tigh in Jour. de Bot. 14: 289. 1900; Pilger in E. & P.

Pfl. Nachtr. 2: 108.

ve fen attempt to ae the genus Drimys into groups was made in

1817 by De Candolle, who erected the Section Drimys on D. axillaris and

the Section Wintera on D. Winteri and three other American species. In

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 11

1824 De Candolle maintained the same classification, except that he used
the sectional name Exudrimys for D. axillaris, The fact that De Candolle
selected the New Zealand species of the Forsters as the basis of his Section
Drimys (or Eudrimys) does not affect the typification of the genus.

No further serious attempt to break up the genus into sections was made
until 1900, when van Tieghem proposed four sections, oe ae to be
based on neither essential a ates nor His
Section Eudrimys is based upon D. Winteri, D. angu ie a a Ba
World species, but, since it excludes D. ae it is not the same entity
as De Candolle’s Section Eudrimys.

Van Tieghem was the first student to realize that D. Winteri and D.
axillaris are not congeneric, and he definitely retained the generic name for
the first of these entities. Drimys axillaris and two other New Zealand
species were referred by Dandy to his newly proposed Pseudowintera
in 1933

KEY TO THE AMERICAN SPECIES

Sepals membranaceous or submembranaceous; petals 4-14, eglandular or sparsely
pellucid-glandular ; stamens 15-40, the connectives eplandular or rarely with a few
very inconspicuous colorless apical glands; stigma usually lateral near apex of
carpel, subsessile or short-stipitate, exceeded or equalled by the body of the aa
ovules 9-26; lower leaf- surface appearing white- or gray-punctate, due to the

a
Ovules (16—)18—26 on elongate placentas; inflorescence always umbellate; leaf-blades
narrowly oblong or elliptic or obovate-lanceolate, usually — 2 by 1.3-3.5 cm.;
Vian Berna ng ez... sr sy2 ornate tae eee eine eee ces wae ina a . D. confertifolia.
Ovules 9-18 on short or slightly elongate placentas; inflorescence Bee or fascicu-
ees or Pony flowers single; leaf-blades various; Chile and adjacent southern
BAe UA Sa teA ranlaes fae olay OCR LAE Rae eRe EOI Sua laiathe grerararemetoin 2. D. Wintert.

surface usually not obviously punctate, the stomatic areas not conspicuously paler
han the rest of the Jeaf-surface
Leaf-blades clliptic-oblong or narrowly so, 5-16(-17.5) by 1-5.5 cm., usually obtuse
ndary

4.5-10(-12) by 5-11(-13) mm.; southern Mexico to western Venezuela and

1) | (ea eee ARR re rere rth erie MIR ROO 3. D. granadensis.
Leaf-blades variously shaped, often arene obovate, sometimes nearly lanceolate,
rarely exceeding 11 by 4 cm., rounded or broadly obtuse or emarginate at apex,

the secondary nerves 6—10(—14) per ary often obscure; lower leaf-surface often
papillate; peduncle less than 40 mm. long; sepals 4-7 by 4-8 mm.; southeastern
Brazil, adjacent Paraguay and Argentine, and Mt. Roraima region.............
caus Did Unna nd dog iehGisuc. aS gt GIN ORIEN Re Aree aT tote AION Apnoea ate . D. brasiliensis.

1. Drimys confertifolia Phil. in Anal. ive! Chile fags re May 1856, in Bot. Zeit.

14:

ept. 1856, in Ann. Sci. Nat. IV. 7: 100.
Drimys fernandeziana Steud. in Flora 139: 408. July ue
Drimys Fernandezianus Miers in Ann. Mag. Nat. berks TI. 2: 48. 1858, eae
Bot. 1: 137. pl. 27B. 1861; Pilger in E, & P . Pfl. Nachtr. 2: 108,

fernandeziana. 1906.

12 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Drimys Winteri var. fernandeziana Steud. ex Reiche, Fl. Chil. 1: 27.1
Drimys Winteri var. confertifolia Johow, Estud. Fl. Juan Fernandez ee 245. 1896;
Skottsb. Nat. Hist. Juan Fern. and Easter Isl. 2: (Phanerog. Juan Fern. Isl.) 127.
21.

Tree, often large, up to 15 m. high, the trunk probably often 50 cm. in
diameter, the branchlets brownish, rugulose, subterete, 2-4 mm. in diameter
toward apices; leaves crowded toward apices of branchlets: petioles rugu-
lose, shallowly canaliculate, 3-14 mm. long, 1-2 mm. in dia meter, often
narrowly winged, slightly swollen at base; leaf-blades coriaceous, dark
brown or olivaceous above when dried, slightly paler to glaucous beneath
and appearing distinctly punctate, narrowly oblong or elliptic or obovate-
lanceolate, (3—)5—12 cm. long, (0.8—)1.3-3.5 cm. broad, attenuate at base
and decurrent on the petiole, obtuse or rounded at apex, slightly recurved

above, prominent beneath, the secondary nerves 6—10 per side, ascending
at an ‘angle of 40-55°, prominulous or rarely immersed on both surfaces,
usually anastomosing toward margin, the veinlets immersed or inconspicu-
ously prominulous beneath; inflorescences clustered at apices of branchlets,
umbellate, the peduncle 7-40 mm. long, the flowers 3—6 per inflorescence,

orbicular-ovate, 6-8 mm. long and broad, inconspicuously apiculate at
apex; petals 7 or 8 (rarely to 12), membranaceous, sparsely and obscurely
pellucid-glandular, oblong, 8-14 mm. long, 2.5-5 mm. broad, obtuse at
apex; stamens 30-40, 3- or 4-seriate, the filaments eglandular, 1.5-3 m
long, the connective eglandular or essentially so, the locules 0.8—1.2 mm.
long; carpels 3—6, obovoid, about 3 mm. long at anthesis, the stigma lateral
near apex, short-stipitate, not exceeding the se of the carpel, the ovules
18-26 (rarely 16) on elongate placentas. (Fic. 2, a-f.)

DisTRIBUTION: Juan Fernandez; the plant is said by Skottsberg to be one of the
commonest forest trees on Masatierra, ranging from 200 m. upward to the highest
ridges, while on Masafuera it is not seen much lower than 500 m. and occurs up to
about 1200 m. In addition to the collections cited below, the species is represented by
material obtained by Philippi (the type collection), Germain, and Johow; Miers cites
Cuming 1328, and Skottsberg lists several of his numbers which are not available to me.

Juan Fernanpez: Masatierra: Bertero 1453 (type coll. of D. fernandeziana
Steud.; also cotype coll. of D. Fernandezianus Miers, M, NY), Moseley (F, GH), Reed
(GH) ; Cumberland Bay, Hastings 255 (NY, UC, US); Salsipuedes, Skottsberg 88 (US),
88b (NY); Portezuelo, Skottsberg 6 (NY); aia of Yunque, Bock 49 (F, GH, M, NY,
US); Masafuera: Chapin 1074 (NY

lo.

This species was independently proposed by Philippi, Steudel, and Miers
within a period of two years; furthermore, Steudel and Miers used the
same specific epithet and based their concepts upon the same collection.

Drimys confertifolia is a biological entity with considerable stability of
number and form of parts. Of all the groups in Drimys in America, this
most obviously merits specific rank and can be submerged in a continental
species only if all the American representatives are combined. Its closest
relative is D. Winteri (and especially var. chilensis), with which it has in
common eglandular stamens, a lateral stigma not exceeding the body of the
carpel, obscurely glandular petals, and a lower leaf-surface with conspicu-
ous stomatic areas. Drimys confertifolia is further characterized by its

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 13

crowded narrow leaves, always umbellate flowers, an essentially fixed num-
ber of petals (usually 7 or 8), few carpels, and numerous ovules. The last is
the most definite character and, in spite of its trivial nature, seems depend-
able enough to be the principal basis of the species.

2. Drimys Winteri J. R. & G. Forst. Char. Gen. 84. 1776.

petioles rugulose, canaliculate, 3-27 mm. lon -
slightly swollen toward base; leaf-blades coriaceous or th

TREN
CA, 4

Fic. 2. a-f. Drimys confertifolia, drawn from Bock 49: a. flowering branchlet, x %;
b. flower, X 1%; c. detached calyx, x 1/2; d. stamens, extrorse and introrse views, X
5; e. carpel, x 5; f. carpel, longitudinal section, x 5. g-k. Drimys Winteri var. andina,
drawn from Werdermann 1245: g. stamens, introrse and extrorse views, x 5; h. carpel,
x 5; i. carpel, longitudinal section, xX 5; j. fruit, showing four mature carpels, x 14;
k. seed, X 3.

pale green to dark brown above when dried, glaucous or at least paler be-
neath and usually appearing distinctly punctate, usually obovate-oblong to
elliptic, (4-)6—-15(-18) cm. long, (1.3—)1.8-6.5(—7) cm. broad, attenuate
to obtuse at base and decurrent on the petiole, obtuse or rounded and some-

10-70 mm g;
obscurely pellucid-glandular, sometimes copiously so, broadly ovate to sub-

14 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

orbicular or reniform, 4-7 mm. long, 4-12 mm. broad, apiculate to rounded
at apex; petals 4— 14, membranaceous, sparsely pellucid- -glandular, oblong
to narrowly obovate, = 20 mm. long, os 6(-7) m road, obtuse at apex;
ni 15-40, 2— 4- seriate, the filaments carnose, eglandular or nearly
, 0.8-3 mm. long, the connective eglandular or rarely with a few very

eG ae colorless apical glands, the Maes 0.5-1 mm. long; elaine
(2—)3-10, obovoid or ellipsoid, 2—3.5 mm. long at anthesis, the stigm
lateral near apex or rarely subterminal, pelteite. ieee or short- “stpilate,
exceeded or equalled by the body of the carpel, the ovules 9-18 on short
or slightly elongate placentas.

DisrripuTIoN: Central Chile (Coquimbo) southward to Cape Horn and in adjacent
Argentine from Neuquén southward. The type was collected by the Forsters on Cook’s

other collections and had been freely mentioned in literature. The old common name

has tonic and antiscorbutic properties and was widely used medicinally in Shien,

I have not been able to limit the species D. Winteri to the southern form
upon which it was originally based, but on the other hand I prefer not to
extend the specific concept to include all the American representatives of
the genus. As mentioned in the preliminary portion of this treatment, the
specimens of Drimys from Chile and adjacent southern Argentine seem
sufficiently distinct from the remainder of the population to merit specific
recognition as D. Winteri. Within this complex, at least three groups are
discernible which appear to merit varietal recognition. Although these three
groups have different aspects which make them recognizable, these aspects
are difficult to analyze and do not have a basis in definite clear-cut
characters.

ESSENTIAL DIAGNOSTIC CHARACTERS OF THE VARIETIES
Tree, ee compact, sometimes up to 17 m. high; leaves crowded on distal portions of
Ta

chlets; branchlets stout, 4-6 mm. in diameter toward apices; petioles stout,
mm. in diameter, 3-15 mm. long; leaf-blades coriaceous or thick-coriaceous,

g
pats mm.; ovules 10-18; southern Chile (south of lat. 42° and adjacent

PTRONUNC 550 Aas bates cea eerste na panies oe tedak voleakyeeeaice ar. punctata
Shrub or al tree, less than 5 m. high, the ening often seen leaves evenly
distributed along branchlets ; branchlets 3-5 m n diameter toward apices; peti-

oles 1-2 mm. in diameter, 5-18 mm. long; Teaf-blades coriaceous, pale green to
olivaceous or pale brown above when dried, conspicuously pale and glaucous be-

or faintly prominulous on both surfaces; flowers single, rarely umbellate; petals
4-9, 8-18 by 2.5-5 mm.; ovules 10-12; nate Chile and adjacent Argentine
(lat. about 38° to 41°) at high devations (760-2300 m.).......... 2b. var. andina.
Shrub or tree 3-15 m. high; leaves evenly dis aie along branchlets; branchlets 3—5
mm. in diameter toward apices; petioles 1-3 mm. in diame er, 5-27 mm. long;
leaf-blades coriaceous, pale or dark green above when dried, glaucous or at least
paler beneath, often exentally hase usually conspicuously revolute at margin

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 15

toward base, the secondary nerves (5—)7-15 per side, erecto-patent or ascending
at an angle of met usually prominulous and obvious on both surfaces; anfio-
rescences umbellat the Pie very rarely single; petals 6-14, 6-20 by 2-5 mm.

ovules 9-16; a Chile (lat. about 30°30’ to 44°40’) at SS ee
AIP CULES sais sers eee eer e ee Co ee rere: Siac ec aiies a ce isis 31d) 9 ahaPern tee arene ogee eraenes chilensis.

2a. Drimys ae var, punctata (Lam.) DC. Reg. Veg. Syst. Nat. 1: 443. 1817,
Prodr. IL: 1824; Hauman in Comun. Mus. Nac. aera oe 2: 48. 1923;
Hauman Ay eee in An. Mus. Nac. Buenos Aires 32: 228

Winteranus cortex a Hist. Gen. Pl. 1858. 1586; oa ane 750: 16052

Laurifolia Magellanica cortice acri Bauhin, Pinax Theatr. Bot. 461. 1623.

Arbor laurifoliae Magellanicae . . . Jonston, Dendr. Hist. Nat. 232. 1662.

Periclymenum rectum fae en rinis . . . Sloan in Phil. Trans. Roy. Soc. London
17 (204) : 923, pl. 1.

Boigue Cinnamomifera a fructu Feuillée, are Méd. . $Y) 6.1725.

Cortex Winteranus Gars. Fig. Pl. Anim. Med. 1: 27. pl. 35B

Drimys Winteri J. R. & G. Forst. Char. Gen. f. 42, m-z. ae Forst. f. in Nova
Acta Reg. Soc. Sci. Ups. 3: 181. 1780; L. f. Suppl. 269. 1781; Lam. Encycl.

Fl. Antarct. 2: 229. 1845; Gay, Fl. Chil. 1: 63. 1845; Carson, I'l. Med. Bot. 1:
pl. 5. 1847; Dec. Bot. Voy. Pole Sud 2: 64. pl. 19 (in Atlas). 1853; A. Gray, Bot.
. Expl. Exped. 1: 24. 1854; Good, Family Flora 2: pl. 89. 1854; Steud. in

Flora 39: 408. 1856; Miers in Ann. Mag. Nat. Hist. III. 2: 45. 1858, Contrib.
Bot. 1: 135. pl. 25A. 1861; Eichl. in Ma oe Fi. Bras. 13(1): 133. 1864; Baill. Hist.
Pl. 1: 157. f. 200-202. 1867-69, Traité Bot. Méd. Phan. 503. f. 1192-1194. 1884,
Dict. Bot. 2: 474, 475. 1886; Franch. in Bot. Miss. Sci. Cap Horn, 322. —
Dujard.-Beaumetz & Egasse, Pl. Méd. 249. f. 312-315. 1889; Prantl in E. & P.N
Pfl, III. 2: 19. 1891; Reiche, Fl. Chil. 1: 26. 1896; Speg. in An. Mus. Nac. eee
Aires 5: 46. 1896; De Wildem. Bot. Phan. Terres Mazel. 18, 102. 1905; Macloskie
in Rep. Princeton Univ. Exped. Patagonia 8: 419. f. 71 (excl. fr.). A Pilger in
E. & P. Nat. Pfl. Nachtr. 2: 108. 1906; Karsten & Schenck, Veg.-Bi ow? pl. 14.
1907; Skottsb. in Kungl. Sv. aoe Handl. 56(5): 226. 1916; a at in Bull.
Soc. Bot. Belg. 58: pl. 11. 19

Winterana aromatica Soland. ex nel in aaa Obs. and Inq. 5: 46. f. 1.
1776; non Drimys aromatica F. v. Muell. (18

Wintera aromatica Murr. Syst. Ver. ed. 14. GR a Apparat. Medic. 4: 557. 1787;
Forst. f. in Comment. Soc. Reg. Sci. Goett. 9: 34. pl. 7. 1787; Willd. Sp. PI. 2:
1239. 1800; Pers. Syn. Pl. 2: 84. 1806; Humb. & Bonpl. Pl. Aequin. 1: 209. 1808;
Nees, Pl. Offic. 2: pl. 372. 1828.

Drimys punctata Lam. Encycl. 2: 330. 1786, Ill. 2: pl. 494. 1797.

Drimys aromatique Descourt. Fl. Pitt. & Med. Ant. 1: 188. pl. 40. 1821.

Drimys Winteri f. magellanica Eichl. in en Fl. Bras. 13(1): 134. ce ks ee

1864; Macloskie in Rep. Princeton Univ. Exped. Patagonia 8: 420

Drimys ‘Qvomatica Descourt. ex Baill. Hist. Pl. 1: 157. 1867-69.

Drimys polymorpha Spach ex Baill. Hist. Pl. 1: 157. 1867-69.

Drimys Winteri var. Morenonis Kuntze, Rev. ie 3(2): 2. 1898; Macloskie in

Rep. Princeton Univ. Exped. Patagonia 8: 420

Drimys Winterana Thell. in Bull. Herb. Boiss. I. 8: ene 1908.

a often compact, sometimes up to 17 m. high, the branchlets stout,
4—6 mm. in diameter near apices; leaves Saleh on distal portions of
Pancho: petioles 3-15 mm. long, stout, 1.5—4 mm. in diameter; leaf-
blades Siemet or thick-coriaceous, dark brown above when dried,
glaucous to pale or dark brown beneath, obovate or obovate-oblong,
(4-)6-13(-14) cm. long, (1.5—)2-5.5(-7) cm. broad, subattenuate to ob-

16 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

tuse at base, usually heel! and narrowly revolute at margin toward base,
the seco ndary nerves 7-11 per side, ascending at an angle of 35-45°,
immersed or faint east or faintly prominulous above, immersed or
prominulous beneath, the veinlets immersed; flowers single, clustered near
apices of branchlets, the yi ape rarely umbellate and then with 2 or 3
flowers on peduncles less than 10 mm. long, the pedicels 10-37 mm. long;
sepals 5—7 mm. long, 5-12 mm. broad: petals 5-7, 8-15 mm. long, 3.5—
6(—7) mm. broad; stamens 20-35, the filaments 0.8- 3 mm. long; carpels
3-9, the ovules 10-18.

DistriBpuTION: Southern Chile from lat. about 42° (Chiloé) southward to Cape
Horn, and in the adjacent Argentine section of Tierra del Fuego. The specimens cited
below represent only a fraction of those recorded, as the plant has been obtained by
essentially all collectors in the region (see above-cited references by Hauman, Spegazzini,
De Wildeman, Skottsberg, etc. It is apparently abundant at sea-level in the Magellan
region, and according to some writers it also ascends the hills to at least several hundred
meters. The following citations are arranged, in general, from northwest to southeast.

Cue: Chiloé: Chiloé I., Huite, Cunningham (NY); Aysen: Wellington
I., Eden Harbor, Ball (NY, US); Magellanes: Cerro Paine, Donat 398 (F, GH,
NY); Straits of Magellan, without detailed seeigg Moreno (NY, type of D. Winteri
var. Morenonis), Lenormand (M), Andersson (NY, US), Douglass (GH); Penins.
Munoz Gamero, Puerto Tamar, Safford 358 (US); ans Cordova, Borja Bay, Lee
(US) ; Brunswick Penins., Fortescue Bay, Safford (NY, US); Port Gallant, Blake
(GH); Punta Arenas, Blake (GH); Port Famine, Lee (US); Hoste I., Hardy Penins.,
Orange Bay, U. S. E Expl. cad (GH, M, NY, US), Hyades (NY), Collector? (F) ;
Hermite I., Hooker (F, G

pean Tierra 3 Fuego: La Maire Straits, Banks & Solander (co-
type coll. of Winterana aromatica, GH, US).

I AMES: Canelo, ouchkouta (wood), liouch (leaves), usskitta, ciila,
shdahlku, shélakudhr. The second and third names were apparently first recorded by
Franchet, the last four by Spegazzini.

The first variety founded upon the Magellanic form of D. Winteri was
var. punctata, based upon D. punctata Lam. This binomial was based upon
a specimen collected by Commerson in the Magellan region, sent to Lamarck
by Jussieu. Lamarck’s description and plate leave no doubt that this
species is identical with the Forsters’. Thus, while the type of the variety
is Commerson’s plant and the type of the species is the Forsters’, there can
be no doubt that the same entity is involved and that var. punctata is to
be construed as the typical variety of the species.

Garsault’s name cited above is not intended as a binomial in the Linnaean
sense, ‘“cortex” not being proposed as a genus. The entire work of Garsault
may be excluded from nomenclatural consideration, since the Linnaean
system of binary nomenclature for species was not consistently employed

(Int. Rules Bot. Nomenclature Art. 68 [4]. 1935). Therefore Thellung’s
combination Drimys Winterana, based on oo “binomial,” is not
acceptable to supplant the binomial D. Winter

A great deal of confusion has been caused by Murray’s substitution of
the binomial Wintera aromatica for Drimys Winteri. Not liking the generic
name Drimys, Murray simply substituted Wintera, which thus can have no
other connotation than as a direct synonym of Drimys J. R. & G. Forst.
Murray’s specific epithet is derived from Winterana aromatica Soland.,

~—

1943 | SMITH, AMERICAN SPECIES OF DRIMYS 17

published in the same year as Drimys Winteri Forst. Which of these two
specific epithets has strict priority is not known to me, but at any rate the
use of Drimys aromatica for the Magellanic plant (as proposed by Baillon
in 1867-69) is excluded by the earlier Drimys aromatica F. v. Muell.
(1860) for an Australian species.

The typical variety of D. Winteri is characterized by its stout branchlets,
its crowded and coriaceous leaves, of which the petioles are thick and the
blades dark brown and usually obovate, with ascending secondaries, and its
usually single flowers with comparatively few petals. These few characters
impart to the plants an impression quite distinct from that given by the
other two varieties described below.
2b. Drimys Winteri var. andina Reiche in Anal. Univ. Chile 100: 535. Apr. 1898, FI.

Chil. 2: 371. 1898.

Drimys Winteri var. quinoensis Kuntze, Rev. Gen. 3(2): 2. Sept. 1898.

Drimys Winteri var. chilensis {. andina Hauman in Comun. Mus. Nac. Buenos Aires
2:50. 1923; Hauman & Irigoyen in An. Mus. Nac. Buenos Aires 32: 228. 1923.

oblong or -lanceolate or elliptic, (4—)6-11.5 cm. long, (1.3—)1.8-4.5 cm.
broad, acute to attenuate at base, not conspicuously recurved at margin
toward base, the secondary nerves 5—7 per side, ascending at an angle of
35-45°, faintly prominulous or immersed on both surfaces, the veinlets
immersed: flowers single, rarely 2—4 in umbels (on peduncles up to 25 mm.
long), clustered near apices of branchlets, the pedicels 10-70 mm. long;
sepals 5-6 mm. long, 5-7 mm. broad; petals 4-9, 8-18 mm. long, 2.5—5 mm.
broad; stamens 15-40, the filaments 1-2.5 mm. long; carpels (2—)3-8,
the ovules 10-12. (Fic. 2, g-k.)

Distrisution: Mountains of south-central Chile (Cautin and Valdivia) and adja-
cent Argentine (common in region of Lake Nahuel Huapi); Reiche also mentions the
variety from Llanquihue, and Hauman reports it from the Rio Aluminié in Neuquén.
The plant is said to occur from 760 to 2300 m., in open woods or in forest, often asso-
ciated with Nothofagus and Araucaria.

Cautin: Rio Quino, Kuntze (NY, type of D. Winteri var. quinoensis,
US); Bafios de Trolguaca, Sargent (A, M); along road from Termas de Trolguaca to
Laguna Malleco, Morrison & Wagenknecht 17480 (GH) ; Volcan de Trolguaca, Pennell
12785 (GH); Volcan Llaima, Werdermann 1245 (A, F, GH, M, NY, UC, US); west
foot of Volcan Llaima, West 4900 (GH, M, UC); Valdivia: Dept. Osorno, Cor-
dillera Pelada, Morrison 17639 (GH)

ARGENTINE: Rio Negro: Region of Lake Nahuel Huapi, Cordini 131 (US);
Puerto Blest and vicinity, Cabrera & Job 268 (NY), Parodi 11783 (GH), West 4730
(GH, M); between Puerto Blest and Lake Todos los Santos, Elwes (A) ; Laguna Frias,
Cerro Riggi, Cabrera 6047 (F); Brazo Viento, Cantaros, Lake Nahuel Huapi, Ljungner
947 (NY).

NATIVE NAME: Canelo.

The two varietal names andina and quinoensis were both published in
1898, but the first has priority. Reiche did not cite a type, but there seems
no doubt that the present variety is the one described by him.

18 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

The var. andina is characterized by its small compact habit; it is dis-
tinguished from var. punctata by having the leaves comparatively spaced,
the branchlets and petioles less robust, and the leaf-blades thinner and
paler in color and more definitely elliptic, without the strongly recurved
basal margins. It is apparently a montane form and in many respects sug-
gests a link between the other two varieties of D. Winteri, but on the whole
it has sufficiently strong characteristics to make recognition advisable. As
a matter of fact, Prof. Bailey’s anatomical study of preparations of the
leaves of this variety incline him to believe it unusually well marked and
homogeneous. He finds it to be more stable and more readily characterized
than the other two varieties of D. Winteri.

A collection from the vicinity of the south shores of Lake Argentine,
Santa Cruz, Argentine (Furlong 62 [GH, NY]) does not fit well into any
of the proposed varieties of D. Winteri, but I believe that it is best placed
with var. andina, although it occurs far south of otherwise known localities
of this'variety. It differs from the rest of the material of this variety in
having the leaf-blades somewhat darker above (as in var. punctata), the
margins slightly more conspicuously revolute toward base, the secondaries
slightly more obvious, the flowers umbellate on peduncles up to 30 mm.
long, the stamens only 14 and with unusually large locules, and the ovules
14, I have not included these extreme variations in the above description
of var. andina. Future collections may permit the more accurate placing
of the Furlong specimen.
2c. Drimys Winteri var. chilensis (DC.) A. Gray, Bot. U. S. Expl. Exped. 1: 24.

1854; Eichl. ex Kuntze, Rev. Gen. 3(2): 2. 1898; Eichl. ex Hauman in Comun.
Mus. Nac. Buenos Aires 2: 49. 1923; Eichl. ex Hauman & Irigoyen in An. Mus.
Nac. Buenos Aires 32: 227. 1923.

Drimys chilensis DC. Reg. Veg. Syst. Nat. 1: 444. 1817; Deless. Ic. Sel. 1: 22.

pl. 83. 1820; DC. Prodr. 1: 78. 1824; Hook. in Bot. Misc. 3: 134, pro parte.

Steud. in Flora 39: 408. 1856; Miers in Ann. . Nat. Hist. III. 2: 47. 1858,
Contrib. Bot. 1: 136. pl. 26C. 1861; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108.
1906

Drimys Winteri sensu Hook. f. in Curt. Bot. Mag. 80: pl. 4800. 1854; Demilly in Rev.
Hort. n. s. 12: 18. f. 3, 4.1912; non J. R. & G. Forst.

Drimys paniculata Steud. in Flora 39: 408. 1856.

Drimys chilensis var. latifolia Miers in Ann. Mag. Nat. Hist. III. 2: 47. 1858,
Contrib. Bot. 1: 136. 1861.

Drimys Winteri {. chilensis Eichl. in Mart. Fl. Bras. 13(1): 135. pl. 30, f. 1. 1864.

Drimys magnoliaefolia Kunth ex Eichl. in Mart. Fl. Bras. 13(1): 135, as synonym.
1864.

Drimys Winteri chilensis Macloskie in Rep. Princeton Univ. Exped. Patagonia 8:
420, excl. spec. 1905.

Shrub or tree 3-15 m. high, the branchlets 3—5 mm. in diameter toward
apices; leaves usually scattered along distal portions of branchlets; petioles
5-27 mm. long, 1-3 mm. in diameter; leaf-blades coriaceous, pale or dark
green above when dried, sometimes fuscous, glaucous or at least paler be-

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 19

per side, erecto-patent or ascending at an angle of 40—60°, usually prominu-
lous and obvious on both surfaces, rarely subimmersed above, the veinlets
immersed or obscurely prominulous on both surfaces; inflorescences often
densely crowded at apices of branchlets, very rarely axillary, umbellate
(flowers occasionally fasciculate, rarely single), the peduncle (0O—)8—50 mm.
long, the flowers (1—)3—8(-12) per inflorescence, the pedicels 10-55 mm.
long, rarely branched and 2-flowered ; sepals 4-7 mm. long, 4-9 mm. broad;
petals 6-14, 6-20 mm. long, 2-5 mm. broad; stamens 24-35, the filaments
1-2.5 mm. long; carpels 4-10, the ovules 9-16.

DistRIBUTION: Central Chile, from Coquimbo (lat. about 30°30’) to Aysen (lat.
about 44°40’), often near the coast, at altitudes up to 700 m. (rarely to 1000 m.). The
plant is said to grow in forests or in woods, often in sheltered locations or near water.

Cute: Coquimbo: Dept. Ovalle, Fray Jorge, Munoz B117 (GH); Acon-

: Valparaiso and vicinity, Cuming 8644 (NY), Ball (NY), Harvey (GH),
Buchtien (US), Claude-Joseph 3692 (US), Behn (F); Las Zorras, near Valparaiso,
Harshberger 1036 (NY, US); Cerro Campana, Philippi & Borchers (F);Santiago:
Nujfioa, vicinity of Santiago, Claude-Joseph 1755 (GH, US); Melipilla, Gay 171
(GH); Esmeralda, near Melipilla, Reed (GH) ; “Winganis,” Hastings 355 (NY, UC,
US); Rio Clarillo, Grandjot (M); Rungue, Cerro El Roble, Montero 173 (GH, M);
Colchagua: Tagua-tagua 229 (type coll. of D. paniculata, M, NY);

- Camarico, Rio Claro, Moreira (GH); Rio Claro, Reiche? (A); Nuble:
Chillan, Philippi (US); near Racinto, West 5117 (GH, M, UC); Concepcion:
Vicinity of Concepcién, Germain (F); Bio-Bio: Angol, Kuntze (NY, US);
Cautin: Temuco, Claude-Joseph 1178 (US); Valdivia: Philippi (GH);
Valdivia and vicinity, Buchtien (UC), Junge (M), Lechler 550 (US); Rio Calle-Calle,
Buchtien (F, US); Corral, Gunckel 1795 (UC), 3498 (M), Thaxter (GH); Amargos,
Gunckel 86 (F); Antilhve, Sargent (A); Panguipulli, Hollermayer 1926 (US);
Chiloé: Dept. Llanquihue, road from Hotel Ensenada to Cochamo and Laguna
Patos, Morrison 17580 (GH); Isla de Chiloé, Castro, Pennell 12605 (GH, NY, US);
Aysen: Ins. Magdalena, Calqueman, Werdermann 73 (A, F, GH, NY, M, UC, US) ;
without definite locality: Dombeys.n. or 601 (type COLL. of D. chilensis,
F), Gay (GH).

CULTIVATED SPECIMENS: Nicholson (cult. Kew) (A); Anderson (cult. J. I. Hort.
Institute, England) (M); cult. Hort. Berol. (M) ; cult. Jardin des Plantes, Caen (US).

NatTIve NAMES: Canelo, boighe.

Although the specimens cited above cannot be given specific recognition,
they appear to represent a well-marked variety of D. Winteri, character-
ized primarily by the slender habit, the numerous and comparatively spread-
ing and obvious secondary nerves of the leaf-blades, the predominantly
umbellate inflorescences, and the more numerous petals. The ranges of this
variety and var. punctata apparently overlap in Chiloé and Aysen, but, in
general, var. chilensis occurs only toward the north of the range of the
species and does not ascend to as high an elevation as is characteristic for
var. andina.

Prof. Bailey calls my attention to the fact that four of the above-cited
specimens (Munoz B117, West 5117, Pennell 12605, and Werdermann 73)
demonstrate anatomical characters in the leaf which are more to be ex-
pected in var. punctata than in var. chilensis. However, in the gross
morphological characters which I have used to separate the two varieties,
the four specimens agree better with var. chilensis than with var. punctata,
that is, they have ample and umbellate inflorescences, comparatively

&
a
x
mi
fa
=
S

20 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

numerous petals, and more or less spreading secondary nerves. Thus it
appears that the anatomical and morphological characters within D.
Winteri are not always correlated in such a way that the two varieties under
discussion are obvious. To be sure, one might expect the occurrence of
unusually variable individuals where the ranges of the two varieties overlap,
and therefore the disagreement of characters found in thé Pennell and
Werdermann specimens, from Chiloé and Aysen, is not unexpected. That
the West and Munoz specimens (the latter the most northern known collec-
tion of D. Winteri) should show anatomical characters of var. punctata
is a phenomenon which I cannot explain at present.

3. Drimys granadensis L. f. Suppl. 269. 1781.

Shrub or small tree up to 13 m. high, the branchlets brownish or cinere-
ous, sometimes glaucous when young, a ea subterete (or short lateral
ones sometimes flattened), 2-7 mm. in diameter toward apices; petioles
rugulose, flattened to obviously ne caida above, often narrowly winged
distally, 4-25 mm. long, 1-3 mm. in diameter; leaf-blades thick-coriaceous
to subcoriaceous, olivaceous to brown above when dried, glaucous or pale
beneath and appearing uniform in color rather than punctate, elliptic-
oblong or spe oe so, 5—-16(—17.5) cm. long, 1—5.5 cm. broad, acute to
attenuate at base and ca on the petiole, usually obtuse at apex and
narrowly recurved at margin, the costa shallowly canaliculate to nearly
plane above, prominent beneath, the secondary nerves 8-19 per side,
ascending to erecto-patent, prominulous to immersed on both surfaces,
often freely anastomosing, the veinlets immersed or prominulous; inflo-
rescences terminal or axillary, umbellate or fasciculate or 1-flowered,
the peduncle often slightly flattened, 7—65(—90) mm. long, the flowers
1—6(—8) per inflorescence, the pedicels 5—60(—90) mm. long; sepals
papyraceous to submembranaceous, more or less opaque, usually oe
ously yellow-glandular, suborbicular- deltoid, 4.5—-10(-12) mm. long, 5—
11(—13) mm. broad, rounded to obscurely ‘apiculate at Sons ‘petals 8-
17(-25), membranaceous or submembranaceous, opaquely yellow-glandu-
lar (sometimes obscurely so), oblong to elliptic- or ovate-oblong, 5—20(—25)
mm. long, 1.5—8(—11) mm. broad, cuneate to obtuse at base, obtuse at
apex; torus short-cylindric, usually conspicuous; stamens 25—50(-—65),
2—4(—5)-seriate, the filaments carnose, somewhat flattened, often sparsely
yellow-glandular, 0.5—3 mm. long, the connective glandular and usually with
a few obvious apical yellow glands, the locules 0.3—-1 mm. long; carpels
3—12(—24), obovoid, 2-5 mm. long at anthesis, usually contracted at base,
rounded at apex, the stigma gee eae or obliquely terminal on a short
stipe 0.3-0.7 mm. long, exceeding the body of the carpel, the ovules 7-12
on short placentas.

ISTRIBUTION: Southern Mexico to western Venezuela and Peru; five varieties are
here circumscribed. The type of the species was collected by Mutis in Colombia; it
falls into the variety grandiflora, which thus, although not based upon the type of the
species, is nevertheless the typical variety.

In the original publication of this species the name was spelled grana-
densis, but in 1817 De Candolle took up the spelling granatensis, which has
been followed by most writers using the name. There appears to be no
reason for this change, and throughout this paper I have used the original

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 21

spelling, regardless of the spelling used by the cited aug in order to
avoid further complications of the synonymy. Various writers have
accredited the specific name to Mutis, but in the original Siblication Mutis
is mentioned merely as the collector.

ESSENTIAL DIAGNOSTIC CHARACTERS OF THE VARIETIES

sie etita tending to be thick (3-7 mm. in diameter toward apices) ; leaf- geet coria-

faces, rarely obscure, freely anastomosing toward margin; inflorescences axillary or
congested toward apices of branchlets but comparatively scattered; flowers usually
umbellate, rarely single; sepals thick in texture, never membranaceous, averaging
large (6-10[-12] by 7-11[-13] mm.) ; petals comparatively few, 8-13(-15), large
(9-20[-25] by 4-8[-11] mm.); stamens 25-50(-60), the anthers comparatively
large, He locules 0.6-1 mm. long; carpels 5-12(-14), 2.5-5 mm. long at anthesis;

ge eee eeu aie area ey Fraiire, re reat agetrsb St aorone see sina nce a. grandiflora.

oa oe e mm, in diameter toward apices; leaf-blades coriaceous, elliptic-oblong,
5 0.5 by 1.5-3.5 cm., rounded or definitely obtuse at apex, the s econdary nerves
eee and scarcely visible; flowers single, aggregated at apices of branchlets ;
sepals submembranaceous, about 8 by 9 mm.; petals ae 10, 9-12 by 3-5 mm.;
stamens about 30, the anthers with locules 0.5—0.6 m aes carpels about 6,
2-2.5 mm. long at anthesis; western Venezuela.............+.+- 3b. var. uniflora.
Branchlets 2-5 mm. in diameter toward apices; as blades dampened narrowly
elliptic-oblong, = aa 16(-17. 2 by 1.5-4.5(—5.5) cm., obtuse or subacute at apex,

the secondary nerves obvious on both Rae or usually so, freely anastomosing

n

comparatively small (4.5-8 by 5-9 mm.) ; petals 9-17, comparatively small (6-17
by 1.5-6 mm.); stamens 25-45(-55), the anthers with locules 0.5-0.8 mm. long;
carpels 5-12, 2-3 mm. long at anthesis; southern Mexico to Costa RGA ree ies, 0
var. mexicana.

oblong, 6-12(-14) by 1.5-4(-5.5) cm., obiuse at apex, the secondary nerves hat
f .

branchlets; flowers umbellate, rarely single; sepals submembranaceous, compara-
tively small (5-6 by 6-7 mm.); petals numerous, 14-25, small (5-8 by 1.5-2.5
mm.) ; stamens numerous, 50-65, the anthers small, with locules 0.3— te mm, long;
carpels 8-24, about 2 mm. long at anthesis; Chiriqui region in Panama...........

bea AR AN ares Keaniaciat sol chat Stents hens Ai eee eR Pe ee See OSE var. ee
Branchiets about 4 mm. in diameter toward apices; leaf-blades coriaceous, oblong-
la late, S5— , obtuse at apex, the secondary nerves faintly prominulous

n both surfaces; inflorescences aggregated at or toward apices of branchlets, the
flowers umbellate; sepals s embranaceous, 8-9 m long and broad; petals
numerous, about 22, the outer and larger ones 1C-13 by 5—7 mm.; stamens about

40, the locules 0.6-0.7 mm. long; carpels 3-7, 2— 2.5 mm. long at ‘ethene: Peru.

he aitasds eet tea ehy aria Goce wos eeeia og eae ATCT ENG EN eRe ayers 6 Sey nars aye 3e. var. peruviana.

3a. Drimys granadensis var. grandiflora Hieron. in Bot. Jahrb. 20: Beibl. 49: 10,
s D. granatensis Mutis var. g. 1895.

Drimys granadensis L. f. Suppl. 269. 1781; Lam. Encycl. 2: 330. 1786; DC. ae Veg.
Syst. Nat. 1: 444. 1817; H. B. K. Nov. Gen. & Sp. 5: 53. 1821; DC. Prodr. 1: 78
1824; Endl. En ue Bot. 430. 1841; Miers in Ann. Mag. Nat. Hist. TIT. 2: 43.
1858, Contrib. Bot. 1: 133. pl. 27A. 1861; Tr. & Pl. in Ann. Sci. Nat. IV. 17: 24.
1862; Cortés, Fl. Colomb. 1: 86. 1898; Pilger in . Aa Nat. Pfl. Nachtr. 2: 108.
1906; U.S. Bur. Pl. Industr. Pl. Immigr. 132: 11

aida | gramadensi hae Syst. Veg. ed. 14. 507. ae. Willd. ee Pl. 2: sere ‘ad

_ Pl. 2: 84. 1806; Humb. & Bonpl. Pl. Aequin. 1: . pl. 58.

22 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

Drimys Winteri {. granadensis Eichl. in Mart. FI. Bras. ena’ 135. pl. 31, f. 1. vets
Drimys Winteri var. sages sues Eichl. ex Dusén in Arc us. Nac. Rio 13:
excl. spec. 1905; Pittier, Man. Pl. Us. Venez. 159. 1926
Tree (or sometimes shrub) up to 13 m. high, hie branchlets stout (3-7
mm. in diameter toward apices) ; leaves scattered along branchlets; petioles
5-15 mm. long, often stout (1-3 mm. in diameter) ; leaf-blades coriaceous.
or thick-coriaceous, usually shining and fuscous above when dried, elliptic-
oblong, (5—) 7-14. 5 cm. long, (1—)1.5-5.5 cm. broad, obtuse or subrounded
at apex, usually sharply but narrowly recurved at margin, often strongly
so toward base, the secondary nerves 8-19 per side, ascending or erecto-
patent at an angle of 40-60°, prominulous or nearly plane above, prominu-
lous and obvious (rarely obscure) beneath, freely anastomosing towar
margin, the veinlets immersed above, faintly prominulous or obscure be-
neath; Ex site ae near apices of branchlets but not crowded at actual
apex, umbellate or sometimes 1-flowered or fasciculate, the peduncle
16- 50(_ 90) mm. long, the flowers 3—6 per inflorescence (or single), the pedi-
a 15-60(—80) mm. long; sepals papyraceous, 6—10(—12) mm. long, 7—
11(-13) mm. broad (or slightly narrower when 3 rather than 2); petals
8-13(-15), 9-20(-25) mm. long, 4-8(—11) mm. broad; stamens 25—
50(—60), the filaments 1.2—-3 mm. long, the locules 0.6-1 mm. long; carpels
5-12(—14), 2.5-5 mm. long at anthesis, the ovules 7-12.
DistrisuTION: Mountainous parts of Colombia, in all three Cordilleras, in temperate
forests, shrub-zone (paramillo), or on slopes of paramos, at altitudes of 1500-3300 m.;
in parts of the range, especially in Cundinamarca. It is reported as eae
ear the tree- ~ a often found in association with such typical paramo plants a
reer of Espele
COLOMBIA: = orte de Santander: Between Pamplona and Toledo
(divide between sage erate and Orinoco drainage), Killip & Smith 19897 (A, GH, NY,
US); Santande Western slope of Paramo Rico, Killip & Smith 17817 (A,
GH, NY, US); uci ieee: Vicinity of Bogota, Triana (NY, US),
Hartweg 877 (NY), Holton 673 (GH, NY), Dawe 140 (US), Ariste-Joseph (US),
Schultze 14 (US), Cuatrecasas 5440 (US); Guadelupe, above Bogota, Ariste-Joseph
(US); San Miguel, W. of the savanna of sg a Cuatrecasas 6687 (US); Monserrate,
near Bogota, Garcia 4813 (US); Fuentes de San Francisco, Ariste-Joseph A106 (GH,
US) ; Sibaté, Popenoe 1109 (US) ; western oe of Paramo de Cruz Verde, Cuatrecasas
343 (US); Paramo de Guasca, Garcia 6278 (US), Killip 34077 (A, US), Balls 5749
(US) 5 above ‘Ubague, Pennell 1897 (GH, NY, US); Usaquén, Garcia weer (US);
“Rosalito,’ near Paramo de Ruiz, Pennell 2963 (GH, M, NY, US);
eae te Vicinity of Medellin, Toro 718 (NY); pent Elena, yee 1202
(US); San Pedro, Daniel & Tomds 1255 (A), 1305 (F); Caldas: “Pinares,
above Salento, Pennell 9222 (GH, NY, US); El Cauca: Rio Paez — Pittier
1367 (US); Mt. El Trueno, Pennell 7542 Leas “Paletara,”’ Pennell 6945 (GH, NY,
US); without definite lo ty : Mutis 1049 (US), 2525 (US),
3839 (TYPE COLL. of D. granadensis L. f., ee "4483 (US), 4609 (US), Purdie (GH).
ATIVE NAMES: Canelo, canela de pdramo, aji, palo de aji, quinon, cupis. Several
writers have briefly mentioned the tonic and stimulant qualities of this plan
The first varietal name which appears referable to the biological entity
described above is Hieronymus’ var. grandiflora. The original description
of this variety clearly indicates that the typical form of the species (as
1In sorting the Mutis collection in the Madrid Herbarium, Mr. E. P. Killip wer pres

arbitrary numbers to the specimens. He has selected the best Mutis specimen of t
species as the type; all the Mutis numbers here cited are essentially identical.

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS 23

represented by Mutis’ material) was under consideration. The type of
the variety is Lehmann 7469 from Sibaté, Dept. Cundinamarca; this lard
must be construed also to include the Mutis type of the species. Dusén’
reference to D. Winteri var. granadensis, although accompanied by Hite:
of Brazilian specimens, appears to be the first use of the epithet granadensis
as a variety; Eichler’s earlier use of this epithet in a trinomial was as a form.
3b. Drimys granadensis var. uniflora (Turcz.) comb. n
oa uniflora Turcz. in Bull. Soc. Nat. Mosc. 27(2) : "280. 1854; Pilger in E. & P.
. Pfl. Nachtr. 2: 108. 1906; Knuth in Rep. Sp. Nov. Beih. 43: 331. 1927.
eet tree (?), the branchlets 2-5 mm. in diameter toward apices; leaves

coriaceous, olivaceous to pale brown above when dried, elliptic-oblong,
5.5-10.5 cm. long, 1.5—3.5 cm. broad, rounded or obtuse at apex, narrowly
recurved ater more conspicuously so toward base, the secondary
nerves 8-12 per side, erecto-patent at an angle of about 45°, usually im-
mersed and obscure on both surfaces, sometimes faintly prominulous be-
neath, the veinlets immersed; flowers single, clustered at apices of branch-
lets, the pedicels 25-50 mm. long: sepals submembranaceous, about 8 mm.
ong and 9 mm. broad; petals about 10, 9-12 mm. long, 3-5 mm. broad;
stamens about 30, the filaments carnose, somewhat flattened, 1—-1.5 mm.
long, the locules 0.5—0.6 mm. long; carpels about 6, 2—2.5 mm. long at
anthesis, the carpel-wall densely glandular, the ovules 9 or 10.
DisTRIBUTION: State of Trujillo, Venezuela; known only from the type collection.
VENEZUELA: Trujillo: Near Agua d’Obispo, alt. about 2700 m., Linden 1444
(TYPE COLL., F, GH).
As pointed out in the preceding diagnoses of varietal characters, var.
uniflora differs from the typical Colombian variety chiefly in its more
slender habit, obscure secondary nerves, single and terminal flowers, thinner
sepals, slightly smaller petals, smaller anther-locules, and smaller carpels at,
anthesis. On the basis of the present material, the Trujillo specimen is
maintained as a variety, but its status cannot be satisfactorily decided with-
out more ample material from the Venezuelan Andes. The weakness of
such a character as single versus umbellate flowers is indicated by such a
collection as Killip & Smith 19897, from the part of Colombia adjacent to
Venezuela. This collection has some branchlets with ample and obviously
umbellate inflorescences, while other branchlets have the flowers con-
sistently solitary, either fasciculate or single. The leaves of this collection
also tend to resemble those of Linden 1444, having the secondaries only
weakly apparent; nevertheless the inflorescences (or single flowers) are
scattered along the branchlets and not aggregated at apices.
3c. Drimys granadensis var. mexicana (DC.) comb. n
Drimys mexicana DC. Reg. Veg. Syst. Nat. 1: 444. re “Moc. & Sessé ex DC. Prodr.
1: 78. 1824; Hemsl. Biol. Centr. Am. Bot. 1: 14. 1879

Drimys granadensis var. nip a? paper aane & Cham. in Linnaea 5: 211.
1830; Hemsl. Biol. Centr. Sti . 1: 14. 1879; non St. Hil.

Drimys Winteri sensu Goyen nl ee ee 1909; Standl. in Contr. U. S. Nat
Herb. 23: 276. 1922, in Field Mus. Publ. Bot. 18: 438. 1937; non J. R. & G. Forst.

Shrub or tree 2-13 m. high; branchlets 2-5 mm. in diameter near apices;

24 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

ii usually scattered on branchlets; petioles (5—)8—25 mm. long, 1-2

n diameter; leaf-blades subcoriaceous, shining or dull and fuscous or
dark ne above, narrowly elliptic-oblong, (5— )7-16(-17.5) cm. long, 1.5—
4.5(—5.5) cm. broad, obtuse or subacute at apex, slightly recurved at
margin, often strongly but narrowly revolute toward base, the secondary
nerves (8—)10-16 per side, ascending or erecto-patent at an angle of 40—
55(—60)°, prominulous or nearly plane above, usually obvious, prominu-
lous and obvious beneath, freely anastomosing near margin, the veinlets
immersed or faintly prominulous; inflorescences aggregated at or near
apices of branchlets, umbellate or rarely 1-flowered, the peduncle 7—
65(—75) mm. long, the flowers (1—)2—6(—8) per inflorescence, the pedicels
5-60 mm. long (up to 90 mm. when flowers are single); sepals submem-
ipeaiiei to papyraceous, 4.5-8 mm. long, 5-9 mm. broad; petals 9-17,
6-17 mm. long, 1.5—-6 mm. broad; ace 25—45(-55), the filaments
he yellow-glandular, often eglandular, 0.7-3 mm. long, the locules

.5-0.8 mm. long, lateral to oblique; carpels 5—i2, 2-3 mm. long at anthesis,
the ovules 7-12. (Fic. 3, b-f.)

BUTION: Southern Mexico (Veracruz and Guerrero) to Costa Rica, in moun-
tains at altitudes of 1150-3000 m.; usually occurring in moist forest, but noted in oak
and pine forest in Guerrero by Hinton; common in Costa Rica.

Mexico: Veracruz: Inter Huatamalco et Tinzutlan (Teziutlan?), Liebmann
1984 (US); Guerrero: Distr. Galeana, Teotepec, Hinton 14441 (F, GH);
Distr. Galeana, Piedra Ancha, Hinton 14235 (GH); Oaxaca: Distr. Feotitlan,
Cumbre de los Frailes, Conzatti 2101 (F, US); Distr. Cuicatlan, Cerro La Raya,
Cuyamecalco, Conzatti 3475 (US); northwestern slope of Mt. Zempoaltepec, Nelson
666 (US); Chiapas: Ghiesbreght 117 (GH, NY); Cerro de Huitepec,
Ghiesbreght 518 (GH. M); Saxchanal, Sierra Madre, Matuda 4287 (M, NY)

78 de Tanetze, Hartweg 444 (NY)

Gu apa: Summit of Sierra de las Minas, vicinity of Finca \Plan-
ados, aaa . (F).

NicaraGcua: No specimens seen; the fact that Goyena mentions D. “Winter?” in his
Flora ie hen bra indicates that he probably saw a specimen of the present variety
from that c

Costa ete Paramos del Abejonal, Tonduz 7897 (US); Tijar, Quirds 152 (F); El
Roble, Stork 2034 (F); Alajuela: Palmira, Alfaro Ruiz, Austin Smith 4139
(F), 4123 (F, M), H524 (F); vicinity of Fraijanes, Standley & Torres 47632 (US);
Volcan de Poas and vicinity, Tonduz 10937 (US), Standley 34604 (NY, US), Stork
2503 ee aig 599 (A, F); Vara Blanca de 7 Skutch 3585 (M, NY);
Here : Volcan de Barba, ee 2130 (US); Cerro de Las Lajas, orth of San
Isidro, es & Valerio 51443 (A, US); Cerros de Zurqui, northeast of San Isidro,
Standley “a Valerio 50414 (US), 50617 (US), 50635 (US); San José: La Palm

e San Ramon, Brenes 4078 (F), 4122 (F), 4452 (F), 5718 (F) ; ty Rosa de Copey,
Preis pie (12174) (GH, NY, US); above Los Lotes, north of El Copey, Standley
42578 (US), 42798 (US); Las Nubes, Standley 38789 ( a Valerio 1398 (F);
Laguna de la Chonta, northeast of Santa Maria de Dota, Standley 42319; Cerro de las
Vueltas, Standley & Valerio 43654 (US); north of El Alto de Cabeza de Vaca, on Rio
Sucio, Dodge & Thomas 4948 (M); a del gains Pittier 7338 (12300) (GH, Set
Cerro Gallito, Valerio 1005 (F); Ca ago: Cartago and vicinity, Stevens 79
(US), Stork 367 (US), 404 (US); Alto oe La Estrella, Standley 39058 (US); satiate
slope of Volcan Irazt, Standley 36628 (US); Santa Clara, Torres 185 (F).

CULTIVATED SPECIMEN: Zabel (cult. Bot. Gart. Muenden) (A).

TIVE NAMES: Chilillo, chachaca, palo picante, palo de chile, muelo (in Mexico) ;
quiebra-muelas, chile, muelo (in Se Rica). Standley reports on native medicinal
uses in the cited discussions of D. Winteri

1943 | SMITH, AMERICAN SPECIES OF DRIMYS 25

While this entity does not appear to merit specific recognition, it never-
theless has certain tendencies which permit its separation from the Colom-
bian variety. In general, it is more slender in habit, with thinner and pro-
portionately narrower leaf-blades, these tending to be more pointed at apex.
The sepals are usually thinner in texture and smaller on the average, the
petals are more numerous (on the whole) and somewhat smaller, the
anther-locules average smaller, and the carpels are smaller at anthesis.

The above-mentioned characters are far from satisfactory as varietal
criteria, but nevertheless the two entities, when ample material is examined,
give different impressions, and one is generally able to place specimens
without knowledge of the geographic source. There appear to be no differ-
ences of consequence between Mexican and Costa Rican specimens.

_ 3. a. Drimys granadensis var. chiriquiensis, drawn from the type: flowering
branchlet, x 14. b-f. Drimys granadensis var. mexicana, drawn from Skutch 3585:
b. flowering branchlet, x 1%; c. flower, with some petals removed, x 1, d. stamens,
extrorse and introrse views, X 5; e. carpel, x 5; f. carpel, longitudinal section, x 5.
g-j. Drimys brasiliensis var. campestris, drawn from Mexia 5791: g. flower, with some

petals removed, - h. stamens, extrorse and introrse views, X 5; 7. carpel, xX 5;
j. carpel, longitudinal section, < 5. k-m. Drimys brasiliensis var. roraimensis, drawn
from the type: k. stamens, extrorse and introrse views, x 5; /. carpel, X 5; m. carpel,
longitudinal section, x 5.

3d. Drimys granadensis var. chiriquiensis var. nov.
Frutex vel arbor parva, ramulis apicem versus 2-4 mm. diametro, saepe
conspicue cicatricosis; foliis apicem ramulorum versus plerumque confertis;

sicco olivaceis vel pallide brunneis, elliptico-oblongis, 6-12 (—14) cm. longis,

immersis; inflorescentiis apicem ramulorum versus axillaribus umbellatis
vel raro 1-floris, pedunculo 10-45 mm. longo, floribus (1—)2—4 per inflore-

26 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

camp pedicellis 7-30 mm. longis; sepalis submembranaceis, 5-6 m

long , 6-7 m mm. latis; petalis 14-25, 5-8 mm. longis, 1.5—2.5 mm. ae
Pe 50-65, 3- = seriatis, filamentis eglandulosis 0.5—1.2 mm. longis,
connectivo glandulas luteas 2 vel 3 interdum obscuras apice gerente, loculis
0.3—0.5 mm. longis; carpellis 8-24 falcato-ellipsoideis vel obovoideis sub
anthesi circiter 2 mm. longis, ovulis 10-12. (Fic. 3, a.

DisTrRIBUTION: Known only from Chiriqui, Panama.

PanaMA: Chiriqui: Bajo Chorro, Boquete District, alt. about 1800 m., David-
son 127 (A, TYPE, F), Jan. 14, 1938 (shrub or tree, in rain-forest; petals white; stamens
yellow), Davidson 328 (A, F, M) (small tree, in rain-forest); between Alto de las
Palmas and top of Cerro la Horqueta, alt. 2100-2268 m., Pittier 3232 (US) (in humid
forest)

The biological entity described above is more closely related to var.
mexicana than to other varieties of D. granadensis. It is characterized by
its comparatively small and aggregated leaves and small floral parts, but
principally by the unusual number of its petals, stamens, and carpels. The
type specimen is extraordinary in having as many as 25 petals, 65 stamens,
and 24 carpels, but other cited specimens show that substantial variation
in these numbers is to be expected. On the whole, the Chiriqui specimens
show such definite trends in the direction indicated above that they seem
worthy of varietal recognition.
3e. Drimys granadensis var. peruviana var. nov.

Frutex ad 70 cm. altus, ramulis apicem versus circiter 4 mm. diametro;
foliis secus ramulos dispositis, petiolis 6-11 mm. longis, 1-2 mm. diametro,
basim versus paullo incrassatis; laminis coriaceis, supra in sicco olivaceis et
nitidis, oblongo-lanceolatis, 5-9 cm. longis, 1-2 cm. latis, apice obtusis,
margine anguste revolutis vel abrupte recurvatis, nervis secundariis utrinse-
cus 12-15 angulo 45—50° erecto-patentibus supra leviter prominulis vel
immersis subtus plerumque prominulis, venulis immersis; inflorescentiis
ramulorum apice vel apicem versus aggregatis umbellatis, pedunculo crasso

mm

longis; sepalis submembranaceis, 8-9 mm. longis et latis; petalis circiter
22 et 3-seriatis, exterioribus 10-13 mm. longis et 5-7 m m. latis, interioribus
paullo minoribus; staminibus circiter 40, filamentis LS mm. longis,

i mm.

longis; carpellis 3—7, sub anthesi 2—2.5 mm. longis, ovulis plerumque 12.
sdugessnaabeleuanee Known only from the type collection, Dept. Cajamarca, Peru.

PERU: Cutervo, trail between Socota and Tambillo, alt.
3200 m., Stork & Horton ee “(UC TYPE), Dec. 14, 1938 caaeiion: shrub to 70
cm. high, in shrub-land; petals white; stamens yellow; leaves nearly pure white be-
neath; only one specimen observed)

The single Peruvian specimen described above is characterized by narrow
oblong-lanceolate leaf-blades, numerous and comparatively short and broad
petals, and comparatively few and small carpels. Its position in D. grana-
densis is indicated by such characters as the glandular anther-connective and
the texture of the lower leaf-surface. It seems amply distinguished from
the bulk of the population referred to this species, but this conclusion
should be further checked when additional collections of Drimys are made
in Peru and Ecuador.

1943] SMITH, AMERICAN SPECIES OF DRIMYS 27

4. gine peru Miers in Ann. Mag. Nat. Hist. III. 2: nee 1858, Contrib. Bot.
1: 1861; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108
oe sen D sensu St. Hil. Pl. Us. Bras. pl. 26-28. oe ot ‘Bras. Merid. 1: 24.
1825; Spach, Hist. Nat. Veg. 7: 437. 1839; non L. f.

Shrub or small tree, the branchlets brownish or cinereous (or suaagietrs
when young), rugulose, subterete, comparatively slender (1.5—5 mm. in di-
ameter toward apices) ; petioles rugulose shallowly canaliculate; leaf. ade:
coriaceous or thin-coriaceous, glaucous beneath or at length pale brown,
often papillate beneath with mate ae shaped or knob-like papillae,

nearly lanceolate, rarely exceeding 11 cm. in length and 4 cm. in breadth,
acute or attenuate at base and decurrent on the petiole, ee or aiieere

umbellate (flowers sometimes single or fasciculate), the peduncle less than
oi mm. long, the flowers up to 6 per inflorescence, the pedicels 5-40 mm.
g (50-80 mm. in var. roraimensis); sepals membranaceous to papyra-
Re usually obscurely glandular, 4-7 mm. long, 4-8 mm. broad, rounded
to apiculate at apex; petals 8-14 (rarely to 20), sony eer ‘opaquely
yellow-glandular, oblong to elliptic-oblong, 6-17 mm. long, 2-6 mm. broad,
obtuse at apex; stamens 18-50, 2—4-seriate, the filaments carnose, pean
flattened, 0.5—2.5 mm. long, the connective glandular, i conspicuously
ial -glandular at apex, the locules ellipsoid, 0.4—0.8 mm. lo ong: carpels
9(—13), obovoid, 1.5-2.5 mm. long at anthesis, rounde led a apex, the
in lateral near apex or subterminal, usually short- ee (sessile in
var. roraimensis), the ovules 6—12, on short placentas

DIsTRIBUTION: Southeastern Brazil ne adjacent Paraguay aid Argentine, with one
variety from Mt. Roraima on the boundary of Venezuela, Brazil, and British Guiana ;
four varieties are recognizable. The type of D. granadensis var. campestris St. Hil.,
for reasons pointed out below, is here selected as the type of D. brasiliensis and its
typical variety (var. campestris).

The Brazilian population of Drimys was first treated by St. Hilaire, who
considered his Brazilian specimens to be conspecific with the Colombian ones
(D. granadensis) and different from the Chilean- Magellanic ones (D.
Winteri). St. Hilaire recognized four Brazilian varieties. He considered
the most common of these to be var. campestris, which is described in detail,
whereas the remaining varieties are discussed only as to their points .
difference. As long as these varieties are considered under D. granadensis,
the question of the typification of a Brazilian species does not arise. But
Miers, in erecting D. brasiliensis, founded it upon St. Hilaire’s concept
(exclusive of var. montana, which Miers considered specifically distinct )
without designating which of St. Hilaire’s varieties he wished to accept as
the type.

The Brazilian plant has not generally been considered as a species, and’
as far as I can ascertain no author has attempted definitely to typify it.
Hauman was content to accept it as outlined by Miers, with the difference
that he again submerged var. montana. Since no author has selected a type
for D. brasiliensis, it therefore becomes necessary to do so. St. Hilaire clearly

28 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

considered var. campestris to be the most common of his Brazilian varieties,
and since this variety is the only one he fully described, I believe that it
may be taken as the principal basis of St. Hilaire’s concept and considered
the type of D. brasiliensis Miers.

Var. campestris is said to grow in Minas Geraes, but no definite locality
is given. It grows “dans les lieux découverts (campos), le plus souvent
sur le bord des ruisseaux.”’ A specimen in the herbarium of the New York
Botanical Garden, collected by St. Hilaire and marked “Drymis Granatensis
var. campestris,” may thus be taken as a portion of the type collection of
D. brasiliensis Miers.

St. Hilaire’s second variety, var. sylvatica, is said to differ from var.
campestris in having its leaf-blades proportionately narrower and its petals
larger. A comparison of St. Hilaire’s two plates discloses only inconsequen-
tial individual differences. A portion of the type collection of this variety
(NY) also shows that the differences between it and var. campestris are
trivial. I have no hesitation in submerging var. sylvatica.

St. Hilaire’s third variety, var. axillaris, is said to occur in the vicinity of
Villa Rica (‘‘sur les montagnes ferrugineuses’”). It is not illustrated, nor
have I seen specimens referable to it, but I find only the most inconse-
quential individual differences pointed out in St. Hilaire’s description.

St. Hilaire’s fourth variety, var. montana, is said to occur in the moun-
tains of the Serra Negra, on the boundary of the States of Minas Geraes
and Rio de Janeiro. It is said to be characterized by small leaves and re-
duced (and sometimes 1-flowered) inflorescences, which are axillary rather
than terminal. These differences also appear to be trivial, and I find no
reason to maintain the variety.

The Brazilian specimens were next considered by Miers, who proposed
the name D. brasiliensis to include three of St. Hilaire’s four varieties
(excluding var. montana). The varietal names were transferred from D
granadensis to D. brasiliensis by Miers without comment. Var. montana
was erected. as a distinct species by Miers, who differentiated it on the
grounds that “. . . it differs from D. brasiliensis, not only in the size and
form of its leaves, but in its axillary inflorescence, and in the number of
the parts of its smaller flowers.” These differences appear to be entirely
inconsequential, the matter of axillary versus terminal inflorescence being
susceptible to great variation in Drimys, and the number and size of floral
parts being dependable only within very broad limits. The variation in the
size and shape of leaves is often conspicuous even on the same individual.
Therefore I have no hesitation in referring D. montana to D. brasiliensis
as a synonym.

The remaining Brazilian specimens available to Miers were placed in
two species, D. retorta and D. angustifolia. Both are very much reduced
in foliage and appear to be reasonably good varieties of D. brasiliensis.

Eichler considered the genus as a whole, referring all the American speci-
mens to D. Winteri. He considered the bulk of the Brazilian material to
be identical with the Colombian plants, which he referred to the forma
granadensis. However, he recognized Miers’ two species D. retorta and

1943 ] SMITH, AMERICAN SPECIES OF DRIMYS a

D. angustifolia to the extent of keeping them as forms of D. Winteri (the
first as f. revoluta) ; both are well illustrated in Flora Brasiliensis.

Hauman, in discussing the occurrence of D. brasiliensis in the Argentine,
recognized two varieties, based respectively on St. Hilaire’s varieties
compestris and montana. The differences pointed out by him seem inconse-
quential when an extensive series of collections is considered.

On the basis of the material available to me, and taking into considera-
tion previous descriptions and illustrations of this complex, I believe that
the bulk of the Brazilian material may be referred to a single a
designated as var. campestris, the typical variety of the species. Miers
two species D. retorta and D. angustifolia are given varietal rank, and a
fourth very distinct variety is described on the basis of collections from
Mt. Roraima.

ESSENTIAL DIAGNOSTIC CHARACTERS OF THE VARIETIES

Leaf- oe narrowly a ia peo or oblong or elliptic, (3—)4-11(-14) cm. long,
(0.8-)1.2-4(-5) cm. broad

florescences usually terminal and umbellate, the peduncle up to 40 mm. long, the
pedicels 5-40 mm. long; stamens spe a truncate connective; stigma short-stipitate ;
Bahia to Paraguay, Misiones, and Rio Grande do Sul.......... 4a. var. campestris.
Leaf-blades narrowly elliptic- or Pigne: lanceolate, (2—)2.5-5.5 cm. long, (0. 2—)0.3- 1. 5

°
_
es

appearing ries the peduncle inconspicuous, up ong, the pedicels

15=, ng; Snes with a truncate connective; stigma short-stipitate; Minas

and Sac 01 C0 one wm Re es Sed cs cot es Oracle ee eae ree 4b. var. retorta.

Leaf- blades ce or linear- lanceolate, 4-10 cm. long, 0.5 —0.6 cm. broad, obtuse or
faintly

p
fe)

usually single; southeastern Brazil ................. ar. eon

Leaf-blades oblong-elliptic, 6-11 cm. long, 2.5-5 cm. broad, obtuse or rounded at apex,

faintly recurved at margin, the secondary nerves usually prominulous on both sides,

spreading at an angle of 55—70°; inflorescences axillary, the flowers single or paired

on a slender peduncle, the pedicels 50-80 mm. long; stamens with the connective
apiculate, densely glandular, and exceeding the locules by about 0.15 mm.; stigm

Sessile s Vit, GROLAlM Aa. -<-2., acesetice icine the apeeehere setts cue sea) (ayers 4d. var. VOrGGERSES.

4a. Drimys declares var. campestris (St. Hil.) Miers in Ann. Mag. Nat. Hist. III.
2: 48. 1858, Contrib. Bot. 1: 137. pl. 25B (as D. brasiliensis). 1861; Hauman in
Cera ee Nac. Buenos Aires 2: 50. 1923; Hauman & Irigoyen in An. Mus
Nac. Buenos Aires 32: 226

Drimys heraae var. campestris St. Hil. Pl. Us. Bras. pl. 26. 1825, Fl. Bras.
Merid. 1 . 1825.

Drimys Petes var. sylvatica St. Hil. Pl. Us. Bras. pl. 27. 1825, Fl. Bras. Merid.

Drimys eianddemd: var. axillaris St. Hil. Pl. Us. Bras. sub pl. 26. 1825, Fl. Bras.
Mer 82

ene Pandenae var. montana St. Hil. Pl. Us. Bras. pl. 28. 1825, Fl. Bras. Merid.

Pans Winteri sensu Vell. Fl. Flum. 240. 1825, Fl. Flum. Ic. 5: pl. 132. 1827; non
eRe rst.
Drimys aa aes sensu Spach, Hist. Nat. Vég. Phan. Atlas, pl. 60. 1846; non L. f.

30 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Drimys sig sete mies ra Miers in Ann. Mag. Nat. Hist. III. 2: 48. 1858,
Contrib. Bot. 1: 137.
Drimys ene te sp pee Miers in Ann. Mag. Nat. Hist. III. 2: 48. 1858,
Contrib. Bot. 1: 1861.
Drimys pessoa a. in Ann. Mag. Nat. Hist. III, 2: sees 1858, Contrib. Bot. 1:
133. 1861; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108.
Drimys Winteri var. semiglobosa Dusén in Arch. Mus. Pe a 13: 1905.
Drimys brasiliensis var. montana Hauman in Comun. Mus. Nac. Buenos Aires 2: 50.
1923; Hauman & Irigoyen in An. Mus. Nac. Buenos Aires 32: 227.
Shrub or tree, up to 13 m. high, the branchlets 2—5 mm. in a
toward apices; leaves scattered or crowded along distal portions of branch-
lets, usually evenly distributed and not clustered in whorls; petioles 3—

leaf-blades coriaceous or thin-coriaceous, pale brown to greenish and usually
shining above when dried, narrowly elliptic-obovate or oblong or elliptic,
(3—)4-11(-14) cm. long, (0.8—)1.2-4(—5) cm. broad, rounded or broadly
obtuse or faintly emarginate at apex, narrowly recurved to conspicuously
revolute at margin, often more obviously so toward base, the secondary
nerves 6—-10(—12) per side, erecto- -patent at an angle of 40- 50°, scarcely

spicuously anastomosing toward margin, the veinlets immersed;
rescences usually aggregated at apices of branchlets, rarely axillary, um-
bellate (flowers rarely single, occasionally fasciculate), the peduncle up
to 40 mm. long, usually obviously flattened, the flowers (1—)2—6 per inflo-
rescence, the pedicels 5-40 mm. long; sepals membranaceous or sub-
membranaceous, obscurely opaque- or pellucid-glandular, suborbicular or
deltoid-orbicular, 4-7 mm. long, 4-8 mm. broad; petals 8-14 (rarely to
20), opaque-yellow-glandular or very sparsely so, 7-17 mm. long, 2—5 mm.
broad; stamens 20—40(—50), the filaments 0.5-2.5 mm. long, the con-
nective yellow-glandular (usually conspicuously so at apex, sometimes
scarcely so); carpels 3—8(—13), the stigma conspicuous, on a stipe usually
0.2—0.5 mm. long, the ovules 6-12. (Fic. 3, g-j.)

DistrRIBUTION: Southeastern Brazil (Bahia to Parana [and Rio Grande do Sul,
according to Eichler]), adjacent Paraguay, and Misiones in northern Argentine, at
altitudes between 800 and 1250 m. (according to collectors’ incomplete data, but prob-
ably found both lower and higher); occurring in forests, woods, or campos, often on
shores of streams; one to be fairly common in parts of Minas Geraes.

Brazit: Bah : Rio de Contas, Bom Jesus, Liitzelburg 268 (NY); Minas
Geraes: St. Hilaire (TYPE COLL. of D. granadensis var. campestris, NY), St. Hilaire
(TYPE COLL. of - granadensis var. sylvatica, NY), Gardner 4402 (M, NY, US),
Claussen (F, NY), 1064 (GH); Rio Tejuco, Ackermann [Mart. Herb. Fl. Bras. 288]

H, M, NY), Vauthier 489 (GH) ; Diamantina, Olaria, Mexia 5791 (A, GH, M, ahs,

Mun. Santa Luzia, Mello Bites 7452 (F); Caldas, Regnell 145 (US), Mosén 331
(NY); Rio de Janeiro: Vargem, Organ Mts., Miers 4606 (US); Sao
Paulo: Lund (NY); Serra de Cubatao, oe ig (GH, NY); Alto da Serra,
Hoehne 1205 (A, US); Butantan, Hoehne 3839 (GH); Jardim Botanico, Sao Paulo
Hoehne 28700 (F, NY); Parana: _ Pinhaes, pees

definite locality: Sellow (M), Burchell 4748 (GH), “Riedel (A), Collector?

Me
ParRAGUAY: Sierra de Amambay, eck 9992 (A), 10586 (A).
RGENT cimens seen, but those cited by Hauman indi-
cate the occurrence of the variety in Misio

1943] SMITH, AMERICAN SPECIES OF DRIMYS 31

NATIVE NAMES: Casca d’Anta, ll curvillo (in Brazil). St. Hilaire has discussed
native uses of this plant, and his data are repeated by subsequent writers

Collectors of the above- fie: specimens note this plant as a inh or tree,
as low as 50 cm. high in open situations on campos and up to 13 m. high in
the forests. Probably this difference in habit has been the principal reason
why St. Hilaire and other writers have distinguished varieties within this
entity, which, on the basis of the material I have seen, cannot satisfactorily
be further divided.

The majority of the specimens which I refer to D. brasiliensis is charac-
terized by having the lower surfaces of the leaf-blades distinctly papillate.
Papillae are obviously present in the specimens of var. retorta and var.
roraimensis, but they are lacking in the following specimens of var. cam pes-
tris: Lutzelberg 268, Claussen 1064, Miers 4606, Burchell 3567, Hoehne
1205 and 28700, and Dusén 14504. The remaining specimens of var. cam-
pestris have the lower leaf-surfaces clearly papillate. Whether the presence
or absence of such papillae is a fundamental character cannot be stated at
present, but quite possibly the population under discussion should be broken
up into varieties on this basis. On the other hand, the papillate and the
non-papillate specimens present no geographic pattern, nor is this char-
acter accompanied by any other. In leaf-size, number and size of floral
parts, etc., there is considerable variation within var. campestris, but such
variation in different organs is not correlated and therefore, in the present
state of our knowledge, not usable for further subspecific division.
4b. Drimys brasiliensis var. retorta (Miers) comb. n

Drimys retorta Miers in Ann. Mag. Nat. Hist. III. 2: a0 ee Mess Bot. 1: 134.

pl. 260B. 1861; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108.

Drimys Winteri {. revoluta Eich]. in Mart. Fl. Bras. 13(1): one # Be f. 2. 1864.

Drimys ledifolia Eichl. in Mart. FI. Bras. 13(1): 136, as synonym.

Shrub to 3 m. high, the branchlets 1.5—4 mm. in diameter ok apices;
leaves crowded on branchlets, especially distally ; petioles 3-11 mm. long,

./-1.5 mm. in diameter; leaf- blades coriaceous, dark green above when
dried, narrowly elliptic- or oblong- lanceolate, (2—)2.5-5.5 cm. long,

growing points, appearing axillary, fabeete (or siete sometimes single) ,
the peduncle rugulose, inconspicuous, up to 15 mm. long or obsolete, the
flowers (1—)2—6 per inflorescence, the pedicels sear 15-33 mm. long:
sepals papyraceous, opaque, obscurely glandular, suborbicular- ovate, 5-6
mm. long, 4-7 mm. broad; petals 8-11, conspicuously opaque- yellow-
glandular, 6-12 mm. long, 25 mm. broad; torus cylindric, conspicuous;
stamens 30-40, the filaments 1.5-2.5 mm. long, eglandular, the connective
copiously and minutely yellow-glandular at apex; carpels 3—7, the stigma
short-stipitate (stipe 0.3-0.7 mm. long), the ovules 6 or 7.

DistriBuTION: Minas Geraes, Brazil; probably restricted in area. Miers also cites
a collection by Bowie from Sao Paulo. Eichler cites Claussen 340 and Sellow, both
from Minas Geraes

Brazit: Minas Geraes: Claussen s. n. or 1503 (coTypE COLL., F); Serra do

32 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Itacolomy, Mun. Ouro Preto, Mello Barreto 9083 (F) (shrub 3 m. high, common;
flowers white).

A few of the specimens referred to the typical variety of D. brasiliensis
(var. campestris) have unusually small and strongly revolute leaf-blades,
thus indicating a transition between that variety and var. retorta. The
existence of such specimens (e. g. Mello Barreto 7452, 10069, Mexia 5791)
indicates that D. retorta can be accepted as no more than a variety, and
probably not a very stable one.
4c. Drimys brasiliensis var. angustifolia (Miers) comb. n

n. Mag. Nat. Hist. ITI. 2 46. 1858, Contrib. Bot.
1: 135. pl. 260A. 1861; Pilger in E. & . Nat. Pfl. Nachtr. 2: 108. 1906.
Drimys Winteri {. angustifolia Eichl. in Mart. Fl. Bras. 13(1): 136. pl. 31, f. 3.

bd,
|
3
—
Rg
=
os
2
SS
=
i)
>
Q
o
=
wn
oa
=)

Branchlets slender; leaves scattered along branchlets dona Mtg
slender, narrowly winged, 8-10 mm. long; leaf-blades elongate- inear-
lanceolate, 4-10 cm. long, 0.5-0. es cm. broad, obtuse or faintly Rescind
at apex, slightly recurved at margin but essentially plane, the secon dary
nerves entirely immersed and obscure; inflorescences aggregated at apices
of branchlets, the flowers single or rarely paired on short inconspicuous
peduncles, the pedicels about 12 mm. long; sepals ovate; petals 9 or 10;
stamens about 18; carpels about 5

STRIBUTION: Known only from a single unnumbered collection of Sellow from
southeastern Brazil, without detailed locality, cited by Miers and Eichler.

I have seen no material of this plant, the above notes being taken from
the descriptions and illustrations of Miers and Eichler. It appears to be
sufficiently distinct to be recognized as a variety of D. brasiliensis. Both
Miers and Eichler discuss it as an even more extreme form than D. retorta.
4d. Drimys brasiliensis var. roraimensis var.

— granadensis sensu Oliver in cc ia. ‘Soc. II. Bot. 2: 271. 1886; N. E.

Br. in Trans. Linn. Soc. I]. Bot. @: 8. 1901; non
Drimss. ales sensu Ule in Bot. ee 52: Beibl As: 49. 1914; Knuth in Rep.
Sp. Beih. 43: 331. 1927; non J. R. & G.

he parva (?), ramulis apicem versus 2—4 mm. diametro; foliis apicem
ramulorum versus confertis; petiolis 5-15 mm. longis, 0 mm. dia
metro; laminis subcoriaceis, supra in sicco fuscis vel fusco-viridibus,
oblongo- -ellipticis, 6-11 cm longis, 2.5-5 cm. latis, apice obtusis vel
rotundatis, margine obscure. recurvatis, nervis secundariis utrinsecus 7-14

angulo 55—70° patentibus utrinque prominulis vel subplanis, venulis immer-
sis vel utrinque leviter prominulis; inflorescentiis ut videtur axillaribus et
solitariis, floribus singularibus vel pedunculo gracili circiter 35 mm. longo
binis, pedicellis gracilibus 50-80 mm. longis; sepalis papyraceis ovato-
deltoideis circiter 5 mm. longis et 6 mm. latis; petalis 10-12 obscure
glandulosis, 10-16 mm. longis, 3-6 mm. latis: staminibus 35—50, filamentis
sparse luteo-glandulosis 1.5—2 mm. longis, connectivo dense glanduloso in
apiculum circiter 0.15 mm. longum dense et minute luteo-glandulosum
ieee, ee circiter 9, stigmate sessili subterminali, ovulis 10-12.
(Fic. 3, k—-m.)

eusebieas Known only from Mt. Roraima, on the Venezuela-Brazil-British
Guiana boundary.

VENEZUELA (or adjacent countries): Amazonas: Mt. Roraima, The Ledge, im

1943] SMITH, AMERICAN SPECIES OF DRIMYS 33

Thurn 242 (US, Type), 1884; Rondon Camp, upper slopes, alt. about 2100 m., Tate
500 (NY) (in humid temperate forest).

The variety described above is quite distinct from other American
material of Drimys, being characterized by the apiculate and very densely
glandular connective of the stamens. The long pedicels are also note-
worthy, as is the fact that the stigma is essentially sessile, whereas in other
specimens from the northern part of the range of the genus the stigma is
short-stipitate. As my description is based upon only two specimens (of
which the Tate collection is sterile), the rigidity of these characters remains
to be seen. The relationship of this Roraima plant seems to be with the
Brazilian rather than the Andean-Mexican species, as evidenced by the
papillate lower leaf-surface and the rounded or broadly obtuse leaf-apex.
It seems conceivable that this Roraima form may prove to merit specific
status when more ample material is available.

EXCLUDED SPECIES
Drimys VASCULARIS Parment. in Bull. Sci. Fr. & Belg. 27: 229, 306. pl. 11, f. 39, 40. 1896.

In a footnote on page 229 of the above-cited publication, Parmentier
states: “Etiquette: ‘Drimys.— Brésil; Martins; cortex aromaticus.’
(Unicum).” The description is detailed as far as anatomical details are
concerned, but the habit-sketch (f. 39) cannot possibly be taken to repre-
sent a species of Drimys. On p. 229 we read: ‘“‘N’ayant eu ama disposition
que des fleurs mal conservées, trés petites, il m’a été impossible d’en
reconnaitre tous les caractéres; j’ai néanmoins pu m/assurer que ce sont bien
des fleurs de Drimys, surtout 4 cause de la forme de leur calice.” In view
of this uncertainty, Parmentier would have been wiser not to have proposed
the species at all; in fact, had Parmentier’s work in its entirety never been
published, taxonomy would have benefited from the omission. Van
Tieghem’s remarks (11: 284, 285) in this connection are of interest.

PRINCIPAL LITERATURE CITED

1. Battey, I. W. a _C. Smitu. Degeneriaceae, a new family of flowering plants
from Fiji. Jour. geet Arb. 23: 356-365. pl. 1-5.

2. CANDOLLE, A. pe. Magnoliaceae (in part). Reg. Veg. Syst: Nat. 1: 442-446.
1817.

_ —————. Magnoliaceae (in part). Prodr. 1: 78.1
Ercuuer, A. G. Winteraceae. Mart. Fl. Bras. as: 127-140. pl. 30-32. 1864.
5. Hauman, L. Note sur le ‘““Drimys eral Forst. et les espéces voisines. Comun.
Mus. Nac. Buenos Aires 2: 45-52.
6. Hooxer, J.D. Magnoliaceae. FI. ae = 229-230. 1845

WwW

e)
fe Hurcnnson, J. The family Winteraceae. w Bull. 1921: 185-191. 1921.
8. Miers, J. On the Winteraceae. Ann an ag. Dee Hist. III. 2: 33-48, 109-115.
rer

9. On the Winteraceae. Contrib. Bot. 1: 123-145. pl. 25-27.

10. SMITH, A. C. Studies of Papuasian plants, V. Jour. Arnold Arb. 28: Oe:
1942.

_ TrecHEM, P. van. Sur les dicotylédones du groupe des Homoxylées. Jour. de Bot.
14; 259-297, 330-361. 1900.

—
_

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

34 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

PLANTAE PAPUANAE ARCHBOLDIANAE, XI*
E. D. MERRILL AND L. M. Perry
With seven text-figures

THIs paper, consisting of a consideration of the Ranunculaceae and the
Begoniaceae, is a continuation of the series being published on the plants
collected by the Archbold Expeditions in New Guinea; a few plants from
the Solomon Islands are included in the discussion.

RANUNCULACEAE

Clematis Linnaeus
Clematis Gouriana Roxb. ex DC. var. malaiana Miq. Ann. Mus. Bot. Lugd.-Bat.
6. 1869.

Clematis Spies Linn. var. javana (DC.) O. Ktze. Verh. Bot. Ver. Prov, Brandenb.
6: 100. ; K. Schum. & Hollr. Fl. Kais. Wilhelms Land 47. 1889; K. Schum, &
Lauterb. “4 Tans Schutzg. Stidsee 311, 1900.

Clematis Gouriana sensu Kaneh. & Hatus. Bot. Mag. Tokyo 52: 354. 1938, non Roxb.

NETHERLANDS NEW GuINEA: Balim River, Brass 11788, ig one 1938, alt. 1600 m.,
climbing over stone walls in secondary forest. British New GuINeEA: Rona, Laloki
River, Brass 3611, April 1933, alt. 450 m., light rain-forests on valle slopes, climbing
over small trees and bushes (branches eoerumaied: pale crinkled leaves).

In view of the consistent texture and lack of nileaanes the leaves in
this species, as represented by a considerable number of collections from
India and China, we believe it is more advantageous to maintain Miquel’s

variety for this material with leaves less coriaceous and obviously pubescent.
Clematis Seager er DC. Reg. Veg. Syst. Nat. 1: . 1817; Benth, 7 rr 137,
7. Muell. Pap. PL. 2: oe 1886; F. M. oa Queensl. Fl. 1: 899

Chonan a DCHA..2

British NEw GuINeEa: a, . River, Tarara, Brass 8423, December 1936,
scrambling in rain-forest second growths

The leaves of the collection are thinly chartaceous and broader than in
most Australian specimens examined, but the flowers of this staminate
material agree very well with those of most collections so designated. The
sepals have a distinct mucro 0.5—1 mm. long.

Clematis papuasiea sp. nov.

Frutex scandens; ramis leviter sulcatis subteretibus: ramulis hornotinis
parce pubescentibus vel glabratis ; foliis 3-foliolatis glabris vel interdum
pilis brevibus paucis conspersis chartaceis vel subcoriaceis, 10-17 cm.
longis et 5—9 cm. latis; foliolis ellipticis vel ovato-ellipticis utrinque angus-
tatis vel basi rotundat ato-cuneatis, apice acuminatis vel acutis, margine
integris, 5-nervatis, nervis prominulis 2 interdum basim + 7 mm. supra

* Botanical i of the Richard Archbold Expeditions. See Jour. Arnold Arb.
23: 383-416.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 35

ortis, venis manifestis, reticulo inconspicuo; petiolulo 1.5—-2.5 cm. longo
parce pubescente vel glabro: petiolo 8-12 cm. longo; paniculis axillaribus
20-25 cm. ee axi, ramulis et pedicellis subcrispe pubescentibus, pedi-
cellis 1.5-2.5 c _ longis: floribus 1—3-fasciculatis; sepalis 4 patentibus
lineari-oblongis, a 8 mm. longis, 2 mm. latis, apice obtusiusculis apiculatis,
extus dense minute pubescentibus; staminibus 3-seriatis, filamentis 3 mm.

6 cm. longum atten

British NEw a “Thu, Vailala River, Brass 1010, February 1926, climbing
over rain-forest see (large vine; leaves pale and shining) ; Buna District, Lane-Poole
179. Sotomon Istanps: Bougainville: Koniguru, Buin, Kajewski 2174, August

in rain-forest; Ysabel: Meringe, Brass 3168 (typE), November 1932, const to 300

rain-forest on limestone hills (scandent; leaves fleshy and very glossy; flowers
white, faintly perfumed) ; same locality, Brass 3411, December 1932, alt. 200 m., rain-
forest clearing, climbing over low bushes (leaves thick, pale, very glossy, with slechily
recurved margins; fruit pale green).

This material has been passing for Clematis Pickeringti A. Gray. With
several Fijian collections for comparison, including Gray’s fragmentary
type, we find ourselves unable to accept the Papuasian plant here cited as
representing that species. The anther-appendages of C. Pickering are
minute (less than 0.5 mm. long) in all the collections we have seen. In
C. papuasica they are more like those of C. aristata DC. but very narrowly
club-shaped towards the tip, while the leaves are larger and quite different
from those of the Australian species. It should be noted that the plumose
style is shorter in the Papuan material and the achenes are smaller; probably
this is owing to a difference in stage of development.

Clematis phanerophlebia sp. nov.

Frutex scandens inflorescentia excepta glaber; ramulis leviter sulcatis;
foliis coriaceis 3-foliolatis vel interdum simplicibus; ee lanceolato-
oblongis vel anguste ellipticis, 5-13 c m. longis et 2—5.5 cm. latis; costa
valida, nervis utrinque 2 subparallelis, fere a_ basi nee ortis plus

connectivo aristato; ovario ovoideo complanato dense pubescente, stylo
6-7 mm. longo piloso, stigmate recurvo glabro; achaeniis circiter 14
ovoideis, 2 mm. longis, 1 mm. latis, pubescentibus, in stylum persistentem
pilis longis plumosum curvatum + 3 cm. longum attenuatis.

NETHERLANDS NEw GUuINEA: 9 km. ee of Lake Habbema, Brass 10772,
October 1938, alt. 2700 m. pier in secondary forest; Bele River, 18 km. north-

36 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

east of Lake Habbema, Brass 11235 (TYPE), November 1938, alt. 2300 m., occasional
in secondary forest.

These collections are unlike others we have examined in their definitely
coriaceous leaves and the conspicuous venation. Not only are the primary
veins prominent, but the horizontal veins are more obvious than usual. The
staminodes do not have the long clavate appendage characteristic of C.
papuasica but are more like those of C. aristata DC
Clematis Archboldiana sp. nov.

Frutex scandens; ramulis novellis, pect petiolulis, axi inflorescentiae,
pedicellis fulvo- flocculoso- tomentosis; ramulis teretibus; foliis coriaceis
3-foliolatis, superioribus interdum simples foliolis ovatis obtusis basi
rotundatis, 2.5-6 cm. longis et 1.2—3.7 c latis, supra glabratis (minute
flocculosis), subtus dense ferrugineo- Flocctiloso- tomentosis: costa valida,
conspicuis vel subobscuris 2-4 cum costa conjuncto; reticulo subobscuro:
petiolo 1.5—5 cm. longo, petiolulo terminali 1—1.5 cm. longo, laterali 0.7—1
cm. longo; inflorescentiis paniculatis axillaribus terminalibus pedicellis

bus vel staminodiis(?) glabris circiter stylos aequantibus, filamentis com-
pressis linearibus 5 mm. longis, antheris vix 1 mm. longis, loculis marginali-

mm. infra apiculum obtusum positis; achaeniis numerosis (+ 28),
immaturis 2 mm. longis et 1 mm. latis, compressis, aie saa apice in
stylum persistentem een vix 2 cm. longum attenu

NETHERLANDS NEW Gurnea: Bele River, 18 km. northeast of on Habbema, Brass
11366 (TYPE), November 1938, alt. 2200 m., common climber in young secondary rain-
forest.

This collection of pistillate material is past the flowering stage and only
by a careful search of the duplicates were we lucky enough to find a flower
showing the few stamens still adhering. These appear to be sterile, but
at least they give a clue to the type of stamens to be expected in the
staminate plant. The general habit of the species is similar to that of the
other species of the region belonging to the Section Flammula Prantl, but
the crisp somewhat flocculose tomentum is unlike the type of pubescence
in any of the species available for examination.

Ranunculus Linnaeus

Five species of Ranunculus Linn. have been reported from New Guinea.
As usual, we have at hand little material for comparison, and again the
original diagnoses sometimes lack definition of specific characters. We
have been unable to apply the descriptions of R. Muelleri Benth. and R.
tridens Ridl. to any of the collections before us. All the species represented
have flowers borne on leafless scapes, petals with a yellow upper surface, and
sepals spreading or ascending but not reflexed. The species are mostly of
Australian alliance.

Ranunculus perpusillus sp. nov.

Planta pumila acaulis: caudice 1—2 cm. longo fibris vestito; foliis omni-

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 37

bus basalibus; petiolo 1.5 cm. longo (basi vaginante 5 mm. longa inclusa)
patenti-piloso; lamina 5—6 mm. longa tripartita, segmentis integris lineari-
bus vel anguste lanceolatis obtusiusculis interdum consperse Pilosis ; scapis
1-2, unifloris aphyllis patenti-pilosis folia subaequantibus vel in fructu
quam follis longioribus ; sepalis 2.5-3 mm. longis, 1 mm. latis, concavis
+ patentibus apicem versus parce pilosulis; petalis 5 mm. longis, [ 1.2m
latis, oblongo-linearibus, apice obtusis, basim versus paullo angustatis, basi
squamula minuta nectarium tegente praeditis: staminibus 5, circiter 2.5
mm. longis; receptaculo brevi; achaeniis paucis + 4, oblique ovatis com-
pressis, stylo apice tantum leviter recurvo.,

NETHERLANDS NEW GUINEA: 2 km. east of Wilhelmina-top, Brass & Mvyer-Drees
10381 (TYPE), September 1938, alt. 3700 m., amongst short grasses (flowers yellow).

This species somewhat resembles Ranunculus Millani F. v. Muell., of
the Australian mountains and Tasmania, but a comparison with authentic
material of that species, in the Gray Herbarium, collected by von Mueller,
shows a real contrast in floral characters. The peduncle of R. Millani is
much shorter than the petioles, the petals are about as broad as long, and
the plant as a whole is less pubescent.

Ranunculus bellus sp. nov.

Planta nana acaulis; rhizomate brevi; foliis dense rosulatis Pave
lamina oblonga vel late lanceolata, 5 mm longa, mm. lata, utrinque hir
suta, sensim in petiolum basi dilatatum 4 mm. longum e et 2 mm. latum extus
adpresse hirsutum confluente; flore sessili terminali; sepalis uninerviis 5
oblongis, 4 mm. longis, 1.2 mm. latis, extus praecipue apicem versus costa
hirsutis; petalis 5 lineari-spathulatis, 5 mm. longis, 0.5 mm. latis, 3-
nervils, ‘nnectario parvo haud obvio ad medium laminae posito; staminibus
5; filamentis 3 mm. longis; achaeniis novellis 12-15, levibus, 1 mm. longis,
in rostrum subaequilongum tenuem fere rectum angustatis, maturis semi-
ellipticis, 2 mm. longis, 1 mm. hi compressis, in rostrum leviter recurva-

; receptaculo hirsuto subplan

NETHERLANDS NEw GuINEA: 7 km. northeast of Wilhelmina-top, Brass & Myer
Drees 9867 (tvPE), September 1938, alt. 3560 m., gregarious in alpine bogs (flat rosettes
1.5—2 cm. diameter; flowers yellow

This plant may be of the ae of Ranunculus recens Kirk of New
Zealand. Both are small plants, but that of New Guinea is smaller and
densely covered with pubescence, the very small leaves are entire, the scape
is apparently absent, and the achenes are cultriform. In this plant the petiole
as such is practically lacking except for the broad sheathing base.

Ranunculus angustipetalus sp. nov.

Planta parva acaulis; Run valde abbreviato; radicibus fibrosis; foliis
omnibus basalibus; lamina crasse chartacea fere glabra, se tg et subtus
praecipue ad costam ale. oblonga, 1-1.5 cm. longa, 0.5 cm. lata, basi
sensim in petiolum brevem (3-8 mm. longum) + adpresse ea nee angus-
tata, basi vaginante 0.8-2 cm. longa extus dense adpresse hirsuta; scapis
solitariis aphyllis unifloris circiter folia aequantibus parce adpresse hir-
tellis; sepalis 5—6 mm. longis, 2 mm. latis, margine membranaceis, ad apicem
angustatum parce hirsuts petalis anguste oblongis, 9-10 mm. longis,
2.5 mm. latis, 1 mm. supra basim squamula nectarium tegente praeditis;

38 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV
staminibus 15—20 circiter 5 mm. longis; achaeniis + 18, stylo apice leviter
recurvo.

NETHERLANDS NEW GUINEA: 7 km. northeast of Wilhelmina-top, Brass & Myer-
Drees 9866A (TYPE), September 1938, alt. 3560 m., few plants in alpine bog (petals
long and narrow, underside red

The fruit of this species is immature. The petioles and basal sheaths are
densely appressed-hirsute, but the leaf-blades are almost glabrous. Another
distinctive character is in the long narrow petals. The species obviously is
of the same group to which Ranunculus amerophyllus F. v. Muell. belongs,
but closer than that we cannot place it.

— amerophyllus F. v. Muell. Trans. Roy. Soc. Vict. ar 1. 1889; Kew
. 1899: 96. 1899; Ridl. Trans. Linn. Soc. Bot, II. 9: 10.

ecceeine New Guinea: Lake Habbema, Brass 9245, aes nn alt. 3225 m.,
forming small carpets on wet grassy shores of lake (lower side of petals purple-brown) ;
7 = northeast of Lippe top, Brass & Myer-Drees ¢792, September 1938, alt. 3560
m., in wet grassy valley; 11 km. northeast of Wilhelmina-top, Brass & Myer-Drees 9709
jomeside also 9740, same locality), September 1938, alt. 3400 m., in grass on moist
western slope (corolla yellow within, the outer upper part dark), British NEw
GutneEA: Mount Albert Edward, auvatinwest slope, Brass 4266, June 1933, alt. 3680 m.,
sporadic on wet grasslands, common (leaves erect; petals bright yellow inside, reddish
brown outside) ; Murray Pass, Wha rton Range, Brass 4647, July 1933, alt. 2840 m
abundant on grasslands (petals reddish brown beneat

The collections from British New Guinea were identified by Mr. C. T.
White, who noted in his list that these were ‘‘a good match for the scrappy
type-material received on loan from the National Herbarium, Melbourne.”
The leaves are 1-1.5 cm. long and 0.3 cm. broad, with petioles 2-3 cm.
long. The collections cited from Netherlands New Guinea are plants
6-15 cm. tall, with leaves 1.5—2 cm. long and 0.6—0.8 cm. broad, and with
petioles 3-9 cm. long. The flowers appear to agree well with those of the
British New Guinea plants, and, although we have no mature achenes,
we believe these all belong to the same species, with the possible exception
of Brass & Mvyer-Drees 9740, a plant with leaves coarsely 3-dentate and
in general more pubescent.

Ranunculus habbemensis sp. nov.

Caudex brevis; foliis omnibus basalibus petiolatis; petiolo 1.5—-4 cm.
longo strigoso; lamina chartacea supra consperse subtus -+ dense strigosa,
variabili, obovato-cuneata vel oblongo-cuneata, 2—3 cm. longa, 0.8—1.8 ¢
lata, apice grosse dentata (dente medio 3-5 mm. * 3-7 mm., dentibus
lateralibus + 2 * 2- ae vel interdum 5-dentata, basi late breviter
cuneata, margine integra; scapis 1-3 aphyllis unifloris 8-18 cm. longis
parce strigosis sepalis ovato-oblongis, 6.5 cm. longis, 2—2.5 cm. latis, extus
strigillosis; petalis 12-14 mm. longis, 5-6 mm. latis, basi squamula
oblonga nectarium tegente praeditis; staminibus + 35, circiter 4 mm.
longis, filamentis complanatis; achaeniis numerosis 2.5 mm. longis, oblique
aes compressis, rostro 1 mm. longo apice leviter recurvato

LANDS NEw GuINEA: Lake Habbema, Brass 9589 (TYPE), heicoss 1938,
alt. ee m., plentiful on alpine grassland.

Possibly this species is close to Ranunculus Muelleri Benth., but the
pubescence is of stiffish hairs, the sepals are half as long as the petals, and
the achenes are recurved at the stigmatic tip of the beak.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 39

Ranunculus perindutus sp. nov.

Planta acaulis; rhizomate descendente; foliis omnibus basalibus charta-
ceis saepe dense adpresse vel subpatenter villosulis: petiolo 2—5 cm. longo;
lamina oblonga vel elliptica, basi obtusa, grosse ee dentata, dente medio
5-6 mm. longo et 3-8 mm. lato , dentibus ceteris paullo minoribus: scapis
unifloris aphyllis 3—7 cm. eli in fructu 10-14 cm. longis, adpresse hirsutis;
sepalis patenti-adscendentibus late ovatis, 6 mm. longis, 3 mm. latis, mar-
gine membranaceis obtusiusculis, Ree hirsutis, intus 3-nerviis; petalis
oblongis, 8-9 mm. longis, 2.5—-3 mm. latis, apice obtusis, basi leviter angus-

3—4 mm. longis; achaeniis numerosis immaturis oblique ovatis, stylo sensim
recurvato.

NETHERLANDS New Gurnea: 11 km. northeast of Wilhelmina-top, Brass & Myer-
Drees 9727, September 1938, alt. 3400 m., rather dry grassy valley (corolla yellowish
within, the outer part reddish): 7 km. morbeas of Wilhelmina-top, Brass & Mvyer-
Drees 10027 (TYPE), September 1938, alt. 3560 m., abundant on grassy slopes (petals
brown beneath) ; 2 km. east of Wilhelmina- top, Peas & Myer-Drees 10132, September
1938, alt. 3800 m., common on grassy banks of stream

This species differs from Ranunculus babes in the finer and more
profuse indument and somewhat in the leaves being more evenly incised-
dentate along the margin, rather than having the dentations confined to
the terminal part of the margin.

Brass 4354, Mount Albert Edward, alt. 3680 m., gregarious on alpine
slopes (peduncles purple; flowers small, pale yellow striped underneath with
purple-brown), is a plant with indument similar to the type, but with stout
petioles, small flowers and achenes. We have been uncertain where to place
the collection.

Ranunculus perindutus var. papuanus var. nov.
A forma typica recedit petiolo longiore, lamina suborbiculari trilobata
lobo terminali 6-12 mm. a o 5-7 mm. lato, lobis lateralibus inciso-
lobatis vel dentatis) palmatim 3-—5-nervata, supra fere glabra (consper-
sissime pilosa), subtus ze hirsuta, petiolo patenti-hirsuto, floribus et
fructibus ut in typo

British New Guinea: Mount Albert Edward, Brass 4353 (type of var., New York
Bot. Gard.), June 1936, alt. 3680 m., common along banks of small alpine stream
(peduncles purple; petals reddish brown beneath).

Ranunculus Lowii Stapf in Hook, Ic. Pl. 23: pl. 2261. 1893; Trans. Linn. Soc. Bot. II.
27. 1894; van Steenis, Bull. Jard. Bot. Buitenz. III. 13: 239. 1934, vel. aff.

NETHERLANDS NEW GurNeEa: 5 miles northeast of Wilhelmina-top, Brass 9395,
August 1938, alt. 3400 m., gregarious along a grassland stream (small yellow flowers) ;
7 km. northeast of Wilhelmina-top, Brass & Myer-Drees 10030, September 1938, alt.
3560 m., common on wet grassy slopes; Bele River, 18 km. northeast of Lake Habbema,
Brass 11422, alt. 2200 m., common on mossy rocks in river and on forest paths.

This material differs from the original description and the one Bornean
specimen examined in having the leaves strigulose-hirsute on both surfaces
and in its somewhat smaller flowers. The cited specimens are also close to
Ranunculus perindutus, described above.

Ranunculus coacervatus sp.
Planta acaulis; ian crasso brevi; foliis omnibus basalibus subcoria-

40 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

ceis; petiolo + 3 cm. longo patenti-villosulo, pilis luteis; lamina 1—1.3 cm.
longa et 1.3-1.8 cm. lata, orbiculari-reniformi, 3—5-lobata, lobis grosse 2-
vel 3-dentatis, utrinque + villosula, margine copiose villosula, venis_ pri-
mariis palmatis: - scapis unifloris 3-5 cm. longis patenti- -villosulis: sepalis
late ovatis leviter concavis, 3-3.5 mm. longis et 2 mm. latis, fere glabris,
pilis paucis conspersis ; petalis oblongis, 6-7 mm. longis et 3 mm. latis,
obtusis, basi angustatis squamula nectarium tegente praeditis; staminibus
+ 20 circiter 3 mm, longis; achaeniis numerosis oblique obovatis com-
pressis, rostro brevi valde recurv

NETHERLANDS NEW GUINEA: 11 cs northeast of Wilhelmina-top, Bra
Drees 9727A (TYPE), September 1938, alt. 3400 m., rather dry grassy nice ant
yellow within, the upper outer part reddish).

Amongst the New Guinean material, this species is perhaps closest to
Ranunculus perindutus, but the leaves are broader than long and differently
lobed, the flowers are smaller, the pubescence is distinctly yellow and widely
spreading, and the achenes are broader toward the apex.

Ranunculus a var. oe (Hook.) Benth. Fl, Austr. 1: 12. 1863;
dway Tas . Fi.

Ranunculus ashen ig uk Bot. 1: 244. 1834; Hook. f. Fl. Tasman. I: 7.

1860.

NetuertAnps New Guinea: Lake Habbema, Brass 9203, August 1938, alt, 3225 m.,
in a mossy bog (small rosette herb; petals brown beneath). British NEw GUINEA:
Mount Albert Edward, Brass 4267, June 1933, alt. 3680 m., rather rare grassland herb
in wet soil (leaves flat-spreading ; petals yellow inside, brown outside) ; same locality,
Brass 4355 (peduncles purplish; flowers deep yellow, shaded purple-brown beneath).

We have named these collections with some hesitancy. They differ from
the original description in not —— a reflexed calyx, and the petals are
longer than the sepals. As far we may judge from the diagnosis of
Ridley’s subvar. papuanus, these collections could not possibly belong to it.

Ranunculus lappaceus var. oe Hook. f. Handb. N. Z. Fl. 7. 1864; Cheese-
n, Man

ma .N. Z. Fl. 4
ogee multiscapus ae i. FI. N. Z. 1: 9, t. 5. 1852.
NETHERLANDS NEW GUIN 9 km. northeast of Lake Habbema, Brass 10546,

October 1938, alt. 2800 m., open bouldery bed of a forest-stream (petals brown be-
neath).

This variety or its neon nas ee previously reported from British New
Guinea, Kew Bull. 1899: 899. This variety differs from the last,
according to the specimens a ng in that the former has leaves definitely
3-foliolate, somewhat thicker in texture, and almost sericeous-villous.
Ranunculus uncostigma sp. nov

Planta acaulis stolonifera; Seite omnibus basalibus chartaceis, Abies
dense maturis parce adpresso-pilosis, longe petiolatis; petiolo 8-15 c
longo; lamina tripartita vel trifoliolata; foliolis petiolulatis; foliolo ae
rotundato-cuneato grosse 3- dentato, foliolis lateralibus interdum bilobatis
ac grosse dentatis; scapis + 16 cm. longis parce et adpresse pilosis; ala-
bastris sub anthesi: sepalis ovato-ellipticis 4 mm. longis obtusis concavis,
margine late membranaceis, fere glabris; petalis 5 mm. longis, 4 mm. latis,
rotundato-ellipticis, basi obtuse angustatis squamula nectarium tegente
praeditis; staminibus + 15 circiter 2 mm. longis; achaeniis numerosis,

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 41

2.5 mm. longis, vix 2 mm. latis, oblique obovatis compressis, rostro brevis-
simo retrorse uncinato.

NETHERLANDS NEW GuINEA: 9 km. northeast . Lake Habbema, Brass 10748
(TYPE), October 1938, on a native clearing in the fore

This species differs from Ranunculus ne var. multiscapus Hook. f.
in the obovate achenes with very short hooked beak. In this character it is
more like R. diffusus DC., but the plants are all scapose.

BEGONIACEAE

Forty-three species of Begonia have been described from Papuasia. Nine
of these are illustrated by partly diagrammatic sketches for the purpose of
comparing modes of inflorescence, and five are represented by habit
sketches. Thus, under present conditions, we find ourselves faced with the
necessity of identifying most of our material from the original descriptions
alone. To aid others in the interpretation of this difficult group, our species-
descriptions are supplemented by text-figures. Some described from stami-
nate material alone are assigned temporarily to the Section Petermannia.
The first two species defined do not seem, in their combination of characters,
to fall entirely within the limits of any Asiatic section given in the latest
work on this family in the Pflanzenfamilien, but are perhaps nearest
Sphenanthera and Platycentrum. Only one species of Symbegonia Warb.
is represented in the material at hand.

Begonia Linnaeus
Begonia physandra sp. nov. Fig. 1, a-e.

Planta acaulis, foliis scapisque floriferis instructa; foliis membranaceis
orbiculari-ovatis, 12-18 cm. longis et 9-16 cm. latis, apice acutis vel breviter
acuminatis, basi vix aequalibus oblique cordatis, margine duplicato-serrato-
denticulatis breviter setulosis interdum sinuato-sublobatis, supra con-
spersissime setulosis (setulis 1 mm. longis), ae prope basim sparsim
hirsutis, nervis basalibus 6 vel 7 bi-trifurcatis, lateralibus utrinsecus 3;
petiolo 10-12 cm. longo fere glabro; stipulis ochraceis: scapo usque 30 cm.

inflorescentiis flores ¢ atque @ gerentibus; bracteis late sa
8-9 mm. longis membranaceis; floribus ¢ : pedicello ae Cm,
tepalis 4, exterioribus ellipticis 1.5 cm. longis et 1.1 cm. latis, oat
Lio Ci; longis et 1.1 cm. latis obovatis retusis; staminibus 60 fasciculatim
in columnam 1.5 mm. connatis; filamentis 1 mm. lo ngis subaequalibus,
connectivo 1 mm. longo apice evidenter inflato, thecis longitudinaliter de-
hiscentibus inter filamentum et connectivum inflatum positis; floribus @
ante anthesim: pedicello + 1 cm. longo; tepalis 4 ut in floribus ¢ ; stylis
3 in columnam brevem connatis deinde in crura spiraliter papillosa par-
titis; capsula verisimiliter matura 1 cm. longa et lata, alis 3 aequalibus
rotundatis 2 mm. medio latis.

British New Gurnea: Mafulu, Brass 5199 (TYPE in NYBG), October 1933, alt.
1250 m., crevices in limestone rock faces, common (plant very fleshy; leaf-stalk and
peduncle reddish pink, lamina pale green; flowers delicate pale pink)

In floral characters this species and Begonia Archboldiana very closely
resemble each other, but they seem to be distinct in vegetative characters;

42 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

the former has not variegated leaves, the hairs on the upper surface of the
leaf are shorter, stouter and much more scattered than in the latter, and
the bracts of the inflorescence are almost three times as broad and probably
white rather than ochraceous.

Begonia Archboldiana sp, nov. Fig. 1, f-k.

Planta acaulis, foliis scapisque floriferis instructa; foliis membranaceis
variegatis orbiculari- ovatis, usque 12 cm. longis et 9. 5. cm. latis, apice acutis
vel breviter acuminatis, basi subinaequalibus leviter oblique cordatis, mar-
gine irregulariter subsinuatis vel interdum sinuato-sublobatis (lobis 5 parvis
duplicato-serrato-dentatis setulosis), supra consperse setulosis (setulis 2—3
mm. longis interdum crispulis), subtus glabris vel costa nervisque prope
basim setulosis; nervis basalibus circiter 6 bifurcatis, lateralibus utrin-
secus 3; petiolo usque 18 cm. longo, glabro vel sparsissime hirtello; Rae:
ochraceis lanceolatis 1 cm. longis apice longiuscule acuminatis ; sca 2

lanceolatis 6 mm. longis acuminatis; floribus ¢ : pedicello 1—2.5 cm. longo;
tepalis 4 glabris, exterioribus ellipticis 1-1.5 cm. longis et 8 mm. latis,
interioribus 1—1.2 cm. longis et 6—8 mm. latis obovatis retusis; staminibus
50 fasciculatim in columnam 1-2 mm. connatis; filamentis | mm. longis

a. habit, x %4; b. flower, x Y; r of stamens, eran d. single ao
enlarged; e. capsule, X 1%. f-k. Eun Arehboldiana Merr. & Perry, drawn from the
type in NYBG: f. habit, x %4; g. & flower, x ™%; A. cluster of i aaa enlarged ;

z. single stamen, enlarged; j7. 2 flower, x 1%; k. bea x WY. lLg@q. Begonia acaulis
Merr. & Perry, drawn from the type: J. habit, x 14; m. 8 flower, x %; n. cluster of
stamens, enlarged; 0. single stamen, enlarged; ». 9 flower, x 1%; qg. capsule, x ™%.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 43

subaequalibus, connectivo 1 mm. longo apice evidenter inflato, thecis longi-
tudinaliter dehiscentibus inter filamentum et connectivum inflatum positis;
floribus @: pedicello + 1 cm. longo; tepalis 4 ut in floribus ¢ ; stylis 3
circiter 3 mm. in toto longis, basi in columnam 1 mm. connatis subinde in
crura 2 mm. longa spiraliter papillosa partitis; ovario 4.5—-6 mm. longo et
4—5 mm. lato, alis 3 circiter 0.5 mm. latis; capsula immatura 7 mm. longa
et lata subrotundata, alis inconspicuis.

British New GurneEa: Bella Vista, Brass 5470 (type in NYBG), November 1933,
alt. 1450 m., wet banks of stream in forest (growing in small clump; leaves variegated
green and brown; petiole and peduncle red; flowers white; seen only in one locality).

This species and Begonia physandra differ from the others we have ex-
amined in that the stamens are at the apex of a short column and the con-
nective of the anthers is inflated at the apex; also the styles seem to be
united almost to the point where the apex is divided into two parts. In
the other species examined, although the three styles may be shortly
connate at the base, they soon separate so that they appear free a short
distance before the appearance of the stigmatic surface, which in most
cases forms a band around the apical branches, these being spirally twisted.
Begonia acaulis sp. nov. § Diploclinium. Fig. 1, l-q.

Planta acaulis, foliis scapisque floriferis instructa; foliis membranaceis
oblique orbiculari-ovatis, apice obtusiusculis, basi inaequalibus oblique
cordatis, margine crenato- sinuatis et duplicato- dentatis ciliatis, supra con-
sperse crispule albido-pilosulis, subtus costa nervisque consperse pilosulis,
nervis basalibus 4-6 bi-tri-furcatis, lateralibus utrinsecus 3; petiolo 5—10
cm. longo consperse piloso; stipulis ochraceis; scapo usque 20 cm. longo
in sicco complanato sparsim piloso; inflorescentiis flores atque @
gerentibus; bracteis ovato-oblongis 4 mm. longis, margine ciliatis; floribus
g : pedicello 1.5 cm. longo; tepalis 4 extus sparsim pilosulis, exterioribus
late ellipticis 1.1 cm. longis, interioribus 9 mm. longis obovatis emarginatis;

. longi

rotundato-cuneatis; floribus @: pedicello 1.5—2 cm. longo; tepalis ut in
floribus ¢ ; stylis 3 vix 3 mm. longis basi connatis, apice in crura 2 circiter
1.5 mm. longa spiraliter papillosa fissis; ovario suborbiculari 0.5 mm. dia-
metro consperse piloso alato; capsula 1 cm. longa et 2 cm. lata, alis 3 valde
inaequalibus quarum 2 minoribus, basi atque apice ultra loculos 2 mm.
productis; alis minoribus aequalibus, obtuse subtriangularibus, margine
superiore 5-6 mm. longis, medio 3 mm. latis, una majore elongata, margine
superiore 1.8 cm. longa, medio 1.2 cm. lata.

a New Gutea: Rona, Laloki River, Brass 3599 (type), March 1933, alt.
450 co on rock plant in light Pan tereaa (whole plant very fleshy; stem,
lanes a ata and lower part of main leaf-nerves reddish; petals pink; fruit
pinkish wh

This species undoubtedly belongs in the same group with Begonia
Sharpeana F. v. Muell., but the latter is a much more pubescent species,
the fruit of which is characterized by a large wing about twice as broad
in proportion to its length as in our species.

Begonia Brassii sp. nov. § Diploclinium. Fig. 2, a-f.
a erecta gracilis; cauli simplice vel apicem versus pauci-ramoso et
crispule hirtello deorsum glabro, nodis imis radicante; foliis tenuiter char-

44 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

taceis fere membranaceis lanceolato-ellipticis, 3.5—-9 cm. longis et 1—3.5 cm.
latis, apice acutiusculis, basi valde obliquis, latere exteriore in lobum
brevem (5 mm. longum) productis nervis basalibus plerumque 3, latere
nue cuneatis vel subobtusis nervis basalibus 1 vel 2, lateralibus 4
ascendentibus, margine irregulariter remotiuscule subduplicato-dentatis
aide we supra glabris, subtus costa nervisque hirtellis; petiolo 1—3
cm. longo hirtello; stipulis ovatis 5 mm. longis
ochraceis; inflorescentiis terminalibus vel in axillibus superioribus 3—5 cm.
longis racemosis flores ¢ atque 2 gerentibus; bracteis albis late ovatis
+ ongis semiamplexicaulibus; floribus ¢ : pedicello circiter 8 mm.
longo glabro; tepalis 4 oblongis exterioribus 8 mm. interioribus 4 mm.
longis; staminibus 4, filamentis 2 mm. longis liberis, antheris 0.5 mm
longis ovatis vel subrotundatis: floribus 9: immaturis; pedicello 5 mm.
longo; tepalis 3 vel 4 ellipticis, 8-9 mm. longis, 4.5 mm. latis, apice obtusis;
stylis 3 erectis liberis quove 2 mm. supra basim partito, brachiis haud tortis
stigmaticis subchartaceis complanatis 2 mm. longis, margine irregulariter
erosis obscure papillosis; ovario 2-3 mm. longo glabro, alis inaequalibus;
capsula cernua 6 mm. longa, alis 2 aequalibus + rotundatis 3 mm. medio
latis, una 1 cm. longa et 4 mm. lata chartacea viridescente.

poner New Guinea: Bele River, 18 km. northeast of Lake Habbema, Brass
11228 (TYP November 1938, alt. 2300 m., forest undergrowth, plentiful in moist
gullies ments white).

The flowers, both staminate and pistillate, of Begonia Brassii and B.
oligandra closely resemble each other in the very few stamens and the
unequal wings of the capsule; two wings are alike and more or less follow
the contour of the ovary, while the third is less developed towards the base
of the ovary, but in the dried flower apparently lies projecting upward
against the tepals, the margin being inconspicuously 2- or 3-dentate some-
what like the margin of a minute leaf. In fruit this wing stands straight
out between the other two and is greenish, whereas the two smaller ones
tend to be straw-colored in the dried specimens. The style-branches in
all the flowers examined are flattened and not in any case twisted, as is
usual in the other species of the genus.

Begonia oligandra sp. nov. § Diploclinium. Fig. 2, g-k.

Herba erecta gracilis ramosa; radice subcrassa; caule rufescente glabro

nodis imis radicante; foliis breviter vel longiuscule petiolatis, petiolo
l

usque 4.5(—8) cm. longa et 2.5(—5) cm. lata, palmatim 4—6-partita, lobis
pinnati-partitis vel bipinnatifidis, lobis ultimis integris vel apicem versus
1- vel 2-inciso-serratis; stipulis ochraceis ovatis + 4 mm. longis caducis;
inflorescentiis racemosis terminalibus et in axillis superioribus, 2—4 cm.
longis, flores ¢ atque @? gerentibus; bracteis late ovatis vel fere orbiculari-
bus circiter 1 cm. longis membranaceis albis; floribus ¢ : pedicello eh
7 mm. longo glabro; tepalis 4 glabris, exterioribus 2 late ovatis 4 m
longis latisque obtusiusculis, interioribus lanceolato-oblongis 3 mm. ies
et 1.5 mm. latis; staminibus 6-8 basi 0.5 mm. connatis, filamentis 2.5 mm.
longis, antheris suborbicularibus 0.5 mm. diametro; floribus 9: bracteis
ongis; pedicellis 5 mm. longis; tepalis 4, 2 majoribus ovatis obtusis
7 mm. longis, 2 minoribus lanceolatis 3-5 mm. longis; stylis 3 ad basim

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 45

liberis, quove 2 mm. supra basim partito, brachiis circiter 2 mm. longis
erectis subtortis papillosis; ovario mm. longo; alis inaequalibus, 2
minoribus angustis 0.5 mm. latis rotundatis, una majore oblique desinente,
pide eos 4 mm. longa inferiore minute 2- vel 3-denticulata.

NETHERLANDS NEw GuINEA: 15 km. southwest of Bernhard Camp, Idenburg
River, ae pps (tyPE), January 1939, alt. 1600 m., on mossy rocks in a rain-forest
stream (flowers white); Valley of Bele River = 20 km. north of Lake Habbema,
Brass 10829, October 1938, alt. 2200 m., plentiful in forest shade on low banks of
stream Gonen white).

Fic. 2. a-f. Begonia oer Merr. & Perry, drawn from the type: a. leaf,

x %; b. & flower, X ™%; c. cluster of stamens, enlarged; d. @ flower, x 1%; e. one
style, enlarged; f. capsule, . 6 —k. Begonia oligandra Merr. & ry, drawn
r e e: g. leaf, x wer, X 1%; 7. cluster of stamens, enlarged;
j. 2 flower, x 1%; k. one style enlarged, l. Begonia otophora Merr. & Perry, drawn
from the ae habit, x %. n. Begonia diffusiflora Merr. & Perry: m. habit, drawn
from the type, x 4; n. ani oe from Brass 6713, X VY.

Among the described Papuasian species, this suggests B. Warburgi
Lauterb. & K. Schum. in the palmate-pinnatifid or bipinnatifid leaves; this
character was so distinctive that Lauterbach & Schumann give only a brief
description of the fruit: capsule broadly winged, red, 12-13 mm. long.

owever, if the wings had been conspicuously unlike in size they would
surely have mentioned it. Unfortunately none of the plants belonging to
the above cited numbers show mature or even partly mature fruits. The
ovary has two wings more or less conforming to its contour and a third
oblique and somewhat elongate wing projecting or directed upward against
the larger tepals in the young flowers, its outer margin is uneven or

46 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

minutely denticulate; the whole is very much like that found in B. Brassii.

e immature flowers do not show papillose styles and sometimes they
appear irregular in form, but at maturity they are similar to those of other
species of Begonia.

Begonia otophora sp. nov. § Petermannia. Fig. 2,1.

Planta 50 cm. alta; ramulis crispe ferrugineo-hirtellis; foliis usque 8 cm.
longis et 3.5 cm. latis, lanceolato-ellipticis, apice acutis, basi latere ex-
teriore angustatis deinde in lobum orbicularem 1—1.5 cm. productis nervis
basalibus 4 vel 5, interiore rotundatis vel obtusis nervis basalibus 1 vel 2,
lateralibus utrinsecus 4 vel 5, margine duplicato- serrato-dentatis vel
irregulariter breviter incisis, setulosis, supra conspersissime setulosis, subtus
costa nervisque crispe hirtellis: petio olo + 5 mm. longo hirtello; 'stipulis
oblongo-lanceolatis 1 cm. longis, apice setosis, ochraceis, glabris: inflore-
scentiis ¢ tantum visis, terminalibus 3 cm. longis immaturis pauciramosis
paucifloris; bracteis ut stipulis: pedicellis 0.8—1.5 cm. longis crispe hirtellis;
tepalis 2 + hirtellis ovatis obtusis 1.1 cm. longis; staminibus paucis circiter
8 basi connatis, antheris oblongis.

NETHERLANDS NEw Guinea: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13218 eau March 1939, alt. 850 m., rare in rain-forest flood-plain (ascending
herb 50 cm. high; upper surface ‘of leaves dark brownish green, the lower surface red;
flowers white).

Begonia diffusiflora sp. nov. § Petermannia, Fig. 2, m,n.

Planta 50 cm. alta; ciate ferrugineo-hirsutis, pilis + 4 mm. lon gis ;
foliis + 16 cm. longis et . latis, ellipticis, utrinque angustatis, apice
acuminatis, acumine + 2 cm. e ngo, ‘basi in naequalibus, in latere exteriore
in lobum brevem 5 mm. longum rotundatum productis nervis basalibus 3,
in nteriore cuneatis vel rotundato-cuneatis nervo basali uno, lateralibus

ores 6 aut @ gerentibus: inflorescentis 4 dichotomo-ramosis 8 cm.
longis et 12 cm. latis + 30-floris, ramulis sparsim hirtellis vel sursum fere
glabris; floribus pedicellatis ; eee 2 suborbicularibus 5—7 mm. diametro,
extus basim versus sparsim pilosis; staminibus 25—30 basi breviter connatis,
filamentis antheris subaequilongis, antheris 1 mm. longis oblongis emargi-
natis, rimis 2 antherae aequilongis; inflorescentiis @ bifloris; pedicellis
3 cm. longis sparsim pilosis; tepalis verisimiliter 5 (3 tantum ‘visis) late
ovatis 1 cm. longis; stylis 3 basi breviter connatis apice in crura 2 brevia
patenti-erecta spiraliter papillosa ea ovario 1.5 cm. longo ellipsoideo;
alis 3 inaequalibus ultra ovarium 3 mm. productis, una majore rotundata
7 mm. medio lata, 2 minoribus ee 5 mm. medio latis.
British New Guinea: Fly River, 528 mile Camp, Brass 6713, 7010, May 1936, alt.
80 m., heavily shaded gully in ridge-forest undergrowth, uncommon (30 cm. high;
flowers pink) ; Palmer _ 2 miles below Black River Junction, Brass 7318 rene
July 1936, alt. 100 m., a rare species, epiphytic or terrestrial in ridge-forests (fleshy
herb 50 cm. high; inition of Rin plants “reddish brown beneath; flowers pink).
In several characters this plant seems to agree with Begonia filibracteosa
Irmsch., but the staminate inflorescence is distinctly dichotomously
branched, the flowers have fewer stamens with longer lines of dehiscence,

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 47

and the @ flowers are practically glabrous, although when younger they
may have had a few scattered hairs.
Begonia calliantha sp. nov. § Petermannia. Fig. 3, a, b.

Planta suffruticosa circiter 2 m. alta; ramulis + dense et crispule rufo-
hirsutis; foliis 11-12.5 cm. longis et 4-5 cm. latis, oblongo-ellipticis vel
ovato- ellipticis, apice acuminatis, acumine + 1 cm. longo, basi inaequalibus,
latere exteriore in lobum brevem (6 mm. longum) rotundatum productis,
nervis basalibus 3 vel 4, inter’ iore late cuneatis vel leviter rotundato-
cuneatis nervis basalibus 2, lateralibus utrinsecus 4 valde ascendentibus,
margine indistincte duplicato-serrato-dentatis setulosis, supra glabris vel
consperse minuteque glanduloso- pubescentibus, subtus dense minute pustu-
latis, costa venisque crispe + dense pilosis; petiolo 0.5—1 cm. longo dense
crispe hirsuto; stipulis 2—2.5 cm. longis, 4—5 mm. latis, lineari-oblongis
apice in setam rs 3 mm. longam sensim angustatis; inflorescentiis terminali-
bus circiter 7 cm. longis racemosis; floribus ¢ tantum visis; pedicellis
2 cm. longis crispe hirtellis, basi bracteis glabris 2.5—3 cm. fae et 1-1.2
cm. latis lanceolato- ellipticis apice setulosis; tepalis 2 glabris 3 cm. longis
et 2.5 cm. latis, ovatis, basi retusis, apice 0 obtusis; staminibus 50-60, a
mentis basi 2 mm. connatis parte superiore liberis, exterioribus 0.5 m
interioribus 1-1.5 mm. longis, antheris 2.5-3 mm. longis oblongis em
rimis 24 antherae aequilongis.

British New Guinea: Mount Tafa, Mavi, Brass 4986 (TYPE in NYBG), September
1933, alt. 2225 m., bed of a small stream in forest (large fleshy shrub about 2 m. high;
branches, petioles and peduncles with reddish brown hairs; flowering bracts suffused
with red; petals carmine)

Begonia Randiana sp. nov. § Petermannia. Fig. 3, c,d

Planta suffruticosa usque 1 m. alta, ramosa, ex toto ferrugineo-hirsuta ;
foliis oblongo-lanceolatis, 5-10 cm. longis et 1.5—3 cm. latis, apice acutis,
basi latere exteriore in lobum brevem (5 mm. longum) cordatum productis
nervis basalibus 2 vel 3, interiore rotundatis nervo basali uno, lateralibus

lanceolatis 8 mm. longis, apice abrupte caudatis; inflorescentiis terminali-
bus flores ¢ atque @ gerentibus + 5 cm. lon ngis secetar floribus ¢:
pedicello 1-2 cm. longo; tepalis 2 ovatis late obtusis 1.5—2 cm. longis;
staminibus 20-25 basi 2 mm. connatis, plamentie ie 1.5 mm. longis,
antheris 1.6 mm. longis oblongis leviter obovatis obtusis, rimis 4 antherae
aequalibus; floribus @: tepalis 5 oblongo-lanceolatis, 1.8 cm. longis et
6 mm. latis, obtusis; stylis 5 mm. longis 3, basi breviter connatis sursum
liberis, apicem versus in crura 2 brevia (2 mm. longa) ascendentia spiraliter
papillosa fissis; ovario late ellipsoideo 8 mm. longo, alis 3 subaequalibus,
margine depen horizontaliter truncatis, basi rotundatis, 4 mm. medio
latis; capsula cernua 1.3 cm. longa, 1.5 cm

British NEw ee NEA: East Mount Tafa, Brass 4136, May 1933, alt. 2100 m., in a
moist forest gully (compact small fleshy shrub + 50 cm. high; branches purple- red and
whole plant covered with long red hairs; petals pinkish white); Mount Tafa, Brass

flowers white) ; Mafulu, Brass 5508, November 1933, alt. 1100 m., floor plant in lower
level forests, rare (low spreading shrub 50-70 cm. high; indumentum red; flowers pale
pink).

48 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

The collection designated as the type differs from the other two chiefly
in its leaf-margins being less deeply incised. In general habit the species
seems to approach closely Begonia fruticella Ridl., but it is quickly dis-
tinguished from that species by having wings similar in size and shape.
Dedicated to Dr. A. L. Rand, assistant leader of the Expedition.

a, 6. Begonia calliantha Merr. & Perry, drawn from the type in NYBG:
a. leaf, showing lower surface, X 1%; b. portion of 2 inflorescence, xX . c,d.
Begonia Randiana Merr. & Perry, drawn from the type: c. habit, showing ¢ inflo-
rescence, X 12; d. Q flower, X “%. e, f. Begonia Richardsoniana Merr. & Perry, drawn
from the type: e. habit, x ™%; f. 2 flower, x 4.

Begonia Richardsoniana sp. nov. § Petermannia. Fig. 3, e, {.
Planta erecta 60-70 cm. alta; ramulis dense ferrugineo-hirsutis; foliis
9-15 cm. longis et 4.5—6 cm. latis oblongo-ellipticis, apice subabrupte

rotundatum cordatum productis nervis basalibus 3 vel 4, intericre cuneatis
vel rotundato-cuneatis nervis basalibus 1 vel 2, lateralibus circiter 5,
utrinque consperse setulosis subtus costa nervisque hirtellis; margine minute
duplicato-serrato-dentatis setulosis; petiolo 1-2 cm. longo hirsuto; stipulis
hirtellis 1 cm. longis lanceolatis, apice 5 mm. caudatis: inflorescentiis axil-
lari : + 4 . longis racemosis: bracteis circiter 6 mm. longis
stipulis similibus; pedicellis rufo-hirtellis; tepalis 2 late rotundatis, 1.5 cm.
longis, 2 cm. latis, consperse hirtellis; staminibus 50-60, filamentis + 1
mm. longis antheris subaequalibus, antheris oblongis emarginatis basi
angustatis, rimis fere 14 antherae aequilongis; inflorescentiis ? : flore unico

1943] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 49

viso consperse setuloso; pedicello 2 cm. longo hirtello; tepalis 5 lanceolato-
ellipticis 1.5 cm. longis ‘ciliato- setulosis; stylis 3 circiter 5 mm. ongis, basi
connatis sursum liberis, apice in crura 2 brevia patentia spiraliter papillosa
fissis; ovario 1.5 cm longo et 0.6 cm. lato, oblongo; alis 3 subaequalibus
subtriangularibus, margine superiore leviter oblique et horizontaliter
truncatis vix 1 cm. longis, basi cuneatis, 3 mm. medio latis, margine
setulosis
ETHERLANDS New GUINEA: 4 km. southwest of Bernhard Camp, Idenburg River,

Brass 13423 (type), March 1939, alt. 850 m., occasional on shady banks of rain-forest
streams (erect to 60-70 cm.; all parts including petals sprinkled with red hairs; @
flowers pink; ¢ pink to white)

Dedicated to Mr. W. B. Richardson, mammologist of the Expedition.
Begonia oxyura sp. nov. § Petermannia. Fig. 4,a,b

Planta ascendens ramosa; ramulis crispe eu pee -hirsutis; foliis char-

taceis lanceolato-ellipticis, 9-16 cm. longis et 3-7 cm eae apice abrupte
aoe acumine 2—3 cm. os basi 7 mm. apicem versus 1
mm. lato, basi valde inaequalibus, uno latere nic: auriculatis, lobo

1.5-2.5 cm. longo, nervis basalibus 4 vel 5, altero rotundatis nervis basali-
bus 2, lateralibus utrinsecus 2 vel 3, margine sinuatis remote dentatis veri-
similiter setulosis utrinque crispe subadpresse hirtellis, subtus costa venisque
dense hirsutis; petiolo + 1 cm. longo hirsuto; stipulis ovatis, 6 mm. longis,
4 mm. latis, hirsutis; inflorescentiis terminalibus + 6 cm. longis, cymoso-
paniculatis, paucifloris, flores ¢ atque & gerentibus pedunculo brevissimo

: tepalis 2 Gare ellipticis, 13 1. cm. longis, 6 mm. latis, hirtellis;
ein + 60 basi breviter connatis, filamentis 1 mm. vel ultra longis,
antheris 1 mm. longis obovatis emarginatis : floribus @ totis hirtellis:
tepalis 5 oblongo- lanceolatis, 1.3 cm. longis; stylis 3, basi breviter connatis

ellipsoideo 1.2 cm. ene et 0.7 cm. lato, alis 3 subaequalibus; ee 233

medio latis.

NETHERLANDS New GuINEA: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13217, 13457 (type), March 1939, alt. 850 m., on banks of a rain-forest stream
(13217: climbing to 3 m.; 13457: herb 60 cm. high; flowers white).

These collections suggest Begonia torricellensis Warb., but they differ in
having very distinctly acuminate leaves and fruits with very much narrower
wings.

Begonia montis Bismarckii Warb. Fl. Deutsch. Schutzgeb. Siidsee Nachtr. 322. 1905.

NETHERLANDS NEw GUINEA: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13397, March 1939, alt. 850 m., common in undergrowth of flood-plain rain-
forest (herb 70-80 cm. high; flowers white or pale pink).

This collection seems to be a very good match for a fragmentary speci-
men collected by Schlechter from the Bismarck Mountains in Northeastern
New Guinea.

Begonia suffrutescens sp. nov. § Petermannia. Fig. 4, c-e.

Planta suffruticosa usque 50 cm. alta ramosa; ramis ramulisque crispule
hirsutis, pilis brunneis; foliis rien 1.5-4.5 cm. longis, 0.5—2 cm. latis,
lanceolatis vel subovatis, apice acutis vel acuminatis interdum obtusiusculis,

50 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

superioribus solitariis vel binis, pedunculis vel pedicellis + crispe hirtellis
2 cm. longis prope basim bracteis minimis ochraceis instructis; tepalis 2
ovatis 1.5—-1.7 cm. longis glabris albis; staminibus 4-8, filamentis 11.5
mm. longis + connatis, antheris -- 2 mm. longis apiculatis, rimis fere
antherae aequalibus; floribus ¢ plerumque solitariis ex axillibus superiori-

Fic. 4. a, b. Begonia oxyura Merr. & Perry: a. habit, drawn from Brass 13217,
xX 2; b. Q flower, drawn from the type, x 4. c-e. Begonia suffrutescens Merr. &
Perry, drawn from the type: c. habit, x 1%; d. Q flower, xX 4; e. capsule, xX V4.

. inflorescence, showing 9 flower and immature ¢ flower at base, X 4. h. Begonia
pinnatifida Merr. & Perry, drawn from the type: habit, x 1%

bus, pedunculo circiter 1.5 cm. longo crispe hirtello; tepalis 5 lanceolato-
ovatis, 9-11 mm. longis, 3-4 mm. latis, glabris, albis; stylis 3 circiter 6 mm.
longis, in parte inferiore 2 mm. connatis, apice in crura 2.5 mm. longa erecta
in summo apice papillosa fissis; ovario ellipsoideo, 7 mm. longo, 4—4.5 mm.
lato, pilis longiusculis sparsim obsito, alis 3 margine superiore truncatis basi
rotundatis, sparsim pilosis, in fructu una ala latiore margine superiore 8
mm., ceteris 5 mm. longis; seminibus minute reticulatis.

ETHERLANDS NEW 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 12030 (tyPE), January 1939, alt. 1750 m., rain-forest, common on banks of

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 51

streams and slopes of ravines (30-50 cm. high; flowers white); 6 km. southwest of
Bernhard Camp, Idenburg River, Brass 12985, ebicsdrats eee alt. 1050 m., on banks
of a rain-forest stream Poca to 20-30 cm.; flowers w

Amongst the descriptions of en Beane. this species most nearly
approaches B. monantha Warb., but in that the stamens are many and the
anthers ovate. We have found no formal description of the fruit of B.
monantha, but Irmscher gives a sketch of it in his article on inflorescences,
Bot. Jahrb. 50. Suppl.: 573, figs. 24 a—c. 1914.
Begonia serraticauda sp. nov. § Petermannia. Fig. 4, f, g.

Planta glabra usque 1 m. alta ramosa; caule in sicco rufo-brunnescente,
nodis eet: foliis tenuiter chartaceis oblongis vel oblongo-ellipticis,
8-18 cm. longis, 2-7 cm . latis, basi cuneato-rotundatis leviter inaequalibus,

structis, apice sensim caudato-acuminatis, acumine 3—4.5 cm. longo basi
4—6 mm. lato, margine indistincte (acuminis distincte) remotiuscule serratis,
nervis primariis in latere exteriore 7—9, interiore 6—8 ascendentibus, venulis
ultimis distincte clevatis; petiolo 5-7 m mm. longo; stipulis oblongo- -lanceo-

—
7
AUT a

oO

wn,

Qa

3

jet)

=e

n

go
=

QO

i.)

Sj

<<

io)

i

n

c

72)

ay
ut

2

°

a.
ion

c

77)
Oc
Le
=3
QO
i
os)
>
oO
=
oO
a,
jee)

truncatis, basi rotundatis; capsula cernua, 1.8 cm. longa, 2—2.3 cm. lata.
NETHERLANDS NEw GUINEA: 6 km. southwest of Bernhard Camp, Idenburg River,

(ascending herb 50-60 cm. high; flowers pink, white or pale pink streaked with darker
pink).

In some characters this species suggests Begonia strictinervis Irmsch.,
but the acumen of the leaves is about twice as long as in the latter species,
the base of the leaf is not sufficiently auricled to show three basal nerves
on the larger side of the blade, the inflorescence is only very few-flowered
(in the specimens cited not more than two were seen), the stamens are
more than twice as many, and the wings do not extend beyond the apex
of the capsule.

Begonia pinnatifida sp. nov. § Petermannia. Fig. 4, h.

Planta glabra ramosa nodis inferioribus ic ns; ramis gracilibus in
sicco + striatis vel angulatis; foliis pene See oe vel fere mem-
branaceis, 4—5.5 cm. longis et 1-1.5 cm. latis, lanceolatis, parte tertia

inferiore excepta pinnatifidis, laciniis Wen vel i: vel 2- dentatis setulosis,
inferioribus circiter 7 mm. longis sursum decrescentibus, terminalibus
linearibus; petiolo 1-1.5 mm. longo; stipulis 8 mm. longis linearibus
longiuscule setulosis; inflorescentiis terminalibus; floribus ¢ non visis;

52 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

pedicello floris @ 1.3 cm. longo; tepalis 5 fere aequalibus late lanceolatis,
circiter 9 mm, longis et 3 mm. latis, apice acutis breviter setulosis; stylis
basi tantum connatis, 3 mm, lo ongis, a apicem versus latioribus, in crura 2
brevia patenti-erecta spiraliter papillosa fissis; ovario ellipsoideo 8 mm.
longo utrinque angustato alis toto circumdato, alis 3 subaequalibus, 11 mm.
longis + 5 mm. medio latis; placentis bilamellatis

British NEw aes Pairies River, 2 miles below Black River Junction, Brass
7051 (TYPE), June 1936, alt. 100 m., gregarious in considerable communities in forest
ground cover of the more elevated ridges (flowers pale pink; only one plant found in

ower

A striking species readily recognized by the rather small almost sessile
lanceolate leaves pinnatifid in the upper two-thirds. The single specimen
collected has only one @ flower, which is pale pink with translucent
strongly veined wings, which project about 1.5 mm. beyond the base and
the apex of the ovary.

Oe stilandra sp. nov. § Petermannia. Fig. 5, a—d.

Planta usque | m. alta glabra; ramulis in sicco leviter striatis supra nodos
nae foliis 8-14 cm. longis, 3—5.5 cm. latis, oblongo-lanceolatis, apice
acuminatis, acumine + 1 cm. longo, basi valde inaequalibus, latere ex-
teriore in lobum 1.5—2.5 cm. longum rotundato-cordatum productis nervis
basalibus 5, interiore rotundatis nervis basalibus 2, lateralibus utrinsecus
2 vel 3, margine duplicato-serrato-dentatis breviter setulosis: petiolo 2—4.5
cm. longo: stipulis caducis haud visis; inflorescentiis termina libus + 9cm
longis pauciramosis flores ¢ atque @ gerentibus; bracteis caducis; flori-
bus ¢: pedicello + 1 cm. lon 0; epaue ¢ ovatis vel ellipticis, 1.7 cm.

S i
antheris 3 mm. longis ee emarginatis; floribus @: pedicello 2(—5)
cm. longo; tepalis 5 lanceolato-ellipticis + 2 cm. lo ngis, apice acutis; stylis
3 circiter 7 mm. longis, basi 1.5 mm. connatis, apice in crura 2 erecta 3 mm.
longa spiraliter papillosa fissis; ovario 1.1 cm. longo ellipsoideo; alis sub-
aequalibus 3, basi rotundatis, ultra ovarium 3 mm, productis, apice hori-
zontaliter truncatis rotundatis, 3 mm. et 4 mm. medio latis.

NETHERLANDS NEw Guinea: 15 km. southwest of Bernhard Camp, Idenburg River,
rating 12301 (type), January 1939, alt. 1750 m., occasional in rain-forest gullies (up to

. high; petioles, nerves and margins of eaves: peduncles and pedicels red; flowers
fon are et with red); 18 km. southwest of Bernhard Camp, Idenburg River, Brass
12499, February 1939, alt. 2000 m., undergrowth of a gully (plant 1 m. high; flowers
pink, streaked with red).

At a glance this species is scarcely to be distinguished from Begonia
simulans Merr. & Perry, but a more careful examination shows the rather
distinct character of the staminal column, the fewer and much longer
anthers, and the less broadly winged fruit.

Begonia simulans sp. nov. § Petermannia. Fig. 5
Planta ultra 35 cm. alta glabra; aia in sicco sulcatis ferrugineis;

basalibus 4 vel 5, interiore rotundatis vel obtusis nervis basalibus 2 vel 3,
lateralibus utrinsecus 3, margine duplicato-serrato-dentatis; costa et petiolo

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 53

angulum + 90° formantibus; ana 1.5-3 cm. longo ; stipulis caducis;
inflorescentiis terminalibus . longis ramosis flores atque
gerentibus; floribus ¢ : pedicello 1— ci cm. longo; tepalis 2 suborbicularibus,
basi cordatis, + 1 cm. longis oe staminibus 70, filamentis exteriori-
S 5 mm. interioribus 1.5 mm. longis, antheris 1.2 mm. longis obo-
vatis -obtusis; floribus @: Saree usque 5 cm. longo; tepalis 5 lanceolato-
ovatis 1-1.2 cm. longis; stylis 3 circiter 6 mm. longis, basi 2 mm. connatis,

Fic. 5. a-d. Begonia stilandra Merr. & Perry, drawn from the oe Saye d):
a. habit, x 1%; b. one stamen, enlarged; c. capsule, x ™%; d. Q flower, drawn from
Brass 12499, x ea e-i. Bego nia simulans Merr. & Perry, drawn from the type: e. leaf,
x “; f. 6 flow x 14; g. single stamen, enlarged; hk. @ flower, x ™%; i. capsule,
x Ye jen. ens Sandy Merr. & a drawn from the type: j. leaf, x %;
k. & flower, x 1%; 1. single stamen, enlarged; m. Q flower, X %; n. capsule, x ™%.

ovarilo
o minoribus, basi ultra
oductis, apice horizontaliter truncatis; capsula 20 cin,
longa et lata, alis 2 minoribus 7 mm. medio latis, una majore 9 mm. medio
lata

apice in crura 2 circiter 2 mm. longa spiraliter papillosa fissis;
0.8-1.4 cm. longo ellipsoideo, alis 3, quarum 2 paull
ovarium + 3 mm. pro

iene New Guinea: Balim River, Brass 11835 (type), December 1933, alt.
1800 m., a few plants in forest shade (branches, petioles and leaf-nerves beneath red:
flowers pink).

The species is to be compared with Begonia naumoniensis Irmsch., with
which it has many characters incommon. It may be readily distinguished,
however, by the smaller leaves with distinctly dentate margin a less
cordate base: the more elongate anthers, and the larger ? flower

o4 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIV

Begonia pediophylla sp. nov. § Petermannia. Fig. 5, j-n.

Planta parva glabra; foliis lanceolato- oblongis, 6-13 cm. longis, 2—4.5
cm. latis, apice abrupte acuminatis, acumine saepe angusto 0.5-1.5 cm
longo, basi inaequalibus, latere exteriore in lobum 1-2.3 cm. longum
rotundatum productis nervis basalibus 3 vel 4, interiore rotundato-cuneatis
vel obtusis nervis basalibus 1 vel 2, lateralibus utrinsecus 3 vel 4, margine
penal serrato-denticulatis ; petio olo 1—-7.5 cm. longo; stipulis lanceolato-
oblongis + 7 mm. longis; inflorescentiis terminalibus et axillaribus pauci-
floris + ° 3 cm. longis, flores ¢ et @ gerentibus; bracteis 1 cm. longis ovato-
oblongis pedicellis 1—1.5 cm. longis; floribus ¢ : tepalis 2 late ovatis 1 cm.

5 ellipticis, 10 mm. ongis, 6 mm. latis; stylis x] circiter 6 mm. een basi
2 mm. connatis, apicem versus latioribus, apice in crura 2 brevia (2 mm
longa) patenti-erecta spiraliter papillosa fissis: ovario | cm. longo 5 mm. lato
ellipsoideo, alis 3 subinaequalibus; capsula 1 .4 cm. longa et lata, alis mar-
gine — subtruncatis rotundatis, basi rotundatis, 2 mm. atque 3 mm.
medio lat

British New Guinea: Wharton Range, Murray Pass, Brass 4577 (TYPE), July 1933,
alt. 2840 m., forest shade (fleshy small shrub; branches, petioles, leaf-margins and
nerves beneath red, also peduncles and pedicels; flowers very pale pink, marked with
darker lines).

The shape of ‘the anthers, with inflated sacs and gaping sutures, is very
much like that pictured for Begonia naumoniensis Irmsch., as is also the
indistinctly denticulate margin of the leaves. The new species differs in the
smaller leaves, the number of stamens, and the much smaller fruits with
narrower wings.

Begonia media sp. nov. § Petermannia. Fig. 6, a-c.

Planta circiter 1 m. alta; ramulis elena parce crispe hirtellis cito
foliis 12-24 cm. longis, 4.5-11.5 cm. latis, ovato- ellipticis vel

basalibus 1-3, lateralibus utrinsecus 3 vel 4, margine duplicato-serrato-den-
tatis, supra glabris, subtus costa nervisque sparsim hirtellis; petiolo 1-3 cm.
longo sparsim crispe hirtello; stipulis oblongo- lanceolatis 1.3 cm. longis

1-1.3 cm. longis, cordatis, glabris vel sparsim hirtellis; staminibus 70-80
basi breviter late connatis; filamentis 0.4-1 mm. longis, exterioribus quam
interioribus brevioribus, antheris 1.6 mm. longis oblongis emarginatis ;
floribus @ non visis; pedicello 2 cm. longo; capsula cernua novella sparsim
hirtella matura fere glabra, 1.5 cm. longa, 1.8 cm. lata; alis 3 subaequalibus
basi rotundatis, ultra ovarium 2 mm. glee margine superiore hori-
zontaliter truncatis rotundatis, medio 5—6 mm. latis
British New Guinea: Mafulu, Brass 5435 eons in NYBG), November 1933, alt.

1250 m., uo in oak forest, rare (about 1 m. high; branches, peduncles, petioles
and pers nerves beneath red; flowers pink).

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 55

Begonia tafaensis sp. nov. § Petermannia. Fig. 6, d-f.

Planta 1 m. alta glabra; ramulis in sicco striatis; foliis 11-21 cm. a
4.5-7 cm. latis, oblongo- ellipticis, apice acuminatis, acumine + 2 ¢
longo, basi valde inaequalibus, latere exteriore in lobum usque 3.5 cm.
longum rotundatum productis nervis basalibus 5, interiore obtusis nervis
basalibus 2, lateralibus utrinsecus 3 vel 4, margine minute denticulatis
subsinuatis; petiolo 2-3 cm. longo; stipulis caducis haud visis ; inflore-

bus, margine superioribus + 6 mm. longis, 4 mm. medio latis, valde venosis.
British NEw Guinea: Mount an. Brass 4017 (TyPE in NYBG), May 1933, alt.
2100 m., fairly common in tall forest seas plant 1 m. or more high; stems, petioles,
peduncles and pedicels red; flowers pink); same locality, Brass 5109, September 1933,
alt. 2400 m., damp places in valley forest, rare (erect fleshy sparsely branched shrub
about 1 m. high ; large pink flowers).

: a—c. Begonia media Merr. & Perry, drawn from the type: a. leaf, x \%;
b. g flower, x 1%; c. capsule, x %. d-f. Begonia tafaensis Merr. & Perry, drawn from
the type (except d): d. leaf, drawn from Brass 5109, X %4; e. & flower, x 4; f. cap-
sule, xX ‘’ g-t. Begonia slowoneaGe Merr. & Perry, drawn from the type: g. leaf,
xX %; h. & flower, x %; i. capsule, x %. j-m. Begonia Alla eae Merr. &
Perry, drawn from the type: j. habit, xX 4; k. 3 flower, x 1%; 1. 2 flower, x 4;

m. capsule, X 14

56 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

This species is fairly close to Begonia media Merr. & Perry, but it lacks
the pubescence, the margin of the leaves more nearly approaches an entire
condition, and the stamens are fewer. It may be near B. Vandewateri Ridl.,
but the plant is glabrous, although the leaf-veins on the lower surface are
somewhat verruculose.

Begonia salomonensis sp. nov. § Petermannia. Fig. 6

Planta 1-2 m. alta glabra; ramulis in sicco alates foliis 10-27 cm.
longis, 5.5—-11.5 cm. ie ovato-ellipticis
acutis, basi latere exteriore in lobum 2—4 cm. longum rotundato-cordatum
productis nervis basalibus 5, interiore rotundatis vel cuneatis nervis basali-
bus 2, lateralibus utrinsecus + 3, margine inconspicue serrato-denticulatis
interdum sinuatis; petiolo 1.5—-4 cm. longo; stipulis haud visis, caducis;
inflorescentia ¢ folio opposita, 8.5 cm. longa, subcymosa; floribus é cir-
citer 20; pedicello — 1.5 cm. longo; tepalis 2 subrotundatis 1.2 cm.
longis; staminibus 35 basi connatis, filamentis brevibus 0.6—0.8 m
longis, antheris een 1.6-1.8 mm. longis emarginatis, rimis incites
0.4 mm, longis; inflorescentia 9: axi 4 mm. longa; floribus @ non visis;
pedicello 2.7 cm. longo; capsula 1.5—1.7 cm. ne. 2.5 cm. lata; alis 3
inaequalibus, basi rotundatis, ultra ovarium 3 mm. productis, in apicem sat
acutum desinentibus, 2 brevioribus 7 mm. medio latis, margine superiore
1 cm. longis, una latiore 9 mm. medio lata, margine superiore 1.5 cm. longa

SoLtomMon Istanps: Ulawa: Brass 2950 (type), October 1932, swampy rain-
forest, on coral limestones (fleshy shrub 1-2 m. high; flowers white; fruits red)
Begonia brachybotrys sp. nov. § Petermannia. Fig. 6, j—m.

Herba 80 em. oe mrs glabro in sicco sulcato; foliis tenuiter chartaceis
glabris, 10-20 c _ long s, 6-16 cm. latis, oblique ovato-ellipticis, apice
abrupte breviter ete basi latere exteriore in lobum latum cordatum
productis, interiore breviore cordato-rotu tundato, costa et petiolo angulum
130-140° formantibus, pa matim 7—9-nerviis, nervis a basi 13 longitudinis
dichotomo-ramosis, margine subsinuatis + rato-denticulatis; petiolo
glabro + 15 cm. longo, folii laminae fere spain stipulis caducis
probabiliter eens similibus; inflorescentiis axillaribus flores ¢ et
gerentibus, axi 1-3 cm. longa, ramis brevissimis; bracteis novellis 1.5 cm.
longis ovatis obtusis caducis; floribus ¢ : pedicello usque 2 cm. longo;
tepalis 2 glabris rotundatis 1 cm. longis latisque; staminibus 50-60, fila-

mentis inaequalibus usque 1.5 mm. longis, antheris 0.8 mm. longis late
ellipticis retusis, rimis antherae subaequilongis; floribus @ : pedicello + 2
cm. longo; tepalis 5 ellipticis 9 mm. longis, apice rotundatis; stylis 3 basi
connatis sursum liberis 3 mm. longis, apice in crura 2 circiter 1 mm, longa
patenti-erecta spiraliter papillosa fissis; ovario 6-7 mm. egy ellipsoideo,
alis 3 subaequalibus; capsula 1.5 cm. longa, 1.6—-1.8 cm. lata, alis apice
oe truncatis obtusis, basi 3 mm. ultra loculos ae 3—4 mm.
medio la

sn New Guinea: Bernhard Camp, Idenburg River, Brass 14112 (TYPE),
April 1939, alt. 55 m., bank of a small stream in rain-forest (herb 80 cm. high; flowers

nk). Britis New Guinea: Mafulu, Brass 5429, November 1933, alt. 1250 m., oak
forest ground cover (about 50 cm. high; stems, petioles and cedunwies red; leaves dark
green; small pink flowers and red fruits).

This species should be compared with Begonia strictiformis Irmsch. from

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 97

the North Celebes; the Papuasian material is glabrous with axillary in-
florescences, fewer stamens, and leaves minutely denticulate.
Begonia novoguineensis sp. nov. § Petermannia. Fig. 7, a, b.

Planta 60 cm. alta; ramulis parce hirsutis dein glabratis; foliis ovato—
ellipticis, 16— 29 cm. ‘longis, 10-15 cm. latis, apice breviter subabrupte
acuminatis, acumine 2 cm. longo basi 1 cm. lato, basi latere exteriore in
lobum 2-5 cm. longum rotundato-cordatum nervis basalibus 5, interiore
rotundatis nervis basalibus 2, lateralibus utrinsecus circiter 5, margine sinu-
atis inconspicue serrato-denticulatis setulosis, supra glabris, subtus costa
nervisque sparsim setulosis; petiolo 2-7 cm. longo parce hirsuto; stipulis
+ 1 cm. longis anguste oblongis, apice setulosis caducis; inflorescentiis 3
tantum visis usque 30 cm. longis, racemoso-panicula tis, axi ramisqu
sparsim hirtellis, ramulis pedicellisque minute glanduloso- pilosis, bracteis
parvis stipulis similibus: pedicello 5-7 mm. longo; tepalis 2 late orbiculari-

Teese minute hirtellis vel glabratis; staminibus circiter 60,
filamentis quam antheris brevioribus, antheris 1 mm. longis oblongo-
obovatis een ae rimis 14 antherae aequilongis.

NETHERLANDS NEw Guinea: Hollandia, Brass 8841 (1ypPE), June 1938, alt. 50 m.,
deep ravine (60 cm. high; leaves red beneath; flowers white).

Among the descriptions of Papuasian species of Begonia, perhaps this is
to be compared to B. brevirtmosa Irmsch., but it is a much less pubescent
plant with about twice as many stamens; the young leaves are rather
densely hirsute but very quickly become glabrate.

Begonia subelliptica sp. nov. § Petermannia. Fig. 7, c-e.

Planta 1 m. alta foliis novellis exceptis glabra; foliis novellis 5 c
longis, ee costa nervisque hirtellis, maturis late obovatis vel ellipticis,

20-25 cm. longis, 13-16 cm. latis, aoe obtusis vel rotundatis, basi latere
exteriore in lobum 5—6 cm. m, longum undato-auriculatum productis nervis
basalibus 6, interiore rotundatis ey basalibus 3 vel 4, lateralibus utrin-
secus 4, margine subsinuatis integris; petiolo 5-7 cm. longo; stipulis
oblongis + 3 cm. longis, apice obtusis, breviter setosis; inflorescentiis 9-15
cm. longis flores ¢ atque @ gerentibus; floribus ¢ in ramis brevissimis
fasciculatis; pedicello usque 2 cm. longo; tepalis 2 suborbicularibus 1.3—1.8
cm. diametro: staminibus 30-35 basi connatis, filamentis 1-1.5 mm. longis,

longa et lata, _ 3 subaequalibus medio 6 mm. latis, basi atque apice
angulis rotundat

NETHERLANDS Mes GuInEA: 4 km. southwest of Bernhard oe Idenburg River,
Brass 13396 (TYPE), March 1939, alt. 850 m., undergrowth of flood-plain rain-forest
(single clump 1 m. high; leaf-veins red bene flowers white, veined with pink).
Begonia Somervillei Hemsl. Kew Bull. 1896: 17. 1896.

Sotomon Istanps: San Cristobal: Waimamura, oe 20644, August 1932,
steep rocky or gravelly slopes sheltered by rain-forest, ver wy fleshy
shrub growing often in dense masses; stems and petioles veddish inflorescence pale
ink).

Described from New Georgia and recorded only by the original de-
scription

58 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Begonia Augustae Irmsch. Bot. Jahrb. 50: 350, f. 2, A-G. 1913.

Br New GutneEA: Oroville Camp, Fly River, Brass 7392, 7417, August 1936,
plentiful on steep river bank of red clay (robust plant over 1 m. high; leaves dark
green, very glossy above; flowers and fruit red in 7392, lowers white in 7417). North-
eastern New Guinea.

This should also be compared with Begonia sogerensis Ridl.

=) >

ZD
/ TaN Mag sey

fy mm
y Pe ae Nee
\ on ee. ees
; .

a, b. Begonia novoguineensis Merr. & Perry, drawn from the type: a. habit,
x 4; b. 6 flower, x “%. c-e. Begonia subelliptica Merr. & Perry, drawn from the
type: c. habit, x 14; d. 6 flower, x 1%; e. capsule, x %. f-j. Symbegonia papuana
Merr. & Perry, drawn from the type: f. leaf, x ™%; g. g flower, x ™%; h. cluster of
stamens, enlarged; 7. 2 flower, x ™%; j. capsule, x 4.

Begonia spilotophylla F. vy. Muell. Descr. Not. Pap. Pl. 4: 67. 1876; D’Albertis,

uin. 2: 398. 1880; non sensu K. Schum. & Lauterb. (fide Irmscher, Bot.

Jahrb. 50: 345. 1913).

British New Guinea: Fly River, 528 mile Camp, Brass 6702, 6730, May 1936, alt.

80 m., occasional in forest undergrowth on ridges, gregarious and very abundant in

semi-shade on steep slopes above river (very conspicuous fleshy herb 30-40 cm. high;
leaves spotted with silver, the veins purple; flowers pink)

papillose lobes; otherwise as in the original description. These parts had
apparently fallen before the original specimens were collected.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XI 99

Symbegonia Warburg
Symbegonia papuana sp. nov. Fig. 7, f-j.
Planta 30-50 cm. alta; caule, ramulis, petiolis et axi inflorescentiae +

dense et crispule ferrugineo- hirtellis foliis chartaceis inaequaliter oblongo-

0.5-1.5 cm. longum rotundatum productis nervis basalibus 3, interiore
rotundatis vel cuneatis nervis basalibus 1 vel 2, lateralibus utrinsecus 3 vel
4, margine remotiuscule denticulatis interdum repandis, supra consperse et
crispule subhirtellis vel pilosis, subtus costa nervisque crispe hirtellis;

inflorescentiis terminalibus usque
7 cm. longis; bracteis glabris late ovatis membranaceis verisimiliter albis;
floribus 8: pedicello brevi; tepalis 2 ovato-ellipticis, circiter 1 cm. longis,
consperse et pain versus dense crispule hirtellis; staminibus + 15, fila-
mentis in columnam brevem connatis, antheris late oblongis obtusis 1 mm.
longis; floribus @: pedicello 0.5 (fructus 1-1.5) cm. longo; alec
campanulato in sicco usque 1.5 c . longo (lobis inclusis 5 mm. longis
obtusis) et 0.8 cm. lato, extus acne crispe pilosulo; stylis 3, basi 1 mm.
connatis, deinde in crura 2 stricto- ae 5 mm. longa torta partitis; ovario
ellipsoideo 1 cm. longo, 0.6 cm. lato, sparsim piloso, alato; capsu m.
longa et lata, alis subtriangularibus aequalibus, basi aren “apice
acutiuscule angulatis apiculatis, margine superiore 6—7 mm. longis, 3 mm.
medio latis.

NETHERLANDS NEw Guinea: 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 12161 (type), January 1939, alt. 1750 m., common in rain-forested gullies (plant
30-50 cm. high; leaf-nerves red beneath; flowers white).

This plant appears to be most like Symbegonia strigosa Warb., but it has
a much shorter indument and larger and differently proportioned fruits.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

60 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

STUDIES IN THE THEACEAE, XIV
NOTES ON THE WEST INDIAN SPECIES OF TERNSTROEMIA

CLARENCE E. KoBuskI

In 1896, Urban (Bot. Jahrb. 21: 521-544) presented an excellent and
rather complete treatment of the West Indian species of Ternstroemia. The
present contribution is hardly more than a review based upon Urban’s work.
As a result of study of the American species as a whole, I feel that Urban
placed too much significance on the extent of connation in the petals and
on other variable characters, using them extensively in his key to the species.
Furthermore, the new species, which were described under the joint author-
ship of Krug and Urban, were often based on minor and variable charac-
ters. Nearly every number collected by C. Wright in Cuba was designated
as a new species or variety. These species are not outstanding and in
several instances have been reduced to synonymy in this paper.

presentation of the South American and the Mexican and Central
American groups (Jour. Arnold Arb, 23: 298-343, 464-478. 1942), com-
plete description of all the species were offered, many of them for the
first time. However, in the present paper, because of Urban’s very care-
fully drawn descriptions, it is unnecessary to continue this practice, and
instead only the salient characteristics, along with discussions of specific
relationships and differences, are recorded. Here, for the first time, to my
knowledge, is mentioned a pubescent species of Ternstroemia, T. pubescens
from Santo Domingo.
KEY TO THE SPECIES
A. Pedicels, bracteoles, calyx-lobes, lower surface of leaves and current year’s growth
covered with a short villous pubescence (Santo Domingo) ...... 1. T. pubescens.
AA. Entire plant strictly glabrous.
Ovary and fruit one- or three-celled.
Cc. re three-celled.
D. Leaves membranaceous, not granular-punctate; stigma subcapitate,
tri-crenate; each ovary cell single-seeded (Trinidad, es Guiana).
z.

ava hidgeig Diavbia lana esleecisie ys eee te a ened eae ners «ead . delicatula.

DD. 7 heavy-coriaceous, densely eine ster bins on ie surfaces ;

three-parted, evolute; ovary cells more than one-seeded (Guade-

lou age art St. Kitts, Dominica)................3. T. elliptica.

CC. Ovary shane Ea teioe & ¥eapnale snare Ns wk eae oe 4. T. parviflora.
BB. Ovary two- or fou

C. Ovary four- arr (Tobago, PEMIdad) .0s'sf<04s sh eee dos 5. T. oligostemon.

CC. Ovary two-ce
D. Pedicel oth ene bracteoles in two opposite pairs eae Rico).......
Sta Sip vara Gre ge bia Goa Heenan ox earaio Goe Fie eins Sarin a ee . heptase pala.
DD. Pedicel with two opposite bracteoles.
. Flowers and fruit sessile or subsessile (Jamaica)....7. T. subsessilis.
EE. ene and fruit distinctly pedicellate.

. Calyx- pi remarkably large, the smallest over 1 cm. long

(1.2-1.7 cm.).

1943 | KOBUSKI, STUDIES IN THE THEACEAE, XIV 61

G. Calyx- ei Egon denticulate, up to 17 mm. long; brac-

ie g, ovate, eglandular: pedicels 1.5—2.0 cm.
(Jam ie Pe a et ee Ce ee ee AA . calycina

GG. ates fens eglandular, 10-12(-14) mm. long; bracteoles
minute, not over 3 mm. long, yaa pee 6-8 cm.

long Glamaica’), x 5<.Ae'< ce, oa, Girne eater aan . I. rostrata.
FF. Calyx-lobes under 1 c g.
G. Calyx-lobes and peicdies eglandular
H. aie elliptic-oblong, 2—3 times fonces than broad, acute
apex, up to 12 cm. long (Porto Rico) .10. T. luquillensis.
HH. pee broadly ovate-elliptic, suborbicular, rounded and
lightly emarginate at apex, 3-4 cm. long, 3-4 cm. wide
CEL ALE ee Soci chats ob do se auacttigh care 11. T. Barkeri.
GG. Calyx-lobes and bracteoles glandular-denticulate.
ee long, conspicuous, linear, foliar, up to 6 m
ng UD) bcs con reat efor eg ee W220 Nashii
HH. eae — over 3 mm. long, not foliar and usually
inconspicuo
I. Largest oe less than 3 cm. long.
. Outer calyx-lobes very small, ca. 2.0-2.5 mm.
long; bracteoles not over 1.5 mm. long, not keeled

GEV ATE ie 53 ce carsaciacdove ib taten av ee 13. T. Selleana.
JJ. Calyx-lobes 4-6 mm. long; bracteoles 2-4 mm.
long, keeled.

. Leaves rounded, spathulate, not revolute; brac-
teoles long- deltoid, ca. 3 mm. long (Haiti).....
ee ee eee er ee 14. T. gracilifolia.
KK. ination thick-coriaceous, ovate or broadly ovate,
inte bracteoles shortly triangular, acute,
m. long (Cuba)........... Poo?" eas.
II. Largest ee well over 3 cm. long.
J. Pedicels less than 1.5 cm. long.
K. Pedicels ae — 0.5-0.9 cm. long, usually
averaging 0
. Leaves omen ‘orbicular (Cuba)
Roea nr aay 3 lg un a iatireia olay ee . I. baracoénsis.
LL. Leaves elliptic-oblong, aeually “tae —3 times
] (Cuba)..... 17. T. cernua.
KK. Pedicels 1.0-1.5 cm. long; calyx- lobes : mm, or
18.

lessyG@Uba) Settee sc . microcalyx.
JJ. Pedicels longer than = cm., usually much longer,
in some cases u

o8c
K. en hg lobes spline sharp- ene (Porto

CO) Pawnee tienen erie were eneioe T. Stahlii.
KK. ae. lobes ee rounded or oe
Leaves with 12 or 13 pairs of veins, clearly

visible on the ae surface pha ieeeienets

CM eOR ey Ar ie Brees Ae Cae.

LL. Leaves with 5-8 pairs of veins, ‘usually in-
conspicuous on the lower surface

. Leaves rounded or very obtuse at the

base ; ae 6-12 mm. long; fruit semi-

oval (Jamaica)........... 21. T. Hartii.

MM. Leaves Baie at the base; petiole 3-7

. long; fruit conical (Cuba, Haiti,

Santo fone seit ae St. Jan, St.

62 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

1. Ternstroemia pubescens, sp. nov.

Ramuli grisei, glabri, hornotinis brunnescentibus breviter villosis. Folia
elliptica vel obovata, apice ramulorum congesta, coriacea vel subcoriacea,
(2.5—) 5-6 cm. longa et 1.5-3.0 cm. lata, supra glabra opaca (ut videtur
rubida), subtus breviter villosa, pallidiora, apice obtusa, basi in petiolum
protracta, margine integerrima vel subrevoluta, costa supra impressa (in
toto manifesta), venis 5 vel 6 paribus, supra obscuris, subtus pauce mani-
festis, petiolis 8-14 mm. longis, villosulis. Flores parvissimi, in foliorum
axillis solitarii, pedicellis 8-14 mm. longis breviter villosis, bracteolis 2
oppositis vel suboppositis ca. 1 mm. minusve longis villosulis triangularibus
vel subtriangularibus sparse glanduloso-denticulatis; sepala 5, imbricata,
subaequalia, exterioribus villosulis ca. 2.5 mm. longis et 2.0 mm, latis

leviter coalita, ca. 3.5 mm. longa, quam calyce pauce longiora; stamina
ca. 30, filamentis antherisque ca. 1 mm. longis; ovarium parvum, 2-loculare,
glabrum, basi 1.5—2.0 mm. diam., in stylum 1.5 mm. longum attenuatum,
stigmatibus bipunctiformibus. Fructus ignotus.

DistrRIBUTION: Santo Domingo.

Santo Dominco: Cordillera Central, Prov. Monte Cristi, Moncion, Lagunas de
Cenobi, common in forest, alt. 1100 m., FE. L. Ekman H-12879 (TYPE, US), June 17,
1929.

The astounding distinguishing feature of this species is the presence of a
short villous pubescence on the lower surface of the leaves, the current
year’s growth of branchlets, the pedicels, bracteoles and calyx-lobes.
This separates the species from all others in the genus. To my knowledge
it is the first record of any kind of pubescence in Ternstroemia. Very
closely related is T. microcalyx Krug & Urban. In the latter species the
sepals are equally as small but are entire, lacking the distinct glandular
denticulations on the margin as in 7. pubescens. Also, the petals in
T. microcalyx are 7 mm. long, twice as long as the petals in T. pubescens,
which barely exceed the calyx-lobes in length. In 7. microcalyx the petiole
is considerably shorter (5-8 mm.) and the attenuation of the leaf-base into
the petiole is more pronounced.

1: 70. 1929.— Kobuski in Jour. Arnold Arb. 23: 308. 1942

Mokofua delicatula (Choisy) O. Kuntze, Rev. Gen. PI. 1: 63. 1891.

Taonabo delicatula (Choisy) Szyszylowicz in Nat. Pflanzenfam. HI. 6: 118. 1893.

DistripuTion: Trinidad, French Guiana.

Trrnipap: Forests near Arima, alt. 600 m., H. F. A. Eggers 1381 (NY, US). Frencu
Gutana: Cayenne, Martin s.n. (isotype, FM; photos, FM, G).

This species is described as having membranaceous, obovate-elliptic
leaves, narrowed at the base into a long petiole (6-15 mm. long), obtuse
or very shortly acuminate at the apex, 5—9 cm. long and 2-4 cm. wide,
quite distinctly crenulate in the upper half, and with lateral nerves prominu-

1943 ] KOBUSKI, STUDIES IN THE THEACEAE, XIV 63

lous on both surfaces. The flowers are fairly numerous, crowded on the
branchlets. The peduncle is slender, 1-2 cm. long, recurved. The 5 sepals
are more or less equal, 5—6 mm. long, obtuse at the apex, with eglandular
entire margins. The ovary is globose-conical, 3-celled, contracted into a
tyle ca. 4 mm. long which is topped by a subcapitate tri-crenate stigma.
Each cell of the fruit is single-seeded.

The membranaceous leaves, the long petiole, the crowded flowers, the
thin, recurved pedicels, the eglandular sepals, and the tri-crenate, ‘sub-
capitate stigma are the distinguishing characters for identification. This
is the only species which is found both in South America and the West
Indies. However, its closest relationship is with T. Browniana Kobuski of
British Guiana. Both have entire, scarious-margined, eglandular, small
(S mm. long) calyx-lobes, 3-celled ovaries.and fruit with a single seed to
each locule, and slender pedicels. Ternstroemia Browniana differs in having
a punctiform rather than a subcapitate, tri-crenate stigma, and in the
coriaceous, veinless leaves, shorter and rounded at the apex.

3. Ternstroemia elliptica Swartz, Prodr. 81. 1788; Fl. Ind. Occ. 2: 929. 1800. —
Vahl, Symb. 2: 61. 1791. — De Candolle in Mém. Soc. Phys. Hist. Nat. Genéve

Jahrb. 21: 535. 1893. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 142. 1925.

Mokofua elliptica (Swartz) O. Kuntze, Rev. Gen. Pl. 1:

Amphania integrifolia Solander Mss. ex oo in Mém. Soc. Phys. Hist. Nat.

Genéve, 1: 410 (Mém. Ternstr. 18). 18

DIsTRIBUTION: Guadeloupe, Martinique, . Kitts, Dominica.

GUADELOUPE: Trois-Riviéres, alt. 300-600 m., Pére Duss 2987 (FM, NY, US), 1893
(grand bel arbre, rare; feuillage trés vert.; dans ‘ik haute région, comme a ie Savane a
Mulets, cet arbre Sak a état @arbriesen rabougri). — Soufriére, somets volcaniques
humides, alt. 1100 , H. Stehlé 322, 1023 (NY), Févr.-Sept. 1936 (2 m., arbrisseau
rabougri a cette altitude). — Savane a Mulets forét dense ventée, alt. 1100 m., H. Stehlé
1524 (US), 8 Février 1937 (rabougrie). Martin1quvE: Bois de Fonds Saint: Denis, de
Case Pilote et de la fontaine Absalon, alt. 300-600 m., Pére Duss 638 (NY, US), 1882-
1883; Pére ae 171 ae Mo, NY, US; not Duss 171, Berlin), 1887. St. Kirts: Upper
donee of Mt. ae tie Bp Britton & J. F. Cowell 532 (NY, US). Dominica: Dr.
Imray 280 (G, neh (fide Urban).

This species is characterized by coriaceous leaves, 4.5—-9.0 cm. long and
2-4 cm. wide, obovate to obovate-elliptic, granular-punctate on both sur-
faces, very obtuse at the apex, contracted at the base into a petiole 5-10
mm. long, the margin revolute, entire, the veins (ca. 7 pairs) obsolete or
prominulous on the lower surface; peduncles 1-2 cm. long, the bracteoles
ovate, the 5 sepals unequal, semiorbicular, 4-7 mm. long, ca. 5 mm. wide,
the outer lobes glandular-denticulate, the petals 5, obovate, 8-10 mm. long,
6-8 mm. wide, the stamens numerous, 2—4-seriate, the filaments unequal,
the inner filaments nearly 3 times longer than the anthers, the outer fila-
ments about equal to anthers in length; ovary semiglobose or conical,
3-celled, tapering through the style to a 3-parted, evolute stigma which sur-
passes considerably the style in diameter; fruit three-celled, up to 18 mm.
diam., the seeds few

The characters significant in identification are: the three-celled ovary

64 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

and fruit, the 3-parted evolute stigma, the long petiole, the granular
punctations on both leaf-surfaces, the very numerous stamens in 2—4 series,
and the eglandular sepals and bracteoles.

Cited above is Duss 171 from various American herbaria. Evidently
this specimen must differ from Duss 171 as found in the herbarium at
Berlin, since Urban (1893) cited the number as belonging to his new species
T. oligostemon, which, in turn, is very different from the present species.

According to the labels and dates, Duss made frequent collections over a
period of years, in different localities, assigning a single number to the
massed collections.

4. haart parviflora Krug & in Bot. Jahrb. 21: 523. 1896. — Melchior
Nat. Pflanzenfam. ed. 2, 21: 142.

DistripuTIon: Cuba.

Cusa: Loma del Pinal Mayan, along road, C. Wright 2110 (1sotypEs, G, M), large
bush, 10 ft.; flowers whitish rose, tinged with yellow at center. rov. Santa
Clara: Palm Barren, N. L. Britton & J. F. Cowell 10178 (NY, US), March 1911. —
Santa Clara to Loma Cruz, N. . poi E. G. Britton & J. F. Cowell 10220 (NY,
US), March 23, 1911.— Palm Barren, N. L. Britton, E. G. Britton & P. Wilson 6168
(NY, US), March 1910 (shrub 2 m.). Prov. Pinar del Rio: on top of
Cajalbana, Bros. Léon & Charles 4955 (NY), April 6, 1915 (shrub 4-5 ine — Hato
Abajo, J. T. Roig 3185 (G, NY, US), Apr. 7, 1924. Prov. Orien
streamlets or edge of deciduous thickets, J. A. Shafer 1282 (FM, NY, US), oP (FM,
NY), 1691 (NY), 3189 (FM, NY, US), Apr.-Dec. 1909 (shrub 1-3 m.).

In his original description, Urban states that the ovary is one-celled,
seven-ovulate. This character seems to hold for the type. However, such
is not the case in all specimens. After dissecting several ovaries from each
specimen, I find that Shafer 1282 and 1691, as well as Britton, Britton and
Cowell 10220, possess single-celled ovaries. Shafer 1424 and 3189, Bros.
Léon and Charles 4955 and Britton, Britton & Wilson 6168 have two-
celled ovaries, while Britton & Cowell 10178 have both single-celled and
two-celled ovaries, with the larger number single-celled. The ovary is
very minute (1 mm. or less long), conical in shape, and tapering into a
style ca. 2 mm. long. Because of the surrounding subligneous calyx-lobes
and the flatness of the ovary, dissections are very difficult.

The pedicels are very slender and are 2.5-3.0 cm. long. The bracteoles
are ca. 2 mm. long, narrow and acute at the apex. Although not constantly
so, there is a tendency for one of the bracteoles to be placed as much as
2 mm. below the sepals. Nearly every specimen shows this arrangement.
The calyx-lobes are 3-5 cm. long and ovate, and the outer lobes are some-
what apiculate. The anthers also are long- apiculate (1 mm.).

Most closely allied to this species is T. microcalyx Kr. & Urb., which
can be separated by the shorter stockier pedicel (1.0—1.5 cm. long), sealed
calyx-lobes, and longer petals. The petals of T. parviflora scarcely
exceed the calyx in length.

5. Ternstroemia oligostemon Krug & Urban in Bot. Jahrb. 21:
Melchior in Nat. Pflanzenfam. ed. 2, 21: 142. 1925. —R.O. Wiliams A Enea
& Tobago, 1:70. 1927.— Kobuski in Jour. Arnold Arb, 23: 327.
DistrRiBUTION: Tobago, Trinidad.

1943 |] KOBUSKI, STUDIES IN THE THEACEAE, XIV 65

Tospaco: The Widow, W. E£. Broadway 4154 (FM, Mo, US), Sept. 29, 1910 (shrub
with white, sweet-smelling flowers).— Easterfield, W. E. Broadway #369 (FM), Dec.
16, 1912 (bark of trunk rough and dark in color; leaves glossy-green).— Slopes of
main ridge above Parlatuvier, relict forest paueieriig cultivations, V. Y. Sandwith 1916
(NY), Oct. 24, 1937 (middle-sized hee with white flowers).— Exact locality missing,
F. A. “Durity” 12620 (NY), Jan. 2

This species is ene by. obovate or narrowly obovate-elliptic
leaves, 6-10(—14) cm. long and 2.5—5.0 cm. wide, shortly or obtusely
acuminate at the apex, long-attenuately tapering at the base into a petiole
8-13 mm. long, the margin subrevolute, crenulate or occasionaliy entire,
frequently glandular, the 10-15 pairs of rather straight veins conspicuous
on the lower surface, sometimes obsolete above; the texture of the leaves
is thick-chartaceous and the surface is free from granular punctations.
The flowers are white with a sweet odor and the pedicels measure 1.0—2.5
cm. in length. The sepals are suborbicular, 5-6 mm. long and about 5 mm.
wide and, like the narrowly ovate bracteoles, devoid of glandular denticula-
tions. The petals are about 7 mm. long. The stamens (ca. 20) are about
5 mm. long, the filaments measuring only 1 mm. in length while the anthers
are 4 mm. long, linear, and taper gradually to the apex. The ovary is
conical, 4-celled, each cell having one or two ovules and tapering into the
style, which is crowned by an entire stigma slightly exceeding the style in
diameter. The fruit is globose, 12-20 mm. in diameter, 4-celled with one
or two seeds in each cell, only one of which usually fully matures.

One of the types of this species, as cited by Krug and Urban, is Pére Duss
171 from Martinique. Before me are several specimens of this collection
(supposed isotypes, FM, Mo, NY, US), which belong not to this species,
but to T. elliptica. Perhaps there may be some confusion in the label of
the Berlin specimen, since all seven sheets of Duss 171 in American
herbaria are true 7. elliptica. On the label of two specimens in the New
York and U. S. National herbaria are two numbers, 171 and 638. There is
no difference in the material, but Duss 638 has been cited by Urban under
T. elliptica. I doubt very much whether 7. oligostemon actually grows in
either Martinique or Guadeloupe, from which it has been cited. Williams
cites material from Trinidad (none of which I have seen), and this, with
that from Tobago, perhaps gives the correct geographical distribution for
the species.

Krug & Urban’s description is very complete. They state, however, that
the ovary and fruit are 2-celled or incompletely 4-celled and that the num-
ber of ovules and seeds in either case is four. All material sectioned by
me showed the ovary and fruit to be clearly 4-celled, with one or two ovules
in each cell. In the first, eight seeds were found, four of which were fully
mature, while the other four, although immature, were of considerable size.
In other cases, a single seed was found in each cell.

The distinguishing characters of 7. oligostemon are the four-celled ovary
and fruit, the eglandular sepals and bracteoles, the entire stigma, and the
10-15 pairs of lateral veins of the thick-chartaceous leaves. Its closest ally
is T. delicatula of Trinidad and French Guiana. This latter species can be

66 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

distinguished by the 3-celled ovary, the tricrenate stigma, and the thin
papery leaves with 7 or 8 pairs of lateral vein

6. Ternstroemia capri do Krug & Urban in Bot. Jahrb. 21: He 1896. — Urban,

nd. Occ. 4: 411. 1910. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 142. 1925.
Taonabo ergo eh & Urb.) Britton in Britton & Wilson, Sci. em Porto Rico
& Virgin Isl. 5:

DISTRIBUTION: nie “i

Porto Rico: — de Luquillo, in Lappe montis Jiminez, P. Sintenis 1425 (1so-
types, FM, G, Mo, NY, US), June-July 1885.— Luquillo Mts., P. Wilson 151 (FM,
NY), July 1902. oe W. E. Hess Ae L. Stevens 2949 (NY), Aug. 28, 1913.

According to the name and Urban’s subsequent interpretation, this species
is characterized by seven sepals and a single pair of bracteoles. However,
to my mind, the more correct interpretation is to consider the flower as
having the customary number of sepals (five) and two pairs of bracteoles.
The outer pair of bracteoles are unequal, opposite, ovate or triangular,
keeled, eglandular, and about 1 mm. long. The bracteoles of the second
or inner pair alternate with those of the outer pair and are subrotund,
measuring 2 mm. or less in length, being also eglandular. The — them-
selves are imbricate, eglandular, se vuln varying from 3—5 mm. in
length. The pedicels are curved, 10-18 ng

Closely allied is T. subsessilis, which can ce separated by the single pair
of bracteoles, the two-parted style, and the subsessile flowers.
7. Ternstroemia subsessilis (Britton), comb. n

Taonabo subsessilis Britton in Britton & Wilson, ‘oe Surv. Porto Rico & Virgin Isl.

5: 581. 1924

DistTRIBUTION: Porto Rico.

Porto Rico: Luquillo Mts., rocky summit of Mt. Yunque, V. L. Britton & E. M.
Bruner 7627 (NY, TyPE), Feb. 1923 (shrub 1 m. high; petals white, 1 cm. long).—
Sierra de Naguabo, Rio Prieto and adjacent ren Seainees ng top of peak, alt. 1000 m.,
J.A. Shafer 3648 (NY, US), Aug. 1914 (small t

This interesting species is coun is ee or subsessile flowers
and fruit. This character, along with the two-parted style, is sufficient to
separate T. subsessilts es all other species of the genus in the West Indies.
The suborbicular eglandular calyx-lobes are very minute, the inner lobes
not over 3 mm. and the outer lobes about 2 mm. long. The bracteoles
resemble the calyx-lobes in shape but seldom measure over | mm. in length.
Britton states that the calyx-lobes (inner) measure 5 mm. and the petals
measure 10 mm. In the type specimen I could find no calyx-lobes measur-
ing more than 3 mm. and the petals were lacking. The fruit is conical,
tapering to a decided point. Because of the lack of material, no dissections
have been made to determine the number of cells of the ovary and the
number of ovules.

8. ee calycina Fawcett & Rendle in Jour. Ne"? 60: i Fl. Jam.
21 926. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 142.

oe Jamaica

Jamaica: W. Harris 10979, 11035 (1sorypres, NY).

Long ovate calyx-lobes (13-17 9-12 mm.) and bracteoles (7-9 mm
long) are the distinctive characteristics of this species. The calyx-lobes

1943] KOBUSKI, STUDIES IN THE THEACEAE, XIV 67

are glandular-denticulate, while the bracteoles are distinctly eglandular.
The pedicels are rather short (1.5-2.0 cm. long), and the leaves are
obovate-elliptic, rounded or very obtuse at the apex, sharply contracted at
the base into a petiole 5-8 mm. long.
9. hace gic rostrata Krug & Urban in Bot. Jahrb. 21: 533. 1896. — Melchior in
Pflanzenfam. ed. 2, 21: 142. 1925.— Fawcett & Rendle, Fl. Jam. 5: 184.
oe

DIsTRIBUTION: Jamaica.

W. Harris 10259 (FM, NY, US); 10317 (FM, NY, US); N. L. Britton
2339 (NY); N.L. Britton & A. Hollick 2168 (NY)

The fruit of this species is spheroidal or roundish-ovoid, rostrate at the
apex. Krug & Urban record the pedicel as 3-5 cm. long. In all the
specimens cited above, the pedicels are 6-8 cm. long, seldom less, and as
much as 3 mm. diameter at the apex. The calyx-lobes are intermediate
between those of 7. granulata and T. calycina (10-12 [-14] 7-10 mm.)
in size and are eglandular. The bracteoles are comparatively minute,
measuring not over 3 mm. in length, and acuminate. The leaves are
obovate to obovate-elliptic, 7-12 cm. long, 3.0—5.5 cm. wide, obtuse, and
occasionally emarginate at the apex, tapering at base into a petiole 5—7 mm.
long.

10. Ternstroemia luquillensis Krug & Urban in Bot. Jahrb. 21: 531. 1896. — Urban
Fl. Ind. Occ. 4: 411. 1910. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 142. 1925.
ee Ce ae Urb.) Britton in Britton & Wilson, Sci. Gury. Porto Rico
& Virgin Isl. 5:
ieee one Porto pie
orto Rico: Sierra de Luquillo, in monte Jimenez, in silvis, P. Sintensis 1523
(1so-SYNTYPES, M, NY, US), 1331 (1so-synTyPEs, G, } S), July-Aug. 1885. —
Sierra de Luquillo, in woods, alt. 600 m., H. F. A. Eggers 1224 (US), May 1883.

Most species of Ternstroemia are characterized by obovate leaves. In
this species the leaves are elliptic, up to 12 cm. long, usually 2.5—3.5 times
longer than broad, acute at the apex rather than obtuse or rounded, with
black punctate dots on the lower surface. The midrib above is impressed
the whole length of the leaf. The margin is entire, flat or slightly recurved.
The pedicels are slender, 3-9 cm. long, usually compressed. The bracteoles
are ovate or suborbicular, 4-5 mm. long, 3-5 mm. wide, with a midrib
evident near the apex and prolonged into a short apicule. The sepals are
unequal, up to 9 mm. long and 11 mm. wide, and eglandular.

11. Ternstroemia Barkeri Ekman & Schmidt in Rep. Spec. Nov. 24: 78. 1927.

DISTRIBUTION: Haiti.

Haiti: Massif de la Hotte, western ae ae top of M. Formond, alt. 2225 m.,
E. L. Ekman H-7483 (isotype, US), Jan. 1

This species is characterized A ae eptie (or rarely broadly ovate-
elliptic) nearly rotund leaves, 3-4 cm. long and 3-4 cm. wide, rounded and
lightly emarginate at the apex, obtusely and abruptly contracted at the
base into a petiole 3-5 mm. long, with the margin revolute, entire or nearly
so, and the veins inconspicuous. The pedicels are 2-4 cm. long. The
bracteoles are unequal, one subrotund, the other triangular, ca. 3 mm. long,

68 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

2-3 mm. wide, entire, not glandular-denticulate. The calyx-lobes are sub-
equal, 8-9 mm. long, 7-8 mm. wide, broadly ovate, the margins scarious
without glandular-denticulations. The petals are lanceolate, acuminate,
12-13 mm. long, connate at base. The stamens number ca. 50; filaments
up to 8 mm. long; anthers ca. 4 mm. long. The ovary is semi-globose,
bi-loculate.

The nearest relative is 7. baracoénsis O. C. Schmidt, which can be sepa-
rated from the present species by the glandular-denticulate bracteoles and
calyx-lobes, the very short pedicels (5—6 mm.), and the conspicuous veins.
12. beeen Nashii Urban in Rep. Spec. Nov. 13: a Auer Fl. Ind. Occ. 8:

36. 1920. — Melchior in Nat. Pflanzenfam. ed. 2, 21:

Risener Haiti.

: Marmelade, Camp No. 4, pineland, alt. 850 m., G. V. Nash & N. Taylor
1307 (1sotyprE, NY), Aug. 1-2, 1905 (shrub 1-2 m.).— Massif du Nord, Marmelade,
Jalousiére, alt. 1000 m., E. L. Ekman H-8254 (US), May 24, 1927

The outstanding feature of this species is the presence of the long brac-
teoles (up to 6 mm.), by which it can be quickly separated from T
peduncularis DC., its nearest relative. The leaves of T. Nashii are coria-
ceous, obovate-elliptic to obovate, 4-7 cm. long and 1.5—3.0 cm. wide,
rounded at the apex and tapering at the base into a petiole 2-3 mm. long,
the margin revolute and glandular-punctate, the veins inconspicuous. The
peduncles (3.5—5.0 cm.) immediately suggest T. peduncularis, but the
linear, glandular-denticulate bracteoles mentioned above distinguish it.
The calyx-lobes are unequal, apiculate, up to 10 mm. long and 7-9 mm.
wide, glandular-denticulate. The ovary and fruit are globular, tapering
into the style.

Urban suggests a relationship between this species and his 7. apleura.
The type specimen resembles the type specimen of 7. apleura. However,
subsequent collections show larger leaves with more conspicuous veining.
13. Ternstroemia Selleana Ekman & Schmidt in Rep. Spec. Nov. 24: 79. 1927.

DistTRipuTIon: Haiti.

Haiti: Massif de la Selle, gr. Crete-au-Piquants, Port-au- ft top of M. Malanga,
alt. 1475 m., E. L. Ekman H-7391 (1sotype, US), Dec. 16, 1926.— Massif de la Selle,
Ganthice, along path Badeau to Saltrou, alt. 2000 m., E. . Ekman H-3101 (US),
Jan. 27, 1925

This species is characterized by small, broadly obovate or suborbicular
leaves, usually 2—3 cm. long (rarely 3.5 cm.) and ca. 2 cm. wide, rounded
(or nearly so) and lightly emarginate at the apex, attenuated at base into
a petiole 4—5 mm. long, the margin revolute, slightly denticulate and plane
toward the apex, the veins 4 or 5 pairs (inconspicuous in Ekman H-3101).
In the type specimen the pedicels vary from 1.0—1.5 cm. in length; how-
ever, in Ekman H-3101 the pedicels are considerably shorter (0.5—0.8 cm.).
The bracteoles are unequal, 1.0-1.5 mm. long, scarious-margined and
sparsely (if at all) glandular-denticulate. The calyx-lobes are unequal
and suborbicular, the outer lobes smaller, ca. 2.5 mm. long, sparsely
glandular-denticulate, the inner lobes 3.0-3.5 mm. long, the margin scari-
ous. The young fruit is conical, up to 1 cm. long and only one-half as wide.

1943 | KOBUSKI, STUDIES IN THE THEACEAE, XIV 69

The nearest relative is T. gracilifolia O. C. Schmidt. This latter species
can be separated by the longer pedicels and bracteoles, the larger, triangular
calyx-lobes, and the semi-globose fruit.

14. Ternstroemia gracilifolia O. C. Schmidt in Rep. Spec. Nov. 22: 95. 1925.

DIstRIBUTION: Haiti.

Hartt: Massif de la Selle, Pétionville, M. La Visite, alt. 2050 m., E. L. Ekman H-1410
(1isoTyPE, US), Aug. 9, 1924

This species is characterized by small, coriaceous, spathulate leaves,
1.5—3.0 cm. long and 1.5—2.2 cm. wide, rounded at the apex, narrowed at
the base into a petiole 2-3 mm. long, the margin both revolute and
glandular-denticulate along the entire length, the veins inconspicuous.
The pedicels are 2.0-3.5 cm. long. The bracteoles are long-deltoid, ca.
3 mm. long, carinate, glandular-denticulate. The calyx-lobes are unequal,
the outer lobes subcordate-triangular, ca. 6 mm. long and 5 mm. wide,
strongly glandular-denticulate, the inner lobes slightly longer, somewhat
apiculate at the apex, and scarious-margined. The immature fruit is
globose.

Closely allied is T. Selleana Ekm. & Schmidt, which can be separated by
the smaller, rounded bracteoles and calyx- -lobes, only sparsely, if at all,
glandular-denticulate, the shorter pedicels, the elongated fruit, and the
leaf margin plane and denticulate only at the apex.

15. rity flavescens Grisebach, Cat. Pl. Cuba, 35. 1866. — Sauvalle, Fl. Cub.
1873. — Urban in Bot. Jahrb. 21: 530. 1896. — Melchior in Nat. Pflanzenfam.
? 21: 142, 1925.

Preiers vee clusiifolia Grisebach, Pl. Wright. 166. 1860. — Non H.B.K

Taonabo flavescens (Griseb.) Szyszylowicz in Nat. Pflanzenfam. III. 6; 189. 1893.

DIsTRIBUTION: Cuba.

Cuspa: Near Mont Verde, C. Wright 1124 (isorypE, NY).— Prov. Oriente:
Camp La Gloria, south of Sierra Moa, J. A. Shafer 8200 (NY), ie 1910 (straggling
shrub 1 m

The very small, thick-coriaceous, ovate to broadly ovate leaves (2.4 x
1.0-1.5 cm.) and the small flowers are the outstanding distinguishing char-
acters of this species. The nearest relative is T. baracoénsis Schmidt, the
leaves of which are considerably larger and subrotund.

16. Ternstroemia baracoénsis O. C. Schmidt in Rep. Spec. Nov. 22: 95. 1925.

DIsTRIBUTION: Cuba.

Cups: Prov. Oriente: as de Cuaba, near Baracoa, in pines, E. L.
Ekman 4230 (photo of isotype, NY), ae 13, 1925.

A photograph of the isotype (Stockholm) with a few leaves (NY) are
the only available material for study. Close observation shows flowers to
be present on the isotype. According to the description and the photo-
graph, the outstanding characteristics are (1) the broadly elliptic or nearly
rounded, coriaceous leaves, 3—5 cm. long and 1.8—3.0 cm. broad, rounded at
the apex, obtusely contracted at the base into the petiole, the margin entire,
slightly revolute, the midrib flat above, fading out toward the apex, raised
below, the veins (5 or 6 pairs) tenuous and conspicuous below; (2) the
pedicel short, 0.5-0.6 cm. long; (3) the bracteoles subcordate(?).

70 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

The nearly rounded leaves, the tenuous veins, and the very short pedicels
are characters which clearly distinguish the species from its Cuban rela-
tives. This species is probably the only Cuban species whose leaves are
not twice as long as broad. Unfortunately, the description of the flowers
was drawn only from buds and proved of little aid. Schmidt remarked that
the bracteoles were subcordate; this seems very dubious

The nearest relative of this species is T. flavescens Griseb. However, the
latter species has much narrower leaves with inconspicuous veining and
longer pedicels (1.0—2.5 cm.).

17. er cernua Grisebach, Cat. Pl. Cuba, 35. 1866. — Sauvalle, Fl. Cub. 10
prude in Bot. Jahrb. 21: 529. 1896.— Melchior in Nat. Pflanzenfam.
2,21:

i cernua Meee ) Szyszylowicz in Nat. Pflanzenfam. III. 6: 189. 1893.

apart Cuba.

CuBA: ov. Pinar del Rio: near Retiro and Sumdero, C. Wright 2112
(ISOTYPES, : "NY, US).

This species is characterized by obovate or elliptic-oblong, coriaceous
leaves, nearly three times longer than broad (5-8 2-3 cm.), with 6-8
pairs of veins visible on the under surface. The petiole exceeds that of
most species in length (5-10 mm.). The short, thick, recurved pedicels
(0.4-1.0 cm. long and 1.5 mm. thick) also distinguish this species. The
sharply pointed triangular bracteoles and the ovate, acuminate calyx-lobes
are added features which separate this species from 7. peduncularis DC.

18. Ternstroemia microcalyx Krug & geet in Bot. Jahrb, 21: 531. 1896.—O. C.
Schmidt in Rep. Spec. Nov. 22: 96. — Melchior in Nat. Pflanzenfam. ed.
2, 21: 142. 1925. — Non Airy-Sha
Ternstroemia elliptica Grisebach, Pl. te 166. 1860, Cat. Pl. Cuba, 35. 1866. —
n Swartz.

Dicahee Cuba, Haiti.

Cupa: Oriente, near La Guinea, C. Wright 1577 oe G, NY), Dec. 17, 1859
(tall shrub). — Southern Oriente a Pico Turquino, Fr. Leon 10737, 11073 (NY),
July 1922.— Oriente, Sierra Maestra, alt. 1200 m., J. ” Roig & Bucher 6901 (NY), July
1935. — Oriente, Sierra Maestra, Bucher (244) 5720 (NY), July 8, 1931. Harri:
Massif du Nord, Anse-i-Foleur, top of Mt. Colombos, alt. 900 m., E. L. Ekman H-4342
(US), _ 20, 1925.— Massif du Nord, Port-de-Paix, summit of Haut-Piton, alt.
1205 m., E. L. Ekman H-4623 (US), Aug. 9, 1925.

The leaves of this species are thin-coriaceous, obovate-elliptic or elliptic-
oblong, 4.5—7.5 cm. long and 1.5—3.0 cm. wide, obtuse or obtusely acuminate
at the apex, narrowing at the base into a petiole 5-8 mm. long. The leaf-
margin is plane and usually entire, with occasional slight evidences of serra-
tion near the apex. The midrib is narrowly and deeply sulcate the whole
length of the leaf above, raised below with 3—5 pairs of obscure nerves
evident on the lower surface. The pedicels are 1.0—-1.5 cm. long and the
small suborbicular bracteoles are less than 1 mm. long. The calyx-lobes are
minute, measuring 3 mm. or less in length, rounded at the apex, with the
margin of the outer lobes entire or occasionally obscurely glandular-denticu-
late. According to a subsequent observation by Urban, the petals far exceed
the calyx in length, measuring 7 mm. long.

1943] KOBUSKI, STUDIES IN THE THEACEAE, XIV 71

A close relative is T. parviflora, which can easily be distinguished by the
usually single-celled ovary and fruit, the longer peduncle (2.5—-3.0 cm
long), and the petals exceeding the le) in length. Another closely related
species is T. pubescens. The calyx-lobes, bracteoles, pedicels, lower surface
of the leaves, and the young branchlets of this latter species are covered
with a soft villous pubescence. Also, the petals measure only 3 mm. long,
barely exceeding the calyx in length, and the bracteoles and outer calyx-
lobes are sharply glandular-denticulate, with the inner lobes scarious and
subfimbriate.

n 1939, Airy-Shaw (Kew Bull. 506) described T. microcalyx from
Sarawak, apparently overlooking Krug & Urban’s earlier species of the
same name. No material of Airy-Shaw’s species is available for my study,
and since, as he states, his new species is so very closely allied to T. Robin-
sonii Merrill from Amboina, I merely wish to draw attention to this later
homonym rather than suggest a new combination.

19. brant oy at ae Krug & Urban in Bot. Jahrb. 21: ae 1896. Weuiiee FI.

d. Occ. 4: 910. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 1925.
Perntroemia tts Krug & Urban in Bot. Jahrb. 21: on a es
Fl. Ind. Occ. 1. ce. — Melchior, ]. c. — Syn

paren: Stahlii (Krug : oe Britton in ie & Wilson, Sci. Surv. Porto Rico
& Virgin Isl. 5: 581

Taonabo pachyphylla fae & Urban) Britton, 1. c.

DIsTRIBUTION: Porto Rico

Porto Rico: WN. L. Britton & E. G. Britton 9512, 9636, 9923, 9924 (NY).—WN. L.
Britton & J. F. Cowell 1556 (NY, US).— WN. L. Britton, F. L. Stevens & W. E. Hess
2445 (NY, US).—E. G. Britton & D. W. Marble 722 (FM, NY, US).— P. Sintenis
349 (G, Isotype of T. pachyphylla).— W. E. Hess 3863 (NY).—A. A. Heller 4545
(AA, FM, G).—J. A. Stevenson 2118 (NY, US).— JL. R. Holdridge 38 (NY).

The leaves are obovate, thick-coriaceous, 4-7 cm. long, rounded at the
apex, tapering at the base into a stout petiole, 3-5 mm. long; the midrib is
quite flat on the upper surface, fading out toward the apex, raised below;
the surface is distinctly granular below; the margin is flat, occasionally
remotely denticulate. The pedicels are stout, 5-12 mm. long. The brac-
teoles are long-triangular, 2.5—3.5 mm. long, 1.0-1.5 mm. wide at the
base, glandular-denticulate. The calyx-lobes are ovate, distinctly acute,
occasionally decidedly acuminate, the outer lobes being glandular-denticu-
late. The sharp apex of the calyx-lobes is an excellent diagnostic character.

Ternstroemia pachyphylla Krug & Urban is undoubtedly a synonym of
this species. In the type specimen the petioles and pedicels are slightly
shorter. However, in all other respects the two species are identical.
Heller 4545, jdenuaed by Urban (1910) as 7. StaAlii, is identical with the
type of T. pachyphylla.

20. oe granulata Krug & Urban in Bot. Jahrb. 21: 534. 1896. — Melchior
n Nat. Pflanzenfam. ed. 2, 21: 142. 1925.— Fawcett & Rendle, Fl. Jam. 5: 182.
ise.

DISTRIBUTION: Jamaica.

Jamaica: Purdie s.n. (probable iso-syntype, G).

This species is characterized by obovate-elliptic leaves, 7-9 cm. long and

72 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

3-4 cm. wide, densely granular-punctate on both surfaces, obtuse at the
apex, tapering quite abruptly at the base into a petiole 8-10 mm. long, with
the midrib evident the entire leaf-length on both surfaces and 12 or 13
pairs of veins, clearly evident on the lower-surface; pedicels 3.5—4.5 cm.
long, the bracteoles 2, suborbicular, 4-5 mm. long, 3—4 mm. wide, glandular-
denticulate, not keeled (Purdie); sepals 5, subequal, somewhat orbicular,
8-9 mm. long, 8-11 mm. wide, glandular- -denticulate (outer sepals) ; petals
5, 12-13 mm. long, 8—9 mm. wide.

Purdie s.n. (G) is probably an iso-syntype of 7. granulata, as originally
formulated by Krug & Urban. This specimen has densely granular-
punctate leaves, from which the species derives its name. However, the
description calls for bracteoles which are more or less carinate on the dorsal
surface. On bracteoles of the specimen cited above there is no evidence
of a keel. Later (1908), Urban determined questionably a specimen col-
lected by Prior near Brownstown. This specimen lacks any vestige of
granular punctation on the leaves, and here also there is no evidence of a
keel on the bracteoles.

Fawcett & Rendle included both these entities under 7. granulata in
their treatment of the Ternstroemiaceae for the Flora of Jamaica. Besides
the leaves, the larger flowers (both calyx and corolla) and the larger
number (12 or 13 pairs) of clearly visible veins should clearly distinguish
this species from 7. Hartii, its nearest relative.

21. oe Hartii Krug & —— in Bot. Jahrb, 21: 532. 1896. — Melchior in
. Pflanzenfam. ed. 2, 21: 1925.— Fawcett & Rendle, Fl. Jam. 5: 182.
ee

DISTRIBUTION: Jamaica.

Jamaica: St. Georges, alt. 650 m., W. Harris 5767 (1so-syNTYPE, FM, NY), June
14, 1895 (tree 8 m.).—Troy, alt. 600 m., W. Harris 8786 (NY), Aug. 1904 (tree

1908 (small tree 6 m. with white fragrant flowers).— Ipswich, St. Elizabeth, alt.
Om., W. Harris 12369 (FM, Mo, NY, US), June 13, 1916 (tree 8 m.).— Summit of
John Crow Peak, vicinity of Cinchona, NV. L. Britton 266 (FM, NY, US), Sept. 1906
Sige 8 m. with white flowers).— Grove Place, near Mile Gully, alt. 500 m., W. Harris
. L. Britton 10621 (FM, NY, US), Sept. 23, 1908 (tree 8 m.).—J. Hart 529
— — Near Brownstown, Prior s.n. (NY).

This species is characterized by obovate or obovate-elliptic leaves,
6-11 cm. long, 3.0-5.5 cm. wide, opaque, obtuse or rounded at the apex,
contracted abruptly or gradually at the base into a petiole 6-12 mm. long,
with the surface free from granular punctations, the margin plane or
slightly revolute and entire, and the veins obscure on both surfaces and,
when visible, few in number; pedicels varying in length, 1.5—5.0 cm. long,
usually somewhat carinate, the bracteoles 2, 3-5 mm. long, 2.5—3.5 mm.
wide, ovate, the margin usually free from glandular denticulations; sepals
5, subequal, suborbicular, 6.0-7.5 mm. long and about as wide, the mar-
gins of the outer sepals sparsely glandular-denticulate, usually entire,
scarious; petals barely exceeding the hag in length (

Several of these numbers (Prior s.n., W. Harris 8786, 8987 ane 12369,
and Harris & Britton 10621) have sn cited by Fawcett & Rendle (FI.

1943 | KOBUSKI, STUDIES IN THE THEACEAE, XIV 73

Jam.) under T. granulata. Prior s.n. was determined questionably by
Urban (1908) as 7. granulata. Ternstroemia granulata can be separated
from the present species by the dense granular punctations covering both
leaf-surfaces, the highly revolute leaf-margin, and the copious (12 or 13
pairs) and clearly visible veins. The flowers are larger, especially as
regards the corolla (12-13 mm.).

22. Ternstroemia peduncularis A. de Candolle in Mém. Soc. Phys. Hist. Nat.
9

Ternstroemia meridionalis Swartz Prodi: 81. 1788. — Non Mut
sient Lai obovalis A. Richard in Sagra, Hist. Phys. Pol. Nat. Cuba, 221, t. 25.
—Walpers, Rep. 5: 130. 1845.— Grisebach, Cat. Pl. Cuba, 36. 1866.—

oe Fl. Cub. 10. 1873.— Krug & Urban in Bot. Jahrb. 21: 524. 1896. —
bags n, FI. Ind. se bai 436. 1920.— Melchior in Nat. Pflanzenfam. ed. 2, 21:
142. 1925. — Syn.

Fe Faust» var. excelsa Grisebach, Cat. Pl. Cuba, 36. 18

Ternstroemia obovalis var. ovulosa Wright in Grisebach, Cat. Pl. oy 36. 1866.

Mokofua obovalis (A. Rich.) O. Kuntze, Rev. ie Pl. 1: 63. 1891.

Mokofua peduncularis (DC.) O. Kuntze, loc

sdnaiciey ae il DC. var. aes Krug & Urban in Bot. Jahrb. 21:
526. 1896. — Syn. no

Ternstroemia obovalis var. a. genuina Krug & Urban, loc. cit. 21: 524. 1896. —

yn. nov.

Ternstroemia obovalis var. 8. Lindenii Krug & Urban, loc. cit. 21: 524. 1896. —
Syn. nov.

Ternstroemia obovalis var. Y. minor Krug & Urban, loc. cit. 21: 525. 1896.—
Syn, nov

Taonabo Leonis Britton & Wilson in Bull. Torrey ve ve 50: 42. 1923.

Taonabo monticola Britton & Wilson, loc. cit. 50:

Ternstroemia apleura Krug & Urban in Bot. Ja sue — 523, 1896. — Melchior in
Nat. Pflanzenfam. ed. 2, 21: 142. 1925.—

Syn
Ternstroemia potrerillensis O. C. Schmidt in e are Nov. 22: 94. 1925.—

yn. nov.
Ternstroemia monticola (Britton & Wilson) Ekman ex Schmidt in Rep. Spec. Nov.
2:95. 1925.—Syn. nov
Ternstroemia Ekmani Schinidt in Rep. Spec. Nov. 22: — Syn.n
Ternstroemia Leonis (Britton & Wilson) Ekman ex aca in Rep. tet Nee: 22:

; — . nov.
Ternstroemia rupicola Ekman ex Schmidt in Rep. Spec. Nov. 22:97, 1925.—
nov

DistripuTion: Cuba, Haiti, Santo Domingo, Porto Rico, St. Jan, St. Eustatius,
aaa? Martinique

CuBA: Prov sc csanta: ‘Cla Bro. Leon 782 (NY), 4642 (NY).— Bros.
Leon o Gate 5291 (NY). mas ae M. Roco 8006, 8040, 8129 (NY). — Bros.
Leon & Clement 6700 (NY).— Bro. Leon & - Acuna 13040 (NY).—E. L. Ekman
14016 (1soTyPE of T. ere NY).—J. T. Roig & M. Cremata 2194, 2220
(US).— A. Luna 552 (NY).—N. L. sie & J. F. Cowell 10263 (NY).—
Smith & A. R. Hodgdon 3205 (G). Prov. 4 nie Pig E. L. Ekman
16388 (1sotyPE of T. Ekmani, NY).—N. L. Bri F. S. e & C. S. Gager 6909
(NY). rov. Oriente: Bro. Leon 10837 oe. Pee bee of T. monticola,

74 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

NY), 10910 (1sotyPE of T. Leonis, NY), 11076 (NY), eer (NY).—C. Wright 1125
(isotypes of T. obovalis var. minor, FM, G, Mo, NY), 21/11 (1sorypes of T. apleura,
G, Mo, NY), 2114 (G, Mo, US). —J. T. Roig & G. C. Bucher 6689 (NY).—G. C.
Bucher 88 (NY). Prov. Camaguey: WN. L. Britton & J. F. Cowell 13277

,G, NY, US). Isle of Pines: A. H. Curtiss 429 (FM, G, Mo, NY, US). —
hee pe 5133 (NY).—N. L. Britton & P. Wilson 15658 (FM, G, NY, US). Hartt:
E. C. Leonard 7146 (NY, on 8849 (G, NY, US).—£. C. & G. M. Leonard 13379
(NY, US), 15825 (G, US). .L. Ekman H-1903 (US), H-8330 (US).—G.V. Nash
816 (NY).—G.V. Nash & } Po Taylor 1311 (FM, NY, US).—O. F. Cook, C. S. Sco-
field & C. B. Doyle 92 (US).—G. 8S. Miller 304 (US). Santo Dominco: E. L. Ekman
H-13043 (US).—M. adhe 1003 (NY), 1294 (FM, G, Mo).— Eggers 2411 (US),
2519 (NY, US).—W. L. Abbott 2226 (US).—H. von Tiirckheim 3401 (FM, G, Mo,
NY, US).—E. J. SA 114 (US), 667 (FM, Mo, NY, US).—C. Wright, C. C. Parry
& H, Brummel 171 (US). Porto Rico: P. Binsands 948 (FM, G, Mo, NY, US).—
N.L. Britton & J. A. Shafer 1565 (FM, Mo, NY, US).—H. A. Gleason & M.T. Cook
R-59 (NY). St. Jan: WN. L. Britton & vie 4. Shafer 532 (FM, NY, US). St.
Eustatius: J. Boldingh 323 (NY). GuapELoUPE: Pére Duss 3651 (FM, NY, US).
MARTINIQUE: Pére Duss 1825 (FM, Mo, NY, US)

After long consideration, I find it expedient to combine under the name
T. peduncularis DC. several formerly recognized species. The best known
of these is JT. obovalis Rich. Urban (Bot. Jahrb. 21: 524-526. 1896)
separates these two species on the granular-punctate surface of the leaves,
the degree of prominence of the lateral veins, the plane or recurved margin
of the leaves, the shape of the bracteoles, and the number of ovules in the
ovary. None of these characters or any others listed by Urban consistently
apply to either of the above mentioned entities. In the Haitian material,
often on the same flower, bracteoles may be found, one of which is large
and triangular-ovate, 4 mm. or more wide (far surpassing the pedicel in
width), while the other is linear-acuminate, scarcely 1 mm. wide. In other
specimens the larger bracteole may be subrotund. In the Santo Domingan
material nearly all specimens have flowers with bracteoles equal or sub-
equal, triangular, ca. 1.5 % 1.5 mm., while in the Cuban material the
bracteoles vary from linear-lanceolate to subrotund.

The pedicel may vary from 2 to 8 cm. in length. Specimens with pedi-
cels 8 cm. long certainly appear vastly different from specimens whose
pedicels measure only 2-3 cm. Yet the gradation through a large series of
specimens shows this character insufficient for specific delimitation.

Variation in veining, coupled with deciduous bracteoles, has been used
as the basis of specific differentiation in 7. apleura. Deciduous bracteoles
would seem to be an excellent specific character. Still, on the isotype of
T. apleura (C. Wright 2111) in the Missouri Botanical Garden all the
bracteoles are present and there is also evidence of venation. In T
peduncularis itself, veining is often obscure, even to the extent found by
Urban in T. apleura. On the other hand, some specimens have veins so

glandular denticulation on the bracteoles and outer calyx-lobes can be
found on a single specimen. Often the age of a specimen may cause this

1943]. KOBUSKI, STUDIES IN THE THEACEAE, XIV 75

variation. On fruiting calyx-lobes, the glandular denticulations of the
slightly scarious margin are often worn off and the margins appear revolute
or fimbriate.

The species T. apleura Krug & Urban, T. Ekmani O. C. Schmidt, T.
Leonis (Britton & Wilson) Ekman, 7. monticola (Britton & Wilson)
Ekman, T. potrerillensis O. C. Schmidt, and T. rupicola O. C. Schmidt are
separated on characters scarcely sufficient for specific delimitation. All
these entities have obovate or obovate-elliptic leaves, rounded, obtuse, or
obtusely acuminate at the apex, and tapering at the base into a petiole
which averages 4-5 mm. long. The leaves of all are coriaceous or occasion-
ally chartaceous-coriaceous, varying in thickness. The margin is usually
entire, with occasional signs of serration, and flat or somewhat revolute.
The pedicels in all the above listed entities (2-3 cm. long) vary little.
The bracteoles (2-3 mm. long) likewise seldom vary more than a single
millimeter either way. They are all triangular or deltoid and occasionally
may be considered semi-ovate (7. rupicola). The calyx-lobes are
always subequal and range from 4 to 7 mm. long. The outer calyx-lobes are
always glandular-denticulate and the inner lobes eglandular. Tern-
stroemia apleura, T. Leonis, T. monticola, and T. rupicola were proposed
without any description of corolla and stamens. In most cases, the corolla
is approximately 6-8 mm. long. Urban (Bot. Jahrb. 21: 521-551. 1896)
based his key, to a considerable degree, on the coalescence of the petals at
the base. The petals measure 6—8 mm. in length, and differentiation by
means of the varying amount of coalescence of petals (as follows: 1/6,
1/4, 1/3, 2/5 or 1/2), is drawing too fine a line, considering the amount
of known variation in the group and also the fact that in the majority of
specimens the petals are lacking. The ovary and fruit characters show
similar lines of variation.

DUBIOUS OR LITTLE KNOWN SPECIES

Ternstroemia buxifolia Ekman & Schmidt in Rep. Spec. Nov. 29: 13. 1931.

DistripuTion: Santo Domingo. No specimens studied.

According to the authors, this species is characterized by small coria-
ceous leaves (1.8-3.0 * 0.9-1.6 cm.), elliptic to obovate, rounded and
lightly emarginate at the apex, narrowed at the base into a petiole 3-5 mm.
long, with the margin recurved and the veins inconspicuous. The pedicel
is short, 4-6 mm. long, and the bracteoles are semi-elliptic, + 2 mm. long,
membranaceous, fimbriate-serrulate. The sepals are unequal, the margin
membranaceous-fimbriate, + 5 mm. long and 4 mm. wide. The petals
are obovate or subobovate-cordate, 5-6 mm. long, and up to 5.5 mm. wide,
with the margin membranaceous and fimbriate. The stamens number
about 25. The ovary is semi-ovate or subconical, ca. 1.5 mm. long, uni-
locular, 6-ovulate, attenuated at the apex into a style about 3.2 mm. long.

The authors of this species suggest a close relationship with the Cuban
T. parviflora Krug & Urban, because of the single-celled ovary.

76 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

EXCLUDED SPECIES
— albo-punctata Grisebach, Cat. Pl. Cuba, 36. 1866 oo albo-
punctata (Grisebach) Krug & Urban in Bot. Jahrb. 21: 537.

Ternstroemia dentata Sprengel ex DeCandolle in Mém. Soc. Phys. Hist. Nat. Geneve,
1: 411 — Ternstr. 19). 1822 = Freziera undulata paces Swartz, FI.
1800.

Ind. Occ. 2:
Ternstroemia nll DeCandolle in Mém. Soc. Phys Nat. Genéve, 1: 411
i. 19). 1822 = Freziera adulats ees Swartz, Fl. Ind. Occ.
2: 974,
Ternstroemia crenata Macfadyen, Fl. Jam. 1: 114. 1837 = Symplocos octopetala

Swartz, Prodr. Veg. Ind. Occ. 110. 1788

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

1943 ] CHASE, PAPUAN GRASSES, III 77

PAPUAN GRASSES COLLECTED BY L. J. BRASS, III*
AGNES CHASE

With four text-figures

THE grasses here enumerated were collected in 1938-39, in Papua and
in Netherlands New Guinea. Part I of Papuan Grasses collected by Mr.
Brass, by A. S. Hitchcock, was published in Brittonia 2: 107-130. 1936,
Part II, by Agnes Chase, in the Journal of the Arnold Arboretum 20:
304-316. 1939. The types of the species here described are deposited in the
Gray Herbarium, with isotypes in the United States National Herbarium.
Species previously collected by Mr. Brass in Papua are not included in the
present paper, except when collected in Netherlands New Guinea. A few
collections by J. and M.S. Clemens or by Mrs. Clemens are included when
they represent species not before collected in New Guinea.

Oreiostachys producta Pilger, Bot. Jahrb. Engler 62: 460. 1929.

NETHERLANDS NEW GuINEA: Bele River, 18 kilometers northeast of Lake Habbema,
alt. 2300 m., Brass 11072; abundant in oe margins, downfall openings, etc., scram-
bling and tangled; flowering specimen. Eighteen kilometers southwest of Bernhard
Camp, Idenburg River, alt. 2150 m., Brass 12662; scrambling to 5-6 m. in mossy
forest, frequent at 1800 m., and on the highest point of the ridge at 2200 m., forming
a high dense undergrowth, practically excluding the usual undergrowth and substage
trees; sterile specim

These ileceons agree well throughout with Pilger’s description. The
palea is 2-keeled and sulcate toward the summit, the rachilla segment is
prolonged and bears a rudimentary floret, the rachilla and rudiment reach-
ing the apex of the palea. Henrard (Blumea 2: 71. 1936), in reestablish-
ing the genus Chloothamnus Buse, reducing Oreiostachys Gamble to it as a
synonym, states that because of the prolonged rachilla segment and rudi-
mentary floret he hesitated to place O. producta in the genus Chloothamnus.
Endemic
Chloothamnus sp.

NETHERLANDS New Guinea: Six kilometers southwest of Bernhard Camp, Idenburg
River, alt. 1450 m., Brass 13020; ‘“‘upper limits of rain-forests; occasional open clumps
of few stout stems up to + 4 cm. diam. at base, 7-8 m. long, upper part weak, resting
on ee trees; internodes up to + 45 cm. long; leaves glaucous below; spikelets
glauc

T he specimen, which consists of an internode and two nodes of a stout
culm and four flowering twigs, agrees with the generic description of
Chioothamnus but differs from any known species in the open panicles, the
stiff branches having pronounced pulvini at their base.

An over mature specimen collected in “Lower regions of British New
Guinea” in 1894 by MacGregor, no. 49, is apparently the same species.

*Botanical Results of the third Archbold Expedition.

78 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Dendrocalamus latifolius a & K. Schum. in K. Schum, & Lauterb, Fl. Deutsch.
Schutzgeb. Siidsee 188. 1901.

NETHERLANDS NEW any Hollandia and vicinity, alt. 10 m., Brass 8894; shrub
about 1 m. tall, plentiful on dry gravel beds in river.

The specimen consists of a single branch of inflorescence which agrees
with Brass 7908, 8518 and 8660 from Papua, excellent specimens of D.
latifolius, but the detached leafy shoot of Brass 8894 has blades only
15—20 mm. wide. Endemic.

Dendrocalamus sp.

NETHERLANDS New GuINEA: Four to six kilometers southwest of Bernhard Camp,
Idenburg River, alt. 850-1200 m., Brass 12825, 13070, 13327; rain-forest, in flood plain
and characteristic in rather open rain-forest, scrambling 2-20 m
Schizostachyum lima (Blanco) Merr. pena ata Bot. 3: 62. 1916.

Bambusa lima Blanco, FI. Filip. ed. 1. 271.

Papua: Middle Fly River, Lake blerat — (listed as Schizostachyum
sp. by Chase, Jour. Arnold Arb. 20: 305. 1939). Philippin
?Schizostachyum sp.

NETHERLANDS New GuINEA: Four kilometers southwest to Bernhard Camp, Iden-
burg River, alt. 900 m., Brass 13743; — o 4-6 m. in mossy forest; clumps of
several stems about 2 cm. diameter at bas

Sterile specimen, the culm thin-walled and roughish as in Schizostachyum,
the overlapping slightly carinate sheaths with conspicuous slender spreading
fragile fimbriae from the shoulder, the fimbriae arising from thick tubercu-
late bases, and with a row of stiff hairs back of the very short ligule;
blades somewhat plicate.

Bromus scopulorum sp. nov.

Perennis; culmi 50-125 cm. alti, tenues; vaginae sparse retrorse-pilosae ;
ligula minuta: laminae planae, 15-20 cm. longae, 2-4 mm. latae; panicula
subsimplex, 12-18 cm. onga, angusta, laxa, ramis tenuibus ‘previbus:
spiculae 2-3 cm.. longae, 5—7-florae; glumae 7-9.5 mm. longae, 1—3- vel
5-nervia, ad marginem sparse pilosae; lemmata 10-13 mm. longa, 5-nervia,
ad marginem dense hirsuta, arista 5-7 mm. longa.

Perennial, loosely tufted with lax-leaved innovations, purplish at base;
culms ascending, 50-125 cm. tall, slender, glabrous; sheaths shorter than
the internodes, thin, the lower sparsely to rather densely’ retrorsely pilose,
becoming shredded, the upper loosely pilose; ligule minute; blades flat,
thin, lax, 15-20 cm. long, 2-4 mm. wide, delicately pilose toward the base,
acuminate; ts purplish brown, nearly simple, 12—18 cm. long, narrow,
lax, nodding, the short slender branches scaberulous; spikelets 2—3 c
long, loosely 5—7-flowered, the slender rachilla segments 2.5—3 mm. lon
pubescent; glumes 7 and 9. 5m mm. long, narrow, acuminate, the first 1- aan
(short lateral nerves developed in some), sparsely pilose along the margin
near the base, the second 3—5-nerved, pilose along the margin nearly to the
summit; lemmas 10-13 mm. long, excluding the awn, 7-nerved, hirsute on
the margin to the summit and at the very base, the back otherwise glabrous,
the nerves scabrous toward the summit, the awn erect, slender, scaberulous,
5—7 mm. lon

s & Myer-Drees 9825, collected in September 1938, on dry ledges of a
limestone cliff, 7 kilometers northeast of top of Mt. Wilhelmina, alt. 3560 m. Nether-
lands New Guinea

1943 | CHASE, PAPUAN GRASSES, III 79

Bromus sp.

NETHERLANDS NEW GUINEA: Lake Habbema, alt. 3225 m., Brass 9127; common in
forest glades and other sheltered places.

Tall slender perennial with nodding panicle.

Brachypodium pubifolium Hitchc. Brittonia 2: 108. 1936.

NETHERLANDS NEw GuINEA: Lake Habbema, 3225 m. camp, Brass 9310; forest
poe Bele River, 18 kilometers, sigs of Lake Habbema, 2200 m. camp, Brass

1426; on wet marshy aie in river. . Wilhelmina, 7-11 kilometers northeast of
tn, alt. 3400-3560 m., Brass & Myer- eas 9714, 9827; rather wet grassy valley and

dry ledge under Bee iane cliff. Endemic.

Festuca nubigena Jungh. Nat. Geneesk. Arch. Neerl. Ind. 2: 51. 1845.

NETHERLANDS NEW GUINEA: Lake Habbema, alt. 3225 m., Brass 9325, 9547; ee
banks of streamlet and scattered over wet peaty flat; a slender tufts 40-80 cm. high.
Mt. Wilhelmina, 11 kilometers northeast of top, alt. 3

8
=
aR
ao
nn
be
s
ay
S
x
sd
®
an
6
ai
#®
N

alpine rane covering marshy hollows, 70-90 cm. high. Same, northern slopes, alt.

050 m., Brass & Myer-Drees 10066; wet shallow soil of old screes, 30 cm. high, Java,
New Gusts,

Festuca nubila Jungh. ex Steud. Syn. Pl. Glum. 1: 315. 1854.

NETHERLANDS NEw GUINEA: Lake Habbema, 3225 m. camp, Brass 9128; common
in moist forest glades. Mt. Wilhelmina, 4-11 kilometers northeast of top, alt. 3400-
3560 m., Brass & Myer-Drees 9715, 9824, 9845; wet grassy valley and forest glades,
tufts up to 1 m. high, ari rena purple. Same, northern slopes, alt. 4050 m., Brass
& wrk Drees 10061, 10071; very common on upper slopes, about 80 cm. high. Same,
alt. 4050 m., Brass & Myer- Drees 10070; common in grass cover of old screes, panicle
greenish purple. Java, New Guinea.

This species is closely related to F. nubigena, but has taller culms, much
longer leaves, — lax panicles and larger spikelets with hispidulous
glumes and lemm
Monostachya oreoboloides (F. Muell.) Hitchc. pone 2: 107. 1936.

Festuca oreoboloides F. Muell. Trans. Roy. Soc. Vict. 1(2): 38. 1889.

NETHERLANDS NEw GuINEA: Lake Habbema, ere m. camp, Brass 9184. Mt.
Wilhelmina, 7-11 kilometers northeast of summit, alt. 3400-3900 m., Brass & Myer-
Drees 9818, 9941; large bright green cushions among taller grasses of open summit.
Same, sae slope, alt. 4100 m., Brass & Myer-Drees 10159. Philippines (Luzon),
New Guin
Poa eis Pilger, Bot. Jahrb. Engler 62: 459. 1929.

NETHERLANDS New GuINEA: Mt. oe 7 kilometers northeast of top, alt.
3560 m., Brass & Myer-Drees 9844, End
Poa egregia sp. nov. eel

Perennis; culmi 18— a cm. alti; vaginae peer ligula 6-8 mm
longa; laminae 10-16 cm. longae, involutae vel planae, 2-3 mm. latae,
subpungentes; panicula laxa, nutans, 9-12 cm. longa, axi ramisque sub-
capillaribus; pedicellis 2—3 mm. longis; spiculae 6-7 mm. longae, 2-florae

ro-

natae; lemmata 5.5-6.5 mm longa, lata, scaberrima, apice dentata vel

minute lobata, mucronata, basi sparse lanata; palea lemmate brevior,
carinis excurrentibus

Tufted perennial; “culms 18-40 cm. tall, scabrous below the panicle,

80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV
leafy at the base; sheaths overlapping, strongly nerved, keeled, smooth,
those toward the base crowded, thin, papery, loose, pale, final ly somewhat

base of the panicle;

shredded, the upper eee reaching nearly to the
ligule 6— 8 mm, long, hyaline, lacerate, often split; blades 10-16 cm. long,

the lower sometimes loosely involute, the others mostly flat, 2-3 mm. wide,
scaberulous on the upper surface, smooth beneath, the apex acuminately
boat-shaped, tipped with a fine stiff point; panicles lax drooping, 9 to 12 cm.
long, the axis and branches subcapillary, flexuous, scabrous, the branches
mostly in pairs, the pairs rather distant, the ultimate branchlets and
pedicels angled, scabrous, 2-3 mm. long, the spikelets a ee ae spike-
lets 6-7 mm. long, 2- flowered (rarely 3-flowered), broadly ovate to
V-shaped; glumes equal, 5.5—6 mm. long, very scabrous, pie apex dentate

SSS

——
><

—
_~

by y
Re on é
aya y
' ¥, y , if
Oh a a te
ae ‘y
COP Mg te J 34
ty i a Py, £4

“LF as

Sse

as

a

>
=

Fic. 1. Poa egregia; spikelet, palea, and floret, x 10; from Brass & Myer-Drees 10206

or minutely lobed, bees ane excurrent as a short mucro; rachilla seg-
; lemmas 5.5-6.5 m

m. long, broad, very scabrous,

ments slender, smoo
strongly nerved, the sa ed. lobed or aide the midnerve excurrent
as a mucro, the base with sparse white cottony hairs 1-2 mm. long; palea
shorter than the lemma, the scabrous keels SS as short mucros, the
internerve toothed at apex, scaberulous on the
Tyre: Brass & Myer-Drees 10206, collected in ea 1938, in “alpine grass-
land; gregarious under shelter of rocks, 4150 m. alt., northern slope of Mt. Wilhelmina,”
Netherlands New Guinea.
There are two other collections of this species from the same ee
Brass & Myer-Drees 10040, 10068, altitudes 4050 m. and 4100 m
oe ee plentiful in wet grassy spots and under rocks on tussock- ane
In no. 10068 a piece of the buried culm with old shredded sheaths
. Nearly

slopes.
is attached at base, giving the appearance of a rhizomatous plant

1943] CHASE, PAPUAN GRASSES, III 81

mature caryopses were found with anthers about 1 mm. long, crushed with
the stigmas at the summit, indicating closed fertilization.

Poa lunata sp. nov. Fic. 2.
ae eres culmi 25-35 cm. alti, erecti; vaginae imbricatae;
ligula 2—4 onga; laminae 8—15 cm. longae, conduplicatae vel involutae:

panicula deuecls: erecta, 6-8 cm. longa, circa 1-1.5 cm. lata, axi
ramisque scabris, pedicellis 3 mm. longis; spiculae 5 mm. longae, circa
1.6 mm. latae, 2-florae; glumae 3.5 et 4.3 mm. longae, sublunatae, carina
ad apicem scaberula; lemmata 3.5—-4 mm. longa, sublunata, acuta, glabra;
palea lemma aequans, hyalina.

Fic. 2 (left). Poa lunata; spikelet, palea, and floret, x 10; from Brass & Myer-Drees
10067.
Fic. 3 (right). Poa multinodis; spikelet, palea, and floret, x 10; from Brass 9584.

Tufted perennial; culms 25-35 cm. tall, erect, scaberulous below the
s

mm. long, hyaline, acute; blades ascending, 8-15 cm. long (uppermost
shorter), folded or loosely involute, slender, smooth, the pis acuminately
boat- shaped: panicle erect, narrow, rather densely flowered, the axis,

wide, elliptic in outline, 2-flowered; glumes 3.5 and 4.3 mm. long, ‘almost
lunate, the keel scaberulous toward the apex; rachilla segments slender,
smooth; lemmas 3.5—4 mm. long, somewhat lunate, especially at maturity,
acute, scaberulous on the keel toward the apex, otherwise glabrous; palea
as long as the lemma, hyaline, the keels scabrous, terminal rachilla segment
nearly as long as the palea.

Tyre: Brass & Myer-Drees 10067, collected in September 1938; ‘common i
small clumps under of old screes, 3950 m. alt., northern slopes of Mt. Wilhelmina,”
Netherlands New Guin
Poa multinodis sp. nov. Bice 3:

Perennis, e rhizomatibus tenuibus; culmi graciles, basi decumbentes,

)

multinodes, vaginae inferiores imbricatae; ligula 1 mm. longa, decurrens;

82 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

laminae planae vel conduplicatae, 2.5-8 cm. longae, 1-1.6 mm. latae;
panicula erecta, 6—7 cm. longa, patens, ramis 5-25 mm. longis, pedicellis
0.5-1 mm. longis: spiculae 3 mm. longae, circa 1.3 mm. latae, 2-florae;
glumae 3 et 3.7 mm. longae; lemmata 2.1—2.4 mm. longa, latiuscula, firma,
acuta; palea lemma aequans, membranacea.

Perennial with slender rhizomes; culms slender, decumbent at base, the
lower part with numerous nodes 5-15 mm. apart, the upper 3 nodes dis-
tant, scaberulous below the panicle; lower sheaths overlapping, si
smooth, the upper about as long as the internodes or shorter; ligule 1
ong, hyaline, decurrent; blades lax, flat or folded, 2.5-9 cm. long, 1- nL U7

m. wide, minutely scaberulous toward the apex; panicle erect, 6—8 cm.
ane open, the axis aa branches filiform, scabrous, the branches 5-25 mm.
long, finally stiffly spreading or reflexed, the pedicels 0.5-1 mm. long
spreading; spikelets 3 mm. long, about 1.3 mm. wide, leaden-purplish at
maturity, 2-flowered; glumes 3 and 3.7 mm. long, the keels convex,
scaberulous toward the apex; rachilla segments slender, smooth; lemmas

m. long, rather broad, firm, acute, scaberulous on the keels toward
the apex, otherwise glabrous, the intermediate nerves often obscure; palea
as long as the lemma, thin-membranaceous, the internerve sulcate toward
the summit, the keels scaberulous.

Type: Brass 9584, collected in eng 1938, “in wet forest glade, Lake Habbema,
3225 m. camp,” Netherlands New Gui
Poa erectifolia Hitchc. Brittonia oe 111. 1936.

NETHERLANDS NEw Guinea: Mt. Wilhelmina, 7 kilometers northeast of top, alt.
3720 m., Brass & Myer-Drees 9945, often the dominant grass on n rather wet open sum-
mits, siact in very stiff small clumps, 30-40 cm. high. emic.

This specimen differs from Brass 4326 (the type) and Brass 4420 in
having glabrous, coarser, longer blades (8-15 cm. long) and a panicle with
longer branches. The decurrent ligule and the spikelets agree with those
of the earlier collections.

Poa crassicaulis Pilger, Bot. Jahrb. Engler 62: 458, 1929.

NETHERLANDS NEW GUINEA: Northern slopes of Mt. Wilhelmina, alt. 3900 m
Brass & Mvyer-Drees 10079; gregarious on edges of earth screes. Lake Habbema,
3225 m. camp, Brass 9338; alpine grassland, several plants on sandy bed of stream.
Endemic.

Brass & Mvyer-Drees 10079 consists of dwarf plants like the type,
Keysser 4 from Sarawaket Mountains, Northeastern New Guinea, and
Brass 4469, previously reported from Papua. Brass 9338, from a lower
altitude, is a well developed plant, the lower sheaths closely imbricate as
in the dwarf plants but the blades 5—7 cm. long. The culms are com-
pressed, 12 and 14 cm. long, and the panicles are narrow, rather dense,
and 8 cm. long.

Poa pilata sp. nov

Perennis, narcosis, 3-4 cm. alta; folia numerosissima; vaginae im-
bricatae; ligula nulla; laminae squarrosae, conduplicatae, 10-20 mm.
eure (plicatae) 0.5- 0. 8 mm. latae; panicula parva; een —4, eee

3.5 mm. longae, circa 1.2 mm. latae, 2—3-florae; glumae 1.6 et
eae latae, 3-nerviae, acutae; lemmata acuta, 2—2.3 mm. longa, nervis
intermediis inconspicuis; palea ‘lemmaque aequalia: segmentum superius
rachillae productum.

1943 ] CHASE, PAPUAN GRASSES, III 83

Perennial in dense tufts, branching at base, glabrous as a whole; culms
3-4 cm. high above the surface of the bog, with old buried culms 1—4 cm
long extending downward from the base; leaves very numerous, about 18—20
to a centimeter; sheaths closely imbricate, thin below and adhering to
the culm; ligule obsolete; blades firm, squarrose, conduplicate, 10-20 mm.
long, 0.5—0.8 mm. wide as folded, the midnerve thick, the apex acute, boat-
shaped; panicle not more than 10 mm. long, of 1 to 4 erect spikelets, the
rather stiff peduncle exserted 3-10 mm., the peduncle, axis, and pedicels
compressed, scabrous; spikelets 3.5 mm. long, about 1.2 mm. wide,
2—3-flowered, glabrous, the rachilla segments rather thick; glumes 1.6 and
2 mm. long, firm, broad, 3-nerved, acute, the second somewhat boat-shaped
at apex; lemmas acute, the intermediate nerves obscure, the lower lemma
2.3 mm. long, the upper about 2 mm.; palea about as long as the lemma;
upper rachilla segment prolonged back of the palea

Type: Brass & Myer-Drees 10153, collected in September 1938, in “alpine grass-
land, abundant on boggy ground, alt. 4100 m., northern slopes of Mt. Wilhelmina,”
Netherlands New Guinea. Also collected at Lake Habbema, 3225 m. camp, Brass 9554,
9580, very abundant and one of the characteristic plants of open alpine peat bogs. In
these two collections the spikelets are defective, either partly eaten away or affected
by fungus. Brass 9942, ‘tufted on barren seepage-wet slopes of sandstone, alt. 3850 m.,
7 kilometers northeast of the Wilhelmina top,” consisting of tufts 1.5-3 cm. high, the
lemmas all fallen, and Brass & Myer-Drees 10205, ‘forming loose clumps under rocks
on alpine grassland, common but mostly sterile, alt. 4150 m., northern slope of Mt.
Wilhelmina,” with long under-surface culms clothed with remnants of sheaths, and
with blades 2-4 cm. long, are doubtfully referred here.

This species is apparently related to Poa crassicaulis Pilger, but is much
more delicate, the blades much finer, acute, the ligule obsolete.

Poa spp.

NETHERLANDS NEw Guinea: Mt. Wilhelmina, northeast of top, alt. 3720 m., Brass
& Myer-Drees 9929, scattered on bare peaty ground on edges of dying forest; alt.
3850 m., Brass & Myer-Drees 10347, on limestone near waterfall.

Both are tufted perennials, with small open panicles with 2-flowered
purple spikelets; the lemmas of no. 9929 are glabrous, those of no. 10347
with very scanty cottony hairs at base.

Deschampsia Klossii Ridley, Kew Bull. 1913: 268. 1913.
NETHERLANDS New Gutnea: Lake Habbema, 3225 m. camp, Brass 9048, 9312;

Habbema, 2800 m. camp, ida 10555, open beds of streams and on landslips. Mt.

Trisetum flavescens (L.) Beauv. var. papillosum Hack. Bull. Herb. Boiss. 7: 702.
1899

NETHERLANDS NEw GuINEA: Lake Habbema, 3225 m. camp, Brass 9118; a few
erect tufts in an old native camp. Japan.

Differs from the species in the strongly papillose-scabrous lemmas.
Agrostis Reinwardtii Van Hall; Miquel, Fl. Ind. Bat. 3: 750. 1855.

NETHERLANDS NEw GuInEA: Lake Habbema, alt. 3225 m., Brass 9050, 9576; com-

84 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

mon tussock grass in alpine grassland. Mt. Wilhelmina, alt. 3400-4080 m., Brass &
Myer-Drees 9674, 9717, 9864, 9928, 10064, 10069, wet valleys and forest glades on
north and northeast slopes of mountain. Jav
— avenacea Gmel. Syst. Nat. 2: 171
Avena nig G. Forst. Fl. bee pene ie 9. 1786. Not Agrostis filiformis
Vill. Hist. Pl. Dauph. 2: 78.

Agrostis sera Willd. ets ff 1: 94. 1809.

NETHERLANDS New Guinea: Mt. Wilhelmina, 2—7 kilometers pr ronpaiar of top, alt.
3650-3700 m., Brass & Myer-Drees 9821, 9969, 10116; along n path and under
native rock shelters. Nine kilometers northeast of Lake ht daa ere 2800 m., Brass
10726; native clearing. Australia, Polynesia; introduced in America
Aulacolepis epileuca (Stapf) Hitchc. Brittonia 2: 117. 1936.

Deyeuxia epileuca Stapf, Trans. Linn. Soc. Bot. II. 4: 7 1894,

Poa epileuca Stapf, Hook. Ic. Pl. 27: pl. 2607, in she, 1899.

NETHERLANDS New GuvutInEA: Lake Habbema, alt. 3225 m., Brass 9125, 9581;
grassland and forest glades. Mt. Wilhelmina, alt. 3400-3900 m., Brass & Myer-Drees
9816, 10074; alpine grassland, 11 kilometers northeast of top. Borneo, New Guinea.
Garnotia Mezii Janowski in Mez, Repert. Sp. Nov. Fedde 17: 86. 1921; 18: 27. 1922.

NETHERLANDS New Guinea: Four kilometers southwest of Bernhard Camp, iden
burg River, alt. 850 m., Brass 13216; rain-forest, common on banks of river. PAPuA
Central Division: Diene, Ononge Road, alt. 50 pay Brass 3818 (listed as G. siete
Brongn. by Hitchcock, Brittonia 2: 118. 1936). Endem
Dichelachne novoguineensis (Pilger) Pilger, Bot. Jat Engler zai 1938.

biel novoguineensis Pilger, Bot. Jahrb. Engler 62: 457.

Pa ral Division: Mt. Albert Edward, alt. 3680 m., tae gr Bg ae
listed by eee Brittonia 2: 117. 1936 as D. sciurea (R. Br.)

LANDS New Gutnea: Nine kilometers northeast of Lake Habbema, a se
Brass 10742; native clearing in forest. Mt. Wilhelmina, alt. 3400 m., eee & Myer.
Drees 9802; sandy banks of stream 11 kilometers northeast of top. Endem

Aristida Cumingiana Trin. & Rupr. Mém. Acad. St. Pétersb. VI. Sci. Nat. 5!: 141
1842.

NETHERLANDS New Guinea: Balim River, alt. 1600 m., Brass 11741; common
along paths on sandy long-deforested slopes. Southweateen Asia, Philippines, New
inea.

Cynodon Barberi Rang. & Tad. Jour. Bombay Nat. Hist. Soc. 24: 846. 1916.
NorTHEASTERN New Gurtnea: Morobe District: Kajabit Mission, Clemens 10473
bis. India; new to New Guinea.

Hierochloé redolens (Vahl) Roem. & er Syst. Veg. 2: 514. 1817.

Holcus redolens Vahl, Symb. Bot. 2:

NETHERLANDS NEW GUINEA: Lake nn 3225 m. camp, Brass 9461; abundant
among tussock grasses in a limestone sinkhole. Mt. Wilhelmina, 7-11 kilometers north-
east of top, alt. 3400-3560 m., Brass & Myer-Drees 9712, 9830; alpine grassland, com-
mon among tussocks in moist situations. Northern slopes of Mt. Wilhelmina, alt.
4000 m., Brass & Myer-Drees 10062; common in wet situations,

Hierochloé Horsfieldii (Kunth) Maxim. Bull. Acad. Sci. St. Petersb. 32: 627. 1888.
Ataxia Horsfieldii Kunth, Rev. bre : 22. 1829, name only; Enum. Pl. 1: 39

8 me only; Kunth ex Horsfield in eres Pl. Jav. 8: pl. 3. 1838-1852.
Miscookla? angusta Hitche. Bitoni : 118. 1936.
NETHERLANDS NEW GUINEA: ke otha 4 3225 m. camp, Brass 9049, 9117,

9577; forest margins, glades and rr shore. NORTHEASTERN NEw GUINEA: Sara
waket, Clemens 7261; bank of rivulet, alpine meadow. Java.

1943] CHASE, PAPUAN GRASSES, III 85

The type of Hierochloé angusta, Brass 4412, Mt. Albert Edward, Papua,
is a slender specimen with glabrous foliage, narrow blades, and narrow
panicle only 6—7 cm. long. The plate of Ataxia Horsfieldii shows a much
larger plant with wider blades and looser panicle. The later New Guinea
collections and additional specimens from Java show that the species is
variable, ranging from narrow-leaved plants like Brass 4412, 9577, and
Clemens 7261, to tall, broad-leaved plants like Brass 9049 and most of the
Java specimens. The panicles range from small and almost spikelike, as
in Brass 4412, Clemens 7261, and two Java specimens, to as much as
20 cm. long with branches to 8 cm. long, as in Brass 9049. The foliage in
the type of H. angusta is glabrous and that of Ataxia Horsfieldii is de-
scribed as glabrous, but some of the specimens from both Java and New
Guinea are retrorsely pubescent on the sheaths and spreading-pubescent
on the upper surface of the blades. This is the type species of the Section
Ataxia (R. Br.) Hack., in which the lower floret is staminate or neuter and
the second neuter and usually without a palea. In Hierochloé proper
the lower florets are staminate.

Arundinella furva sp. nov. Fic. 4.

Perennis; culmi ascendentes vel erecti, 75-95 cm. alti; nodi strigosi;
vaginae apice et margine hirsutae; ligula 0.5 mm. longa, longe ciliata;
laminae planae, 5.5-10.5 cm. longae, 5-10 mm. latae; panicula erecta,
10-16 cm. longa, 1-1.5 cm. lata, fusca, axi ramisque 'scabris, pedicellis
1-3 mm. longis; spiculae 5-6 mm. longae: glumae acuminatae, primo
3-nervio, secundo paulo eee 5-nervio; lemma sterile acuminatum,
3-nervium; fructus 3 mm. lon .6 mm. latus, fuscus, callo piloso, lem-
mate bidentato, aristato, ati ye mm. longa, prope basim geniculata.

Perennial with extravaginal innovations, sometimes rooting at the lower
nodes; culms ascending to erect, 75-95 cm. tall, terete, firm, smooth; nodes
strigose; sheaths hirsute on the collar and on the margin toward the summit,
otherwise glabrous, the lower pies Cake ligule a firm membrane 0.5
mm. long, ciliate with white hairs 3—5 mm. long, the hairs brittle and
broken off in old sheaths; blades Peeing: firm, flat or the margins in-
volute toward the apex, 5.5-10.5 cm, long, 5-10 mm. wide, slightly rounded
at base, acuminate, the margins ha ae ‘and sparsely stiffly ciliate toward
the base, the hairs mostly broken off; panicle long-exserted, erect, 10-16
cm. long, 1—-1.5 cm. wide, dusky purplish brown, the axis and branches erect,
angled, scabrous, the branches slender, the lower distant, naked at base;
pedicels 1-3 mm. long, angled, scabrous; spikelets 5-6 mm. long; glumes
firm, acuminate, the first 3-nerved, the nerves scabrous, the second about
1 mm. longer than the first, 5- nerved, the nerves glabrous: sterile lemma
slightly exceeding the first glume, acuminate, 3- nerved, its palea shorter,
empty; fruit 3 mm. long, 0.6 mm. wide, dark purplish brown, minutely
Se acpi roughened, the oa pilose, the ‘hairs white, 0.5—1 mm. long, the

ma bidentate, the awn 3—4 mm. long, twisted at ‘base, geniculate just
Raa the base

(eae 5826a, collected in March 1937, at 2600 m. alt., Sarawaket, Morobe

District, Northeastern New Guinea. A second collection from the same region is

Clemens 5250, from “open place, Basu Tamanac, Sarawaket,” alt. 2300 m., Jan. 31,

86 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

This species belongs in Arundinella proper as limited by Keng (Nat.
Cent. Univ. Science Reports, Biology 2: 20. 1936). It somewhat re-
sembles A. fuscata Nees, but is taller and nearly glabrous, with longer
spikelets and much larger fruit. It differs from all the species known in
the very short twisted column of the awn, the geniculum being scarcely
1 mm. above the base of the awn.

Fic. 4. Arundinella furva; pas first glume, second glume, sterile floret, and fruit,
10; from Clemens 5826a

Digitaria violascens Link; Hort. Berol. 1: 229.
Paspalum chinensis Nees in Hook. & Arn. Bot. Beeches Aa 5 231, 1836.
sa minutiflorum Steud. Syn. Pl. Glum. 1: 17. 1854; not P. minutiflorum
. 1831.

Piel errim A. Camus, Not. Syst. Lecomte 4: 48. 1923; not D. chinensis
Hornem. 1819.

NETHERLANDS New GuINeEA: Balim ease alt. 1600 m., ibe — sandy soil
on deforested a. ot common. NorTHEASTERN New Guinea: Morobe iidet:
Ogao, alt. 2000 m., Clemens 10359. ane and southern Asia, East Indies, Philippines ;
introduced in poms tropics and subtropics.
Panicum auritum Presl, Rel. Haenk. 7 305. 1830.

se sccaeastti New GuIneEA: Bernhard Camp, Idenburg River, alt. 50 m., Bras

3942; co n swamp forests of river silt Ath. ascending to 2 m. East Indies,

southeastern ren to Philippines.
Panicum zizanioides H.B.K. Nov. Gen. & Sp. 1: 100. 1815.

NETHERLANDS New GvuINEA: Bernhard Camp, Idenburg River, alt. , Brass
13946; abundant in semi-open swampy forest of river flood plain; saat sr ascend-
ing to 1m. Tropical America and Africa, India, New Guinea.

Ichnanthus vicinus (F. M. Bailey) Merr. Enum. Phil. Fl. Pl. 1: 70. 1922.
anicum vicinum F. M. Bailey, Syn. Queensl. FI. Seay “ 82. 1890.
NETHERLANDS ine GuinEA: Hollandia and vicinity, alt. 50 m., Brass 8909;

1943 ] CHASE, PAPUAN GRASSES, III 87

abundant on shaded sandy banks in moist ravine. Four kilometers southwest of Bern-
hard Camp, Idenburg River, alt. 850 m., Brass 13210; occasional on flooded rocky banks
of river. Southern Asia, East Indies, coaied: Australia.

ee contracta (Wight & Arn.) Hitchc. Mem. Bishop Mus. 8: 199. f. 90.

ee contractum Wight & Arn. Linnaea 10: Litt. 117. 1836.
NETHERLANDS NEw GutneEa: Nine kilometers northeast of Lake Habbema, alt.

o abou

m., Brass 11824; plentiful on sandy vai deforested bee erect tufts 40 cm. tall,
ee Asia, Japan, East Indies, Philippine
Isachne albens Trin. Icon. Gram. pl. 85. 1827.

NETHERLANDS NEw GuIneEa: Bele River, 18 kilometers northeast of Lake Habbema
alt. 2200 m., Brass 11559; common in open secondary forest, scrambling or in suberect
clumps 1-1.5 m, tall. Southern Asia and East Indies
Isachne Brassii Hitchc. Proc. Linn. Soc. N. S. Wales 54: 146. 1929.

ERLANDS NEw Gu : Bernhard Camp, wie River, Brass 14055 [no data
on this label, other than aries locality]. Endemi

Plant larger than previous collections, he culms to 50 cm. long, blades
to 7 cm. long and 7 mm. wide, the panicle 7 cm. long.

Isachne Brassii Hitchc. var. villosa Hitchc. Brittonia 2: 123. 1936.
NETHERLANDS NEW GuINEA: Fifteen ann southwest of Bernhard ean
s

open rock slide; more densely villous than usual. Endemi
Isachne globosa (Thunb.) Kuntze, Rev. Gen. Pl. 2: 778. 1891.
Milium globosum Thunb. FI. Japon. 49. 1784
Tsachne australis R. Br. Prod. Fl. Nov. Holl. 196. 1810.
NETHERLANDS NEw GUuINEA: Nine kilometers | northeast of Lake Habbema, alt.

and in old gardens. Balim River, alt. 1600 m., Brass 11823; na kras on deforested
slopes. China to Philippines and East Indies, New Guinea and Aus

Isachne grisea K. Schum. in Schum. & Lauterb. Nachtr. Fl. Deutsch. Schutzgeb.
Siidsee 57. 1905.

Papua: Central Division: Murray Pass, Brass 4642 (listed by Hitchcock, Brittonia
2: 123. 1936, under J. ae ela Hack.). NORTHEASTERN NEW GUINEA: Morobe Dis-
trict: Sattelberg, J. & M. S. Clemens 312; forest, Sarawaket, J. & M. S. Clemens
097. Endemic.

Isachne micrantha Merr. Philip. Jour. Sci. Bot. 5: 168. 1910.

RLANDS NEw GuINEA: Lake Habbema, alt. 3225 Brass 9556; locally
Prarenee on open bi ground, prostrate. Nine kilometers elke of La ie Hab-
bema, alt. 2800 , Brass 10740; plentiful in native clearing in forest. Eighteen
ieee ceiheee of Lake Habbema, alt. 2200 m., Brass 11583; abundant in native
gardens. Four kilometers southwest of Bernhard Camp, Idenburg tenes alt. 850 m.,
Brass 13201; matted on flood-washed rocks on river bank. NorTHEASTERN NEw
GuInEA: Morobe District: Yunzaing, Clemens, Sept. 1936; wet ae a Philippines,

a

Isachne scabrosa Hook. f. Fl. Brit. Ind. 7: 23. 1896.
Papua: Central Division: Mt. Tafa, alt. 2400 m., Brass 4871; plentiful on open

88 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

banks of small stream flowing over landslide debris. Big aiteriaee New GUvuINeEA:
Morobe District: Ogeramnang, alt. 1800 m., J. & M. S. Clemens 6958a; wet open
forest trail. India, New Guinea

Isachne stricta Elmer, Leaflets Philip. Bot. 2: 463. 1908.

NorTHEASTERN New GuINEA: Morobe District: Yunzaing, — oe a & M.S.
Clemens 4105; in seepage over rock slope, rare. Ogeramnang, alt. 1700 m., Clemens
402. Samanzing, alt. 1500-1800 m., Clemens 9208, 9239; ‘eo 5 cepetteae f rivulet.
Philippines, New Guinea

The only collection previously known (the type), from Dumaguete
(Cuernos Mts.), Island of Negros, Philippines, is very immature and has
longer blades than the overmature Clemens specimens, but on the whole
they agree very closely.

Echinochloa crusgalli (L.) sera = Agrost. 53, 161, 169. 1812.
Panicum crusgalli L. Sp. Pl. 175;
NETHERLANDS NEW GUINEA: aL River, alt. 1600 m., Brass 11793, 11803; com-
mon in ditches, erect, 1 m. tall. Nearly awnless form. Warm regions of both hemi-
spheres.

Imperata cylindrica (L.) Beauv. Ess. — 177. 1812.

Lagurus cylindricus L. Syst. Nat. ed. 10. 2: 1759

Imperata arundinacea Cyrillo, Pl. Rar. ae. z 27. 1792.

NETHERLANDS New Guinea: Northern slopes of Mt. Wilhelmina, alt. 3200 m.,
Brass & Myer-Drees 10235; covering beaches of gravel and sand, Wamena River. Nine
kilometers northeast of Lake Habbema, alt. 2600 m., Brass 10902; on old landslip in
forest. Bele River, 18 kilometers northeast of Lake Habbema, alt. 2200 m., Brass
11491; dominant grass on old garden lands. Warm regions of the eastern hemisphere.

Saccharum spontaneum L, Mant. PI. 2: 183. 1771.

NETHERLANDS NEw GutneEa: Eighteen kilometers northeast of Lake Habbema, alt.
2200 m., Brass 11374; very abundant in tall thickets on abandoned garden land,
2.5-3 m. high, Bele River. Balim River, alt. 1600 m., Brass 11778; abundant on
alluvial soil of river banks. Four kilometers southwest of Bernhard Camp, Idenburg
River, alt. 850 m., Brass 13264; colonizing sand and gravel beaches. Bernhard Camp,
Idenburg River, alt. 50 m., Brass 13791; in dense pure stands, 7-8 m. high, on recent
silt deposits of river banks. Warm regions of Asia, the East Indies, Philippines.

jaca spontaneum var. klagha (Jungh.) Hack. in DC. Monogr. Phan. 6: 116.
1889.

Saccharum klagha Jungh. ex Steud. Syn. Pl. Glum. 1: 405. 1854.

NETHERLANDS New Guinea: Hollandia and vicinity, Brass 8920. Java.

Differs from the species in the slightly larger spikelets with shorter less
silky hairs.

Eulalia leptostachys (Pilger) comb. n

Pollinia leptostachys Pilger, Bot. Jahrb. Engler 52: 170. 1914.

NETHERLANDS NEW GUINEA: siess kilometers northeast of Lake Habbema, alt.
2600 m., Brass 10901; a few tufts on dry open landslip in forest. Bele River, 18
lometers northeast of Lake Habbema, alt. 2200 m., Brass 11363; covering high,
gravelly, formerly forested banks of river, ascending and tangled. eee River, alt.
1600 m., Brass 11825; locally abundant on long-deforested slopes. Endem
Microstegium gratum (Hack.) A. Camus, Ann. Soc. te Lyon 68: 201. 1921.

Pollinia grata Hack. in DC. Monogr. Phan. 6: 175.

Eulalia grata Kuntze, Rev. Gen. Pl. 2: 775. 1891

1943 | CHASE, PAPUAN GRASSES, III 89
Four kilometers southwest of Bernhard Camp, Iden-

NETHERLANDS NEW GUINEA:
ae seral rain-forest on sandy

burg River, alt. 850 m., Brass 13720; occasional in
flood banks. Southern China to India and East Indie
Andropogon intermedius R. Br. Prodr. Fl. Nov. Holl. 202. 1810.

Balim River, alt. 1600 m., Brass 11616, 11805, 11845,

NETHERLANDS NEW GUINEA:
China to India, East Indies, Philippines and Australia.

common on deforested slopes.

U.S. NaTIonaL HERBARIUM,
WasuincrTon, D. C.

90 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

NEW PHANEROGAMS FROM MEXICO, V

IvaAN M. JOHNSTON
Milla Bryani, sp. nov
Planta 3—9 dm. sia erecta; foliis se a revolutis gue sone laevi-
bus 2-4 dm. longis; scapo apice 1—3-floro saepe infra medium aspero;
bracteis subulatis 5-8 mm. longis ; floribus albis lineis set ee 5 notatis
16-18 cm. longis 1- “3 (saepe ca. 2) cm. longe hapa gator 4 gracili
basim versus 1.5 mm, crasso summum ad apicem sub anth
crasso, lobis alee subaequalibus 14-18 mm. longis 4.5—6. 6 mm. latis
medium versus vel paulo supra medium latioribus utrinque contractis;
staminibus exsertis Pts ad faucem affixis; antheris oblongis in sicco
ca. 3 mm. longis (sub aqua ad 7 mm. longis) extus vesciculari- rugulosis;
mm.

apicem versus gradatim attenuatis; filamentis a 2 (vel 3) in faucem
corollae decurrentes gerentibus.

Coanutmta: Western base of Picacho del Fuste, on mountain side, Johnston 8364;
head of Canon del Cuervo Chico, rocky slopes and crests, Johnston 8529 (TypE, Gray
Herb.) ; Corte Blanco fork of Charretarras Canyon, Sierra Madera, grassy rocky flat
forming in oak-thickets, Johnston 9123; ridge of Sierra de la Fragua north of Puerto
Colorado, rocky soil in openings among brush and pines, Johnston 8777; Canyon del
Agua, Sierra Madera, Muller 3203

Differing decisively from és widely ranging Milla biflora Cav. in its
well developed exserted filaments and more slender and elongate flower.
The filaments bear two plates of tissue on their inner surface. These plates
narrow as they extend upward toward the base of the anther. The summit
of the immature ovary bears three lines of minute conic trichomes. These
are more prominent than those frequently present on the ovary of M.
biflora. The pale green lines down the middle of each perianth-segment
of M. biflora are usually evidently three-nerved. In the proposed species
these nerves are very inconspicuous. Milla biflora has been collected in
various parts of Chihuahua, but the only station in Coahuila known to me
is in the Sierra de Hechiceros on the Chihuahua border. The species here
proposed ranges well to the east of the known range of M. biflora and
appears to be endemic to the limestone mountains of north-central Coahuila.

I have associated with this species the name of my companion ae
much of my field work in Coahuila in 1941, Prof. Kirk Bryan of the
partment of Geology and Geography of Fucourd University. His ade.
lating companionship and his geologist’s pick, frequently used for botanical
purposes, contributed greatly to the pleasure and botanical success of the
weeks we travelled together. It is a pleasure to associate his name with
one of the plants he helped me collect.

Nolina texana Wats., var. compacta (Trel.), comb. no
N. erumpens compacta Trel. Proc. Amer. Philos. Soc. 50: 418. 1911.

1943 ] JOHNSTON, NEW PHANEROGAMS FROM MEXICO, V 91

N. affinis Trel. ].c. 417.

N. caudata Trel. l.c. 417.

Differing from typical V. texana, of the eastern portions of the Edwards
Plateau, in its more robust habit and in the more elongate, much larger, and
more densely and abundantly branched female inflorescences. The ascend-
ing branches of the inflorescence are much more rigid and much more
abundantly and stiffly short-branched. Except for Thompson’s collection
from Marathon (which is typical NV. erumpens), all of the material of the

var. compacta originally cited by Trelease belongs to the present plant.
Trelease’s N. affinis is a mixture of the present variety and N. micrantha.
The actual type-sheet of N. affinis, at St. Louis, however, entirely repre-
sents the present variety. The var. compacta ranges from the western
parts of the Edwards Plateau, in Texas, west to southern Arizona and south
into northern Chihuahua
Nolina micrantha, sp. nov.
Planta acaulis 6-10 dm. alta; foltis linearibus numerosis duris 5—13 dm.

laevibus vel scabris; panicula 4-6 dm. longa 10-15 cm. diametro; ramis
rigidis numerosis laevibus gracilibus ascendentibus vel fete ascendentibus,
superioribus saepe simplicibus ad 15 cm. longis, inferioribus saepe 10-15
cm. longis ramosis (ramulis 2-12 gracilibus ascendentibus 3-7 cm
yarn bracteis 1—3 dm. longis basi scariosis dilatis ad 1 cm. latis alibi fili-
formibus; floribus ex axillis fing secundariarum minutarum scario-
sarum erosarum 1-2 vel rariter 3 mm. longarum erumpentibus; perianthii
lobis 2—2.5 mm. longis oblongis apice ae masculi ochroleucis, feminei
costa rosea mediali notatis; capsula purpurascenti 3-4 mm. alta 5—6 mm.
lata, pedicello 2.5-4 mm. longo ca. 2 mm. infra perianthum articulato;
seminibus globosis ca. 3 mm. diametro capsulas chartaceas inflatas dirum-
pentibus.

CuinvuaHua: Vicinity of Santa Eulalia, 1908, Palmer 139; rocky hills, Bachimba
Canyon south of Chihuahua, Pringle 2 in pt. (G); Organos, base of grassy hills with
large oaks, Stewart & Johnston 2072; Sierra Hechiceros, Rancho Encampanada, sunny
hillside, Stewart 193. Coanvutta: Sierra Hechiceros, vicinity of Rancho El Tule, rocky
slopes and flats, Stewart 490 and Johnston & Muller 1326 (type, Gray Herb.).

A well marked species, probably most closely related to V. texana, from
which it differs in its distinctly smaller flowers, looser larger inflorescence
with less rigid, less twiggy, more slender and elongate branches, pur-
purascent capsules, and minute less lacerate bractlets.

Trelease’s original publication of NV. affinis is based upon specimens rep-
resenting both the present species and Nolina texana var. compacta. The
material at St. Louis, which Trelease designated as the type of his V. affinis,
is representative of N. texana var. compacta. It is a composite sheet of
Pringle 1 and 2. At the Gray Herbarium the sheet of Pringle 2 is half
fruiting material of N. texana and half the present plant. The Gray
Herbarium sheet of Pringle 1 is entirely male flowering material of NV.
texana. According to Pringle’s field notes his two sheets, nos. 7 and 2,
were collected April 2 and May 22, 1885, in Bachimba Canyon, which is
about 35 km. southeast of Chihuahua City.

92 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Dasylirion heteracanthum, sp. nov.

Planta robusta D. leiophyllo affinis; foliis 8-11 dm. longis supra basim
ampliatam 2-3 cm. latis plus minusve opacis viridibus, margine spinis
appa g divaricatis rectis vel antrorse arcuatis munitis, spinis gracilibus

longis 5-15 mm. distantibus saepe lutescentibus rariter ——
versus i sea fructu skies ek oe 6-7 mm. longo 4-5 m
lato, sinibus apicalibus latis haud profu

CoauuIta: Western base of Picacho del Fuste, trunk up to 3 ft. tall, scape 10-15 ft.
tall, sites on rocky slopes and flats, Soknsian 8428 (type, Gray Herb.). TExas:
Persim n Gap area, Brewster Co., July 30, 1938, Sperry 1321; Chisos Mts., July 5,
1931, Mueller 7958.

In Texas this plant occupies the area between D. texanum Scheele, of the
Edwards Plateau, and D. letophyllum Trel., of western trans-Pecos Texas.
It has the fruit of D. texanum but much longer and broader non-glossy
leaves, of which the marginal thorns are distinctly divergent and straight
or weakly curved, rather than ascending, stout, and strongly up-curved.
The proposed species differs from D. leiophyllum in its somewhat broader
non-lustrous leaves and in the absence of abundant stout strongly curved
retrorse marginal thorns. Its fruit may be similar to those of D. texanum
and D. leiophyllum or narrower with a deep apical notch.

Dasylirion Stewartii, sp. no

Planta robusta; foliis metralibus supra basim 2-3 cm. lata viridibus vix
lucentibus margine spinis validis brunnescentibus conspicue retrorseque
curvatis armatis; fructibus a 7 mm. longis 4-5 mm. latis obovatis apice
profunde et anguste emarginatis, alis 1-2 mm. latis.

HUILA: Vicinity of Santa Elena, eastern foothills of the Sierra de las Cruces,
1941, Stewart 823 (type, Gray Herb.) and 841; 7 km. north of Santa Elena, Johnston
& Muller 331.

There is only one described species of Dasylirion with retrorse thorns on
the leaf-margins, D. leiophyllum, of western trans-Pecos Texas and
Chihuahua. From that species, D. Stewarti differs conspicuously in its
larger size, its very much coarser and more strongly armed non-lustrous
leaves, its more deeply notched fruit, and its detached more south-
easterly range. The plants of the Sierra de las Cruces, where it abounds
on the limestone foothills, have a trunk becoming a meter tall and a large
head of light green non-lustrous leaves. The flowering stalk becomes
several meters tall. There is a vinata at San José, at the southeastern base
of the Sierra de las Cruces, where for many years D. Stewartii has been
used for the preparation of the alcoholic liquor ‘‘Sotol.”

To a botanist who has not observed and lived with Dasylirion in the field,
the fact that various species differ in the direction of their leaf-thorns may
seem of minor importance. To those living with these plants the differences
are very real indeed. In areas having species with antrorseiy armed leaves,
horsemen and cattle brush these plants with impunity. One may even
grasp a handful of their leaves to help one up a steep slope. The collect-
ing of the stem-crown for sotol-making or the gathering of the flowering
stems (‘‘garochas’’) for their many uses (from corks to building material)
is a simple task. About the Sierra de las Cruces, however, where D.

1943 ] JOHNSTON, NEW PHANEROGAMS FROM MEXICO, V 93

Stewart is the only species, all this is very different, for the thousands
of cats-claws on the leaves of that species rake savagely the flanks of any
animal brushing this plant. Any rider, unprotected by heavy leather chaps,
has his legs and thighs unmercifully clawed and perhaps his clothes torn,
if he is brushed, while on some narrow trail, against one of the massive
clusters of viciously armed leaves. One who collects this plant, either for
the herbarium or for the making of sotol, unless he is very cautious and
deliberate, bears some scratches after his encounter with the plant. Because
of its retrorse leaf-thorns, the plant has won the same respect that is
accorded to such cacti as Opuntia. Direction of leaf-thorns do make a
difference!

With this handsome plant I have associated the name of my good friend,
Mr. Robert M. Stewart, of Santa Elena. Through his effort the area about
Santa Elena has now become botanically the best explored in all of northern
Coahuila. It is a pleasure to associate his name with one of the most
common and conspicuous plants in that area.

Dasylirion Stewartii, var. glaucum, var. nov

A forma typica differt foliis pallidis conspicue glaucis.

CoAHUILA: 3 m. northwest of El Oro, on road to Esmeralda, 1939, White 1970.
CutHuAHuA: Mouth of Canon del Rayo, Sierra Diablo, 1941, Stewart 957 (TYPE, Gray
Herb.)

The above cited collections have the same fruit as the plant about Santa
Elena, but their leaves are very glaucous rather than pale green. Perhaps
also the thorns may be less coarse and the leaf-margin between them more
denticulate. This glaucous plant is so strikingly different that I believe it
merits a name.

Phorodendron flavum, sp. nov.

Planta dioica fulvescens, partibus junioribus omnino velutinis_pilis
stellatis minutis mollibus fulvis dense obtectis, partibus vetustioribus tarde
subglabrescentibus pilis sparsis pallidioribus donatis; caulibus lignosis
rigidis saltem ad 3 dm. longis ascendenter ramosis, internodiis 1-3 cm.
longis inferioribus 4-7 mm. crassis; foliis oblongo- vel ovato-ellipticis quam
internodiis saepe longioribus, majoribus 3.5—3.8 cm. longis 1.3—1.5 cm. latis,
1.5—2.3 mm. longe et 3—4 mm. late petiolatis; lamina crasse coriacea fulve
fusco-viridescenti, medium versus vel paulo infra medium latiore, basi
rotundata vel angulata in petiolum abrupte contracta, apice rotunda vel

non raro plus minusve acuta, subtus (in sicco) plus minusve prominenter
costata; spicis femineis (masculis non visis) ad axillas solitariis vel pluri-
bus, ad anthesim 10-14 m. longis, fructiferis ca. 2 cm. longis, 3- vel raro

5- vel 6- articulatis; Herbie in series 2 vel 3 dispositis 8-12 pro articulo;
baccis subglobosis ca. 4 mm. diametro glabris laevibus perianthio clauso
basi ciliato coronatis.
ae Vicinity of Durango, on Quercus, Nov. 1896, Palmer 777 (TYPE, Gray
. Coanurta: Sierra Negras, 9 km. south of Parras, on Quercus, Stanford,
Retae-tova ¢& Northcraft cee hills 11 km. northeast of Jimulco, on Quercus, Stanford,
Retherford & Northcraft 7
This species ae to the Boreales-Pluriseriales-Flavescentes of Tre-
lease. The type-collection is cited and illustrated by Trelease, Monogr.

94. JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

Phor. 42, tab. 41b (1916), as an aberrant form of P. tomentosum (DC.)
Engelm. I do not believe that the species proposed here is very closely
ae to P. tomentosum. ‘That species, based upon material collected near
Catorce, S.L.P., “supra Mimoseas,” is the common parasite on Leguminosae
in the intermontane desert valleys of northern Mexico. It is a grayish green
plant with proportionately broader leaves and a thinner grayish indument.
The present species is a montane plant, growing on oaks, and apparently
reaching southwestern Coahuila from the highlands of Durango. Its very
dark green herbage, its more coriaceous, more elongate leaves, and its
abundant softer tawny velvety indument set it off pi cane ay from
P.tomentosum. The three specimens cited form a very uniform series

Gilia Stewartii, sp. nov.

Planta erecta 1-3 dm. alta, ut videtur biennis, stricte ascendenter ramosa
minute inconspicueque glandulifera basim versus fruticulosa; foliis numero-
sis firmulis pinnatipartitis; foliis radicalibus congestis sub anthesi saepe
delapsis 2—5 cm. longis saepe bipinnatisectis, lobis distantibus saepe 1 mm.
latis raro latioribus; foliis caulinis saepe pinnatifidis sursum reductis, infra
medium caulis 3—5 cm. longis, lobis saepe linearibus 3- vel 4-jugatis 5—20
cm. longis ca. 1 mm. latis; floribus numerosis laxe paniculatis; pedicellis
1-3 cm. longis ascendentibus rigidulis glandulis minutis dense obsitis;
calyce fere ad basim partito 4-5 mm. longo, lobis cuneatis strictis margine

corolla lilacina vel violacea usque ad basim fere partita quam calyce plus
duplo longiore, lobis 7-12 mm. longis 2—5 mm. latis elongatis medium
versus latioribus, apice acutis vel late acutis, basi subito in unguem angus-
tatis; filamentis exsertis filiformibus quam antheris oblongis valde longiori-
bus; ovario glabro; capsula ellipsoideo-ovoidea apicem lobis calycis attin-
gente vel paulo superante; seminibus numerosis minutis sub aqua mucila-
ginosis.

6 mi. north of Hot aera Brewster Co., Innes & Warnock 546; stony
hills near Oultman Canyon, Hudspeth Co., Chas. Wright. pial ices ills near
Chihuahua, Aug. 11, 1885, Pringle 530; 12 mi. south of Camargo, White 2196; 1 km.
north of Victoria, Stewart é& Johnston 2004; Canon del Rayo, Sierra Diablo, Stewart
904 and 950. Duran nco: Cerro de San Tenacio, July 1910, Purpus 4595. COAHUILA:

km. north of Eutimias, Stewart 1750; 15 km. east of La India, Llano de Guaje,
Stewart 1184; near Tinaja Blanca, Sierra de las Cruces, Stewart 2235; Santa Elena,
Sierra de las Cruces, Johnston & Muller 236; south base of Picacho de San José,
Johnston & Muller 801; 5 mi. south of San José, Johnston & Muller 1258 (TYPE,
Gray Herb.) ; Cafion del Gringo, Sierra Planchada, Stewart 1049; Canon de Hidalgo,
Sierra Mojada, Stewart 1096; Sierra Mojada, Jones 348; Tanque Jerico, north margin
of Canada de Cuervo Grande, Johnston 8345; Puerto Colorado, Johnston 8708;

Soledad, 1880, Palmer 845; Saltillo, 1898, Palmer 312; Saltillo, 1878, Parry 8; 2 mi.
west of Saltillo, White 1060; 4 mi. north of Pena, Johnston 7718. Hipartco: Ixmiquil-
pan, 1905, Purpus 1398. GuANAJUATO: Jaral, 1887, Schumann 351

A well marked species of the desert limestone mountains, growing on
slopes and on outwash near their bases. It appears to favor silty soils. The
lilac or pale violet or lavender corolla-lobes are elongate and taper to a
point. The shape and color of its corolla-lobes, as well as its deeply cut
calyx, which is equalled or over-topped by the capsule, readily distinguish

1943] JOHNSTON, NEW PHANEROGAMS FROM MEXICO, V 95

the new species from G. rigidula Benth. Its pinnate leaves, with few distant
slender elongate lobes, as well as its glanduliferous calyx, its abundant basal
branching, and its lower stature, readily distinguish it from G. incisa Benth.,
probably its closest relative. Materi al from the eastern half of Coahuila
appears to have smaller corollas (7-8 mm. long) than that from other areas.
The corollas are usually 9-12 mm. long. The species is named in honor of
Mr. Robert Stewart, who has made numerous collections of this charming
little plant.

Gilia platyloba, sp. nov.

Planta erecta 2-3 dm. alta, ut videtur biennis, stricte ascendenter ramosa
minute inconspicueque glandulifera ceterum glabra basim versus fruticu-
losa; foliis firmulis; foliis radicalibus congestis sub anthesi saepe delapsis;
foliis caulinis pinnatis vel bipinnatis, infra medium caulis 2—6 cm. longis
1—2.5 cm. latis, lobis 2—4-jugatis linearibus vel anguste cuneatis integris vel
sparse breviterque incisis; floribus numerosis laxe paniculatis, pedicellis
1—3 cm. longis rigidulis dense glanduliferis; Si fere ad basim partito 5—6
mm. longo, lobis strictis cuneatis margine membranaceis partibus mediali-
bus viridibus trinervatis et glanduliferis; ea caerulea usque ad basim
partita quam calyce plus quam duplo longiore, lobis obovatis 13-17 mm
longis 7—9 mm. latis supra medium latioribus apice rotundis vel obtusis basi
in unguem angustatis; filamentis exsertis filiformibus quam antheris luteis
oblongis valde longioribus; capsula late ellipsoideo-ovoidea apice lobos
calycis attingentibus vel paulo superantibus; seminibus minutis numerosis
sub aqua mucilaginosis.

Coanumta: Saltillo, 1898, oc 799 (tTypE, Gray Herb.); Fraile, Stanford,

ZACARECAS: Con ncepcion del Oro, 1904, Palmer 282; 18 km. west of ga oa .
Oro, poi cone ee é& Northcraft 602; Cedros, Lloyd 85. Duranco: Dura
1896, Palm

Related to 2 Stewartii, from which it differs chiefly in the shape of
the corolla-lobes and their larger size and dark blue color. From G. rigidula
Benth., with which it has been confused, the new species is readily dis-
tinguished by its cleft, inconspicuously glanduliferous calyx, and its more
elongate capsule which equals or surpasses the calyx in length. From G.
incisum Benth., another relative, it differs in its pinnate leaves and very
large corollas.

Gilia aggregata Spreng., var. texana Speeoer , comb. nov.

Callisteris texana Greene, Leaflets 1: 160. 190

This is the variant of G. aggregata aad in rocky arroyos in the oak-belt
of the Sierra Madre of Tamaulipas and Nuevo Leon, in the desert ranges of
Coahuila, and north into trans-Pecos Texas.

Gilia calothyrsa, sp. nov.

Herba biennis 3—5 dm. alta basim versus ramosa; ramis pluribus erectis
Strictis pilis minutis crispis albis mollibus — foliis pinnatifidis;
foliis caulinis inferioribus 2—3 cm. longis 1.5—2 cm. latis, rhachi lineari
utrinque segmenta 3 vel 4 anguste linearia 0.5.0.7 m mm. lata gerente; foliis
superioribus gradatim reductis supremis simplicibus linearibus; floribus in
glomerulis 3—10-floris sessilibus vel ad 6 mm. longe pedunculatis gestis, in

96 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

thyrsum elongatum 10-18 cm. longum subsecundum dispositis; calyce
sparse inconspicueque glandulifero 5-7 mm. longo, lobis 2-3 mm. longis
subulatis spinescentibus herbaceis quam tubo membranaceo brevioribus;
corolla violaceo-purpurea saepe 1.5 (raro ad 2) cm. longa salviformi, tubo
2-3 mm. crasso plus minusve curvato, limbo 10-13 mm. diametro, lobis
patentibus ovatis 3-5 mm. latis quam tubo duplo longioribus apice rotundis
apiculatis; staminibus 5 ad tubum corollae valde inaequaliter affixis haud
vel vix exsertis; stylo vix exserto basim versus pilis brevibus sparsis ornato;
capsula et seminibus ignotis.

Coanvutta: Sierra de las Cruces, rocky slope about the summit of the highest peaks,
flower light purple, Stewart 1044 (Type, Gray Herb.). Curuuanua: Just east of
Organos, local on rocky flat, flowers violet, Stewart & Johnston 2054; Sierra de los
Organos, LeSueur 1391.

A species related to G. Macombi Torr., G. Thurberi Torr., and G.
Pringlei Gray, and occurring in an area to the east of that occupied by
these species. It is probably closest to G. Pringlei, from which it differs in
‘ts coarser more branched stems, smaller leaves, and short salverform
(rather than trumpet-shaped) corollas with flat rounded apiculate (rather
than curved lance-ovate attenuate) lobes. It is a beautiful and attractive
plant and merits cultivation.

Nama Marshii (Standley), comb. nov.

Nama biflorum var. Marshii Standley, Field. Mus. Pub. Bot. 22: 167. 1940.

This species is most closely related to NV. propinquum Mort. & Hitchc.,
having similar long-petiolate leaves and more or less cordate blades. The
leaves, however, are thinner, green, and somewhat shaggy villous or gla-
brescent. The pedicels are more elongate and the stems are more slender
and flaccid.

Nama serpylloides Gray, var. confertum, var. nov.

A varietate typica differt floribus confertis; pedicellis crassioribus 1—2
mm. longis maturitate vix elongatis; foliis cum pilis gracilioribus longioribus
abundantibus velutinis; caulibus rigidioribus.

Coanutta: 2 miles west of Cuatro Cienegas, spreading over a low bank of alkaline
gypseous soil, leaves fleshy, 1938, Johnston 7126 (type, Gray Herb.) ; a mile west o

porate :

J ine
lent, corolla pale pink, 1941, Johnston 8868; Cuatro Cienegas, 1939, Marsh 2016

A form evidently related to N. serpylloides var. velutinum Hitchc. (to
which it was referred by the author of that variety), but differing in its
coarser more loosely branched and more rigid branches, its denser yellowish
(rather than grayish) velutinous indument, and especially in its perma-
nently short-pedicellate congested flowers. In typical N. serpylloides and
in the var. velutinum, the pedicels are slender, eventually spreading, and
become 1-2 cm. long. The leaf-blade of the var. confertum is more succu-
Jent than in the var. velutinum, and perhaps even more revolute and be-
coming more pronouncedly boat-shaped.

Phacelia infundibuliformis Torr., var. phanerandra, var. nov.

A varietate typica differt filamentis lobos corollae evidenter sed breviter

superantibus antheras aurantiacas conspicuas proferentibus.

1943 JOHNSTON, NEW PHANEROGAMS FROM MEXICO, V 97

Coauvutta: Sierra de las Cruces near Tinaja Blanca, frequent on arroyo-banks,
March 12, 1942, Stewart 2241 (type, Gray Herb.). Texas: 14 mi. east of Castolon,
Brewster Co., frequent along creek, Cutler 749.

This variety occurs far to the east of the known stations for typical
G. infundibuliformis and appears to be a geographic race distinguished by
its protruding stamens. In other characters it agrees closely with the
typical form of the species.

Phacelia robusta (Macbride), comb. nov,

Phacelia integrifolia var. robusta Macbr. Contr. Gray Herb. 49: 25. 1917.

A coarse glandular herb on rocky places, along arroyos, and about cliffs.
It has been associated with P. integrifolia Torr., from which it differs in
being non-gypsophilous and in having larger salverform (rather than sub-
tubular) corollas and larger non-corrugated seeds. From trans-Pecos
Texas (Chisos and Chinati mountains) it extends far south in Coahuila
and Chihuahua. Under the name “P. integrifolia var. arenicola,” Brand,
Pflanzenfam. 59 (IV. 251): 81, f. 17 (1913), has given a mediocre illus-
tration of P. robusta, probably based on material collected near Chihuahua
City by Pringle or by Palmer.

Phacelia pallida, sp. nov.

longe petiolatis, superioribus duplo minoribus onge petiolatis;
cymis terminalibus pluribus densifloribus; calyce ca. 4 mm. longo, lobis
oblongis ciliatis dorso glanduliferis cum setis vestitis ca. 1 mm. longe pedi-
cellatis, fructiferis ad 6 mm. longis, lobis spathulatis capsulam evidenter
superantibus; capsula ovoidea ca. 4 mm. longa hispidula; seminibus 4
nigris ca. 3 mm. longis corrugatis; corolla ca. 6 mm. longa, tubo basi 1.5
mm, apice 3-4 mm. crasso, 4 mm. longo, lobis ad 2 mm. longis ascendenti-
bus; staminibus longe exsertis.

Coanurita: Gypsum beds on the escarpment of Cafiada Oscuro near Tanque La
Luz, fleshy grayish non-glandular herb confined to gypsum, corolla lavender-white,
Johnston 8486 (typE, Gray Herb.).

A relative of P. integrifolia Torr., characterized by its practically gland-
less herbage, branching habit, large gray pallid frequently lobed petiolate
leaves, and few terminal cymes. Many of the lower leaves have a pair of
small lobes borne on the petiole just below, and separated from, the blade
proper.

Phacelia petiolata, sp. nov.

Herba glandulifera 1-3 dm. alta e radice lignosa annua erumpens; cauli-
bus pluribus erectis vel decumbentibus sparse ramosis foliosis pilis minutis
vix rigidis abundantibus glanduliferis et pilis longioribus divergentibus
sparsioribus vestitis; foliis numerosis evidenter petiolatis; lamina glandulis
minutis sessilibus utrinque obsita pilis erectis vel ascendentibus gracilibus
vestita, in ambitu late ovata vel elliptica, medium versus latiore, basi late

98 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

acuta, obtusa vel reniformi, margine irregulariter crenata et non raro
paulo lobulata; circinnis saepe terminalibus solitaribus vel geminatis densi-
floris; calyce pilis brevibus rigidulis ascendentibus vestito glandulifero, ad
anthesin ca. 3 mm. lon ngo ca. 0.8 mm. longe pedicellato, fructiferi ca. 5 mm.
longo quam capsula conspicue longiore, lobis spathulatis ad apicem 1.3—1.7
mm. latis; corolla 6-8 mm. longa, tubo 4—5 mm. longo cylindrico apice
3.5-4 mm. diametro basi 1 5 mm. crasso, lobis reas mm. diametro ascendenti-
bus; staminibus longe exsertis; seminibus nigris 2-3 mm. longis margine
carinaque corrugatis; capsula ovata ca. 3 mm. longa hispidula.

Coanvuimta: San Lorenzo de la Laguna, 1880, Palmer 851; 12 mi. south of Ojinaga,
abundant on bank of saline clays, corolla pale lilac, Johnston 8040 (TypE, Gray Herb.) ;
11.5 mi. south of Ojinaga, a few plants about a limestone ledge in deep arroyo,
Johnston 8036.

A relative of Phacelia integrifolia, readily recognized by its broadly
elliptic distinctly petiolate leaves, much branched low growth-habit, thick-
ened woody root, sparse cymes, and black corrugated seeds. The lilac
corolla has a moderately ampliated tube and throat.

Phacelia teucriifolia, sp. nov.

Annua herbacea erecta; caulibus solitariis vel pluribus saepe simplicibus
10-25 cm. longis rectis rigidulis pallidulis, cum pilis minutis gracilibus
adpressis inconspicuis sparse vestitis; foliis basalibus ignotis; foliis caulinis
4—7-pinnatifidis vel pinnato-lobatis adpresse minuteque ieeechei apie ort
inferioribus petiolatis 4-6 cm. longis in ambitu oblanceolatis, mediis e
superioribus sessilibus gradatim reductis in ambitu oblongis vel og.
summum ad 4 cm. longis, lobos 3—5-jugos ovatos vel saepe lanceolatos
gerentibus; cymis racemiformibus maturitate laxifloris; pedicellis 8-12 m
longis ascendentibus rectis vel curvatis; calyce sub anthesi 5—6 mm. aie
ca. 1.5 mm. lato, lobis oblongo- lanceolatis cee hispidulis margine cilio-
latis, fructiferis herbaceis accrescentibus 7—12(—15) mm. longis erectis vel
ascendentibus quam capsula subduplo vel plus duplo longlon bus corolla
rotato-campanulata 7-10 mm. longa 12-14 mm. diametro, lobis latis
rotundis integris ascendentibus; staminibus sparse ciliatis 6-9 mm. longis;
ovario dense piloso; stylo usque ad medium partito ciliolato; capsula sub-
globosa 4-6 mm. longa sparse adpresse hispidula; seminibus 10—20 irregu-
lariter prismaticis 1.5—2.5 mm. longis nigris papillatis irregulariter foveo-

tis.

Coanutta: Muzquiz, 1935, Marsh 138a; Muzquiz, April 12, 1936, Marsh 2120
(type, Gray Herb.) and 2135. Texas: 4.7 mi. west of Menard, 1929, Cory 640;
Tarrant County, 1923, Ruth 459.

Related to P. strictiflora Gray, and ranging to the west and south of
that species. It differs from its relative in its erect or strictly ascending
stems, less conspicuous less dense paler non-viscidulous indument of more
appressed paler hairs, and its loosely ascending rather than strict fruiting
pedicels. The plant dries a light green. The color, texture, and shape of
its stem-leaves are suggestive of those found in forms of Teucrium
cubense L.

ARNOLD ARBORETUM,
HaArvarD UNIVERSITY.

1943] LI, SCHIZOMUSSAENDA 99

SCHIZOMUSSAENDA, A NEW GENUS OF THE RUBIACEAE
Hu1-Lin Lr

IN connection with a study of the accumulated unnamed collections of
Chinese material in the herbarium of the Arnold Arboretum, my attention
was called to a series of specimens which had been variously named. Some
of them had been correctly determined as Mussaenda dehiscens Craib of
the Rubiaceae, but others, although manifestly representing the same
species, had been referred to Emmenopterys Rehderi Metcalf of the Rubia-
ceae and Schizophragma macrosepalum Hu of the Saxifragaceae. In view
of the anomalous position of this dehiscent-fruited rubiaceous species in
Mussaenda, a genus characterized by the indehiscent more or less fleshy
berry-like fruits, a critical examination was made of all available material.
As a result of this study, I have concluded that a distinct genus is rep-
resented, which is here described. At the same time the synonymy of the
species is adjusted. All specimens cited are deposited in the herbarium of
the Arnold Arboretum. This study was made possible by a grant from
the Milton Fund of Harvard University to Dr. E. D. Merrill of the Arnold
Arboretum.

Schizomussaenda gen. nov.

Inflorescentia terminalis cymosa trichotomo-racemosa, ob florum quo-
rundam calycis lobum unicum in laminam amplam dilatatum conspicue
4—6-bracteata, ramulis ultimis erectis paribrnibus subscorpioideis vel
scorpioideis, bracteis bracteolisque lineari-lanceolatis pilosulis persistenti-
bus vel deciduis. Flores sessiles vel subsessiles, calycis tubo elongato-
turbinato puberulo, margine 5-lobato, lobo inte rdum unico in laminam

amplam petiolatam producto, lamina elliptica vel ovata, alba, apice acuta,

basi cuneata rotundata vel subcordata, 5-nervia, in nervis tantum obscure

puberula, laminae petiolo supra canaliculato, subtus sulcato, calycis lobis
normalibus lanceolato-oblongis acutis erectis; corolla hypocrateriformi,
g

een peltata carnosa. ale nigro-brunnea, turbinata vel oblongo-
obovata, apice loculicide dehiscens; seminibus numerosis, minutis, testa
fragili brunnea foveolata.

Frutex, ramis quadrangularibus pallide brunneo-cinereis parce lenti-
cellatis, ramulis dense subadpresse pilosis; foliis tenuiter chartaceis na
tico- oblongis vel oblanceolatis, apice acuminatis, basi cuneatis ve
cordatis, supra viridibus, subtus pallidioribus, costa a nervisque supra rae
elevatis pilosis, subtus prominentibus laxe pilosis, nervis lateralibus utrinse-
cus 8-10 adscendentibus prope margine arcuatis et conjunctis, anastomo-
santibus, rete venularum utrinque subconspicuo; petiolis canaliculatis

100 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

adpresse pilosis; een caudato-lanceolatis integris adpresse pilosis, per-
sistentibus vel dec

ne species in nee Burma, northern Siam, Tonkin, and southwestern
and southern China (Yunnan, Kwangsi, and Kwangtung).

Schizomussaenda dehiscens (Craib) comb, n

Mussaenda dehiscens Craib in Kew Bull. 1916: 263. 1916; Pitard in Lecomte, FI.
Gén. Indo-Chine 3: 174, fig. 12, 2-3. 1923; Chun in Sunyatsenia 1: 306, fig. 5.
1934.

Mussaenda Roxburghii sensu Drake del Castillo in Jour. de Bot. 9: 216. 1895; non
Hook. f. (1880).

Mussaenda glabra sensu Pierre ex Pitard in Lecomte, He ae Indo-Chine 3: 174.

n syn. sub Mussaenda dehiscente; non be (17

Greenea tegen eae ex Pitard in Leco omte . Gén. eae 3: 174. 1923,

nome

Schizophragma eS I Hu n Jour. Arnold Arb, ll: 48, 1080: Hu & Chun,
Ic. Pl. Sin. 4: 13, t. 163. 1935.

ulin Rikdert Metcalf in Lingnan Sci. Jour. 11: 528. 1932.

Bu Keng Tung Territory, Meh Lui watershed, J. F. Rock 2303 (syntype of
Pasa Rehderi Metcalf), Feb. 10, 1922. Sram: Chiengmai Prov., between
Meh Soi and Hue San, J. F. Rock 1854 (syntype of Emmenopterys Rehderi Metcalf),
Jan. 5, 1922. Inpo-Cuina: Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T.
Tsang 26900, Oct. 16-22, 1936, 29128, May 21-31, 1939, 29248, June 10-22, 1939,
Tonkin, Tien-yen, Kau Nea Shan, Ww. T. Tsang 27253, Jan. 1-9, 1937, 30521, Sept. 23 -
Oct. 7, 1940; Tonkin, Dam-ha, Se i Wong Mo Shan, Lung Wan Village, W. T. Tsang
29897, 30004, May 19 - July 5, 1940, 30214, July 18 — Sept. 9, 1940; Sontoy, A. Pételot
6600, Dec: 2, 1940. CHIN Yunn nan: Menglieh, A. Henry 12825 (isoparatype of
Mussaenda dehiscens Craib) : Pieng-pien Hsien, H. 7. Tsai 61574, Aug. 22, 1934; Fo-hai,
C. W. Wang 74949, July 1936; Che-li Hsien, Sheau-meng-yeang, C. W. Wane 75983,
Sept. 1936; * Che- li Hsien, You-louh-shan, C. W. Wang 78096, Sept. 1936; Jenneyeh
Hsien, Meng- la, C. W. Wang 80681, Nov. 1936; Kwangsi: Shih Wan Tai Shan,
south of Macias. R. C. Ching 7871 (merotype of Schizophragma pile i Hu),
Oct. 15, 1928; Shang-sze District, Shih Wan Tai Shan, near Hoh Lung Village, W. T.
Tsang 22627, July 3, 1933; Shih Wan Tai Shan, Nam She Village, Ww. T. Tsang 24509,
24576, Oct. 22-31, 1934; Chen-pien District, S. P. Ko 56004, Nov. 4, 1935; Kwang -
tung: Ma Hou Ho, Shih Wan Tai Shan, H. Y. Liang 69546, July 11, 1937.

Craig described Mussaenda dehiscens in 1916, basing it on Kerr 2522
from Siam and also Henry 12825 from Yunnan and Balansa 2683 and 2684
and Wilson 13642 from Tonkin. Among these cited specimens, the Henr
number is the only one actually available to me for study. This particular
sheet was also studied by Hutchinson and was then determined by him
simply as Mussaenda sp. in connection with his investigation of the
Chinese species of that genus, published in Sarg. Pl. Wils. 3: 395—400.
1916. Inthe same year Craib published his Mussaenda dehiscens, Craib’s
name was accepted for the Indo-Chinese plants in 1923 by Pitard, who
also published Pierre’s manuscript name Greenea hoaensis in synonymy.
Chun redescribed the species on the basis of Kwangtung material in 1934,
but cited no specimens. Not having seen Craib’s type, and noticing certain
differences between the Kwangtung plant and several particulars of Craib’s
and Pitard’s descriptions, he was somewhat doubtful as to the disposition

of his material. With ample material from the different localities now
available for study, I can only conclude that all the specimens above cited
represent a single species.

1943 ] LI, SCHIZOMUSSAENDA 101

After a critical examination of the material and an examination of all
species of Mussaenda in the herbarium of the Arnoid Arboretum, it is felt
that a separate genus should be established for this species. The relation-
ship of this proposed new genus, Schizomussaenda, to Mussaenda is close,
but the inclusion of the plant in the latter, in view of the manifest differ-
ences, is scarcely warranted. In vegetative as well as in floral characters,
it resembles certain species of Mussaenda. The large petaloid calyx-lobe is
a character common to both genera and increases their superficial re-
semblance. However, this new genus differs manifestly from Mussaenda
in the elongated scorpioid branches of the inflorescence and in its loculi-
cidally 2-valved capsular fruits. The inflorescences of Mussaenda are
strictly cymose, and the fruits are fleshy berries, indehiscent and areolate
at the top. The calyx-teeth are frequently persistent in Schizomussaenda,
while they are usually deciduous in Mussaenda.

Hu placed the plant under discussion in the Saxifragaceae when he
described it as Schizophragma macrosepalum in 1930; it was well illus-
trated by him in 1935. His descriptions are brief, but from the excellent
illustration and from an examination of fragments of the type in the herba-
rium of the Arnold Arboretum, it is possible to determine the status of his
species; it is in all respects the same as Schizomussaenda dehiscens (Craib)
Li. Type material of Emmenopterys Rehderi Metcalf is in the herbarium
of the Arnold Arboretum. All specimens so named by Metcalf represent
Schizomussaenda dehiscens (Craib) Li. Emmenopterys Oliver is mani-
festly different from either this new genus or Mussaenda, notably in its
paniculate cymes, large broad infundibular corolla with imbricate lobes,
rounded ciliate calyx-lobes, large capsular fruits, and winged seeds. The
genus Emmenopterys therefore remains a monotypic one. The distribution
of the single species, EK. Henryi Oliver, represented in this herbarium by
numerous specimens from Chekiang, Anhwei, Kiangsi, Hunan, Hupeh,
Szechuan, Yunnan, Kweichow, and Fukien, is limited to eastern, central,
and southwestern China.

In the scorpioid-cymose inflorescences Schizomussaenda suggests
Greenea Wight & Arn., in which genus Pierre placed it. Greenea has longer
and more distinctly scorpioid inflorescence-branches, while its floral and
fruit structures are very different. The fruits of both are two-valved, but
in Greenea they are small, globose, completely and septicidally separating
into two valves, with each valve again splitting at its tip, while in Schizo-
mussaenda the fruits are larger, oblong, and loculicidally dehiscent for only
about one-third their length. In Greenea the flowers are white, the calyx-
tube globose, and the corolla-tube glabrous within, with the lobes twisted
in the bud; in Schizomussaenda the flowers are yellow, the calyx-tube
elongate, and the corolla-tube villose inside and with valvate lobes. In
Greenea one of the calyx-lobes is slightly longer than the others, but none
of them is accrescent, and the large petaloid sepal, so characteristic of
Mussaenda, Emmenopterys, and Schizomussaenda, is entirely lacking.

Summarizing the above discussion, this new genus is related on the one

102 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

hand in its vegetative and floral characters to Mussaenda, differing in the
inflorescences and fruits; in its general facies, except in the scorpioid
arrangement of its flowers, it is strongly suggestive of Mussaenda, in the
latter character suggesting Greenea of the Rondeletieae. Schizomussaenda
is evidently much closer to Mussaenda than to Greenea, but in its dehiscent
capsular fruits it transcends the characters of the Mussaendeae; however,
it seems best to retain it in the Mussaendeae in spite of its dehiscent fruits.
In its combined characters it seems to be remote from the Rondeletieae.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

1943 | BALL, SALIX FLORIDANA CHAPMAN 103

SALIX FLORIDANA CHAPMAN, A VALID SPECIES
CARLETON R. BALL
With one plate

Tue purpose of this paper is to bring together the synonymy of Salzx
floridana Chapman and to discuss the various treatments of this species.
An amplified description, based on all available collections, is given.

Salix floridana Chapman, Flora Southern U. S. 430. ne

Salix Chapmanii Small, Man. Southeastern ee: 14.

Salix astatulana Murrill & Palmer in Jour. Arnold Pe 32; 580. 1941.

In 1860 Dr. A. W. Chapman of pea described “Salix Floridana,
n. sp.” (1. c., above), and he repeated the description without change on
page 430 of Edition 2 (1883) and page 453 of Edition 3 (1897), as follows:

*“S, Floridana n. sp. Leaves ovate-lanceolate, acute, smooth above,
glaucous beneath, finely serrate, rounded at the ae the petioles pubescent ;
stipules small, caducous: fruiting ament oblong, dense; capsule ovate-
lanceolate, smooth. — Rocky banks, West Florida, fruiting in April. —
Shrub 3°-12° high, Leaves thin, 2’ 3 long. Fruiting aments 2’—3’ long, 1’
in diameter, enveloped in the copious wool of the seeds. Flowers not seen.”’

The leaf description (‘“ovate-lanceolate, acute, . . . rounded at the
base . . .”), taken with that of the glabrous capsules, indicates a willow
unlike any other species known in the southern states. Apparently the
species was not collected again in the 41 years from 1860 to 1900, inclusive.

With the twentieth century, there began a second and more confused
41-year period for S. floridana. In 1903, Small described it (Flora SE.
U. S. 342; also ed. 2, 1913) much as Chapman had and credited it, for
some reason, to middle as well as western Florida. Chapman, lacking
staminate specimens, made no attempt to indicate relationships. Small
keyed it as having two stamens and placed it next to S. cordata.

In 1902, Dr. Roland M. Harper collected his no. 1381 in Pulaski Co.,
Ga., and, in 1938, his no. 3634 in a ‘“‘semi-calcareous swamp of small creek
about % mile east of Cedar Springs, Early Co., Ga., April 11, 1938.”
Both numbers represent full-grown foliage and were distributed (usually
2 sheets of 1381) as S. floridana. The writer has seen these collections in
each of five herbaria (Field Museum, Gray, Mo. Bot. Gard., N. Y. Bot.
Gard., and U. S. Nat.). One sheet of 1381 carries vigorous shoots with
enormous leaves, reaching a maximum size of 6 & 18cm. Early in 1904,
Dr. Harper published notes (Bull. Torrey Bot. Club 31: 21-22. 1904) on
his no. 1381, but he does not mention the larger leaves, lacking on Chap-
man’s fruiting specimen.

“Salix Floridana Chapm. I refer tentatively to this almost unknown
species specimens collected in wet woods at the outer (eastern) base of the

104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

sand-hills of the Ocmulgee in Pulaski County below Hawkinsville, June
27 (no. 1381). My specimens were slender erect trees about ten feet tall
and two inches in diameter, with tasteless bark roughish below and smooth
above. The twigs were very brittle not only at the bases but ee
No trace of fruit was found, but my material, as far as it goes, seems to be
a perfect match for the type-specimen in the Torrey Herbarium, collected
by Dr. Chapman in West Florida.

In 1913, Small published two more small volumes, Shrubs of Florida and
Florida Trees. Salix floridana Chapman was described oo in both
(pages 9 and 13, respectively), and in the second Small adds: “. . . found
in southern Georgia.” He probably refers to Harper’s collection i in Pulaski
County, which is just south of the center of the state.

Schneider, in January, 1918 (Bot. Gaz. 65: 21) lists S. floridana as a
synonym of S. longipes Shuttleworth. In 1919, he says (Jour. Arnold Arb.
1: 25) that S. longipes “has been again described by Chapman (1860) as
S. floridana and by Small (1913) as S. amphibia.” He gives no reason for
his conclusions and his annotation on the Torrey Herbarium isotype is dated
1919, after the above-cited papers were printed. Neither the Gray
Herbarium isotype nor the Arnold Arboretum sheet of Harper 1381 are
annotated at all by Schneider.

In 1933, Small (Man. SE. Flora 414) described S. Chapmanii in essen-
tially the same words previously used in describing S. floridana, and on
p. 1504 he says: “Type, Middle Fla., Chapman, in herb. C(olumbia)
U(niversity).” This specimen is the type of S. floridana. In this Manual,
Small lists S. floridana as a synonym of S. longipes. He nowhere explains
how the type of a synonymous species can be made the type of a later new
species. Small placed his species in the Cordatae, with two stamens.

In 1941, Murrill and Palmer described and figured Salix astatulana
sp. nov.,! discovered by Dr. Murrill in Lake and Levy Counties of Central
Florida. The authors state that their species ‘“‘appears to be most nearly
related to Salix longipes’’ but is “‘a very different plant in its foliage,” which
is true also of S. floridana Chapman. The description of leaves, aments,
and fruits closely parallels the description of S. floridana Chapman

The writer borrowed the pistillate and staminate types from the Arnold
Arboretum, and portions of the isotypes were generously contributed by
Dr. Murrill, of the University of Florida. ‘The types consist of several
twig fragments plus detached mature leaves obviously from other plants or
branches. The flowering specimens represent the juvenile stages of S.
floridana, as the Chapman type and the Harper collections represent
maturity.

The following amplified description is drawn up from all available
authentic material, including two isotypes of S. floridana (and therefore
of S. Chapmanii) from the New York Botanical Garden and the Gray
Herbarium, Harper 1381 and 3634 from the five herbaria named above,
and the types of S. astatulana from the Arnold Arboretum and fragments

1Murrill, William A., and Ernest : ee A New Willow from Florida. Jour.
Arnold Arb, 22: 580-581. 7 fig. Oct.,

1943 ] BALL, SALIX FLORIDANA CHAPMAN 105

of the isotypes from the University of Florida. The type specimen of
S. floridana cannot be located at the Missouri Botanical Garden.

Shrub or small tree, 2—4 m. tall; branchlets brittle (Harper), the older
yellowish brown or grayish brown and sparsely pubescent or glabrous, the
younger dark brown to blackish (as dried) and more or less pubescent;
bud-scales 2-5 mm. long, colored and clothed as the branchlets; stipules,
on vigorous shoots, 5-10 mm. long, semilunate, glandular- serrulate. eaves
lanceolate, broadly lanceolate, or ovate- lanceolate, mostly 5-12 cm. long
and 2—4 cm. wide, on puberulent or pubescent petioles 0.7-1.5 cm. long
(not 4 cm. as indicated in text-figure of S. astatulana); the lowest leaves
on a given branchlet ovate, obtuse, 1.5-3 cm. long; the next higher or
median leaves oval or elliptic-oval or somewhat obovate, obtuse to acute,
5—7 cm. long by 2.5—3 cm. wide; the remainder or normal leaves lanceolate
to broadly lanceolate, acuminate, and 5—12 cm. long by 2.5—4 cm. wide or,
on vigorous shoots (Harper 1381), broadly lanceolate and acuminate, 12—16
cm. long and 4—5 cm. wide, or the largest broadly elliptic and acute, 17-18
cm. long and 6 cm. wide, on puberulent petioles up to 2.5 or 3 cm. long; all
leaves rounded to truncate to somewhat cordate at base, glandular-crenate-
serrulate on the margins (serrulations averaging 3.5 to 7 per cm. on smaller

and 2.5—4 per cm. on larger leaves), sometimes with small lobes near the
eee of the "Bide (as in S. pseudomonticola Ball) or with solitary glands or
isolated pieces of glandular foliaceous tissue (remnants of once basal lobes)
gis a m. long, on the sides of the petiole 3-7 mm. below the blade and
in me usually densely pubescent while unfolding, becoming pro-
gressively ees to glabrate with age; the yellowish midribs prominent

eath and usually remaining finely pubescent above, especially toward
‘he ae Sales beneath, becoming dark green above in age, the greenish
yellow secondaries and tertiaries forming a coarse raised network on both
surfaces as in S. discolor but never forming the fine flat mosaic of vein-
islets so characteristic of S. longipes, S. amygdaloides, S. nigra, and related
species.

Aments coetaneous, leafy- pedunculate, 3.5-6 cm. long, 1.5 cm. wide in
flower, the pistillate 2—2.5 cm. wide in fruit; peduncles 0.5—1 cm. long,
pubescent, bearing 1-3 small leaves 1-3 cm. lon ng; flower-scales oblong-
obovate, about 2 mm. long, yellowish to yellowish brown, rounded or
sometimes truncate at apex, glabrate outside except on margin and base,
densely villose inside; stamens 3 or 4 to 6 (sometimes only 2), filaments
5-7 mm. long, sparsely pilose at base; capsules ovate-lanceolate, 5-8 mm.
long, glabrous, brown at maturity, the style 0.1-0.2 mm. long, bifid, the
stigmas very short, the pedicel 2-4 mm. long, sparsely pilose, becoming
glabrate; gland 1, ventral, 0.6—1 mm. long, stout, cylindric.

Salix floridana has lanceolate to ovate-lanceolate or broadly elliptic
leaves with coarse venation, while the leaves of S. longipes are linear-
lanceolate to narrowly lanceolate and with a very fine mosaic of vein-islets.
The capsules of S. floridana are larger, the stamens fewer, and the flower-
scales more glabrate. It is not certain that S. floridana is most nearly akin
to S. longipes and other members of the Section Bonplandianae, It prob-
ably belongs to the Section Triandrae, represented by S. triandra L. of
Eurasia and other species.

106 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

In any case, S. floridana is a species apparently on the verge of ex-
tinction. The type locality (per Gray Herbarium isotype) is Marianna,
Jackson Co., northern Florida. It has been found once in each of two
counties in southern Georgia, and in two counties of central Florida.
Dr. Harper says the Pulaski County location probably has been destroyed.
The species is relatively rare, if existent, at the other localities mentioned.

EXPLANATION OF PLATE

Salix floridana Chapman; isotype in Gray Herbarium.

U.S. DEPARTMENT OF AGRICULTURE,
WASHINGTON, D. C.

Jour. Arnotp Ars. Voi. XXIV PLATE I

2! asl ale! als

tia
ly

tobatadedatadelbedadntS fad

bithe

shuld

Lal

ea a

|

9

R R. Ba ase
|
ix flori ApRAn
read. pf if -
avidently an isotype with the label 43

; se ate 4
ah owing re type locality "Uarianna® :

ich is in Jacicson Sounty acninst the s

Toa and Alabaua State boun laries. ee

Stig Cote Re BU ae

SALIX FLORIDANA Chapman

|

1943] ASMOUS, KARELIN AND KIRILOV 107

KARELIN (1801-1872) AND KIRILOV (1821 - 1842)
EXPLORERS OF SIBERIA AND MIDDLE ASIA

VLADIMIR C, ASMOUS

On September 11, 1842,* there died in a small town, Arzamas, in central
Russia, after a short illness, a student of St. Petersburg University—Ivan
Petrovich Kirilov. He was on his way by stage from Siberia to Moscow,
and in the absence of any relatives the police buried his remains; his death
was unnoticed, and it seemed that he was forgotten by everybod

One hundred years have passed, but the name of this youth still lives
in the annals of the history of science. It is true that his name is usually
connected with that of his senior partner, the more widely known explorer
and botanist, Grigorij Silych Karelin. Their joint contributions to science
are great and their famous expedition of 1840-45 into the Altai, Dzungaria,
and Semirechensk regions is justly considered one of the most fruitful and
important of the nineteenth century. Without exaggeration it may be said
that they discovered and explored an enormous region for science.

These two members of one of the most perfect teams of botanists were
very different men by reason of their age, social position, and temperament,
and they met by mere chance. Karelin came to St. Petersburg in 1837 and
there met a modest, lonesome, and apparently very poor student. Talking
with him, he was struck by his intelligence, seriousness, and deep knowledge
of botany. He recognized in this boy a future scientist of great promise,
and, being a man of generous heart and quick impulse, he immediately took
him under his protection.

“Don’t be surprised, my dear,” he wrote to his wife, “I am bringing
with me a young man, whom I want you to take into our family as our son.
He is I. P. Kirilov, an orphan, a poor devil, a great lover of botany, and a
student at St. Petersburg University. He is like a fair girl in character but
is unusually clever...’ From that time “Vanichka” was a member of
the hospitable family of the Karelins and followed his “daddy,” as he
affectionately called Karelin, in all his travels. A boundless enthusiasm
for the study of nature united these men.

Very little is known about the short life of Kirilov. He was born in 1821
(or 1822) in Yalturovsk in Siberia, where his father was a chief of police;
he studied in Tobolsk and later in Irkutsk, where he met the well known
botanist N. S. Turczaninow, who instructed him in botany from 1835 to
1837. He collected with Turczaninow on the southern shores of Lake
Baikal in 1835, and the next year he made another expedition into the
western Baikal Mountains. In the preface to his Flora Baikalensi-Daurica,
Turczaninow says (p. 20): “Mr. Kirilow a rempli cette mission avec un zéle

*All dates in this paper are according to the Julian calendar (old style), which was
in general use in Russia before the revolution of 1917.

108 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

et un succés remarquables, ayant trouvé plusieurs especes échappés a son
prédécesseur.” In 1837 Turczaninow brought Kirilov to St. Petersburg
and helped him to enter the University.

Karelin esteemed his young assistant very highly. “Best regards to you,”
he wrote to A. V. Richter, “from I. P. Kirilov, the most industrious, most
energetic, and the smartest of all students that I have ever met in this
world.” In the expedition of 1840-45, Kirilov was in charge of botanical
collections; he made almost all the identifications, and the well known
Enumeratio plantarum was written by him under the supervision of Karelin
and Turczaninow. There is no doubt that in the untimely death of Kirilov
science lost a botanist of unusual promise.

The life of Grigorij Silych Karelin is more colorful than that of his young
assistant. He was born in January, 1801, in Petersburg Province, a son of
the conductor of the concert band, but he lost his parents when he was
eight years old. He was placed by his elder brother in the First Cadet
Corps, from which he was graduated with distinction in 1817 as a second
lieutenant of artillery. He was assigned to the Office of Military Settle-
ments, but his chance of making a brilliant military career was ruined when
he wrote some derisive verses, supplemented by a cartoon, about the sec-
retary of war, Count Arakcheev, a very powerful man in the last years of
the reign of Alexander I. This became known to the “dreadful count,” and
Karelin was sent on February 20, 1822, into virtual exile to the small town
of Orenburg on the southeastern border of Russia.

This exile seemed to be a disaster, but actually it gave Karelin a chance
to become a scientist and explorer. Deprived of the comfort and gay life of
the capital, he turned to the study of natural history, to which he always
had a strong inclination. He was fortunate in finding an excellent teacher
in the person of E. F. Eversmann, a future professor of natural history at
Kazan University. Due to his extraordinary abilities, Karelin became in
a comparatively short time an expert in botany, zoology, and mineralogy,
and soon he started field work in natural history. He corresponded with
many outstanding Russian botanists, such as Ledebour, Fischer, Meyer,
Bunge, and Turczaninow, and sent them specimens from the Caspian region.
His collections of Caspian plants are preserved in the herbarium of the St.
Petersburg Botanical Garden.

Besides his natural history studies, Karelin was engaged in a number of
other activities which prove his unusual versatility. Fortunately two con-
secutive governors of the Orenburg region, Count Sukhtelen and Count V.
Perovsky, both capable administrators, recognized the outstanding abilities
of Karelin and utilized them fully in the development of that half-civilized
part of Russia.

Karelin made surveys and descriptions of various Caspian regions, ex-
plored mineralogical resources, and made a study of steel and munition
plants in the lower Ural region. In 1826 he was permitted to resign from
the military service. After that, he made a journey with Professor Evers-
mann into the Kirghiz steppes (the so-called Bukeev Orda) and made a
good map of that region. In 1829 he accompanied the expedition of the

1943 | ASMOUS, KARELIN AND KIRILOV 109

Norwegian Professor Christopher Hansteen and Lieutenant Due for as-
tronomical and geodetical observations in the lower Volga and the Ural
regions and travelled in Bashkiria and Orenburg and Perm Provinces. In
1831 he explored the basin of the Tobol River. He also successfully man-
aged intricate diplomatic relations with semi-independent khans of the
Kirghiz steppes. It may be noted here that Karelin was very popular with
nomadic tribes of the Caspian region. He possessed all the qualities which
appeal to the imagination of oriental peoples: he was a good looking, power-
fully built, energetic man, a skillful hunter and horseman, and, although he
was generous and mild in dealing with other people and had a very gay
and pleasant disposition, he was always ready to defend himself vigorously
against any aggression. He inspired not only love but also respect and fear.

In 1831 Karelin was employed by the Asiatic section of the State Depart-
ment and soon started a series of his explorations of the Caspian Sea regions,
which separated him from his family for months at a time, for he had been
married some years and was the father of two daughters at that time.

In 1832 he was at the head of an expedition including four ships and a
detachment of one hundred and seventy Ural Cossacks. He explored the
eastern shores of the Caspian Sea, made a number of good maps, and pre-
pared descriptions of the adjoining regions.

Karelin in 1834 commanded another expedition in the same region, with
a special mission to build a fortress on the eastern shore of the Caspian Sea
in order to protect Russian settlements from the raids of nomadic tribes
of the Kirghiz steppes and Chiwa. After careful investigations he selected
Kara-su Bay for that purpose and on May 2, 1834, founded a fortress
named by him Novo-Aleksandrovsk, which was completed, supplied, and
garrisoned in less than three months. This enterprise was not only a com-
plete military success, but it did not disturb the relations between Russia
and Chiwa, because diplomatic parleys conducted by the emissaries of
Karelin with the Khan of Chiwa prevented the bloodshed which was feared.

In 1836 Karelin was charged with the exploration of the eastern and
southern shores and the islands of the Caspian Sea. He corrected the maps
of these regions, travelled far inland, explored the steppes of Turkmenia,
and made a survey of the old estuary of ancient Oxus or Amu-Darya.
This particular map is a very valuable contribution to the science of
geography. Finally he visited the shores of Asterabad Bay and laid the
groundwork for the establishment of regular trade relations between Russia
and Persia.

All these varied activities never prevented Karelin from continuing his
scientific researches; he prepared extensive botanical and zoological col-
lections, while his diaries include valuable data on the flora and fauna of
the regions visited. His travels had made the name of Karelin widely
known and he was elected to membership in many scientific societies. The
Russian government recognized his services and rewarded him with a grant
of 6,000 roubles and a pension of 900 roubles a year. Emperor Nicholas I
invited him for a lunch and gave him a valuable diamond ring as his per-
sonal gift.

110 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

No wonder that in 1839 the oldest scientific society in Russia, the Société
de Naturalistes de Moscou, invited him to take charge of a scientific expe-
dition for the exploration of the Altai, Tarbagatai, and Sayan Mountains,
and Dzungaria. Karelin willingly accepted and soon an agreement was
reached as to the itinerary of the expedition and its financial support.
Karelin was to receive 3,500 roubles a year plus all expenses of the expe-
dition. Besides that, he retained his regular salary from the Finance De-
partment, where he ‘had been employed since 1838. He was to collect
botanical, zoological, mineralogical, and geological specimens, ship his col-
lections every two months, and submit a monthly report. The agreement
was for a period of two years. It may be noted here that, although the
Sayan Mountains were included in the itinerary of the expedition and there
are some references in literature concerning Karelin’s travels in this region,
neither he nor any of the members of his expedition ever visited that region.
The researches of Lipsky, Romanovsky, and Lipschitz and the study of
Karelin’s diaries prove this conclusively.

After careful preparations, which took several months, Karelin started
his famous expedition on March 15, 1840, from Orenburg, going eastward
through Troitzk, Petropavlovsk, and Omsk to Semipalatinsk. In addition
to Kirilov, the party included a Cossack officer, G. A. Maslennikov, a very
skillful hunter, and G. D. Karaulov, a zoological collector and preparator.

I make no attempt to give a detailed itinerary of Karelin’s travels, but
will merely summarize his explorations for each year. Fortunately he
wrote many letters, sent his reports regularly to the Society of Naturalists
of Moscow, and kept a diary of his travels. These letters and some of his
reports are samples of brilliant, colorful scientific language. It is to be
regretted that Karelin published so little and that all his manuscripts were
destroyed by a fire in 1872.

In 1840 Karelin explored the Tarbagatai and Narym Mountains and the
Semipalatinsk region. Kiriloy made independent explorations to Lake
Nor-Zaissan and to the northern slopes of the Tarbagatai Range. The re-
sults of the first year were very good. Fifty-two boxes of collections, in-
cluding 38,000 botanical specimens representing 1,127 species, were sent
to Moscow.

In 1841 the party explored the Alatau Mountains and the Semirechensk
region. Besides his regular staff, Karelin had about fifty Cossacks who
served as a convoy to the expedition, but he trained some of them as col-
lectors and hunters. Of these men we should mention especially Captain
S. M. Abakumov, who became a proficient botanical and ornithological
collector. The results of the second year were even better than those of the
first, although the expedition was many times attacked by “barantachi”
(rebellious Kirghizs). About 55,000 botanical specimens were sent to Mos-
cow. In the fall Kirilov was sent with some collections to Krasnoyarsk,
where he worked under direction of Turczaninow on the preparation of the
Enumeratio plantarum. The correspondence between Karelin and Turc-
zaninow proves that the latter contributed greatly to this work, which is
considered one of the best in Russian botanical literature. It lists 1891

1943 | ASMOUS, KARELIN AND KIRILOV 111

species and includes descriptions of eight new genera and 220 new species.
Lipsky is of opinion that the influence and constructive criticism of Turc-
zaninow, the best Russian botanist of that time, added very much to the
value of this work.

In 1842 Karelin was again in the Alatau Mountains, in eastern Altai, and
in Dzungaria (Bayan Aul and Karakalinsk regions). All his diaries of this
year are missing; probably they were lost with other papers (letters, reports,
surveys) which Kirilov had with him when he died so suddenly at Arzamas,
apparently of cholera. This tragedy, which deprived the expedition of a
key assistant, was a severe personal blow to Karelin, who loved Kirilov no
less than his own children. Sophia G. Karelina says in her recollections
that her father was so distressed on receiving the news of Kirilov’s death
that he became hysterical, tried to drive his head against the wall, and his
companions had to keep him under close watch for many days, fearing that
he might commit suicide. After this violent period he became apathetic
and morose and was not able to work. This depressed condition continued
for some months but gradually passed. After the death of Kirilov, the
work on the identification and description of the plants, which had been
done mostly in the winter months, stopped. The botanical collections of
1842-44 were partly described later by S. S. Shchegleev.* It is a peculiar
fact that after the death of Kirilov, Karelin never published a botanical
paper in the remaining 30 years of his life (1842-72). This does not of
necessity mean that he suddenly lost interest in botany. Although he is
usually considered as a zoologist par excellence, this is hardly correct. At
least he considered himself a botanist in the first place, and his earlier
publications, his correspondence with the best Russian botanists of the time,
and especially his diaries, where botanical observations always play a
dominant role, prove that he was right in that respect. It must be borne in
mind that he was manifestly an outdoor type, a man of action, a born
traveller and explorer, and not a cabinet worker. He started to write
regularly when he was over fifty years old and partially incapacitated as a
traveller by a severe attack of rheumatism.

In 1843 Karelin made an expedition through the Kokbetinsk region to
Lake Nor-Zaissan, described it, and explored the Upper Irtysh for 400 miles
from its source to Ust’ Kamenogorsk. There are no data on the number
of specimens in the collection; it is known only that it contained 1678
species. Maslennikov made a separate expedition to the sources of the
Lepsa River in the Alatau Mountains, where he collected zoological and
botanical specimens.

*Shchegleev, S. S. [Stschégléew, S. S.]
1854a. as as ad Floram Altaicam. Dissertatio. i-ii, i-iv, 1-119.
e chapter is in Latin, the rest in Russian; supplement to Karelin &
Kirilov 1841 & 1
1854b. peers ie a la flore Altaique. Bull. Soc. Nat. Moscou 27(1):

Li
“t jel of a chapter in Latin in the original work, with a preface
in Frenc

112 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Karelin explored the basin of Buchtarma River in 1844 and travelled in
Dzungaria and in the border regions of northeastern China. Maslennikov
again visited the Alatau Mountains and collected plants in the valleys of
the Lepsa, Djamantash, and Tentek Rivers. There are no data on the size
of the botanical collections assembled.

The expeditions having continued for almost five years, Karelin en-
thusiastically insisted that the work should be prolonged for at least one
year more. This suggestion met a very unfavorable reception from all sides.
The Society of Naturalists of Moscow helped Karelin in many ways and
gave its moral support, but, after spending about 18,000 roubles on the
expeditions, it discontinued financial aid in the middle of 1842. The only
source of money was Karelin’s salary, but even this meager supply became
uncertain. The Secretary of Finance, from whose department Karelin was
drawing his regular salary of 3,000 roubles, frowned on his scientific ac-
tivities and advised him many times to return to his regular work in St.
Petersburg, but Karelin always refused on some pretext. The long con-
troversy came to a sudden end when the Finance Department asked Prince
Gorchakov, Governor of Western Siberia, to induce Karelin to return to
St. Petersburg. The Governor, who had a very strong personal dislike of
Karelin, was very willing to do this. He provided a police escort and sent
him out of Siberia in June, 1845. This was the sad end of the famous
expedition. How it could happen that the famous explorer, honored per-
sonally by Emperor Nicholas I, was deported in such a humiliating way,
by an arrogant provincial official, is very strange. Lipsky, the best biog-
rapher of Karelin, hints vaguely that there must have been some less
obvious factors. A very careful study of all available documents relating
to this deportation does not support this allegation, and we do not feel
justified in making further speculations on that subject. As an aftermath
of this unhappy affair, Karelin was retired from the Finance Department
on November 14, 1845.

There are no complete data on the collections of Karelin and Kirilov,
which are considered the richest ever made in that formerly almost unknown
region. The dossier of that expedition, which was at one time in the
archives of the Society of Naturalists of Moscow, was lost. Fortunately
many copies of letters, reports, and some manuscripts, which were in the
possession of Karelin’s daughters, were saved. In one of his letters to
Count Stroganov, Karelin gives the following data on his collections for
three years (from May 9, 1840 to March 16, 1843): animals — 240,
birds — 1669, insects — 9766, fishes — 8, snakes — 34, lizards — 169,
plants — 90,142, seeds — 442, and minerals — 474. It should be noted
that the number of plants in this particular list is apparently wrong; it
must be larger. The number of botanical specimens for the first two years
alone (1840-41) totaled 93,000. These collections are incorporated in the
herbaria of Moscow and St. Petersburg Universities, the St. Petersburg
Botanical Garden, and the Academy of Sciences. Many thousands of
duplicates were distributed to the principal herbaria of Europe, as well
as to some in America, thus making Karelin’s expedition widely known.

1943] | ASMOUS, KARELIN AND KIRILOV 113

After his forced retirement, Karelin joined his family, which had moved
in 1842 from Orenburg to his small estate in Trubitsyno, near Moscow.
Here he spent six years in the company of his wife and four daughters,
working on his collections. It was apparently a happy period of his life,
because the recollections of his daughters give us a picture of him as a
very devoted father and husband and a man of gay, pleasant disposition,
who loved rural life and its simple pleasures. But still his inborn longing
for the study of nature and new places was stronger than his family ties.
In 1852 he wrote to Count Perovsky, then the Governor of the Orenburg
region: ‘For six years I have been living in Moscow and its vicinity and I
feel now an irresistible desire to travel once more. I want to make a
short trip this summer through the Ural steppes to the Caspian Sea. My
Sua is to explore the Inder Lake and the northern shores of the Caspian

. .” He asked the permission of Perovsky to make this trip, and,
pecan being granted, Karelin departed on July 20, 1852, telling his
daughters: ‘In six months, children, I shall be back, and now good-bye.”

He never came back. He lived for twenty years a virtual hermit in the
small town of Guriev, travelled occasionally in the Ural Cossack region,
made some ornithological observations, and worked very assiduously on
the description of his travels. It is known that he prepared for publication
eleven large volumes of manuscripts, but lack of financial support pre-
vented their publication. He suffered very much from severe rheumatism
and was bedridden for the last two or three years of his life. His house
and all his manuscripts and collections were destroyed by fire in the summer
of 1872. This was too much for an old and sick man, for he failed to
survive this final blow and soon died, on December 17, 1872

Some facts regarding Karelin’s life are strange and can hardly be ex-
plained in a satisfactory way, especially his self-exile to Guriev. The
persons closest to him, his daughters, frankly confess that they were at a loss
to find any reasons for the sad, solitary end of his life. There is no doubt
that Karelin was a very complex human being and not as well balanced and
happy as he seemed to be. The only personal comment regarding this ast
period of his life is the following excerpt from one of his last articles
written in Guriev in 1868: “In 1852 I came for a short visit to the estuary
of the Ural River, in order to make some observations on the migration,
breeding, and moulting of birds. But I found such a wide field for other
scientific explorations and such peace and freedom, which I needed for
the arranging of the extensive materials collected on my Blo that
instead of two years I stayed here more than sixteen years . . .

This and another sentence in Karelin’s letter of June 30, 1841, to Count
Stroganov give us some hints for the better understanding of the character
of this remarkable man. In that letter Karelin wrote: “My expedition is
a great success. I am writing to you from the height of 9,000 feet, in the
neighborhood of icebergs and eternal snow, but surrounded by such wonder-
ful and rare flowers that some days I walk hatless in an ecstasy of
reverie...” It seems to us that nature was the only great and real love of

114 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

this enthusiastic scientist. On the altar of this passionate, all consuming
love he sacrificed his happiness and his life.

BIBLIOGRAPHY OF WORKS BY KARELIN AND KIRILOV

ELI, G. S.
1829. pera gee des Coléopteéres ~ par Mr. Karélin dans les steppes des Kirguises,
ntre le Volga et l’Oural. Bull. Soc. Nat. Moscou 1: 169-170.
1833. On the seal trade in the northeastern part of the Caspian Sea
In Russian; reprinted in Zapisk. Russk. Geogr. Obshch. Geogr. 10; ne 153.

1883.
1839. Enumerats plantarum quas in Turcomania et Persia Boreali legit. Bull. Soc.
Nat. Moscou 12: 141-177.
A eae arranged list of 973 speci
1840. Lettres de Mr. Karéline, voyageur de la Société, Bull. Soc. Nat. Moscou 13:
495- ae

Three letters containing notes on the vegetation,
1841a. Pers et Suchtelenia, ae nova aectaiti a G. Karelin descripta.
Bul c. Nat. mescirs ri 7. pl. 1
ves of Turcoma
1841b. Travels of Mr. Karelin i in he altal and Sayan Mountains (sic!). Otechestven-
yia Zapiski 16 (Sect. ““Smies 1-4.
ch Russian; this anonymous article i is an excerpt of the next entry; no men-
of Sayan Mountains in the text.
1841c. Vosare de Mr. a Bull. Soc. Nat. Moscou 14: 559-573.
Six letters of Kare

1842. Herrn Karelin’s nen, zur sen: des og ge — Altai und der
Sajanischen Gebirge. Arch. Wiss. Kunde Russland 2: 399.

German translation of some reports and a a Kardin, with some
additional notes.

1843. Turkmenien oder das Land in Osten des oy es Meeres. (Nach Karelin’s

tizen). Arch. Wiss. Kunde Russland 3:
: full German translation by W. Schott te ani 2 of Karelin’s work which
was published in Russian 40 years later; see Karelin 1883.

1847. Extrait du journal d’un voyage, fait en Djoungarie ou Sungarie par Grégoire
Karélin en 1841. In Jubilaeum semisaecularem doctoris ce et philoso-
phiae Gotthelf Fischer de Waldheim celebrant sodales Societatis Caesareae
Naturae Scrutatorum Mosquensis die x/xxii Februarii anno 1847. Art. 2, 1-17.
pl. 1, 2. in folio.

Description of two days’ travel in the Alatau Mountains, with spear eee
of mae and illustration of Waldheimia eget and A plotaxis a
a reviewed by V. Lipsky in his Fl. Asiae Med., pp. 55-58. 1902

1867-— a fot of the article by A. Riabinin on ore — a the ie
Cossack reine. Uralskiia Voiskovyia ara ti 1867: 43: 3-7, :
5-7, no. 52: 6-9; 1868: no. 1: ee no. 7: 4-7, no. 10: Pe . no,:.212 6. no,
22: 2-6, no. 51: 6-9, no. 52: 6-7, no. 53: tie it.

In Russian; for amplification see ene entr

1875. Review of the article by A. Riabinin, “Natur al products of the Ural Cossack

“ice excerpted from his in “Ural Cossack region,” 2 pts. 1866. Trav.

Soc. Nat. St. Petersb. 6: 186-29
In ae aggre se jeans by Prof. A. N. Beketov and
_N. Bogdanov. Ani tant work with a quite misleading title; includes

enumeration of minerals, ee (3 lists), od animals and birds (345 species),
mposed by Karelin.
1883. Travels of G. S. Karelin in the rai Sea. Zapisk. Russk. Geogr. Obshch.
Geog

Edited by Prof. M. N. Bo ‘ase Descript ion in Russian of Karelin’s ex-
peditions of 1832 and 1836; includes an enumeration of 280 plants.

1943] ASMOUS, KARELIN AND KIRILOV 115

1889. A journal of scientific investigations of part of the Kirghiz Steppes made by
ouncilor Karelin. Izv. Russk. Geogr. Obshch. 25: 503-512.
fa Russian; a copy of a report by Karelin on his exploration of ‘the sources
of the Tobol ives in 1831, under command of Col. Gens.
1905. A bole autobiographical note. In V. Lipsky, Fl. Asiae Med. pp. 599-601.
Composed on January 5, 1852, in Tarusovo. found and ibaa by Lipsky.
KareELin, G. S., and I. P. Krrimov.
1841. Enumeati plantarum anno 1840 in regionibus Altaicis et confinibus collec-
arum. Bull. Soc. Nat. Moscou 14: 369-459, 703-870.
Reieediakads of 959 species, including two new genera, 88 new species, and
e new varieties.
1842a. Tetras generum plantarum novorum, “7 nk Compositarum Rossiae indi-
gen . Bull. Soc. Nat. Moscou 15: 128.
Cancrinia » Waldheimia, Richteria, a de gen. nov. from the
Alatau Mountai
1842b. Enumeratio Sern um in desertis Songoriae orientalis et in jugo summarum
Alpium rae — 1841 collectarum. Bull. Soc. Nat. Moscou 15: 129-180,
321-453, 503
eda 932 species, including description of 6 new genera (Pachy-
pteris, Spirorhynchus, Cryptospora, Bryomorpha, Ammolirion, and Henningia)
and 132 new species and varieties.
1843. Nova plantarum genera Rossiae indigena. Ann. Sci. Nat. II. Bot. 19: 311-316.
Includes description of some new genera and species from Dzungaria; for
amplification see Karelin and Kirilov 1842b.

BIOGRAPHIES OF KARELIN AND KIRILOV
ANONYMOUS.
1897. Karelin, Grigorij Silych. Russkil Biograficheskii Slovar’ [8]: 521-522.
BEKETO EAN:
1873. ont note on Karelin. Trav. Soc. Nat. St. Pétersb. 4:
1895. Karelin, Grigorij Silych. Entsikl. Slovar’ Brockhaus & Efron cf 492.
Bocpanov, M.N.
1875. Survey of the expeditions and natural history investigations in the Aralo-
aspian region. Trudy Aralo-Kaspiisk. Ekspedits. 1: 1-53.
In Russian; oo biographical notes and data on the expeditions of

Karelin, pp. :
1883. Brief tei sketch of Karelin. Zapisk. Russk. Geogr. Obshch. Geogr.
10 in preface.
See Karelin 1883, in preceding bibliography of his works.
Borop1n, I. P.
1908 Collectors and collections of the Siberian flora. Trav. Mus. Bot. Acad. Sci.
—245.
Includes biographical notes and data on collections of Karelin — pp. 42-43,
and Kirilov — pp. 46-47.
HAnsTEEN, C.
1854. Reise-Erinnerungen aus Sibirien von Prof. Christoph Hansteen. Deutsch von
Dr. N. Sebald. i—viii, 1-215.
Translation of several articles published in Norsk Folke-Kalender 1849-55,
describing the author’s travel in Siberia; for biographical data on Karelin see

chap

1857. Souvenirs d’un voyage en Sibérie par Christophe Hansteen .. . Traduits du
norvégien par M-me Colban et revus par MM. Sédillot et Da: la-Roquette.
I-xv, 1-428. map

116 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Translation of the articles published in a Folke-Kalender 1849-57; pp.
315-373 concern Karelin and his explora
1859. rye Erindringer af Christopher erg i-xviii, 1-387. : ;
This is a revised and more complete edition of the tens s “Recollections” ;
chapters 25 and 26 include biographical notes and data on Karelin’s explora-
tions.

Karetina, 5. G.
1873. Russian man in the eastern border-land: Sketch of the life and the activities
we S. Karelin. Russkii Arkhiv 1]: 1318-1344, 01313-01319.
very va — biography by the daughter of Karelin; reprinted in V.
a, 1905, pp. 609-630, with addition of pp. 637-651
KarELINA, S. G., E. G. BEKETova, and A. P. Bocpanovy.
1889. Grigorij Silych Karelin. Izv. Obshch. Estestv. Antrop. & Etnogr. 57: 1-4.

Reprinted in V. Lipsky 1905, pp. 604-608; first two authors are the daugh-
ters of Karelin
KoseEko, D. F
1890. Tavel of Karelin in oo aeaiiad Sea in 1836. Zapisk. Vost. Otd. Russk.
Arkheol. Obshch. 5: 79-
In Russian; eiehiigs some references to the vegetation. Reprint 1-6.

1890.
LINDEMANN, E.
1884. Dritter Bericht iiber den Bestand meines Herbariums. Bull. Soc. Nat. Moscou
Zz 65-312.

Slides brief pane data on Karelin under no. 366 (p. 309), and on
Kirilov under no. 383 (p.

Lipscuit7z, S. J.
1940. Société de naturalistes de Moscou (1805-1940). Apergu historique. 1-131.
—4|. 14 portraits.
A historical sketch in — including some data on the expedition of
Karelin and Kirilov, pp. 49-55.

Lipsky, V. I.
190Sa. To ee Silych, his life and ene bs V. Lipsky, Flora Asiae Mediae,
Tiflissk. Bot. Sad. 7(3):
ere a ae most exhaustive ke — nexedtin including autobiography
ian his biography by S. G. Karelina, itineraries, hence lene letters,
n extensive a by E. O. Romanovsky ted 1- 190
1905b. Kirilov, Ivan Petrovich. In V. Lipsky, Flora Asiae Media - 3, pp. ee
A biographical sketch in Russian, based partially on a letter of N.
Turczaninow to Karelin

Romanovsky, E. O.
190Sa. A conclusion to the letter of Prince P. D. Gorchakov to Karelin. In V. Lipsky,
perp ereeginer is 3, pp. 726-742
udy in Russian of the letters and a pertaining to the deporta-
of Karelin from Semipalatinsk in S
1905b. “Karelia” 1829-1905. In V. Ae Flora Asiae Mediae pt. 3, pp.
787

y complete bibliography containing (1) works by Karelin, and (2)
papers describing the life of Karelin or based on his works and collections.

1943 ] ASMOUS, KARELIN AND KIRILOV 117

LIST OF PLANTS NAMED IN HONOR OF G. S. KARELIN
Genus: Karelinia Less. (= Pluchea Cass.) in Compositae.
Acantholimon Karelini Bu unge, Alfredia Karelini Led., Ammodendron

; y.,

Astragalus Karelini Fisch., Atraphaxis Karelini Jaub. et
pach, Carex Karelini Meinsh., Chara Karelini Less., Cousinia Karelini Less., Corydalis
Kareliniana Turcz., Echinospermum Karelini Fisch. et Mey., Ferula Karelini Bunge,
Frittillaria Karelini Bak. F. gibbosa), Halimocnemis Karelini Moq. Tand., Karelinia

petalum Karelini Fisch., Salix Karelini Turcz., Saussurea Karelinii Stsch., Suaeda Kare-
liniana Fzl., Statice Rarebni Stsch., Tamarix Karelini Bunge, Zygophyllum Karelini
Fisch,

LIST OF PLANTS NAMED IN HONOR OF I. P. KIRILOV*

Genus: Kirilowia Bunge (= Kirilovia Lindl.) in Chenopodiaceae.

Species: Carex Kirilowii Turcz., Halimocnemis Kirilowii Fenzl, Kirilowia eriantha
Bunge, K. pilosa Bunge (Panderia pilosa), K. villosa Benth. & Hook., Lepidium Kirilowii
Trautv., Pyrethrum Kirilowii Turcz. (Tridactylina Kirilowii C. H. Schultz), Salix
Kirilowiana Stsch.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

*Borodin made a mistake in listing in his Collectors and collections of the Siberian
flora (p. 47) Sedum Kirilowii Regel and Senecio Kirilowii Turcz. as plants named in
iri : :

coveries in China, pp. 346-352), were named in honor of Porfirij ae aah
(1801-1864), a een of the 11th Russian Ecclesiastic Mission to China, who
travelled with Bunge in 1830 through Mongolia to Peking and who ener in Chine
from 1830 to 1841

Herder, in his Biographische Notizen (in Bot. Jahrb. 9: 438. nia cites baal
Joh. Porph.” This mysterious collector is really a mixture of the two above-named
Kirilovs. Biographical data and even first and middle names are ao mixed u

JOURNAL

OF THE

ARNOLD ARBORETUM

Vout. XXIV APRIL, 1943 NuMBER 2

TAXONOMIC NOTES ON THE OLD WORLD SPECIES
OF WINTERACEAE

A. C. SMITH
With six text-figures

As implied by the title, this treatment is not monographic, its scope
being primarily to bring together references to the taxonomic literature, to
indicate the acceptable name for each species, to discuss the typification of
species and genera, to mention the known distribution, and to cite the
specimens available in American herbaria. From some regions there is a
deplorable lack of available herbarium material in this country, and conse-
quently my remarks must often be based upon those of previous workers.
Some revision of specific lines will be inevitable when European herbaria
can again be consulted, and doubtless a certain amount of revision will be
dictated by future collections throughout the range of the family. For
some of the genera in certain regions, such as New Guinea and Australia,
enough material is available to make feasible the preparation of preliminary
keys, but these are presented with the reservation that modification and
expansion will be essential.

Prof. I. W. Bailey and the writer have in preparation a treatment of the
inter-relationships of the six genera which make up the Winteraceae and
the place of the family in the Ranales. Consequently, in the present paper
I omit comprehensive generic descriptions and detailed discussions of gen-
eric relationships. During the preparation of this work, I have repeatedly
called upon Prof. Bailey for advice, and many of the conclusions expressed
are the result of his painstaking preparation and examination of material
for microscope study. It is hoped that the present paper will provide an
outline of the classification of the family and will bring up to date the
sometimes complicated synonymy.

The Winteraceae has an interesting bihemispheric and presumably
palaeoantarctic distribution, which I have recently discussed (5). In the
Old World, species are found in the area roughly outlined by the Philip-
pines, Borneo, New Guinea, the Solomon Islands, New Caledonia, New

120 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Zealand, Tasmania, and eastern Australia, All of the six known genera

occur in this region, while only one, Drimys, extends to America, where

it is found from Cape Horn to sonthenn Mexico. According to my interpre-
tation, the family now contains about 88 species, of which only four are

American. The type genus is Drimys (of which Wintera is a synonym),

and the type species of Drimys is D. Winteri J. R. & G. Forst. of southern

South America. This species, therefore, is the nomenclatural basis of the

family.

Many students have included Drimys and Illicium L. in the same family,

subfamily, or tribe. There appear to be no sound morphological, anatomi-

cal, or genetic reasons for this broad concept. The wood structure, nodal
anatomy, pollen grain, carpellary structure, and many other characters of

Illictum remove it from the Winteraceae. Whitaker (8) has pointed out

that Jllictum cytologically bears no resemblance to either Drimys or mem-

bers of the Magnoliaceae, being suggestive, in its chromosome number, of

Schizandra and Kadsura. ‘Therefore the genus is not considered in con-

nection with the present study; whether it should be placed in the

Schizandraceae or in an independent family must be decided by future

study

The directors and curators of the following institutions have kindly
loaned herbarium material, which is cited in this paper as indicated by the
parenthetical letters: Arnold Arboretum (A), Field Museum of Natural

History (F), Gray Herbarium (GH), Missouri Botanical Garden (M),

New York Botanical Garden (NY), University of California (UC), and

U.S. National Museum (US).

WINTERACEAE Lindl. Nat. Syst. Bot. ed. 2. 17, pro parte. 1836; Miers in Ann. Mag.
Nat. Hist. HI. 2: 33, pro parte. non Contrib. Bot. 1: 123, pro parte. 1861;
Eichl. in Mart. FI. ins, 13(1): 127. 1864; Hutchinson in Kew Bull. 1921: 185,
pro parte. 1921, Fam. FI. Pl. Dicot. 81, pro parte. 1926.

Wintereae . Br. ex DC. Reg. Veg. Syst. Nat. 1: 548, pro parte. 1817; Lindl. Introd.
Nat. Syst. Bot. 26, pro parte. 1830, Nixus Pl. 9, pro parte. 18

Pt le Trib. Jllicteae DC. Prodr. 1: 77, pro parte. 1824

Magnoliaceae Trib. Wintereae R. Br. ex Meisn, Pl. Vasc. Gen. 3 (pars alt. 5), pro
parte. 1836; Benth. & Hook. f. Gen. Pl. 1: 17, pro parte. 1862.

Magnoliaceae Subordo Illicieae DC. ex Endl. Gen. Pl. 838, pro parte. 1839.

Magnoliaceae Trib. Illicieae Sect. Winterineae Spach, Hist. Nat. Veg. 7: 435. 1839.

Magnotacen — Wintereae Lindl. Veg. Kingd. ed. 2. 419, pro parte. 1847, ed. 3. 419,

Pant oie 4 ‘Wicieae Prantl in E. & P. Nat. Pfl. III. 2: 18, pro parte. 1
Magnoliaceae Subfam. Drimyoideae Harms in Ber, Deutsch. Bot. Gesell. 16: 358.
1897.

Drimytacées v. Tiegh. in Jour. de Bot. 14: 354.
Drimytacées Trib. Drimytées and Trib. oe v. Tiegh. in Jour. de Bot.
14: 354. 1900

Drimytaceae Diels i in Bot. Jahrb. 55: 13
— Subfam. Drimydoideae nents in Vaxternas Liv 5: 349, pro parte.

The aia interpretations of the taxonomic position of the group now
known as the Winteraceae are indicated by the above synonymy. Most of
the early writers linked Drimys and IJllicium in the same tribe or section,

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 121

but the classification of Spach, in 1839, is interesting because of his pro-
posal of different sections of the Tribe Illicieae for these two genera.
Harms’ concept of Magnoliaceae Subfam. Drimyoideae (1897) is synony-
mous with the Winteraceae in the modern sense. The most precise delimi-
tation of the family, however, has been that of van Tieghem (6), who
apparently neglected to use the Latin form of his ““Drimytacées” anywhere
in his treatment.

Fic. 1. Approximate distribution of the genera of Winteraceae. From Goode’s series
of base maps, no. 108

KEY TO THE GENERA IN THE OLp WorLp

Calyx submembranaceous or papyraceous, calyptrate, completely enclosing the bud, at
length splitting into 2 or 3 ee an em often caducous; plants dioecious or polygamo-
dioecious; oi aa free, with the igmatic surface extended along the ventral suture;

closing th ; plants hermap ree odite ; carpe s often wit a short ae ridge
(or elliptic or subpeltate stigma in Lee filaments carnose, flattened.
Carpels free, sometim mes appressed-contiguous (in Bubbia and Exospermum), not
developing into a syncarp.
Inflorescence terminal or eae cham the flowers or inflorescence-rays aggre-
gated around the growing poin
Anther- fetes apical, horizontal or eer at the apex of distally enlarged
fi

laments, not exceeded by the connectiv
Carpels usually free even in young flowers, rarely appressed-contiguous at
anthesis ; aay ie essentially corresponding to the eis stig-
MALICISUTLACE = iar. fps z cteys, had sees eione coe sa write ars . Bubbia.
Gace: ed) -contiguous, free only after anthesis; placentation pee and
Scattered: on distal’Suriacesi<..c sci ascte ee eee ct care oe xospermum.,

Anther-locules vertical, extrorse-lateral, exceeded apically by the connective .

ee em re ety hic: Se eee Belliolum.
Inflorescence axillary; flowers fasciculate (rarely solitary), often arising from
branchlets of iia VeOAlS = Crowt le nee erste eee ices 3-3 . Pseudowintera.

Carpels firmly concrescent, the placentation primarily on the dorsal surface; fruit a
syncar 6. Zygogynum.

iz JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

1. DRIMYS
Drimys J. R. & G. Forst. Char. Gen. 83. 1776.

The Old World representatives of the genus all belong in the Section
Tasmannia, which is separable from the American Section Wintera as
pointed out in my recent treatment of that group (4: 10). I have already
listed the principal references to the genus Drimys as a whole, and conse-
quently the following references pertain only to the Old World portion of
the genus.

Drimys Sect. Tasmannia (R. Br.) F. v. Muell. Pl. Indig. Col. Vict. 1: 20. 1860; Baill.
Hist. Pl. 1: 160. ag 69.

Tasmannia R. Br. ex DC. Reg. Veg. Syst. Nat. 1: 445, 547. 1817, ex DC. Prodr.
1: 78. 1824; Lindl. Nat. Syst. Bot. ed. 2. 17. 1836; Meisn. Pl. Vasc. Gen. 3 (pars
alt. 5). 1836; Spach, Hist. Nat. Veg. 7: 433. 1839; Endl. Gen. Pl. 838. 1839,
Enchir. Bot. 428. 1841; Lindl. Veg. Kingd. ed. 2. 419. 1847; Hook. f. Fl. Tasm.
1:10. 1855; Miers in Ann. Mag. Nat. Hist. III. 2: 109. 1858, Contrib. Bot.
1: 138. 1861.

Drimys J. R. & G. Forst. sensu Benth. Fl. Austral. 1: 49. 1863; F. M. Bailey, Queensl.
Fl. 1: 18. 1899.

Drimys Sect. Winterana Baill. Hist. Pl. 1: 160. 1867-69.

Drimys Sect. Eudrimys v. Tiegh. in Jour. de Bot. 14: 288, pro aa 1900; Pilger

E. & P. Nat. Pfl. Nachtr. 2: 108, pro parte. 1906; non DC. (18

Drimys occurs in the Old World in Australia and Tasmania, a Guinea,
the Philippines, Borneo, Celebes, and Amboina. Its occurrence in other
high eastern Malayan islands is possible, but thus far no other regions
are represented in the herbaria examined by me, nor have published records
of other occurrences been found. It is a typically montane genus, reaching
elevations of 3800 m. in New Guinea and Borneo, but descending to sea-level
in the southern part of its range in Tasmania. Six species are known from
Australia and Tasmania, 29 from New Guinea, and one from the Phil-
ippines-Amboina region. In order to make this treatment more usable,
I propose to discuss and key the Australian and the New Guinean-Malayan
species separately.

Bs e genus. Tasmannia was originally based on 7. aromatica R. Br. and

. insipida R. Br., Australian species which are referable to Drimys
i (Poir.) Baill. and D. insipida (R. Br.) Pilger respectively. The
Sections Tasmannia and Wintera are sharply differentiated, but their
common origin is indisputable and the separating characters do not seem
generic in quality. The total variability of Drimys in the Old World,
especially in New Guinea, is considerable, particularly when the region is
compared to America, where, in an area extending from southern Mexico
to Cape Horn, no more than four species can be recognized. I have
attempted to use essentially similar standards for specific delimitation in
the two hemispheres. In spite of the greater number of Old World species,
it is much easier to draw specific lines in Sect. Tasmannia than in Sect.
Wintera, while intraspecific variability seems much less pronounced in the
Old World than in the New. From this fact one might assume that isolat-
ing mechanisms have been at work longer in Sect. Tasmannia than in

1943] SMITH, OLD WORLD SPECIES OF WINTERACEAE 123

Sect. Wintera and that extensive interchange of genetic material between
parts of the population ceased earlier in the Old World than in the New.

Following De Candolle’s treatments, many subsequent writers main-
tained the genus Tasmannia as distinct from Drimys, but modern students
have submerged it. I believe that it is well retained as a section with a
status similar to that of the American Section Wintera (Murr.) DC.
Baillon proposed to divide Drimys into four sections, but one of these,
Sect. Winterana, is difficult to typify from his brief treatment. I assume
that he meant to base it on “Winterania” lanceolata Poir., an Australian
species, and consequently I list it as a synonym of Sect. Tasmannia. Van
Tieghem’s four sections of Drimys are all based exclusively on American
species with the exception of Sect. Eudrimys, which, curiously enough,
includes the most extreme forms of the genus. The Old World representa-
tives of van Tieghem’s Section Exdrimys are referable to Mueller’s Sect.
Tasmannia, but the actual type of Sect. Eudrimys v. Tiegh. is the
American D. Winteri.

AUSTRALIAN SPECIES

The six recognizable Australian species are quite distinct from the New
Guinean representatives, all being characterized by having their pistillate
flowers lacking stamens (a feature of only a few New Guinean species).
However, the Australian species do not appear to form a single coherent
group, and it is not to be assumed that they were derived from a single
recent prototype. Drimys lanceolata and its two allies (D. stipitata and
D. Vickeriana) are entirely distinct from any New Guinean species and
appear to have no close relatives; they have probably been isolated for a
long time from the main trends of evolution in the Section Tasmannia.

The only Australian species which appears to have a close affinity. with
the New Guinean species is, as might be expected, D. membranea, of
Queensland, which has much in common with D. hatamensis Becc. This
relationship has already been suggested by Diels (in Bot. Jahrb. 54: 242.
1916). However, D. hatamensis (like its closest New Guinean allies
D. dictyophlebia Diels and D. coriacea Pulle) has pistillate flowers with
at least three carpels, while those of D. membranea have a single carpel.
There are also differences in the shape of the perianth-parts and fruits,
but in foliage the Queensland species and D. hatamensis are remarkably
similar.

Ridley (in Trans. Linn. Soc. II. Bot. 9: 12. 1916) has pointed out the
possible relationship of his D. densifolia to D. insipida (i. e. D. dipetala),
but the Australian species differs in its narrower leaves, larger petals, more
numerous stamens in staminate flowers, and usually solitary carpel. The
actual relationship of these two species is probably quite remote, in spite
of the similarity of their leaf-bases.

It seems desirable to redescribe three of the older Australian species and
one novelty, but I do not find it necessary to repeat Vickery’s excellent
descriptions (7) of D. purpurascens and D. stipitata.

124 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Key TO THE AUSTRALIAN SPECIES

Leaf-blades with (5—) 7-18 short anastomosing secondary nerves, these erecto-patent at
an angle of (20-)30-45°, the blades (4—)6—20 cm. long; petals 2, very rarely 3.
Leaf-blades gradually narcowed toward a suddenly obtuse and often auriculate base ;

carpels 1 (rarely 2) in both staminate and pistillate flowers; fruit 1-carpellate,
ellipsoid, usually 13-20 mm. long, obtuse at base, with (8—-)15-27 seeds ........
sori dshssdic asian ssi spiisg lds AR Hops 4ON AND GRIT SRS Sg GAA: dee ghatat fry CI Biet eAET aa 1. D. insipida.
Leaf-blades acute to attenuate at base
Petals up to 8 mm. long at anthesis; filaments 0.53.5 mm. long; carpels 1 (rarely
2 or 3) in both staminate and pistillate flowers; aul | 1-carpellate, often nearly
subglobose, not more than 10 mm. long, rounded at t base, ee with 4-7
BABS) foo anaia Gun arers cea icuhepnaaie ew aase space gape Re De ata Ses 2D. mbranea.
Petals 10-12 mm. long; filaments 2-6 mm. long; carpels 2-8 in bo th peers and
pistillate flowers ; a uit 2-8 (usually 3- or 4-)-carpellate, “ay carpels oblong-
obose, 10-15 mm. long, short-stipitate .............+-. . D. purpurascens.
Leaf- tere mane 3-7 ease secondary nerves, these sharply ais ing at an angle of
0-20° or completely immersed and obscure, the blades not exceeding 11 cm. in
non ie to attenuate at base, — auriculate
Sepals 3.5-6 mm. in diameter; petals 4-9 mm. long, 1.5-3.5 mm. broad; stamens in
staminate flowers 15-45; fruiting — with at least 6 seeds; leaf- blades rarelv
less than 4 cm. long and 1 cm
Petals fic staminate flowers 5-8, n pistilate cage usually 4; carpels in both
minate wn pistillate ee 1, rarely 2 or 3, subglobose or ellipsoid, sessile,
oa stigm ridge occupying the dite pa and ventral edge; fruit
1-carpe nate, pene subglobose, rounded at base .......... 4, D. lanceolata.
Petals 2; carpels in staminate flowers 1 or 2, in pistillate flowers 2-8, obovoid-
ellipsoid, obviously stipitate, the stigmatic ridge obliquely apical or extending
base of ovary; fruit 2—-8(usually 3- or 4-)-carpellate, the carpels oblong-
ellipsoid, conspicuously BUTERA CES rs tndeaaarnris,abensiataroirere eee 5. D. stipitata.
Sepals 2.5-3 mm. in diameter; petals 2, not exceeding 3 mm. in length and 1 mm. in
breadth at anthesis; samen s in staminate flowers 10-12; carpel in staminate
flowers 1, subsessile; fruit 1-carpellate, subglobose, with about 3 seeds; leaves
pace the blades small, up to 16 mm. long and 5 mm. broad ..........
scenes pussies counsel fo SRE eget onabenstahey atevance ace creterreesene iar cota tahwaane agmiaae Si miaverate 6. D. Vickeriana.

1. Drimys insipida (R. Br.) Pilger in E. & P. . Pfl. Nachtr. 2: 108. 1906; Druce
in h. Cl. Brit. Isles aaae. er 1017; Domin in Bibl. Bot.
22 [Heit 89]: 115. 1925; Vickery in Proc. Linn. San ~ S. Wales 62: 82. 1937.

Tasmannia monticola A. Rich. Sant. Aabedlaly. 50. ‘18s te = 19. 1833) ;
Muell. Pl. Indig. Col. Vict. 1: 21, as synonym
Drimys dipetala F. v. Muell. Pl. indi, Col. Vict, le 21. 1860; Benth. FI, Austral.
49 i

27: 227, 301. 1896; F. M. Peon Queensl. Fl. 1:18. 1899; Pilger in E. & P.

Nat. Pfl. Nachtr. 2: 108. 1906; F. M. Bailey, peer _ Queensl, Pl. 21. f. 7.
1913; Maiden & Betche, ee N. S. Wales Pl. 79.
Shrub or small tree, to 3 m. or more high, the branchlets subterete,

lets; petioles rugulose, shallowly canaliculate, often inconspicuous, 0.5—4
mm. long, 0.7—2 . in diameter; leaf- blades chartaceous, brownish or

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 125

DRIMYS
* INSIPIDA eee ih cee ;
+ MEMBRANEA ;
PURPURASCENS
LANCEOLATA
STIPITATA
VICKERIANA

x

>

4

/

Fic. 2. Approximate known distribution of the Australian and Tasmanian species of
Drimys. From Goode’s series of base maps, no. 107.

hie and faintly Sr riadiee or obscure on both surfaces; flowers
single, numerous, congested around the growing point of branchlets, at
length pseudolateral, subtended by numerous bracts, these papyraceous,
oblong, 5-11 mm. long, obtuse, soon caducous, the pedicels slender, 10—25
mm. long at anthesis (sometimes shorter in pistillate plants), up to 30 mm.

126 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

long in fruit; staminate flowers: sepals 2, submembranaceous, densely but
obscurely yello ow-glandular, ee deltoid, i 7mm. long and broad, obtusely
apiculate at apex; petals submembranaceous, eglandular to densely
glandular- punctate, Leeann or spatulate-linear, 7.5-13 mm. long, 1.5-2.5
mm. broad, conspic uously narrowed at base, ‘obtuse at apex; stamens
usually 30-38, 3- or 4-seriate, the filaments subcarnose, ligulate, obscurely
pellucid- glandular, 1-5 mm. long, the locules 1-2 mm. long; carpel 1
(rarely 2), sterile, ellipsoid, slightly falcate, often yellow-glandular, 2.5—4
mm. long at anthesis, the stigmatic ridge elongate: pistillate flowers: sepals
as in staminate but usually longer than broad, obtuse at apex; petals
similar but smaller, up to 10 mm. long, 1-1.5 mm. broad; stamens none;
carpel 1, ellipsoid, about 3 mm. long at anthesis, obtuse at apex, the
stigmatic ridge extending from apex nearly to base, the ovules about 28;
fruit 1-carpellate, ellipsoid, at maturity (9—)13—20 mm. long and (S—)7—10
mm. broad, obtuse at base and apex, obscurely marked re by the
stigmatic ridge, the pericarp carnose, copiously yellow- -glandular, the seeds
(8—) 15-27, loosely imbedded in sparse mucilaginous pulp, coriaceous, black,
shining, obovoid, slightly or strongly falcate, 3-3.5 mm. long, 1.5— c am
broad, acute at base, rounded at apex.

DistriBuTION: Southeastern Queensland and eastern New South Wales, from about
lat. 24° to 35°30’, from sea-level to about 1500 m. in mountains; usually reported as
occurring in rain- forest. The type was collected by Brown at or near Port Jackson.

Austratia: Queensland: Fraser Island, Epps 229 (NY); Tinana Creek,
White 3476 (A); Imbil, Wilson (A); Eumundi, Simmonds (A); Tallebudgera, White
1866 (A); — Plateau, Lamington National Park, White 6062 (A, NY); eae a
ham’s Gap, Main Range, White 6862 (A, NY); Mt. Spurgeon, White 10737 (A);
Greville, he. 9949 (A), Everist 556 (A); Tambourine Mt., White 3568 (A)% National

ee neni Range, White (A); Mistake Range, White (A), Bailey (US); New
So Wales: Richmond River, Henderson (UC); Upper Williams River at
ce White 11473 (A); Wentworth Falls, Maiden (UC), Burges (NY); Moonam-
bale, Maiden (A); Port Tackson, Caley (A)

NaTIVE NAME: Pepper shrub.

White reports the sepals as red, the petals paler red, and the stamens
yellow; other collectors have not mentioned the flower-color. The fruit
is usually noted as purple to black at maturity, but some collectors indi-
cate that it is white (probably only when juvenile).

The first occurrence of the name Tasmannia dipetala in literature is
De Candolle’s reference to it, in the Prodromus, as a synonym of T. insipida.
In transferring the species to Drimys, Mueller unfortunately selected the
epithet dipetala and has been followed by many students. The correct
transfer of the epithet insipida was apparently first made in 1906, Subse-
quently to Pilger’s publication of the binomial Drimys insipida, both Druce
and Domin proposed the combination as new.

Richard’s description and illustration of Tasmannia monticola portray
an apparently hermaphrodite flower, with numerous stamens and two fertile
carpels. In staminate flowers of D. insipida one occasionally finds two
carpels, but these are always sterile; the fact that Richard illustrates
ovules suggests either faulty observation or a very unusual individual, since
there can be no doubt of the identity of his plant with D. insipida.

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 127

2. Drimys membranea F. v. Muell. Fragm. Phyt. Austr. 5: 175. 1866; F. M. Bailey,
Syn. Queensl. Fl. 5. 1883; Parment. in Bull. Sci. Fr. & Belg. 27: 227, 302, as
D. membranacea. 1896; F. M. Bailey, Queensl. Fl. 1:18. 1899; Pilger in E. & P
Nat. Pfl. Nachtr. - 108, as D. membranacea. 1906; F. M. Railey, Compr. Cat.
Queensl. Pl. 21. 19

Shrub or tree up to ee m. nigh, the branchlets subterete or slightly angled,
rugulose, brownish, 1.5—-4 mm. in diameter near apices; leaves often more
or less congested toward oe ‘of branchlets, the petioles stout, 1-2 mm
broad, often flattened and narrowly winged, 1-7 mm. long; leaf- blades
chartaceous or subcoriaceous, brownish when dried, narrowly obovate,
(4—)6-13 cm. long, (1—)1.5-4.3 cm. broad, gradually attenuate at base
and decurrent on the petiole, obtusely cuspidate or short-acuminate at apex,
slightly recurved at margin, the costa broad, slightly raised or shallowly
canaliculate above, usually subprominent beneath, the secondary nerves

(S5—)8—-14 per side, erecto-patent at an angle of 30—45°, prominulous on
ee surfaces, freely anastomosing near margin, the veinlets reticulate,
faintly prominulous on both surfaces; flowers single, clustered aroun nd
growing point of branchlets, at length hcalciol ara subtended by bracts,
these papyraceous, oblong, about 5 mm. long, soon caducous, the pedicels
slender, 12-33 mm. long (sometimes shorter in pistillate plants) ; staminate
flowers: sepals 2, submembranaceous, eglandular or obscurely yellow-
glandular, suborbicular-deltoid, 4-5.5 mm. long and broad, obtuse or
rounded at apex; petals 2, similar to sepals in texture, obovate-linear, about
& mm. long and 2—3.5 mm. broad at anthesis, narrowed to base, obtuse or
rounded at apex; stamens 25-35, 3- or 4-seriate, the filaments subcarnose,
subterete-ligulate, obscurely yellow-glandular, 0.5—3.5 mm. long, the locules
0.7-1.3 mm. long; carpels 1 or 2 (or 3, ex Mueller), sterile, obovoid-
ellipsoid, 2.5-3 mm. long at anthesis, the stigmatic ridge obvious, elongate;
pistillate flowers: sepals and petals similar to staminate in texture, the
sepals ovate-oblong, 3.5-6 mm. long, 2.5-3 mm. broad, the petals about
& mm. long and 2—2.5 mm. broad; stamens none; carpel 1, ellipsoid, 2.5—4
mm. long at anthesis, short-stipitate, the stigmatic ridge obvious, extending
from apex nearly to base, the ovules 14-36 (or possibly sometimes more,
with some aborted) ; fruit 1-carpellate, oblong-ellipsoid to nearly subglobose,
6-10 mm. long and 5—7 mm. broad at maturity, rounded at base and apex,
the stigmatic ridge elongate, obscure, the pericarp carnose, obscurely
yellow-glandular, the seeds usually 4—7 (with many aborted), black, shin-
ing, obovoid, slightly falcate, 3-3. 5 mm. bee and about 2 mm. broad at
maturity, subacute at base, rounded at apex.

DISTRIBUTION: Eastern Queensland, from about lat. 16° to 18°, at elevations of
800-1600 m. (or perhaps oe down to sea-level); in rain- foret or low bush,
sometimes common (Kajew

AusTRALIA: Queensland: Thornton Peak, Brass 2291 (A); Gadgarra, Peer-
amon, Atherton, Kajewski 1065 (A, NY, UC); Mt. Bartle Frere, Kajewski 1291
(A, NY); Bellenden Ker, near summit, White (A); Rockingham Bay, Dallachy
(GH, NY):

The sepals and petals are reported as white by Brass, green by Kajewski;
the mature fruit is said to be black. The young leaves, bracts, and some-
times the whole plant are often tinged with purple.

Mueller’s original description is based on an apparently staminate speci-
men collected by Dallachy “ad fontes fluminis Mackay-River,’ whereas

128 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

the Dallachy specimens cited above, from Rockingham Bay, are pistillate.
The type specimen must have very young leaves, as none of those seen by
me could be considered membranaceous. Mueller describes the petals as
2 or 3, but all my specimens have them 2

Drimys membranea is doubtless a close relative of D. insipida, which it
resembles in leaf-shape and texture, differing primarily in the quite different
leaf-base. Both species have the sepals and petals of the pistillate flowers
narrower than those of the staminate. In D. membranea both sepals and
petals are shorter and usually proportionately narrower. ‘The fruit of

. membranea appears to be consistently smaller and with fewer seeds
than that of D. insipida.

3. Drimys purpurascens Vickery in gine Linn. Soc. N. S. Wales 62: 78. f. 1; pl. 5.
1937; Fraser & Vickery, 1]. c. 288.

DistriputTion: Thus far known sue — a restricted area on the Mt. Royal
Ranges in the vicinity of the Barrington Tops, New South Wales, at an altitude of
1350-1500 m

AustTrRaALIA: New South Wales: Barrington Tops, Boorman (US), White
11472 (A)

The original description is very complete, being based upon several col-
lections from the same locality; the type is Fraser & Vickery (in May,

193 The species is readily distinguished from its allies, D. instpida and
D. membranea, by its more numerous carpels and several- carpellate fruits,
as well as by its usually broader leaf-blades and broad subalate petioles.
White indicates that the species is very common locally in both rain-forest
and Eucalyptus forest.

4. — lanceolata (Poir.) Baill. Hist. Pl. 1: 159. f. 205-207. 1867-69; Parment.

Bull. Sci. Fr. & Belg. 27: 225, as synonym. 1896; Pilger in E. & P. Nat. Pfl.

Nacht 2: 108. 1906; Ewart, FI. Vict. 517. 1930; — in Proc, Linn. Soc. N.S.
Wales 62: 82. 1937; Fraser & Vickery, ]. c. 288

Winterania lanceolata Poir. Encycl. 8: 799. 1808.

Tasmannia aromatica R. Br. ex DC. Reg. Veg. Syst. Nat. 1: 445. 1817; Deless.
Ic. Sel. 1: 22. pl. 84. 1820; DC. Prodr. 1: 78. 1824; Lindl. Bot. Reg. 3 l. 43.
1845; Hook. f. Fl. Tasm. 1: 11. 1855; Miers in Ann. Mag. Nat. Hist. IIT. ‘3: 110.
1858, Contrib. Bot. 1: 139. 1861; Baill. Hist. Pl. 1: 159, as synonym. 1867-69;
Meredith, Bush Friends Tasm. Ser Ult. pl. 17. 1891.

Drimys aromatica F. v. Muell. Pl. Indig. Col. Vict. 1: 20. 1860; Benth. Fl. Austral.

63% lg v. Muell. Nat. Pl. Vict. 1: 19, 187. f. 44. 1879, Key Syst. Vict.
Pi: 2: 6. pl. 3. 1885, op. cit. 1: 121. 1888; C. Moore, Handb. FI. N. S. Wales 13.
1893 ; said in Bull. Sci. Fr. & Belg. 27: 225, 298. pl. 11, f. 41. 1896; Rodway,
Tasm. Fl. 5. 1903; De Wildem. in Ic. Sel. Hort. Then. i 127. pl. 191. 1906;
Maiden & Betche, Census of N. S. Wales Pl. 79. 1916; Ewart, Handb. For. Trees
Vict. For. 116. 1925; Johnstone in Jour. Roy. Hort. Soc. 62: 96. f. 25. 1937.

Drimys aromatica var. aromatica Parment. in Bull. Sci. Fr. & Belg. 27: 226. 1896.

?Drimys xerophylla Parment. in Bull. Sci. Fr. & Belg. 27: 226, 299, nomen sub-
nudum. 1896; Vickery in Proc. Linn. Soc. N. S. Wales 62: 83, as synonym. 1937.

hrub or small tree, 2-10 m. high, the branchlets aceon or slightly
angled, rugulose, reddish brown or purplish, slender, 1-3 mm. in diameter
toward apices; leaves scattered, the petioles rugulose, dusllowis canalicu-
late, 2-15 mm. long, 0.7—-2 mm. in diameter, often swollen at base; leaf-
blades subcoriaceous or chartaceous, pale green or yellow-green when dried,
oblanceolate or narrowly elliptic-obovate, (3—)4-11 cm. long, 0.6—-3 cm.

1943] SMITH, OLD WORLD SPECIES OF WINTERACEAE 129

broad, attenuate at base and decurrent on the petiole, obtuse or subacute at
apex, narrowly recurved at margin, the costa subplane or slightly raised
above, more obviously elevated beneath, the secondary nerves 3—7 per side,
elongate, sharply ascending at an angle of 10-20°, inconspicuously anas-
tomosing toward margin, immersed or prominulous above, prominulous
beneath, the veinlets immersed or obscurely prominulous beneath: flowers
single, aggregated around growing point of branchlets, at length sa
lateral, subtended by bracts, these papyraceous, oblong, obtuse, 4-13 m

long, soon caducous, the pedicels slender, 8—25 mm. long (staminate eae
ers) or 4-12 mm. long (pistillate flowers and fruits) : staminate flowers:
sepals 2 (rarely 3), membranaceous, densely but obscurely pellucid-

petals 5—8, resembling sepals in texture, linear-oblong or narrowly obovate,
4-9 mm. long, 1.5—3.5 mm. broad, obtuse at apex; stamens 15-28, 2- or
3- ua! iy filaments eglandular, 0.7-3.5 mm. long, the locules ellipsoid,
m. long, the carpel 1 (rarely 2 or perhaps 3), sterile, ellipsoid,
0.7—-1 mm. long at anthesis, the stigmatic ridge obvious, occupying entire
rounded apex and extending to base; pistillate flowers: sepals as in stami-
nate or ma, narrower; petals 4 (in all available specimens), as in
staminate but 3.5—5.5 mm. long and 0.7—2 mm. broad; stamens none;
carpel 1 (rarely 2), subglobose or ellipsoid, about 1.5 mm. in diameter at
anthesis, rounded at base and apex, the stigmatic ridge extending over apex
along entire ventral edge, the ovules 10-18; fruit 1-carpellate, oblong-
subglobose or subglobose, often appearing obscurely bilobed, 5-10 mm.
long and 5—8 mm. broad at maturity, rounded at base and apex, the stig-
matic ridge elongate, the pericarp subcarnose, rugulose, obscurely yellow-
glandular, the seeds 6—18, dark castaneous or black, shining, obovoid,
strongly falcate, 2.5-3.5 mm. long, 1.3—2 mm. broad, acute at base, rounded
at apex
DistrRIBUTION: New South Wales, Victoria, and Tasmania, extending southward
from about lat. 31°, at elevations up to 1200 m. in the north, occurring down to sea-
level in Tasmania; usually found in wet forest in New South Wales and Victoria;
ane in Tasmants and forming dense thickets on margins of streams in rich humid

VT eared LIA: South Wales: Hastings River, Moore (GH); Braid-
wood, Baker mise ‘aie ioal Mt., Braidwood, Boorman (US); Bago, de Beuzeville
201 (A); Tumbarumba, Cambage (GH); Victoria: Kuntze 20113 (NY); Mt.
Baw Baw, Melvin (GH); Mt. Mueller, near Mt. Baw Baw, Mueller (GH);
Tasmania: R. Browns. n. or 2918 (type coll. of Tasmannia aromatica, F, GH,
M, NY), Gunn (A, F, GH, M, US), Archer (GH); Golden Valley, toward Westbury
Road, Rodway 125 (A); Mt. Wellington, Gunn 777 (GH, M), Oldfield (US);

without definite locality: Labillardiére (tyPE coLL., GH),
Paterson (A).

NATIVE NAMES: Pepper-tree, native pepper, mountain pepper, wild pepper-corn.

Mueller describes the petals as white with a red or green streak down
the middle; colored plates show the petals as white, but none of the speci-
mens seen by me have adequate color-notes.

Many of the descriptions cited above, such as those of Hooker, Mueller,
and De Wildeman, are based on a more comprehensive concept than
admitted in my description; these authors include in the species material
from the Victorian Alps which is much reduced in foliage and floral char-

130 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

acters and which I describe below as D. Vickeriana. The species which
Vickery has described as D. stipitata is also included in some of the early
concepts of the present species, as indicated by references to the petals as
2-8 in number; I believe that D. /anceolata never has fewer than 4 petals.
Mueller describes the ovaries as 1-11 and in some of his plates shows
4 carpels; his concept doubtless included specimens which are referable
to D. stipitata, but even there no more than 8 carpels have otherwise been
reported.
5. Drimys stipitata Vickery in Proc. Linn. Soc. N. S. Wales 62: = ‘2.1937:

Drimys aromatica var. pedunculata Maiden in Agric. Gaz. S. Wales 5: 600

(Dorrigo For. Res. 8). 1894; Maiden & Betche, Census N. Ps Wales Pl. 79, 1916.

DISTRIBUTION: Fairly common in onan New South Wales between about
lat. 29° and 32°, at elevations of 600-120 and also occurring near the boundary
with Victoria perhaps at slightly higher ita, probably to be found elsewhere in
New South Wales and in adjacent Victor

AustrRALiA: New South Wa ; es: Dorrigo Forest Reserve, Beilsdown
Creek, Maiden in 1893 (type coll. of D. aromatica var. pedunculata, A); Dorrigo State
Forest, toward Wild Cattle Creek, White 7572 (A, NY); Mt. oo i line to
about 2100 m., Maiden & Forsyth (UC); Snowy River, Cheel 813 (A

The original description is ample and is based upon mee ee
collections, the type of the species being Maiden (in 1895), collected at
Guy Fawkes. Although D. stipitata is scarcely distinguishable from
lanceolata in foliage, the differences in number of petals and carpels, and
especially in the shape of the carpels, are quite obvious. The carpels of
the present species are conspicuously stipitate in fruiting specimens, and
this character is also apparent in both staminate and pistillate flowers.
However, there can be no doubt that D. lanceolata and D. stipitata are
close relatives, and I believe that they are too far separated in Vickery’s
key (7: 83). The pistillate flowers, as in other Australian species, lack
stamens, while carpels in the staminate flowers (1 or 2 in my material)
are sterile.
6. Drimys Vickeriana sp. nov. Fu

Drimys aromatica var. alpina commen in “Bull Sci. Fr. & Belg. 27: 226, 300, nomen

subnudum, 1896.
Drimys ismcuoiote var. parvifolia Vickery in Proc. Linn. Soc. N. S. Wales 62: 83.

Frutex compactus, ramis crassis multiramulosis, ramulis subteretibus
cinereis apicem versus 1—3 mm. diametro cicatricibus foliorum delapsorum
copiose ornatis; foliis dense congestis, petiolis rugulosis semiteretibus 1—3
mm. longis 0.8—1 mm. diametro, laminis coriaceis siccitate fuscis oblongis
vel anguste ellipticis, 8-16 mm. longis, 2~5 mm. latis, basi acutis vel

secundariis utrinsecus circiter 3 adscendentibus utrinque immersis vel supra
inconspicue impressis, venulis obscuris; floribus singulis aggregatis termi-
nalibus demum hep ae ei bracteis submembranaceis pellucido-
glandulosis obovatis, 3—4 mm. longis, 2—2.5 mm. latis, apice rotundatis, mox
caducis; pedicellis gracilibus teretibus 3— 5 mm. longis; floribus ¢_ solis
visis: sepalis 2 submembranaceis eglandulosis suborbicularibus 2.5-3 mm

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 131

diametro, apice rotundatis; petalis 2 ee obovato-oblongis,
sub ant thesi 2.5—3 mm. longis et circiter 1 mm. apice rotundatis;
staminibus 10-12, 2-seriatis, filamentis cia “egandulos 0.5-1. 5
mm. longis, loculis ellipsoideis 0.6-0.8 mm. longis; ca arpello unico sterili
ellipsoideo sub anthesi circiter 1 mm. longo, basi et apice obtuso, carina
stigmatum apicali-ventrali circiter 0.5 mm. longa; fructibus 1- -carpellatis
subglobosis maturitate 4-5 mm. diametro, pericarpio subcarnoso ruguloso

4) y Mi] 5,
ei pine
re A] Wb
57) we
Fee
Zs Go

Fic. 3. a-e. Drimys Vickeriana, drawn from the type: a. flowering branchlet, x 2;
b. staminate flower, x 3; c. stamens, introrse and extrorse views, x 5; d. sterile carpel,
x 5; e. sterile carpel, longitudinal section, x 5. f-i. Drimys microphylla, drawn from
the type: f. flowering branchlet, x 4; g. staminate flower, xX 2; h. stamens, extrorse

and introrse views, X 5; i. sterile carpel, X 5. j-n. Drimys obovata, drawn from the
type: 7. flowering agence x 4; k. young staminate flower, x 14; J. mature stam-
inate flower, X 13; m. stamens, introrse and extrorse views, x 5; n. sterile carpel, x

SO .eP: slaps Brassii, ae from the type: o. fruit, showing three carpels, x 1;
p. seed

obscure glanduloso, seminibus circiter 3 castaneis nitidis obovoideis leviter
falcatis, 2.5-3 mm. longis, 1.5—2 mm. latis, basi obtusis, apice rotundatis.

DISTRIBUTION: eee portion of the Australian Alps, Victoria, at elevations of
about 1200-1500 m

AUSTRALIA: Vic ria: Mt. Baw Baw, Mueller (US); Mt. Mueller, near Mt.
Baw Baw, J. G. rae (cam Ge pee in 1893 (GH, Type, UC), French in 1895
(GH, NY, UC).

No specimen is cited with Parmentier’s brief note on D. aromatica var.
alpina, but the specimen I have selected as the type of the new species
bears this herbarium name; the type of D. lanceolata var. parvifolia is
J. Staer in April 1911, from the Upper Yarra. In order to avoid ambiguity
I have thought it better to propose the above-described entity as a new

132 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

species rather than to take up one of the varietal names. The species is
named for Miss Joyce W. Vickery, in recognition of her work on the
Australian Drimytes. Additional specimens are cited with her varietal
description, and she describes the leaf-blades as 8-23 mm. long, but none of
those available to me exceed 16 mm
Drimys Vickertana is very distinct in its compact habit, small crowded
leaves, small floral parts, reduced number of stamens and seeds, etc. Like
lisemataten it has a subglobose 1-carpellate fruit which is rounded at
base, while it resembles D. stipitata in having only two petals. It appears
to me to be one of the most distinct species of the genus, and I am unable
to consider it merely a variety of either of its relatives.
NEW GUINEAN AND MALAYAN SPECIES
The Section Tasmannia reaches its greatest development in New
Guinea, in its total variability, number of species, and number of indi-
viduals. Until the extensive exploration of New Guinea began, no more
than forty years ago, the genus was supposed to occur there only sparsely,
as represented by a very few collections referred to D. hatamensis Becc.
and D. piperita Hook. f. We are now aware that the genus is one of the
predominant elements in many of the montane regions of New Guinea,
having been collected at all elevations between 800 and 3800 m._ Its
occurrence toward the lower limit of this altitudinal belt is presumably
sparse, and therefore it was not observed in quantity until explorers gained
access to the interior mountains. In this treatment I am able to recognize
29 species from New Guinea, although many of these are known to me
only from the original descriptions. This number contrasts with the six
species known from Australia and Tasmania and the single species known
from the Philippines, Borneo, Celebes, and perhaps Amboina.
Key TO THE NEw GUINEAN AND MALAYAN SPECIES
sara agar oS probably always ann tic. the leaf-blades 5-10 mm. long.
Pet . 2: t 3 mm. long; stamens about 8, 1- or 2-seriate; leaf- gene 5-7 mm.
SRENSSAASAES SHG Mecane BAe a: chimp: eats Show civaus cick Sutny: bean edrand-denesnonbanasionss is . vaccinioides.

oe 6, 4-5 mm. long; stamens 14-18, usually 3-seriate; leaf- ae (5- )6-10 mm.

MAIN nic ce vas tonetssseasd ws onatieaarcsdal ate ata ome ere) paw dearaaaeea seen eee sions microphylla
Small-leaved species, sometimes epiphytic, usually sclerophyllous, ‘the leaf-blades
0.9—)1—4 (rarely to 5.5) cm. long.

Leaf-blades (9—)10-18 mm. long, 4-12 mm. broad; petals 2-4, up to 8 mm. long;
stamens in staminate flowers 18-27, in hermaphrodite flowers 13-18 ..........
gs esigel ead dees calmed eaghs Ses aos MO eaery anastasia acas is dheS ante siadein tact aedinneye raise aniceasneos 9. D. buxifolia.

Doubtfully distinct from preceding ..................00-0000.. 10. D. Versteegii.
11. D. reducta.
Leaf-blades (1.5—)2—4(—5.5) cm. lon

Petals none (rarely 1); stamens in dtewinate flowers 4—6; leaf- ore lanceolate-
oblong, obtusely acuminate at apex........... 00.00 cee ueee oligandra.
Petals stamens in staminate flowers probably 12-27; leaf- blades elliptic- or

oblong-obovate, obtuse or rounded at apex.
Leaf- bois with obvious venation, (10—)15—25 mm. broad; pee in pistillate
owers none; petals 4—-4.5 mm. long; carpels 2-4; ovules - 1G: oieiow sinus

Leaf-blades with immersed venation.

1943] SMITH, OLD WORLD SPECIES OF WINTERACEAE 133

Carpels in hermaphrodite flowers 5 or 6; stamens et ape blades 15-22
mm. broad

scstalfesatcesaitecnee-diiy Wiaterslaat € sieang a at quanta abeeaane a D. pittosporoides.
Carpels os fruit) 1-3; seeds pa 24; petals in Meare ‘flowers 5.5-7 mm.
long; leaf-blades (4—) 6-13 he OAC) cc. ceo nets. ne be Brassii.

Petals 5 or more (rarely 4) ; aa in ret flowers 12-30 or
Leaf-blades fistulose, strongly revolute at margins, 1-5 mm. broad; petals 10-12,
the outer ones Broadest and sepaloid ...... 0.1.08 seed e.0 0 16. D. fistulosa.
Leaf-blades more or less flattened, at least 5 mm. broad
Flowers large, the petals about 18 mm. long and 7 mm. broad; stamens in
i owers about 30, the carpel solitary .......... 17. D. elongata.
Flowers smaller, the petals less than 10 mm. long.

t or more, variable in size, the outer and inner series smaller than
the middle series; stamens about 25; leaf-blades 5-10 mm. broad, with
Obsolete: Nerves: . so. oe shavs cae d dade vanessa s onlides . D. myrtoides.

ee 7-14; stamens in staminate flowers 19-22; leaf- blades 12-23 mm.
— Ree etait Any ain & oan a ou abate cane mneene enna . D. arfakensis.
mee

Leaf- ‘ae obovate, about 2 cm. long and 1 cm. pies ie coriaceous ;

tals linear-oblong or subspatulate, about 4 m mm.
<0: (0 Lee ee IO ee oC RC ea ae er are ere a are “19. D . parviflora.

Leaf- ae elliptic-obovate, 1.5-3 cm. long, 1-2 cm. broad, thick-

coriaceous, with reticulate-prominulous nerves; petals oblanceolate or

narrowly oblong, 3-5 mm. long, about 2.5 mm. Be ete eet ess

FTN RE ee ee O54D: ed aan
Leat- ares elongate-obovate or oblong-spatulate, 2- a cm, long,

ad, coriaceous; petals narrowly spatulate, a mm. ae

oe MAT DLOAG cia rec: k aoracchsa. sarane capaci eats hee .D. Lamii.

ee verticillate, or at east ee pcan ks in clusters of Ne a aid 7 or 8.
Pedicels about 1 cm. long; largest petals about 6 mm. long and : 1.7 mm. bro

léat-blades. attenuate at base? . 4 )4.00204 3080.2. San beets aD: ee ledlaic.
Pedicels 6—7 cm. long; largest sate about 14 mm. long and 5 mm. broad; leaf-
blades broadly rounded at base ........... 0.000 ce eeeueceeeeee 23. D. rosea.

tie alternate or subopposite, not vert.
s 5 or more (sometimes 4 i ye
he small, the petals 4— ‘i mm. long at anthesis, the stamens in staminate
flowers usually 15-22.
Petals 7-14; leaf-blades obovate-lanceolate, up to 6 cm. long hae 2.3 cm.
BLO AC: <i boa Scien te ee cancun oc arene 4. D. arfakensis.
sen 4-6; -_ blades narrowly oblanceolate-oblong, 6-10 cm. long, 1.3-2.5
dibs osier anita, 3 acd Ren TE Ss cect ote a, hare 25. D. reticulata.
Flowers aoe the petals 6-14 mm. long at anthesis, the stamens in staminate

flowers 25-65 (or rarely m
Leaves em whitish net the secondary nerves about 7 pairs; flowers
large, the petals in staminate flowers 8 or 9, about 10 ey ONE: “cic:

Rte bcc ares gla) ig Acs ho caaa eae me aa RRR aT: emasisviale alana <-D pp em
lueaves petiolate (petioles at least 3 mm. long), slightly es beneath or
concolorous.

Secondary nerves and veinlets sharply insculpted above; sepals 8-10 mm.
long and broad; petals 5-7 in staminate flowers, 10-14 mm. long .
ies a 76S SASS Shay ailsereva) Sayan shee aes ae en Pee ete NPR Ree ac . D. macra antha.
reise nerves and veinlets Shae on both surfaces, rarely slightly

impressed above; sepals 3-6.5 mm. long and broad (rarely to 8 mm.) ;
stig 6-8 (-—11) in staminate flowers, usually 6-12 mm. long ........
a eile ycdiiay ah os Sy! xs as 8 Gesaile, SR MRS a deal ea Tolle Inlado ce ReAeeneie tele re’ y Yankee 6 /oo ois 28. D. piperita
Petals
Leaf- oe narrow, 1-2.2 cm. broad, 4-8 cm. long, the veinlet-reticulation

prominulous on neers surfaces.

134 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Apex of leaf-blades abruptly long-acuminate, the texture genie
]

pedicels: 3=4: Cri JONG 6s. eiceisermersis ins aibudiartesen ae aoeietent ame ete 29. D. acutifolia.

Apex of leaf-blades obtusish, the texture coriaceous; — 1.8-2 cm. long

slp a Sys be -as4 hs da ngTav Sapam lab aapadasa Sino Malas DRAPER ane aRiaLekeonipae . D. Beccariana.

Apex of leaf-blades acuminate, the texture coriaceous; olivia 1.5-1.7 cm.

NTR: ceesens atetvaxe ewarehomge Shots ee ators cis se ee ee aK 31. D. cyclopum.
Leaf-blades broader, usually more than 3 cm. broad and 10 cm. —.

Leaf-blades obtuse and auriculate at base ................. . D. densifolia.

Leat-blades gradually narrowed to base and decurrent on a tiole.
Veinlet-reticulation of the leaf-blades only faintly eipeapeaies sometimes
obsolete above, the costa raised above; stamens in staminate flowers
On ee obovata.
Veinlet. reticulation of the leaf-blades conspicuously prominulous, at least
ath; stamens in staminate flowers fewer than 35.
Leaf- “blades chartaceous, the costa nearly plane or slightly impressed
seni oe lets 2-4 mm. in diameter toward apex; pedicels 7-25
wesvauiits se-elt, g csetsse eee lecpe Reed se area RN SOROS 34. D. hatamensis.
Leaf- ‘lads ally coriaceous, the costa impressed above; branchlets not
e than 4 mm. in diameter toward apex; ner’ 30-35 mm. long
santana sibue haaeAdeh Su tee euaue-easen ag aleen ae sah oara eNeets aD: dictyophlebia.
Leaf- nce thick-coriaceous, the nerves impressed on ye upper surface;
anchlets about 7 mm. in diameter toward apex; pedicels to 30 mm.
PVT scxoestvatigdesaa repair aah ore aie aa ators ane Haren Gamoade asararen 36. D. coriacea.
7. Drimys vaccinioides Ridley in Trans. Linn. Soc. II. Bot. 9:13. pl. 1, f. 1-6. 1916.
DiIstRIBUTION: Netherlands New Guinea, known only from the type collection, made
by Kloss on the Wollaston Expedition at an altitude of about 3180 m. on Mt. Carstensz.
This species and the following are sharply characterized by their small-
leaved epiphytic habit and small flowers; they are quite unmistakable
among the New Guinean Drimytes. Differences between the two species
are chiefly of degree, but the fact that D. vaccinioides has 2 petals and
D. microphylla 4-6 petals seems to indicate that they are not conspecific.
8. Drimys microphylla A. C. Sm. in nie Arnold Arb, 23: 418, 1942. Fic. 3, f-i.
DistriBUTION: Netherlands New Guinea, known only from the type collection,
Brass 12006 (A), from the Idenburg River region at 1800 m.
9. Drimys buxifolia Ridley in cor Linn. Soc. II. Bot. 9: 13. 1916; A. C. Sm. in
Jour. Arnold Arb, 23: 419
Drimys hatamensis sensu F. v. ey in Trans. Roy. Soc. Vict. 1(2): 1. 1889; non

ecc.

DistriBpuTION: Netherlands and British New Guinea, known from the type collec-
tion (Kloss, Mt. Carstensz) and Brass 4239, 4322, and 4602 (all A, NY) from the Cen-
tral Division, British New Guinea; also collected by MacGregor (ex F. v. Muell.) ; alt.
2500-3680 m.

I discussed the variation in this species in 1942 and expressed doubt of
the specific status of the two following entities, having seen authentic
material of neither of them.

10. Drimys Versteegii Diels in Nova Guin. Bot. 14: 77. 1924.

DistriBuTION: Netherlands New Guinea, known pa from the type collection,

oo Gruppe, auf offenem Gelande, bei 3100 m. i. M. (Versteeg apud Pulle n.
a

11. on, reducta Diels in Nova Guin. Bot. 14: 77. 1924.
DISTRIBUTION: Netherlands vs chistes ee bes from the type collection,
“Gipfel des Wichmann-Berges, 3000 m.. . (Pul 976).’

1943] SMITH, OLD WORLD SPECIES OF WINTERACEAE 135

Diels has noted the close relationship of this species and the preceding
to D. buxifolia Ridley. From a comparison of the original descriptions of
the three entities, it is difficult to point out characters which will serve to
separate them.

12. Drimys oligandra A. C. Sm. in Jour. Arnold Arb. 23: 420. 1942.

DistripuTion: Netherlands New Guinea, known only from the type collection,
Brass 12975 (A), from the Idenburg River region at 1300 m
13. Drimys rubiginosa A. C. Sm. in Jour. Arnold Arb. 23: 420. 1942.

DistriBpuTION: Netherlands New Guinea, known from the type collection, Brass
12629 (A), from the Idenburg River region at 2150 m. Probably also represented by
Brass 9104 (A) from Lake Habbema, 3225 m
14. Drimys pittosporoides Diels in Nova Guin. Bot. 14: 76. 1924.

DistriBuTION: Netherlands New Guinea, reported only from - type: eollection,
Lam 2167, “Central-Gebirge, Fuss des Doorman- Gipfels, 3250 m.

This species is known to me only from the original ae from
which I conclude that it is closely related only to the following. It is
possible that the montane small-leaved species of New Guinea will prove
to be less stable than supposed when ample material is available, at which
time the specific lines will need reconsideration.

15. Drimys Brassii A. C. Sm. in Jour. Arnold Arb. 23: 421. 1942. Fic. 3, 0, p.

DistrisutTion: Netherlands New Guinea, in the Lake Habbema and Mt. Wilhelmina
region, alt. 3000-3800 m. Represented by Brass 9068 (TYPE), 9536, 10671, Brass &
Myer-Drees 10126, 10303 (all A), and probably also by Brass & Myer-Drees 10111 and
10309 (both A)

16. Drimys fistulosa Diels in Nova Guin. Bot. 14: 78. 1924.

DistrisuTION: Netherlands New Guinea, reported from two collections, oe 1615
and 1653, “Central-Gebirge, unterhalb des Doorman-Gipfels,” alt. 3250-3500

From the original description this appears to be a very sietinct: species,
characterized by its very narrow revolute-margined leaf-blades and its
numerous petals of diverse sizes.

17. Drimys elongata Ridley in Hook. Ic. Pl. 31: pl. 3051. 1916, in Trans. Linn. Soc.
II. Bot. 9: 12. 1916.

DISTRIBUTION: Netherlands New Guinea, reported only from the collection of Kloss
on Mt. Carstensz, alt. 750-1650 m. In the second publication cited it is implied that
the species is represented by two collections, both apparently unnumbered.

Drimys elongata appears to be a well-marked species, by virtue of its
large flowers and elongate pedicels. Its relationship is presanaably wit
the following.

18. Drimys myrtoides Diels in Bot. Jahrb. 54: 241. 1916.

DistRIBUTION: Northeastern New Guinea, reported from several collections (by
Ledermann, Schlechter, and Schultze Jena) from the Sepik region and the Torriceili
Mts., at altitudes of 800-2070 m. The type collection is Ledermann 12877.

Drimys myrtoides is said to be characterized by its epiphytic small-
leaved habit and its numerous petals of diverse sizes. One would antici-
pate the discovery of this species in the neighboring part of Netherlands
New Guinea, but I feel reasonably sure that it is not represented in the
collections of the Archbold Expeditions.

136 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XxIV

Diels has also proposed a variety gracilis (in Bot. Jahrb. 54: 242. 1916),
based on Legermninn, 8430 (type coll.) and 11424, from the Sepik region
at 1050 m.; it is said to be more slender (‘zierlicher’’) in all parts than
the typical form of the species.

19. Drimys parviflora Ridley in aie Linn. Soc. II. Bot. 9: 12. 1916.

Bubbia parviflora Burtt in Hook. Ic. Pl. 34: sub pl. 33175. 1936.

DistripuTion: Netherlands New Guca recorded only from the type collection,
Kloss, Mt. Carstensz, alt. about 2500-333

Burtt has transferred Ridley’s species to Bubbia without comment, but
the original description contains no suggestion that a species of Bubbia
is represented. The facts that the leaves are very small (2 by 1 cm.)
and that the sepals are 2 and connate seem to indicate the place of the
plant in Drimys. However, Burtt has doubtless seen the specimen and
has a good reason for his transfer; if this is correct, Ridley’s description
must be quite inaccurate. For the time being I treat the species on the
basis of its original description, which leads me to believe that it is a close
relative of the two following.

20. Drimys pachyphylla Diels in Nova Guin. Bot. 14: 78. 1924.

DistripuTion: Netherlands New Gui “Central-Gebirge, am Doorman-Gipfel,
3260 m. u. M.,” Lam 1812. Probably i Lam 1707 from the same locality and
altitude.

21. Drimys Lamii Diels in Nova Guin. Bot. 14: 77, 1924.

DistripuTION: Netherlands New Guinea, “Unterhalb des Doorman-Gipfels, an of-
fenen oe oe der Waldgrenze, 2900 m.,” Lam 1871; also Lam 1928, same lo-
cality, alt.

22. fe perros Pulle in Nova Guin. Bot. 8: 633. 1912; Diels in Nova Guin.
Cy) : 78.

DISTRIBUTION: ie New Guinea, ‘‘auf dem Gipfel des Hellwig-Gebirges in
c. 2000 m. u. M.,” von Romer 1214 and 1318 (ex Pulle) ; same locality, alt. 1800-2600
m., Pulle 585, 734, 735, 918 (ex Diels).

deearding to the descriptions, this species and the following seem well
distinguished by having their leaves at least pseudoverticillate; Diels
remarks that the leaves of D. verticillata are not strictly in whorls.

23. Drimys rosea Ridley in Trans. Linn. Soc. IT. Bot, 9: 11. 1916.

DistrruTion: Netherlands N ep sits cen only from the type collection,
made by Kloss on Mt. Carstensz, “alk about 3030 m
24. Drimys arfakensis ear Phyt. Fl. Arfak Mts, 135. 1917; A. C. Sm. in Jour.

Arnold Arb, 23: 423.

DIstRIBUTION: Nethe a New Guinea, Arfak Mts., alt. 1800-2400 m., known
from Gibbs 5533, the type, and Kanehira & Hatusima 13408 (A).

Not having seen the type collection, I am not certain of the identity of
the Kanehira & Hatusima plant, which differs from the description as noted
in 1942. Although Gibbs states that the flowers of her specimen are
staminate, the fact that she mentions the ovules and does not describe the
stamens leads me to believe that she saw only pistillate flowers. The species
has small flowers and comparatively small leaves, but its affinities seem to
be with my large-leaved group, as indicated in my key to species.

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 137

25. Drimys reticulata Diels in Bot. Jahrb. 54: 242. 1916; A. C. Sm. in Jour. Arnold
423. 1942
DistrRiBuTION: Northeastern New Guinea, Sepik region, alt. 1400-1500 m. (Leder-
mann 12433, type coll.), and the adjacent Idenburg River region of Netherlands New
Guinea (Brass 11857, 12149, 12494 [all A]), alt. 1800-2150 m.
26. Drimys grandiflora Ridley in Trans. Linn. Soc. II. Bot. 9: 11. 1916.
DistripuTion: Netherlands New Guinea, reported only from the type collection,
made by Kloss on Mt. Carstensz, alt. about 3200 m
27. Sacer macrantha A. C. Sm. in Jour. Arnold Arb. 23: 422. 1942.
DistripuTion: British New Guinea, ile only from Brass 4519 (A, TYPE, NY),
Wharton Range, Central Division, alt. 28
28. Drimys piperita Hook. f. in Hook. Ic. Pl. 9: pl. 896. 1852; Becc. Malesia 1: 185.

1907; Merrill & gs in Philip. Jour. Sci. Bot. 5: 349. 1910; Merr. Enum.
Philip. Fl. Pl. 2: 23;

Tasmannia he laa ey in Ann. Mag. Nat. Hist. III. 2: 110. 1858, Contrib. Bot.
1: 140.

Drimys ive F. v. Muell. Pl. Indig. Col. Vict. 1: 21, sphalm for D. piperita.
1860.

Dioecious shrub or small tree, up to 4 m. high or more, the aaa
subterete, striate-rugulose, 2—5 mm. in see near apices, dark brow

urplish, sometimes glaucous when young; leaves scattered along pete:
lets, a Deanna rugulose, shallowly cannot often narrowly winged,
(5-)7 —14 mm. long, usually stout, (0.5—-)1—2 mm. in diameter; leaf-blades

eee or thin-coriaceous, often papyraceous or submembranaceous
when young, dark brown or dark olivaceous when dried, often glaucous
beneath and frequently with a thin wax-like coating (this smooth, dispersed-
punctate, sooner or later completely lost), oblong-obovate or narrowly ellip-
tic, (4-)6-17 cm. long, (1.5—)2.5-6(—7) cm. broad, gradually narrowed

apex, narrowly recurved at margin, the costa shallowly canaliculate or
slightly iced and flattened above, prominent beneath, the secondary nerves
(6—) 10-14 per side, spreading at an angle of 45—65° , sharply raised on both
surfaces or sometimes slightly impressed above, copiously anastomosing
toward margin, the veinlets forming an intricate and obvious reticulum,
usually sharply prominulous on both surfaces, rarely somewhat obscure or
slightly impressed above; flowers usually numerous, aggregated around
growing point of branchlets, at length pseudolateral, solitary or fasciculate
in clusters of 2-4, subtended diss bracts, these often numerous, papyraceous,
oblong- -deltoid, 7-15 mm. , 3-8 mm rae acute, soon caducous, leav-
ing obvious scars; ee anaes 10-37 m long; staminate flowers:
sepals 2, submembranaceous, sparsely ganar broadly elliptic or sub-
oo 4—6.5 mm. long, 3.5-6 mm. broad, obtuse or obtusely apiculate

t apex; petals 6-8 ee to 11), membranaceous, sparsely glandular,
Re 6-12 mm. long and 1.5—4 mm. broad at ‘anthesis, rounded or
obtuse at apex; stamens 25—60, 3—5-seriate, the filaments slightly flattened,

1-3, sterile, obovoid, 1.5—2 mm. long at anthesis, obtuse at apex, short-

138 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

stipitate at base, the stigmatic ridge conspicuous, extending over apex and

own the ventral edge nearly to base; pistillate flowers: sepals as in stami-
nate but slightly smaller; petals 5—7, similar to those of staminate flowers
but not exceeding 7.5 by 2.5 mm.; stamens none; carpels 2—5, obovoid,
22.5 mm. long at anthesis, similar in shape to those of staminate flowers,
the ovules 12~—30; fruits 1—5-carpellate, the carpels ellipsoid, at maturity
5-15 mm. long and 3-8 mm. broad, obtuse to short-stipitate at base, obtuse
or rounded at apex, obviously marked by the stigmatic ridge, the pericarp
carnose, the seeds usually 12-30, castaneous or black, obovoid, slightly or
strongly falcate, 2-3 mm. long, 1-1.5 mm. broad, subacute at base,
rounded at apex.

DistriBuTION: Philippine Islands, Borneo, _ Celebes (also eo from Amboina
by Beccari), at elevations of 1000-2900 m. (up to 3800 m. on Mt. Kinabalu) ; usually
reported as sei in mossy-forest and ie in soon rene The type is from
Mt. Kinabalu, Borne

PHILipPINE ISLANDS: Luzon: Ilocos Norte Prov., Mt. a Ramos 33329
(NY), 33353 (A, GH, UC, US); Benguet Prov., Loker 21 (M, NY, US), Clemens
17150a (UC); Pauai, Sinise 32050 (A, UC); Suyoc to Pauai, phe 4782 (NY. OS):
Mt. Pulog, Ramos & Edano 44911 (UC); Mt. regi Ramos & Edano
40407 (A, UC); Bontoc Prov., Clemens 2326 (UC); . Pukis, Ramos & Edano
37822 (A); Mt. Caua, Ramos é& Edano 38017 (A, Pad Pauai Benguet to Mt.
Data, Clemens 7326 (UC); Nueva Vizcaya Prov., Mt. Alzapan, Ramos & Edano
45739 (UC); Nueva Ecija Prov., Mt. Umingan, Ramos & Edano 26297 (A,
UC, US); Zambales Prov., Ramos 5025 (NY, US); Bataan Prov., Mt. Mariveles
and upper Lamao R., Williams 745 (NY, US), 754 (GH, NY, US), Whitford 149
(NY, US), 1103 (NY, US), Copeland 260 (US), Elmer 6817 (NY), Borden 2093
(NY, US); Rizal Prov., Loher 14054 ‘< 14060 (UC), 14418 (A), Angilog, Loher
5511 (US); Mabiluang, Loher 14442 (A, UC); Montalban, Loher (UC), 12189 (A,
UC); Tabayas Prov., Mt. Camatis, Rake yess (A), 4964 (A); Mt. Binuang, Ramos
& Edano 28572 (A); Mt. Banahao, Gates 7192 (F), Loher 13678 (A, UC); Laguna
Prov., Mt. Banahao, Ramos 19583 (NY, US), Sulit 30071 (UC), Loher 5512 (US);
Camarines Sur Prov., Mt. Isarog, Edano 76264 (NY), 76247 (NY); Mindoro:
Mt. Halcon, Merrill 6134 (NY, US); Leyte: Wenzel 778 (A, GH, M, US);
Negros: Canlaon Volcano, Merrill 248 (US) ; ‘Desmavucte, Cuernos Mts., Elmer
9912 (A, M, NY, US); Min F| anao: Bukidnon Prov., Mt. Lipa, Edano 38561
(A, GH, UC); Mt. Can eh Ramos & Edano 38897 (A), 38905 (A, UC); Agusan
Prov., Cabadbaran, Mt. Urdaneta, Elmer 13799 (A, F, GH, M, NY, UC, US); Davao
Prov., Kanehira 2692 (NY); Mt Apo, Williams 2553 (A, NY), Mearns (US), Cope-
land 1065 (US), Elmer 11410 (M, NY, US), Clemens 15609 (UC), 15610 (

Borneo: British North Borneo: Mt. Kinabalu, Low (UC. TYPE
cCOLL.), Clemens 10564 (A, UC), 10687 (A), ied (A, = 31950 (A, UC), 50632
(UC), 50987 (A), Grtenatt 44 (A), 48 (A), 76 (A); Sarawak: Mt. Murud,
Mjoberg 101 (UC), 102 (UC); Mt. Poi, Fe 193 (A, NY, UC); Mt. Dulit,
Richards 1645 (A), 2507 (A).

CEL Gowa, Lembaja, Neth. Ind. For. Serv. 20554 (A, NY); Gowa to Mt.
ionbelasai Neth. Ind. For. Serv. s.n. (A

Na NAMES: In Philippine dialects, as recorded by Merrill, 1923: Amutitin
(Ieorot), ge (Manobo), inotétan (Igorot), el (Bontok), malagus (Bagobo).

The above redescription of D. piperita, based on abundant material,
seems advisable, since the only other descriptions are those of Hooker and
Miers, both based entirely on the type collection. This is the only Asiatic
species of Drimys known to occur outside of New Guinea and Australia.
In view of the great diversity of the genus in those regions, it is a striking

1943 | SMITH, OLD WORLD SPECIES OF WINTERACEAE 139

contrast to find that the material from the Philippines, Borneo, and Celebes
is remarkably constant in its salient characters. I find no basis for the
further division of D. piperita, although two Philippine specimens not
cited above (Ramos & Edano 30731 [| A, GH, UC, US], from Mt. Madiass,
Panay, and Elmer 7747 [A, M, NY, US], from Lucban, Tayabas Prov.,
Luzon) appear to represent an extreme form. These specimens have
unusually large sepals (up to 8 by 7 mm.), as many as 11 petals (whereas
no more than 8 were found in the remaining material), which are up to 16
by 5.5 mm., and numerous stamens (up to 100). Since these two speci-
mens are otherwise identical with the bulk of the material, it seems likely
that they represent only a local and aberrant form.

The occurrence of D. piperita on Amboina is recorded by Beccari, whose
determination is very likely correct; however, this station should be
verified. It seems possible that the species will also be found on other
high islands in the region. Mention of the occurrence of D. piperita in
New Guinea probably dates from Mueller’s record of it in 1889; I believe
that Mueller actually had specimens of D. hatamensis, for no New Guinean
collections referable to D. piperita are available to me nor were any cited
by Diels.

The relationship of D. piperita is with the New Guinean species with
large leaves and 5 or more petals, especially D. macrantha and D. grandi-
flora, and to a lesser extent D. arfakensis and D. reticulata. Of these New
Guinean species, only D. macrantha is sufficiently similar to D. piperita
to cause any doubt of its specific status; for the time being I believe that
these two species are amply distinguished, but it must be kept in mind
that the interior of New Guinea is still largely unknown and that future
collections may cause students to extend the range of D. piperita to that
island.

Collectors of material of D. piperita have indicated that the petals are
white, the stamens yellow, and the fruit at first red, finally black, with a
deep purple bitter juice.

29. Drimys acutifolia Pulle in Nova Guin. Bot. 8: 633. 1912; Diels in Nova Guin.
Bot. 14: 76. 1924

Distripution: Netherlands New Guinea, “Vorgebirge des Hellwig-Gebirges (Erica-
i m. ii. d. M.,” von Rémer 1044 and 1045 (ex Pulle) ; ‘““Perameles-
Gebirge, 1100 m. ii. M.,” Pulle 482 and 483 (ex Diels).

This species and the two following, according to the descriptions, appear
to differ from each other in minor characters only. A comparison of the
various collections is highly desirable.

30. Drimys Beceariana Gibbs, Phyt. Fl. Arfak Mts. 133. f. 9. 1917; Diels in Nova
Guin. Bot. 14: 75. 1924.
STRIBUTION: Netherlands New Guinea, reported only from the Arfak Mts. as
represented by Gibbs 5651 (type coll.) and Gjellerup 1204, alt. 2500-2700 m.
31. Drimys cyclopum Diels in Nova Guin. Bot. 14: 76. 1924.
T ton: Netherlands New Guinea, apparently known only from the type
collection, Gjellerup 549, from the Cyclops Mts., alt. 1800 m.

Reports of this species from British New Guinea are referred to D.

hatamensis Becc.

140 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

32. Drimys densifolia Ridley in Trans. Linn. Soc. II. Bot. 9: 12, 1916.
DistripuTion: Netherlands New Guinea, recorded only from the type collection,
made by Kloss on Mt. Carstensz, alt. about 3175-3330 m
This species is presumably amply differentiated from its allies by having
its leaf-blades obtuse and auriculate at base, somewhat like those of the

Australian D. insipida (R. Br.) Pilger, which in other respects is not a

very close relative of D. densifolia.

33. Drimys obovata A. C. Sm. in Jour. Arnold Arb. 23: 424. 1942. Fic. 3, j-n.

DistrIBUTION: Netherlands New Guinea, vicinity of Lake Habbema and the Bele

River, alt. 2200-2800 m., represented by Brass 10567, 10570, 11295 (type), and 11312

(all A)

34. Drimys hatamensis Becc. Malesia 1: 185. 1877; Parment. in Bull, Sci, Fr. &
Belg. 27: 227, 301. pl. 10, f. 38. 1896; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108.
1906; Diels in Bot. Jahrb. 54: 242. 1916; A. C. Sm. in Jour. Armold Arb. 23: 425.

2.

194
scl —— sensu F. v. Muell. in Trans. Roy. Soc. Vict. 1(2):1. 1889; non
Hoo

brimss copa sensu Lane-Poole, Rep. For. Res. Papua 86. 1925; White & Fran-
iels.

n Proc. Roy. Soc. Queensl. 38: 228. 1927; non
een ‘Nethrands Pier aaah and British New Guinea. The type is a
Beccari collection from the Arfak Mts. at 2000 m.; also from the Arfak Mts. are

Kanehira & Hatusima 13785 ae 13935 (both A). For the probable occurrence of this
species elsewhere | in New Guinea and discussions of its status, see Diels in 1916 and my
notes in

35. Drimys dictyophlebia Diels in Nova Guin. Bot. 14: 75. 1924.

DistripuTIon: Netherlands New Guinea, represented by the type collection, Pulle
845, from the Hellwig Mts. at 1900 m., and probably also by Brass 13704 (A), from
the Idenburg River region at 700 m.

36. 7 coriacea Pulle in Nova Guin. Bot. 8: 634. 1912; Diels in Nova Guin.
Bot. 14: 75, 1924

DISTRIBUTION: Neth erlands New Guinea, Hellwig en ig 2000-2600 m., von
Rémer ae 1281 (ex Pulle), Pulle 577, 595, 958, 959 (ex Diels)

According to the original description, this species is ae characterized
by its thick-coriaceous leaves and very stout branchlets.

Ov_p WorLD SPECIES OF DRIMYS EXCLUDED FROM THE GENUS
Many workers in this group have accepted Drimys in a very broad sense,
taking its limits to be essentially those of the family. Therefore, practi-
cally all of the early species were first described in Drimys ; these binomials
are referred to the appropriate species of the other five genera in the
following pages. The species listed immediately below, however, should
apparently be removed from the family altogether.
Drimys MUvELLeri Parment. in Bull. Sci. Fr. & Belg. 27: 227, 300. pl. 10, f. 36, 37,
nomen subnudum. 1896; Vickery in Proc. Linn. Soc. N. S. Wales 62: 83. 1937.
Drimys intermedia Parment. in Bull. Sci. Fr. & Belg. 27: 223, 224, sphalm for D.
Muelleri. 1896; Vickery in Proc. Linn, Soc. N. S. Wales 62: 83. 1937.
The secondary wood of this species does not suggest a species of Drimys,
as pointed out by van pupil
esc OBLONGA S. Moore in Jour. . 59: 302. = Hypsophila ae
v. Muell. (Caan waste to Ea in ae Bot. 71: 45.

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 141

2. BUBBIA
rene v. Tiegh. in Jour. de Bot. 14: 278, 293. 1900; Pilger in E. & P. Nat. Pf.
htr. 2: 108. 1906; Hutchinson in Kew B Bull. 1921: 190, 1921; ee in Jour.
ba 72:40. 1934; Vickery in Proc. Linn. Soc. N. S. Wales 62: 83. 1937.
ete a Eububbia v. ae in Jour. de Bot. 14: 294. 1900; ie in E. & P.
. Nachtr. 2: 109. 190
ies ae Monoclada v. Tish in Jour. de Bot. 14: 294. 1900; Pilger in E. & P.
Nat. Pfl. Nachtr. 2: 10
se Sect. Diploclada v. Tish in Jour. de Bot. 14: 294. 1900; Pilger in E. & P.
. Nachtr. 2: 10
Perth Lauterb. in ey on & Lauterb, FI. peta Schutzg. Siidsee Nachtr.
; Engl. in E. & P. Nat. Pfl. ed. 2. 21: 1925

ane is here interpreted to include 30 species, which are distributed
eight in New Caledonia, one in Lord Howe Island, two in Queensland, and
nineteen in New Guinea. The first species referable to the group was
described in 1869 by Mueller (Drimys Howeana) and this is also the
genotype. The genus is readily separated from Drimys on characters per-
taining to the calyx, from Zygogynum by its separate carpels, from Pseudo-
wintera by its terminal inflorescence, from Belliolum by its stamens, and
from Exospermum by its placentation and free carpels (which are rarely
appressed-contiguous in young flowers). In many respects Bubbia appears
to have retained the hypothetical primitive characters of the family better
than the other genera, although its carpellary characters are diverse and
probably more highly evolved than those of Drimys Sect. Tasmannia.

Van Tieghem’s three sections are based on the greater or lesser degree of
branching of the primary rays of the inflorescence, a character which
appears of no more than specific value. I find it impossible to propose
sectional segregations within the genus, as the inter-relationships of the
various species are highly complex. Possibly a classification may eventu-
ally be based upon characters of the carpel, such as the position and extent
of the stigmatic ridge and the extent of placental areas. The key to the
New Guinean species proposed below is entirely artificial.

SPECIES OF NEw CALEDONIA AND Lorp Howe IsLanp

On the basis of herbarium material available in America, it is impossible
properly to evaluate the eight species of Bubbia reported from New Cale-
donia and the species from Lord Howe Island. The original descriptions
of these species are for the most part inadequate, and a consideration of
their status must await examination of the collections in the herbaria at
Paris and the British Museum. However, a few brief notes on these
entities are given below, in order to bring together references to literature.
For lack of a better method, I discuss the names in chronological order.

1. Bubbia Howeana (F. v. Muell.) v. Tiegh. in Jour. de Bot. 14: 293. 1900; Pilger
in E. & P. Nat. Pfl. Nachtr. 2: 109. 1906; Vickery in Proc. Linn. Soc. N. S.
Wales 62: 84. 1937

Drimys Howeana F. v. Muell. ae Phyt. Austral. 7: 17. 1869; Parment. in Bull.

Sci. Fr. & Belg. 27: 230, 307. 18
Drimys insularis Baill. ex F. v. Nail Fragm,. Phyt. Austral. i Ba nomen. 1875;
Parment. in Bull. Sci. Fr. & Belg. 27: 230, 307, as synonym.

142 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Bubbia nih - Tiegh. in vee ey — 14: 293, nomen. 1900; Pilger in E. & P.
P r. 2: 109, nom

mriocniaet Howe cates sone aan by C. Moore.

I have seen no material of Bubdbia from Lord Howe Island, and the origi-
nal description of Drimys Howeana by Mueller does not permit the accurate
placing of the species, although its generic identity is beyond doubt. This
species was selected by van Tieghem as the type of Bubbia. Drimys
insularis has never been adequately described, and I do not question
Vickery’s reference (7: 84) of it to synonymy under Bubbia Howeana.
Bubbia Muelleri was named, but not described, by van Tieghem on the
basis of Mueller’s discussion (Fragm. Phyt. Austral. 7:17. 1869) of a
second plant from Lord Howe Island, known in fruit only and said by
Mueller to be perhaps conspecific with Drimys Howeana. The name
Bubbia Muelleri has no status of consequence, and until collections offer
proof to the contrary, one may assume that there is only one species of
Bubbia on Lord Howe Island.

2. Bubbia Balansae (Baill.) v. eo in Jour. de Bot. 14: 293, 1900; Pilger in E. &
P. Nat. Pfl. Nachtr. 2: 109.
Drimys Balansos erm in "Adansonia 10: 335. 1873; Guillaumin in Ann. Mus. Col.

Marseille II. 9:
DistRIBUTION: New ae nia, apparently known only from the type collection,
Balansa 1844, from Mt. Humboldt, alt. 1100 m

Van Tieghem (6: 294) proposes his Seve. Monoclada on this species
alone. The species is said to be characterized by its very small leaves and
flowers and the small number of its inflorescence-rays, each of which bears
only two flowers at its summit.

3: = Deplanchei v. oe in Jour. de Bot. 14: 293. 1900; Pilger in E. & P.

Pfl. Nachtr. 2: 109. 1906.
Brion hhc abe ex v. Tiegh., J. c., as synonym.
DistRIBUTION: New ries thus far reported only from the type collection,

Vieillard 2290; from Wagap
This species is ae Paarieneed by van Tieghem as having its flowers
arranged in a simple umbel; it was the only species of the genus known to
him with this character and he placed it alone in his Sect. Eububbia. This
choice of a sectional name is unfortunate, since elsewhere (6: 278) van
Tieghem clearly states that the type-species of the genus Bubdia is
B. Howeana.
4. ~— auriculata vy. Tiegh. in _ de Bot. 14: 293. 1900; Pilger in E. & P
at. Pfl. Nachtr. 2: 109. 1906; A. C. Sm. in Jour. Arnold Arb. 23: 438. 1942.
ao ev geammiigil ae ex A aeaved in Bull. Sci. Fr. & Belg, 27: 231, 308. pl.
10, f. 34, nomen subnudum. 1896; Vieill. ex v. Tiegh. in a - Bot. 14: 293,
192

Bubbia gets Dandy in Jour. Bot. 72: 40. 1934; elie in ok Ic. Pl. 34:
sub pl. 3315
DistriputTion: New Caledonia, reported from the type collection, Vieillard 2280
(GH) fro a and also from Compton 1551 (ex Bak. f.) or 1581 (ex Dandy)
from ae
This is one of the species which Burtt (1) believes to weaken the generic

1943 | SMITH, OLD WORLD SPECIES OF WINTERACEAE 143

distinctions between Bubbia and Belliolum. I have recently (3: 438) dis-
cussed the points raised by him and also questioned Dandy’s acceptance of
Parmentier’s specific epithet. Bubbia auriculata is readily distinguished by
its long narrow subsessile leaf-blades with auriculate subamplexicaul bases.
5. Bubbia heteroneura vy. Tiegh. in Jour. de Bot. 14: 294. 1900; Pilger in E. & P.
Nat. Pfl. Nachtr. 2: 109. 1906
Drimys heteroneura v. Tiegh. ex ate f. in Jour. Linn. Soc. Bot. 45: 267. 1921.
DistrIBUTION: New Caledonia, recorded by van Tieghem from Vieillard 20 (type

coll.) and Deplanche 293, is ag from Puepo; doubtfully reported by Baker from
Compton 1130 from Mt. Can

Van Tieghem’s en is very inadequate, but, since he mentions a
few details of the leaf and cites specimens, the publication must be con-
sidered valid.
6. Bubbia isoneura v. a in Jour. de Bot. 14: 294. 1900; Pilger in E. & P. Nat.

htr. 2: 109.

Seen New Pasi reported only from the type collection, Vzeillard 17
(GH), from Wagape.

From the original brief descriptions of the two species, this and B.
heteroneura appear to be weakly differentiated. Burtt (1) has reduced
B. isoneura to Bubbia crassifolia (i.e. Belliolum crasstfolium). The avail-

sponding to the obliquely ventral stigmatic ridge. In this character,
therefore, the plant would appear to be properly placed in Bubbdia, although
verification from staminal characters is desirable. In the carpels of
Belliolum crassifolium, as represented by Schlechter 15348, the ovules are
in two rows which are slightly removed laterally from the ventral suture of
the carpel. This tendency is probably characteristic of Belliolum, to which,
on the basis of its stamens, Schlechter 15348 belongs. The two specimens
under discussion show slight intangible differences in foliage; on the basis
of the evidence now available I cannot agree with Burtt’s reduction, and
therefore I list Bubbia isoneura as an independent species.

7. Bubbia Comptonii (Bak. f.) Dandy in Jour. Bot. 72: 41. HoH

Drimys Comptonii Bak. f. in Jour. Linn. Soc. Bot. 45: 267. 19

DistripuTion: New Caledonia, reported only from the type en Compton
1815, from Mt. Panié, alt. 750-1200 m.

In describing three new species of this complex from New Caledonia,
Baker, although his descriptions are fairly complete, omits the staminal
details which are necessary to distinguish species of Bubbia from those of
Belliolum. In transferring the specific epithets to Bubbia, Dandy has
doubtless considered this point and has found that Baker’s species belong
in Bubbia in van Tieghem’s original sense, not as later expanded by Burtt.
It is significant that Baker does not compare any of his three species with
any of van Tieghem’s; consequently one should see type material of all
the New Caledonian species before reaching conclusions as to their specific
status.

According to the original description, B. Comptonii is characterized by
having its stamens often reduced to 3 and its carpel usually solitary.

144 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

8. Bubbia odorata (Bak. f.) Dandy in Jour. Bot. 72: 41. 1934.

Drimys odorata Bak. f. in Jour. Linn. Soc. Bot. 45: 268. 1921.

DistripuTIOoN: New Caledonia, reported only from the ty be a Compton
1982 (ex Baker) or 1983 (ex Dandy), from Tonine, alt. abov

Baker remarks, “‘The noticeable features of this species are nite oblanceo-
late glabrous leaves with a thick midrib, the flowers with a strong sweet
scent, the petals being white with a purple patch towards the base, the
generally 5 stamens and 2 carpels.”’

9. Bubbia pauciflora (Bak. f.) Dandy in Jour. Bot. 72: 41, 1934,

Drimys pauciflora Bak. f. in Jour. Linn. Soc. Bot. 45: 268. 1921.

ISTRIBUTION: New Caledonia, reported only from the type aes Compton
1761 (ex Baker) or 1768 fax Dandy), from Mt. Panie, alt. about 450 n

Baker remarks, “Easily distinguished by the slender leaves euch attenu-
ate below, with very slender indistinct lateral nerves, the few flowers on
long pedicels, and generally 3 carpels.”

AUSTRALIAN SPECIES

The genus Bubba apparently has a very limited range in Australia, thus
far being known only from eastern Queensland approximately between
latitudes 16° and 18°. To the single species previously reported from
this region, B. semecarpoides, I here add a second, which is probably a
montane derivative from a common ancestor. The two species are essen-
tially similar in fundamental details, but I believe that the points brought
out in the following key are of specific value. The specimen from Bellenden
Ker which Domin refers to Drimys semecarpoides should be re-examined
in comparison with my new species. As the descriptions of Mueller and
Bailey are incomplete, I have redescribed B. semecarpoides from more
recent material.

The two Australian species do not show a close affinity to any of the
New Guinean species, having floral characters somewhat resembling those
of B. oligocarpa (Schlecht.) Burtt and its allies, but in foliage they are
more suggestive of some of the New Caledonian species.

KrEyY TO THE AUSTRALIAN SPECIES

Petioles stout, 2-3 mm. in diameter, 15-27 mm. long; leaf-blades 12-20 cm. long,
3-7 cm. broad, the costa impressed above; inflorescence compound, the primary
wees 2- or 3-flowered; stamens 25-32; carpels 5-8 per flower, Meg ovules 10-16;

eeds usually 10-12 at maturity ......... 0.0.0... 0000 eae B. semecarpoides.
ee slender, 1-1.5 mm. in diameter, 5-20 mm, long; leaf- in 6-12 cm. long,
1.5—4.5 cm. broad, the costa subplane aes hae Rea a the flowers rarely
paired on short peduncles; ens 19; carpels 3 or 4 per flower, the ovules
7-10; seeds 3-6 at abe igs Scie a anetavnahe Wanches uNh shag aa eon aioe « 11. B. Whiteana.

10. eam semecarpoides (F. v. Muell.) Burtt in Hook. Ic. Pl. 34: sub pl. 3315.

ae oe F. v. Muell. in Vict. Nat. 8:15. 1891, in Bot. Centralbl.
46: 204. ; F. M. Bailey, Queensl. Fl. 1: 19. 1899, Compr. Cat. Queensl. PI.

21. 1913; pee in Bibl. Bot. 22[Heft 89]: 115. 1925.
Tree to 20 m. high, the branchlets stout (4-7 mm. in diameter toward
apices), subterete, brownish or cinereous; leaves aggregated toward apices
of branchlets, the petioles rugulose, semiterete, 15-27 mm. long, stout (2—3

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 145

mm, in diameter), the leaf-blades subcoriaceous or chartaceous, dark oliva-
ceous or brownish when dried, concolorous or conspicuously glaucous be-
neath, oblong- or elliptic- obovate, 12-20 cm. long, 3—7 cm. broad, gradually
narrowed to an attenuate base and decurrent on the petiole, rounded or
broadly obtuse at apex, narrowly recurved at margin, especially toward
base, finely rugulose on both surfaces, the costa impressed above, prominent
beneath, the secondary nerves 10— 18 per side, erecto-patent at an angle of
45— 60°, obscurely anastomosing toward margin, slightly prominulous on
both surfaces or nearly obscure, the veinlets immersed; inflorescence pseudo-
sich sessile, the primary rays apparently about 4, 2.5-8 cm. long includ-

g flowers or ‘fruits, once- or rarely twice- branched, 2- or 3-flowered,
er ee papillose, the bracts and bracteoles soon caducous, the pedicels
3—5 mm. long before anthesis, up to 15 mm. long in fruit; calyx papyraceous,
rotate, deeply 2—4-lobed, the lobes sparsely glandular, ovate-deltoid, 2—3
mm. long and broad, obtuse: petals immature in our specimen (coll. White)
but apparently several, carnose; stamens 25-32, 2- or 3-seriate, up to 1.3
mm. long (immature), the filaments subcarnose, flattened, ‘broadened
distally, sparsely yellow-glandular, the locules apical, obliquely horizontal,
0.3-0.5 mm. long; carpels 5-8, obovoid, 1.5-2 mm. long slightly before
anthesis, contracted toward base, the stigmatic ridge subapical, 0.3-0.6 mm.
long, the ovules 10-16, on short ventral-apical placentas; carpels in fruit
5-8, usually 3 or 4 maturing, the others abortive, occasionally only
maturing: mature carpels obovoid, up to 12 mm. long (excl. stipe) and
10 mm. broad, the basal stipe stout, about 2 mm. thick and long, the apex
rounded, the stigmatic ridge inconspicuous, subapical; pericarp coriaceous,
1—-1.5 mm. thick, obscurely rugulose without; seeds usually 10-12 at ma-
turity, closely appressed, oblong-obovoid, slightly falcate, about 5 mm. long
nd 2.5 mm. thick, subacute at base, rounded at apex

DISTRIBUTION: Northeastern Queensland, apparently limited to the region from the
Atherton Tableland to the vicinity of Rockingham Bay, lat. about 17-18°, at alti-
tudes up to 700 m. The type was collected by W. Sayer on “Russell’s Creek,” a locality
I have not located on modern maps but which is probably near Rockingham Bay.
Other collections from this region which have been cited by Bailey and Domin were
made by Dallachy, W. Hill, and Domin, the latter from Bellenden Ker

onjie, ego Tableland, White (A); East

Malanda, Atherton Tableland, Kajewski ay (A, NY) (common in rain-forest).

Although I have not seen authentic can of Mueller’s species, it is
obvious from his original description that he had the large-leaved species
described above and not the following.

11. Bubbia Whiteana sp. nov. Fic. 4, a-f.

Arbor ad 8 m. alta, ramulis subteretibus brunneis crassis, apicem versus
3—4 mm. diametro; foliis secus ramulos copiose dispersis; petiolis rugulosis
corifereuiie 5-20 mm. longis, 1-1.5 mm. diametro; laminis coriaceis sic-

citate olivaceis, utrinque conspicue rugulosis, subtus plerumque glaucis,
elliptico- obovatis, 6-12 cm. longis, 1.5—4.5 cm. latis, basi attenuatis et in
petiolum decurrentibus, apice rotundatis vel obtusis, margine recurvatis vel
conspicue revolutis, costa supra subplana subtus prominente, nervis secun-
dariis utrinsecus 8-15 obscuris angulo 45—60° a costa abeuntibus, rete
venularum immerso; inflorescentia pseudoterminali simplici, floribus 3—6
apicem ramulorum circa dispositis raro pedunculo brevi binis, pedicellis

146 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

gracilibus papillosis sub anthesi 8-18 sub fructu ad 25 mm. longis; calyce

apyraceo rotato profunde 2- vel 3-lobato, lobis parce glandulosis ovato-
deltoideis, 2—2.5 mm. longis, 2.5—4 mm. latis, apice obtusis; petalis 6 vel 7
subcarnosis oblongis vel obovato-oblongis, apice rotundatis vel obtusis, ex-
terioribus sub anthesi 6—6.5 mm. longis et 2.5-3.5 mm. latis, interioribus
paullo minoribus; staminibus 18 vel 19, 2-seriatis, sub anthesi 1.5-2 mm.

Fic. 4. a-f. Bubbia Whiteana, drawn from the type: a. fruiting branchlet, x 3;
b. flower, with two petals removed, x 2; c. stamens, extrorse and introrse views, x 5;
d. carpel, X 5; e. fruit, with one mature carpel, x 1; f. seed, x 2. g-i. Bubbia
Clemensiae, drawn from the type: g. inflorescence and leaf, x 4; h. flower, past
anthesis, with petals and stamens fallen, x 1; i. stamens, introrse and extrorse views,
x 3. j-m. Bubbia Archboldiana, drawn feat the type: j. flowering branchlet,

k. flower, with two petals samoyrel. x 14; 1. stamens, extrorse and introrse views, < 5;
m. carpel,

longis, filamentis complanatis apicem versus incrassatis, pas aie
0.3-0.5 mm. longis, oe vel leviter obliquis; carpellis 3 vel
obovoideis at anthesi 1.5—2 mm. longis, basi contractis, carina aie
subapicali 0.3-0.5 mm. longa, ovulis 7— 10, placentis ‘brevibus ventrali-
apicalibus; carpellis fructiferis maturitate 1-3 subglobosis vel obovoideis,
8-11 mm. diametro, basi breviter stipitatis (stipite circiter 2 mm. longo et
diametro), apice rotundatis, carina stigmatum obscura brevi subapicali,
pericarpio carnoso demum coriaceo 0.5—2 mm. crasso extus obscure rugu-
loso, seminibus maturitate 3-6 nigris oblongo-obovoideis leviter falcatis,
4.5—-5 mm. longis, 2—2.5 mm. crassis, basi obtusis, apice rotundatis.
DISTRIBUTION: Known only from the type locality in northeastern Queensland, lat.
about 16° 15’.
AustraLtia: Queensland: Thornton Peak (Mt. Alexander), Daintree River

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 147

a alt. 1200-1350 m., Brass 2278 (A, Type), Mar. 14, 1932 (small tree of the low

rubs near summit; jenues silver-gray beneath, much recurved at margin; flowers
ee on red pedicds: fruits black), Kajewski 1495 (A, NY) (small gnarled tree up
to 8 m. high, common in poor scrub on top of mountain; leaves silver beneath; ped-
icels brown; petals cream-green; fruits black when ripe).

Although the new species is doubtless a close relative of B. semecarpoides,
I believe that it is worthy of specific rank on the basis of its substantially
smaller leaves, simpler inflorescence, and fewer stamens, carpels, ovules,
and seeds. The species is named for Dr. C. T. White, of the Botanic
Gardens of Brisbane, who has contributed much to our knowledge of
Queensland plants.

NEw GUINEAN SPECIES

Bubbia apparently reaches its greatest development in New Guinea,
where 19 species are now known. ‘The total variability of the New Guinean
population considerably exceeds that of the New Caledonian, as indicated
in the extremes of carpel-structure, fruit-size and shape, number and surface
of seeds, number of stamens, and types of foliage. Many of the New
Guinean species are known from single collections, and some of these I
have not seen; therefore the following key is based to a certain extent
on descriptions, but in general these are ample. This treatment will need
considerable revision when more ample material is available, for it seems
certain that additional species will be discovered and that some of the
existing ones will need amplification. Bubbia is less important than Drimys
in New Guinea as an element of the vegetation, if one may judge from the
existing collections. It occurs at lower elevations, usually between 400
and 2800 m., but sometimes as low as 100 m. and in one species as high
as 3600 m. One may assume, from field notes that the species are usually
of scattered occurrence and are never dominant, as are certain species of
Drimys at high elevations.

Key TO THE NEw GUINEAN SPECIES

Leaf-blades comparatively small, 6.5-14 cm. long, 2.5-4.5 cm. broad; inflorescence com-
paratively few-flowered, the primary rays simple or once- or twice-branched.

Lateral nerves of the leaf-blades inconspicuous, prominulous or slightly prominent,

the blades not bullate.

ma nee simple, about 3-flowered ; anita about 3-lobed; petals about 4 mm.

long; carpel 1; leaf-blades papyraceous ............... 12. B. Ledermannii.

mie once- or twice-branched, seer flowered; calyx s- 7-lobed; petals

8-11 mm. long; carpels 3 or 4, adnate at anthesis, at mee oe leaf-blades

thick=coriaceOus: oc:<2-45, 40 sie pete ee ore gio ye ain a) slo ake B. pachy antha.

Lateral nerves of the leaf-blades whibasd prominent beneath a impress above,

blades: bullate-and coriaceous... cscs onc corer ee eas ones 14. B. bullata.

Leaf- sues larger, 14-40 cm. long, 5- . cm. broad, rarely slightly smaller ; ee

ny-flowered, the primary rays 2- or 3-times branched (essentially simple in

renee nerves of the leaf-blades leaving the costa at an angle of 50-70(-75)°.
Primary lateral nerves of the leaf-blades 20-40, the blades aeons Siete:
15. B. polyneura.
Primary lateral nerves of the leaf- as 8-20 (about 22 in nos. 20 ond 22), the
blades coriaceous or chartaceou

Oo 6 eee: OOLE Eh Ret hae byes 18) 40 Seo OST S68 S18 ew Se oe Bele ee 6 8 66 Be 8 Ole

148 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Flowers large, the petals usually 6 in number, 11-17 mm. long, 5-12 mm. broad;
calyx valent 6- KG ae stamens 100-125, 5- or 6-seriate; primary rays
of -infloréscence: 3°08 40.4 45.) 00s s Gass esse eee abieaa Se 16. B. Clemensiae.

Flowers arma oe eit less than 10 mm. long; calyx 2- or 3- lobed (sub-

; stamens not more than 35, usually 2- or 3-seriate.
Petals 5- 10 stigma strictly apical, not extending down the ventral edge of
the carpel.
faa rays of inflorescence 1-4; petals apparently 5-8, hated ex-
ceeding 8 mm. in length and 3.5 mm. in breadth; stam
sage 1 or 2; stamens mare 17 or 18 Gadine 10- ae cnr wlkdes
cm. lo 7-12 cm. :
Cal 3-lobed; petals 6 or 7, about 6 mm. long and 3.5 mm. broa
nther- sae eh ebliaueW apical-lateral; carpels 2 ..17. B. raed
nes essentially ees and entire at margin; petals 5, not more than
5 g and 3 mm. broad; anther- locules Pace —
SOUCATY: ccccdetind epee eae e ee se baie aerate apace es 18. B. m arpa.
Carpels 3 or 4; stamens 12--16, the anther- ce maria 1,
Leaf-blades 15-18 cm. long, 4-6 cm. broad; petals 8, the outer ones
ree : 5 mm. long and 2 mm. eee Fane a 12; carpels
AONALC sca. ciae nes acetaae ee isaeieeseasnesee . B. montana.
Leaf- hee pee cm. long, 9-11 cm. broad, ued and farinose-
ceriferous beneath; stamens 14-16; carpels 3, free; fruit sub-
Ne up to 3 cm. in diameter, the seeds conspicuously plicate-
PUBOSE ea hod 6p ous e poieiae BAGS s Bee Eee ay ora a ee ces 20. B. longifolia.
Primary rays fe inflorescence 6-8; petals 8-10, the outer ones 8-10 mm. long
and about 6 mm. broad; stamens 22-35, the anther-locules ener
carpels 3-5, free; leat- blades 14-22 cm. jong, 5-7.5 cm. broa
ed ave atdr any Balen pie apdl dere Seas 4 Sa ats ate tele ate ere ghens ara ese 21. B. sylvestris.
Petals 4 or 5 (not known in nos. 23 and 24); stigma apical and also extend-
ing at . part of the way down the ventral edge of the car
Primary lateral nerves of the leaf-blades about 22; flowers _ the petals
mm. long; primary rays of inSosesrenns 5 or 6; stamens about
20} Carpels, 3 OF 4 2s... iwdwqacnatw sieeve aa wae ca od 22, umbellata.
Primary lateral nerves of Baas leaf-blades 8-15; flowers larger, the petals
probably at least 5 mm. lon
Petioles (0.8—)1-2 cm. i. ong; stigmatic ridge inconspicuous, apparently
occupying less than half of both apical and ventral faces of carpel,
3-5 mm. long in fruit; seeds 2-11; rays of inflorescence 3-6
Leat- blades coriaceous, the veinlets asially immersed; rays of fruiting
inflorescence 2- or 3-times branched; carpels 4—6
ia’seG ad centage Wate 6 HARES A Rie Sete Vea aineinseos bs SOEMUUT BENS:
Leaf-blades chartaceous, the veinlets usually prominulous on both s
aces; rays of fruiting inflorescence essentially simple or once-
branched; carpels probably 2 in flower, often solitary in : uit .
shssaisiu ore eats atheig sich Sie Sid idle arsine @arb.e7euie otedsrate were esee 24, elon ca,
Petioles mare less than 1 cm. long; . ridge elongate, ab eon
both apical and ventral faces of c
Pricey rays of inflorescence 6— 11; epee . or 10; stigmatic ucla

locule straight, the ovules about 16, biseriate ..25 rchboldiana
Primary rays of inflorescence 3-5; carpel 1; stigmatic ridge extending

curved, the ovules or more, several-seriate; fruit to 4 cm. b
m., the seeds numerous ..............00000005 26. B. megacarpa.
Lateral nerves of the leaf-blades widely spreading, leaving the costa at an angle of
70-85°; venation conspicuous on both surfaces; stigmatic eta strictly apical.

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 149

Petioles 1-1.8 cm. long; leaf-blades comparatively narrow, 4-6.5 cm. broad, the
primary lateral nerves 15-20 per side; rays of inflorescence 10-13

al nerves more aes 20; rays of inflorescence appantntlg less than 8.
a blades oblanceolate, 6-10 cm. broad, the primary lateral nerves 25 or fewer
per side; petioles 1.5—2.5 cm. long.
Primary rays of ee stout, to 12 cm. long; stamens ie -30; petals

Fs—3 (0) a yoed (0) 1 alee Ree ae ae ee anne ad PEE or 28. Pvcaicthyrsa.
Primary iadg of inflorescence slender, 12-15 cm. long; stamens 12-22; petals
[11 (0) 1 | en eee serie ae 29. B. sororia.

Leaf- “iaes | siete. elliptic, 10-14 cm. broad, the primary lateral nerves 25-35
side; petioles about 3 cm. long; primary rays of inflorescence to 15 cm

Eee ee eateries ae 6 5:5 's, 05 28:8 5's Bein oe 30. B. calophylla.
12. Bubbia Ledermannii (Diels) Burtt in Hook. Ic. Pl. 34: sub. pl. 3315, as B.
Ledermanni. 1936.

Drimys Ledermannii Diels in Bot. Jahrb. 54: 243. 1916.
DISTRIBUTION: Northeastern New Guinea, ns i from Ledermann 8990 (type
coll.) and 8973, from the Sepik region, alt. about
13. Bubbia pachyantha A. C. Sm. in Jour. Arnold Arb. 23: 428. 1942.
DistrRIBUTION: British New Guinea, known only from the type ornare hag
4371 (A, TYPE, NY), from Mt. Albert Edward, Central Division, alt. 3550-3
14. Bubbia bullata (Diels) A. C. Sm. in Jour. Faas Arb. 23: 426. 1942.
se bullata Diels in Bot. Jahrb. 54: 243
STRIBUTION: Northeastern New Guinea, ee vied from the type collection,
Pe ati eee 342, from the Kaiserin aceasta River region
15. Bubbia polyneura (Diels) Burtt in Hook. Ic. ha sub. pl. 3315. 1936.
Drimys polyneura Diels in Bot. Jahrb. 54: 244.
DistriBuTIoN: Northeastern New Guinea, nee only as the type collection,
Ledermann 8986, from the Sepik River region, alt. about
16. Bubbia Clemensiae A. C. Sm. in Jour. Arnold Arb. 23: 431. 1942. Fic. 4, g-i.
ISTRIBUTION: Northeastern New Guinea, known es from Clemens 4596 (A) and
5157 (A, TYPE), from the Morobe District, alt. 1750-1
17. Bubbia oligocarpa (Schlecht.) Burtt in Hook. Ic. Pl. 34: sub pl. 3315. om
Drimys oligoca fie a cht. in Bot. Jahrb. 50: 71. f. 7. 1913; Diels in Nova Gui
B

STRIBUTION: eid and Netherlands New Guinea, reported originally from
Schlechter 16470 (type coLy., UC), from Wobbe, Northeastern New Guinea, alt. about
400 m., and ce 281, from Taua, Netherlands New Guinea. Diels adds the
following records: m 794, 1165, and 1225, from the Mamberamo region of Nether-
lands New Guinea, — Fey 2CO m

In keying this species I fage relied upon the original description and the
type collection, which has about 18 stamens. Diels reports that the
Lam collections have 10-20 stamens and have leaves which are variable
in width. Even including this variation in one’s concept, the ae
remains clearly separable from its closest allies, which are the thre
following species in this treatment.

18. Bubbia monocarpa A. C. Sm. in Jour. Arnold Arb. 23: 428. 1942.

ISTRIBUTION: Netherlands New Guinea, known only ie Kanehira & Hatusima
12105 (A, TYPE), from Dalman, inland from Nabire, alt. 400

150 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

In this species and its close relatives (nos. 17-21 in my key) the stig-
matic ridge is strictly apical and the ovules are pendulous.

. Bubbia montana (Lauterb.) A. C. Sm. in Jour. Arnold Arb. 23: 426

Be ree montanus Lauterb. in K. Schum. & Lauterb. FI. ice ite

iidsee oe 319. 1905, in Bot. Jahrb. 58: 1S. f. 4.1922; Engl. in E. & P. Nat.
Pfl. ed. 2. 21: 229. f. 100. 1925; Burtt in Kew Bull. 1938: 458. 1938.

Di BUTION: Northeastern New aa sone only from the type collection,
Suhles bles 13984, from the Bismarck Mts., a 20

The monotypic genus ee rial placed in the Guttiferae,
was first referred to the Winteraceae by B rtt
20. Bubbia longifolia A. C. Sm. in Jour. Arnold Arb. 23: 429. 1942.

DISTRIBUTION: Netherlands i Guinea, known only from Brass 13868 (A, TYPE),
from the Idenburg River, alt.

. Bubbia sylvestris A. C. _in Jour. Arnold Arb. 23: 430.

gchar Nethe she ae Guinea, Morobe District, alt. 01800 m., known
from Clemens 4122 (A), 4463 (A), 41142 (A, TYPE), and sil a 5008 (A) and
41800 (A) (alt. 750-1350 m.).

22. Bubbia umbellata (Ridley) ae in eae nes 72:41. 1934.

Drimys umbellata Ridley in Trans. Linn. . Bot. 9: 11. 1916.

DistriBuTION: Netherlands New Guinea, gees only from the type collection,
made by Kloss in the Otakwa River region, south of Mt. Carstensz, alt. about 1200 m
23. Bubbia idenburgensis A. C. Sm. in Jour. Arnold Arb. 23: 432. 1942.

DISTRIBUTION: Netherlands New Guinea, ee River region, alt. 900-1250 m.,
known from Brass 13028 (A, TYPE) and 13313 (A).

24. Bubbia glauca A. C. Sm. in Jour, Arnold Arb, 23: 433. 1942.

DisTRIBUTION: British New Guinea, known ce from Brass 7191 (A, TYPE), from

the upper Fly River region, Western Division, alt. 100

25. Bubbia Archboldiana A. C. Sm. in Jour. Arnold Arb, 23: 433. 1942. Fic. 4,

j-m.
DistrIBUTION: Netherlands New Guinea, known only from Brass 12712 (A, TYPE),
from the Idenburg River region, alt. 2100 m
26. Bubbia megacarpa A. C. Sm. in Jour. Arnold Arb. 23: 434. 1942.
DistriBuTION: Netherlands New Guinea, ert only from Brass 10249 (A, TYPE),
from ie vicinity of Lake Habbema, alt. 280
27. Bubbia argentea A. C. Sm. in Jour. Arnold Arb. 23: 436, 1942.
DistrRIBuTION: British New Guinea, known see eae Brass 4740 (A, NY, TYPE),
from the Wharton Range, Central Division, alt.
28. Bubbia calothyrsa (Diels) A. C. Sm. in Jour. ere Arb. 23: 427. 1942.
sting calothyrsa Diels in Bot. Jahrb. 54: 244.

RIBUTION: Northeastern New Guinea, ae from Ledermann 11028 (type
coll, ear 11166, from the Sepik ae alt. 1300-1350 m.; probably also Ledermann
12978, from the same region, alt. 14
29. Bubbia sororia (Diels) A. C. Sm. ee i Arb. 23: 427. 1942.

Drimys sororia Diels in Bot. Jahrb; 54: 245.

DistripuTion: Northeastern New Guinea, ed from oer 11661, 11898
(type coll.), and 12141, from the Sepik region, alt. 1900-2070
30. Bubbia calophylla A. C. Sm. in Jour. Arnold Arb. 23: 436. 1942.

DistRIBUTION: North New Guinea, known only from Clemens 5061 (A,
TYPE), from the Morobe ae alt. about 1800 m

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 151

3. BELLIOLUM
sect are v. Tiegh. in Jour. de Bot. 14: 278, 330. 1900; Pilger in E. & P. Nat. Pf.

Nachtr. 2: 109. 1906; Hutchinson in Kew Bull. 1921: 190. 1921.

Sune Sect. Saidounay: Baill. in Adansonia 8: 200. 1867, Hist. Pl. 1: 159, 160.

ay ae Sect, Monocladiscum v. Tiegh. in Jour. de Bot. 14: 331. 1900; Pilger
in E. & P. Nat. Pfl. Nachtr. 2: 109. 1906.

Belliolum Sect. Dicladiscum v. oe in Jour. de Bot. 14: 331. 1900; Pilger in E.
& P. Nat. Pfl. Nachtr. 2: 109. 19

Belliolum is thus far known to om represented by eight species, four of
which occur in New Caledonia and the remainder in the Solomon Islands.
However, several of these species are known in fruiting condition only, and,
since the distinction between Bubbia and Belliolum depends primarily upon
staminal characters, there is reason to question the generic disposition of
these species. The geographic distribution of Belliolum forms a curious
contrast to that of Bubbia; both genera are known from New Caledonia,
but the first is lacking from New Guinea and the second is lacking from
the Solomons. To be sure, future collections may modify this picture,
and no conclusions should be drawn from our present incomplete knowledge
of this distribution.

In proposing the genus (6: 278), van Tieghem states that it is typified
by two species of Baillon, Drimys crassifolia and D. Pancheri, but in a
subsequent footnote (6: 331) he states that, of the four species he refers
to Belliolum, the flowers of only B. Pancheri are known. It is evident, there-
fore, that van Tieghem’s concept of Belliolum is based primarily upon
B. Pancheri and that this may be designated as the type species.

The first description of a plant belonging to this group is Baillon’s de-
scription of Drimys crassifolia, which is proposed as the type of Drimys
Sect. Sarcodrimys. This fact has no bearing upon the designation of a
lectotype for Belliolum v. Tiegh. Van Tieghem’s two sections are based
upon the degree of branching of the inflorescence, but this is surely merely
a detail of specific value and demonstrates no basic cleavage in the genus.

Burtt (1) has discussed in some detail the reasons for his reduction of
Belliolum to Bubbia, and I have elsewhere (3: 437-438) expressed the
tentative opinion that the two genera are maintainable. The latter arrange-
ment is continued in the present treatment, but the question can be finally
settled only by the examination of more abundant material than is now
available.

NEw CALEDONIAN SPECIES

To the four species, all from New Caledonia, upon which Belliolum was
originally based, no more from that region have been added up to the
present. As I have seen type material of only one of these species, a final
evaluation of them is impossible, and it is even uncertain whether all of
them belong in the genus, since Bubbia and Belliolum are not positively to
be distinguished in the absence of stamens. The New Caledonian species
are listed in the order suggested for them by van Tieghem.

a ee Alege (Baill.) v. oes in cae de Bot. 14: 330. 1900; Pilger in
wtak: . Pfl. Nachtr. 2: 109.

152 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Drimys Pancheri Baill. in Adansonia 10: 336. 1873; ena in ge Mus, Col.
Marseille IT. 9: 95. 1911; Bak. f. in Jour. Linn. Soe. B t. 45: 267. 1921

Bubbia Pancheri Burtt in Hook, Ic. Pl. 34: sub pl. 3315. aces

DistrRIBUTION: New Caledonia; in the original publication Baillon cites 6 specimens
(of cepa hana and Balansa) without indicating a type, but perhaps, because
of the spe epithet, one should designate the Pancher collection (‘inter sylvas, ad
300 metr. a *) as the actual type. Guillaumin and Baker have added other specimens
in their citations, but van Tieghem has definitely removed two of Baillon’s original
specimens from this concept and referred them to Belliolum rivulare and Bubbia
isoneura respective

Belliolum Panchen is the single species referred by van Tieghem to his

a

rescence simply umbellate. Baillon’s description of the stamens demon-
strates beyond doubt that the species falls into Belliolum rather than
Bubbia.
2. ee crassifolium (Baill.) v. Sao in Jour. de Bot. 14: 330. 1900; Pilger
. Nat. Pfl. Nachtr. 2: 109. 19
Drimys onasel li Baill. in Adansonia 8. 199, Bg Hist. Pl. 1: 159. 1867-69;
Guillaumin in Ann. Mus. Col. ee II. 9: 911.

Bubbia crassifolia Burtt in Hook. Ic. Pl. 34: ty pl. ear 1936.

DistripuTion: New Caledonia; a type and only specimen originally cited by
Baillon is an unnumbered collection of Vieillard from Balade. Guillaumin cites nine

Tieghem, some es hoc had cited several of them as representing three other species
of say and Bubbia.
ve seen no medal which I can refer with certainty to Belliolum
Peso Schlechter 15348 (A, GH), which has been referred to this
species by Guillaumin, Burtt, i the present writer (3: 437), does not
agree in all respects with Baillon’s description, but on the other hand it
almost certainly represents none of the other three New Caledonian species
referred to Belliolum in the present treatment.
3. Belliolum Vieillardi vy. Tiegh. in Jour. de Bot. 14: 331. 1900; Pilger in E. & P.
Nachtr. 2: 109, as B. Acetate 1906.
Drimys Vieillardi Baill. ex v. Tiegh. in Jour. de Bot. 14: as synonym. 1900.
DistriBuTION: New Caledonia; van te cites ea 16 5 and 47, from Balade.
The species is said to be characterized by its stout branches and large
leaves, but a comparison of van Tieghem’s brief description with Baillon’s
description of Drimys crassifolia is not too convincing; a comparison of
type material is obviously desirable.
4, ar rivulare v. Tiegh. in Jour. de Bot. 14: 331. 1900; Pilger in E. & P. Nat.
Nachtr, 2: 109. 1906.
Pre rivularis Vieill. ex Parment. in Bull. Sci. Fr. & Belg. 27: 229, 306. pl. 10, f.
men subnudum, 1896; Vieill. ex v. Tiegh. in Jour. de Bot, 14: 331, as

,

900
Bubbia rivulavis Burtt in Hook. Ic. Pl. 34: sub pl. 3315. 1936.

BUTION: New Caledonia; known with certainty only from the type collec-
tion, Vieillard 2278 (F, GH), from Wagape. Parmentier has based his concept in par
on a collection by Pancher (1. c. 230), which may not be conspecific with the Vieillard
specimen

The type collection bears young fruiting carpels, which are fairly numer-

1943 | SMITH, OLD WORLD SPECIES OF WINTERACEAE 153

ous (5—8 per flower) and have a short subapical stigmatic ridge and ventral
placentas. On the basis of the latter character, the species may fall into
either Belliolum or Bubbia, and final determination must await the collec-
tion of flowering specimens. The small leaves, ascending secondaries, and
large freely branching inflorescences characterize the species, which is quite
distinct.

his species and the two preceding were placed by van Tieghem in his
Section Dicladiscum, with the primary rays of the inflorescence branched
twice. Actually, the inflorescence-rays of B. rivulare are 2, 3, or 4 times
branched

SoLtomMon ISLANDS SPECIES

The four species of Belliolum from the Solomon Islands have all been
recently described. As two of them are known in fruiting condition only,
a dependable key cannot at present be proposed. However, characters of
foliage and fruit amply differentiate the species, and the flowers of the first
two species listed below indicate that good floral characters may also be ex-
pected in the others. For discussions of specific characters, the reader is
referred to my earlier treatment (3). The species of this region are all
characterized by having carpels with a short apical stigmatic ridge and
horizontal placentas situated near the middle; that is, the placentas do
not correspond in position to the external stigmatic surface. On the basis
of present evidence, I judge that the New Caledonian species have a more
primitive type of carpel, with the stigmatic ridge obliquely apical and the
placentas more nearly corresponding to this position. Future study may
indicate these characters to be a basis for sectional differentiation.

5. Belliolum haplopus (Burtt) A. C. Sm. in Jour. aon Arb. 23: 438. 1942.

Bubbia haplopus Burtt in Hook. Ic. Pl. 34: pl. 3315. 1936.

.DistripuTion: Solomon Islands; the type eee is Waterhouse 90 (F, NY),
from Bougainville, while other collections are Kajewski 1658, 1994, and 2007 ot A)
from Bougainville, ae Brass 2959 (A), from Ulawa. The species occurs in rain-forest
at altitudes up to m
6. Belliolum Burttianum A. C. Sm. in Jour. Arnold Arb. 23: 439. 1942. Fic. 5,

a-e.

DistRIBUTION: Solomon Islands, known only from the type, Kajewski 1680 (A),
from Bougainville, alt. 950 m
7. Belliolum gracile A. C. Sm. in Jour. Arnold Arb, 23: 439. 1942.

DistripuTION: Solomon Islands; the type is Brass 2898 (A), from San Cristoval,
other collections are Brass 3063 and 3063A (both A), from San Cristoval, and Kajewskt
2630 (A), from Guadalcanal. The species grows in forest at 900-1700
8. Belliolum Kajewskii A. €. Sm. in Jour. Arnold Arb. 23: 440. 1942.

DistrisuTION: Solomon Islands, known only from Kajewski 2099 (A, TYPE), from
Bougainville, and Kajewski 2574 (A), from Guadalcanal, at altitudes of 1200-1500 m.

4. PSEUDOWINTERA
Pseudowintera Dandy in Jour. Bot. bie fae 1933.
Drimys J. R. & G. Forst. Char. Gen. 83, pro gard excl. D. Winteri. 1776; sensu
Cheesem. Man. N. Zeal. FI. 29. 1906, a 2. 455. 1925.

piers sensu Forst. f. Fl. Ins. Austr. bial 42. en v. Tiegh. in Jour.

277, 290. 1900; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 108. 1906; bald aoa
in Rew Bull. 1921: 190. 1921; non Murray (1784).

154 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Drimys Sect. Drimys DC. Reg. Veg. Syst. Nat. 1: 442. 1817.

Drimys Sect. Eudrimys DC. Prodr. 1: 78. 1824; Baill. Hist. Pl. 1: 158, 160, 1867-69.

Wintera Sect. Euwintera v. Tiegh. in Jour. de Bot. 14: 291. 1900; Pilger in E. & P.

Nat. Pfl. Nachtr. 2: 108. 1906.
Wintera Sect. Pleurowintera v. Tiegh. in Jour. de Bot. 14: 291. 1900; Pilger in F.
P. Nat. Pfl. Nachtr. 2: 108. 1906.

The generic status of the New Zealand species of the Winteraceae has

been subject to confusing vicissitudes. Drimys J. R. & G. Forst. was based

Fic. 5. a-e. Belliolum Burttianum, drawn from the type: a. flowering branchlet,
x &; 5. flower, x 4%; c. stamens, extrorse and introrse views, x 5; d. carpel, x 3;
e. carpel, longitudinal section, x 3. f-j. Pseudowintera axillaris var. typica, drawn
from Cheeseman (Coromandel): f. flowering branchlet, x 4; g. flower, x 14; A.
stamens, extrorse and introrse views, X 5; 7. carpel, X 5; 7. carpel, longitudinal section,
x 5. k, l. Pseudowintera axillaris var. colorata, drawn from Cockayne 183: k. fruit,

with two mature carpels, x 14; l. seed, x

in part upon the New Zealand D. axillaris, but van Tieghem, who first
definitely broke up the inclusive generic concept, took the Magellanic
D. Winteri as the genotype, leaving the New Zealand D. axillaris without
a generic name. The fact that De Candolle had earlier selected D. axillaris
as representing Drimys Sect. Drimys (or Sect. Eudrimys) does not affect
the typification of the genus. In 1786, the younger Forster, following
Murray, apparently abandoned the generic name Drimys and took up the
later Wintera Murtr., applying it to the New Zealand species. In so doing,
Forster did not propose Wintera as a new genus. However, van Tieghem

1943 | SMITH, OLD WORLD SPECIES OF WINTERACEAE 155

applied to the generic concept based on Drimys axillaris the name “Wintera
Forster, non Murray,” remarking (6: 277) that: “Le nom de Wintera
Murray n’existe pas.” This curious conclusion does not alter the fact
that Wintera is a direct synonym of Drimys and is applicable only to
the concept based on D. Winteri. Dandy (2) has discussed the problem
and has quite properly proposed the new generic name Pseudowzntera for
the New Zealand species, selecting Drimys axillaris as the type species.

Pseudowintera is not closely related to Drimys, having a type of calyx,
stamen, and carpel much more suggestive of the genus Bubbia, which is
certainly its closest ally. On floral characters it is difficult satisfactorily to
separate these two genera, but the inflorescence of Bubdia is always termi-
nal or pseudoterminal, the primary rays of the inflorescence (or the flowers,
when these are single) being arranged around the growing point of the
branchlets. When the growing point protrudes through the inflorescence,
this may persist for a short while in a pseudolateral position, as in Drimys,
but at its inception the inflorescence is essentially terminal. In Pseudo-
wintera, on the other hand, the inflorescence is axillary at its inception,
and flowers arise from a lateral position often on branchlets of several
years’ growth. Furthermore, the flowers are comparatively reduced in
size and the inflorescence is greatly compacted, while the small leaves give
a distinct facies to the New Zealand species. The wood-ray is of a different
aspect, as will be discussed in a future consideration of intergeneric
relationships.

Van Tieghem’s two sections are based upon the supposed position of the
inflorescence, which he states to be both axillary and terminal in Sect.
Euwintera (as represented by Wintera terminalis v. Tiegh.) and only
axillary in Sect. Pleurowintera (as represented by three other species).
Dandy points out that W. terminalis is merely an inconsequential form of
Pseudowintera axillaris ; van Tieghem’s sections have no real foundation.

Current opinion among students of this group recognizes three species,
but I am unable to distinguish Pseudowintera colorata from P. axillaris as
a species, and therefore I find the genus to consist of only two species, one
of which has two varieties.

KEY TO THE SPECIES

Erect bushy shrub or small tree 2-10 m. high, me Sie cinereous or purplish or
ownish; leaf-blades usually 3-11 cm. long; flowers in fascicles of 2-10, rarely
solitary ; einle 4-6.5 mm. long; stamens ee ay as few as 6; ie usually
2 or 3 (rarely 1, 4, or 3) UES irate the Mog Mure etree a eens ava oo cserranee oie age ae xillaris.
Compact shrub to 2 m. high, the eta reddish; leaves crowded, the ae closely
appressed to branchlets, the blades 2-3 cm. long; flowers often solitary, sometimes
paired; petals about 3 mm. long; stamens 5 or 6; carpel solitary ..2. P. Traversiz.

1. Pseudowintera axillaris (J. R. & G. Forst.) Dandy in Jour. Bot. 71: 121. 1933.

Glabrous shrub or tree, up to 10 m. high, the branchlets subterete, rugu-
lose, cinereous or purplish or brownish, slender, 1-3 mm. in diameter toward
apices; petioles rugulose, shallowly canaliculate, 5- 10 mm. long, 0.8-1.5
mm. in diameter; leaf-blades subcoriaceous, rarely subpapyraceous, oliva-
ceous or dark brown above when dried, usually paler beneath and copiously
white- or brownish-punctate, usually glaucous beneath in var. colorata and

156 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

often coated with a smooth thin layer of wax, obovoid-elliptic or elliptic,
(1.5—)3-11(-12.5) cm. long, (1—)1.5—5 cm. broad, acute at base and de-
current on the petiole, obtuse at apex, essentially plane or narrowly recurved
at margin, the costa plane or slightly grooved above, prominent beneath,
the secondary nerves 4—12 per side, erecto-patent, copiously anastomosing,
prominulous or nearly plane above, prominulous beneath, the veinlets form-
ing a fine reticulum, prominulous or obscure on both surfaces: inflorescences
axillary or arising from defoliate branchlets, the flowers fasciculate in
clusters of 2-10, rarely solitary, usually borne on pulvinate glomerules, sub-
tended by minute deltoid subcoriaceous bracts; pedicels slender, 5—15(—21 es
mm. long; calyx papyraceous, essentially eglandular, rotate, 2, 5-3.5 m

in diameter, entire or irregularly crenate or shallowly 2- lobed, the lobes
broadly ovate, up to 1.5 mm. long and 2.5 mm. broad, rounded or obtuse at
apex, entire or slightly crenulate or shallowly 3-dentate: petals 5 or 6, sub-
membranaceous or thin-carnose, copiously opaque-glandular, oblong or
obovate-oblong, 4—6.5 mm. long, 2—3.5 mm. broad (inner ones sometimes
reduced to 1-2 mm. in width), obtuse or rounded at apex; soon
(6-)10-18, often 2-seriate, oblong-obovate, flattened, carnose, 1.2—2 m

long, the filaments narrowed at base, broadened to 0.5-1 mm. distally.
yellow-glandular distally, the pollen-sacs oblique on the distal margin,
ellipsoid, 0.3—0.5 mm. long, dehiscing by lateral-apical clefts, contiguous
or slightly separated but not exceeded by the truncate apex of filament;
carpels usually 2 or 3 (rarely 1, 4, or 5), obovoid, 1.3-2 mm. long at
anthesis, rounded at apex, the stigmatic ridge short, linear-oblong, 0.2—0.5
mm. long, strictly apical or obliquely subapical, the ovary-wall densely
glandular, the ovules 8-10, pendulous from short subapical or obliquely

the stigmatic ridge essentially apical, inconspicuous, the pericarp subcar-
nose, densely glandular, rugulose without, the seeds (2—)3—6 at maturity,
obovoid, 3-4 mm. long, 1.5—2.5 mm. broad, obtuse at base, rounded at apex.

In view of the difficulty one has in separating herbarium specimens of
Pseudowintera axillaris and P. colorata, an examination of works on the
New Zealand flora in which this problem is considered is of interest. Raoul,
in 1846, described the new species Drimys colorata without comparing it
with D. axillaris. Hooker, in 1852, merely remarks: “I cannot distinguish
the D. colorata of Raoul from Forster’s plant.” Kirk, in 1889, reduced
Drimys colorata to a variety of D. axillaris, pointing out that the former
is essentially a southern plant in New Zealand and the latter essentially
northern, although the two overlap in the Wellington region; he further
remarks that “the characters stated above [| var. colorata| pass into those
of the typical form by almost imperceptible gradations.” Cheeseman, in
1906, retains both species but remarks: “I have considerable hesitation in
re-establishing this [ Drimys colorata| as a species. It is certainly very
close to the preceding [ D. axillaris |, and in the dried state it is often diffi-
cult to separate the two. But in the field it can always be readily dis-
tinguished, and all my correspondents regard it as distinct. The two
species grow intermixed in many localities in the Wellington and Nelson
Districts.” Cockayne, in 1928, retains both species, pointing out the dis-
tributional differences and stating that Drimys axillaris is much taller than

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 157

D. colorata and has larger, glossy, dark green leaves, rather than yellowish
green leaves, which are blotched red or purple and are glaucous beneath.

The prevalent modern opinion seems to be that the two species are dis-
tinguishable and are good entities, although it is admitted that hybridiza-
tion occurs in the region where the two ranges overlap. Naturally, New
Zealand botanists who have observed the genus in nature are best qualified
to judge how distinct the two plants actually are, and for that reason I hesi-
tate to go back to the earlier opinions and combine them. However, a
careful examination of the cited specimens shows that there are absolutely
no floral distinctions, with the possible exception of a slight and undepend-
able tendency toward more entire calyces in Pseudowintera colorata. As
to the differences in habit and foliage pointed out by various students,
these are scarcely noticeable in herbarium material, although the extreme
forms are of course easily distinguished. For instance, the leaf-blades of
such specimens as Colenso and Cockayne 183 are grayish white beneath,
and one would have no hesitation in referring them to Raoul’s species. But
the leaf-blades of Cockayne 3470 and the Setchell collection are only
slightly paler beneath and sometimes nearly concolorous; yet these also
doubtless represent Raoul’s species. The best distinction one can make
between the two groups, I believe, is based on tendencies toward a smaller
habit, smaller leaves, and paler lower leaf-surfaces in Pseudowintera
colorata. If these characters could be linked with any pertaining to the
inflorescence, however inconspicuous, one would feel justified in retaining
both pel but my present opinion is that only one specific entity can
be adm

In view rol the differences pertaining to habit and leaf color, however, I
cannot altogether ignore the entity based on Drimys coloraca, especially
as this has a more or less distinct geographic range and is apparently readily
recognized in a living condition. Therefore I propose varietal combinations
for the two entities, one based on the Forsters’ type and the other on Raoul’s
species. The description which I have given above is comprehensive,
while the few points which differentiate the varieties are ig out below.
la. Pseudowintera axillaris var. typica nom. Fic. 5,

Drimys axillaris J. R. & G. Forst. Char, Gen. vi . 42, i te ‘Forst. f. in Nova

Acta Reg. Soc. Sci. Ups. 3: 182. 1780; L. f. Suppl. 270. 1781; Lam. Encyel.

Hook. Ic. Pl. 6: pl. 576. 1843; Raoul, Choix de Pl. Nouv. Za, 47. 1846; Hook.
f. Fl. Nov. Zel. 1:12. 1852; ‘Miers in Ann. Mag. Nat. Hist. III. 2: 43. 1858,
Contrib. Bot. L: 132. 1861; Hook. f. Handb. N. Zeal. Fl. 10. 1864; Baill. Hist.
Pl. 1: 158. f. 203, 204, 1867-69; Kirk, Forest Fl. N. Zeal. pl. 1. 1889; Featon, Art
Alb. N. -Zealo:Fl.. 12. pl. -5, f. 3: 18895 Kirk: ae Fl. N. Zeal. 22. 1899;
Cheesem. Man. N. Zeal. Fl. 29. 1906, ed. 2. 456

aye axillaris Forst. f. Fl. Ins. Austr. Prodr. i is: Willd. Sp. Pl. 2: 1240.

; Pers. Syn. Pl. 2: 84. 1806; v. Tiegh. in Jour. de Bot. 14: 290. 1900; Pilger

i. . & P. Nat. Pfl. Nachtr. 2: 108. 1906; Cockayne in Bull. N. Zeal. State For.
Serv. 4(2): 43. 1928, in Engl. & Drude, Veg. der Erde ed. 2. 14: 125. 1928.

Wintera eis V. eis in Jour. de Bot. 14: 291. 1900; Pilger in E. & P. Nat.
Pfl. Nachtr. 2: 1906.

Pseudowintera rnin Dandy in Jour. Bot. 71: 121. 1933.

158 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XxIV

4-10 m. high; leaf-blades (3—)4—11(-12.5) cm. long, often paler
beneath but scarcely glaucous, the waxy layer inconspicuous, the secondary
nerves usually 6-12, the veinlets usually prominulous but often obscure on
both surfaces; calyx crenulate or bilobed, rarely entire.

DistrisuTion: North Island (from Ahipara and Bay of Islands [lat. about 35°]
of

parently often: occurring in mixed beech forest, or occasionally in pure beech forest
(ex Cockayne). The type was collected by the Forsters, but no definite locality was
noted.

EW ZEALAND: North Island: Coromandel, Cheeseman (US); Mt.
Egmont Ranges, nha (A); Ohakune, Oliver (UC) ; emo Travers (GH);
Hunua, Kirk (A Tee te 347 (GH, US); withou definite lo-
cality: Eli aie (F

NATIVE NAME: ee ies

Wintera terminalis v. Tiegh. is based on a specimen collected by Sinclair,
without definite locality, which is said to differ from the other New Zealand
species in having its inflorescences both axillary and terminal; upon this
species van Tieghem based his Section Euwintera. Although the original
description is quite inadequate, I cite this as a synonym of the Forsters’
concept on the authority of Dandy.

1b. Pseudowintera axillaris var, colorata (Raoul) comb. Fic. 5, k, 1.

Drimys colorata Raoul in Ann, Sci. Nat. III, 2: 121. 1844, "“Choix de PI, pean -Zél.

24. pl. 23. 1846; Parment. in oe Sci. Fr. & Belg. 27: 227, 303. 1896; Cheesem.

Man. N. Zeal. Fl. 30. 1906, ed. 925,

id Ng Lene var. colorata Ei, tian Fl. N. Zeal. pl. 2. 1889, Students’ FI.

N. Zeal. 2

Wintera ae Raoul ex v. Tiegh. in Jour. de Bot. 14: 290. 1900; Pilger in E. &

P. Nat. Pfl. Nachtr. 2: 108, 1906; Cockayne in Bull. N. re State For. Serv.

4(2): 43, 1928, in Engl. & Drude, Veg. der Erde ed. 2. 14: 1928

Wintera monogyna v. tr in Tour. de Bot. 14: 291. 1900; Se in E. & P. Nat.

Pfl. Nachtr. 2:

Pseudowintera phe a in Jour. Bot. 71: 121. 1933.

ect bushy shrub or small tree, usually not exceeding 4 m. in height;

_ blades (1.5—)3—9 cm. long, sometimes blotched with red or purple,

ually glaucous beneath, often conspicuously so, the waxy layer often

ee the secondary nerves usually 4—6, the veinlets usually obscure
on both surfaces; calyx often entire, rarely crenulate or bilobed.

ISTRIBUTION: North, South, and Stewart Islands, said to occur from Patetere
Plateau and Rotorua (lat. eee 38°) southward, at altitudes of sea-level (toward
the south and in Stewart Island, where it is reported as common) to about 1000 m.
usually subalpine or montane in North Island. The type is the Raoul collection ited
below.

New ZEALAND: Poketitiri, Meebold ete are > orth Island: Kirk
(GH); Wellington, Travers (GH); South Is nd: Otira Valley, Cockayne
183 (GH), 3470 (NY); Akaroa, Banks fairer Sad (TYPE COLL., GH, US), Kirk
(A, F, M); Mt. Sinclair, Banks Peninsula, Meebold i. (NY); Hunter Hills, South
Canterbury, Anderson 213 (A, M, NY, UC, US); ke Manipouri, Setchell (UC);
without definite locality: Colne in Védel (US), Oliver (UC).

NATIVE NAMES: Pepper-tree, craoutink (ex Raoul). In discussing Drimys axillaris,
Featon uses the names pepper-tree and Maori painkiller, which are more likely to refer
to var. colorata

1943 ] SMITH, OLD WORLD SPECIES OF WINTERACEAE 159

Dandy has referred Wintera monogyna v. Tiegh. to the synonymy of
Pseudowintera Traversii, but I am more inclined to believe that it repre-
sents the present variety. It is said to differ from Raoul’s species in its
smaller and more rounded leaves and especially in its single carpel. The
latter character has been noted in some specimens of var. colorata, and the
leaf-shape is too variable to be of much consequence. Van Tieghem does
not mention the stamens as being exceptionally few in number, a point
which he probably would have emphasized if his plant represented P.
Traversii. Furthermore the type of Wintera monogyna was collected by
Hombron at Akaroa (the type locality of Raoul’s species), whereas P.
Traversii has not been authentically reported from that part of the South
Island.

2. Pseudowintera Traversii (Buchanan) Dandy in Jour. Bot. 71: 122.
Hymenanthera Traversii Buchanan in Trans. New Zeal. Inst. 15: 339. 4 1-1b.
1883.

Drimys Traversii Kirk in Trans. New Zeal. Inst. 30: 379. 1898; Cheesem. Man. N.
Zeal. Fl. 30. 1906, ed. 2. 456. 1925, = N. Zeal. Fl. 1: pl. 8

Wintera Traversii Cockayne in Bull. N. Zeal. State For. Sere, A(2): 43. 1928, in

DistRIBUTION: South Island, apparently limited to the northwestern portion, from
near the northern tip of the island southward to the Buller River region, probably not
occurring much farther south than latitude 42°; altitude up to about 900 m. The
type was collected by Travers in the Collingwood district.

Having seen no specimens referable to this species, I am acquainted with
it only through the above references. The descriptions of Kirk and
Cheeseman (in 1906) are quite adequate, and an excellent plate was pub-
lished by Cheeseman in 1914. The species is characterized by its com-
pact habit, being a shrub from 0.7 to 2 m. high, often straggling or semi-
prostrate, with reddish branches and branchlets. The leaves are crowded
and more or less imbricate, with petioles closely appressed to the branchlets
and with coriaceous blades which are 2—3 cm. long, 0.8-1.3 cm. broad,
and glaucous beneath. The pedicels are often solitary, sometimes paired ;
the calyx is entire, the petals 5 or 6 and about 3 mm. long, the stamens
5 or 6 and uniseriate, and the carpel solitary.

5. EXOSPERMUM
Exospermum vy. Tiegh. in Jour. de Bot. 14: 279, 333, 1900; Pilger in & P. Nat.
Pfl. Nachtr. 2: 109. 1906; Hutchinson in Kew Bull. 1921: 190. 19

Exospermum is known only from the two New Caledonian ae upon
which it was originally founded; the type is E. stipitatum. Van Tieghem’s
classification (6: 354) has the genus most closely related to Zygogynum,
with which it has in common “carpelles unis, 4 placentation médiane.”” In
his discussion, however, van Tieghem points out that the carpels of
Exospermum are only superficially united and not firmly concrescent as
in Zygogynum, while the placentation of E. Lecarti (6: 339) is “en méme
temps marginale, latérale et médiane pour chaque carpelle .. .”’ In effect,
the genus is more suggestive of Bubbia than of Zygogynum and might con-
ceivably be combined with the former, although for the present I feel
justified in retaining it as outlined by van Tieghem.

160 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

1, a stipitatum (Baill.) v. ig ex Pilger in . ' P. Nat. Pfl. Nachtr.
. 1906; Hutchinson in Kew B 1921: 190. fig. Fic. 6, a-e.
Peart stipitatum say in »Adanson 10: 334. 1873; vids in Ann. Mus.
Col. Marseille II. 9:
Drimys neo- Sree te ex Baill. in Adansonia 10: 335, as synonym. 1873.
Drimys Lenormandii Vieill. ex Parment. in Bull. Sci. Fr. & Belg. 27: 231, 308. pl.
0, f. 35, pl. 11, f. 42, 43, nomen subnudum. 1 1896; Vieill. ex v. Tiegh. in Jour.
Bot. 14: 333, as synonym. 1 ; Vieill. ex Guillaumin in Ann, Mus. Col.
Marseille II. 9: 95, nomen. 1911.
Drimys austro-caledonicus Vieill. (pro parte) ex v. Tiegh. in Jour. de Bot. 14: 333,
nonym. 1900

DistRipuTION: New Caledonia; the type was collected by Vietllard at Wagap (ex
Baillon), or “dans les montagnes de Ti-Ouaka prés de Wagape” (ex van Tieghem) ;
a single leaf of this is available (A). Van Tieghem also cites Vieillard 2281 (GH), the
source of the name Drimys Lenormandii; from a comparison of the foliage, one may
suspect that this is the same collection as the type, which was unnumbered.

Van Tieghem’s description of this species is very detailed and accurate,
although my observations of the gynaecium lead me to believe that he
attaches too much importance to the regularity of the carpel-arrangement.
In one flower available to me there are 6 carpels in a single whorl, in
second flower there are 7 carpels, of which one has apparently been pushed
into the center by pressure. Van Tieghem reports the carpels as occurring
in two whorls of 3-5 carpels each. No petals are available on our material,
but van Tieghem describes these as occurring in three tetramerous whorls,
implying a degree of regularity which is not found in the related genera.
In this species the placentation is said to be limited to the external face
of the carpels, but actually the ovules also occupy a portion of the distal
lateral faces and the apical-ventral angle.

In his treatment of the plant, van Tieghem neglected to make the actual
combination Exospermum stipitatum, apparently through an oversight,
for he repeatedly mentions ‘“Exosperme stipité’ and gives Zygogynum
stipitatum as a synonym. It appears that Pilger was the first to use the
correct Latin binomial.

2. Exospermum Leearti v. Racioy in Jour. de Bot. 14: 334. 1900; Pilger in E. & P.
Nat. Pfl. Nachtr. 2: 10
TRIBUTION: New Ca ae. lent ed only from the type collection, Lécart 141,
Bese detinite locality; probably a suman leaf (A) under this name, but te
Lécart 144, is actually from the same collection.

Van Tieghem mentions that this is ees in foliage to the preceding,
but differs from it by its usually solitary and short-pedicellate flowers, fewer
carpels, and more extensive placental surface. The ovules are said to be
situated at the internal angle of the carpel, on the lateral faces, and also
on the external face. This is interpreted, by van Tieghem, as a transitional
stage between the carpel of Bubbia and Belliolum (which he supposes
always to have the ovulation along the ventral angle) and that of Exo-
Spermum stipitatum, in which the ovules are supposedly situated only on
the external face of the carpel. The carpels of E. Lecarti are only weakly
united, and in this respect the species is similar to some species is Bubbia,
i.e. B. montana (Lauterb.) A. C. Sm. and B. pachyantha A. C. Sm. It

1943 | SMITH, OLD WORLD SPECIES OF WINTERACEAE 161

appears that the ovulation in Bubbia and Belliolum is more diverse than
supposed by van Tieghem, and the placental surface is not always restricted
to the ventral angle of the carpel in those genera. Consequently, the
primary characters upon which Exospermum is founded do not sharply
distinguish the genus from Bubbia and Belliolum, although the stamens
amply separate it from the latter. However, I do not propose to unite
Exospermum with Bubbia; although it doubtless represents an extr
trend from a Bubbia-like ancestor, it seems to be sufficiently characterized
by its ovulation to merit generic status. However, I believe that its closest
alliance is with Bubdia rather than with Zygogynum.

a)
3
oO

1G. 6. a-e. Exospermum stipitatum, drawn from Vieillard 2281: a. branchlet with
two flowers past anthesis, x 4; b. flower, past anthesis, the petals and stamens having
fallen, X 2; c. stamens, extrorse and introrse views, x 5; d. carpel, x 2; e. carpel,

Zygogynum Vieillardi, drawn from Franc 1740: f. branchlet with solitary terminal
young flower, x 4; g. flower with calyx and petals removed, xX 12; h. stamens,
extrorse and introrse views, X 5; 7. longitudinal section of young flower, showing
torus with stamens and fused carpels, < 2; j. enlarged section of 7, showing two stig-
mas, one locule (with one row of ovules removed), and the connecting canal, x 4.

6. ZYGOGYNUM
Zygogynum Baill. in Adansonia 7: 298. 1867, Hist. Pl. 1: 160, 190. 1867-69; Prantl
in E. . Nat. Pfl. III. 2: 19. 1891; v. Tiegh. in Jour. de Bot. 14: 279, 340.
1900; Pilger in E. & P. Nat. Pfl. Nachtr. 2: 109. 1906; Hutchinson in Kew Bull.
1921: 191. 1921.

162 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

Zygogynum Sect. Monanthum v. — in Jour. de Bot. 14: 341. 1900; Pilger in
E. & P. Nat. Pfl. Nachtr. 2: 109.

nd Nh ie Sect Pleianthum v Tiegh. in Jour. de Bot. 14: 341. 1900; Pilger in
E. & P. . Pfl. Nachtr. 2: 109.

Sena ie been generally eee by workers on the New Cale-
donian flora as a good genus. It is indeed the most distinct genus of the
family in many respects, with a unique type of carpellary specialization.
No species have been added to Zygogynum since it was amplified by van
Tieghem to include six species. In Zygogynum the carpels are firmly
concrescent and the ovules are limited to the external face of the locule;
van Tieghem interprets these facts as a continuation of the trend begun
in Exospermum, which may indeed be true, although the remarkable
syncarpy of Zygogynum is only faintly suggested by the loosely coherent
carpels of Exospermum. ‘The two genera have a common aspect in their
reduced inflorescences, the flowers being either solitary and terminal or
few in a terminal umbel (perhaps more properly described as single and
clustered around the growing point).

The two sections proposed by van Tieghem are based on solitary versus
aggregated flowers, a division which seems arbitrary and not very useful,
since the flowers in the second group are often reduced to one soon after
anthesis. The principal basis of speciation may eventually be found to
occur in the gynaecium, as suggested by van Tieghem (6: 347-348), the
number and disposition of the carpels varying from species to species.

Baillon’s paper discussing his new species does not include a formal
generic description, nor does the generic name appear as an independent
entity in the text, being first mentioned as such in a brief summary (op. cit.
oi However, the description of Zygogynum Vieillardi may be taken
as a descriptio generico-s pecifica.

The species are here briefly discussed in the order proposed by van
Tieghem.

1. Zygogynum Vieillardi Baill. in Adansonia 7: 298. pl. 4. 1867, Hist. Pl. 1: 161.

Z. Viedlardit. 1906; Guillaumin in Ann. Mus. Col. Marseille II. 9:95, 1911.
Fic. 6

DistrIBUuTION: New Caledonia; the type collection is pes 187, from moun-

tains near Balade. Guillaumin adds citations of Pancher 17 and 283 from Mt. Koghi,

neither of which I have seen, and Balansa 2763 from Mt. Mou , of which a fragment
(A) seems very possibly to represent a species, although it is also suggestive of Z.
Bailloni. The only specimen in American herbaria which I can positively refer to
Z. Vieillardi is Franc 1740 (A, UC), one Prony.
The species is well described and figured by Baillon; it is characterized
by its 10-12 concrescent carpels with elliptic or subcapitate stigmas.
7: saa Bailloni v. Tiegh. in Jour. de Bot. 14: 340. 1900; Pilger in E. & P.
. Pfl. Nachtr. 2: 109, as Z. Baillonii. 1906
ISTRIBUTION: New Caledonia; the type was -gollected by Pancher in 1870, ap-
parently without number or definite locality, and is represented by a leaf (A) which
M

Mou, a sterile specimen which
may be referred to Z. Bailloni with reasonable certainty.

1943] SMITH, OLD WORLD SPECIES OF WINTERACEAE 163

Zygogynum Bailoni is compared by van Tieghem to Z. Viellardi, from
which it differs in having its leaf-blades more coriaceous and glossy, and in
having 4 carpels with sessile linear stigmas rather than 10-12 carpels with
elliptic stigmas. This is the only species of the genus elaborated in detail
by van Tieghem (6: 341-345); it is said to have a very short and stout
pedicel, 8 petals in two whorls, and carpels concrescent except at the apices.
The difference in stigmatic characters between this species and the remain-
ing species of Zygogynum causes van Tieghem (6: 348) to speculate on its
generic status. On the basis of his discussion, one might assume that
Z. Bailloni is the most primitive species of the genus in its retention of a
linear stigmatic ridge.

3: ga a Seay bicolor v. es in Jour. de Bot. 14: 341. 1900; Pilger in E. & P.
. Pfl. Nachtr. 2: 10

oe New ae ee only from the type collection, Lécart 41
(A, leaf), without definite locality.

Zygogynum bicolor is said by van Tieghem to resemble 2. Viedllardi in
its solitary flower on a short stout pedicel, but it is distinguished by having
its leaf-blades very pale beneath and is further characterized by its very
numerous stamens, 147-168 in 7 or 8 whorls (6: 346).

4. oe pomiferum Baill. in apne 10: 334. 1873; v. Tiegh. in Jour
ot. 14: 340. 1900; Pilger in E. & P. Nat. Pfl. Nachtr. 2: ‘109. 1906 ; ene
in Ann. Mus. Col. Marseille a 9: 95. ae :

DistRIBUTION: New Caledonia; the species was originally based by Baillon on two
collections, Balansa 2328 (A, TYPE COLL.), from Kanala (alt. 800 m.), and 2804, from

the type, deferring the second to his Z. Balansae. Guillaumin cites a collection of
Lecard from Bourail as Sune Z. pomiferum.

The original description is ample, the species being distinguishable from
all others except the following by its comparatively broad leaf-blades. The
type collection is in fruit only.
cy mee ynum stoma ae v. Tiegh. in ia de Bot. 14: 340. a bey in E.& P

Nat. Pfl. Nachtr. 2: 109, 1906; Bak. f. in Jour. Linn. Soc . 45: 268. 1921.

DistRIBUTION: New ledonia; the type collection is Balansa 2804, from Mt. Mou,
alt. 700 m. Baker cites Compton 1776, from Mt. Panié, as representing co species.

Zygogynum Balansae is compared by van Tieghem with Z. pomiferum ;
it has similarly large leaves and flowers grouped in a simple sessile umbel,
but it is said to differ in having the leaf-blades less coriaceous, longer, nar-
rower, distally attenuate, and with the lateral nerves more distant and
- iculées.” It is further characterized by its very numerous carpels
and large ee these being up to 4 cm. in diameter, according to van
Tieghem (6: 3
6. ae te He ee v. Tiegh. in Jour. de ae 14: 341, as Z. spatulatum.

900; Pilger in E. & P t. Pfl. Nachtr. 2: 109. 1906.

pent austro- caledonicus aa (pro parte) ex v. ee in Jour. de Bot. 14: 341,

as synonym

DIsTRIBUTION: New release reported only from the original collection, Vieillard
2266 (GH, TyPE COLL.) from Wagape.

164 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

Zygogynum spathulatum resembles Z. pomiferum and Z. Balansae in

having its flowers aggregated, but it is distinguished by its narrow spatulate
leaf-blades, which are rounded at apex and attenuate at base. The carpels
are about 20 in number.

pm WH

an

~a

PRINCIPAL LITERATURE CITED

Burtt, B. L. Bubbia haplopus. Hook. Ic. Pl. 34: pl. 3315 (pp. 1-3). 1936.
Danpy, J. E. The Winteraceae of New Zealand. Jour. Bot. 71: 119-122. 1933.
Smiru, A. C. Studies of Papuasian Plants, V. Jour. Arnold Arb. 23: 417-443. 1942.
The American species of Drimys. Jour. Arnold Arb. 24: 1-33. f. 1-3.
1943.
——————.. La Distribution géographique et ]’Histoire des Winteraceae. Boissiera,
1943 (in ei

; a P. vA Sur = ch lédones du groupe des Homoxylées. Jour. de Bot.

14: 259-297, "4 361.

. ieee, Joyce W. Two new species and one new variety of Drimys Forst., with

notes on the species of Drimys and Bubbia van Tiegh. of south-eastern Australi
and Lord Howe Island. Proc. Linn. Soc. N. S. Wales 62: 78-84. f. 1,
193

. bearer: T. W. Chromosome number and relationship in the Magnoliales. Jour.

d Arb. 14: 376-385. f. 1-4; pl. 80. 1933

ARNOLD ARBORETUM,

Harvarp UNIVERSITY.

1943 | CROIZAT, AMERICAN EUPHORBIACEAE 165

NOVELTIES IN AMERICAN EUPHORBIACEAE
LEON CROIZAT

THE new species, trinomials, and records in this work have accumulated
with few exceptions in the herbarium of the Arnold Arboretum of Harvard
University in the course of routine determinations. They are published
primarily to validate the manuscript names under which many of them
have been cited in letters to various correspondents. Some of the entities
dealt with here belong to critical groups which are worthy of more detailed
consideration, or which should be critically revised. Unless otherwise
stated, all the types are deposited in the herbarium of the Arnold Arboretum.

Ditaxis Vahl ex A. de Jussieu
Te Se eases (Muell.-Arg.) Pax & Hoffm. in Pflanzenr. 57 (IV. 147. vi): 65.
; O'Donnel & Lourt. in Lilloa 8: 67. 1942
Arasthamnia. breviramea Muell.-Arg. in Linnaea 34: 146. 1865, in DC. Prodr.
15(2)3 6.

ParaGuay: Fortin Lopez de Filippis, Rojas 8438; Fortin Tte. Montania, Rojas 8479.

The record is apparently new for Paraguay. The material cited agrees
with a photograph of the Herzog specimen from Bolivia identified as
Ditaxis breviramea by Pax, in Med. Rijks. Herb. 40: 22. 1921, and repro-
duced as plate 1 in the excellent monograph by O’Donnel & Lourteig.

Bernardia Adanson
Bernardia Gentryana sp. nov.

Arbuscula ad 5 m. alta, innovationibus strictis stellato- vel fasciculato-
puberulis. Foliis tenuiter membranaceis ovatis vel ovato-lanceolatis eke
stellato-puberulis vel glabratis margine inaequaliter dentate: serratis, 3.5—
cm. longis, 1.5—2.5 cm. latis, nervis adscendentibus ca. 7-jugis eraciibus,
glandulis cicatricoso- crateriformibus i in basi laminae 2, petiolo ca. 3—5 mm
longo. Floribus ¢ ignotis. Floribus @ subsolitariis brevissime pedicel-
latis, perianthii squamulis 8-10 in serie duplici, ovatis puberulis costulatis
a mm. longis latisque, ovario globuloso depresso ad 2 mm. lato fere
totidem longo pallide luteo-tomentello, stylis 3 quove lunulato apice in
laciniis 3(—6) partito ad 1.5 mm. m. longo: capsula submatura c mm.

magna, semine cordiformi 5 mm. longo, 3.5-4 mm. lato, pallide brunneo-
marmorato

Mexico: Sinaloa: Nuevo Mundo, Gentry 5372 (TYPE in Gray Herb.).

This cannot be Bernardia aspera Pax & Hoffm., B. incana Mort., B.
mexicana Muell.-Arg. (at least as illustrated by Pringle 3700), B. obovata
Johnst., or B. viridis Millsp., as it disagrees with each one of them in its
thinly membranous, ovate to ovate-lanceolate leaves, which are ultimately
larger in size. Bernardia Brandegei Millsp. ex Brandeg., in Proc. Cal. Acad.
Sci. IT. 3: 172. 1891, is a nomen nudum replaced by B. viridis at publica-
tion, as shown by the notes that appear on the photograph of the type of
B. viridis Millsp. preserved in the Gray Herbarium.

166 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Bernardia amazonica sp. nov.
Fruticulus vel frutex ad 1 m. altus, ramulis strictis puberulis. Foliis
lanceolatis vel obovato-lanceolatis Gene chartaceis vel subcoriaceis 5—10

2 in cymulis sessilibus axillaribus, perianthii squamulis in serie duplici
9 vel 10 ovatis margine integro hispidis, 2 mm. longis, 1-2 mm. latis, ovario
vix 1 mm. longo latoque cum stylis confluente, stylis 3 ad | 1 mm. longis
apice divaricato-partitis.

VENEZUELA: Amazonas: Puerto Ayacucho, Williams 13142.

This was distributed as representing B. Jacquiniana Muell.-Arg., which
is a very different species. It suggests B. axillaris (Spreng.) Muell.-Arg.
in habit and foliage, but manifestly differs from that species in the venation,
the texture of the leaf, and in its indumentum. As shown by a Brazilian
specimen collected by Riedel, in our herbarium, B. axillaris has glabrous
leaves with an immersed venation and differently shaped marginal teeth.

Alchornea Swartz
Alchornea cyclophylla sp. nov.

Arbor videtur glabra vel glabrescens. Foliis subcoriaceis brunneis
utrinque glabratis in axillis nervorum barbulatis late rotundato-ovatis,
apice breviter apiculato-acuminatis, basi leviter cordatis, margine obscure
crenato-serratis, ca. 15 cm. longis, 12—13 cm. latis, nervis ca. 4- vel 5-jugis
adscendentibus, primo jugo laminae tertium superum attingente caeteris
penninerviis, glandulis in lamina hic inde sparsis parvis, petiolo 5.5—-8 cm.
longo, stipulis subglandulosis minimis deciduis. Inflorescentiis ¢ ignotis.
Inflorescentiis 2 spicatis puberulis vel glabratis, floribus solitariis vel binis
ee bracteolatis, calycis lobis triangulari-acuminatis 4( 5), vix

nm. longis latisque, ovario ovoideo glabrescente, tenello vix nes i
salevile demum glabro, 2-cocco, stylis 2 integerrimis carnosis epapillosis
ad 8 mm. longis ail basim ipsissimam liberis

Costa Rica: Alajuela: Palmira, Austin Smith 28706.

This differs from A. latifolia Sw. in the styles being cleft to the base and
in its foliage. The Central American species are much in need of a critical
study.

Cleidion Blume
Cleidion denticulatum Standl. in Field Mus. Publ. Bot. 4: 218. 1929.

The holotype, Cooper 12239, is poor. Its crowded malpighiaceous hairs
and the large, easily separable bundles of the columella are strongly rem-
iniscent of Bernardia. The loose seeds on the type-sheet, however, have
the characters of Cleidion. Better material is needed to determine the
correct generic position of this aneen
Cleidion oblongifelium (Standl.) comb. n

Alchornea oblongifolia Standl. in Carnegie Tnst, Wash. Publ. 461: 66, 1935.

The @ flower ending the rachis of the inflorescence has manifestly cleft

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 167

styles, such as characterize Cleidion. The comparatively long and narrow
leaves conform to those of this genus. If generic value is denied to the
nature of the style, this being entire in Alchornea, and cleft in Cleidion, it
becomes impossible to separate the two groups which both tradition and an
aggregate of characters have rightly maintained as distinct genera. This
plant is endemic to British Honduras.

Cleidion prealtum sp. nov.

Arbor ad 30-35 m. alta, innovationibus parcius puberulis citissime glab-
ratis vel glabris, cortice pallide brunneo sat lenticellato. Foliis 7-14 c
longis, 2—5 cm. latis, obovato-oblongis vel ellipticis, apice breviter acumi-
natis, basi plus minusve rotundatis, junioribus membranaceis brunneis vel
olivaceis, adultis (e folio singulo) forsan coriaceis, subtus pallidioribus, in
venis leviter puberulis glabratisve supra glabratis vel glabris secus costam
hic inde pustulato-glandulosis, venis gracilibus ca. 6-jugis, petiolis herbaceis
puberulis 2—4 cm. longis, stipulis setaceis ad 3-4 mm. longis. Inflorescentiis
8 spicatis gracillimis 8-10 cm. longis, floribus in cymulis saepius pauci-
floris secus rhachem dissitis; perianthio subsessili 3- vel 4-lobo valde delicato
puberulo ca. 3.5 mm. lato totidemque longo vel minore, staminibus ca. 50,
antheris lateraliter dehiscentibus, filamentis ca. 2 mm. longis; pistillodio,
glandulis nullis. Inflorescentiis 9 spicis simplicibus ad 3 cm. longis, subtus
bracteis ca. 2 vel 3, linearibus vel subsetaceis, 2 mm. longis, in axillis in-
ear semper flore evoluto carentibus, apice flore terminatis; perianthio

a. 3-4.5 mm. lato, lobis 5 vel 6 interdum 2 plus minusve inter se adnatis,
(ie eu ances ad 2 mm. longis, basi etsi disco nullo tumido-saccatis ;
ovario ovoideo ca. 3 mm. longo et 2 mm. lato, luteo-tomentello, stylis 3 quove
7 intus longe grosseque papilloso, papillis haud processiformibus.

zonas: Municipality of Humayta, near Tres Casas, Krukoff
6458 Q ees 6357 86,6391 6.

Also here probably belong Krukoff 6602, 6649, and 6570, all from the
Municipality of Humayta near Livramento on the Rio Livramento. The
material cited was distributed as representing Alchornea Hilariana Baill.
and A. brachygyne Pax & Hoffm., the misdeterminations being probably
suggested by photographs illustrating plants with similar foliage. The 9
flower and the leaves of C. prealtum are strongly reminiscent of Epiprinus
Griff. from tropical Asia, a genus which stands out as a natural unit of few
species but is otherwise closely allied with Cleidion.

Cleidion Woodsonianum sp. nov.

Arbuscula 4—5-metralis, pilis brevissimis in innovationibus am
glabra. Foliis utrinque acuminatis ellipticis, 7-11 cm. longis, 1.5-4 c
latis, apice breviter acuminato-caudatis, basi truncatis vel vix re.
latis, hic inde in lamina ipsa glandulosis, pallide olivaceis, supra sub lente

0.5-1.5 cm, longo. Inflorescentia ¢ ignota. Tao @ ad 15 cm.
longa, gracillima, floribus 3 vel 4 tantum, bracteis minutissimis lineari-
triangularibus fultis; perianthii lobis (videtur) 5 minimis lineari-subulatis,
petalis glandulisque nullis, ovario puberulo ovoideo ca. 3 mm. longo
totidemque lato, stylis 3 ca. 8 mm. longis quove fere ad basim partito;
fructu capsulari, columella gracili delapsa ad 4-5 mm. longa, epicarpio

168 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

sublevi olivaceo sub lente puberulo, semine ovoideo subquadrangulo ca.
4 mm. magno apice acutato.

PanaMA: Canal Zone: Vicinity of the Salamanca Hydrographic Station, Rio
Pequeni, Woodson, Allen & Seibert 1587.

This suggests certain species of the Old World, such as C. leptostachyum
Pax & Hoffm. from the Fiji Islands. The type-material was originally dis-
tributed with doubt as representing Alchornea costaricensis Pax & Hoffm.

Jatropha Linnaeus
Jatropha hippocastanifolia sp. nov.

Frutex fere metralis. Foliis fere ad basim partitis, 5-7-lobis, margine
duplicato-dentatis vel denticulatis vel dentato-serratis longiuscule ciliato-
glandulosis, pallide olivaceis utrinque pilis albidis simplicibus sat longis
mollisque indutis, 15 cm. longis, 18 cm. latis, juvenilibus multo minoribus,
lobo medio maximo cum lateralibus imbricante 12 cm. lon ngo 6 cm. lato ca.
12-nervio,Mobis caeteris minoribus externis ca. 4—5 cm. longis 2.5 cm. latis
vel minoribus, petiolo herbaceo ad 17 cm. longo aeque ac lamina induto,
hic inde glandulis fasciculatis capitulatisve ornato, basi stipulis glandulosis
dissectis insignito. Inflorescentiis cymosis gracilibus velutinosis ad 15 cm.
longis, bracteis longe piloso-glandulosis vix 1 cm. longis 2 mm. latis cpp
Floribus ¢ albo- -hispidis pubescentibusve, sepalis ca. 2—3 mm. longis 1 m
latis integris vel margine denticulis ae glandulosis vel se err
notatis, petalis atro-purpureis ca. 5 mm. Jongis 3 mm. latis basi abrupte
coarctatis vel angustatis pilosulis, ieee ae 10 in acie duplici filamentis
breviter liberis demum in columnam connatis, submaturis ad 6 mm. longis,
glandulis rotundatis discretis 5 oppositisepalis. Floribus @ ut 4, sepalis

semper neque raro tantum ciliato-glandulosis, ovario globuloso costulato hic

inde pilis albidis simplicibus ornato, levi, ca. 2 mm. longo atoque, disco
irregulariter interrupto subintegrove margine hic inde bilabiato vix 0.75
mm. alto, stylis pro more gracilibus glabris ad 2.5 mm. en stigmatibus
primum partitis dein crure quove iterum 2- vel 3-lobulat

Paracuay: Chaco Paraguayo: Oruro, Rojas 8559,

To discuss this new species with finality, a critical consideration should
be given of several species, including J. gossypiifolia L. My new species
suggests the entity called J. intercedens by Pax but differs from it in three
characters that, considered jointly, have specific significance. These char-
acters are: (1) the style, which has bipartite, usually suberect stigmas in
J. gossypufolia and its allies, has here spreading, 2- or 3-lobulate stigmas;
(2) the foliage; the leaf-lobes of J. hippocastanifolia are much narrowed
at the base and enlarged at the apex, the shape of the leaf being accordingly
unlike that of J. gossypiifolia and its allies; (3) the dark-purplish petals.
Jatropha Paxi

Jatropha Aabellifolia on & Hoffm. in Pflanzenr. 42 (IV. 147): 52. 1910, non

Steud. 1840
w name is needed, as J. flabellifolia Pax & neernat is a later homonym
of J. flabellifolia (Pohl) Steud, , Nomencl. ed. 2, 1: 799. 1840.
Jatropha Hoffmanniae nom
sepa sae aera ‘Pax & Hoffm. in Pflanzenr. 85 (IV. 147): 191. 1924,
on Brandeg. 1920.

1943 | CROIZAT, AMERICAN EUPHORBIACEAE 169

Brandegee’s J. longepedunculata, in Univ. Cal. Publ. Bot. 7: 328. 1920,
is a mere orthographic variant of J. longipedunculata Pax & Hoffm., which,
consequently, is a later homonym.

Micrandra Bentham

Micrandra Benth. (1854) has been proposed for a a by Mans-
feld, in Kew Bull. 434. 1935, against Micrandra Benn. (1844).
Micrandra santanderensis sp. nov.

Arbor in sylvis primaevis, innovationibus glabratis cea es Foliis el-
lipticis apice plus minusve breviter acuminatis basi cuneato-rotundatis vel
rotundatis firme chartaceis, 7-14 cm. longis, 2.5—-6 cm. . latis, brunneis vel
obscure olivaceis, margine integris, totis nucle nervosis, nervis utrinque
ca. 8-jugis anastomosantibus obscuris, primo jugo adscendente caeteris
latius patentibus, in axillis conferte penicillatis, petiolo 2—7 cm. longo, apice
sub lamina glandulis 2 pustulosis ornato, stipulis triangularibus minimis.
. ee racemosis ut videtur polygamis ad 25-35 cm. longis. Floribus

: sepalis 5 puberulis triangulari-rotundatis vix 1 mm. eng petalis 5 in
at brunnescentibus quam sepalis duplo longioribus, staminibus ca. 5.

ribus 2 post anthesim tantum visis, involucris floralibus illis ¢ (ut
videtur) similibus, staminodiis ca. 5 setaceis oa disco minimo
parcissime glan uloso, ovario truncato-ovoideo ca. 3 mm. longo et 2 mm
lato, stylis in ovarium enifuent ib brevissimis ae vix bilobis.

ain MBIA: Santande Vicinity of Puerto Berrio, between Carare and ed

ena Rivers, alt. 100-700 m., , Haught 2189 (TYPE); vicinity of Barranca Berm
Spaeniae Valley, between Sogamoso and Carare hives alt. 100-500 m., Haught er

The two specimens here cited appear to differ at first on account of their
differently colored foliage, but there are no floral or other differences. The
conspicuous axillary tufts of hairs on the leaves of this new species are not
found on M. elata Muell.-Arg., which is endemic to southern Brazil, or on
M. siphonioides Benth., to judge from the photographs of the type-
specimens which I have seen through the kindness of Mr. B. A. Krukoff.

Manihot Miller

Manihot orinocensis sp. nov

Specimen suppeditat eid mancum quam frustulum vix melius totum
glaberrimum cortice tenello haar subnitido elenticellato. Foliis pro more
generis parvis ca. 6 cm. latis, 5—6 cm. longis, 3-lobatis, lobis margine in-
tegris revolutisve, mediano elliptico- -lanceolato ca. 3 cm. longo et 1.5 cm,
lato, lateralibus basi plus minusve anisophyllis caeterum lanceolatis, medio
sat dilatatis, ca. 2.5-3 cm. longis et 1.5 cm. latis, subtus glaucescentibus,
supra ruguloso- impressis, membranaceis, nervis ca. 8—10-jugis patentibus,
petiolo ca. 2.5-3 cm. longo, stipulis ut videtur subnullis vel nullis. Perian-
thio ¢ primum in nflato, in anthesi plena a age ca. 10 cm. longo fere
totidem lato pallido, lobis late lanceolatis ad 3 . longis, disco bene
lobulato, staminibus ca. 5-7 mm. longis, plus minis ‘10. Caetera desunt.

VENEZUELA: Amazonas: Upper Orinoco, Fu uer Ayacucho, “a tree 4 m.
high, growing on granitic rocks around Pt». Ayacuch ae 13132.

The type specimen, hardly better than a scrap, was sent by Mr. H. Pittier
with the suggestion that the species may be new. I deferred its publication,

170 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

wishing to check it against material representing M. saxicola Lanj., the
description of which suggested that it was uncomfortably close to the pre-
sumed new entity. Stahel 107, “cultivated from cuttings in the Agricultural
Experiment Garden in Paramaribo, Surinam,” identified at distribution as
M. saxicola, is certainly not M. orinocensis, which differs in the texture and
lobing of its leaves.

Tetrorchidium Poeppig

Poeppig & Endlicher are said to have published this genus in 1842 (see
Pax & Hoffmann in Engl. & Prantl, Nat. Pflanzenfam. ed. 2. 19C: 184.
1931), which is erroneous. The genus was actually published in 1845 by
Poeppig alone, which is fully established by the title-page of the third
volume of the Nova Genera ac Species Plantarum and associated data.
Tetrorchidium gorgonae sp. nov.

Arbuscula vel frutex 3—4 m. altus, innovationibus puberulis indumento
pabgieeicn adpresso simplici gees glabrescentibus griseis. Foliis 10-22
cm. lon 2.5—6.5 cm. latis, utrinque parce pubescentibus vel glabratis,
supra viridibus, subtus pallide viriditas subnitidis, oblanceolatis sat abrupte
caudato-acuminatis obvie distanterque serrato- dentatis, dentibus subcallosis
utrinque ca. 4 vel 5, nervis ascendentibus vel patulis utrinque ca. 6-jugis,
sub margine anastomosatis, trabeculis conspicuis, petiolo 1-2 cm. longo

apice utrinque glandulis baculiformibus luteis insignito parcius hispido;
stipulis late triangularibus, vix ultra 1 mm. longis totidemque latis. In-
Horescentia in fructu tantum visa gracili ad 5 cm. longa; perianthio sub-

—2.5 mm. lato, pedicello ca. 2 mm. longo, columella acuminata
ad 5 mm. longa, semine lenticulari ca. 5 mm - magno, testa nigra profundius
ates foveolata, in arillo sat tenui roseato immersa; epicarpio immaturo
(ut videtur) le

COLOMBIA: ae rino: Island of Gorgona, Killip & Garcia 33214 (tTyPE in U.S.
Nat. Herb.).

A very distinct species, with comparatively narrow, small leaves, their
margins with sharp teeth. In these characters it may be at once distin.
guished from 7. macrophyllum Muell.-Arg., T. rubrivenium Poepp., and the
species next described.

Tetrorchidium boyacanum sp.

Arbuscula glaberrima, a eae subherbaceis. Foliis utrinque virid-
ibus, 9-12 cm. longis, 4.5—5.5 cm. latis, obovatis vel elliptico-obovatis, apice
breviter apiculatis, basi longiuscule cuneatis, firme chartaceis, nervis
utrinque 6- vel 7-jugis jugo basali valde adscendente caeteris plus minusve
patulis sub margine obscure anastomosatis, laminae margine glandulis pus-
tulosis paucioribus in quarto supero praesertim insignito quapropter folii
apice primo intuitu plus minusve profunde dentato-serrato, petiolo canalicu-
lato glandulis 2 pustulosis anticis ornato 1-1.5 cm. longo. Inflorescentia
§ tantum visa, ad 15 cm. longa, gracili, ramosa, glaberrima; floribus ca.
3—6 in axilla Sah minutae utrinque glandulis luteis notata, calyce sub-
sessili ca. 2.5 mm. lato, lobis 3 rotundato-ovatis subcucullatis ca. 1—1.5 mm
longis totidemque lati, staminibus 3, pistillodio (?).

CotomsiA: Boyaca: El Umbo, 130 miles north of Bogota, “in high thick forest,
tree 12-16 {t., 4-6 a diam.,” Lani 47

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 171

This somewhat suggests 7. rubrivenium Poepp., but it is altogether un-
like that species in its leaf-margins.
Tetrorchidium popayanense sp. nov.

Arbuscula videtur, innovationibus subherbaceis, strigulosis, indumento
perbrevi luteo valde adpresso. Foliis pallide olivaceis, 16-23 cm. m. longis,
7—9 cm. latis, longe obovatis vel oblanceolato-ellipticis, apice petie: acumi-
natis, basi longe cuneatis, firme chartaceis subtus sub lente vix pubescen-
tibus, supra glabratis vel glabris subnitidis, nervis manifestis ca. 7—-9-jugis,

eglanduloso pubescente 2-3 cm. longo. Inflorescentia gracili pubescente
7-10 cm. longa parcius ramosa, é tanto visa. Floribus subsolitariis, calyce
ca. 2.5 mm. lato pubescente pedicello ca. 1.5—2 mm. longo fulto, lobis 3
ovato-acuminatis, ca. 1-1.25 mm. longis, 1 mm. basi latis, staminibus 3
oppositilobis sessilibus vel subsessilibus ligulam centralem minutam (an
pistillodium?) circumdantibus; bractea sub pedicello ane uaten Cer: Bea
mm. lata glandulis insignita cum illis in lamina congruent

CotomsiA: El Cauca: Highlands of Popayan, (?) Rio ae oe Lehmann
B7664.

The type locality is not better indicated. The characters of the foliage
suggest those of T. euriphyllum Stand]. of Central America, but the pri-
mary nerves are less numerous and the leaf-blade is differently shaped,
more markedly cuneate at the base, and with different marginal glands;
the glands on the floral bracts are also different.

Tetrorchidium jamaicense sp. nov.

Arbor ad 8 m. alta glaberrima. Foliis integris primum tenuiter membra-
naceis demum firme Lane eae ellipticis vel elliptico-obovatis, 7-12 cm
34.5 cm. latis, apice brevius rotundato-acuminatis, basi longe

)
glandulis cicatricosis ornato, stipulis minimis vel subnullis. Inflorescentia
6 tantum visa more generis gracili, effusa glabra parcius ramosa ad 18 cm.
longa; floribus glomerulatis ad 4—8, glomerulis sat distantibus (proximis in
thachi inferiore 1-2 cm. remotis), perianthio ca. 5 mm. lato, lobis 3 ellip-
ticis vel obovato-ellipticis margine intus subplicatis, staminibus sessilibus
3(—4), antheris more generis sat magnis.

Jamaica: Road to Holly Mount, Harris 8991.

This new species is certainly the plant which Fawcett & Rendle (FI. Jam.
4(2):317. f. 105. 1920) and Pax & Hoffmann (in Engl. & Prantl, Nat.
Pflanzenfam. ed. 2. 19C: 186. 1931) identify as representing T. rubri-
venium Poepp. It differs from Poeppig’s species, however, as represented
by Poeppig 1915 and Klug 3713, in having glabrous inflorescences, elliptic
rather than triangular perianth-lobes, and entire leaves which are more
coriaceous.

Sapium P. Browne
Sapium Bourgeaui sp. nov
Arbor vel frutex glaberrimus. Foliis ellipticis, apice recurvatim glandu-

172 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

loso-cucullatis utrinque fere aequo jure rotundato-acuminatis, supra obscure
subtus pallide olivaceis, margine sat conferte aristato-serratis, aristulis
nigricantibus marcescentibus, hic inde glandulis pustulosis ornatis, ca. 15
cm. longis, 3—4 cm. latis, nervis delicatis at perspicuis latius adscendentibus
ca. 15-jugis, petiolo 3-4 cm. longo apice glandulis 2 patentibus conicis
onusto, stipulis late rotundato-auriculatis margine valde coriaceo-scariosis.
Capsulae coccis delapsis (an revera hujus loci? soluti tantum adsunt) duris,
ca. 10 mm. longis, semine arillo secedibili rubro induto ca. 7 mm. longo
latoque.

Mexico: Veracruz: Orizaba, Santa Ana, Bourgeau 3020.

This collection is mentioned by Hemsley in the notes under S. mexicanum,
in Hook, Ic. 27: sub pl. 2680, p. 2. 1901. Hemsley states that it is neither
that species nor S. lateriflorum Hemsl. It is certainly not S. macrocarpum
Muell.-Arg. (S. mexicanum Hemsl.), S. appendiculatum Pax & Hoffm., or
S. pedicellatum Huber. The reflexed glandular tip of the leaf is a distinctive
character.
ie ick Cuatrecasasii sp. nov.

magna, innovationibus cicatricosis glabris saltem sub apice gaat
eine glandulosis magnis ornatis. Foliis 7-15 cm. longis, 2.5—5.5 ¢
latis, subcoriaceis, apice fere planis vix emarginatis, venulosis, supra otis
olivaceis, subtus brunneo-lutescentibus, ellipticis utrinque sat obtuse acu-
minatis glaberrimis, margine totis crenato-glandulosis crenis haud profundis
glandulis marginalibus hic inde obviis, nervis gracilibus patentibus saepius
haud anastomosantibus utrinque ca. 20 ultimis vix perspicuis, petiolo

florescentia spicata 2-sexuali sat crassa. Floribus ¢ ca. 12 in axilla squamu-
lae cujusvis glandula utrinque valde carnosa ad 4—5 mm. longa, perianthio
longe campanulato ca. 2-3 mm. longo et 2 mm. lato, lobo antico apap
staminibus 2 ad 3 mm. longis. Floribus 9? in anthesi ignotis submaturis ad
10 mm. longis 5 mm. latis, ovario glaberrimo nigro ovoideo stylorum cica-
trice subproducta notato, perianthii lobis 3 ovatis subscariosis intus medio
crasse costatis, glandula utrinque ellipsoidea sat magna.

CoLtomBIA: Putumayo: Valle de Sibundoy, alt. 2200 m., Cuatrecasas 11671.

This does not agree with any other species known to me from the region.
The comparatively short petioles, the robust spikes, and the produced scars
left by the fallen style on the ovaries are characteristic.

Sapium myrmecophilum sp. nov.

Arbuscula semimetralis. Foliis 5-8 cm. longis, 1-3 cm. latis, ellipticis
subcoriaceis, apice subplanis vel retusis, utrinque brevius acumin atis, ob-
scure conferteque crenulatis raro hic inde glandulosis, nervis paeacune ca,
15-jugis, petiolo vix 1 cm. longo vel minore glandulis 2 conicis patentibus
pustulosis, stipulis auriculatis. Inflorescentia spicata bisexuali ad 10-15
cm. longa. Floribus ¢ ca. 9 vel 10 in axilla bracteolae latae parvae,
perianthio aperiente vix 1—1.5 mm. longo, staminibus 2. Floribus ? sub-
sessilibus in anthesi haud visis: perianthii lobis scariosis (videtur) 3, ovario
maturescente globuloso depresso ca. 0.5 cm. magno, apice cicatrice minima
haud producta notato.

Coromsia: El Vichada: 60km. south of Orocué, Haught 2772.

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 173

This is described as a spreading shrub growing on ant-hills in the open
llanos. It is characterized by the close leaf-crenation and the very short
petioles.
eae Bed ae ieee in aia Ic. 27: pl. 2678. 1901; Huber in Bull. Herb.

s. II. 6: 439. f. 32.
Sapam hamatum ae “Ar Pax & Hoffm. in Pflanzenr. 52 (IV. 147. v): 229.
fig. 43 D,E. 1912. Syn
ce biglandulosum var. Dias Muell.-Arg. in Linnaea 32: 116. 1863.
Peru: Huanuco: Between Huanuco and Pampayacu, Kanehira 12; Pampa-
yacu, eanehie iva 28; San Martin: Zepelacio, Klug 3374; Loreto: Lower

Rio Huallaga, Williams 4771.

This is another of the entities which have been confused under the loosely
applied name S. Hippomane. The identifications are made on the strength
of the illustrations and the descriptions. The leaves are longer, narrower
and thinner than are those of S. Marmieri. A specimen from Colombia,
collected by Cuatrecasas along the Rio Guamues, Putumayo, may belong
here, but it lacks 9 flowers and fruits. Pax & Hoffmann err in replacing
S. Poeppigiti Hemsl. (1901) by S. hamatum (Muell.-Arg.) Pax & Hoffm.
(1912). The name which is published first in a given rank has priority,
and cannot be replaced by a later combination even though the basinym,
in this case a trinomial, happens to be the oldest name.

Sapium aereum Kl. ex Muell.-Arg. in Linnaea 32: 119. 1863; Pax & Hoffm. in
Pflanzenr. 52 (IV. 147. v): 233.1912

Brazit: Amazonas: MHumayta, Krukoff 6158, 6296, 6307; Sao Paulo de
Olivenga, Krukoff 8311.

This is an exceptionally critical entity which it would be desirable to
compare with Klotzsch’s own specimen. Only the last of the collections
cited was distributed as S. aereum, all others being referred to S. Hippo-
mane, While it seems quite likely that S. glandulosum (L.) Morong (S.
Hippomane G. F. W. Mey. et auct.) and S. aereum are close, they appear
to differ in the texture of the leaf and, to judge from the fruits of Pittier
11832 and Krukoff 8311, also in the fruit, this being possibly larger in 8.
aereum and somewhat differently shaped. Ina very definite sense S. aereum
connects S. glandulosum and S. Marmieri, differing from the latter primarily
in its leaves on the whole being smaller and narrower. Good fruiting ma-
terial is essential to a final elaboration of all these entities.

The vernacular name “Tapuru” appears on the label of Krukoff 6158.
A specimen which might belong here and strongly suggests the characters
described and illustrated for S. Taburu Ule (in Tropenfl. 9: Beih. 6:13.
fig. 3 D, E. 1905) is Klug 1668, Colombia, Putumayo, Umbria. I strongly
suspect that eventually it may be shown that S. aereum and S. Taburu are
uncomfortably close if not identical. Pax & Hoffmann express the belief,
Pflanzenr. 52 (IV. 147. v): 232.1912, that S. Taburu is hardly separable
from the entity they identify as S. Hippomane. In my opinion, this is the
result of the two authors lacking a clear understanding of the ranges of
these species. Unquestionably, a great simplification of the taxonomy of
this group follows if S. glandulosum (S. Hippomane) is excluded from the
Amazonian ranges of Brazil, Colombia and Peru.

174 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Sie ae capa Huber in Bol. Mus. Goeldi 3: 367. 1902, in Bull. Herb. Boiss. I.
6: 1906; Pax & Hoffm, in Pflanzenr. 52 (IV. 147. v): 256. 1912.
ee seis Gleason in Bull, Torrey Club 60: 364. 1933. Syn. Nov.

eRU: Loreto: Alto Rio Itaya, Williams 3490; Lower Rio Huallaga, Williams
4904, Killip & Smith 29265. Brazit: Matto Grosso: source of the Jatuarana
River, Krukoff 1656 (type collection of S. Leitera Gleason) ; aed Rio Purus,
Krukoff 5717; Amazonas: Sao Paulo de Olivenga, Krukoff 8098, 8428. COLOMBIA:
Putumayo: Puerto Ospina, Cuatrecasas 10784; Tolima: Curvas de Gualanday
(Ibagué-Girardot), Pérez-Arbelaez & Cuatrecasas 6490.

This species is of economic importance as a potential source of rubber,
for which it has been tapped in the past. The type was collected in the
region of the Rio Ucayali and the Rio Huallaga in Amazonian Peru. I
match the descriptions with Williams 3490 and 4904, which I accept as
representing this species. The leaf is essentially oblong to elliptic and more
or less obtusely rounded at the tip in these two specimens which, once
again to judge from the descriptions and the figures (see for instance
Hemsley in Hook. Ic. 29: pl. 2899. 1909), can hardly be separated from
S. eglandulosum Ule in Tropenfl. 9: Beih. 6: 14. 1905. The foliage of
Killip & Smith 29265 differs from that of the Williams specimens in being
distinctly rounded to short round-elliptic, with the tip of the blade often
retuse, but the remaining characters agree so well that, the range being in
common, it must be considered that all these specimens are conspecific.
As is well-known, considerable foliar differences are apt to occur in the same
species of Sapium depending upon conditions of growth.

All the collections cited from Brazil have leaves that tend to match those
of Killip & Smith 29265, but leaves of a pattern intermediate between those
of this specimen and the Williams material are present in Krukoff 8428.
The Colombian collections well match those from Peru, the record from
Tolima being interesting as an extension of the range, heretofore supposedly
restricted to the Andean regions of Peru, Brazil, and Colombia. Sapium
utile Preuss, to judge from Hemsley’s plate (in Hook. Ic. 29: pl. 2896.
1909), is represented in our herbarium by a Lehmann specimen from
Colombia, forests of Choc6-Micay and Timbiqui, which differs from S.
Marmieri in the more or less evidently serrulate leaf-margins as well as its
longer and narrower leaves. It is altogether likely that the records of
S. Hippomane from Peru, Brazil, and Colombia are based to a large extent
on misdeterminations of S. Marmieri and its allies.

Sapium aucuparium Jacq. Enum. Pl. Carib. 31. 1760, p. p. typ., excl. syn. Plumier.

Sapium jamaicense Sw. Adn. Bot. 62. 1829. Syn. Nov

Jacquin’s binomial has been extensively eee the discussion of
Hemsley, in Hook, Ic. 27: pl. 2650. 1901, and the treatment of Pax &
Hoffmann, in Pflanzenr. 52 (IV. 147. v): 229. 1912, merely adding to the
confusion.

The original publication reads as follows: “aucuparium. I. SAPIUM.
Plum. ic. 171. f. 2. Brown. Jam. 1. p. 338.,” the generic name being followed
by the conventional abbreviation to designate a woody perennial. - the
time when this publication was issued, one of its synonyms, “Plum. ic. 171.

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 175

f. 2.,” had already been cited by Linnaeus under Hippomane glandulosa,
Sp. Pl. 1191. 1753, which leaves Jacquin’s binomial standing solely upon
Browne’s “Sapium 1. Arboreum foliis ellipticis glabris, petiolis biglandulis,
floribus spicatis.” (Hist. Jam. 338. 1756). This synonym has been over-
looked by most authors, but not by Fawcett & Rendle, Fl. Jam. 4 (2): 325.
1920, who place it in the synonmy of S. jamaicense Sw. This binomial,
consequently, falls under S. aucuparium Jacq.

hree years after the publication of S. aucuparium, as quoted, Jacquin
again dealt with the same binomial, Select. Amer. Hist. 249. pl. 158. 1763,
repeating the original references but adding Hippomane glandulosa Linn.
and Plukenet’s “Tithymalus arbor americana, mali medicae folio . .
Jacquin’s illustration and description here apply to a very different plant
than the one he identified in 1760 as S. aucuparium. Naturally, this plant
cannot bear the binomial which was misapplied to it by Jacquin and very
numerous other authors. I discuss it as S. biglandulosum (L.) Muell.-Arg.

”)

Sapium biglandulosum (L.) Muell.-Arg. in Linnaea 32: 116. 1863, excl. var. fere

omn.

Hippomane biglandulosa L. Sp. Pl. ed. 2. 1431. 1763, p. p. typ., quoad syn. Jacq.

Sapium aucuparium Jacq. Select. Amer. Hist. 249. pl. 158. 1763, quoad ic. descr.

excl. syn. omn., non S. ee Jacq. Enum. PI. Carib. 31. oy

roe: talicifolium H. B. K. Nov. Gen. & Sp. 2: 52. 1817. Syn. N

Sapium Moritzianum Kl. in Seem. Bot. Voy. Herald 100. 1852; Huber i in Bull. Herb.
Boiss. II. 6: 358. f. 19. 1906; Pax & Hoffm. in Pflanzenr. 52 (IV. 147. v): 230.
1912. be Nov.

Panama: Coclé: EI Valle de Anton, Las Uvas, Allen 2575; Perlas Islands, Pedro
Gonzales, die rh CotomsBiA: Magdalena: siti? Marta, H.H. Smith 1916;
Boyaca: Orocué, Haught 2826. VENEZUELA: Ara : Ca rbanchito, Pittier
11802, Cagua, pei 12291; Distrito Federal: he Pittier 12401

Linnaeus is the author of both Hippomane glandulosa, 1753, and Hip-
pomane biglandulosa, 1763. Although the latter epithet might have been
mistakenly applied for the former, it proves impossible to treat it as a clear
unintentional error under the current Rules of Nomenclature, particularly
so in that it has been extensively used in botanical literature under different
generic names. The correct application of S. glandulosum (L.) Morong is
discussed below.

The original publication of H. biglandulosa reads as follows: “Hippomane
biglandulosa foliis ovato-oblongis basi biglandulosis. Sapium arboreum,
foliis ellipticis glabris, petiolis biglandulis, floribus spicatis. Brown. jam.
338. Sapium aucuparium. Jacq. amer. 31. t. 158. Mancanilla lauri foliis

oblongis. Plum. gen. 50. ic. 171. f. 2. Tithymalus arbor americana, mali
medicae foliis amplioribus tenuissime crenatis succo maxime venenoso.
Pluk. alm. 369. t. 229. f.8. Habitat in America calidiore.” The synonyms

and S. aucuparium, to which they belong. Since no specimen of H. glandu-
losa or H. biglandulosa is extant in the Linnaean herbarium (see Jacks. Ind.
Linn. Herb. 86. 1912), H. biglandulosa rests solely upon the plate and
description of S. aucuparium Jacq., 1763 non 1760. It is fortunate that

176 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Jacquin’s misapplication can easily be corrected by the reinstatement of
S. biglandulosum (L.) Muell.-Arg. in the sense here proposed.

Mueller-Argoviensis followed S. biglandulosum with a trinomial 2 Meveri-
anum, essentially based upon Meyer’s S. Hippomane. ‘This is taxo-
nomically not correct, but, as is well-known, a new combination stands
(Art. 54, Amsterdam Rules 1935) even if it involves a misapplied specimen.
Accordingly, the type of S. diglandulosum is the plant figured by Jacquin,
growing near Cartagena in Colombia, ‘‘inque ipso suburbio Xiximani ante
macellum.”’ The type-variety, on the contrary, is S. biglandulosum var.
Meverianum.

Sapium salicifolium H. B. K. has generally been treated as a doubtful
synonym of S. Moritzianum. The type material was collected at Morales
on the banks of the Rio Magdalena, and is without @ flowers or fruits.
Its description so closely agrees with the material I have seen that I do not
hesitate to accept it as representing S. biglandulosum. ‘The existence of
some varieties under this binomial is likely. The Venezuelan plant, for
instance, would seem to have slightly different leaves on the whole, as noted
by Huber, in Bull. Herb. Boiss. II. 6: 358. 1906, in his discussion of
S. aucuparium.,

Sapium glandulosum (L.) Morong in Britt. & Mor. in Ann. N. Y. Acad. Sci. 7: 227.

Hippomane tiedisiag L. Sp. Pl. 1191. 1753, p. p. typ., quoad syn. Pluk.,
Sapium eles e G. F. W. Mey. Prim. FI. Esseq. 275. 1818; aie in oes
Herb. Boiss. IL "6: 360, - 21, 1906; Pax & Hoffm. in Pflanzenr. 52 (IV. 147. v
231; ie p. p. Syn, N

— at Mul “Are in Linnaea 34: 217. 1865; Hemsl. in Hook. Ic. 27:

pl. 2650. 1900. Syn.

Persil net teas Huber i in Bull. Herb. Boiss. II. 6: 362. f. 22. 1906. Syn. Nov.

BARBADOS: _ Forester’s Hall Eggers 7238. Vereen Distrito
Federal: Naiguta, Pittier 7

Hippomane glandulosa L. rests exclusively upon two synonyms, “Man-
canilla lauri foliis oblongis. Plum. gen. 50” and ‘‘Tithymalus arbor ameri-
cana, mali medicae foliis amplioribus tenuissime crenatis, succo maxime
venenoso. Pluk. alm 369. t. 229. f. 8. Raj. suppl. 428.” Urban used one
of these polynomials, Plumier’s Mancanilla, in the synonymy of his own
S. caribeum, Symb. Ant. 3: 309. 1902, and referred the other, op. cit. 306,
to S. Hippomane G. F. W. Meyer. Sapium Hippomane Meyer was a new
name for Hippomane biglandulosa L., the Linnaean binomial being cited
by Meyer under his own.

Under the current Rules, the correct transfer of Hippomane biglandulosa
to Sapium can be effected only by publishing S. biglandulosum, which was
done by Mueller-Argoviensis but not by Meyer. Meyer’s name, conse-
quently, is illegitimate, and falls now under S. glandulosum (L.) Morong,
which is typified by Plukenet’s Tithymalus as interpreted by Urban. Here
also belong S. suberosum Muell.-Arg., based on a diseased condition of the
entity under discussion. Huber comments that the entities he understands
as S. Hemsleyanum and S. Hippomane are not certainly distinct as species,
but might easily prove to be varieties. This is possible, but the material

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE Lie

I have seen so far is too scanty to justify the publication of trinomials in
this difficult group; Eggers 7238, from Barbados, cited by Pax & Hoffmann,
is certainly very close to Pittier 11832 from the coast of Venezuela.

I cannot follow Pax & Hoffmann in accepting S. glandulosum (which
they call S. Hippomane) for the Amazonian forms of Brazil and Peru.
Everything indicates that Pax & Hoffmann confuse S. glandulosum with
S. Marmiert, The former would seem to have an essentially coastal range,
restricted to the Guianas, Venezuela, and some of the West Indian islands,
while the latter, as noticed elsewhere in this paper, is primarily an Ama-
zonian type.

Index Kewensis lists S. glandulosum Morong as an error for S. biglandu-
losum Muell.-Arg. and accepts as valid the combination of Druce in Rep.
Bot. Exch. Club Brit. Isl. 1913, 3: 423. 1914. I cannot accept this
interpretation. It is true that Morong treats H. glandulosa and H. biglan-
dulosa as synonymous, but the combination he made conforms with the
requirements of Art. 44[2] of the current Rules of Nomenclature in being
followed by a full reference to Hippomane glandulosa L.; the remaining
two citations can be excluded as misapplications without affecting in the
slightest the validity of the new combination. The fact that Morong’s
transfer was effected for a misapplied specimen does not make this transfer
illegitimate.

I have not seen authentic material of S. obtusilobum Muell.-Arg., but
Huber’s illustration, in Bull. Herb. Boiss. II. 6: 357. f. 17. 1906, suggésts
that this species might fall here as a trinomial if not as a straight synonym.

Sebastiania Sprengel
Sebastiania huallagensis sp. nov.

Arbuscula ca. 6-metralis innovationibus glaberrimis. Foliis elliptico-
lanceolatis, apice sat abrupte acuminatis, basi cuneatis, margine integris,
6-9 cm. longis, 2.5—-3.5 cm. latis, nervis gracilibus ca. 10-14- -jugis, petiolo
ca. 1.5 cm. longo apice glandulis 2 a en pustulosis sat magnis or-
nato. a ace ae bisexuali ad 8 cm. longa. Floribus ¢ ca. 6-8
in axilla squamulae ca. 1.5 mm. lata, aes minuto subsessili 3-lobato,
staminibus aitenilcbic 3 basi connatis. Floribus Q singulis pedicello ca.
0.5 cm. longo, ovario levissimo ca. 4 mm. longo et 1.5 mm. lato, basi
squamis imbricatis 3—5 circumdato, in stylis 2—3 crassiusculis divergentibus
abeunte.

Peru: San Martin: Juan Jui, Alto Rio Huallaga, Klug 4243.

Distributed as representing Alchornea sp. ?, which it is certainly not.
The sum of the characters suggests Sebastiania, but better material is
needed to confirm this disposition of it.

Sebastiania anisandra (Griseb.) Lillo in Seg. Contr. Conoc. Pe nl 16. 1917.

Actinostemon anisandrus Pax in Pflanzenr. 52 (IV. 147. v):

Dactylostemon anisandrus Griseb. in Abh. Ges. Wiss. Gotting. oe o 1879.

Paracuay: Alto Paraguay: San Lazaro, Rojas 5490.

The Rojas specimen, collected “entre rocas calcareas semi-sombrias,”’
is a good match for an Argentine specimen, Venturi 1350, from Tucuman,

178 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Famailla, originally distributed as representing S. anisandra. A younger
state of the same species is apparently represented by Venturi 5349,
identified by Lillo himself. This record seems to be a new one for the
flora of Paraguay.

Euphorbia Linnaeus (excl. Chamaesyce)
Euphorbia pelerer Vell. Fl. Flum. 5: pl. 74. 1827; Muell.-Arg. in Mart. Fl. Bras.
11(2

Euphorbia ae minor Muell. -Arg. op. cit. 689. Syn. No
— anomala Salzm. ex Boiss. in DC. Prodr, ea or 1862; Boiss. Ic. Euph.
5. pl. 38. 1866

mee is a widespread species, very close to E. lancifolia Schlect. of Mexico
and Central America. It is convenient to break it up in three subspecies
with a broad geographical background.

Euphorbia insulana subsp. typica subsp. nov.

Cyathiis ca. 2.5 mm. longis, inflorescentiis saepissime abbreviatis bracteis
subfoliaceis.

RAzIL: Ceara: Maranguapé, Drouet 2594; Parana: Guaratuba, Dusén

13518; Sao Paulo: Ilha Queimada, Gehrt 4579.

The type is Vellozo’s figure. Gehrt 4579 is altogether true to Mueller’s
description of minor, but impresses me as having been taken from a weak
shoot of the plant.

Euphorbia insulana subsp. tovarensis (Boiss.) comb. n
Euphorbia tovarensis Boiss. Cent. Euph. 19. 1860; in DC. "Prodr. 15(2): 59. 1862.
CotomsBiA: Santander: Between El Roble and Tona, Killip & Smith 19427.
This trinomial is very near F. lancifolia, resembling it in the rather

diffuse and robust inflorescence and in the comparatively large cyathia.

Bang 2208, collected at an unrecorded locality in Bolivia, also belongs

here; this saree is erroneously listed as FE. geniculata in Bull. N. Y.

Bot. Gard. 4: 441

Euphorbia insulana subsp. pileomayensis subsp. n
A formis totis foliis bractealibus discedit saepius longe ellipticis, cyathiis

minoribus, inflorescentiis magis delicatis.

ParaGuay: Pilcomayo River, Morong 867 (tTypE); Between Rio Apa and
Rio Aquidaban, Fiebrig 4393. ARGENTINA: Formosa, (?) Jorgensen 3081; Chaco
(Argentina ?): Fontana, Meyer 232

The habit sets this new subspecies rather sharply apart from the others,
but a broad concept of specific limits, necessary in this case, forbids its
being treated as a full-fledged species. Euphorbia Mandoniana Boiss.,
of which I have seen only the meagerest scraps, may prove to be an extreme
alpine form of this complex.

Euphorbia Huanchahana (KI. & Garcke) Boiss. in DC. Prodr. 15(2) +: 103. 1862.
Tithymalus huanchahanus K\. & Garcke in Abhandl. Akad. Wiss. Berlin 71. 1860.
As in many species of this genus in the South American range, it proves

to be impossible to adopt for this entity a narrow concept of specific

limits. The material I have at hand indicates the existence of two main
aggregates, one localized in Peru, the other in southern Bolivia and northern

1943 | CROIZAT, AMERICAN EUPHORBIACEAE 179

Argentina. In addition, the aggregate of Peru and Bolivia is divided into
two forms, one glabrous, the other rather pubescent, which bear to each
other a varietal relationship

Euphorbia Huanchahana subsp. typica subsp. nov.

Foliis minutis saepissime margine serratis glabris

I have seen a photograph of the type, ollie in “Canta Prov. Peru.”
This material is to all appearances well matched by a Peruvian specimen
from Matucana, Dept. Lima, Savatier 1356.

Euphorbia Huanchahana subsp. penazuelensis subsp. nov.

Foliis carnosulis margine integris subintegrisve indumento sat conferto
a subsp. typica discedit.

ARGENTINA: Tucuman: Sierra Calchaquies, Pefias Azules, 3400 m., Burkart
5306 (TYPE); Catamarca: Andalgala, Cerro Negro, alt. 3500 m., Jorgensen 1232.

The habit is that of an alpine plant, the branches being rosulate and the
rootstock much thickened.

Euphorbia Huanchahana var. peperomioides var. nov.

Pusilla, plus minusve rosulata a subsp. penazuelensi, cujus est, qua-
cumque notis caeteris totis optime convenit glabritie primo intuitu recedit.

Borivia: Mandon 1068.

This variety belongs to subsp. penazuelensis, with which it agrees in
habit and foliage, differing only in indumentum. The Mandon material
I have seen is probably identical with the Weddell specimen from Bolivia
cited by Boissier.
gers feck ea Lam. Enc. Méth. 2: 421. ae — in DC. Prodr. 15(2):

; Muell.-Arg. in Mart. Fl. Bras. 11(2):

ae species is closely allied to E. sean i. emend. Spreng., which
ranges throughout Argentina and Chile. It is restricted in its range to the
regions immediately adjacent to the mouth of the Rio de la Plata in
Uruguay and Argentina. It falls into two readily separable groups.
Euphorbia caespitosa subsp. typica subsp. nov.

Foliis obovato-ellipticis vel spathulatis, apice rotundatis.

Urucuay: Vicinity of Montevideo, Fruchard 182, Arechevaleta 5194a.

The cited asco agrees perfectly with Lamarck’s type in the herbarium
of the Paris Muse
Euphorbia caespitosa subsp. ventanicola subsp. nov.

Cum subsp. typica in floralibus optime convenit, sed foliis apice bene
acuminatis, tees natali alieno primo intuitu distincta

: Aires: Cerro Naposta, Sierra de la Ventana, Von
Rentzell 1082 (TYPE) ; Pigiié, Burkart 4706.

This is a well-marked form, readily recognizable by the characters of the
foliage and by the range, which centers in the Sierra de la Ventana. It is
likely that this plant is included in part by Boissier in his concept of
E. portulacoides acutifolia Boiss., in DC. Prodr. 15(2): 103. 1862, and is
the one not altogether correctly identified as E. caespitosa by Spegazzini,
Contr. Fl. Sierra Vent. 54. 1896; Contr. Fl. Tandil 47. 1901. E£. portula-

180 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

codes Spr. [sic] normalis O. Ktze. is based upon a plant collected in the

Tandil, O. Kuntze, Rev. Gen. 3: 286. 1898, which I have not seen but

which most likely belongs here. This trinomial is validly published and

must be used for the typical form of E. portulacoides L. emend. Spreng.,

despite Kuntze’s probable misapplication.

sce sciadophila Boiss. in DC. Prodr. 15(2): 57. 1862; Muell.-Arg. in Mart.
as, 11(2): 687. pl. 97. 1874.

Pekin Tucuman: La Criolla, Rodriguez 502.

This species is common in southeastern Brazil and in Paraguay, but I
have seen only the specimen cited from Argentina. The record seems to
be new. The affinities of E. sciadophila with the Peruvian E. adtanthoides
Lam. require careful study.

Euphorbia spathulata Lam. Enc. Méth. 2: 428. 1788; Boiss. in DC. Prodr. 15(2):
> Muell.-Arg. in Mart. Fl. Bras. 11(2): 701. 1874; Croiz. in Torreya

42: 16. 1942, in nota.
Euphorbia a bi Fisch. & Mey. in Ind. Sem. Hort. Petrop. 2: 37. 1836;

Boiss. in DC. Prodr. 15(2): 135. 1862; Nort. in Missouri Bot. Gard. Rept. LL:
106. pl. 22, os oo. Wheeler in Kearn. & Peebl. Fl. Pl. Arizona 539. 1942.
Syn. Nov.

The suspicion I have already voiced that E. spathulata is merely an
introduced weed in the regions of the La Plata and that it is the same as
E. dictyosperma Fisch. & Mey. of the southeastern United States is con-
firmed. No differences can be found to separate such specimens, for
instance, as Culwell & Timmons 3065 (Central North Texas) and
Lombardo 3903 (Montevideo), Scala 90, and Burkart 3747 (Mouth of
the Parana).

Euphorbia invaginata sp. nov.

Herbacea, glabra, caulibus fistulosis striatis hic inde ceraceis. Foliis
carnosulis ligulatis vel longius elliptico-obovatis subeveniis, margine integ-
ris, apice mucronatis, basi longe acuminatis, epetiolatis, 5-6 cm. longis,
1-1. 75 cm. latis, nervo medio validiusculo, stipulis petiolaribus subnullis vel
nullis. Inflorescentiis terminalibus effusis ad 15 cm. longis bracteato-

c

longis, bracteis subpetaloideis pallidis hic inde albicantibus tenuissime ve-
nulosis late ovato-cordatis ad 2 cm. longis totidemque latis mucronulatis.
Cyathiis pions ee singulis bracteis occultatis invaginatisque, ad 4 mm.
longis et 2 mm. latis, cylindrico-campanulatis, nectariis 4 vel 5 stipitatis
uae eee corrugatis parvis appendice petaloidea nulla, pedicello
ca. 1.5—2 mm. longo; capsula glabra levi, coccis delapsis ad 7 mm. longis
angustis, semine valde elongato angusto ambitu tetragono, vix 1.5 mm
crasso, 6 mm. longo, arillo albicante hic inde granulato-leproso, caruncula
rotundata bene umbonata stipitataque

Paraguay: Chaco: Picuyba, Rojas 7268.

This new species belongs to Boissier’s Sect. Stachydium, which includes
E. comosa Vell., E. lupulina Boiss., E. Gollmeriana Boiss., E. foltiflua Ule,
and the African E. phyvlloclada Boiss. From all the American species F.
invaginata differs in the length of the seed. In foliage it most closely
resembles E. Gollmeriana Boiss. and E. folitflua Ule.

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 181

Euphorbia aureocincta sp. nov.

Herbacea hirta fistulosa, serius glabrata. Foliis (ut videtur) miro modo
ludentibus, nunc more Amaranthi ssp. obcuneatis vel grosse quadrangulis,
margine profundius irregulariter lobulato-dentatis, 3-6 cm. longis, 2-3 cm.
latis, tum exquisite elliptico-lanceolatis margine integris 5-12 cm. longis
0.5-1.5 cm. latis, apice acutis, basi breviter cuneato-angustatis, petiolo
semper brevi vix 1.5 cm. longo hirtulo. Inflorescentiis coarctato-capitulatis,
bracteis lineari-lanceolatis acutis, 3-7 cm. longis, 0.5—1 cm. latis, integer-
rimis, basi pulchre aureis; cyathio ca. 3 mm. longo fauce 2.5-3.5 mm. lato,
lobis lacerato-ciliatis, nectario unico sat plano, ovario in anthesi subincluso,
capsula depresse rotundato-trigona ca. 5 mm. longa et lata, stylis vix 1.5
mm. longis ad tertium inferum partitis; semine 3.5 mm. longo, 2 mm. lato,
quadrangulo, arillo albicante vel pallide brunneo toto induto, basi truncato,
apice longiuscule acuminato, sub apicem atque ad medium leviter constricto-
zonato, hic inde verruculoso-lineato.

ParaGuay: Carapegua, Callistro, Rojas 3379 (type). ARGENTINA: Jujuy:
Quebrada del Chanhi, Schreiter 10990.

This is a well-marked form, but its ultimate rank, whether binomial or
trinomial, is a matter of speculation. It belongs to the group of E. elliptica
Lam. (E. geniculata Ort.; E. prunifolia Jacq.), and its taxonomic status
would seem to match exactly that of E. zonosperma Muell.-Arg. Unlike
that species, which is widely distributed, E. aureocincta appears to be
restricted to Paraguay, Argentina, and possibly Bolivia. The Schreiter
specimen cited above requires verification, for it represents a state with
leaves of amaranthoid pattern not resembling, at a glance, the typical
form, Rojas 3379, which has only narrowly lanceolate leaves. However,
both in Rojas 3379 and Schreiter 10990 the floral parts are identical, and
some of the leaves are very similar. Euphorbia heterophylla L. B elliptica
f. hirticaulis O. Kuntze, Rev. Gen. 3: 286. 1891, probably belongs here.

Euphorbia acerensis Boiss. in DC. Prodr. 15(2) : 55. 1862.

ARGENTINA: Tucuman: Villa Lujan, Venturi 524 (? 324); Salta: Rio
Toro y Rio Blanco, Vattuone 17.

The cited material was misdetermined as representing E. adianthoides
Lam. All the species in this group are closely related, and the existence of
intermediates between E. acerensis Boiss. and E. Poeppigi Boiss., which
ranges from the Amazonian regions of Peru to Bolivia, is probable.
Euphorbia pentadactyla Griseb. in Abhandl. Gesell. Wiss. Goettingen 24: 63, 1879.

ParaGuAy: Gran Chaco: Carandaity, Rojas 7287.

This species resembles /. aureocincta Croiz. but is easily recognized as
distinct on account of its long filiform simple styles. Tshe record seems to
be new for Paraguay. Earlier records are all from Argentina.

Euphorbia Milii Des Moul. in Bull. Hist. Nat. Soc. Linn, Bordeaux 1: 27. pl. 1.
1826; Desf. Cat. Hort. Paris. ed. 3. 475. 1829; Croiz. in Jour. Arnold Arb. 21:

Euphorbia splendens Boj. ex Hook. in Bot. Mag. 56: pl. 2902. 1829; Denis, Euph.
Iles Austr. Afr. 82. 1922.
Paracuay: Asuncion, cultivated in the Botanical Gardens, Rojas 1264.

182 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

The reasons calling for the reinstatement of Des Moulins’ neglected
name have been given in my paper cited above.

Euphorbia Hinkleyorum t. M. Johnst. in Contr. Gray Herb. n. s. 70: 72. 1924.

ARGENTINA: Jujuy: Tilcara, Cerro Pefia Alta, Venturi 49106.

The classical locality is Mt. Chachani, near Arequipa, Peru. This is a
new record for Argentina, and the species may be expected from Bolivia.
The forms in this group bear an interesting relationship to E. claytonioides
N.E. Br. of Angola in West Africa.

Euphorbia pampeana Speg. in Rev. Jard. Zool. Buenos-Aires 1: 30. 1893.

Urucuay: Canelones: Las Brujas, Lombardo 1959.

This is probably a new record for Uruguay. The polymorphism of this
species under conditions of experimental cultivation is hardly credible; the
leaves vary from obovate to narrow-linear and from manifestly pubescent
to fully glabrous, as I have observed the species in cultivation.

Euphorbia a Mart. in Spix & Mart. Reise Brasil 2: 612. 1828, im nota;
Boiss. in DC. Prodr. 15(2): 176. 1862; Muell.-Arg. in Mart. Fl. Bras. L1(2):
692. pl. 4 1874; Mansf. in Monatschr. Kakt. -Gesell. 3: 244, 1931

Euphorbia vhinsaloides Glaz. in Bull. Soc. Bot. France 59(Mém. 3g): 638. 1912,

nec alior.

Brazit: Bahia: Queimadas, Rose & Russell 19848 (in herb. N. Y. Bot. Gard.).

The peculiar phosphorescent sap of this plant, noted by Martius and by
Glaziou, is often mentioned in the literature, but I have seen only the cited
specimen which may belong here. The place of publication of the binomial
is variously reported, but I am satisfied that the reference alae above is
correct, for the Latin description is given in the footnote cited; “‘p. 726,”
cite by Boissier, Index Kewensis, and most authors, merely contains
Martius’s comments on the phosphorescent properties of the latex.
Mansfeld puts this species in the Sect. Pteroneurae together with E. Weber-
baueri Mansf., E. Sipolisii N. E. Br., and E. pteroneura Berger. My
understanding, on the contrary, is that E. phosphorea has a distinct position
of its own.

Euphorbia orizabae Boiss. in DC. Prodr. 15(2) : 147. 1862.

GUATEMALA: Quiché: Nebaj, 6400 ft., Skutch 1734; Chimaltenango:
Cerro de Tecpam, alt. 2400-2700 m., Standley 61046; Quezaltenango:
Ostuncalco, alt. 2700 m., Standley 66410.

The record is apparently new for Central America. The peculiar velu-
tinous indumentum of the branchlets and floral parts immediately separates
this species from the forms around E. campestris Cham. & Schlecht.

Chamaesyce S. Gray emend. Croizat

The difference in habit between Chamaesyce and Euphorbia in a narrow
sense is said by Wheeler, in Rhodora 43: 99. 1941, to have been caused
by a process of reduction in the main axis, as follows, “When by pro-
gressive reduction of the main axis subg. Chamaesyce finally arrived at
the habit of branching after the first pair of true leaves appeared, the plant
was obviously too small to produce all the elaborate foods necessary for a

1943 | CROIZAT, AMERICAN EUPHORBIACEAE “183

production of a cyathium with its reproductive structures requiring abun-
dant protein, fats, and carbohydrates; so we find that the cyathium which
would otherwise terminate the main axis is omitted.”

This account does not require explicit refutation for the benefit of any-
body acquainted with plant physiology. The ultimate destination of food
is ruled in living organisms by highly complex metabolic equations, and
no plant is ever so bereft of “abundant protein, fats, and carbohydrates”
as to be incapable of yielding one flower, or cyathium, in lieu of one or
several vegetative buds. The Chamaesyce which, according to the explana-
tion just quoted, is so weak (let us notice: phylogenetically) as to deny
itself the luxury of an apical reproductive structure, is vital enough to
produce up to five or six buds set around the portion of the stem which
ought to bear the apical cyathium but is said to be incapable of doing so.
These buds, in their turn, may yield an internode which is immediately
floriferous, so that several cyathia may be brought forth immediately above
the point at which not a single one could arise on account of the lack of
proteins, fats, and carbohydrates, if the explanation of Wheeler were to
be accepted.

This is not all; Chamaesyce includes at least one-third of the species
commonly treated as Euphorbia and is the most widespread of the Euphor-
biaceae, with the exception, perhaps, of Euphorbia Sect. Tithymalus in the
sense of Boissier. Its vitality is astounding, and its morphologic range
exceedingly varied, including fugacious annuals barely a few inches long,
and trees in which a true woody trunk appears formed by the ultimate
fusion of the internodal growth peculiar to the group. There is not the
slightest evidence to favor the belief that this group has been derived in
evolution from some other aggregate already differentiated as Euphorbia
in the modern sense; its range, morphology, and physiology point to its
being one of the archetypes of the Euphorbiaceae, certainly not a mori-
bund offshoot of some “Section” of the Linnaean genus. The interpetiolar
stipules of Chamaesyce do not seem to be homologous with ordinary
stipules but to have arisen in evolution by the reduction and specialization
of a quaternate foliar verticil, this in itself being an indication that the
theory of progressive reduction advanced by Wheeler neglects the fact that
specialization and differential growth, rather than reduction, are involved
in the shortening of the axes of greens It will be obvious that,
treated as a section, a subgenus, or a s, Chamaesyce is not to be
interpreted as suggested by Wheeler; 7 ms abate morphology, and
life-history contradict this interpretation on the strength of factors which
have nothing to do with the taxonomic and nomenclatural preferences of
an author.

Chamaesyce may be treated in subordination under Euphorbia by any
botanist who accepts traditional values as absolute, and it is not my inten-
tion to dispute the legitimacy of such a point of view so long as it is know-
ingly held. I accept Chamaesyce as a genus for the following reasons:
(1) it includes not less than 600 species and manifestly stands for one of
the largest aggregates in the Euphorbiaceae; (2) the difference is fully as

184 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

great between Euphorbia and Chamaesyce as it is between Mallotus and
Macaranga, Glochidion and Phyllanthus, Alchornea and Cleidion, Cnido-
scolus and Jatropha, and the like (this vital fact is generally unknown to
local students of Euphorbia); (3) the peculiarities of the stem-abortion
of Chamaesyce, so far dismissed as “habit,” are of far-reaching phylo-
genetic and morphological significance; (4) the species under C pragniees
with few exceptions (probably not more than 10-15 species in the
called by Boissier Euphorbia sect. Zygophyllidium) are readily identifiable
in the herbarium; (5) the characters of the nectaries on the cyathium,
which are currently used to segregate from Euphorbia about 50 species of
Monadenium and Synadenium and a single species of Diplocyathium (see
Pax & Hoffm. in Engl. & Prantl, Nat. Pflanzenfam. ed. 2. 19e: 43-44.
1931), could consistently be used to break up the Linnaean genus into
many genera (Dactylanthes, for instance), thus reintroducing in its classi-
fication the confusion which Boissier sought to eliminate. To reject spuri-
ous “floral characters,” full of unwelcome possibilities for classification,
and to take up in their stead broad morphologic and phylogenetic concepts
is sound and conservative taxonomy.
Chamaesyce chamaerrhodos (Boiss.) comb. nov.

Euphorbia chamaerrhodos Boiss. Cent. Euph.. . 1860, in DC. Prodr. 15(2): 51

1862, Ic. Euph. 13. pl. 25. 1866.1

ParaGuaAy: Carapegua, Rojas 3352.

The collection cited is excellent evidence of the dimorphism of this
species, the crowded short florigerous axes being unlike the long and sterile
shoots. The former strongly suggest the growth of Chamaesyce
potentilloides (Boiss.) comb. nov. (Euphorbia potentilloides Boiss.),
the latter that of C. Airta (L.) Millspaugh.

Chamaesyce Selloi (Boiss.) comb. n

Euphorbia Selloi Boiss. in DC. Prod. 15(2): 50. 1862, Ic. Euph. 13. pl. 22. 1866.

This is a collective species with numerous forms. The one described
below differs from the type, illustrated by Sel/o 170, in the characters of
the seed
Chamaesyce Selloi var. brevisemina var. nov.

mine ovoideo potius quam trigono-acuminato in faciebus inter costulas
iano potius quam impresso a formis typicis recedit.

ARGENTINA: Entre Rios: Concordia, Burkart 822.

Chamaesyce Meyeniana (KI.) comb. nov.
a hig KI. in Nova Acta Acad. Leop.-Carol. Nat. Cur, 19: Suppl.
3 (Meyen. Obs. Bot.) ; Boiss. in DC. Prodr. 15(2): 42. 1862.
aiestiilee Chaco Paraguayo: Puerto Casado, Rojas 2171; Isla Poi,
Rojas 7070.
The former specimen particularly is an excellent match for such Bolivian

1See Briquet, in Bull. Soc. Bot. Suisse 50a: 57, footn. 1. 1940, for the date of this
work. Pritzel errs in both editions of the “Thesaurus,” giving the date for the Icones
as . The “Centuria Euphorbiarum” is overlooked by Pritzel and by most
bibliographers.

1943 J CROIZAT, AMERICAN EUPHORBIACEAE 185

collections as Pentland (?) 109 and D’Orbigny 1207. ‘This is a new
record for Paraguay.
Chamaesyce Ejichleri (Muell.-Arg.) comb. no

Euphorbia Eichleri Muell.-Arg. in Jour. Bot. “12: 232. 1874.

Paracuay: Chaco: Puerto Casado, Rojas 2170; Loma Pora, Rojas 2969.
ARGENTINA: Tucuman: Tapia, Venturi 2320; alta: Oran, Schreiter 10991.

This appears to be a new record for Paraguay. The identifications were
made on the basis of a photograph of Lorentz 301, in the Delessert
herbarium, and the description.

Chamaesyce Lorentzii (Muell.-Arg.) comb.

Euphorbia Lorentzii Muell.-Arg. in Jour. oe 12: 231. 1874.

Urucuay: locality °?, Arechevaleta 5192 a. ARGENTINA: e
Belgrano, Parodi 9879; Tigre, Parodi 11095, Hicken 441, Burkart 5711; Los Talas,
Marelli 39; Belgrano Bits, Burkart 3632; Delta Parana, Rapa 8357.

The record for Uruguay is new, I believe. The determinations were
based on a photograph of Lorentz 466, from the Berlin herbarium, and the
description. This species tends to be restricted to very moist habitats.
Chamaesyce emarginata (KI. & Garcke) comb.

Anisophyllum emarginatum Kl. & Garcke in Abhandl, Akad. Berlin 24.

Euphorbia emarginata Boiss. in DC. Prodr. 15(2): 32. 1862; Muell. a in oes FI.

Bras. 11(2): 681. 1874.

Urucuay: Rio Negro: Isla del Pedion, Rosengutt B 1472. ARGENTINA:
Entre Rios: Gualeguaycht, Burkart 4146.

The Burkart record requires confirmation, as the determination was
effected from fragmentary material. The identifications were made from
the descriptions and on the basis of a photograph of Sellow, the type
specimen in the Berlin herbarium.

Chamaesyce hirtella (Boiss.) comb. n

Euphorbia hirtella Boiss. Cent. Euph. : 1860, in DC. Prodr. 15(2): 24. 1862.

Urucuay: Canelones: Las Brujas, Lombardo 1958.

This is an exceedingly critical entity which probably connects two or
three species that may be discussed later. The record is apparently new
for Uruguay.

Chamaesyce Duckei sp. nov.

Perennis basi lignosa, caulibus stricte adscendentibus eee in-
novationibus molliter albo-lanulosis citius glabra atis. Foliis novellis mem-
branaceis denon albo-lanulosis vel glabratis, manifeste petiolatis, eee
ad 1.5 cm. longa, ca. 0.3 cm. lata, lanceolata vel elliptico-lanceolata vel
anisophylla, margine subintegra, petiolo gracillimo ca. 2-3 mm. longo,
stipulis linearibus vel triangularibus minutis. Cyathiis subsolitariis longe
campanulatis ad 1-1.5 mm. longis, fauce ca. 1 mm. la is, nectariis ellip-
soideis centro impressis, appendicibus petaloideis plus minusve profunde
laciniato-sectis albicantibus, flore @ obpyriformi albicante tomentello vel
lanuloso, capsula submatura glabrescente ovoideo-trigona ad 2 mm. longa
et 1.5 mm. lata, stylis gracilibus ad 1.5 mm. longis ut videtur integris.

Brazit: Para: Furnas, on the Middle Tapajoz, Ducke 18534.

The characters of this plant are outstanding, and that it represents a new

186 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

species seems to be obvious. The material, however, is hardly satisfactory
for a generalized description, because it shows a stage in which the new
growth is barely beginning, but the old branchlets have already lost their
leaves. It is altogether likely that the leaves and stipules of a free grown
shoot will not be found to agree closely with those here described.
Chamaesyce Barberiana sp. nov.

Herba annua vel potius perennans vix ultra pedalis suberecta multicaulis,
caulibus in sicco stramineis vel pallide brunneis crispule albido- puberulis,
internodiis ipsis maximis vix 3.5—4 cm. longis, stipulis interpetiolaribus in
laciniis subintegris acutatis 3 vel 4 dissectis; foliis more generis basi aniso-
phyllis, apice obtuse acuminatis vel rotun datis, membranaceis, 0.5—2 cm
longis, 0.5-1 cm. latis, ellipticis vel rot tundato- ellipticis nequaquam

tilagineo sat obtuse distanterque serrato. Inflorescentiis apicalibus capituli-
formibus ca. 1—1.5 cm. longis fere totidem latis, confertis minutissime
bracteolatis saepius dichotomis dein iterum 2- vel 3- " artitis: cyathio _
vel puberulo longiuscule campanulato raro subinflato c

mm, lato, nectariis diminutis appendice rotundata albicante ee dae
lobis minutissimis subtruncatis, floribus ¢ paucis; capsula evidenter longi-
ore quam lata ca. 2 mm. longa, basi ca. 1.25 mm. lata, stylis vix 1 mm.

constrictis, columella gracili ca. 1.5 mm. longa
rubello ca. 1—-1.25 mm. longo, 0.4—-0.6 mm. basi lato, faciebus transverse
ruguloso-insculptis.

Paraguay: Chaco Paraguayo: Irendagué, Rojas 7213 (TYPE). ARGEN-
TINA: Santiago del Estero: C.. Pellegrini, Venturi 5663
Tucuman: Burruyaco, Venturi 769 Entre Rios: Parana, Burkart
439; San Luis: Sierra del Gigante, pe 67; Cordoba: Casquin (?),
Rodrigo 251.

This in the main is the entity which I have mistaken, in Lilloa 6: 299.
1941, for C. indica (Lam.) Croizat. My error is not entirely unaccount-
able, because the vegetative parts of these plants are practically identical.
However, the seeds of the two species, which only recently I have had the
yas of studying to my satisfaction, are altogether unlike. Seeds
re iam Oe ica are more or less ovoid and dark in color, while those of
tg pe ie are narrow, pointed, and reddish brown.

Chamaesyce Barberiana is a strong species which closely resembles no
other of its range. In Boissier’s monograph it would take its place immedi-
ately next to C. Berteriana (Balb.) Millsp. of the West Indies. The
specific name honors Dr. Andrés Barbero, President of the Sociedad
Cientifica del Paraguay, to whom so much is owed by all students of the
natural history of that Republic.

Chamaesyce portucasadiana sp. nov.

Planta certissime perennis e caudice lignoso ramos plures duros repentes
brevissime albicanti-tomentellos vel rarius glabratos edens ad 30 cm. longos
et ultra, stipulis setaceis triangularibus inconspicuis. Foliis saepius valide
costatis ellipticis vel ovato- ellipticis, sealed petiolulatis, 0.4—-1 cm.
longis, 0.2—0.7 cm. latis, plus minusve profundius serratis adpresse setulosis.
Cyathiis cupuliformibus in axillis singulis vel subsingulis, ca. 2 mm. longis

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 187

et latis, puberulis, nectariis late albo-appendiculatis, ovario rotundato-
trigono albicante tomentello vix 1.5 mm. longo latoque, stylis brevibus apice
bilobis

Paraguay: Chaco Paraguayo: Puerto Casado, Rojas 2152.

This plant was originally identified as representing grag ee thymtifolta
L., a determination probably influenced by Chodat & Hassler’s earlier
acceptance of this species for the region. A full discussion of FE. thymtfolia
is here impossible, but on the basis of Metz 67, an Asiatic specimen which
Boissier cites under that binomial, in DC. Prodr. 15(2): 47. 1862, it is
obvious that C. portucasadiana has characters wholly incompatible with
those of Linnaeus’ species as represented by the Metz collection. The
description of E. argillicola Chod. & Hassl., if at all correct, cannot
apply here.

Chamaesyce oranensis sp. nov.

Perennans lignescens, rosulata vel repens suberecta, tota hispidulo-
velutinosa pallide olivacea vel grisea, internodiis pro more nec u
aos maximis 1—2 cm. longis. Foliis rotundato-ellipticis, 0.5—1 cm. longis,

m. latis, velutino-puberulis, margine sat grosse serratis, subsessilibus,
tiple setaceis minutis deciduis. Cyathiis in eta pluribus aggregatis

campanulatis valde tomentellis ca. 1.5 mm. longis, nectariis rotundatis
minimis appendice petaloidea subnulla, flore @ ae stylis brevissimis
glabris partitis; capsula ovato-trigona ca. 1.25 mm. longa 1 mm. basi plus

minusve lata, semine ellipsoideo ee ie nee ruguloso ca.
1—1.25 mm. longo : = 75 mm. lato.

ARGENTINA: Sa a: Oran, Venturi 5555 (TYPE). ParaGuay: Chaco
Paraguayo: Nee Rojas 7224.

The floral characters of this new species are not outstanding, but the
habit is distinctive and is immediately recognizable. The internodes are
usually only 0.5 cm. long, and the stems become manifestly woody with
age.

Chamaesyce catamarcensis sp. nov.

Humilis glaberrima, caules verosimiliter annuos e radice perenni edens.
Foliis crassiusculis integerrimis linearibus, apice obtuse rotundato-apiculatis,
7-14 mm. longis, 1-1.5 mm. latis, petiolo vix 1-2 mm. longo, stipulis inter-
petiolaribus fimbriatis minutis. Cyathiis i in axillis singulis ob internodiorum
brevitatem in pseudocymulis apicalibus congestis vix 1 mm. longis, nectariis
4 vel 5 exappendiculatis vel parcissime appendiculatis in involucrum Jonge
decurrentibus, staminibus paucis; flore @ elongato trigono glaberrimo,
stylis 3 brevissimis partitis; capsula matura ca. 2 mm. longa et 1.25 mm.
lata, gynophoro 1.5—-2 mm. longo, semine quadrangulo apice valde acumi-
nato basi truncato, in one rugis profundis horizontalibus ad 6-10
ornato, arillo albicante, testa rubrobrunnea.

ARGENTINA: Catamarca: Andalgala, Jorgensen 1621.

This resembles C. caecorum but has blunter leaves and a different seed.
Chamaesyce caecorum (Boiss.) comb.

Euphorbia caecorum Mart. ex Boiss. in DC. Prodr. 15(2): 51. 1862; Muell.-Arg. in

. Fl. Bras. 11(2): 675. pl. 92. 1874; oe Ic. Euph. eee 23. 1866; Chod.
& Hassl. in Bull. Herb. Boiss. II. 5: 681. 1905.

fp
+

188 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

The spelling caecorum is to be retained as the one used by Boissier in
the original publication. Mueller’s reference to the place of publication is
garbled, confusing as it does the unpublished “Pl. Med. Bras. t. 73 ined.,”
cited by Boissier, and the “Icones Euphorbiarum.”

This species is frequent in Brazil and probably not rare in Paraguay,
witness: Rojas 6339, Sierra de Amambay. I have so far not seen it from
Argentina. The ternate and quaternate verticils, illustrated by Mueller
and Boissier, on the lower nodes suggest a theoretical primitive condition,
antedating the transformation of two leaves of the verticil into interpetiolar
stipules.

Chamaesyce hirta (L.) Millsp. in Field Mus. Publ. Bot. 2: 303. 1909.
Euphorbia hirta L. Sp. Pl. 454. 1753; Boiss. in DC. Prodr. 15(2): 21. 1862 (as E.
pilulifera); Wheel. in Contr. Gray Herb. 127: 67. 1939, in Rhodora 43: 169.

This widespread weed has been confused both in herbaria and in the
literature with C. pilulifera (L.) Small. This confusion arose through
accepting a concept of E. pilulifera L. based on the plant described in the
Amoenitates Academicae 3: 114. 1756, rather than on that of the Species
Plantarum (1753), which has priority. As Boissier points out (op. cit. 20),
the plant originally determined as E. pilulifera in the Linnaean herbarium
actually represents E. parviflora L., which was not published until 1759.
To E. pilulifera L. and the combinations based upon it, E. parviflora L.
must be added as a synonym.

Chamaesyce hirta L. subsp. procumbens (Boiss.) Croiz. in Lilloa 6: 299. 1941.

Chamaesyce hirta L. var. procumbens (Boiss.) Mold. in Rev. Sudam, Bot. 6: 178.

1940.

ArcENTINA: Buenos Aires: Villa Ortuzar, Parodi 12819; Tucum an:
Villa Lujan, Venturi 167, Trancas, Venturi 4386, Tapia, Rodriguez 526; °& :
Oran, Rodriguez 96, Candelaria, Venturi 3659; Cordo ba: Unquillo, Bruch 5005.
CULTIVATED: Croizat s. n.

This characteristic form is weaker and smaller than the typical plant
and has a fairly thickly arillate seed, the testa of which is dusty-grayish
rather than brick-colored. It is particularly abundant in Argentina, the
collections cited being representative. In some of its most diffuse states (for
instance, Rodriguez 526, Bruch 5005, and Croizat s.n.) this entity is close
to Chamaesyce microcephala (Boiss.) comb. nov. (Euphorbia micro-
cephala Boiss. in DC. Prodr. 15(2): 1262. 1866), which in its turn does
not seem to differ enough from the form called by Wheeler E. Airta var.
destituta, in Contr. Gray Herb. 127: 70. pl. 4, C 1. 1929).

Chamaesyce hirta var. laeticineta var. nov.

Nectariis saepius appendicibus petaloideis albicantibus sat magnis in-
signitis, foliis saepius sub apicem rhombeo-dilatatis

Paracuay: Chaco Paraguayo: Puerto Casado, Rojas 2819.

I have not seen material representing Euphorbia Karwinskyi Boiss.,
which, according to Wheeler, in Contr. Gray Herb. 127: 71. 1939, should
not be far remote from E. kirta var. nocens Wheel. and somewhat suggests
this new variety in the descriptions. In its most characteristic state this

1943 ] CROIZAT, AMERICAN EUPHORBIACEAE 189

variety is easily recognizable by the white petaloid appendages of the
nectaries on the cyathium.
Chamaesyce serpens (H. B. K.) Small, Fl. Southeast. U. S. 709, 1333. 1903.

Euphorbia sag H. B. K. Nov. Gen. & Sp. 2: 41 [folio], 52 sedges 1817; Boiss.

. 15(2): 29. 1862; Wheel. in Contr. Gray Herb. 136: 198. 1941.

PARAGUAY: < 4 aco Paraguayo: Lopez de Filippis, Rojas 8278; Puerto
Casado, Rojas 2161.

The first of these specimens is an absolute match of the typical plant
collected at Cumana. In this plant the stipules definitely tend to be
triangular-truncate, not laciniate-partite. Rojas 2161 is a microphyllous
state and evidently a perennial from a comparatively thick rootstock. This
suggests that the species is annual only where conditions are unfavorable.
Chamaesyce serpens var. montevidensis (Boiss.) comb.

Euphorbia ovalifolia montevidensis Boiss. in DC. Prodr. 15(2) : 43. 1862.

Euphorbia serpens var. fissistipula Thell. in Bull. Herb. Boiss. Il. 7: 755. 1907. Syn.

Nov.

Urucuay: Sha an Casaretto 453 (type number) ; Arechevaleta 5204, Lombardo
222, Legrand 394. ARGENTINA: Buenos Aires: Mar de la Plata, Hicken
642; Loberia, fen (Alboff) s. n.

It is possible that C. serpens and C. ovalifolia cannot be distinguished
with finality as separate species, but it seems clear that the var. monte-
vidensis rather agrees with the former than with the latter on account of
the habit and foliage and the less evolute petaloid appendages.

Lombardo 222 bears the local name ‘‘Yerba Meona.”’ This same name
is given by Larrahaga, Escr. D. A. Larrafiaga, Inst. Geogr. Uruguay 2: 165.
1923, to his E. diuretica, which is described altogether too briefly but is said
to be “pubescens.” Clearly, E. diuretica is not E. serpens or any of its
forms, for these are glabrous. The binomials of Larrafiaga are published
with descriptions so sketchy that, in this group, it proves impossible to place
them without access to authentic material for study and comparison.

EXCLUDED FROM THE EUPHORBIACEAE
Ayenia pusilla L. Syst. ed. 10. 1247. 1759
Tragia eee Hert. in Rev. Sudan: Bot. 3: 166. 1936, nomen; op. cit. 5:34.
7. Syn. Nov.

Iam a to Senor A. Lombardo for data, drawings and notes which
provide definite proof that Herter’s species is not euphorbiaceous but is a
well-known sterculiaceous plant, Ayenia pusilla L. The reduction here
made is based on Chebataroff 6352 in our herbarium, bearing the original
sketches and comments of Sefor Lombardo.

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

190 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

THE FAMILY HIMANTANDRACEAE
I. W. BatLey, CHARLOTTE G. Nast, AND A. C. SMITH
With six plates

THE present paper is the second of a proposed series discussing the inter-
relationships of the families of woody Ranales. We have already briefly
discussed the position of the Himantandraceae as a relative of the Mag-
noliaceae and the Degeneriaceae (1). These three families form a compact
group within the Ranales, being more closely related to each other, on
the basis of important morphological details, than any one of them is to
other families

The most important contributions to our knowledge of the Himan-
tandraceae were made by Diels (3, 4, 5), with whose conclusions (5: 134)
that the closest relative of the group is the Magnoliaceae we are in essential
agreement. Diels has discussed Himantandra in considerable detail, and
therefore we shall emphasize those points which he was unable fully to
observe, and especially those characteristics of Himantandra which we
interpret differently.

The first part of this paper presents a diagnosis of the technical char-
acters of Himantandra, the sole genus of the family, and the two known
species. Only the essential citations to literature are given, as fuller cita-
tions were recently listed (7) and the status of the generic name was dis-
cussed. In the second part of this paper we shall examine the internal
morphological features of the genus, with special reference to points not
made entirely clear by Diels. Specimens cited in this treatment are de-
posited in the herbarium of the Arnold Arboretum. The morphological
and anatomical portions of this study have been prepared by the first two
authors, the taxonomic portion by the third author, while the conclusions
are the result of collaborative discussions.

I

Himantandra F. v. Muell. Pap. Pl. 2: 54. 1890

Galbulimima F. M. Bailey in Queensl. i Age Bot. Bull. 9: 5. 1894.

Trees, the branchlets slender, subterete or faintly angled distally, densely
lepidote; scales covering the young branchlets, petioles, lower surface of
leaf-blades, and external parts of inflorescence (except stamens and stami-
nodes), these scales peltate, eee rae. iarectgue dark at center,
paler toward margin, the stalk very minute, the body composed of 30-56
radiating flattened laterally coalescent haire: stipules none; leaves alter-
nate, simple, pinnate-nerved; petioles slender, rugulose; leaf-blades coria-
ceous or thin-coriaceous, entire and faintly recurved or plane at margin, the
costa prominent beneath, the secondary nerves 7—16 per side (interspersed

1943 ] BAILEY, NAST, & SMITH, HIMANTANDRACEAE 191

with other similar but weaker or obscure laterals), spreading, anastomosing
toward margin (either freely or obscurely so); flowering shoots axillary,
customarily with 1 terminal flower, with 2 (sometimes 3) alternating
bracts, the bracts subcoriaceous, oblong, 1-3 mm. long, occasionally folia-
ceous, with obscure axillary buds, these buds rarely developing into sub-
sidiary flowering axes with two scales and an apical flower; pedicel similar
to the flowering shoot in texture; calyx subcoriaceous, ovoid-conical, obtuse
or umbonate at apex, calyptrate, rupturing along an irregular line near
base and leaving a small undulate or irregularly lobed calycine remnant
attached to the torus, glabrous within, densely and uniformly lepidote
without; corolla similar to calyx in texture, shape, and indument, slightly
smaller than and closely enveloped by the calyx, similarly calyptrate; torus
carnose, flaring to the attachment of the calyx and corolla, thence columnar

2-seriate, castaneous, carnose, sharply reflexed after anthesis; stamens
numerous, several- seriate, closely appressed, similar to outer staminodes i in
texture and shape, rapidly elongating and reflexed after dehiscence of the
perianth, the pollen-sacs 4, paired, extrorse, immersed in the sporophyll-
tissue, linear, obtuse at base and apex, dehiscing ee alee inner

staminodes about 13-20, 1-3-seriate, similar in texture to the outer stami-
nodes, See cae gradually narrowed to an acute ne usually erect
and more r less coherent at base; carpels spirally oe on the conical

texture, sometimes cohering in a gelatinous mass, the locule 1, the ovule 1
(rarely 2, but the second seldom developing), anatropous, attached to the
ventral margin at various levels in different carpels; fruit an ellipsoid or
subglobose syncarp, up to 25 mm. in diameter at maturity, rounded or
obtuse at base and apex, the pericarp coriaceous, 0.5—1. eUnueK. red,
rugulose when dried, fesidets without, the scales long persistent, the carpels
completely coalesced, often imbricate and superposed in 2 or 3 ranks, the
exterior ones appearing shorter than the interior, the dissepiments coria-
ceous, persistent, the endocarp cartilaginous, the seeds solitary Sy possibly
rarely 2), with oily endosperm and small embryo near the hilum

KEY TO THE SPECIES

Leaf-blades usually oblong-elliptic, (6—) 7-15 cm. long, (3—)4—7 cm. broad (2-2.5 times
as long as broad), obtuse to acute at base, rounded to acute at apex; scales on

at least contiguous, only rarely not touching; outer staminodes about 12-23;
stamens 90-130, 6-25 mm. long; inner staminodes 15-20; alee 9-15; New
Guinea (and probably North Moluccas). .........+-+.+0005s A. Belocavecns.
Leaf-blades oblong-lanceolate, 7-11.5 cm. long, 2-4 cm. broad pan 3 times longer

outer staminodes about 7; stamens about 40, 6-9 mm. long; inner staminodes
about.13% carpels/=10; Queensland. «<< sales. sce nae ca eee rene 2. H. baccata.

192 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

. Himantandra Belgraveana (F. v. Muell.) F. v. Muell. Pap. Pl. 2: 54. 1890.
Eupomatia Belgraveana F. v. Mu ie in Austral. Jour. Pharm. 2: 4. 1887, in Bot.

—_

Galbulimima Belgraveana Sprague in Jour. Bot. 60: 138.
Himantandra nitida Bak. f. & Norman in Jour. si ie ae 2. 1923.
Galbulimima nitida Sprague in Jour. Bot. 61: 200.

Tall tree, up to 25 m. or more high; ious straight or subflexuose,
2-4 mm. in diameter toward apices, om brown or castaneous or at lengt th
fusco-cinereous; scales 0.15—0.4 mm, in diameter, crowded, usually com-
pletely obscuring the surface, the aoe of adjacent scales often imbri-
cate; petioles 1-2 mm. in diameter, 8-24 mm. long, deeply or shallowly
canaliculate; leaf-blades coriaceous, brown or dark olivaceous, smooth or
d

sely and minutely papillose, and shining or dull abov e when dried,
castaneous- or fuscous-lepidote beneath, ee rae or obovate- )-
elliptic, (6-)7—15 cm. long, (3—)4—7 cm. broad, obtuse to acute at base,

rounded to acute and sometimes ieee G mucronulate at apex, the
costa impressed or deeply canaliculate above, the secondary nerves 8-16

prominulous or immersed beneath, the veinlets immersed, occasionally form-
ing a faintly prominulous reticulum above and rarely beneath: flowering
shoots 1.5—6 cm. long in senaget of flower or fruit, the vegetative portion
slender, rugulose, 1-2 mm. in diameter, up to 25 mm. long, the pedicel
gradually swollen distally . 3 mm. in diameter, 8-16 mm. long at anthesis,
up to 20 mm. long in fruit; calyx 9-16 mm. long and 8-13 mm. in diameter
at anthesis; outer staminodes oa 12-23, 1- or 2-seriate, oblong-ligulate,
5- mm. long at anthesis and probably ‘often longer, L. 5—2 mm. broad
near base, gradually narrowed to a sharp and often unequally ae or
rostrate apex, sometimes sparsely pelucid-g landular; stamens 90-1
usually 6- or 7-seriate, 6-25 mm. long, 1.5- ? mm. broad near base, often
with numerous and obvious sclereids and obscurely or obviously striate,
the pollen-sacs 0.8—-1.8 mm. long, the lower edge 0.6—3 mm. distant any om
the base of the cea inner staminodes 15-20, 1—3-seriate, 5—7 n
long, 0.5—1.3 mm. broad near base; carpels 9-15, 4— 6 mm, long at ciiese
the ovary 1.5— - mm. long; fruit with seeds (in all our specimens ) solitary,
flattened, _submembranaceous, suborbicular, 3-4 mm. in diameter, appar-
ently ster

eee New Guinea, and probably also some of the islands to the west.
Diels (5: 131) reports that Watbur rg (no. 17770) collected loose flowers which are
probably referable to the species at Sibela, on Batjan, an island ar of Halmahera ;
extension of typical New Guinean elements to this region is frequ

NETHERLANDS New Gutnea: Biak I., Seroei, alt. about 50 m., Neth. Ind.
For. Serv. 30722, 30898; Japen I., Seroei, alt. about 370 m., Ne th. Ind. For.
Serv. 30406 ; 6-15 km. southwest of Bernhard Camp, Idenburg River, alt. 1300-1800 m.,
Brass & Peas 11195 (tree 25 m. high, frequent in primary forest on slopes of a
ridge, the tru 34 cm. diam., the crown not wide-spreading; bark 5 mm. thick, dark

brown, fairly smooth; wood white; fruits red), Brass 12103 (tree to 25 m. high,
abundant in mossy forest flowers white; fruits orange- -brown), Brass & Versteegh
12572 (tree 21*m. high, occasional in primary forest on slopes of a ridge, the trunk

51 cm. diam., the crown fairly small; bark 9 mm. thick, gray, smooth; wood light
brown; flowers yellow; fruits green). NorTHEASTERN NEW rey Sepik River
region, Hauptlager Malu, alt. 50-100 m., Ledermann 10884a (frag.) ; Morobe District:
Yunzaing, alt. 1200-1350 m., Clemens 3586, 3678, 6498, 6503 (large trees, the trunk to

m.; fruit red) ; Ogeramnang, alt. 1750-1800 m., Clemens 4991, 5538; between

1943] BAILEY, NAST, & SMITH, HIMANTANDRACEAE 193

Ogeramnang and Tobou, alt. 1500-1800 m., Clemens 6584a; Matap, alt. 1500-1800 m.,
Clemens 11100, 41200 (tall trees; ‘inflorescence’ brown; flowers russet-green).
British NEw GuINEA: Central Division, Mt. Tafa, alt. 2400 m., Brass 4916 (tall
tree, plentiful in valley forests, with slender trunk and thinly foliaged crown; staminodes
and stamens cream-colored; fruit brown).

In studying the above-cited specimens, we have noted certain differences
which we have tried to correlate with the different geographic areas, think-
ing that more than one nomenclatural division of the genus might be discern-
ible in New Guinea. However, our attempts to divide the New Guinean
population have not succeeded, and we have reached the conclusion that
only one species is represented.

The material from the Morobe District has the upper surface of the
leaves usually dull and conspicuously rugulose-papillose, while the remain-
ing collections have this surface comparatively shining and essentially
smooth. However, there are exceptions to this generalization, and the
texture of the upper surface appears to be subject to individual variation,
possibly being dependent upon the size and distribution of stone-cells. The
degree to which the secondary nerves are immersed is also subject to great
individual variation, although in general the Morobe collections have more
completely immersed nerves. When young, the leaves are infolded with the
two halves of the upper surface closely appressed. Although mature leaves
are always strictly glabrous on the upper surface, these young folded leaves
sometimes bear small many-branched stellate hairs before they open. Such
hairs are often found in material from the Morobe District, but they are
apparently always lacking in the other specimens cited.

On the basis of size and number of floral parts, no important differences
are found among the available collections, the species being very variable
in this respect. The greatest variability is found in the size of the stamens
and outer staminodes, which elongate rapidly after the dehiscence of the
corolla. The longest stamens we have observed are 18 mm. long, but Diels
(4) portrays the stamens of the type collection as about 25 mm. long, and
we have no reason to doubt the accuracy of this observation.

From our study of the available material and the earlier descriptions,
therefore, we are inclined to believe that Himantandra is represented in
New Guinea by only one species.

Himantandra Belgraveana has as its type a specimen collected by Forbes
(no. 759 according to Mueller, no. 795 according to Diels [4] and Baker
[2]), collected in the vicinity of Sogere, British New Guinea. The original
description does not give dimensions, but Diels’ description in 1912 (4) is
adequate. Himantandra nitida is based upon Forbes 828a from the same
region; according to Baker and Norman this differs from the earlier species
“by the shining broader coriaceous leaves and much longer stamens.” A
comparison of the description of H. nitida with our concept of H. Bel-
graveana does not demonstrate differences of any consequence.

2. Himantandra baccata (F. M. Bailey) Diels in Bot. Jahrb. 55: 128. 1917.

Galbulimima baccata F. M. Bailey in Queensl. Dept. Agr. Bot. Bull. 9: 5. 1894.

Tree up to 17 m. high, the branchlets 1.5-3 mm. in diameter toward

194 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

apices, brownish; scales 0.13—0.17 mm. in diameter, comparatively scat-
tered, the margins of adjacent scales not imbricate, rarely contiguous;
petioles 0.7-1 mm. in diameter, 8-20 mm. long, shallowly canaliculate;
leaf-blades thin-coriaceous, dark brown and shining above when dried,
castaneous-lepidote beneath, oblong-lanceolate, 7—-11.5 cm. long, 2—4 cm.
broad, acute to attenuate at base, subacute to short-acuminate at apex, the
costa shallowly impressed or slightly raised above, the secondary nerves
7-10 per side, prominulous above, less conspicuously so beneath, the vein-
lets immersed or faintly uaa ie above; flowering shoots up to 2 cm.
long at anthesis, the vegetative portion angled, 5-13 mm. long, the pedicel
shorter; calyx up to 10 mm. long and slightly less in diameter at anthesis;
outer staminodes about 7, l-seriate, lanceolate, 4-6 mm. long or paged
longer after anthesis, a stamens about 40, several-seriate, 6—9 m

long, the pollen-sac s 1. 5—2 mm. long; inner staminodes about 13, 6-8 mm
ious narrowed at sae aaa 7-10, about 2 mm. long at anthesis: fruit
with seeds “with a loose outer ragged coat; testa smooth, cartilaginous;
albumen copious, oily. Embryo not particularly small near the hilum,
apical with reference to the position of the seed in the berry.” (ex
F. M. Bailey)

DISTRIBUTION: Queensland, Australia.

AUSTRALIA: Queensland: North Queensland, Gadgarra, Peeramon, Atherton,
White 1561. In addition to the preceding specimen, which is the only one we have seen,
the following are cited by other writers, all from Queensland: Eumundi, Arundeil
(tyPE); Boar Pocket and Evelyn, Heberton District, J. F. Bailey; Kin Kin, North
Coast Line, Francis.

Our description is based primarily upon the White collection, which is
in fruit; we have also incorporated the characters and dimensions recorded
by F. M. Bailey and Sprague (see Smith [7] for citations).

INADEQUATELY KNOWN SPECIES
HIMANTANDRA PARVIFOLIA Bak. f. & Norman in Jour. Bot. see ee 2.1923.

Galbulimima parvifolia Sprague in Jour. Bot. 61: 200.

This species, known to us only from the end eer is based on
Forbes 355, from Meroka, British New Guinea. The leaves described seem
closer to those of H. baccata than to those of H. Belgraveana, but they
are even smaller than any described for H. baccata, being similar in pro-
portions. The species is said to differ from H. baccata in its smaller leaves
and flowers, but the dimensions given for the flowers do not indicate this
to be the case. No numbers of floral parts are given.

From the locality, one would expect this to be a depauperate form of H.
Belgraveana, but the leaf-proportions do not suggest this. If it represents
H. baccata, the occurrence of this species in New Guinea will be noteworthy.
Himantandra parvifolia may quite possibly be a good third species of the
genus, but for the time being we are inclined to believe it an extreme varia-
tion of H. Belgraveana, which, as illustrated by the specimens cited above,
seems best interpreted as a very variable species.

II

As indicated above, we are much better acquainted with the New

1943] BAILEY, NAST, & SMITH, HIMANTANDRACEAE 195

Guinean species than with the Australian, and consequently the following
notes are based primarily upon H. Belgraveana, of which we have ample
recently collected material. The Australian species and H. Belgraveana are
very closely related and show only minor differences, and for this reason
we do not doubt that remarks on the morphology of one species apply
equally well to the other.

Stem. In Himantandra baccata and H. Belgraveana, as in Degeneria
and the Magnoliaceae,! the primary vascular cylinder is a dictyostele,
being constituted of discrete bundles that are separated by relatively wide
gaps. Each bundle is capped externally by slender thick-walled fibers and
is subtended internally by vertically elongated, thin-walled parenchyma.
During the earlier stages of the formation of the secondary body, the
external arcs of fibers tend to become united into a continuous ring of
sclerenchyma by the sclerification of the intervening arcs of parenchyma.
The bulk of the pith is composed of large comparatively thin-walled cells,
but nests and transversely oriented plates of sclereids are of common
occurrence, particularly in the nodal parts of the stem. The cortex ‘s
characterized by having numerous spherical secretory cells and more or
less abundant sclereids. Crystalliferous parenchyma occurs in the cortex,
phloem, and pith, usually in close association with the sclerenchyma. Each
small crystal-bearing cell or chamber contains a single rhombohedral crystal
of calcium oxalate that is jacketed by a thick sheath of lignified cellulose.
As noted by Diels (5), the origin of the cork is superficial, probably
hypodermal.

The rays of the first-formed secondary xylem are narrow, with a high
ratio of uniseriate to biseriate, and are markedly heterogeneous. On the con-
trary, in wood from large stems (fig. /0), there is a high ratio of fusiform,
nearly homogeneous triseriate and tetraseriate rays, and uniseriate rays are
much reduced in size and number. The vessels of the first-formed sec-
ondary xylem are smaller, more numerous, and occur in more extensive
radial seriations than they do in the later- formed wood (fig. 9). Further-
more, the vessels of the metaxylem and of the first-formed secondary xylem
commonly exhibit a higher ratio of scalariform to porous perforations and
of scalariform and opposite to alternate lateral pitting than do the vessels of
the later-formed wood, where scalariform perforations and transitional types
of lateral pitting are evanescent or vestigial. It should be noted in this
connection that the discrepancies in Diels’ (5) and McLaughlin’s (6) de-
scriptions of the wood of Himantandra may have been due in part to differ-
ences in the type of material examined by them. ‘The wood fibers of
Himantandra are comparatively thin-walled fiber tracheids, having small
circular bordered pits. The wood parenchyma is of a broad banded apo-
tracheal type (fig. 9). More or less numerous strands of crystal-bearing
cells occur in association with the wood parenchyma.

1Whenever mentioned in the following pages, the family Magnoliaceae is intended 1

the restricted sense of Dandy, Hutchinson, and others, viz. exclusive ot the W ae
Illicium, Schizandraceae, and Tetracentron.

196 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

In herbarium specimens of Himantandra, there is less conspicuous flaring
of the rays in the secondary phloem than in comparable material of
Degeneria and of most Magnoliaceae, and stratified hard and soft bast are
less precociously developed. It is significant, however, that in larger stems
the phloem is distinctly stratified and has flaring rays. Furthermore, the
sieve tubes are of the same structural type as in the Magnoliaceae and
Degeneria. Crystalliferous parenchyma occurs along the surfaces of the
hard bast.

LEAF AND NODAL ANATOMY. In Himantandra Belgraveana and H.
baccata, three traces enter the base of the petiole, leaving three gaps in the
cauline vascular cylinder, i. e. the stems have typically trilacunar nodes.
The three traces divide forming 6-8 vascular bundles that become oriented
into a more or less cylindrical foliar dictyostele (fig. 8). As in Degeneria
and the Magnoliaceae, the vascular strands that branch outward from the
median trace are segregated in opposite sides of the foliar dictyostele. In
other words, one or more of them retain a normal orientation of xylem and
phloem and form part of the abaxial surface of the foliar stele, whereas the
remaining ones develop an inverted orientation of xylem and phloem and
form part of the adaxial surface of the foliar vascular cylinder. Thus, the
vascularization of the petiole and midrib is of a fundamentally different type
than that which occurs in such ranalian plants as Tetracentron, where a
medullated foliar dictyostele is formed by the closure of an adaxially ex-
panding arc of vascular tissue.

The young leaves of Himantandra are adaxially folded, i. e. conduplicate
(fig. 3). They do not unfold until they have attained a considerable size,
not infrequently a length of 8 centimeters or more in the case of the larger-
leaved specimens. The exposed abaxial surfaces of the conduplicate leaves,
from very early stages of their development, are provided with a dense
coating of peltate scales (fig. 3). These scales are persistent on the un-
folded mature leaves (fig. 1), but are smaller and less crowded in H. baccata
than in H. Belgraveana. The adaxial or upper surfaces of mature leaves
of all investigated specimens of Himantandra are glabrous, but the immature
leaves of certain collections of H. Belgraveana, viz. Clemens 3586, 3678,
4991, 5538, 6498, 6503, and 6584a, bears scales or stellate hairs on their
ventral surface during certain stages of their conduplicate development.
The ray cells of these scales or stellate hairs are not firmly coherent and
drop off during subsequent development of the leaf. However, the basal
cells or stalks are persistent and are more or less widely scattered among
the epidermal cells of the upper surface of the mature leaf.

The stomata of both 1. baccata and H. Belgraveana have a very peculiar
and highly characteristic distribution. As shown in fig. 2, they occur in
discrete, nearly circular clusters that subtend each of the peltate scales on
the lower surface of the leaf. Crystal-bearing cells, of the same morpho-
logical type as in the stem, are more or less abundant in the leaf. They
tend to occur characteristically in pairs or small clusters in the lower
epidermis (fig. 7), and in strands along the sclerenchymatous sheaths of the

1943] BAILEY, NAST, & SMITH, HIMANTANDRACEAE 197

veins and veinlets. Numerous clusters or nests of sclereids are scattered
throughout the mesophyll of all the Clemens collections of H. Belgraveana,
but they are absent or of less frequent occurrence in other material ex-
amined by us. Spherical secretory cells are abundantly developed in the
leaf, as in the other organs of the plant.

FLOWERING sHOoOoTS. The solitary bisexual flowers of Himantandra, as
of Degeneria and certain genera of the Magnoliaceae, e. g. Michelia and
Elmerrillia, are borne at the apex of axillary shoots. These flowering
axillary shoots are provided with two (occasionally three) scales which
have more or less rudimentary buds in their axils. Since the scales fre-
quently develop into typical leaves, they may be interpreted as reduced
foliar organs. In exceptional instances, one of the buds develops a sub-
sidiary flowering axis bearing two scales and an apical flower. The
flowers are separated from the upper scale or leaf by an internode of con-
siderable length, which may be designated as the pedicel in contrast to the
essentially vegetative nodes and internodes that subtend it. The pedicel
flares toward the base of the torus, forming a circular flange (figs. 11, 12),
to which the calyptrate calyx is attached. A second internal flange pro-
vides an attachment for the calyptrate corolla. The carpel-bearing, cone-
like apex of the torus (fig. 12) projects from the concave upper surface of
the broadly columnar part of the receptacle, to which the stamens and
staminodes are attached.

The axillary flowering shoot contains a dictyostele of many small bundles
(figs. 15-18) similar to that of a typical vegetative branch, except for a
short distance in the basal part of the pedicel. Here the bundles are
constricted into four large vascular strands (fig. 17). Three traces (figs.
15, 16) enter! the first bract, leaving three gaps in the stele, just as in the
case of the leaves of ordinary vegetative shoots. The three traces of the
second bract (or leaf) initiate their departure at a slightly higher level
and from the opposite half of the stele. Although these traces fluctuate
considerably in their subsequent behavior, two of them (commonly the
median and one lateral) tend to divide, forming two additional traces that
extend upward through the cortex of the pedicel (figs. 15-18). A third
set of three traces departs from the stele just above the level of the first
node (fig. 16). These traces are detached from the same half of the stele
as the three traces of the first scale and extend upward through the cortex
of the second internode, the second node, and the pedicel (figs. 16-18).
Thus, there are five cortical bundles in the base of the pedicel, to which is
soon added a sixth bundle which departs from the same side of the stele
as the median trace of the second scale (figs. 16, 17). As will be shown
subsequently, these six cortical bundles of the pedicel vascularize the outer
calyptra.

Since the foliar appendages of Himantandra have a % phyllotaxy and
are attached to trilacunar nodes, it is evident that the flowering shoot and

1The terminology used is purely descriptive and bears no ontogenetic implications
regarding downward or upward development of procambium, phloem, and xylem.

198 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

pedicel have four distinct sets of three traces, each set attached alternately
to opposite sides of the stele. The lateral traces of the fourth set exhibit
more or less conspicuous fusion to the traces of the second set. ‘Therefore,
the outer calyptra of the flower represents a pair of fused appendages,
either bracts or sepals. It should be emphasized in this connection, how-
ever, that if the homologous appendages of Degeneria are typical sepals,
the outer calyptra of Himantandra should similarly be designated as calyx.
Furthermore, the calyptrate calyx of Drimys is obviously composed of
two fused sepals.

The vascularization of the torus is extremely complex and _ variable.
There is a network of variously oriented bundles which divide, anastomose,
redivide, reanastomose, and shift position throughout the torus. The
number and arrangement of the bundles varies to a certain extent in differ-
ent flowers, indicating a lack of stabilization within the ebay A basic
or average condition is, therefore, illustrated in figs. 18-23. The six
cortical bundles of the pedicel (fig. 18) either bifurcate or break up into
several branches (fig. 19). These branches divide laterally into smaller
strands (fig. 20) and internally into strands which extend upward through
the base of the torus (fig. 21). The lateral strands may divide or anasto-
mose in the base of the outer calyptra, but eventually they enter its free
part as numerous small strands variable in size and number (fig. 21).
The internally directed branches are usually eight in number, exhibiting
considerable diversity in their relations to the branches of the six cortical
bundles of the pedicels. For example, cortical bundle no. 1 in figs. 18-20
bifurcates laterally, one branch of which forms an internal strand. Cortical
bundle no. 2 divides laterally and produces two internally directed strands,
whereas cortical bundle no. 3 gives rise to no internal strands. To these
eight peripheral strands, four additional strands are added from the central
stele (blackened strands of figs. 19-21), making an outer ring of twelve
strands in the part of the torus subtending the inner calyptra (fig. 2
These twelve bundles divide laterally (fig. 22), and some of them may be
joined by a few tracheal elements to an average of five smaller internal
bundles (unstippled strands of figs. 20-22) which depart from the stele at
a higher level than the four bundles referred to above. The five smaller
internal strands, however, are only feebly and temporarily attached to the
vascular system of the inner calyptra and subsequently extend upward
through the torus (fig. 23). The lateral branches of the twelve bundles of
the inner calyptra divide and anastomose laterally, giving rise to an indeti-
nite number of small strands which enter the free part of the inner
calyptra (fig. 23). Simultaneously with the lateral divisions, a few inter-
nally and upwardly directed branches are formed. Thus, at the bases of
both the outer and the inner calyptras, a certain amount of vascular tissue
remains in the torus to become traces or parts of traces for the succeeding
appendages.

The basic pattern in the vascularization of the inner calyptra evidently
consists of four sets of three traces. There are apparently four median

1943 ] BAILEY, NAST, & SMITH, HIMANTANDRACEAE 199

traces that are attached to the central dictyostele and four pairs of
lateral traces that are joined to the cortical system of calycine bundles.
Such an interpretation is strengthened by comparisons with the flowering
axis of Magnolia, Liriodendron, and other Magnoliaceae, where complex
systems of cortical bundles are characteristically present. In these mag-
noliaceous plants, the dorsal trace of the 3-veined carpels is attached to
the central dictyostele, whereas the two lateral traces tie into the cortical
system of vascular bundles. This indicates that the inner calyptra of
Himantandra is composed of four fused appendages, and comparisons with
homologous members of Degeneria demonstrate that it is a corolla. We are
unable to follow Diels (5) in homologizing the calyptras of Himantandra
with the bud-scales of Michelia Figo (Lour.) Spreng., since the scales of
both vegetative buds and flower-buds are clearly of stipular origin in the
Magnoliaceae.

In a former comparison between the floral axes of Degeneriaceae, Mag-
noliaceae, and Himantandraceae (Bailey and Smith, 1), we failed to rec-
ognize fully the vascular complexities of Himantandra. Subsequent de-
tailed investigations of more abundant and adequate material have shown
that this genus resembles the Magnoliaceae rather than Degeneria,
although its vascular complexities appear to be less stereotyped and stable
than in many Magnoliaceae. The flowering shoots and pedicels of
Degeneria do not have complicating systems of cortical bundles.

STAMENS AND STAMINODES. The columnar region of the torus upon
which the stamens and staminodes are borne contains the terminus of the
increasingly complex network of vascular strands. The strands in this
region of the torus comprise (1) upward extensions of the inwardly directed
branches of the corollaceous vascularization (figs. 22, 23), (2) upward
extensions of the bundles which become temporarily attached to this system,
and (3) additional traces detached from the central dictyostele, which loses
its identity as a cylinder by the branching, rebranching, and dispersal of
its constituent bundles. Traces from the strands in the peripheral regions
vascularize the outer staminodes and lower stamens. The upper stamens
and inner staminodes have traces that arise from the dispersed bundles of
the dictyostele. Three traces from three separate strands enter the base
of each fertile microsporophyll and likewise a majority of the sterile ones.
In the case of the broad outermost staminodes and the innermost awl-
shaped ones, the traces are sometimes reduced to two or one.

The stamens of Himantandra are not differentiated into filament, anther,
and connective, and are best described as much elongated, narrowly lanceo-
late sporophylls (fig. 25). This was recognized by Diels, who states (5:

9): “Es ist unangebracht, bei diesen Sporophyllen uberhaupt von Kon-
nektiv oder Anthere zu reden. Denn der Blattcharakter ist kaum
gestort, ” Each microsporophyll bears two pairs of vertically elongated
Spakancia Aha are immersed beneath the abaxial surface of the sporophyll.
Dehiscence is longitudinal and extrorse. In transverse sections cut at the
level of the sporangia (fig. 5), the microsporophylls of Himantandra exhibit

Oo
=}

200 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

close similarities to those of Degeneria, not only as regards their general
topographical features, but also concerning specific peculiarities of their
endothecia. Three veins enter the base of the microsporophylls as in
Degeneria, but there is more extensive branching of the veins in the
sporophylls of the Himantandraceae than in the much shorter micro-
sporophylls of the Degeneriaceae. Although the details of the vasculariza-
tion fluctuate considerably from flower to flower of different collections and
from stamen to stamen of the same flower, the marginal veins rarely extend
beyond the lower third of the sporophylls of Himantandra (fig. 25) and
tend to anastomose with the median vein or its branches just above the
level of the sporangia. The paired sporangia are situated between the
median and marginal veins and thus, as in Degeneria, cannot be regarded
as slightly displaced marginal, or terminal, structures. Furthermore, in
the Himantandraceae, as in the Degeneriaceae, the veins and veinlets are
not directed toward the sporangia and do not establish connections with
the endothecia.

The staminodes of Himantandra resemble the microsporophylls in gen-
eral form and texture (figs. 24, 26), but their median and lateral veins
commonly exhibit less extensive branching. Furthermore, in the outermost
broad sterile sporophylls and the innermost awl-shaped ones, the lateral
veins frequently are much reduced in length or are absent. The sterile
sporophylls resemble the fertile ones in having numerous spherical secretory
cells, more or less abundant nests of sclereids, and in being devoid of
peltate scales, which are such characteristic features of the other organs of
the plant.

The occurrence of staminodes within the fertile microsporophylls has
been noted in Degeneria, which, like Himantandra, differs from the Mag-
noliaceae in this respect (Bailey and Smith, 1). The fact that in Himan-
tandra sterile microsporophylls occur outside, as well as inside, the fertile
ones does not appear to be of great significance, as in all respects except
their sterility these staminodes are similar to the stamens. To interpret
the outer staminodes as petals, Sprague (7) seems to have no justification.
Therefore, we are in agreement with Diels (5: 129) in interpreting these
outer appendages as sterile microsporophylls.

The pollen of Himantandra is provided with a single germinal furrow
and therefore is of the monocolpate type. As in the Magnoliaceae, the form
and the dimensions of the pollen fluctuate during expansion and contrac-
tion of the grains. When fully expanded the pollen tends to be nearly
spherical, with diameters of from 30 to 38 microns. As the tenuous floor
of the furrow invaginates during contraction, the grains become ellipsoidal,
whereas when it evaginates the outlines of the grains become triangular in
certain planes of orientation. The exine is thin and comparatively homo-
geneous, but, as in the case of many of the so-called smooth exines of
Magnoliaceae, it exhibits minute pits.or granulations when examined under
high magnification in lactic acid.

Carpets. The young carpels, like the young leaves, of Himantandra

1943 ] BAILEY, NAST, & SMITH, HIMANTANDRACEAE 201

are adaxially folded, viz. conduplicate, but the margins and the ventral
surfaces of the free parts of the carpels (fig. 6) commonly are less closely
approximated than those of the immature leaves (fig. 3). The adaxially
oriented margins at the base of the carpels are adnate to the cone-shaped
apex of the torus (fig. 7). At this level of the floral axis, there is more or
less lateral concrescence of carpels, a tendency which becomes markedly
intensified during the development of the fruits. The free parts of the
carpels between the level of adnation and the base of the style not infre-
quently remain open at anthesis (fig. 4). The glandular cells of the plume-
like style (figs. 13, 14) extend downward along the free margins and
adjacent ventral surfaces of the carpel to the level of attachment of the
large, much flattened, anatropous ovule. A transverse section of this free
open part of the carpel (fig. 4) resembles a transverse section of the
megasporophyll of Degeneria except that the ovule is attached closer to
the margins of the conduplicate carpels than are the numerous ovules of
Degeneria. In the Degeneriaceae, one is concerned with a seemingly primi-
tive, conduplicate, 3-veined megasporophyll of comparatively unmodified
form, bearing numerous ovules on its adaxial or ventral surface and having
stigmatic structures along its margins and adjacent parts of its free ventral
surfaces. With the reduction in the number of ovules to a single one
(rarely two) in Himantandra, there appears to have been a concomitant
narrowing of the sterilized upper *5 to 4 of the megasporophyll. This
style-like projection beyond the broader base of the sporophyll still
exhibits a conduplicate structure in transverse sections and retains its stig-
matic margins. It should be noted in this connection that the styles of the
Magnoliaceae likewise exhibit a conduplicate structure.

The free dorsal surfaces of the carpels of Himantandra are coated below
the level of the style with numerous dark brown peltate scales (figs. 4, 7,
11, 13, 14). The abaxial parts of the carpels contain more or less numerous
nests of sclereids, such as occur in the tissue of the torus (fig. 7). Spherical
secretory cells are of common occurrence in the carpels, as in other parts
of the flower. The level of attachment and the orientation of the large,
much flattened, anatropous ovule fluctuate considerably from carpel to
carpel and from flower to flower of different collections of H. Belgraveana.
Thus, the ovule may be attached at a higher level where the carpel is open
(fig. 4), or at a lower one where the carpel is adnate to the torus.

The vascular system in the cone-shaped, carpel-bearing, apical part of
the torus is simple in contrast to its complexity in subtending regions. The
vascular strands remaining after departure of traces to fertile and sterile
microsporophylls briefly reassemble at the top of the columnar part of the
torus into a weakly defined cylinder of bundles, most of which are dorsal
bundles of carpels. The dorsal traces of all of the carpels, except the 2 or
3 distal ones, enter the megasporophylls without branching. A variable
number of bundles (4—8) left in the center of the torus divide in various
ways to form two (rarely one) small ventral traces for each of the carpels.
The dorsal bundles of the 2 or 3 uppermost carpels arise also from these

202 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

central strands, but these dorsals first give off ventral traces to lower
carpels of the same orthostichies. The carpel of Himantandra, therefore,
is a modified 3-veined megasporophyll having a well-developed dorsal vein
and two more or less reduced ventral ones. The dorsal vein extends upward
as far as the middle or lower third of the style. It is much enlarged for a
short distance in the region of the junction of the style and ovary. The
ventral veins occasionally terminate in the ovule (especially when one
ventral only is present), but usually they continue short distances in the
margins of the carpels above the attachment of the ovule (fig. 14). How-
ever, the ventral veins rarely if ever extend upward into the style.

Fruit. Soon after anthesis, the styles apparently atrophy and the basal
portions of the carpels become increasingly concrescent. The mature fruit
is a subglobose or somewhat elongated syncarp, upon the surface of which
the outlines of the outermost imbricate carpels can be only indistinctly, if
at all, perceived. The whole exterior surface is more or less persistently
lepidote. The lower carpels appear shorter than those at the apex of the
fruit, the constituent carpels being irregular in shape and apparently often
distorted by mutual pressure. The original conical apical portion of the
torus elongates, carrying the distal carpels upward and thus somewhat
distorting the spiral arrangement. Sometimes the carpels are 2- or
3-ranked and strictly superposed.

All of the fruits available to us are dried, in which condition they are
hard and coriaceous; according to Diels (5: 129) they are fleshy when
fresh and somewhat suggestive of the syncarps of Annona. The inner walls
of the carpels thicken after anthesis and in dried fruits form coriaceous
dissepiments. During development of the fruit the carpels are apparently
under lateral pressure, and the locules eventually appear as mere slits, much
narrower in proportion than they are in the flower (fig. 7). The seeds, in
all specimens examined by us, are solitary, greatly flattened, submem-
branous and suborbicular. In dried material we have not been able to
perceive whether such seeds are fertile, and for the present we can neither
add to nor verify Diels’ statements (5: 130).

RELATIONSHIPS OF THE HIMANTANDRACEAE. In studying the relation-
ships of families, it is essential to weigh evidence from all organs and parts
of the plants. In the past, excessive emphasis has not infrequently been
placed upon similarities between one or two morphological features without
regard to outstanding differences in other parts of the plants, or con-
versely to stress differences in one organ or tissue without allowing for
similarities in other organs or tissues. It should be noted in this connec-
tion that certain of the morphological similarities within the Ranales appear
to be due to retentions of structures that characterized the primitive
ranalian stock, whereas others represent parallel specializations from a
common ancestry. Thus, the retention of vesselless xylem, in itself, does
not provide adequate evidence for combining Trochodencron, Tetracentron,
and the Winteraceae in an independent order, viz. Homoxylées of van
Tieghem (9). Nor are the specialized calyptrate structures of Drimys,

1943 | BAILEY, NAST, & SMITH, HIMANTANDRACEAE 203

Eupomatia, and Himantandra necessarily indicative of close relationship.
Furthermore, certain superficial similarities or differences between specific
organs prove to have been misleading when the ontogeny and the internal
structure of these organs are carefully investigated.

There are numerous similarities between the Degeneriaceae, Magnolia-
ceae, and Himantandraceae. Many of these similarities (e. g. presence of
spherical secretory cells, sclerenchymatous medullary diaphragms, stratified
phloem, wood fibers with small bordered pits, superficial origin of periderm,
monocolpate pollen, stomata with subsidiary cells oriented parallel to the
guard cells, etc.), taken independently, are not necessarily indicative of
close relationships, since they occur in other representatives of the Ranales.
However, the.totality of the similarities does indicate that the three fam-
ilies are more closely related to one another than to other ranalian families.
In fact, the three families form a compact group within the Ranales com-
parable to that composed of the Monimiaceae, Lauraceae, Gomortegaceae,
and Hernandiaceae.

The structure of the stem, including the cortex, pith, and vascular tissues,
is of a basically similar type in the Degeneriaceae, Magnoliaceae, and
Himantandraceae, and differs markedly from that which occurs in other
ranalian families, with the possible exception of the Annonaceae. The wood
of Degeneria is of a comparatively primitive type, whereas that of
Himantandra, with its transitions to porous perforations and alternate
lateral pitting of the vessels, is obviously more highly specialized. The
woods of the numerous representatives of the Magnoliaceae provide a
graded series of transitions between these structural extremes. However,
the range of morphological variability of the stem is no greater than in
single families or even genera of the dicotyledons and therefore, by itself,
does not provide cogent arguments for differentiating the plants into three
families or even for excluding them from close relationship to the
Annonaceae.

The vascularization of the leaf in the Degeneriaceae, Magnoliaceae, and
Himantandraceae is of a characteristic and basically similar type and serves
to differentiate the three families from other ranalian families (including
the Annonaceae) which have secretory cells and monocolpate pollen.
Throughout the Magnoliaceae the vascularization of the foliar organs is
complicated by the presence of stipules and provides a reliable means for
differentiating the vegetative shoots of magnoliaceous plants from those of
Degeneria and Himantandra. ‘The peltate scales, crystalliferous paren-
chyma, and peculiar stomatal arrangements of Himantandra differentiate
its vegetative organs from those of Degeneria and the Magnoliaceae. It
should be admitted, however, that such differences in the vegetative organs,
by themselves, do not afford a thoroughly reliable argument for segregating
the plants into separate families rather than into tribes of a single family.
Only when combined with outstanding differences in the reproductive
organs is there a summation of evidence in favor of separate families.

The flowers of Degeneria and Himantandra, as of Michelia and Elmer-

204 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

rillia, are borne at the apex of axillary shoots. These shoots, as the terminal
flowering ones of Magnolia and Liriodendron, are essentially vegetative,
since they exhibit various stages in the reduction of typical leaves to scales
(Himantandra) or to stipular bud-scales (Magnoliaceae). The flowers of
Degeneria and Himantandra have clearly differentiated pedicels, whereas
those of the Magnoliaceae are sessile on the last vegetative node. The
flowers of Degeneria are provided with distinct sepals and petals, those of
Himantandra with a calyptrate corolla enclosed within a calyptrate calyx,
and those of the Magnoliaceae with tepals or subsimilar sepals and petals.
The tepals of the Magnoliaceae usually have a conspicuously petaloid
texture and internal structure, whereas the homologous parts of Degeneria
and Himantandra are coriaceous and provided with very numerous nests
of sclereids. The immature flowers of the latter genera are not enclosed
within a bud, whereas those of the Magnoliaceae are enveloped within one
or more pairs of stipular bud-scales.

There are no staminodes in the flowers of Magnoliaceae, whereas
Degeneria has numerous inner staminodes and Himantandra both inner
and outer ones. The microsporophylls of Degeneria and Himantandra are
not differentiated into filament, anther, and connective, and their sporangia
are immersed beneath the abaxial surface of the sporophyll. On the con-
trary, the microsporophylls of the Magnoliaceae are typical stamens with
conspicuous protuberant anthers, but they tend to retain the 3-veined type
of vascularization that characterizes both the fertile and sterile sporophylls
of Degeneria and Himantandra.

The floral axis of the Degeneriaceae and Himantandraceae, unlike that
of most Magnoliaceae, is short, and the torus is characterized by having a
conspicuous concavity. In Himantandra the cone-shaped, carpel-bearing
apex of the torus projects beyond this concavity, whereas in Degeneria the
solitary carpel is attached within it. In the latter genus, the carpel is a
3-veined, conduplicate megasporophyll of relatively unmodified form,
bearing numerous ovules on its morphologically adaxial surface. The
attachment of the ovules is remote from the free stigmatic margins of the
sporophyll. The numerous (rarely reduced to two, e. g. Pachylarnax)
carpels of Himantandra and the Magnoliaceae have well-differentiated
styles, which are plume-like in Himantandra and commonly provided with
more or less decurrent stigmatic surfaces in Magnoliaceae. As contrasted

Such morphological divergences are impressive, but a detailed study of
the carpels of Himantandra indicates that they probably represent speciali-
zations of the 3-veined, conduplicate type of carpel encountered in De-
generia. With reduction in the number of ovules to one or a few more or
less basally attached ones, there appears to have been a concomitant narrow-
ing of the upper sterilized part of the conduplicate megasporophyll, forming
styles which retain a conduplicate structure and stigmatic margins. Reduc-
tion of the “decurrent” stigmatic surfaces in certain of the Magnoliaceae

1943 | BAILEY, NAST, & SMITH, HIMANTANDRACEAE 205

leads to the formation of a style with a nearly apical stigmatic surface.
In Himantandra, as in many Magnoliaceae, there is more or less adnation
and concrescence of carpels both preceding and following anthesis. In
Degeneria, part of the maturing seeds are attached by slender much
elongated funicles, suggestive of the suspended seeds of certain Magnolia-
ceae. The pollen of the Degeneriaceae, Magnoliaceae, and Himantandra-
ceae is of a similar monocolpate type, that of Himantandra more closely
resembling the pollen of certain Magnoliaceae than of Degeneria.

Outstanding differences in the carpel, calyx, and corolla render difficult
the inclusion of Degeneria and Himantandra in a single family, in spite of
obvious similarities in the form of the floral axis, the stamens, and the
staminodes. Furthermore, numerous floral differences form a serious
obstacle to including these genera in the Magnoliaceae. Aside from certain
similarities in the pollen, in the vascularization of the stamens, and in the
carpels of Himantandra, there is scant floral evidence for inferring close
relationship to the Magnoliaceae. Such evidence is amply provided, how-
ever, by the vegetative organs. Thus, the summation of evidence from
both vegetative and reproductive organs indicates that in the Degeneria-
ceae, Himantandraceae, and Magnoliaceae we are concerned with three
distinct but closely related families. As will be shown in subsequent papers,
similar summations of evidence indicate that such ranalian plants as the
Winteraceae, Jllictum, the Schizandraceae, and Tetracentron are only
remotely elated to this compact group of chrea families. To include them
within the Magnoliaceae, as some investigators have done, broadens this
family even beyond the limits of a natural sub-order.

PRINCIPAL LITERATURE CITED

. Batrey, I. W., and A. C. Smiru. Degeneriaceae, a new family of flowering plants
from Fiji. Jour. Arnold Arb, 23: 356-365. pl. 1-5. 1942
Baker, E. G. Himantandraceae (of Dr. H. O. Forbes’s New Guinea Plants).
pee Bot. 61: Suppl. 2-3. 1923.
. Diets, L. nie primitive Rana der australischen Flora. Bot. Jahrb. 48: Beibl.
107: 7-13. 1912.
4. ———— Die Anon von Papuasien. Bot. Jahrb. 49: 113-167 (Himantandra,
pp. 164-165. f. 6).
5. ————— Ueber die ee eerie a it Verbreitung und ihre systematische
Stellung. Bot. Jahrb. 55: 134. f. 1. 1917.
. McLaucuHiiy, R. P. ea anatomy fe the woods of the Magnoliales. Trop.
Woods 34: 3-39. 1933
. SmitH, A.C. A comentinniell note on the Himantandraceae. Jour. Arnold Arb.
23: 366-368. 1942
. SpracuE, T. A. Galbulimima baccata F. M. Bailey. Hook. Ic. Pl. 31: pl. 3001.
915.

—_

dR

Ww

a

~

1°,¢)

. TIEGHEM, P. vAN. Sur - Perera du groupe des Homoxylées. Jour. de Bot.
14: 259- As 330- 361.

‘oO

206 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

EXPLANATION OF PLATES

1 plates illustrate Himantandra Belgraveana (F. v. Muell.) F. v. Mue The
hee are een raphed from or drawn from various nest were these being tail
n each ¢
PLaTE I

Fic. 1. Clemens 6584a. Dorsal surface of partially cleared leaf, showing peltate
scales . crystal-bearing cells of the epidermis, x 260. Fic. Brass & Versteegh
11195. ower epidermis of fully cleared leaf, showing circular dae of stomata,
x 260

Prate II

a 3. Clemens 5538. Transverse section of young conduplicate leaf, x 34.
Clemens 11100. Transverse section of flower above the level of adnation of

the “conduplicate carpels, eri free stigmatic margins and the attachment of an
ovule, x 100. 1c. 5. s & Versteegh 11195. Transverse ieee of a fertile
microsporophyll, showing as sporangia and four short arcs of endothecia, x 100.

PriaTE III

Clemens 3678, Transverse section of immature flower, showing open con-
duplicate carpels, = 37. Fic. 7. Brass & Versteegh 11195. Transverse section of
older flower, showing adnation and coalescence of carpels, x 37

PLATE IV

Fic. Ledermann 10884a. Transverse section of basal part of the midrib, show-
ing foliar vascular dictyostele, x 50. Fic. 9. Y. U. 15717. Transverse section of
secondary xylem from a large stem, x 50. a 10. Y.U. 15717. Tangential longi-
tudinal section of the same piece of wood, x

PLATE V
Fic. 11. Brass 12103. Flower with calyptrate calyx, sealekeg corolla, pre ome
and stamens removed, approx. xX 10. Fic. 12. Half o e flower as in
viewed from cut auirkace, showing shape of torus and attached ince and distal Be
approx. X 10. Fic. 13. Clemens 3586. Carpel with short and comparatively smooth
style, approx. X 15. Fic. 14. Carpel from flower of fig. 11, showing plumose style,
position of ovule (micropyle, micr.), ventral bundle, ven. bn., and ovule trace, ov. tr.
approx. x 15.
PiaTE VI
Fics. 15-23. Diagrams of successive segments through flowering shoot pe base of
torus, showing average or basic vascular condition in Himantandra r bract
traces, lo. br. trs.; upper bract traces, up. br. trs.; bud trace, bu. tr.; cotical bundles,
cor. bn.; outer calyptra, o. cal.; inner calyptra, in. cal. Fic. 24. ce 12572. Outer
staminode, approx. 6. Fic. 25. Brass 12572. Stamen, approx. x 6. Fic. 26.
Brass 12572. Inner staminode, approx. x 6.

BroLocicaAL LABORATORIES AND ARNOLD ARBORETUM
HARVARD UNIVERSITY.

Jour. Arno.ip Ars. Vou. XXIV PLaTE I

Oy J G
« ae a
—_ ae as SS fo}

HIMANTANDRA BELGRAVEANA

PLATE I]

Jour. Arnotp Ars. Vot. XXIV

yeas

,

De wet
IRS

:

ty
vas a

oe Re
..

he

Ly

SN

ae
>

Ul

be A
OTL

RS
Sa stg

.

RAVEANA

HiMANTANDRA BELG

PLATE III

XXIV

Jour. Arnoutp Arp. Vou.

aK

eC

SARA Sz,

iy

2
Y

7

SS
ms
re & Y

ie

s
=

oh
ma

o>, i

exe:

6

rf,
<a?

=

ze:

ck

433

2

So. an fh AN

~ or er .:
Risso 4
efoto Sa

e3)

ahs

oatels :
Heke

s

Oye

Seay

isis:

o3

a
a

Sie

i

fost

4
“4

rag

ta,

it

es
Sires

Goa ee

KS

*

LA
an

Fes
ay?

-,*.

20 bat

2,

ian)

ANA

3ELGRAVE

“ANDRA |

HIMANT

PLATE IV

XXIV

Jour. Arnotp Ars. Vou.

SEE
ee

i
i

FTE PORE

—" aS a

— _— = a pe ere ey he hy

— 2 a

——— = =< ——
a
mOZiied 378. —

= a =< we.
Bee eee
Secstee
AI

«
|

3ELGRAVEANA

“ANDRA |

IMANT

{

I

Jour. Arnotp Ars. Vou. XXIV

Prate V

HIMANTANDRA BELGRAVEANA

Jour. Arnotp Ars. Vor. XXIV Pirate VI

H IMANTANDRA BELGRAVEANA

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XII 207

PLANTAE PAPUANAE ARCHBOLDIANAE, XII*
E. D. MERRILL AND L. M. PERRY

In continuation of our work on the plants collected by the Archbold Ex-
peditions in New Guinea, this article consists of a miscellany of notes and
a few new species in the Cruciferae, Violaceae, Passifloraceae and
Apocynaceae.

CRUCIFERAE

Cardamine Linnaeus
Cardamine papuana (Lauterb.) O. E. Schulz, Bot. Jahrb. 55: 271. 1918.
Cardamine africana subsp. borbonica var. papuana Lauterb. in K. Schum. & Lauterb.
Fl. Deutsch. Schutzgeb. Siidsee Nachtr. 271. 190

NETHERLANDS New GuINEA: 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 12164, eee 1939, alt. 1600 m., rain-forest, common on rocks in small streams
(tufts 15-25 cm. high; flowers white) ; 6 km. southwest of Bernhard Camp, Idenburg
River, Brass 12922, February 1939, alt. 1500 m., banks of a forest stream. Bririsi
New Guinea: Mafulu, Brass 5540, Sentember November 1933, alt. 1250 m., wet
rocks of a forest stream, rare (flowers whi te

We have accepted Schulz’s peensiioD of the Papuan material. Whether
it is really more than a localized form of the wide-ranging Cardamine
africana L. is impossible to determine with our limited collections.
Cardamine altigena Schlechter ex Schulz, Bot. Jahrb. 62: 479. 1929.

NETHERLANDS NEW GUINEA: Lake Habbema, Brass 9285, August 1938, alt. 3225 m.,
with Marchantia in a long-grass marsh (flowers white) ; same locality, Brass 9472, on
a mossy bank; 9 km. 7 of Lake Habbema, Brass 10553, October 1938, alt.
2800 m., stony ad of a stream, rare; 7 km. northeast of Wilhelmina-top, Brass &
Myer-Drees 10026, September . alt. 3560 m., wet grassy cliffs, rare.

Siliques 1.5—2 cm. long, 0.2 cm. wide, attenuate into a style 1-2 mm. long;
stigma 0.5 mm. broad; seeds oblong, 1.6 mm. long, 1 mm. wide.

Papuzilla Ridley
slater magica Ridley, Trans. Linn. Soc. II. Bot. 9: 17. ¢. 1, figs. 7-14. 1916;
Schulz, Nat. Pflanzenfam. ed. 2, 17b: 275, 410. 1936

NE RLANDS NEW GUINEA: 2 km. east of Wilhelmina-top, Brass & Mvyer-Drees
9986, 10334, September 1938, alt. 3800 m., alpine grassland, massed on limestone rocks,
also wet open places along small waterfall (flowers greenis ish to violet and white) ;
northern slopes of Mt. Wilhelmina, Brass & Myer-Drees 10053, alt. 3900 m., prostrate
in dense mats or forming loose cushions on screes (flowers and fruits purple).

Since the genus appears to be known only from the type-material of the
species, we have placed these collections on record. They agree pretty well
with the plate of the original material, but the pubescence of the stem
varies from minutely pubescent to glabrous, and the leaves are often 7—9-
incised-dentate.

eae aaa of the Richard Archbold Expeditions. See Jour. Arnold Arb.
24: 34-59.

208 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

VIOLACEAE

Rinorea Aublet
Rinorea cha pete ee Merr. Philip. Jour. Sci. Bot. 12: 286. 1917, Enum.
Philip. Fl. Pl. 3: 192:
Pentaloba i re ante Bull. Soc. Nat. Mosc. 27(2): 341. 1854.
Sotomon Istanps: Florida (N’Gela): northern end of island, Brass 3509,

flowers yellow-green .

This specimen compares reasonably well with the isotype in our herba-
rium. In the latter the anthers are a little larger and, in part of the
inflorescences, the pedicels are scarred at the base, but this may be owing
to the age of the growth. The inflorescences on the year old twigs have
pedicels apparently scarred at the base; the scarred part we take to be
the very short persistent axis of the old inflorescence. In the collection from
the Solomon Islands the inflorescences are all on new growth and do not
show this character.

Rinorea salomonensis (Rechinger) Melchior, Nat. Pflanzenfam. ed. 2, 21: 352.
1925.

gains pies Rechinger, Rep. Sp. Nov. eet Sau 1912, Denkschr. Math.-
Akad. Wiss. Wien 89: 579. t. 6, f. 11 B.

Secu dine Bougainville: Karngu, cn Kajewski 2295, October
1930, alt. 5O m., rain-forest, common pana tree up to 15 m. high; petals
white with cream-colored edges; fruit somewhat triangular, 6 mm. long, 7 mm. diam-

eter, with small appendage 2 mm

In the specimen above cited there are, on the lower surface, minute tufts
of brownish hairs in the axils between the primary veins and the midrib.
The petals, about 5 mm. long, in full grown flowers are about twice as long
as the sepals. The fruit has reticulate valves with 3 mottled brown seeds
about 4 mm. in diameter. In general habit the species closely resembles
Rinorea carolinensis Kaneh., as pictured in Bot. Mag. Tokyo 48: 922. f. 8.
1934, but the anther-appendages are different; in the former they are
simply acute at the apex, not lobed or erose as in Kanehira’s species.

Hybanthus Jacquin
Hybanthus sae apps (Linn.) F. v. Muell. Fragm. Phytogr. Austr. 10: 81.
1877, Pap. Pl. 2: 4. 18
Viola enneasperma Linn. =. Pl. L: 937.
adietg enneaspermum Vent. Jard. aaa “sub 27. 1803-04; Merr. Enum. Philip.
. 3: 106. 1923.
aire New Guinea: Western Division, Penzara, between Morehead and Wassi
Kussa Rivers, Brass 8434, December 1936, occasional on savanna-forest ridges
The species has once before been reported from New Guinea, but modern
workers on the family in this region seem to have overlooked the record.
The species is also reported from the Philippines under /onidium.

Agatea A. Gray
Agatea macrobotrys Lauterb. & K. Schum. Fl. Deutsch. Schutzgeb. Siidsee 453. t. 14.
1900; Melchior, Nat. Pflanzenfam. ed. 2, 21: 360. f. 157 D. 1925, Bot. Jahrb.
62: 373. 192

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XII 209

agian New Gurnea: Bernhard Camp, Idenburg River, Brass 14057, April
1939, alt. 5 , flooded rain- os of river plain (large climber in dense mareinl
growths of ier flowers pur

This specimen agrees with the original description in all details except
that the ovary is sparsely hairy. In the previous records of this species
only the type from Northeastern New Guinea has been cited.

Agatea salomonensis sp. nov.

Frutex scandens; ramulis maturis glabris novellis + pubescentibus:
foliis chartaceis utrinque glabris manifeste reticulatis petiolatis, petiolo
glabrato 2.5-3 cm. longo, stipulis minutis subulatis, lamina late ovata
7-11 cm. longa 5.5—7.5 cm. lata, basi rotundata apice acute breviterque
acuminata margine subintegra vel minute denticulata, venis primariis
utrinsecus 6 vel 7 utrinque manifestis; inflorescentiis axillaribus panicu-
latis + 20 cm. longis, rhachi ramulis pedicellisque + patenti-pubescentibus,
bracteis bracteolisque minutis; sepalis circiter 2 mm. longis oblongis obtusis
ciliatis; petalis ciliolatis: posterioribus ea 4 mm. longis, lateralibus
5 mm. longis 3 mm. latis, labello 8—9 . longo medio valde ligulato con-
stricto 2 mm. lato, parte distali cabdolabuifort) + 6 mm. longa lataque
emarginata extus glabra, intus apice et margine lato involuto excepto villosa,

nectivo in appendicem petaloideam 1.5 mm. latam 2 mm. longam producto;
ovario globoso ad basim piloso; fructibus immaturis.

SoLtomMon IsLtanps: Bougainville: Karngu, Buin, Kajewski 2309 (TYPE),
October 1930, alt. 50 m., rain-forest, common (vine climbing rain-forest trees; petals
white, the inbellum with a touch of purp

Agatea salomonensis is closely “allied to A. macrobotrys Lauterb. &
K. Schum. The flowers are larger, the pubescence of the inflorescence is
more ee the leaves are much more shortly acuminate or scarcely more
than ac

Viola Linnaeus
Viola betonicifolia Sm. subsp. — ae .) W. ig Bot. Jahrb. 54,
Beibl. 120: 166. 1917; Melchior, Bot. Jahrb. 62: 374.

NETHERLANDS NEw Guinea: Balim River, Brass 11653, Decent 1938, alt. 1600 m.,
occasional on grassy deforested slopes (flowers pale, almost w

This species is widespread from the Himalayas pian but it has not
previously been reported from Netherlands New Guinea

Viola lunata Rid]. Trans. Linn. Soc. II. Bot. 9: 17. 1916; Melchior, Bot. Jahrb. 62:
374. 1929

NETHERLANDS NEW GuINEA: Southern slopes of Grand Valley, Brass 9521 (coll.

same locality, Brass 4756, on ‘banks of stream (flowers pale purple) ; Vanapa Walley.

210 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Urunu, Brass 4800, August 1933, rare in swampy hollows on open grassland (flowers.
pale purple, marked with darker lines).

Apparently fairly common in the mountains of New Guinea; recorded
by Melchior from the Arfak Mountains, the Carstensz Range (type-
locality) and Mount Sarawaket.

Viola diffusa Ging. subsp. tenuis (Benth.) W. Becker, Philip. Jour. Sci. 19: 714.
1921; Merr. Enum, Philip. Fl. Pl. 3: 105. 1923
Viola tenuis Benth. in Hook. Lond. Jour. Bot. 1: “482. 1842.
NETHERLANDS NEw GuINEA: Bele River, 18 km. northeast of Lake Habbema, Brass
1469, November 1938, alt. 2200 m., rare herb growing under Jmperata on | formerly
cultivated slopes (flowers pale violet). Southern China, Formosa, Philippi

The only collection of this subspecies we have seen is a ae anne
from Hongkong. The New Guinean material differs in being a plant with
smaller leaves and somewhat more narrowly winged petioles longer than
the leaf-blades; the flower, however, seems to agree reasonably well with
the description of that of the species.

PASSIFLORACEAE

Hollrungia K. Schumann
Hollrungia ara K. Schum. Bot. Jahrb. 9: 212. 1888; K. Schum. & Hollr.
Fl. Kaiser Wilhelms Land 82. 1889; K. Schum. & Lauterb. Fl. Deutsch. Schutzg.
Siidsee Wek 1900; pier Nat. Pflanzentam. 3(6a): 86. fig. 25, E, F. 1893;
Harms, op. cit. ed. 2, 21: 495. fig. 278, E, F. 1925.
NETHERLANDS NEW GUINEA: : km. southwest of Bernhard aii Idenburg River,
Brass 12880, February 1939, alt. 1200 m., rain-forest canopy liane
Apparently anal kom only from the type- nese collected in
Northeastern New Gui
APOCYNACEAE

The study of the Apocynaceae has brought to light some range-extensions
and a few new species. Carruthersia, a Polynesian and Philippine genus, is
reported for the first time from the Solomon Islands, and Bleekeria Hasskarl
has been re-established.
Clitandropsis S. Moore
Clitandropeis noyo-guineensis (Wernh.) S. Moore ex Markgr. Nov. Guin. 14: 279.
1926; Marker. Bot. Jahrb, 61: 174. 192

NETHERLANDS NEW GUINEA: Bele River, 18 km. northeast of Lake Habbema,
Brass 11441. British New Guinea: Matulu, Brass 5253; Tarara, Brass 86706.
NorTHEASTERN NEw GuINnEA: Morobe District, pe 3014, 4523, 5266, 5419, 5443,

6462, 11096. SoLtomon Istanps: Bougainville Koniguru, Buin . Kates
2075, October 1930, alt. 900 m., common in rain- as ie ers cream- colored, strongly
scented; fruit greenish yellow, ovoid, x 2.3 em.); Guadalcanal: Uulolo,

Tutuve Mountain, Kajewski 2572, 2614, pen 1931, alt. 1200 m., rain-forest, common
(vine climbing well into the tops of rain-forest trees; fruit orange, 4 cm. long, with a
sharp apex, 2 cm. diam

For the present we are assigning all these collections to Clitandropsis
novo-guineensis (Wernh.) S. Moore ex Markgr. In the Solomon Islands
material the fruits are a little smaller, but we are not inclined to do more
than give a provisional determination at present. We have not found the

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XII 211

generic lines as drawn between Melodinus, Clitandropsis and Pseudo-
Willughbeia very satisfactory in determining our rather scanty material.

Alstonia R. Brown

Alstonia mperpadiitheny Wall. List no. 1648. 1829, nomen nudum; A. DC. Prodr, 8:
409. 1844; vela

So anges Pe ps: Bougainville: Kugumaru, Buin, Kajewski 1834, June
1930, alt. ne rain-forest, common (medium-sized tree up to 20 m. high; follicles
up to 50 cm. oe

This ae . seemingly new for the Solomon Islands.

Alstonia rey F. v. Muell. Pap. Pl. 5: 91. 1877; Rehder in C. T. White, Jour.
Arn, Arb, 10: 260. 1929

SOLOMON ISLANDS: Cu Bahan Berande, Kajewski 2449, January 1931,
sea-level, rain- oe common (large tree up to 25 cm. high with very ae buttresses ;
follicles 28 cm. long, 5 mm. diameter; timber used in ie houses); San Cris-
to bad: ~ Kira, Kira: Brass 3014, October 1932, coastal rain-forests, common (large
tree with grey scaly bark; leaves thin, the midrib almost white).

The species was described from New Guinea and has since been reported
from Queensland.

Alstonia Reineckeana Lauterb. Bot. Jahrb. 41: 233. 1908; Stapf in Setchell, Amer.
moa 58, pl. 12, fig. A. 1924; Christophersen, B. P. Bish. Mus. Bull. 128: 177.

SoLtomon IsLtanps: Ysabel: Tiratona, Brass 3404, December 1932, alt. 600 m.,
rain-forest (small tree with shining leaves and cream-colored flowers); Guadal-
canal: Uulolo, Tutuve Mountain, Kajewski 2575, April 1931, common in rain-
forest (small tree up to 10 m. high, with showy w bite faintly scented flowers).

These flowering collections are a reasonably good match for Christopher-
sen 1903 from Samoa (of which we have only a fruiting specimen) and also
for Setchell’s plate of this Samoan species. The leaves are larger than

sen’s collection. They differ from the New Caledonian material of Alstonia
plumosa Labill. in the chartaceous texture of the leaves, the slightly more
remote lateral veins, and the tendency to larger flowers. Some Kajewski
collections from the New Hebrides previously recorded as Alstonia villosa
f. calvescens Markgr. are certainly of this alliance and probably conspecific.

Alyxia R. Brown

Alyxia floribunda Markgr. Bot. Jahrb. 61: 184. 1927.
NETHERLANDS NEw GuINeEA: Bele River, 18 km. northeast of Lake Habbema, Brass
1439, November 1938, alt. 2200 m., occasional in forest undergrowth on steep lime-
“a slopes (large cocamibline shrub; panicles white).

This is an entirely glabrous fruiting specimen which seems most like the
above-named species. It differs from the isotype in that the branchlets are
less definitely angled and the leaves are 3-8 mm. petiolate.

Alyxia pugio Markgr. Bot. Jahrb. 61: 182. 1927.

NETHERLANDS NEW GuINEA: 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 12060, aan 1939, alt. 1800 m., mossy forest seral growths (slender scrambling
shrub; flowers white).

212 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Except for the broader and shorter leaves (4-6 1-3 cm.), this collec-
tion suits the description of Alyxia pugio Markgr. from the mossy forest of
the Sepik Territory.

Alyxia subalpina Markgr. Bot. Jahrb. 61: 183. 1927.
ERLANDS New GuINEA: 18 km. southwest of Bernhard Camp, Idenburg River,

Camp, Idenburg River, Brass 12361, January sa alt. 1700 m., epiphytic on a tall
tree in mossy forest (shortly alas slender shru

The first two cited collections vary from “ original description in having
slightly narrower leaves, longer peduncles (5-7 mm.) and pedicels (4—5
mm.), shorter corolla-tube (2.5 mm.), and shorter stigmatic appendages.
These differences, however, are so small that they cannot be considered
specific without actual comparison with the type-specimen. Brass 12361
differs in being more freely growing, with larger leaves and longer inter-
nodes, whereas the other two are compact with short branches. Possibly
two species are represented, but, lacking flowers in the last number, we
suspect that the contrast represents different phases of growth or age.
Alyxia fragrans sp. nov

Frutex scandens; amills novellis puberulis; foliis quaternatim verti-
cillatis glabris coriaceis anguste obovato-ellipticis, basi cuneatis breviter
acutiusculis vel brevissime et obtuse acuminatis, 5-9 cm. longis 2.5—4.5 cm.
latis, nervis lateralibus utrinque manifestis 2— 3 mm. inter se distantibus
rectis late patentibus; petiolo 5-9 mm. longo; inflorescentiis axillaribus

longo puberulo, ramulis 0.5-2.5 (in fructu 3) cm. longis pubescentibus
angulatis; bracteis lanceolato-ovatis acutis 3 mm. longis interdum carinatis;
lobis calycis ovatis 3.5 mm. longis subcarinatis extus pubescentibus intus
glabris; corolla albida, tubo 4 mm. longo infra stamina 1 mm. et inter
antheras piloso, lobis oblongis 1.5 mm. longis; staminibus in medio tubo
insertis, filamentis circiter 0.4 mm. longis glabris, antheris lineari-oblongis
stylo 1 mm. longo; stigmate ellipsoideo apice
setoso, ovario dense piloso; fructu aurantiaco; mericarpiorum articulis
1 vel 2, ellipsoideis, + 16 mm. longis et 12 mm. crassis, stipite circiter
7mm

Gecusce New Guinea: Bele River, 18 km. northeast of Lake Habbema, Brass
11577 (TYPE), 11300, November 1938, alt. 2350 m., scrambling in a forest opening and
frequent in n undergrowth under oaks Cakes shrub; leaves stiff, convex; inflo-
rescence white; ek fragrant; fruit orange-colored).

This species is closely allied to Alyxia scabrida Markgr. but is i
distinguished by the considerably larger and fewer flowers of the
florescence.

acne —— Marker. Nov. Guin. Bot. 14: 280. 1926, Bot. Jahrb. 61: 183. 1927,

eae = GuINEA: Mt. Tafa, Brass 5011, May-September 1933, alt. 2400
common throughout forests (slender high climbing liane; corolla-tube pale brown, the
lobes cream-colored; fruits green)

The collection differs from Brass 12679 (from Netherlands New Guinea),

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XII 213

which we have accepted as typical, chiefly in its lack of pubescence; also the
leaves are less revolute on drying. Lacking further material and the privi-
lege of examining types, the determination can only be accepted pro-
visionally.
Ochrosia Jussieu
Ochrosia glomerata (BI.) F. v. Muell. Fragm. Phytogr. 7: 130. 1871; Markgr. Bot.
Jahrb. 61: 191. 1927.

SOLOMON IsLanps: Ysabel: Tatamba, Brass 3442, January 1933, alt. 50 m.,, hill-
forests, common (tall tree with brown fissured bark, yellow when cut; wood hard,
yellow; leaves smooth and shining; flowers white).

This collection is a reasonably good match for one so determined by
Valeton in 1912, cultivated at the Botanic Garden of Buitenzorg. Accord-
ing to Markgraf the species is known from the Philippines and New
Guinea. ;

Bleekeria Hasskarl

Although, for the most part, we have accepted generic names in current
usage, on account of the difficulty which we encountered in determining a
fruiting specimen from the Solomon Islands we have looked into the nomen-
clatural problems of Ochrosia Juss. sensu lato. Markgraf, Bot. Jahrb. OL:
192-194. 1927, in a reasonably adequate discussion of the generic limits
of the genus, decided that the two sections of Ochrosia Juss. as accepted
by Valeton (as subgenera) and by F. v. Mueller (as sections) possessed
generic value. Retaining the epithet Ochrosta Juss. for one group, he pro-
posed the new name Excavatia Markgr. for the other, the sectional name
Lactaria not being acceptable as a genus on account of having been coined
for the true Ochrosia Juss.; further, Markgraf indicated that all the names
given as synonyms of Ochrosia Juss. were to be considered in the same way.
In so doing, he must have overlooked the little mentioned but greatly
detailed description of Hasskarl’s genus Bleekeria, based according to
Hasskarl himself on the species Bleekeria calocarpa ‘“‘(e Bleekeria kalokarpa
sumptus).” Valeton, commenting on Ochrosia calocarpa Miq. (a transfer
from Hasskarl’s species), made this rather significant remark on the generic
limits of Bleekeria: “introduisant toutefois dans la description du genre
quelques-unes des propriétés qui n’appartiennent qu’a cette espece.” That
is to say the type of the genus Bleekeria Hassk. has to be Bleekeria calo-
carpa Hasskarl, and the secondary application of this generic name to
Rumphius’ species is incidental and has nothing whatever to do with the
typification of the genus. In addition, the fact that, four years later,
Hasskarl considered the same two species under the genus name Lactaria
is irrevelant, according to the present International Rules of Nomenclature.
In 1923 Koidzumi, Bot. Mag. Tokyo 37: 48-52, apparently divided
Ochrosia Juss. into two parts, using the generic name Lactaria Rumph. for
the true Ochrosia Juss. and Bleekeria Hassk. for the Section Lactaria
F. v. Muell. (subg. Valeton). The name Lactaria Rumph. cannot super-
sede Ochrosia Juss., for it is a later designation first described by Rafinesque
(1838) under the binomial system; but Bleekeria Hasskarl, Retz. 1:

214 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

38-40. 1855, must replace Excavatia Markgraf, Bot. Jahrb. 61: 194. 1927.
Bleekeria solomonensis sp. nov.

Arbor usque 15 m. alta glabra; ramulis teretibus novellis —
foliis oppositis vel ternatis ee oblongis utrinque angustatis, 9-16 ¢c
longis 3.5—5.5 cm. latis, basi cuneatis apice subabrupte acuminatis, acumine
5-10 mm. longo acutiusculo, nervis lateralibus crebris subtransversis

utrinque manifestis nervo marginali conjunctis; pet tiolo 1-1.5 cm. longo;
inflorescentiis — in axillis verticilli summi foliorum; pedunc ulo com-

+ 1.5 cm. longo compresso vel subalato; alabastro tantum viso
bracteato; lobis oie 5 lanceolato-ovatis, oe obtusiusculis, intus non
glandulosis, 2 mm. longis 1 mm. latis; tu bo corollae 6 mm. longo, lobis
7 mm. longis invicem sese dextrorsum hai us; antheris vix 1.5 mm.
longis lanceolatis, filamentis 0.5 mm. longis infra faucem insertis; stylo

mm. longo; stigmate turbinato-conico; fructu apocarpo, spiral es

2 ovoideis obtusis compressis in sicco versus apicem alatis, -— 3.5 cm. longis
et 2.5 cm. latis et 1 cm. crassis, obscure laxe reticulatis sub (eke striate
granulosis; excavationibus mesocarpii 2.2 cm. longis 0.5 cm. latis; semine

unico elliptico plano 1.7 cm. longo 0.8 cm. lato

Sotomon Istanps: Bougainville: Kugumaru, Buin, Kajewski 1859 (TYPE),
June 1930, alt. 150 m., rain-forest, common (small tree up to 15 m. high; fruits boat-
shaped, borne in pedis)

This species appears to be most like Bleekeria mariannensis (DC.)
Koidz., but the latter has a narrower oblong fruit with more pointed apex
and lateral wings. In B. solomonensis the dried ine is winged at the apex
and the wing extends down on both sides about 24 of the length of the fruit.
Bleekeria minima (Markgr.) comb. n

Excavatia minima Markgr. in Merr. & on Jour. Arnold Arb. 21: 199. 1940.

Type-collection: Brass 8512.

Miecrechites Miquel
Micrechites Archboldiana sp. nov.

Frutex scandens, ramulis ultimis et bracteis inflorescentiae puberulis
exceptis glaber; foliis coriaceis 9-17 cm. longis 3—8 cm. latis ellipticis, basi
late cuneatis vel obtusis apice acuminatis, acumine 1—1.5 cm. longo, nervis
lateralibus utrinsecus 7-10 oblique adscendentibus, reticulo inconspicuo;
petiolo 1-2 cm. longo; inflorescentiis axillaribus et terminalibus paniculatis,
8-15 cm. longis 6-10 cm. latis, 4- vel 5-dichotomis; ramulis ultimis brevissi-
mis, bracteatis, bracteis rotundatis circiter 1 mm. magnis; lobis calycis
rotundatis 1 mm. longis ciliolatis interdum puberulis; corolla hypocrateri-
formi, tubo inflato sub faucem paullo angustato, 3.5 mm. longo 1.5 mm.
lato, lobis sinistrorsum obliquis, parte inferiore oblongis dextrorsum tegenti-
bus, parte superiore linearibus undulatis 3.5 mm. longis, fauce dense barba-
tis; antheris 1.5 mm. longis, inclusis, filamentis 1.5 mm. supra basim tubi
affixis usque ad basim corollae decurrentibus, glabris; stigmate conico basi
annulo angusto cincto; ovario dense setuloso 1—1.5 mm. longo, disco quam
ovario paullo breviore 5-partito; mericarpiis <n cylindricis horizontali-
ter patentibus immaturis 9 cm. longis et 6 mm. crassis; seminibus lineari-
oblongis 2 cm. longis 2.5 mm. latis erostratis apice coma brunnea 2—2.5 cm.
longa coronatis.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XII 215

NETHERLANDS NEW GuIinEA: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13071, 13491 (type), March 1939, alt. 850 m., rain-forest, common (canopy liane;
corolla yellow with red lobes; fruit immature).

The general habit of this species is like that of Markgraf’s Papuechites,
but the seeds are unbeaked, with a terminal coma, whereas those of
Papuechites are narrowed into an elongated somewhat filiform beak covered
with long spreading hairs.

Carruthersia Seemann
Carruthersia Brassii sp. nov.

Frutex scandens glaber vel consperse pilosus, ramulis teretibus; foliis
oppositis ovato-ellipticis basi retusis vel leviter cordatis apice longiuscule
acuminatis, 5-14 cm. longis 3-8 cm. latis, utrinque glabris (vel consperse
pilosis ) brunneis subdiscoloribus, nervis lateralibus utrinsecus 9-11 utrinque
perspicuis patenti-adscendentibus marginem versus arcuatim conjunctis,
reticulo conferto subtus in sicco atro-brunneo; petiolo 1—2.5 cm. longo;
inflorescentiis axillaribus terminalibusque 6-10 cm. longis, paniculatis.
lobis calycis et bracteis ciliatis exceptis glabris; lobis calycis 1.5 mm. longis
ovatis obtusiusculis intus pluriglandulosis; corolla Aepoer ten iors fauce
pubescente, tubo 9 mm. longo ad antheras paullum inflato extus bro
intus pubescente, lobis eae oe 7-8 mm. longis 4 mm. latis,
obliquis; staminibus circiter 2 mm. supra basim tubi insertis, filamentis vix
1 mm. longis pubescentibus, sais circiter 2 mn 1. longis apiculatis liberis:
disci squamulis tantum 2 oppositis inter pe ovario apocarpo; stylo
1 mm. longo, stigmate anguste conoideo.

Sotomon Istanps: Guadalcanal: Sorvorhio basin, Kajewski 2702, January
1932, alt. 180 m., rain-forest, common (vine with white flowers); San Cristobal:
Huro River, Brass 2609 (type), August 1932, lowland rain-forests (scandent; sap
milky; leaves coriaceous, pale below flowers white).

The genus is already known to occur in the Philippines and Polynesia.

This species from the Solomon Islands is closely allied to C. Macgregori
clan: but the inflorescence of the latter is constantly minutely pubescent

and the flower is a little smaller than in the new species.

Wrightia R. Brown
Wrightia laevis Hook. f. Fl. Brit. Ind. 3: 654. 1882; Markgr. Bot. Jahrb. 61: 212.
1927.

NETHERLANDS New Guinea: Bernhard Camp, Idenburg River, Brass 13893, an

white). British New Guinea: Mt. Tafa, Brass 5570, May-September 1933, alt.
.. rain-forest on low ridges, common (large tree; trunk corrugated; bark pale
brown, “slightly flaky ; foliage pale green; flowers profuse, pale yellow, perfumed).
Markgraf reported this species from the Bismarck Archipelago, but we
have not found any other record of it from Papuasia, although Markgraf
determined the above-cited collection from Mt. Tafa

Parsonsia R. Brown

Parsonsia curvisepala K. Schum. Bot. Jahrb. 9: 215. 1888; Markgr. Bot. Jahrb.
61: 219. 1927.

SOLOMON Istanps: San Cristobal: Waimamura, Brass 2653, August 1932,

216 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

lowland rain-forest (one plant seen; large climber; leaves thin, very much paler below,
the margins wrinkled; flowers yellow; follicles striate, green; coma on seeds pale
brown; sap colorless, slightly viscid)

This specimen differs from those collected in New Guinea chiefly in the
larger leaves (6-8 *& 2.5—4.5 cm.).

Parsonsia bananas Hook. & Arn. Bot. Beechey’s Voy. 197. 1836; Merr. Brittonia
1: 236. 1933; Kaneh. & Hatus. Bot. Mag. Tokyo 53: 939.
vance pte Wall. List no. 1631. 1829, nomen pas G. Don, Gen. Syst.
4:8 ; Marker. Bot. Jahrb. 61: 217. 1927.

SotomMon Istanps: Bougainville: Karngu, Buin, Kajewski 2230, October
1930, sea-level, rain-forest, common (vine with green flowers; fruits 19.5 cm. long,
1 cm. diam.); Ysabel: Meringe, Brass 3161, November 1932, twining on fore-
shore trees, common; San Cristobal: Kira Kira, Brass 2771, August 1932,
large climber on beach trees, common.

We have not found this widespread species previously recorded from the
Solomon Islands.

Parsonsia lata Markgr. Bot. Jahrb, 61: 221, 1927.

NETHERLANDS NEw GUINEA: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13267, March 1939, alt. 850 m., abundant in rain-forest of river plains (large
canopy liane; flowers yellow). Sotomon Istanps: Guadalcanal: Berande,
Kajewski 2434, January 1931, sea-level, common in rain-forest (vine with cream-
colored flowers; follicles 12 cm. long, at widest part 3.6 cm. wide, tapering slightly,
green, covered with fine brown hairs).

In the Solomon Islands collection the leaves are a little broader in propor-
tion to their length and tend to be less pubescent above than in the material
from New Guinea. The species has been collected from Northeastern New
Guinea.

Parsonsia mollissima pape ) eal Bot. Jahrb. 61: 220. 1927.

British New Gui Palmer River, 2 miles below junction Black River, Brass
7089 (det. Marker), Tate on ak 100 m., climbing in forest substage (upper leaf
surface gloss

The ae was collected in Netherlands New Guinea; the species prob-

ably occurs also in Northeastern New Guinea, cf. Mar ker. A
Parsonsia flavescens sp. nov
utex scandens; felis oppositis subcoriaceis = Puy ovato-
ellipticis magaitudine ludentibus, 6-7 cm. longis 1.5—2.3 . latis vel
7-9 cm. longis 4.5—5 cm. latis, basi obtusis vel rotundatis, eee acutis vel
obtuse acuminatis mucronatis, margine + recurvatis, utrinque glabris
subtus minute granulatis, nervis lateralibus + 9 oblique patentibus utrinque

scente; inflorescentiis axillaribus plurifloris 3—4.5 cm. longis puberulis;
pedicello + ongo; calycis lobis ovatis acutiusculis 2 mm. longis
intus pauciglandulosis; corolla flava; tubo 4.5 mm. longo extus glabro intus

sub antheris minute piloso, lobis 3 mm. lo oe 1.5 mm. latis oblongo-
lanceolatis; filamentis circiter 2 mm. longis, in medio tubo insertis genicu-
latis pubescentibus, antheris 4 mm. longis; stigmate 1.4 mm. longo obtuse
conico in basi annulato; ovario glabro 0.5 mm. longo disco glabro aequilongo
cincto.

NETHERLANDS NEw Guinea: Bele River, 18 km. northeast of Lake Habbema, Brass

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XII ole

11573, November 1938, alt. 2350 m., on open face of cliff (small twiner; sap not milky ;
flowers yellow) ; Balim River, Brass 11647 (type), December 1938, alt. 1600 m., climb-
ing in sparse second growths on deforested slopes (sap not milky; flowers yellow).

Among the New Guinean species this seems to approach Parsonsia diverst-
folia (Warb.) Markgr. most closely; the plant, however, is not hispidulous,
the leaves are coriaceous, the flower is twice as large as the dimensions
given for Warburg’s species, and the disk is entire.

Parsonsia rubra Kaneh. & Hatus. ms.

Scandens; foliis coriaceis glabris lanceolato-oblongis, 6-12 cm. longis
1.5-4.5 cm. latis, basi rotundatis apice acuminatis, nervis lateralibus
utrinsecus + 6 supra impressis subtus prominulis; petiolo 1.5—2 cm. longo;
inflorescentiis terminalibus usque 11 cm. longis; ramulis puberulis; corolla
rubra extus glabra.

NETHERLANDS New Guinea: Hollandia, Brass 8991, July 1938, alt. 100 m., rain-
forest ees canopy liane; flowers red).

We have been able to match this collection only with Kanehira & Hatu-
sima 12220 from Netherlands New Guinea. We know that the description
of this species was already in manuscript before Pearl Harbor and we
anticipate that it is now published, although unavailable to us at present.

owever, not being able to cite the place of publication, and to protect
our use of the name, we have appended a brief Latin description.

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

218 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

NEW AND NOTEWORTHY POLYPETALAE FROM
BRITISH GUIANA

N. Y. SANDWITH

Tue following descriptions, new combinations, and notes are based on
material of Polypetalae collection recently in British Guiana, the speci-
mens, unless otherwise stated, being deposited in the herbarium of the
Royal Botanic Gardens, Kew.

DILLENIACEAE

Doliocarpus (Subgen, Calinea, Sect. Eudoliocarpus) savannarum sp. nov.

Ob habitum eae: folia, flores subsessiles, ovarium dense pilosum
valde distincta; a D. virgato ‘Sagot e descriptione forsan affini foliorum
paginae inferioris ne debiliter piloso nec tomentoso-puberulo, nervis
lateralibus semper paucioribus statim distinguitur.

Frutex humilis, erectus vel procumbens, haud scandens; sosaper subteretes,
3-6 mm. diametro, apicem versus foliiferi ac ibi adpresse pilosi, novelli
fere subsericei. Stipulae lanceolatae vel lanceolato- iss seh obtusae, 1 cm.
longae, basi ad 5 mm. latae, adpresse pilosae. Folia elliptica vel elliptico-
oblonga vel oblonga, apice breviter late cuspidato-acuminata, basi cuneata
vel maxima obtusa subrotundata, 6-19 cm. longa, 2.5—9.2 cm. lata, valde
coriacea, siccitate brunnea vel olivaceo- shee subintegra sed dimidio
Superiore ob nervos laterales excurrentes distincte denticulata, subtus plus
minusve nitidula, supra glabra, subtus secus costam nervosque laterales pilis
debilibus satis sparse sed distincte induta ceterum glabra, costa nervisque
primariis utroque latere 9-11 sursum arcuatis supra canaliculato-i -Impressis
subtus prominentibus, venis subhorizontaliter abeuntibus cum venulis intri-
cate reticulatis omnibus supra vix prominulis sed obviis subtus prominenti-
bus; petiolus adpresse pilosus vel glabrescens, supra canaliculatus subalatus,
1-3 cm. longus, ad 3 mm. crassus. Flores in fasciculos plerumque sub
foliis dispositi, fere sessiles, sean pubescentibus vix ad 1.5 mm. longis.
Sepala orbiculari-ovata, apice rotundata, concava, 4.5—5 mm. longa, 3.5—4.2
mm. rai — pubescentia, nek glabra. Petala obovata, ad 6 mm. longa,
ad 3.2 ata. Stamina filamentis ad 5 mm. longis apice in connectivum
ae en caehn loculi 0.5 mm. longi. Ovarium late ovoideo-sub-
globosum, dense pilosum; stylus q: 5 mm. longus, page pilosus, stig-
mate peltato. Fructus kermesinus, globosus, 8-10 mm. diametro, satis
sparse pubescens vel glabrescens, mau eae ce RE coronatus.

British GuIANA: Kaieteur Savannah, c. 400 m., fl. Sept.-Oct. 1881, Jenman 1038
(typus) ; ibid., fr. Sept. 5, 1937, Sandwith 1377. Noted by Jenman as 3-5 ft. high,
with erect ee ath. ii by Sandwith as a trailing shrub with crimson meted growing
on sandy bushy groun

Doliocarpus eta Sagot, which was based on imperfect material col-
lected in French Guiana by Perrottet and Mélinon, was described by its
author as having inter alia: ‘rami lignosi, validi, recti, virgati. Folia .
superne glabrescentia, inferne breviter tomentoso- -puberula. Nervi lat-

1943 ] SANDWITH, POLYPETALAE FROM BRITISH GUIANA 219

erales, recti, numerosi.’”’ The specimens were said to be easily recognizable
on account of the peculiar facies. M. Raymond Benoist, in a review of
the Dilleniaceae of French Guiana in Bull. Soc. Bot. Fr. IV. 13: 400-401.
1913, slightly amplified the description of D. virgatus, to which he referred
an additional flowering collection from Para, remarking that the latter bore
fully expanded flowers, whereas the collections from French Guiana were
in young fruit, so that it was impossible to be certain whether the stamens
were folded back or erect in the bud-stage. On this account he still treated
D. virgatus as an espéce a caractéres insuffisamment connus, and did not
place it in either Section Othlis or Section Calinea. The important points
to be noted in Benoist’s amplified description are the pubescent lower
surface of the leaves and the 12-17 pairs of main lateral nerves, characters
which sufficiently distinguish D. virgatus from the plant of the Kaieteur
Savanna
More ere mention has been made of D. virgatus as an ally, in
Section Eudoliocarpus of Subgen. Calinea, of D. paraénsis Sleumer in Rep.
Sp. Nov. 39: 45-46. 1935. The latter species evidently differs from
D. savannarum in the short (6-10 mm.) petiole, the quite glabrous leaves
with 7 pairs of main lateral nerves and the reticulation obscure on the lower
surface, the glabrous oblong sepals, and the much larger (about 1 cm. X
6-8 mm.) petals.
FLACOURTIACEAE

Ryania pyrifera (L. C. Rich.) Uitt. & Sleumer var. subuliflora var. nov.

typica sepalis angustioribus lineari-lanceolatis plerumque
longioribus sursum saepius sensim longe attenuatis 2—4 cm. lon ngis 3-8 mm.
latis differt. Folia subtus secus costam nervosque primarios stellato-
pubescentia, ceterum ee vel glabra

British GUIANA: Essequibo River, in alist forest, Labbakbra Creek, Tiger
Creek, August 26, 1937, peer 1211 (typus); Demerara River, May 1889, Jenman

853; Mazaruni- Kuribrone Divide, in wallaba forest, Forest Dept. 893; Bartica-
Potaro road, 83rd milepost, in clump wallaba bush, June 1933, Tutin 216 (Herb. Mus.
Brit. and Kew)

A small tree of the lowest storey of the forests of the white sand areas,
especially in wallaba forest, up to 20 ft. high, less than 1 inch in diameter;
sepals pale greenish white; filaments white, pinkish or red at the base.

This has been written up and distributed as a new species under the name
adopted here for the lower rank; the writer is convinced that it does not
deserve a higher status on account of the variability in the shape and
dimensions of the sepals which can be seen on a really extensive gathering
from a single tree. In Trinidad, where typical R. pyrifera is plentiful,
specimens have been collected (Fendler 203, Broadway 9358) on which
some of the sepals match some of those on the type collection of var.
subuliflora.
ee hea var. ble (Migq.) Sleumer ex Sleumer & Uitt. in Pulle, FI.

me 3:

ee ne iy a densely stellate-pubescent or tomentose all over
the lower surface, is known in British Guiana from the following collections:
Berbice River, Forest Dept. 774; Demerara River, Jenman 3888, Persaud

220 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

174, Forest Dept. 774, 2452; Pomeroon District, Tapakooma Creek, Jen-

man 6616. This, from the evidence of field notes, is an undergrowth tree

of brown sand areas, while the color of the flower is similar to that of var.

subuliflora.

nn bg Sane Eichl. in Monte Fl. Bras. 13(1): 491. 1871; Sleumer & Uitt. in
e, Fl. Suriname 3: 28

i. species, which a ener prominulous ultimate veinlets on the
lower surface of the leaves and, above all, a conspicuously raised disk and
a long stipe to the ovary, has been twice collected in British Guiana:
Buruma Creek, Kibilibiri Creek, Berbice River, 1919, Forest Dept. 7744;
Upper Demerara River, Sept. 1887, Jenman 4098. This, again, is a shrub
or small tree.

GUTTIFERAE
Caraipa simplicior sp. nov.

Inter species foliis inflorescentiisque pubescentibus praeditas ob inflo-
rescentias axillares atque terminales valde abbreviatas reductas, scilicet
cymas plerumque 1-—3-floras gerentes, praeterea ob sepala magna singularis,
C. grandifoliae Mart. forsan affinis sed forma inflorescentiae praeterea foliis
petalisque minoribus distinguitur.

Arbor satis grandis, 27 m. alta; ramuli angulati superne furfuraceo-tomen-
telli. Folia lanceolata vel anguste elliptico-oblonga, apice acuta, obtusa vel
brevissime cuspidata, basi cuneata sed saepe obtusa vel fere rotundata tum
abrupte in petiolum decurrentia, 3.5-11 cm. longa, 1.3—4.3 cm. lata, illa
inflorescentias subtendentia sensim reducta, coriacea, marginibus revolutis
ac apice saepe recurvato fissoque, supra subopaca minute haud dense sed
regulariter stellato-puberula, subtus plus minusve incana indumento lepidi-
formi arcte minute stellato-tomentella, costa nervisque supra canaliculato-
impressis subtus (costa valde) prominentibus, nervis peaerve utroque

costae — 16-20 patulo-adscendentibus subrectis sed marginem versus
sursum arcuatis, nervis secundariis crebris supra eae haud_ obviis
subtus uated sed inconspicuis; petiolus 4-10 mm. longus. J/nflore-

scentiae e cymis axillaribus atque terminalibus eae simplicibus
1—3-floris rarius compositis 4-floris brevipedunculatis ad 2.5 cm. longis
ubique praesertim sepalis ferrugineo-stellato- a constantes; brac-
teae conspicuae, ovatae vel ovato-oblongae, 5-10 mm. longae, ad 3.5 mm.
latae, nonnunquam majores foliaceae; bracteolae similes sed paulo minores;
pedicelli 3-5 mm. longi rigidi. epala basi tantum connata, nna ovata

apice obtusa, basi rotundata, 5.5-6 mm. longa, 5 mm. lata, n onnunquam
extra plus minusve carinata, utrinque stellato- tomentella. Petala obovata,
unilateraliter in appendicem latam membranaceam pr ta, hac inclusa

longa, 0.6-0.8 cm. lata, partibus crassioribus utrinque dense
stellato-puberulis partibus membranaceis glabris ciliatis. Stamina longiora
ad 7 mm. longa; antherae forma generis propria obovato-rhomboidea, cir-
citer 0.75 mm. longae atque latae. Ovarium pluricostatum, 2.75 mm.
diametro, stellato-tomentellum; stylus glaber, 3.5 mm. longus, clavatus,
apicem versus sensim incrassatus. Fructus non visus.

BriTisH GUIANA-VENEZUELA BOUNDARY REGION: Roraima, slopes near Arabupu, in
swampy soil in mixed forest, 1380 m., Dec. 15, 1938, ras Dept. 2823 (field no. P. 51)
(TypuS) ; tree 90 ft. high, 12 in, diam., with scaly bak like that of Sawarri (Caryocar
nuciferum) ; fls. white, cup-shaped; calyx green; petals buff; stamens white

1943 | SANDWITH, POLYPETALAE FROM BRITISH GUIANA 221

This species somewhat resembles C. minor Huber and C. parvifolia Aubl.
in its leaves. Caraipa minor, however, has the leaves densely pilosulous
beneath with simple hairs, while its sepals are connate for at least half their
ength. Caraipa parvifolia Aubl., the type of which consists of leaves and
branchlets alone, has relatively shorter petioles and, if other French Guiana
sheets are correctly identified with it, a terminal compound inflorescence
with small sepals (see Jour. Bot. 80: 53. 1942)

MALPIGHIACEAE
Byrsonima incarnata sp. nov.; Subgen. Macrozeugma, Sect. Colobotheca, Subsect.
Atricho oe seriel Sienolens. ex clavi cl. Niedenzu in Engler, Pflanzenreich 94
(IV. 141). 1928.

Ut videtur ees B. perseifoliam Griseb. atque B. frondosam Mart. ex
Juss. ponenda, ab utraque ovario tomentoso, praeterea ab illa antherae
connectivi appendicula loculos longe superante, ab hac foliis majoribus
senectute tantum fere glabris statim distinguitur; revera, ut videtur, B.
Schomburgkianae eee magis affinis, a qua petiolis longioribus, bracteolis
ovatis, antherae loculis glabris, ovario tomentoso distinguitur.

Arbor r parva vix mediocris, circiter 13 m. alta, ramulis summis teretibus
hornotinis dense anos ferrugineo- pilosis. Stipulae ovatae, obtusae,
4.5—-6.5 mm. longae, extra indumento ramulorum indutae. Folia obovata,
obovato- elliptica vel elliptica, apice rotundata brevissime cuspidata vel in
exemplis maximis ellipticis magis attenuata ac acuminata, basi in petiolum
attenuata cuneata, 11-26 cm. longa, 5.4—10.5 cm. lata, satis tenuiter coria-
cea, supra siccitate brunneo- nigrescentia india subtus pallidiora, supra
sparse cinereo-pubescentia vel plus minusve glabra subtus juventute satis
copiose pubescentia demum fere glabra, costa nervisque primarlis supra
prominulis subtus prominentibus, nervis primariis utroque costae latere
12-18 patulis vel patulo- ascendentibus et satis longe a margine anastomo-
santibus, nervis secundariis supra inconspicuis vel impressis subtus venu-
liscum planis vel tenuiter prominulis satis conspicuis ; petiolus Paras
ramulorum praeditus, demum glabrescens, 1—2.5 cm. longus. wt 10—
18 cm. longi; rhachis angulata, satis dense subadpresse pallide ferrugineo-
pilosa, nisi ah versus “densiflora; pee anguste lanceolato-oblongae,
obtusae, 5-7 mm. longae, basi ad 2 . latae, patulae vel patentes et
sursum Saas demum deciduae cae praesertim secus medium fere
subsericeo- pubescentes: bracteolae dimidio breviores, ovatae, obtusae,
extra pubescentes; pedicelli villosulo-ferrugineo-tomentosi, 5-9 mm. longi.
Alabastra carnea. Sepala glandulas roseo-carneas 2.5-3 mm. longas sicci-
tate albo-flavescentes gerentia; lamina deltoideo-ovata, obtusa, 2.5-3.5 mm
longa, 2.5 mm. lata, utrinque pubescenti-tomentosa. Petala alba roseo-
suffusa, vel pallide rosea, ungue 3—3.5 mm. longo, lamina orbiculari-cordata
vel (in ‘petalis minoribus) reniformi- acute? 5.5-8 mm. longa, 6.5—8 mm.
lata. Torus pilosus. Stamina glabra, fi amentis brevissimis ae sub-

glabris apice obtusis rotundatis neque mucronatis nec aristatis, connectivi
5 ina valde evoluta crassa oblonga vel conoidea obtusa glabra aed

mm. lata loculos longe plus quam duplo superante. Ovarium ovoideu
ae tomentosum, 2 mm. longum, 1.75 mm. diametro; styli glabri, rae
2 mm. longi. Fructus non visus.

222 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

British Gutana: Moraballi Creek, Essequibo River, in greenheart forest on brown
loam, Oct. 13, 1938, Fanshawe in Forest Dept. 2737 (tTypus) ; Bartica-Potaro road, near
14th milepost, in wallaba forest on white sand, Aug. 19, 1937, Sandwith 1145. Each of
the trees from which these collections were made was about 50 ft. high and 4 inches
in diameter. The calyx is pale olive or brown, with waxy vermilion glands, while the
petals are pale pink or white with a tinge of pink.

The fact that this tree is almost certainly an ally of B. Schomburgkiana,
a savannah species with hairy anther-loculi in Niedenzu’s Subsect. Dasyv-
theca of the Subgenus Macrozeugma, shows that this author’s choice of
characters in the anthers for constituting major divisions of his key results
in an artificial arrangement which separates species naturally allied.

The addition of this species and of B. Poeppigiana Juss. (see Kew Bull.
1937: 102. 1937, and Lloydia 2: 188. 1939) brings the number of species
of Byrsonima known to occur in the Colony up to fourteen. The key
offered by the writer in Kew Bull. 1935: 314-315. 1935 must now be
emended for the inclusion of these two species, as follows:

Anther Joculi not horned:
Mature leaves glabrous or nearly so; petals white to deep rose
aves subsessile, rounded and more or less cordate at the bins: ae rounded o

emarginate at the apex, ..........cc cece rec ceccccacecens ; aaa.

Leaves distinctly petiolate, attenuate and cuneate at the base:
Leaves commonly or always exceeding 9 cm. in length; bracts linear to narrowly

oblong-lanceolate

Petioles less than 1 cm. long; bracteoles linear-oblong; anther site pilosulous.
Ligier giv atesia GASSES GE eh eu Gin cee a yea eee Cele B. cami
Petioles 1 cm. or more long; bracteoles ovate; anther loculi ern toa mae

Leaves usually less than 9 cm. long; bracts ovate or ovate-lanceolate. .........
Sab) asic aisig as sia Gutseregami eo aes B. eugeniifolia, concinna and bracteolaris.
Mature ph not glabrous or, if nearly so, then petals yellow; petals bright yellow
except in B. chalchophylla:
Bracts conspicuously curled backw
Leaves reddish rusty-tomentose ee petioles commonly more than 1.5 cm.
long.

Gia inds 5,955 y/o Glavara' ae) Saheia ahaa & «ewe ee eae eae RSS ea sees . Aerugo.

Leaves sparsely pubescent or glabrescent beneath; petioles rarely up to 1.5 cm.
SONG s~ 5.5 ale inte inn 514 eae vas ostream ie Gig! ates pao ats ee . Poeppigiana.
Bracts not conspicuously curled backwards. ....... Species as in the previous key.

Byrsonima Poeppigiana is a species of Amazonian Brazil and Peru which
has recently been found as a small river-bank tree in the far interior of
British Guiana, towards the Brazilian frontier.

Kostermans, in Pulle, Fl. Suriname 2: 243. 1936, and in his notes on
the new and critical species of Surinam Malpighiaceae (in Meded. Bot.
Mus. en Herb. Utrecht 25: 10-12. 1936), reduces Bentham’s B. rugosa to
B. stipulacea Juss., of which he regards it as merely a form with gland-
bearing sepals.

Banisteriopsis — (Tr. & Pl.) comb. nov
Banisteria elegans Tr. & Pl. in Ann. Sci. Nat. IV. 18: 322. 1862; Nied. in Engler,
Pflanzenreich 93 (IV. 141): 412. 1928.
Banisteriopsis elegans subsp. cordata (Nied.) comb. no
Banisteria elegans subsp. cordata Nied. in Verz. Vorles. we reaeen ts W.-S. 1912-—
1913: 10. 1912, in Engler, Pflanzenreich 93 (IV. 141): 413.

1943 | SANDWITH, POLYPETALAE FROM BRITISH GUIANA 223

Banisteriopsis elegans subsp. cordata var. ciliata (Nied.) comb. no
Banisteria elegans subsp. cordata var. ciliata Nied. in Verz. Vorles, Ak. Braunsberg
W.-S. 1912—1913: 10. 1912, in Engler, Pflanzenreich 93 (IV. 141): 413. 1928.

Banisteriopsis elegans subsp. cordata var. pulcherrima var. nov.

Var. ciliatae (Nied.) Sandwith affinis foliis magis coriaceis basi pro-
fundius eximie auriculato-cordatis, bracteis inflorescentiae irregulariter
longius fimbriatis apice non ae pectinatis fimbriis nonnullis ad 4 mm.
longis, floribus speciosioribus diffe

British GUIANA: Kaieteur a. c. 400 m., fl. and fr. Sept. 8, 1937, Sandwith
1430 (tTyPuS) ; ibid., Sept.-Oct. 1881, Jenman 1256; ibid., Aug. 1933, Tutin 694 (Herb.
Mus. Brit. and Kem) ibid., Feb. 20, 1939, Sir Ceuta Evans 12; Membaro Creek,
Mazaruni River, Feb. 14, 1939, Forest Dept. 2845 (field no. P. 73); Roraima, “Our
House,” Dec. 10, 1884, 7m Thurn 136, Sets A and C.

This is one of the most beautiful of all the remarkable plants found on
the hard sandstone conglomerate of the Kaieteur Savannah. It is a shrub
trailing over bushes and small trees to a length of some 12 feet. The
flowers are borne in showy lax sprays and simulate those of a species of
Oncidium. They are at least 2.5 cm. across, conspicuously zygomorphic,
with bright yellow fringed petals, the anterior one striped with crimson
veins in the lower half, and white stigmas. The peculiar long cilia of the
bracts of the inflorescence are crimson. The fruiting thyrse is scarcely less
attractive than the flowering sprays, since the large wings of the samaras
are pink,

It is by no means certain that this beautiful plant, with its very definite
range which is shared by so many others, does not deserve a higher rank
than is proposed for it; but this could not satisfactorily be accorded without
a thorough review of Niedenzu’s treatment of the aggregate species B.
elegans, and insufficient material has so far been accumulated for this
purpose from other parts of the range. Moreover, there is the further ques-
tion of the relationship of B. elegans with B. leptocarpa (Benth.) R. O.
Williams, to which Kostermans has recently reduced its subsp. cordata var.
ciliata (see his notes on new and critical species of Surinam Malpighiaceae,
in Meded. Bot. Mus. en Herb. Utrecht 25: 8. 1936). If this reduction be
correct, then the case for raising this new var. pulcherrima to specific rank
becomes stronger, since B. leptocarpa, which is a very common species in
British Guiana and occurs with var. pulcherrima on the Kaieteur Savannah,
is remarkably distinct from it in the field, having a much denser more
complex inflorescence, with smaller flowers and rather pale yellow petals.
But the writer is by no means convinced that the type collection of B.
elegans subsp. cordata var. ciliata, viz. H. H. Smith 1509 from Santa Marta,
Colombia, should be so summarily reduced to B. leptocarpa, since it does
not show this characteristic inflorescence.

RUTACEAE
Rhabdodendron amazonicum (Benth.) Huber in Bol. Mus. Goeldi 5: 427. 1909;
Ducke in Arch. Jard. Bot. Rio 3: 181. 1922

In the opinion of the writer, aiodon sylvestre Gleason (in Bull.

Torrey Club 54: 608. 1927) cannot be separated from Rhabdodendron

224 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

amazonicum, a frequent small tree of campos and campinas in Amazonian
Brazil. The variability of that species was discussed by Ducke in 1922
and is apparent in several British Guiana collections, especially i in the size,
the texture, and the prominence of the venation of the leaves, the ole
taken by their lateral nerves, the degree of pubescence on the inflorescence,
and the length of the pedicels. In British Guiana this species occurs as a
small to medium-sized tree, on white or dark brown sand. The following
collections are in the Kew Herbarium.

British Gut1ana: Hill slope on brown sand, Warunana Creek, Ituni River, Berbice
River, Forest Dept. 450; slope on white sand, Kuruabaru River, Demerara River, Forest
Dept. 450A; in wallaba forest, Moraballi Creek, Essequibo River, Forest Dept. 2747;
common in “clump wallaba” (Dicymbe sees eben bush, 83 miles along Bartica-Potaro
road, Tutin 243 (Herb. Mus. Brit. and Kew); in forest on white sand between
Kangaruma and Garraway Stream, Potaro River, Sandwith 1237; in dense forest
between Kangaruma and Potaro Landing, Gleason 211 (type coll. of Lecostemon
sylvestre). Forest Dept. 2747 (Moraballi Creek) was taken from a tree 60 ft. high and
12 inches in diameter; the other collections were all from small trees of the undergrowth.

The genus Rhabdodendron Gilg & Pilger and the question of its identity
with Lecostemon DC. were fully discussed by Huber in Bol. Mus. Goeldi
3: 424-431. 1909, where a new tribe in Rutaceae was proposed for
Rhabdodendron.

Rhabdodendron Gardnerianum (Benth.) comb. no
Lecostemon Gardnerianum Benth. in Hook. Jour. an & Kew Misc. 5: 295. 1853;
yok. f. in Mart. Fl. Bras. 14(2): 54. 1867; Huber in Bol. Mus. Goeldi 5: 427.
1909.

Brazit: Bahia: Banks of the Rio Preto, Sept. 1839, Gardner 2814, “a small
tree 12-16 feet high.” "This locality lies in the northwestern corner of the modern State
of Bahia; it was incorrectly placed by Bentham in the State of Piauhy, while in
Gardner’s day it apparently - within the eee la limits of that of Pernambuco
(see Gardner’s “Travels,” ed. 1. 298, 309, and m

Evidently a member . “his genus, aa strikingly distinct on account of
the small leaves, but not included in the conspectus of Rhabdodendron
given by Huber, and not mentioned by Engler in the latest treatment of
Rutaceae in Engler & Prantl, Nat. Pflanzenfam: ed. 2. 19A: 358. 1931,
where Huber’s tribe Rhabdodendreae is accepte

OCHNACEAE

Ouratea cataractarum sp. no

Inter congeneros ee foliorum inflorescentiaeque, pedicellis longis
rectis valde distincta

rutex vel arbor. parva, glabra, ramulis summis tenuibus aa 3m

diametro, fence brevibus. Stzpulae lanceolatae, 3-5.5 mm. lo one
longitudinaliter striatae. Folia anguste salicifolia, lineari-elliptica vel
lineari-lanceolata vel anguste oblanceolata, in apicem acutum longe attenu-
ata, seve acuminata, basi in petiolum longe acute attenuata atque
decurrentia, 4.5-12.5 cm. lon nga, 0.7-2.2 cm. lata, basin versus integra
superne tenuiter ater obscure (apicem versus ernie obtuse serrulata,
marginibus cartilagineis subrevolutis, coriacea, supra nitida, siccitate oli-
vacea vel subtus brunnea, costa supra prominula subtus prominente, nervis
venisque utrinque subaequaliter prominulis, nervis primariis utroque costae
latere 6-8 primo sursum arcuatis tum praerupte ascendentibus et in mar-

1943 ] SANDWITH, POLYPETALAE FROM BRITISH GUIANA 225

ginem demum provectis, venulis subhorizontalibus crebris obviis; petiolus
4-8 mm. longus. ube centia terminalis, erecta, paniculata, pyramidalis,
4-8 cm. longa, ad 10 cm. lata, ramis satis paucifloris patulis vel patulo-
pee ne He rectis angulatis inferioribus ad 4.5 cm. longis; bracteae brac-
teolaeque deciduae, haud visae; pedicelli ascendentes, recti, graciles, 1.2—1.9
cm. longi, superiores apice ramorum corymbos praebentes. Alabastra

)

u
lata, concava, subcoriacea, longitudinaliter tenuiter striata, interiora mar-
ginibus late membranaceo- scariosis. Petala laete flava, obovato- spathulata,
unguiculata, 8-9 mm. longa, 5.3—6.5 mm. lata. Antherae 6.3—7 mm. longze.
Torus sub anthesi 0.8 mm. longus, sub Ran maturis turbinato- sub-
globosus 4 mm. longus 4.5 mm. latus. Carpidia 5, sub anthesi 0.8 mm.
longa, matura (drupae) ellipsoidea vel obroted: ellipsoidea 8-8.5 m
longa ad 5.5 mm. diametro.

Britis GUIANA: Potaro River, Amatuk a on rocks by the fall, fl. Aug. 31,
1937, Sandwith 1244 (typus); Kaieteur Savannah, fl. et fr. Sept.-Oct. 1881, Jenman
890, 1213 (fr.), 1264. A shrub or small tree, with beet yellow sepals and petals.

This is a very pretty small tree, presumably to be classed as one of
Dr. Van Steenis’ “rheophytes” on account of the willow-like leaves and the
habitat — see Bull. Jard. Bot. Buitenzorg III. 12: 174-175. 1932. The
same morphological adaptation is seen in the beautiful Hirtella angustissima
Sandwith, which grows in the same locality at Amatuk and in many other
spots along the banks of the Potaro.

Ouratea cernuifiora sp. nov.

Ob racemos axillares, flores nutantes cernuos, calycem omnino integrum
coriaceum sub anthesi aliens ad basin in lobos 2—3 concavos fissum, scilicet
sepalis omnino secum concretis nec etiamnunc alabastro singillatim dis-
tinguendis imbricatis posted sejunctis, peregregia, O. vaccinioidi (St. Hil.
& 2 i) ae forsan affinis

‘rutex vel arbor parva, glabra, ramulis rigidis 2.5—5 mm. diametro densi-
foliatis ever brevibus. Stipulae lanceolatae, 10-11 mm. longae, longi-
tudinaliter striatae. Folia elliptica vel obovato-elliptica, apice obtusa
nonnunquam rotundata, summo apice ipso semper plus minusve levissime
emarginato necnon decur vo pagina superiore concavo, basi attenuata
cuneata ac in petiolum decurrentia, 2.5—-10.5 cm. longa, 1-5 cm. lata,
integra vel sub lente dimidio superiore obsolete remote undulato-serrulata,
marginibus cartilagineis, utrinque plus minusve nitida, siccitate brunneo-
olivacea, costa supra secus sulcum paginae tenuiter saepe acute Lar
subtus prominente inferne lata obtusa superne angustata acutata, nerv
venulisque supra immersis em vix vel haud distinguendis subtus nunc
prominulis nunc primariis exceptis obscuris atque plus minusve immersis,
primariis utroque costae la a circiter 8-10 primo arcuatis tum sursum
(inferioribus longe, superioribus aoe brevius) marginem versus prae-
rupte ascendentibus, intermediis osta exorientibus numerosis, venulis
creberrimis subhorizontalibus pacallclicque, petiolus supra canaliculatus,

—10 mm. longus. Inflorescentia e racemis simplicibus ex axillis foliorum

dentibus; racemi 3—7.5 cm. longi, superne satis densiflori; bracteae bracteo-

226 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

oe shone vel ovato- ee obtusae, nonnunquam emarginatae,
75 . longae, ad 2 mm. latae, longitudinaliter striatae ; pedicelli
a 3- o mm. longi. ee cernui, nutantes, siccitate (etiam in vivo ?)
saepe secundi. Alabastra ovoidea, apice brevi iter acuminata vel obtusata,
ad 6.5 mm. longa, ad 5 mm. diametro. Cie egregius, omnino coriaceus,
integer, sepalis secum omnino concretis haud singillatim distinguendis nec
ut in ceteris speciebus visis distinctis atque marginibus scariosis imbricatis,
sub anthesi ad basin in lobos 2 vel 3 valde concavos ovatos vel ap een
obtusos rotundatos vel subacutos atque plus minusve apiculatos 6—7.5 m
longos 5.5-8 mm. latos longitudinaliter striatos marginibus aieres
fissus. Petala obovato- -spathulata, elliptico-oblonga vel oblonga, apice acute

Antherae 4.3-6 mm. longae. Torus sub anthesi 0.8 mm. ea sub car-
pidiis maturis turbinato-pulvinatus 2-4 mm. longus circiter 3.5 mm. latus.
Carpidia 5, sub anthesi 0.8 mm. longa, matura obovoideo-subglobosa 5—7
mm. longa 4- 5.5 mm. diametro.

British Guana: Kaieteur Savannah, Potaro River, fl. Sept.-Oct. 1881, Jenman
863, 1042 (TYPUS), 1210, “a spreading shrub or small tree, about 10 ft. high;” Francois
Creek, Mahaicony River, Demerara Co., fr. March — padi in Forest Dept. 2367,
“shrub 6-10 ft. high, abundant in s ai savannah, soil swampy with peaty humus
overlying white sand. Peduncle, calyx and fruit deen ie seeds dark Fe greyish-
green or pale yellow.”

An outstanding species of Ouratea on account of the ascending axillary
racemes with nodding flowers on recurved pedicels and the very peculiar
calyx with completely fused and indistinguishable sepals, ruptured at the
time of flowering into 2 or 3 concave coriaceous lobes. A tendency to a
gradation towards this phenomenon can be observed in certain other species,
for instance, on some specimens of O. acuminata (DC.) Engl., in which the
inner sepals adhere so closely to the coriaceous outer members that their
broadly scarious free margins cannot be separated and distinguished with-
out dissection and examination of the inner side of the whole calyx.

A possible ally of this plant is O. vaccinioides (St. Hil. & Tul.) Engl.,
of the States of Bahia, Minas Geraes, and Rio de Janeiro, in Central Brazil:
it has somewhat similar but uniformly small leaves, terminal panicles of
nodding more or less secund flowers, and distinct sepals.

This is the second instance of a rare undescribed plant growing both on
the British Guiana peneplain, on the low-lying scrub savannah of the
Mahaicony River, and on the Kaieteur Savannah at the edge of the
Pakaraima escarpment; the previous example was the malpighiaceous
Byrsonima eugenitfolia Sandwith.

Several other undescribed Ourateae from British Guiana are represented
in the Kew Herbarium, but they are more critical than the two very distinct
species described above, and a review of the members of the genus occurring
in the Colony cannot be prepared without the examination of specimens and
literature which are at present unavailable. The number of species of
Ouratea found in British Guiana is probably not less than twenty.

Royat Botanic GARDENS,
Kew.

1943] JOHNSTON, SPECIES FROM MEXICO AND UNITED STATES, I 227

NOTEWORTHY SPECIES FROM MEXICO AND ADJACENT
UNITED STATES, I

IvAN M. JOHNSTON

SINCE, in continuation of studies of the flora of northern Mexico, many
noteworthy species from the adjacent parts of the United States will be
discussed, it seems advisable to begin a new series under the above more
inclusive title to replace my ‘““New Phanerogams from Mexico,” of which
no. V was published in this Journal, 24: 90-98. 1943. Unless otherwise
noted, the material cited in this series is deposited in the Gray Herbarium
or in the herbarium of the Arnold Arboretum.

Atriplex prosopidum, sp. nov.

Frutex monoicus rotundus pallidus 5-12 dm. altus ramosissimus; ramulis
gracilibus foliosis ascendentibus numerosis teretibus pallidis vesiculosis;
foliis integerrimis numerosis alternis oblanceolatis ad late oblanceolatis vel
oblongo-obovatis, 4-plo ad subduplo longioribus quam latis, 15-20(—24)
mm. longis, 3—6(—7) mm. latis, supra medium vel medium versus latioribus,
deinde basim versus in petiolum 1-2 mm. longum gradatim attenuatis,
costatis sed enervatis, indumento pallido vesiculoso tectis, apice obtusis vel
subretusis; floribus masculis flavis in panicula conspicua terminali subnuda
moniliforma dispositis; floribus femineis abundantibus glomeratis in spica
plus minusve interrupta infra medium foliato-bracteata dispositis; brac-
teis fructiferis 4-alatis dense furfuraceis summum ad apicem connatis, cor-
pore (alis 1-3 mm. latis exclusis) 1-2 mm. longe stipitato os 5 mm. longo
infra medium crassiore apice in rostrum 1—1.5 mm. longum angustum vel
triangularem protracto, corpore alato a latere viso triangulari vel triangulari-
ovato, 2.5—6 mm. lato, supra medium saepe latiore, basi rotundo vel obtuso,
margine alis promi inenter paucidentatis; seminibus brunneis ad 2 mm
longis, radicula lateraliter erecta.

Coauutta: About 10 miles north of Cuatro Cienegas, Wynd 742 and 744; south of
El Oso, rounded bush 2-3 ft. tall, Johnston 8877 (tTypr, Gray Herb.); near Flores,
plant bushy, somewhat globose, 1—4 ft. tall, Johnston 8875; 12 miles north of Monclova,
bush 3-4 ft. tall, Johnston 7187

This species appears to be a relative of A. canescens, but differs in selec-
tion of habitat, indument, form of growth, color of herbage, shape of leaves,
and size and shape of fruit. The fruit has four wings as in A. canescens, but
it is smaller, distinctly more furfuraceous, and short-stipitate. About Flores
and El Oso, on the road from Cuatro Cienegas north to Ocampo (in the
area where Wynd also collected the plant), I was able to compare living
specimens of A. prosopidum and A. canescens. The former is a pallid
rounded somewhat globose bush growing with Suaeda. Atriplex canescens
did not associate with the Suaeda and differed conspicuously from A.
prosopidum in being taller, stricter in branching and erect, and a light ashy

228 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

green rather than pallid. The indument of A. prosopidum is similar to that
of A. obovate and A. Nuttallii, and hence very much denser and much paler
than in A. canescens. North of Cuatro Cienegas and north of Monclova
the plant grows abundantly in silty, somewhat gypseous and saline valley
soils supporting a luxuriant growth of Prosopis glandulosa. About Flores
and El Oso A. canescens was rare in the areas where A. prosopidum abound-
ed, perhaps because of the gypseous content of the soil.

Suaeda jacoensis, sp. nov.
Planta succulenta annua vel perennis 1~3 dm, alta glabra pallide viridis,

apiculatis, inferioribus conspicue oppositis, ceterum alternis; floribus axil-
laribus sessilibus glomeratis subspicatis; bracteis hyalinis triangularibus
vel oblongis dentatis vel lacerato-dentatis haud conspicuis; calyce fructifero
latiore quam lato (sine appendiculis 1—1.5 mm. diametro) depresso asym-
metrico, lobis deltoideis margine scariosis dorso irregulariter tuberculato-
cristatis vel carinatis, basi non raro transverse alatis vel gibbosis; seminibus
horizontalibus 1.1-1.3 mm. longis nitidis atro-brunneis; stylis 2.

CoaHuILA: Salt-flats at the southeastern end of Laguna de Jaco, common, Johnston
& Muller 1087, Stewart & Johnston 1975 (type, Gray Herb.) and 1976.

Related to Suaeda mexicana, from which it differs in its lower stature,
more slender and smaller leaves (which are conspicuously opposite below),
more slender somewhat purplish usually verrucose stems, and irregularly
cristate and keeled mature calyx-lobes. The perianth-lobes become un-
equally cristate and carinate at maturity, some being almost eg cece
others corky-tuberculate, while others may have a vertical as well a
transverse basal keel that may become corky-thickened. In S. mexicana “all
the perianth-lobes become corky-thickened and prolonged off obliquely at
their base. The proposed species may appear as a perennial, with the
stems arising from a small caudex crowning a coarse persisting root. Most
of the plants seen, however, appear to be annual. The species is very com-
mon on the flats at the southeastern end of Lake Jaco. The soil here is
strongly saline and also gypsiferous.

Suaeda nigrescens, sp. nov.

Frutex 3—6 dm. altus ascendenter ramosus erectus vel decumbens in sicco
saepe nigrescens; ramulis gracilibus rigidulis fruticosis ramulosis angulatis,
pilis minutis brunnescentibus saepe abundantibus donatis, in vivo saepe
purpurascentibus; foliis ee ot plus minusve glaucis glabris 3-5(-8)
mm. longis 1(—2) mm. latis, plus minusve compressis lineari- oblongis vel
oblanceolatis numerosis confertis, medium versus latioribus, apice obtusis
vel rotundis, basi angustatis ; floribus glomeratis basim versus foliorum
superio rum vix conspicuorum enatis homomorphis subspicatis; bracteis
scariosis triangularibus vel ovato-triangularibus sparse dentatis incon-
spicuis haud persistentibus; calyce fructifero glabro turbinato 1.5—2 mm
diametro, lobis deltoideis convexis succulentis; stylis 2 vel 3; seminibus
nigris nitidis oblique ovoideis crassis erectis vel horizontalibus ca. 1 mm.
longis.

1943] JOHNSTON, SPECIES FROM MEXICO AND UNITED STATES, I 229

CoAHUILA: 4 mi. southwest of Hermanas, saline flat, Johnston 7074; 8 mi. north
of Avalos, saline flat, Johnston 7340; 12 mi. north of La Ventura, saline flat, Johnston
7650 (type, Gray Herb.). San Luis Potosi: 6 mi. northwest of Cedral, saline flat,
Johnston 7603; San Miguel, saline flat, Johnston 7619; San Domingo Lake, Lundell
5576, Hacienda del Salado, Dec. 1827, Berlandier 1345; San Luis Potosi to Tampico,
Palmer 778%

Suaeda nigrescens, var. glabra, var. nov.

A varietate typica differt ramulis omnino glabris saepe glaucis, raro
ramulis junioribus pilis sparsissimis pallidulis donatis.

CurHvuaHva: Meoqui, 1935, LeSueur 197. Coanutta: about 30 mi. south of Sierra
Mojada, 1937, Wynd 771. Texas: along irrigation ditch, Rio Grande Valley, near old
Fort Quitman, Waterfall 3993a (typr, Gray Herb.) ; saline plains of Rio Grande [near
Eagle Pass], 1848, Wright; salt flats near Pilgrim, Gonzales Co., Cory 18992 and
19213. New Mexico: White Sands, Otero Co., Archer 7334; Hagerman, Chaves Co.,
Benke 5023

The present plant is that treated by Standley, in No. Am. Fl. 21:90
(1916) and in Bull. Torr. Cl. 44: 428 (1917), as indistinguishable from
the Old World Suaeda fruticosa (L.) Forsk. That species was based on
plants from the Mediterranean coast of France and Spain. It differs from
our present plant in aspect, as well as in the more slender and elongate
styles and less persistent more scarious bracteoles. Our species behaves as
a native plant and I am convinced that it is indigenous. It is related to
S. fruticosa, but so are such generally recognized West American species
as S. Torreyana Wats., S. Palmeri Standl., S. ramosissima (Standl.) John-
ston, etc. In fact, our plant is as closely related to these American con-
geners as it is to the Old World plant. Unless most of the American
members of Suaeda Section Salsina are to be submerged in S. fruticosa, the
present species must also be accorded recognition.

I can find no available published names for the present plant. Suaeda
fruticosa var. multiflora Torr. (1857) is based upon a collection by Bigelow,
from “Llano Estacado,” or, to be more precise, from the region about
Tucumcari, New Mexico. This plant may represent S. nigrescens var.
glabra, but because of Suaeda multiflora Phil., a Chilean species, the name
is unavailable for our North American plant. Suaeda Moquini (Torr.)
Greene, based upon the casually published Chenopodina Moquini Torr.
(1858), is properly discarded as a hyponym by Standley, in Bull. Torr. Cl.
44: 428 (1917). Standley, in No. Am, FI. 21:90 (1916), cites S. inter-
media Wats. (1879) asa synonym of S. fruticosa. That species, based upon
cited material from Utah, collected by Ward, Parry, Hooker and Gray, is
not S. nigrescens. Since Parry 84, from “Central Utah,” is the only cited
collection of S. intermedia with good flowers and is the only one on whose
sheet Watson wrote his binomial, I am taking that plant as the type of
S. intermedia Wats. This type and the other cotypes all appear to repre-
sent the common Utah plant which most recent authors, including Standley,
have called S. Torreyana Wats. Though it seems probable that Watson
proposed S. Torreyana to include American plants formerly placed in
S. fruticosa, usage has centered his binomial on Great Basin plants, and
Standley has typified it with the material collected on the west shore of the

230 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Great Salt Lake by Stanbury. Swaeda diffusa Wats. was based upon eleven
collections, of which the two collections of Charles Wright represent S.
nigrescens, and of which Gregg 458 represents S. suffrutescens var. detonsa.
Seven of the original eleven collections came from Utah and Arizona and
appear to represent the plant commonly called S. Torreyana. In describing
S. diffusa, Watson drew heavily on his collection no. 996 and his account
of it in the Botany of the King Expedition, p. 294. That specimen is logi-
cally the type of his S. diffusa. It was collected in Truckee Valley and is
conspecific with Heller 8650, also from that area, and obviously belongs to
the aggregate called “S. Torreyana.” Standley cites Dondia conferta Small
as a synonym of S. fruticosa. That species, based on plants collected on
the coastal flats at Corpus Christi, Texas, differs from S. migrescens in its
spreading less fruticose habit, large flowers, and conspicuous leafy bracts.
It is the sea-coast plant found on the coast of Texas and eastern Mexico and
in the West Indies. It should be called:
Suaeda conferta (Small), comb,

Dondia conferta Small, in Bull. 5 Y. Bot. Gard. 1: 280. 1899.
Suaeda suffrutescens Wats., var. detonsa, var. nov.

A varietate genuina differt foliis glabris, caulibus non raro glabrescen-
ibus.

CoAHUILA: 3 mi. west of Cuatro Cienegas, saline flats, loosely and widely branched,
1—5 ft. tall, frequently supported by other bushes, Johnston 7128 (Type, Gray Herb.) ;
3 mi. south of Cuatro Cienegas, saline land, low shrub, White 1913; Cuatro Cienegas,
Marsh 2042; Perros Bravos, Bota 3 ft. tall, abundant, Sept. 20, 1848, Gregg 458;
Saltillo, July 1880, Palmer 1167; 5 mi. west of Viesca, plant erect, Johnston 7739;
Bolson de Mapimi, April 15, 1847, Gregg #49.

Suaeda suffrutescens is the most common and mostly widely distributed
member of its genus in trans-Pecos Texas and along the Rio Grande and
Pecos valleys in adjacent New Mexico. It extends south in northern
Coahuila and Chihuahua. If not restricted to saline gypseous soils, it at
least appears strongly to favor that substratum. ‘The species is readily
recognized because of its pubescence, abundant staminate flowers, and very
moderately fleshy usually hairy perianth-lobes.

As originally described, the species was a complex. The specimens orig-
inally cited are Berlandier 1345, representing S. nigrescens from San Luis
Potosi, and Wright 578 in part and Emory, representive of the species as
here taken. In subsequent usage the name S. suffrutescens has been applied
to the present plant of Texas and to another, more western species, also with
hairy herbage. The latter, S. ramosissima (Standl.) Johnston, was segre-
gated and named by Standley, who redefined S. suffrutescens and gave its
type-locality as ‘“‘Western Texas.” I have selected Wright 578 in part as
the type-collection. This is part of a mixed sheet containing specimens
obtained by Charles Wright at three different localities in trans-Pecos Texas
and representing three different species, S. suffrutescens Wats., S. duripes
Johnston, and S. nigrescens var. glabra Johnston. These were obtained at
(1) Escondido Springs, about 20 miles east of Ft. Stockton, Aug. 17, 1849,
(2) near Ft. Stockton, Aug. 18, 1849, and (3) near San Elizario, Oct. 4,

1943] JOHNSTON, SPECIES FROM MEXICO AND UNITED STATES, I 231

1849. Just which species was obtained at which locality cannot now be
determined. The largest specimen on the sheet represents S. suffrutescens
as here accepted. It was distinguished from the other material on the sheet
by Watson and labeled by him “‘S. suffrutescens, W.” It is a form of the
species with the leaves only sparsely hairy. Watson’s reference to the
material collected by “Emory” is unintelligible unless it refers to the ma-
terial treated by Torrey in the Botany of the Mexican Boundary, p. 184,
as Suaeda fruticosa var. multiflora. Torrey cites collections of Bigelow
from Presidio del Norte (i.e. Ojinaga) which probably represent S. suffru-
tescens, the common species at that locality, and collections of Thurber
from the Rio Pecos, which perhaps represent S. duripes.

The plant of southern Coahuila, which I have distinguished as S. suffru-
tescens var. detonsa, differs from the more northern typical form of the
species in having green glabrous leaves that contrast noticeably with the
hairy stems. Its stems are usually densely hairy, occasionally, however,
they are only sparsely so. Among the specimens which I have cited as
representing the var. detonsa, Gregg 458 was referred to S. diffusa and
Gregg’s collection from the ‘‘Bolson de Mapimi” was referred to S. Tor-
reyana when Watson originally published these latter species.

Suaeda duripes, sp. nov.

Planta succulenta herbacea vel fruticosa 5-30 cm. alta; ramis gracilibus
ut videtur ascendentibus vel decumbentibus, laevibus vel obscure vesicu-
lato-tuberculatis, glabris vel pilis caer minutissimis a ae foliis
succulentis linearibus vel oblongo-obovatis 3-7 mm. longis 1—3 mm. latis
compressis, basim versus oradlanitn Hehe. haud petiolatis, late affixis a
basi usque ad 1—-1.5 mm. supra basim induratis et in caule persistentibus,
apice obtusis vel rotundis; glomerulis plurifloris, basi foliis haud vel breviter
longioribus basim versus induratis enatis; bracteolis inconspicuis mox
deciduis membranaceis triangularibus vel lanceolato-triangularibus saepe
integris; perianthio infra medium lobato sub anthesi succulento maturitate
vesiculoso et suberoso-incrassato depresso ca. 2—2.5 mm. diametro, lobis
ahaa convexis; seminibus nigris erectis nitidis laevibus

TEa Pecos River, in salt soil, ee 1850, Thurber 114 mes Gray Herb.) ;
trans- ee Texas, 1849, Wright 878 in

A well marked species, nee jeer to Reeves and Pecos Counties,
Texas, and characterized by thickened and persistent leaf-bases and the
corky-thickened lobes of the fruiting calyx. The collection of Charles
Wright is part of the mixed sheet which I have described above in my dis-
cussion of S. suffrutescens. It consists of small apparently annual plants less
than a decimeter tall. Since Wright’s field notes give his collection from
San Elizario as a small bush, it is obvious that his material of S. duripes
was collected either near the present Ft. Stockton or about 20 miles to the
eastward, along the old San Antonio-El Paso wagon road, at the watering
place which Wright called Escondido Springs. Thurber’s material was
collected along the northern wagon road to El Paso which crossed the
Pecos about 50 miles south of the New Mexico boundary and then went up
the west side of the river to Delaware Creek before continuing westward.

232 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Tidestromia carnosa (Steyerm.), comb. nov.
Cladothrix lanuginosa var. carnosa Steyerm., in Ann. Mo. Bot. Gard. 19: 38%
2

1 ‘

ar wince lanuginosa var, carnosa Cory, in Rhodora 38: 405 (1936).

This plant appears to be restricted to gypseous saline Upper Cretaceous
clays and is known only from Brewster County, Texas, and south of
Ojinaga, Chihuahua. South of Ojinaga I found it growing in close prox-
imity to T. lanuginosa. The two species were ‘0 completely different in
appearance and so constant in their differences that it seems surprising that
they have continued to be treated only as variety and species. From the
widely ranging 7. /anuginosa the present species differs in the very fleshy
yellowish green nearly glabrous herbage, more brittle stems, and more
indurate involucres bearing subsessile leaves. These characters are con-
stant in all the herbarium material I have examined. The closest relative
of T. carnosa is T. tenella Johnston, of the gypseous soils in western
Coahuila south of Laguna del Rey, from which it differs in having much
coarser stems and leaves, flowers that dry brownish rather than white, and
coarsely branched rather than minutely barbellate hairs along the stem
and on the flowers. The three species 7. lanuginosa, T. carnosa, and T.
tenella are all annual herbs. The remaining species of the genus are all
strong-rooted perennials.

Tidestromia suffruticosa, var, coahuilana, var. nov.

A varietate typica differt indumento laeviore minus pruinoso, foliis in-
Be heer petiolatis, floribus mox glabratis; caulibus rigidis distincte
fruticosis

Cocina Sierra Cruces, 5 km. west of Picacho San José, limestone hillside,
Stewart 820; Sierra Cruces, 5 mi. north of Santa Elena, stony bank, Johnston & Muller
1014 (rype, Gray Herb.); Sierra del Pino, Cafon Ybarra, dry hillside, Stewart 1853;
Lomas del Aparejo, east side of Llano de Guaje, Johnston & Muller 773; south end of
Sierra del Pino, northeast of Armendais, rocky flat, Johnston & Muller 363; western

ocorro, dry arroyo banks, Johnston 8844; Sierra Baila. Oct. 1910,
Paris 4927.

Typical 7. suffruticosa was based upon material collected by Wright on
mountain sides near Frontera, July 19, 1851, and on hills about 30 miles
west of the Pecos, Aug. 18, 1849. Subsequent collections have been made
in Brewster County, Texas, and in Dona Ana County, New Mexico. This
typical form, of Texas and New Mexico, differs from the plant of central
and western Coahuila in its looser pruinose indument, much less woody
stems, permanently hairy flowers, and shorter petioles on the involucrate
leaves. These differences are not always strongly developed, but together
they add up to appreciable differences in the gross aspects of the two plants.
In Coahuila the plant is most commonly found on planed down upper
Cretaceous beds at the base of anticlinal mountains and is usually associated
with Dyssodia acerosa and Coldenia hispidissima and other indicators of
somewhat gypseous soil. The localities at which the typical form of 7.
suffruticosa has been collected suggest that it too may select rocky soils
containing at least some gypsum

1943] JOHNSTON, SPECIES FROM MEXICO AND UNITED STATES, I rE)

Tidestromia rhizomatosa sp. nov.

Planta perennis prostrata ramosa grisea pilis ramosis vestita; rhizomati-
bus laevibus brunneis gracilibus; caulibus 1-2 mm. crassis griseis 1-2 dm
longis basi gemmis albis parvis donatis, internodiis 5-20 mm. longis; foliis
carnosis saepe plus minusve reflexis et plano-convexis subtus conspicue
crassinerviis, inferioribus obovatis 5-12 mm. longis 5-10 mm. latis, maturi-
tate plus minusve glabrescentibus, supra medium latioribus deinde basim
versus gradatim attenuatis, apice rotundis vel obtusis, superioribus ovatis
sessilibus; glomerulis 1- vel 2-floris; floribus ee 3 mm. longis; lobis
perianthii obtusiusculis ovato-oblongis ad 3 mm. longis 1-1.2 mm. latis
quam bracteis lateralibus circa duplo ee ena pee basim versus glaber-
rimis brunnescentibus alibi tomentosis; tubo staminali 0.4-0.5 mm. alto
staminodia 5 et filamentas 5 gerentibus; filamentis subulatis 0.7—0.9(—1.2)
mm. longis, antheras ad 1 mm. longas 4—5-plo longiores quam latas geren-
tibus; staminodeis subulatis quam filamentis 1/5—1/2 brevioribus; ovario
globoso; apie profunde bilobato; semine ad 1.3 mm. longo erecto
brunnescente ovo

CoAaHUILA: Near a Anteojo, about 3 miles west of Cuatro Cienegas, with Allenrolfea
on a silty saline, probably also gypseous, flat just east of the pond, locally abundant,
decidedly prostrate, grayish, rhizomatous, leaves fleshy and more or less reflexed,
Sept. 5, 1941, Johnston 8873 (vTypr, Gray Herb.).

A perennial species with the stems flat on the soil and spreading under-
ground by slender smooth rhizomes. The leaves are fleshy, thinly clothed
with trichomes, and more or less plano-convex and reflexed. They are very
strongly veined beneath. The species is most closely related to 7. gemmata,
from which it differs in its rhizomes, lank more elongate stems, more or less
glabrescent more fleshy reflexed leaves. slightly larger dower: and the
development of staminodes. The well ‘developed staminodes distinguish
the species from all its congeners save only TJ. oblongifolia (Wats.)
Standley, of California and Arizona. The plant is probably a halophytic
gypsophile.

Tidestromia gemmata, sp. nov.

Planta perennis pilis ramosis dense vestita, e radice valida palari profunda
apice gemmis conspicuis albo-tomentosis donata et non raro caudicem com-
pactam proferente oriens; caulibus laxe ramosis prostratis vel decumbenti-
bus rubiginosis pilis griseis vestitis 1.5-2.3 mm. crassis 1-1.5 dm. longis,
internodiis 1—2.5 cm. longis; foliis juvenilibus albis vel cinereis, maton
dilute flavescentibus; foliis inferioribus conspicue petiolatis, lamina
12-15(—23) mm. longa 1—1.6 mm. lata, subtus prominenter nervata, apice
obtusa vel rotundata, basi truncata vel obtusa in petiolum 4-8 (- 10) mm
longum abrupte contracta: foliis superioribus numerosis ovatis subsessilibus
sursum gradatim reductis: glomerulis 2—4-floris; floribus lacteis 2.5 mm.
longis; lobis perianthii ad 2 mm, longis lanceolato- oblongis vel ovato-
oblongis obtusiusculis quam bracteis lateralibus circa duplo longioribus,
extus basim versus glabris alibi praesertim medium versus tomentosis; tubo
staminali 0.6-0.7 mm. alto; staminodeis nullis; filamentis 0.8—1.2 mm.
longis; antheris oblongis 0.8— ‘1 mm gi

CoaHUILA: Just south of ‘apenas Viejo, restricted to gypsiferous shales, ed
trate, common, Sept. 22, 1941, Johnston 9363 (TYPE, Gray Herb.) ; just east of Amer

234 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

canos, on cemented gravels capping gypsum, prostrate, locally common, Sept. 23, 1941,
Johnston 9379; 20 km. southeast of Rancho Alegre on road to Acatita, on flats, pros-
trate, common, Sept. 20, 1942, Stewart 2668.

This species superficially resembles coarse plants of T. lanuginosa, but
differs in its coarse very strong tap-root, conspicuous large cottony buds
borne near the surface of the soil, slightly more fruticulose reddish stems,
and thicker more strongly veined leaves. I first recognized the distinctness
of the species near Matrimonio, where I found it growing near specimens
of the common and widely distributed annual, 7. /anuginosa. I suspect that
the Tidestromias I noted as growing in gypsum seams in the shales south
of Laguna de Leche may also belong to T. gemmata.

Spiraea Northcraftii, sp. nov.

Frutex; ramulis gracilibus rigidis laxe ascendentibus vel subdivaricatis
glaberrimis, vetustioribus griseis subsp:nescentibus, juventute castaneis;
foliis minutis numerosis crassis laevibus enervatis obscure costatis pallidulis
glaberrimis solitariis vel plus minusve fasciculatis integerrimis oblanceolatis,
~8 mm. longis 1—2.5 mm. latis, apicem versus latioribus deinde basim
versus in petiolum 0.5—1 mm. longum subcastaneum gradatim contractis,
apice rotundis vel obtusis abrupte conspicueque mucronatis; inflorescentia

terminali subcorymbosa 8—15-flora plus minusve hemisphaerica ca. 8 mm.
diametro glabra; bracteis linearibus 1—-1.5 mm. longis, pedicellis 1—-1.5 mm.
longis, hypanthio 1.5 mm. alto et diametro, lobis triangularibus ad 1 mm.
longis valvatis intus sparse tomentulosis; petalis albis 1.5-2 mm. longis
1—-1.3 mm. latis, margine erosis; staminibus 5, cna linearibus 1.5
mm, longis; carpellis 2 vel 3 glaberrimis: folliculis non v

CoauvuiLa: Summit of Picacho de Jimulco, 13 km. east of aes in thick under-
brush in woodland association with oak, pine and juniper, June 29, 1941, Stanford,
Retherford & Northcraft 115 (tvpr, Gray Herb.).

A microphyllous shrub related to S. Hartwegiana Rydb., of Hidalgo,
Puebla and Oaxaca, from which it differs in being smaller in all parts, and
in having a shorter less elongate hypanthium and a dense few-flowered sub-
corymbose rather than an elongate many-flowered racemose inflorescence.
Vauquelinia scape eiaaen sp. nov.

altus mulis a grisea glabra obtectis; foliis apice
ramulorum confertis beers oblongis 2—5 cm. longis 4-7 mm. latis, basim
versus in petiolum 2—4 mm. pea pallidum abrupte contractis, supra
viridibus inconspicue sparseque puberulentis canaliculatis sed inconspicue
minuteque reticulatis, subtus pallidis tomentulosis pilis mollibus contortis
brevibus albis dense vestitis maturitate aliquantulum deciduis, sc
utrinque dentibus 15-20 parvis ascendentibus donatis, apice obtusis;
corymbo terminali 2-3 cm. diametro 10—30-floro, ramulis Saute as be
hypanthio hemisphaerico 3-4 mm. diametro dense puberulente; sepalis

ovatis intus tomentulosis; petalis ellipticis albis ad 4 mm. longis ca. 2 mm.
latis integerrimis; staminibus 12-15, filamentis ROT ca. 3 mm. longis;
capsulis submaturis globoso- ovoideis ca. 3 mm. di

OAHUILA: Sierra Jimulco, about 11 km. northeast of ae Ico, in rolling hills

covered with Agave, Yucca and low mesquites, Stanford, Retherford & Northcraft 87
(TyPE, Gray Herb.).

A species related to V. californica, from which it differs in its smaller

1943] JOHNSTON, SPECIES FROM MEXICO AND UNITED STATES, I 235

narrow elongate short-petiolate leaves clustered on short-shoots, its per-
manently tomentulose lower leaf-faces, and its small few-flowered corymbs.
Thamnosma Stanfordii, sp. nov.

aes 6 dm. altus ramosissimus glaberrimus; ramulis teretibus tubercu-
latis 1-2 mm. crassis foliosis evanescenter glaucis foliis numerosis confertis
aoe oblanceolatis vel linearibus compressis 7-12 mm. longis 0.8—1.5 mm.
latis, medium versus usque ad apicem versus latioribus, costatis sed ener-
vatis, j juventute subglaucis, margine minute sed distincte sinuatis; floribus
1 vel 2 extra-axillaribus apicem ramuli versus gestis; pedicellis 1.5—2 mm.
longis, fructiferis ad 5 mm. longis: calyce ca. 4 mm. longo infra medium
lobato, lobis triangulari-oblongis ca. 3 mm. longis obtusis purpurascentibus
basi ad 2 mm. latis; corolla alba purpurascente; petalis 4, ad 10 mm. longis,
oblongo-ovatis (infra medium ad 4 mm. latis) acutis; filamentis 4 exteriori-
bus ad 5 mm. longis; filamentis 4 oppositipetalis 3 mm. longis linearibus;
antheris ellipsoideis ca. 1.5 mm. longis; stylo stamina superante; capsu ula
stipite ad 1 mm. jongo incluso 1 cm. longa, ca. 6 mm. lata, apice ca. 2 mm.
profunde bilobat

CoAHUILA: sh de Jimulco, 11 km. northeast of Jimulco, rolling hills with Agave,
Yucca and low mesquites, plant 2 ft. tall, fl. purple, fennel-like odor, June 27, 1941
Stanford, Retherford & Northcraft 16 (vtypre, Gray Herb.).

A very distinct addition to this small genus of North America and Africa.
In its strongly biseriate stamens and in its bushy growth-habit it departs
widely from the three American species previously described. It is a much-
branched slender-stemmed bush with numerous leaves. The leaves some-
what suggest those of JT. texana. The large flowers superficially resemble
those of T. montana. The capsule most suggests that of 7. texana but is
larger and more elongate. The stamens in 7. texana and T. montana
(flowers not known in 7. ¢rifoliata) are subequal. In 7. Stanford the
filaments of the inner whorl of stamens are very conspicuously shorter than
those of the outer whorl.

Pseudosmodingium ? anomalum, sp. nov.

Frutex erectus ca. 2.5 m. altus; ramulis vetustioribus cortice grisea glabra
obtectis, hornotinis dense puberulentis: foliis submaturis (vetustioribus non
visis ) simplicibus angustissimis integerrimis minute puberulentis apice ra-

mulorum confertis, 2-5 cm. longis 2-3 mm. latis, utrinque attenuatis
costatis sed haud vel vix nervatis, 1-2 mm. longe petiolatis: paniculis sub
anthesi (fructiferis non visis) 4- 5 cm. longis elongatis sparsifloris sparse
ramosis minute sparseque puberulentis quam foliis subduplo longioribus,
axi paniculae gracillimo parte 1/3 raro ad 1/2 inferiore ae deinde sursum
ramulos 3 vel 4 gracillimos 3-12 mm. distantes ascendentes sparsifloros
3-10 mm. longos simplices vel raro pauciramosos gerente; bracteis lineari-
bus 1-2 mm. longis; ae minutis 5-meris; pedicellis gracillimis 1-3 mm.
longis; sepalis ca. 0.6 mm. longis et latis triangularibus apice rotundis basi
abrupte constrictis; ate imbricatis ellipticis vel elliptico-ovatis, 1.5 mm.
longis 1.2 mm. la tis, medium versus vel infra medium latioribus, apice
rotundis, basi subtruncatis : disco patelliformi; antheris ca. 0.5 mm. longis
et latis, basi cordatis, apice ‘yotundis: filamentis subulatis ca. 0.6 mm. longis
latere exteriore disci affixis; ovario sessili glabro stigmatibus 3 tuberculatis
terminato; fructu ignoto.

236 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Coanutta: Sierra de Jimulco, about 11 km. northeast of Jimulco, along arroyo in
rolling hills covered with Agave, Yucca and low mesquites, plant uncommon, erect,

8 ft. tall, June 28, 1941, Stanford, Retherford & Northcraft 73 (type, Gray Her

A deciduous shrub with slender woody stems bearing scattered aur of
elongate simple entire leaves on short-shoots. The type material shows
nearly fully grown new leaves and well developed inflorescences with the
flowers at anthesis. Without fruit the generic position of the plant is
necessarily uncertain. I have placed it in Pseudosmodingium because in
that genus I find leaves clustered on similar short-shoots, panicles of similar
position and form, and flowers of similar appearance and structure. Prof.
Irving W. Bailey reports that the wood of the plant is also suggestive of the
genus. The described species of Pseudosmodingium, however, have well
developed imparipinnate leaves. The leaves are distinctly simple in the
present plant. Their obscure veining shows that they are not phyllodal
and the absence of a secondary basal articulation gives no indication that
they might be the terminal leaflet of an otherwise suppressed pinnate leaf.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

1943 ] JOHNSTON, PACIFIC RAILROAD REPORTS 237

PUBLICATION DATES FOR THE BOTANICAL PARTS OF
THE PACIFIC RAILROAD REPORTS

IvAN M. JOHNSTON

N the present paper I present such data as I have been able to assemble
concerning the exact dates of publication of the various botanical papers
contained in the Pacific Railroad Reports. These reports, based upon
explorations in the western United States between 1853 and 1855, and pub-
lished by the War Department under the lengthy title, “Reports of Ex-
plorations and Surveys for a Railroad Route from the Mississippi River
to the Pacific Ocean,” include important botanical papers by Torrey, Gray,
Engelmann, and others, in which were first described a large number of
the characteristic West American plants. A study of the various volumes
of the Pacific Railroad Reports reveals puzzling discrepancies between
dates found on the title pages of the volumes, those on the initial leaf of
the separate reports within each volume, and those found scattered through
the text. In a search for precise information regarding the dates of pub-
lication of the botanical portions of the Pacific Railroad Reports, I have
consulted the Historic Letter File at the Gray Herbarium and searched
for contemporary mention of these papers. The letters of John Torrey and
George Engelmann have supplied much detailed information.1 The pub-
lished proceedings of the Academy of Natural Sciences of Philadelphia, the
American Philosophical Society, and the Academy of Science of St. Louis
have also provided exact dates at which copies of the completed volumes of
the Reports had been distributed from Washington. From these sources it
has been possible to assign reasonably exact dates to the various botanical
reports, accurate in most cases to within a month or two. In less dis-
turbed times, when at least Torrey’s letters at Kew and St. Louis and Gray’s
letters at Kew, New York, and St. Louis can be examined, it seems probable
that additional information may be found which will establish an even
more precise dating for these papers.

1Contemporary letters also om to date Torrey’s and Engelmann’s important botani-

cal reports in vol. 2 of Emory’s Report on the United States and Mexican Boundary
rvey. Torrey wrote Gray, on Jan. 10, 1859, that he was preparing a list of errata
from page-proof of his Mexican Boundary Report. Engelmann wrote Gray on Apr 5
1859, that he had ordered separates of his account of ihe boundary Cacti but did
not know if his report had been printed. On June 7, 1859, Torrey was expecting his
pHinted report. Schott (in a letter at the N. Y. caine al Garden, fide notation in the
Gray Herbarium copy) stated that the botanical reports were issued before April 21,
1859, On June 2, 1859, Engelmann wrote Gray that he had seen the printed report,
y

undistributed. The second volume of the Boundary Survey was obviously issued in
May or late April, 1859. No advance separates of the botanical reports were issued.

238 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

The Pacific Railroad Reports appeared in two editions, first in octavo
and later, much enlarged in scope, in quarto. In explanation of this pro-
cedure Torrey wrote Gray, on Sept. 14, 1854, “This [Lieut. Whipple’s
report | will be printed in the ordinary pub. doc, & then a revised edition
will be ordered in which our illustrations can come & any additional de-
scriptions & observations that may be ready. He says this is the only way,
or the Natural History may be thrown out altogether.”

The octavo first edition of the Pacific Railroad Reports appeared as House
Executive Document no. 129, 33rd Congress, Ist Session. Announced as
a three-volume work, only two volumes of text were published. There are
only two botanical reports in the octavo edition, Torrey’s catalogue of the
plants collected on the Pope Expedition and Bigelow’s account of the forest
trees and vegetation observed during the Whipple Expedition. These bo-
tanical reports were not illustrated. Bigelow’s report was reprinted ap-
parently without change, in the second, quarto, edition of the Reports.
The account of the plants collected on the Pope Expedition, a taxonomic
paper, was subsequently much changed in the second edition and merits
special comment.

The botanical report of the Pope Expedition appeared on pp. 307-324
of part 2, of volume 2, of the octavo edition. Althought Torrey wrote
Gray, on Nov. 4, 1854, just after he had sent the manuscript of this report
to Washington, that, “You have had a larger share in the Catalogue than
I ..,” the catalogue as published gives Torrey as sole author. This first
edition of the botanical report on the Pope plants has become a forgotten
item in the literature of West American botany. There is no copy of the
report at the Gray Herbarium, and Dr. H. W. Rickett (in lit. Aug., 1942)
writes that there is none in the library of the New York Botanical Garden.
In their later writings Torrey and Gray, apparently considering the octavo
report as a preliminary one and superseded by the enlarged and changed
quarto edition published about two years later, invariably cited only the
second edition of the Pope Report. Later botanists, unaware of the early
edition, have done the same. The first edition of the Pope Report can be
dated reasonably well. The publication of the octavo edition of the Pacific
Railroad Reports is noted in the American Journal of Science (20: 299)
for September 1855. Engelmann, however, saw the publication several
months earlier, for, writing Gray on June 12, 1855, he states, “I see in
some of the Pacific Railroad Reports Torrey has mentioned without de-
scribing several new Euphorbia coll. before by Wright or Fendler — going
ahead of me — but it serves me right.” The first edition of the Pope Report,
accordingly, must have appeared before the middle of 1855, probably in
the spring of that vear.

The following three specific names, none listed in Index Kewensis, were
published in the first edition of Pope’s Report and abandoned in the second
edition: Ehretia ? hispida, nomen (p. 320), Stegnocarpus ? Ctocarya
(p. 320), and Camassia Gawleri (p. 323). In the second edition they are
replaced by the following newly published names: Eddya his pidissima,

1943 | JOHNSTON, PACIFIC RAILROAD REPORTS 239

Stegnocarpus canescens, and Camassia Fraseri. The names Eritrichium
crassisepalum (p. 321), Euphorbia Wrightii (p. 321), Euphorbia dilatata
(p. 321), Euphorbia albomarginata (p. 321), and Euphorbia Fendleri
(p. 321), nomina nuda in the first edition, were supplied with descriptions
in the second edition. The well-known species Selenia dissecta (p. 308),
Stenandrium barbatum (p. 317), and Pentstemon Fendleri (p. 318) were
well-described in the first edition of the report. The species Astrophyllum
dumosum, Ammoselinum Popei, Phacelia Popet, Eritrichium pusillum, and
Ptilocalyx Greggii, published in the second edition, are not mentioned, at
least by name, in the first edition.

The second edition of the Pacific Railroad Survey Reports was published
in sumptuous quarto volumes. The text of the first edition was reprinted
with few changes, and many new special reports with many plates were
added. This second edition is the one represented in most libraries. It
appeared in two forms, differing only in title-page, as Senate Executive
Document no. 78, and as House Executive Document no. 91, both of the
33rd Congress, 2nd Session. As originally planned the work contained
eleven volumes. Subsequently two more volumes (numbered vol, 12, pt.
1, and 12, pt. 2) were added and published as House Executive Document
56, 36th Congress, Ist Session. In the completed work, botanical reports
are found in volumes no. 2, 4, 5, 6, 7, and 122.

Volume 2 of the quarto reports contains the botanical reports for the
Beckwith and the Gunnison Expeditions, and the second edition of the
botanical report for the Pope Expedition. These reports are by Torrey and
Gray. The complete volume was issued at Washington in 1857. Within
this volume Captain Pope’s “Explanatory Note to the Geological Report”
bears the printed date Feb. 18, 1857. Torrey, in a letter to Engelmann
dated July 30, 1857 (quoted by Rodgers in his book “John Torrey,” p.
248), speaks of the Beckwith and Pope reports as “contained in a volume
[ of the Pacific Railroad Reports] just published . .” On October 6G. 1857.
the volume had been received from Washington and accessioned at the
Philadeiphia Academy.

Torrey and Gray had reprints of the botanical reports published in vol. 2
before the end of June 1857. On Oct. 20, 1855, Torrey had written Gray
that the botany of the Beckwith (and Gunnison) report was printed and
that separates were ordered. In his letter of Jan. 9, 1856, he stated that
the botany of the Pope report was not yet printed, and on March 12, 1856,
he asked if Sprague was still at work on the plates for that report. The
authors seem to have received the printed plates for the Pope Report
shortly before Torrey’s letter of June 23,1857. At that time Torrey wrote,
“You probably rec’d from me, lately, a parcel containing 50 sets of plates
for Bot. Pope. You can return the 10 extra copies of the Beckwith letter-
press, or I will send you the plates for them — just as you like. My parcel
for England will go soon. A friend will take care of it. I will send Pope's
and Beckwith’s Rep. to Hooker, etc.” The botanical parts of the Pope and
Beckwith-Gunnison reports are apparently those acknowledged and com-
mented upon by Sir William Hooker in his letter to Gray, dated Nov. 27,

240 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

1857. These same reports (with complete text and plates) were reviewed
by Hooker in the final issue of the Kew Journal of Botany, 9: 376 (Dec.
1857). He states that the botany of the Beckwith-Gunnison reports was
“Published January 7, 1857.’ No such date is given for the botany of the
Pope Report. Torrey’s letter of June 23, 1857 (already quoted above)
does seem to infer that the complete botany of the Beckwith-Gunnison
report and at least the text of the botany of the Pope Report had been in
the hands of the authors for some time. While Hooker may have been
correct in stating that the botanical part of the Beckwith-Gunnison Report
had been “published January 7, 1857,” there is no evidence that it was
distributed by the authors until after June 1857, and little if at all before
the time when the whole of vol. 2 was available to the general public at
Washington. I believe that the effective dates of publication of the bo-
tanical reports for the Beckwith-Gunnison and Pope expeditions is either
June or early July, 1857. This date is two years later than the date of
publication for the botanical appendix in the first (octavo) edition of Pope’s

Volume 4 of the quarto reports contains the important botanical cat-
alogues by Torrey, Engelmann, Bigelow, and Sullivant, based upon
material assembled during Whipple’s Expedition.

On May 22, 1857, Torrey wrote Gray that the text of his part of the
Whipple Report was printed and that he was preparing the index, and on
July 22, 1857, that his extra copies were ready but the plates were still
unprinted. On Aug. 12, 1857, he wrote, ‘A day or two ago I was surprised
to get from Dr. Bigelow a printed copy of the Bot. of Whipple’s Exped.
containing our portion, together with the Cactaceae, an introductory article
on the Bot. Geography explored & a memoir on the principal forest trees
found on the route .. . I have 150 copies of the plates of Cactaceae, which
were intended for extra copies of text that Dr. B. promised, long ago, to
have struck off. I did not learn till yesterday that the lithographer had
printed these. My own extra copies of Bot. have not yet arrived, & I
rather think that they may include Bigelow’s articles.” On Aug. 22, 1857,
having just returned from Montreal, Torrey wrote, ‘““My extra copies of
Plant. Whipple have not arrived, but I found a single one (sent by Express )
on my table this morning. It contains Bigelow’s two reports & Sullivant’s
Mosses, the latter not in a previous copy sent two weeks ago. I am morti-
fied to find so many typographical errors . . . Perhaps the Superintendent
of Public Printing will authorize the insertion of the errata list in all copies.”
On Sept. 2, 1857, Torrey received word from Washington that 150 copies
of the botanical report had been shipped to him. These arrived by Sept.
10th, when he wrote, “They do not contain Bigelow’s article, Cactaceae,
nor Sullivant’s mosses. So if the Cactaceae have not been received for
Engelmann we must fall back on Bigelow for these, who has 200 copies of
the entire Botany. I have 200 copies of the Cact. plates . .. We can dis-
tribute our part of the Botany without Bigelow’s, & your 50 copies shall
be sent as soon as I can get them ready — but you had better not distribute
till we get errata printed.” Engelmann, travelling in Europe, did not get

1943 ] JOHNSTON, PACIFIC RAILROAD REPORTS 241

his reprints until May, 1858. C. W. Short (Louisville, Kentucky) wrote
Gray on Sept. 10, 1857, “Mr. Sullivant has been so good as to send me a
copy of his very beautiful Mosses of Whipple’s survey.’ Although the
volume of the Pacific Railroad Reports containing the Whipple Reports
was only received at the Boston Society of Natural History on Jan. 1, 1858,
at the Philadelphia Academy on Jan. 5, 1858, and at the St. Louis Academy
on March 8, 1858, and was first reviewed in the American Journal of Science
(25: 149) for Jan., 1858, the evidence is clear that Torrey, Gray, Bigelow,
and Sullivant had sufficient copies of the botanical portions of the Whipple
Report to establish Sept. 1857 as the effective date of issuance for this
important botanical volume.

Volume 5 of the quarto reports contains the two botanical reports based
on the collections of the Williamson Expeditions, one by Torrey, the other
by Durand & Hilgard. The substance of the Durand & Hilgard report first
appeared as “Plantae Heermannianae”’ in the Journal of the Philadelphia
Academy 3: 37-46 (Nov. 1854). <A reprint of this article was sent Gray
by Durand with a covering letter dated Dec. 4, 1854. The revised report
on the Heermann collections was in print at Washington before Sept. 2,
1857, for on that date Torrey wrote that, by some mistake, he had received
150 copies of Durand & Hilgard’s report intended for the authors.

Torrey’s report, on the plants collected by Blake during the Williamson
Expeditions, was probably printed much later. Concerning this report there
is only one reference in his letters which may be significant. On June 12,
1858, he wrote Gray, “As to those plates of Blake and Antisell’s Repts. I
don’t mean to let you pay for any unless I learn that some extra copies of
the letter press can be obtained from the Public Printer.” The whole
volume, containing Torrey’s report, was displayed at the St. Louis Academy
on March 22, 1858. There is no reason for believing that Torrey received
advance reprints. The effective date of Torrey’s report is probably the
date of issuance of the complete volume in Washington, about Feb., 1858.

Volume 6 of the quarto reports contains the botanical reports resulting
from the Williamson & Abbot Expeditions in 1855. The botanical reports
were organized and partially written by J.S. Newberry. On July 23, 1857,
Newberry, in ill health, wrote Gray, asking him to read all the proofs of the
botanical report which would soon become available. On Debte- 2, 1057,
after learning that Newberry was to join a new expedition in the West,
Torrey wrote Gray, “What is to be done with Newberry’s Report? If he
goes with Ives you will probably attend to proof reading...” On April
5, 1858, the completed volume was displayed at the session of the St.
Louis Academy. The volume probably first appeared in Washington in
March, 1858.

Volume 7 of the quarto reports contains Torrey’s report on the collections
of Parke’s Expedition. On Sept. 10, 1857, Torrey wrote, “Two days ago I
rec'd proofsheets of a small report that I prepared for Antisell (Parke’s
Exped.). An officer saw them & begged .. . that I might see them before
being worked off. They permitted him to do so but said, if the sheets were
not returned by next mail, the printer should proceed without corrections.

242 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

Part of the work had already been printed & I did not know that it was
in press!” Torrey’s letter of June 12, 1858, already quoted in connection
with his account of Blake’s plants (vol. 5), is the only other reference to
this report in his letters to Gray. F. W. Vaughn, writing for A. A. Hum-
phreys of the Office of the Pacific Railroad Surveys, wrote Gray on May 8,
1858, that volumes 5, 6, and 7 of the reports were being sent him. Volume
7 was available at the Philadelphia Academy May 11, 1858, and was dis-
played at the St. Louis Academy on May 17, 1858. The volume probably
came from the press in Washington in April 1858.

Volume 12, part 2, of the quarto reports contained the final expanded
report of Stevens’ explorations across the northern United States. The
botanical papers it contains are written by Gray and Cooper. I have no
detailed information concerning this report. ‘The publication of the report
was authorized by the U. S. House of Representatives on March 25, 1860.
The volume was received at the Boston Society of Natural History on Feb.
12, 1861. It was displayed at the St. Louis Academy on June 17, 1861.
The Library of Harvard University did not receive its copy until Aug. 19,
1861. The volume, accordingly, was published probably late in 1860 or
in Jan. 1861.

The dates of publication for the botanical papers published in the two
editions of the Pacific Railroad Reports may be summarized as follows:

First Epition (in octavo)

Volume 2: Botany of the Pope Report, by Torrey before June 1855

SECOND EpiTIon (in quarto)

Volume 2: Botany of Beckwith and Gunnison Re-
ports, by Torrey & Gray about the middle of 1857
Botany of Pope Report (2nd edition), by
Torrey & Gray about the middle of 1857
Volume 4: Botany of the Whipple Report advance reprints Sept. 1857
Volume 5: Botany of Williamson Report: Heermann
collections, by Durand & Hilgard advance reprints Sept. 1857
Blake collections, by Torrey about Feb. 1858
Volume 6: Botany of Williamson & Abbot Report about March 1858
Volume 7: Botany of Parke Report, by Torrey about April 1858
Volume 127: Botany of Stevens Report Jan. 1861 or late 1860

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXIV JULY, 1943 NuMBER 3

A SECOND SUMMARY OF THE SCROPHULARIACEAE
OF NEW GUINEA*

FrRANcisS W. PENNELL

With five plates

Tue third series of Archbold Scrophulariaceae from New Guinea differs
from its predecessors in coming from the western or Netherlands half of the
island.!. The specimens have again been gathered by Mr. L. J. Brass, but
this time some were collected jointly with Dr. E. Myer-Drees, a forester
who accompanied the expedition on behalf of the Netherlands government.
The expedition entered from the northern coast and climbed Mt. Wil-
helmina in the Snow Mountains. The Scrophulariaceae found at high alti-
tudes, njostly species of Hebe and Euphrasia, seem to be wholly endemic,
and show what a break must occur in the highest zones of vegetation
between this mountain and the previously visited Carstensz Peak in the
Nassau Range, several degrees to the westward.*

Since political conditions will probably prohibit for some years further
botanical exploration of the mountains of New Guinea, it seems appropriate
to present now a summary of our present knowledge of the Scrophulariaceae
of the island. It is nineteen years since the last such record, “Die
Scrophulariaceen Papuasiens,’’® by Dr. R. Schlechter, a study enumerating
26 species in all, 19 of which were lowland and 7 alpine. With continued
exploration, figures for both the lowland and the highland species have
increased, but the lowland mainly by the addition of widespread Indo-
Malayan plants, while the highland count builds up by endemic restricted
species. The figures now are 34 for the lowland and 25 for the highland,
making a present total of 59 species. The sharp separation of lowland

*Botanical Results of the Richard Archbold Expedition.

1For account of the Third Archbold Expedition, which visited New Guinea in 1938,
see Richard Archbold’s “Unknown New Guinea” in Nat. Geogr. Mag. 79: 315-344.
1941.

“See “Report on the Botany of the Wollaston Expedition to re as Guinea,
1912-13,” by Henry N. Ridley, in Trans. Linn. Soc. II. Bot. 9: 1-284.
3In Bot. Jahrb. 59: 99-117. 1924.

244 aeNehy OF THE ARNOLD ARBORETUM [ VOL. XXIV

and highland floras for this family in New Guinea, first pointed out by
Schlechter and discussed in my reports on the Scrophulariaceae of the First
and Second Archbold Expeditions," is made more emphatic as collections
accumulate. In this family it is a tribal rather than a generic matter, the
Gratioleae and Buchnereae being lowland, and the Veroniceae and Euphra-
sieae alpine.

To make the present summary as complete as possible, it wiil be necessary
to consider again some problems based upon the earlier expeditions and
therefore concerning plants of eastern New Guinea. Also, there have
reached me, either directly or through the Arnold Arboretum, various
collections made by Mrs. Mary Strong Clemens in the mountains of the
Morobe District or the Saruwaket Range, both being in Northeastern New
Guinea. These will also be reported in the present paper

But this material, altogether, does not contain many collections: and it
would have been indeed desirable to have seen what specimens are in the
herbarium at Buitenzorg, Java, and in the herbaria of Germany and
England. But, even with those, any enumeration now is but a milestone
directing to the task yet before us in adequately collecting for study the
New Guinean flora. There are still undescribed lowland oe
though these will be vastly outnumbered by the alpine species of Hebe a
Euphrasia certain to be revealed as expeditions reach more isolated moun-
tain areas.

in previous papers, the present summary will present keys when
Paley needed. AddiGons to the island’s flora will be starred. It hardly
seems necessary to explain the obvious abbreviations to herbaria: AA —
Arnold Arboretum; ANSP — Academy of Natural Sciences of Philadel-
phia; GH — Gray Herbarium: Mich — University of Michigan.

Kry TO THE GENERA IN NEW GUINEA

A. Corolla with the upper lobes external, overlapping in the bud; stigmas distinct ;
leaves opposite or ternate, often glandular-punctate ; lowland. (Antirrhinoideae-
tigrraee .

_ Anther-cells separated on short arms of the connective ; leaf-blades glandular-

punctate.

C. Anthers all developed, polleniferous; capsule globose to ovoid; sepals alike ;
bracts leaf-like, the inflorescence foliose.

D. Bracteoles none; sepals distinct ; capsule depressed-globose, equally oi
ticidal and loculicidal; seeds longitudinally lined; corolla 4— Ay
leaf-blades ovate, petioled...........0 00. eee eee eee ees . Poa

DD. Bracteoles developed, a pair beneath the calyx; sepals foien proxim mide
ee globose to ovoid, primarily loculicidal ; seeds vestigially reticulate

mooth; corolla 7-15 mm. long; leaf-blades various. .2. Li imno phila.

CC. See. eee developed, one cell of those on longer pair of filaments
abortive; capsule conic, twice as long as wide, septicidally and then locu-

spikes or spike-like racemesS............. 250-0 eeeereeeess acu
BB. pees cells contiguous.

4First Expedition reported in Brittonia 2: 177-188. 1936, and 3:95. 1938; Second
Expedition in Jour. Arnold Arb. 20: 75-84. 193

1943] PENNELL, SCROPHULARIACEAE OF NEW GUINEA 245

C. el spanen developed, a pair at base of the unequal sepals; corolla white,

with violet lines; anther-cells parallel; capsule about equally septicidal and

loculicidal : leaf-blades glandular-punctate...................... ella.

CC. Bracteoles none; corolla more or less violet-blue or purplish; ant ther- ree
ee leaf-blades not (or in some species of Lindernia slightly

puncta

D. ee filaments straight, from their lowest contact free from the

corolla- sai capsule loculicidal; sepals united proximally; bracts

SUISUN UCN tear ner nte tre ns <2." 5, chin ily tere tae rie eee

azus.
DD. pens filaments adnate to corolla as antero-lateral ridges, distally

free and each sharply bent back upon itself

(this appearing as a knob-
like process),

thence abruptly ascending; capsule aaa
; capsule . long,

O mm. long, ie purple ; ee sessile,

bare lateral nate persists as a thi

to linear-cylindric; corolla smaller; leaves petioled or sessile, the

united pala plane, the calyx, if as long
in ; bracts all pina usually leaf-

, ODO) eran oicheaye aes 7. Lindernia.

FF. Sepals united nearly pegs, the calyx exceeding the capsule
a

racts foliose or more usually subulate,
the pedicels (by eee vi of internodes) sometimes in fours.

AA. Corolla with the lower or lateral lobes external, ov iscsi: in the bid
wholly united, capitate or punctiform; leaf-blades not punctate. aiehineutholtens),
B. Upper lobes of eee simply spreading, distinct or united;

anthers all dis-
tinct, glabrous med; — many, wingless
C. Sepals pecans eee two, t ostero-lateral pair alone developed;
anther-cells equal;

corolla ‘hula
smooth; pedicels not bracteolate; alpine. (Veroniceae).
DD: ee 25-28 mm. long, the lobes five, much shorter than the tube;
owers solitary, sessile; stems woody................... . Detzneria.
DD. Corolla ami the lobes (by of the posterior) appearing four,
; flowers racemose, pedicelled.
s woody, erect or diffuse...10. Hebe.
stems herbaceous, prostrate

CC. Sepals united at least half their length;

rotate; seeds flattened,

E. Capsule turgid, septicidal ; rei
EE. Capsule flattened, loculicidal; stems herbaceous, prostrate........
. Veronica.
; stamens four, both pairs a ieoued:
anther-cells not equal; corolla campanulate or salverform, the tu
than the five distinct lobes; capsule loculicidal; seeds reticulate; Seite
bibracteolate; Jowland. (Buchnereae).

Corolla campanulate, yellow;
tubular

elongat

anther-cells two, unequal; calyx not
EB. Calyx symmetrical, externally glabrate, internally densely and softly
hairy. the hairs appearing as a fine cilia ation o
lobes; corolla 5-6 mm. long, sprea i
long, wholly loculicidal; leaves linear, entire or with a pair of slender r
DS rete eae agree ses aa ces ovr ee sa canch Hearne Re ee 12. Sopubia.
EE. Calyx asymmetrical. deeply cleft anteriorly and distally so upcurved
that the short free lobes appear as tips to upper side of corolla and
pea externally densely hispid, seeicunially glabrous; corolla 17-20
mm. long, more narrowly campanulat

ate; capsule 7-8 mm. long, sec-
ee also somewhat septicidal;

leaves linear- ans entire.
Bene ASE Aer ARTO He UY Peale er, Gecartand Sercabinns seinen SR ane Centranthera.

246 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIV

DD. Corolla capeiiaiaes the narrow throat hairy within; anther-cell one, the
other lost by abortion; calyx tubular, more or less ribb
E. Corolla-tube straight, the corolla pink, bluish or white; ribs of calyx
relatively faint; leaf-blades narrowly elliptic to linear-lanceolate. .
E Gace Blnaty Sey © PEaeR Re Ae Ee eae Ue ba ween ae eS 14. Buc enn:

15. .
BB. Upper lobes of corolla united to form a concave galea, the white or violet
corolla strongly zygomorphic; anthers cohering or distinct, the cells hairy
and unequal, awned; capsule loculicidal; seeds few, with slightly raised longi-
tudinal wings (or wing-like ridges) ; alpine. (Gicheusens) ....16. Euphrasia.
*1. Poarium Desvaux
(Genotype, P. veronicoides Desv., the only original species)

This generic name, appearing in Hamilton’s Prodromus Plantarum Indiae
Occidentalis, p. 46, 1825, is the oldest applicable to the group to which I
have applied the recent name Lendneria Minod of 1918. It is a genus of
some 20 species, distinguishable from Stemodia L., also neotropical, by
seeds lined or ridged instead of vestigially reticulate, capsules thinner-
‘walled, bracteoles wanting instead of present beneath sepals, leaf-blades
petioled instead of clasping, and usually (but not constantly) by corolla-
throat hairy within on the upper instead of the lower side. The following
much-described species, a widespread tropical weed, proves to be also in
New Guinea.

*1. Poarium verticillatum (Miller) a »
Erinus verticillatus Miller, Gard. Dic oe 1768. Collected . Vera Cruz
n William Houstoun; ae seen at see Museum in 0.

This was later described as Capraria humilis Ait. (1789), as collected in
the East Indies by J. G. Koenig; again as Stemodia parviflora Ait. (1812),
its best known name and actually based on Erinus verticillatus ; and finally
as Poarium veronicoides Desv. from Hispaniola, the type of which, in the
Museum d’Histoire Naturelle at ike I saw in 1930.

NorTHEASTERN NEW GUINEA: robe District, Lae, Mary S. Clemens 10447
(AA) ; cae Station, mission Bist Mary S. Clemens on May 29, 1940 (ANSP).

2. Limnophila R. Brown

Schlechter (in Bot. Jahrb. 59: 100. 1924) credits four species to New
Guinea and the Bismarck Archipelago, distinguishing them by key and
giving their occurrence. These are all Indo-Malayan and are presumably
immigrants to New Guinea from farther west. Two, L. aromatica (Lam.)
Merr. and L. fragrans (Forst.) Seem., have been already reported from
earlier Archbold Expeditions in Brittonia 2: 181. 1936, and in Jour. Arnold
Arb. 20: 76. 1939. Of these L. aromatica appears to occur at low alti-
tudes throughout the island, and it was collected in 1938 on the Balim
River (Brass 11827). L. rugosa (Roth) Merr., probably equally wide-
spread, has been gathered at Hombron Bluff, British New Guinea, by
L. J. Brass (no. 1651), and at the Kajabit Mission, Morobe District, North-
eastern New Guipea, by M. S. Clemens (no. 10437K), the latter getting

1943 | PENNELL, SCROPHULARIACEAE OF NEW GUINEA 247

it from “girls who prize the fragrant leaves,’ and the former stating that
the natives carry it in their armbands, using it at dances and feasts. Brass
reports it as smelling “‘strongly of aniseed” and describes the flower as
with “lavender blue mouth, tube yellow.”’ The remaining species is the
widespread oriental aquatic, L. sessiliflora (Vahl) Blume.

3. Adenosma R. Brown

Schlechter (1. c. 102) credits two species to New Guinea, A. javanicum
(Bl.) Koord. and A. papuanum Schlechter, to which in 1939 I added A.
ternatum and now propose A. punctatum. The first species, an Indo-
Malayan plant remarkable for the resemblance of its calyx to the Bacopa
alliance of neotropical America, is only remotely related to the others. The
four may be distinguished as follows:

A. Sepals hirsute, not reticulate, ae outer lanceolate to oblong-lanceolate, 2 to 4
times the width of the inner two; flowers in definite spikes, the bracts much
smaller than the leaves; capsule ae on septal line; leaf-blades dentate; stem
erect.

B. Root perennial; stem hirsute, not glandular.
C. Leaves ternate, the blades minutely glandular-punctate beneath, 2—2.5 cm.
long; corolla 10 mm. long, externally glabrous; capsule 7-8 mm. long....
Reh MeO GEN eo 868 AE EMS PS a hs cans eaves os GSR uO ROE RRR le gAe
CC. Leaves opposite, the blades with larger glandular dots, 1.2-1.5 cm. long;
corolla 7-8 mm. long, externally finely pubescent; capsule 5 mm. long..
Ss uate ene Nain Rapti eBccaniata tet Ss) a) aotibh.ancdiva ob ah aahaar aaa ue ayaa ye Raa Aga Ze A punctatum.
BB. Root annual; stem finely pilose with glandular hairs; jeatepiades 1.8-2.9 cm.
long; corolla 12 mm. long, externally finely glandular-pubescent...........
Shey Re eal eee end ASR Si 5) ce dodel as ceed NOt RE apuanum.,
AA. pee finely pubescent, strongly reticulate, the outer widely ovate- cordate, many
width of the inner two; flowers isolated in axils of the functioning
eee capsule not sulcate over septum; leaf-blades crenate; stem extensively
Fe) 039 0 Career Tene (arr OA Irae ae a th Fe AAS
1. Adenosma ternatum Pennell in Jour. Arnold Arb. 20: 77, 1939.

Collected by L. J. Brass (no. 7816) on the Fly River in British New

yuinea

*2. Adenosma punctatum sp. nov.
Herba perennis, 3—4 dm. alta, ramosa; caulis et rami teretiusculi, hir-
ee eglandulosi; folia opposita, petiolata, ovata, obtusa, basi rotundato-
neata, margine dentata, utrinque hirsuto- pubescentia, laminis 1.2—1.5
cm. longis, 9-12 mm, latis, petiolo 5-10 mm. longo; flores yes plus
minusve remoti, alii in racemum spiciformem congesti; bracteolae duae,
lineares, calyce ‘breviores: sepala inaequalia, lanceolata vel ae ee
lata, 5-6 mm. longa; corolla 7-8 mm. longa, tubulosa, bilabiata, extus
minute pubescens, labio supero apice retuso, labio infero hirsuto, trilobato,
lobis retusis; stamina glabra, connectivis globosis, antherarum superarum
cellulis aequalibus, sed inferarum uno rudimento inaequalibus; capsula
5 mm. longa, nigra, latere ad septum sulcata, loculicida et tardius etiam

oe semina 0.3 mm. longa, ovoidea, minute lineat

t seen, but aiecae bases with several stems ascending there-
from ni ote a perennial duration. Ste m 3-4 ee tall or more, branched
(many branches abbreviated, forming eile leafy facies but the upper

248 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIV

branches ascending-spreading). Stem and branches slightly angled, laxly
foliate. Leaves opposite, petiolate, the blades ovate, obtuse, dentate, above
pilose-hirsute, beneath paler and glandular-punctate, hirsute on the larger
veins, 1.2—1.5 cm. long, 9-12 mm. wide, at base widely cuneate or rounded
to hirsute petioles 5-10 mm. long. Bracts smaller, the upper nearly
spatulate and not exceeding the sepals. Lower flowers few and isolated, the
upper numerous and forming short spike-like racemes. Lower pedicels
becoming 2 mm. long, the upper shorter. Bracteoles linear, shorter than
the calyx. Sepals unequal, lanceolate, the inner nearly linear, all obtuse,
5-6 mm. long, hirsute, the uppermost longest and widest, oblong-lanceolate.
Corolla 7-8 mm. long, tubular-bilabiate (without record of color), exter-
nally finely pubescent; upper lip projecting-arched, its lobes rounded,
scarcely distinct at the retuse apex; lower lip 2-ridged and pilose-hirsute
thereon, distally with three retuse lobes. Filaments glabrous, the anterior
pair longer; connective knob-like; postero-laterals with two nearly equal
anther-cells, antero-laterals with one anther-cell developed, the other rudi-
mentary. Stigmas wide, apparently plate-like, the apex of the fused styles
thick-winged. Capsule 5 mm. long, conic, distally attenuate, black, firm,
laterally furrowed on septum, loculicidal and secondarily somewhat sep-
ticidal. Seeds 0.3 mm. long, ovoid, brown, minutely lined.

NorTHEASTERN New GutInEA: Morobe District, Malolo Mountains, alt. 240-270
m., Mary S. Clemens 4359 (AA, TYPE), November 1936, in late flower an it

The name refers to the strong dotting of the leaves, but there are other
species, especially A. glutinosum (L.) Druce, that show it as pronouncedly.
3. Adenosma papuanum Schlechter in Bot. Jahrb, 59: 103. 1924.

Northeastern New Guinea.
4. Adenosma javanicum (Blume) Koorders. See Schlechter, ]. ¢. 102. 1924.

Northeastern New Guinea.

4. Mella Vandelli

Differs from Bacopa Aubl. (1775) and Herpestis Gaertn. fil. (1807) in
definitely bilabiate corollas, pinnately veined leaf-blades, and erect habit.
I have previously used for this dominantly neotropical genus the name
Caconapea Cham. (1833), the earliest typifiable generic name, since Mella
(Vand. FI. Lusit. et Bras. 43. 1788) appeared without citation of species.
The latter would therefore be invalid under the American Code but valid
under the new International Rules of Botanical Nomenclature. There are
a few oriental species, the following extending to New Guinea and northern
Australia.

1. Mella floribunda (R. Br.) comb. nov.

Herpestis floribunda R. Br. Prodr. Fl. Nov. Holl. 1: 442. 1810.

Originally described from tropical northeastern Australia. Collected in
British New Guinea by L. J. Brass, no. 3730, and reported as Bacopa
floribunda in Brittonia 2: 181. 1936.

*5. Mazus Loureiro

Both Brass’ and Mrs. Clemens’ collections bring this genus into the flora
of New Guinea, its representative being the common tropical weed that has

1943 ] PENNELL, SCROPHULARIACEAE OF NEW GUINEA 249

been included in M. japonicus (Thunb.) Kuntze, but which seems to be

separable by the following contrast:
A. Leaf-blades widely oblong to nearly oval, cuneately narrowed to petioles often
as long as the blades; lower pedicels usually longer than the calyces; plants
(stems, leaves, pedicels, and calyces) minutely, and scarcely or ae Lp aeeriee
pubescent, or (especially the leaves) glabrous.................45. 1. japonicus.

. Leaf-blades oblanceolate to narrowly obovate, primarily runcinate- aor attenu-
ate to ill-defined petioles; lower pedicels usua a about equaling ay calyces; plants
hirsute-pubescent, the hairs largely glandiferous............... . goodenifolius.

ca
>

Mazus japonicus, described from oad is ue a a of warm tem-
perate climates, and has become adventive to the New World. M. goodent-
folius is more definitely tropical and oriental. I had suspected that it
would prove to be Loureiro’s rugosus from Cochinchina, but his description
of the leaf-shape precludes that possibility.

*1. Mazus goodenifolius (Hornemann) comb. n
Gratiola goodenifolia Hornemann, Enum. PI. Han 19. 1807. Original pee
ot seen, but its diagnosis? quoted in Willd. . Pl. Hort. Berol.
where it is included as a synonym of ipa ne Willd., there ieee
I suspect from the alteration of Willdenow's primary
diagnosis to “H. foliis obovatis basi integerrimis, calycibus patulibus pedunculisque
glabris,” that the latter was based primarily upon material of M. japonicus, since
the omitted account of an incised-serrate narrow leaf-blade would denote the

only be to M. japonicus. In both the corolla would be bicolorous, the upper lip
blue, the lower white.

An oriental weed, widespread over New Guinea. Collected in 1938 by
L. J. Brass in Netherlands New Guinea, no. 11403 from Bele River, north-
east of Lake Habbema, and no. 12925 from Bernhard Camp on the Idenburg
River. Also gathered by Mary S. Clemens in the Morobe District ot
Northeastern New Guinea, no. 4975 in 1937 from Ogeramnang, and
no. 11286F in 1940 from Matap.

6. Artanema D. Don
(Genotype, Torenia fimbriata Hook., from Australia.)
1. Artanema longifolium (L.) Vatke.

Earlier than my record (in Jour. Arnold Arb, 20: 78. 1939), this wide-
ranging Indo-Malayan plant had been gathered by Branderhorst in 1907
at Merauke in southeastern Netherlands New Guinea and reported by Diels
and Lanjouw in Nova Guinea Bot. 14: 534. 1928

7. Lindernia Allioni

With the including of //ysanthes Raf. and the arising of several problems
of interpretation, it seems desirable to consider this genus rather fully.
The following key distinguishes all species that I have seen from New
Guinea.

A. Capsule globose to ee hig not or only slightly exceeding the calyx,

mesa filaments bearing a
B. Sepals united about 24 ae ie PlaDrOUSs caressa aeresus 1. L. crustaceu.
5“Gratiola Neca ia foliis obovato-lanceolatis inciso-serratis basi integerrimis.
racemis terminalib

250

JOURNAL OF THE ARNOLD ARBORETUM [vOL, xxiv

BB. Sepals distinct nearly to bas

s oval to van crenate-serrate to nearly entire, rounded or sub-
cordate at base; bracts oval, the flowers axillary; capsule semiglobose ;
corolla 9-10 mm. long; sepals and angles of stem hirsute with long non-
glandular hairs, the divaricate pedicels nearly glabrous........ 2. L. hirta.

CC. Leaf-blades ovate to nearly oblong, serrate-dentate; bracts subulate, the
rs in evident racemes; capsule ellipsoid-ovoid to ovoid; corolla, 3-5
mm, long; pedicels glandular-puberulent or -pubescen ae

D. Sepals hirsute with mostly glandless hairs; entire surface of stem and
leaves hirsute with spreading glandless hairs, the main blades usually 3-6
cm. cra and narrowed at base, only the uppermost rounded-cordate

ANC, SCSSUCl 6 co 55 65 545 hale ete bea ee ee alee aee 3. -s vi
DD. Sepals peeeee glandular-puberulent to glabrous;

wi

are r (1-2 cm. ie only the lower narrowed at base
unded to sessile bases............... cece eceecveees 4. L.

AA. aie cae longer than wide, much exceeding the sepals, which are distinct

arly to base.

B Leaf-blades sessile or oo corolla 5-10 mm. long; stems ascending or
erect, often rooting at low es
C. Blades ovate to at ea crenate, sage or cordate at base; anterior

. lon 5. L. cordifolia.

filaments anther-bearing ; osteila a whale acaie ba =e
anterior fila-

. Blades oblong to oblanceolate, eens narrowed at base;
ments without anthers; corolla smaller

D. ee of as satiny: to dentate, the teeth appressed and not pro-

longed; corolla Win 5 VOT Sooty ask aa ok bo 6. hk ws ae 6. L. anagallis.

DD. en of ee cea and seed serrate, the teeth spreading and

ed into setaceous tips; corolla 5 mm. long......... . L. ciliata.

BB. Leaf- blades ane? elliptic, sharply serrate, but the teeth not setaceous-

tipped; corolla 13-15 mm. long; anterior filaments ee “arapee stems

extensively aL a flowering portions abruptly erect...... . L. antipoda.

QO
o)

1. Lindernia crustacea (L.) F. v. Muell.

This widespread tropical weed, cited by Schlechter (1. c. 106) and me

(in Brittonia 2: 181. 1936, in Jour. Arnold Arb. 20: 79. 1939) from many

localities throughout New Guinea, is among
tions from }

Mrs. Clemens’ recent collec-
Morobe; there are also several collections made by Netherlands

botanists in western New Guinea at the Gray Herbarium.

. Lindernia hirta (Cham. & Schlecht. :, comb, nov,

Gratiola pusilla Willd. cur. L. Sp. Plant. ed. 4. 1: 105. 1797. ‘Habitat in ae ad
scaturigines.” D ve of a sa plant of the species now consider Basis
of Lindernia pusilla Willd.) Schlechter in Bot. Jahrb. 59: 106. on “unfor-
tunately preoccupied by L. pusilla (Thunb.) Merr. in Philip. Jour. Sci. Bot. 11:
312. 1916 [1917]. Although intended for the present species and citing Gratiola
pusilla Willd. as a synonym, the latter was based upon Selago pusilla Thunb.
ieee Plant. Cap. 99. 1794-1800,” the part containing: p. 99 actually published
in 1800; Thunberg’s brief diagnosis “‘S. foliis ovatis serratis pilosis, flore terminali
subsolitario” is impossible for our plant, as this has leaf-blades crenate to entire

and an inflorescence of several pairs of long-pediceled flowers, while its occurrence

is far from South Africa
Torenia hirta Cham. & Sctiles
collected by Chamisso & Eschscholtz on the Romanzoff Expedition.
ies now considered.
—— scabra Benth. Scroph. Indicae 36. 1835. The name long applied to this

ht. in Linnaea 2: 571. 1827. “E. Luconia retulimus’”’;
Clearly the

1943 ] PENNELL, SCROPHULARIACEAE OF NEW GUINEA Zl

Recorded by Schlechter (1. c. 107) from the Bismarck Archipelago, and
recently collected at Boana, Morobe District, Northeastern New Guinea, by
ary S. Clemens, no. 41468a (Mich) ; also at Balim River, Netherlands

Ne Guinea, by L. J. Brass, no. 71821 (AA, ANSP).

¥3: Tes viscosa (Hornemann) Boldingh, es mnigineaiigenilit Java 165. 1916,
omen; Merrill, Enum. Philip. Fl. Pl. 3: 4 1923, cum syn
Gratiote viscosa Hornemann, Enum. PI. Hort. po 19. 1807. Descriptio seen as
quoted in account of Hornemannia viscosa (Hornem.) Willd. Enum. or
Berol. 654, 1809, where its habitat was given as “in India orientali.” ieee

Vandellia hirsuta Buch. Ham. ex Benth. — Indicae 36. 1835. The specific epi-
het long applied to this species, and m rare at in its allusion to the
copious hairiness rather than to the yp Sie glandular
A common Indo-Malayan species occurring east to western New Guinea,
as shown by specimens in the Gray Herbarium from Manokoeari (R. F.
Janowsky 502) and Piorienbivak near the Mamberamo River (H. J. Lam
509), Netherlands New Guinea.
*4. Lindernia papuana sp. nov. Pirate I, A.

Herba annua, 0.5-1 dm. longa, ramosissima; caulis quadrangulatus,
angulis hispidulus; folia oblongo-ovata, acuta, serrato-dentata, glabrata
(subtus proxime venis hispidula), laminis 1-2 cm. longis, 6-8 mm. latis,

6In Boldingh’s Coane voor de Landbouwstreken op Java,” Batavia, i-viii. 1-204,
October, 1916, a number of , binomials appear in his key to the families, genera,
and species but in no case are dies names indicated as new, nor in any case is a name-
bringing synonym given; they are merely followed by the abbreviation “Bold.,” and

“Bloeiwijze bijna geheel zonder bladen.................... Lindernia viscosa Bold.

Bloemen in de bladokzels van goed ontwikkelde bladen....Lindernia pusilla Bold.”

In so far as Boldingh’s binomials were actually new, they were entered in Index

Kewensis, Suppl. 9, 1935. Those in the Scrophulariaceae are Jlysanthes procumbens
Bold. p. 164, and Lindernia viscosa Bold., L. pusilla Bold., Microcarpaea minima Bol

and Stemodia verticillata Bold. p. 165. It is manifest that Boldingh attempted ie

if these Boldingh names of 1916 should be accepted as validly published, due to their
presentation in a most unorthodox manner. His method of publication transcends the
provision of the International sai of Botanical Nomenclature that requires, in trans-
ferring specific names from one genus to another, that the name-bringing synonym must
be cited. They may be ete as illegitimate names, but in view of the fact that
ae are now listed in Index Kewensis, Suppl. 9, 1938, sometimes with the name-bring-
ing synonym added, it does seem desirable to accept them and to complete the record
as has been done in hans ase. To complete the record in all cases will be difficult, for
in the case of Lindernia oere Bold., who can say whether it was his intent to base
this on Gratiola ete Willd. (1797), or on Selago pusilla Thunb. (1800), or on both?
The Willdenow and Thunberg entities are actually two entirely different species; the
first is a Lindernia, but the second apparently does not belong in this genus. As pub-
lished by me in 1916 [1917], Lindernia pusilla Merr. must be typified by Selago pusilla
Thunb., and the entry in Index Kewensis, Suppl. 9, 1938, for the pe aecuae Boldingh-
Merrill transfers cites Thunberg’s name as the one synonym. — E. D. Mer

252 JOURNAL OF THE ARNOLD ARBORETUM (VOL, XXIV

basi rotundatis, sessilibus; racemus terminalis, florum 3-7 paribus com-
positus, sparse glandulari- puberulentus: bracteae subulatae; pedicelli 5-8
mm. longi, adscendentes; sepala 2.5 mm. longa, lineari-lanceolata, acuta,
glabrata; corolla 3-4 mm. longa, alba, glabra, labio supero rotundato fere
integro, labio infero trilobato, lobis triangulari-ovatis; stamina glabra,
omnia antherifera; capsula 2.5 mm. longa, ellipsoideo-ovoidea, glabra;
semina 0.25 mm. longa, flava, cylindrica, truncata, obsolete reticulata.

Low much-branched herb. Stems diffusely ascending, 0.5—1 dm. long,
four-angled, the angles slightly winged and bearing ascending short stiff
glandless hairs. Leaf-blades oblong-ovate, serrate-dentate, obscurely and
callosely ciliate, glabrate, beneath slightly paler and proximally with veins
somewhat stiff-pubescent, mostly 1-2 cm. long, 6-8 mm. wide, rounded to
sessile bases, only the lower attenuate proximally. Inflorescence a raceme
of 3 to 7 pairs of flowers, the rachis and pedicels sparsely glandular-
puberulent, the latter ascending, 5—8 mm. long; bracts (above the short
wide lowermost one) subulate, 0.5-1.5 mm. long. Sepals 2.5 mm. long,
linear-lanceolate, acute, minutely and obscurely pilose to glabrous. Corolla
3-4 mm. long, whit te, glabrous, the tube slightly widened horizontally,
anteriorly 2-ridged; upper lip scarcely retuse at the truncately rounded
apex, the lower slightly longer, the lobes triangular-ovate, widely deflexed-
spreading. All filaments antheriferous, the posterior short, ascending, the
anterior forming by their attachment the antero-lateral ridges of the corolla
that distally project as minute rounded knobs, the free distal portions of
the filaments arched-ascending. Capsule 2.5 mm. long, ellipsoid-ovoid,
glabrous. Seeds 0.25 mm. long, yellow, cylindric, truncate, obsoletely
reticulate.

NORTHEASTERN NEw GuINEA: Morobe District, Kajabit Mission, alt. about 300 n
Mary S. Clemens 10907 (ANSP, type, AA), Dec. 25, 1939, in flower and fruit alee
loose spreading mats along margin of slow rivulet). NETHERLANDS NEW rts
Near Pioniorbivak, alt. 60 m., region of Mamberamo River, H. J. Lam 511 (GH

In Schlechter’s account this was called Lindernia trichotoma nae
Schlechter, based upon Tittmannia trichotoma Benth. in Wall. Numer. List
no. 3943. 1831, a nomen nudum. That would be identical with the Indian
species long known as Vandellia multiflora (Roxb.) G. Don, based upon
Torenia multiflora Roxb. Fl. Ind. 3: 96. 1832, a plant described as with
smooth stem and merely subserrate leaves, agreeing with J. D. Hooker’s
statement that the leaves of the Bengal plant are entire or sinuate. Evi-
dently the Papuan plant, with its stiffly hairy stem and rather sharply
serrate-dentate leaves, cannot be this, nor do I identify it with any other
known species.

5. Lindernia cordifolia (Colsmann) ag
Gratiola cordifolia Colsmann, Prod he Grat. 15. 1793; repr. in Archiv Bot.
244. 17 Habitat in graminosis humidiusculis Zeylon{ale la D. Konig
collecta}.” Leat- blades described as ‘‘obsolete serratis.”

The plant, represented by Brass 5918, from Dagwa, Oriomo River,
British New Guinea, appears to be this well-known Indo-Malayan species
rather than Lindernia angustifolia (Benth.) Wettst., as identified by m
in Brittonia 2: 182. 193
6. Lindernia anagallis (Burman f.) comb. nov.

Ruellia anagallis Burman f. Fl. Ind. 135. 1768. “Habitat in Java and Amboina.”

1943 J PENNELL, SCROPHULARIACEAE OF NEW GUINEA 253

Based upon plants collected in Java by Kleinhof and on Rumphius’ description
and rupee of the Amboina form, the latter because of the illustration being
taken as the type. Unfortunately, the Academy lacks vol. 5 of Rumphius’
Herbarium Amboinense, but Merrill’s Interpretation of that work (p. 467. 1917)
has identified Rumphius’ plant as the species now considered. Also, Dr. Merrill
has kindly supplied me with a tracing of Rumphius’ ae In a special paper on
the species proposed by Burman f. (in Philip. Jour. Sci. 19: 381. 1921), he says
of Ruellia anagallis: ‘“‘Burman’s species is clearly identical with the very common
and widely distributed form currently known as Bonnaya veronicaefolia Spreng.”
(As will be shown below, I think npn me was mistaken in identifving as this
plant the older Ruellia antipoda L.; nor is it Gratiola veronicifolia Retz., either).

A common Indo-Malayan weed, e euine east to New Guinea. This is
the plant called //ysanthes veronicaefolia (Retz.) Urb. by Schlechter (1. c.
108), who cites specimens from many parts of New Guinea, while I (in
Brittonia 2: 182. 1936, in Jour. Arnold Arb. 20: 81. 1939) have cited
alas oe e from British New Guinea under the names “‘veronicae-
folia” and “antipoda.”’ Further collections are now at hand, viz. Clemens
10830a and 40615 (Mich) from the Morobe District of Northeastern
New Guinea, and several specimens (GH) made by Netherlands collectors
in western New Guinea.

. Lindernia ciliata (Colsmann) comb. n

Gratiola ciliata Colsmann, Prodr. Descr. ‘Grat . 1793; repr. in Archiv Bot. 2: 243.

1799. “E. Java. Konigius non invenit.’ fe ecellent description of the species
now considered.

A common Indo-Malayan plant that appears barely to reach New
Guinea. This was reported by Schlechter (1. c. 109) as Jlvsanthes ciliata
from ‘“‘Waighiou” Island, on the northern shore of Netherlands New

ulnea

8. Lindernia antipoda (L.) Alst

Ruellia antipoda L. Sp. Pl. 635. nie. “Habit Indiis.”” Based primarily upon
account in L. Fl. Zeyl. 106. 1748. The ‘ara ne cern states that the
leaves are oval, acutely and deeply serrate especially toward apex, at base nar-

rowed, the stem is repent, the pediceled flowers several (5) to a aaah the cap-
sule thrice as long as the cre and ne? ere rang large (being like those
shown in Rheede’s Hort. Malabar. 9: tab. 58. , but four times larger).
All these details are clearly distinctive e the species now considered.

li veronicifolia Retz. Observ. Bot. 4: 8. 1786. Clearly identical with Ruellia

ntipoda, as was recognized by Vahl, adele Plant. 1805

Gratiot ae Colsmann, Prodr. Descr. Grat. 12. 1793; eo in Archiv Bot.

2: 243. 1799 a oar in Java et India orientali.” Clearly distinctive of the

species now consider The name chosen leads me to suppose that Colsmann
realized the identity a this with Ruellia ~— Based upon this is /lysanthes
ruelloides (Colsmann )* Schlechter in Bot. Jahrb. 59: 109. 1924, the name used

sidered.
eee reptans Roxb. FI. Ind. cur. Carey 1: 140. 1820. “A native of the Moluccas ;

clon oe ee based upon this, has been the name longest used for
this sp
A common qeaehisevan weed, extending east to New Guinea. Re-

TWith citation of “Konig” rather than “Colsmann.”’

254 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIV

ported by Schlechter (1. c. 109) from Netherlands and Northeastern New
Guinea, and by me (in Jour. Arnold Arb. 20: 81. 1939) from British
New Guinea. Recently collected by Mrs. Clemens, at Wantoal, no. 41093
(AA), and at Boana, no. 41083 (Mich), both in the Morobe District of
Northeastern New Guinea.

8. Torenia Linnaeus

This extensive Oriental genus is represented by at least three species in
New Guinea.

Although a sharp line between Lindernia (in its present enlarged sense)
and Torenia is not easy to draw, I think that the two should be considered
as generically distinct. Usually they are readily distinguishable by the

calyx consisting of separate or only slightly united plane sepals in
Li ndernia, whereas in Torenia the sepals are united nearly throughout and
have the midvein of each raised into a ridge or a wing. But, in Lindernia,
L. crustacea has sepals united over half their length, while i in Torenia the
sepal-ridges may be mere ribs, as in the case of 7. crenata, below. In
Torenia the corolla is usually larger and with angular open throat, whereas

tracted therefrom, in which the bracts are linear and much smaller than
the foliage leaves, whereas in Lindernia the bracts are usually foliose. In
our first species, Torenia crenata, the small flowers, unwinged calyx, and
foliose bracts all led me to consider it a Lindernia, but I now incline to
place generic value on another feature of the calyx. In Lindernia the
sepals, whether distinct or joined, do not invest the capsule, but have their
tips somewhat spreading, whereas in Torenia the sepals do invest the
capsule, being curved about it and with their connivent tips projecting
above it. On this understanding our first species will find its place in
Torenia. It shows that in that genus too there is a tendency toward the
suppression of the anthers of the anterior pair of stamens.
A. Calyx hirsute, the mid-veins of the sepals merely rib-like; corolla 7-8 mm
long, violet; attached portions of anterior filaments projecting as short knobs;

clearly racemose; leaf-blades crenate, cuneately narrowed to the pee petioles ;
stem erect, with ascending branches................ 0.0000 cece euee .T. crenata.
AA. Calyx finely pubescent or glabrous, _ the mid-veins of the sepals a into thin

cels less than twice the length of the calyces; bracts subulate, the inflorescence
(by suppression of an internode) usually seeming a 4-flowered cyme; leaf-blades
dentate, the larger truncate to petioles at least 14 length of blades.
B. Corolla 15-17 mm. long, violet or white; fruiting calyx anvie! ovoid,
ee 16 mm. long, with 5 wings; leaf-blades acute, serrate-dentate, ‘ cm.
ong; stem erect, laxly branched............... 0.000000 eeu Zi Tr violacea.
BB. raed 10 mm. long, pink; fruiting calyx peed circular, 6-8 mm. long,
with 3 wings; leaf-blades obtuse, crenate-dentate, 1-2 cm. long; stem exten-
sively repent, the flowering branches ascending or eek, a ae T. polvgonoides.

1. Torenia crenata (Pennell) comb. nov. Prate I, B.
Lindernia crenata Pennell in Jour. Arnold Arb. 20:79. 1939. Lake Daviumbu,

1943 | PENNELL, SCROPHULARIACEAE OF NEW GUINEA 255

Middle Fly mr British New Guinea, L. J. Brass 7824. Known only from the
original collect
2. Torenia violacea (Azaola) comb. n
Mimulus violaceus Azaolo ex Blanco, Fi. Filip. ed. 2. 357. 1845. ‘“Descubierta v
descrita por el Sr. Azaola, ex Calauan For identification of the species now
considered see Merrill, Species Binncoanné 347. 1918. Although not adopted by

Torenia exappendiculata Regel in Acta Hort. Petrop. 5: 271
Parisiensi accepimus. Patria verosimiliter India orientalis.’ Described as with
small white flowers.
Torenia peduncularis Benth. in Wall. Numer. List, no. 3956. 1831, nomen nudum
(‘““Maulmyne in Martabanica 1827”); Hook. f. Fl. Brit. a a 276. 1884. The
latter cites specimens from Nepal to Assam and Penang, and credits a: species
also to Java and the Philippine Islands. Flowers described as ‘
with white-flowered synonym, 7. alba Ham., from Wallich’s List. Torenia
peduncularis is the name under which our species has been universally known
A weedy plant, widespread over the Indo-Malayan subregion, and douse
fully extending east to New Guinea. Reported by Schlechter (1. c. 104)
fram Netherlands and Northeastern New Guinea; a single collection from
the vicinity of Manokoeari, ee New Guinea (R. F. Janowsky

506), is in the Gray Herbarium. This, however, is a small plant, less
than 1 dm. tall, more hairy, ae smaller, narrower, more closely serrate
leaf-blades, and smaller corollas (10-12 mm. long) and calyces (11-12
mm. long). Perhaps it will prove distinct from the Indo-Malayan species,
as is suggested by Schlechter’s comment that it seldom exceeds 1.5 dm
in height, a small stature for 7. violacea. The only other specimen seen
associable with it is from Bali (C. R. G. J. van Steenis 7588), at the Gray
Herbarium. The problem can be solved only by someone with much more
material for consideration than is available in this country.

uo}
=
Las)
=
>

3. Torenia polygonoides Bent
Torenia polygonoides Benth. in Wall. Numer. ae no. 3897. nomen nudum
(‘Singapur 1822”), Scroph. Indicae 39. 1835. ‘Hab. ad ca Wallich.”

A widespread Indo-Malayan species. Reported by Schlechter (1. c. 104)
from Northeastern and by me (in Jour. Arnold Arb. 20: 78. 1939) from
British New Guinea.

9. Detzneria Schlechter
(Genotype, D. tubata Diels, of New Guinea)
A monotypic genus of Northeastern New Guinea.
1. Detzneria tubata Diels in Bot. Jahrb. 62: 491. 1929.

Saruwaket Range, Northeastern New Guinea, at 3600 to 4000 meters
altitude.

10. Hebe Commerson

This austral genus, so highly developed in New Zealand, is evidently
one of the most characteristic groups of alpine heights in New Guinea.
With the first species published in 1889 and only four described before
1930 the Archbold Expeditions have raised the number from New Guinea
to twelve. Many more will appear as other alpine areas are reached, as

256

JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIV

each isolated highland has evidently ia its own species. The
following key contrasts those known to
A. Leaf-blades serrate-dentate, the teeth Aaron or reer petioles more spreading ;

aa naaa of stem and inflorescence of up
. Serrations spreading, the leaf-blades pn eoiet peduncles 10-30 m

than twice length of bracts; sepals ee re to narrowly ae
dish.

C. Corolla red, its lobes as wide as or wider than long; sepals recdigge ovate;
peduncles spreading; leaf-blades narrowly elliptic, 1.5—2 cm. long, on evi-
dent aaa petioles; rachis and pedicels pubescent ee upcurved

. A. rubra.

brown hair
>. Corolla ae pink, its lobes longer than wide; peduncles OT ea sepals

2)
‘o)

Pe ee

and leaf-blades elliptic-lanceolate, 0.5-1 cm. long, on ascending petioles;
rachis and pedicels pubescent with: upcurved brownish or grayish hairs. .

hymelacoides.

BB. Serrations ascending or appressed, the leaf-blades more obscurely toothed;
leaves loosely set, the blades not or hess nga longer than the intienndes:

peduncles sscandins:

mes the length of the bracts;

30-50 . lon
sepals narrowly obovate. elliptic to ces ie corolla : Hité...4.45%

Pr ree ae ee ee SSeS

. A. albiflora.

AA. Leaf-blades crenate-serrate to entire, the teeth rounded
B. Leaf-blades oval to elliptic- ablane evidently peat serrate or cren

Rac

his and pedicels with brown spreading ri petioles very ren and

wide, ascending against stem; leaves not glutinous.
D. Corolla and sepals externally pubescent ; ane lanceolate- to orbicular-

ovate

81.5 CM: JONG kes ocd o dh etoesesasioenewss 4. H. Lendenfeldit.

DD. Corolla sclehnally glabrous; sepals glabrous or ciliate.

E. Cor
the rachis elongating; leaf-blades flat.
F. Sepals partially ciliate to usually glabrous; pedicels equaling or

olla pinkish or white, 8-14 mm. long; peduncles 20-40 mm. long,

longer than the bractlets; leaf-blades elliptic to elliptic-oblong,
7-16 mm. long; stem distally; and inflorescence rufous-pubescent.

flattened-constricted; leaf-blades crenate- 2 about %

their length, the midrib evident............ H. polyphylla.

GG. Racemes fewer-flowered, the pedicels ne mm. long; cap-
— been than long; aa blades ce beyond the middle,

idrib scarcely evident.............. 6. H. carstensensts.

FF. eeu ciliate throug aa sarees shorter than the bractlets;
eaf-blades oval or elliptic, 6-9 mm. long; stem and inflorescence

more finely pubescent se with paler yi eee . ciliata.

EE. Corolla dark purple, 7-8 mm. long; racemes 1-3- flowered ; peduncles

5 m. long, the rachis shost: the pedicels reaching 5—7 mm. long;
agen Sroecait rounded to base; na blades crenate-serrate above
neate base, somewhat involute................ 8.

t rigi
CC. Rachis and ‘pedicels with ascending or puennone hairs; leaves glutinous

enuis
Hairs of inflorescence brown, upcurved- ascending; sepals ciliate and

pubescent on midrib externally; corolla red, 9-10 mm. long; leaf-

ciliate; stems esac erect, 2—3 dm. tall, the nine as long as the inter-
od 9. H. Br

assit.

RVs d Gira isa Sadat bs a ea Bcaloa pe carck vo va wow Ey SHES pe EIS o.0e wpe eee

‘This assumed to be probably true for Hebe Lendenfeldii, described as with leaves
“on very short stalks,” and for H. carstensensis with leaves “‘sessile

1943 | PENNELL, SCROPHULARIACEAE OF NEW GUINEA 297

DD. Hairs of inflorescence grayish, minute, incurved-appressed ;" sepals
minutely ciliolate; leaf-blades 0.6-0.7 cm. long, the petioles more evi-
dent, ascending-spreading, le ie stem slender, diffuse, smaller, the
leaves shorter ie n the intern
E. Corolla white, 6-7 mm. ~~ : flowers several, in a pair of long-
aici racemes at the apex of the uniformly finely pubescent
FE RO Ee ee ee IST CU Occ Boe 1. tenuis.
EE. pie purple (magenta), 8-9 mm. ae ; flower solitary10 at the
apex of the bifariously puberulent stem........ 11. H. Vanderwateri.
BB. ee blades narrowly lanceolate, obscurely beer to nearly entire; corolla
mm. long; stems and branches densely foliose, me and densely
See eee ee ase conde Se cynesctycercoutnw eevee Agee . A. diosmoides.
1. Hebe rubra Pennell in Brittonia 2: 184. 1936. Pratt II, A.
British New Gurtnea: Mt. Albert Edward, L. J. Brass 4295 (type) and 5677;
Murray Pass, Wharton Range, Brass 4620,
2. Hebe thymelaeoides Pennell in Brittonia 2: 186. 1936. Pirate II, B.
British NEw Guinea: Mt. Albert Edward, L. J. Brass 4296 and 4297 (TYPE).
*3. Hebe albiflora sp. nov. Prate III, A.
Frutex ramosissimus, 6-9 dm. altus; caulis pilis adscendentibus-incurvis
folia oblongo-lanceolata, acuta, serrato-dentata, fere glabra,
subtus pallida, majora 2 cm. longa, 7-8 mm. lata; racemi 6-8 floribus

pubescent with brownish upcurved Pre, the old bark becoming glabrate.
Leaves numerous, not or only slightly longer than the internodes, the blades
oblong- lanceolate, acute, serrate-dentate throughout with low teeth, flat,
dark green and glabrous above, beneath pale and with pilose midrib, 2 em.

long, 7-8 mm. wide, at base rounded to pubescent petioles 2-3 mm. long.
Racemes brown- pubescent (with incurved hairs), 6—8-flowered, elongated
the peduncle becoming 30 to 50 mm. long, the pedicels 5-8 mm. long,

slightly exceeding the lance-oblong bractlets. Sepals 4—4.5 mm. long,
elliptic-oblanceolate, acute, glabrous. Corolla 7-8 mm. long, white, widely
campanulate, glabrous throughout, the upper lip obovate-circular, the three
lower lobes slightly narrower, nearly elliptic. Stamens glabrous through-
out, the anthers about half the length of the filaments. Capsule 4 mm.
long, globose-ovoid. Seeds 0.5 mm. long, circular, flattened, brown.

NorTHEASTERN NEw GUINEA: ae District, Ulap Trail, Matap, alt. 1500-1800
m., Mary S. Clemens 11347 (AA, TYPE), Feb—April 1940, in flower.

4. Hebe Lendenfeldii (F. v. Muell.) Pennell in Brittonia 2: 184. 1936.

Based upon Veronica Lendenfeldii F. v. Muell. in Trans. . Soc.
Victoria 1(2): 29. 1889, describing a plant collected by William
MacGregor in 1889 on Mt. Victoria in British New Guinea.

*5. Hebe polyphylla sp. nov. Prate III, B.

Frutex ramosissimus, 2—8 dm. altus; caulis pilis patentibus ferrugineus;
folia multa, elliptico- ies obtusa, crenato-dentata, glabra, majora
0.7-1.6 cm. longa, 5—7 mm. lata; racemi 5-10 floribus compositi; sepala

9Pedicels of Hebe Vanderwateri described as ‘“‘fein bestaubt.”
10“Semper ?”, asks Wernham.

258 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

4-5 mm. longa, oblonga, irregulariter ciliata; corolla 8-11 mm. longa,
alba, late campanulata, glabra; capsula 2—4 mm. longa, obovoidea.

Stiff, much-branched shrub, 2-8 dm. tall, the upper branches ascending,
densely foliose, the lower slender, diffusely spreading, with longer inter-
nodes and pairs of minute subulate leaves. Stem distally pubescent with
ferruginous spreading hairs, below that bifariously so but on the proximal
half glabrate or glabrous. Leaves numerous, and close-set by reason of
the short internodes, the lower stem marked by the slightly raised petiole-
rudiments of the fallen leaves: blades elliptic-oblong, obtusely rounded,
crenate-dentate, flat, glabrous, 0.7-1.6 cm. long, 5-7 mm. wide, at base
semi-petiolate (cuneately narrowed with entire ciliate margins) to a clasp-
ing base. Racemes brown-pubescent, 5—10-flowered, elongated, the
peduncle becoming 20 to 40 mm. long, the pedicels 3-18 mm. long, about
equaling to much longer than the linear-oblong bractlets. Calyx spongy
(as judged by irregular wrinkling in drying) at base, the sepals (free
above base) oblong, rounded, irregularly ciliate (mostly at apex), 4-6 mm.
long. Corolla 8-15 mm. long, white, pinkish-tinged, widely campanulate,
glabrous throughout, the upper lip nearly circular, the three lower lobes
obovate-semicircular. Stamens glabrous, the anthers about 1 the length
of the filaments. Capsule 2-6 mm. long, obovoid.

NetHerRLANDS New GuINEA: 5 miles northeast of summit of Mt. Wilhelmina, alt.
3440 m., L. J. Brass 9401 (AA, TypE, ANSP), Aug. 1938, in flower and fruit (along
streams in grassland) ; Snow Mountains, on and near Mt. Wilhelmina (seen also from
3 miles east of that summit), Brass 9414 (AA, ANSP), and 7 km. northeast of summit,
Brass & Myer-Drees 9934 (AA, ANSP).

6. Hebe carstensensis (Wernham) Diels in Bot. Jahrb. 62: 491. 1929.

Based upon Veronica carstensensis Wernham in Trans. Linn. Soc. II. Bot.
9: 121. 1916, describing a plant collected by C. Boden Kloss in 1912-13
on Carstensz Peak, Snow Mountains (Nassau Range), Netherlands New
Guinea.

*7. Hebe ciliata sp. nov. Prate IV, A.

Frutex ramosus, 2—6 dm. altus; caulis pilis brevibus patentibus brunneus ;
folia multa, ovalia vel inferiora elliptica, obtusa, crenato-dentata, glabra
glabratave, majora 0.6-0.9 cm. longa, 5-7 mm. lata; racemi 4-8 floribus
compositi; sepala 4-5 mm. longa, obovato-oblonga, ciliata; corolla 7—8
mm. longa, pallide violacea vel alba, late campanulata, glabra; capsula
2-3 mm. longa, obovoidea.

Stiff shrub, 2-6 dm. tall, the rigid stems ultimately erect, the upper

?

branches ascending-erect and densely foliose, the lower slender, diffusely

sometimes pubescent beneath on the obscure midrib), 6-9 mm. long, 5—7
mm. wide, at base widely cuneate to the ill-defined short entire ciliate
petioles. | Racemes brownish-pubescent, 4—8-flowered, elongated, the
peduncle becoming 20-25 mm. long, the pedicels 2—5(—7) mm. long, mostly
shorter than the obovate-oblong bractlets. Calyx spongy at base, the
sepals 4-5 mm. long, obovate-oblong, obtusely rounded, ciliate throughout.

1943 ] PENNELL, SCROPHULARIACEAE OF NEW GUINEA 259

Corolla 7-8 mm. long, pale violet-purple to white, widely campanulate,
glabrous throughout, the upper lip nearly circular, the three lower lobes
nearly as large. Stamens glabrous, the anther less than half the length of
the filament. Capsule 2-3 mm. long, obovoid.

NETHERLANDS NEw Guinea: Mt. Wilhelmina, alt. 3400 m., L. J. Brass & E. Myer-
Drees 9682 (AA, TyPE, ANSP), Sept. 1938, in flower and fruit (rather wet grassy place
along a small river, the whole plant apparently somewhat violet-tinged, the corolla being
pale violet within and nearly white externally); Mt. Wilhelmina, alt. 3900-4250 m.,
alpine grassland, northern slopes, Brass & Myer-Drees 10092, 10094 (both AA, ANSP),
southern slopes, Brass & Mvyer-Drees 10098, 10101, 10104 (all AA, ANSP). No.
10092 differs from the others in the ciliation of the sepals being shorter and partial
(as in other species), but the leaf-blades are oval as in H. ciliata. No. 10094 is said to
be “common on old rock screes,” and no. 10/01 “plentiful on rocky limestone slopes.”
*8. Hebe rigida sp. nov. PLATE IV, B.

Frutex ramosissimus, rigidus; caulis pilis brevibus patentibus pubescens ;
folia multa, oblongo- -elliptica, obtusa, crenato-dentata, glabra, majora
0.5-0.7 cm. longa, 3-4 mm. lata; racemi 1-3 floribus compost sepala
4-5 mm. longa, anguste oblonga, glabra; corolla 7-8 mm. longa, atro-
purpurea, late campanulata, glabra; capsula 4 mm. longa, Bbovaides,

Stiff shrub, 2—4 dm. tall, with many rigidly erect stems from the slightly
spreading rhizomatous base, these distally with many ascending-erect black-
ish purple branches. Stem pubescent with short spreading hairs, tardily
glabrescent below. Leaves numerous and close-set (the short internodes
of the lower part of the stem marked as in the other species), the blades
oblong- eens obtuse, crenate-dentate, Pee: (with hardly evident

midrib), 5—7 mm. long, 3-4 mm. wide, at base cuneate to the ill-defined
short entire aie petioles, both blades a ana involutely hollowed.
Racemes pubescent with short hairs, 1—3-flowered, short, the peduncle

4-12 mm. long, the pedicels 3—5 mm. long, hota than or equaling the
linear-oblong bractlets. Calyx not spongy at base, the sepals 4-5 mm.
long, narrowly oblong, obtusely rounded, the margin entire and glabrous.
Corolla 7-8 mm. long, dark purple, widely campanulate, glabrous through-
out, the upper lip nearly circular, the three lower lobes similar and nearly
as large. Stamens eet the anthers less than half the length of the
filaments. Capsule 4 mm. long, obovoid.

NETHERLANDS ae Guinea: Northern slope of Mt. Wilhelmina, alt. 4100 m.,
L. J. Brass & E. Myer-Drees 10090 (ANSP, type, AA), Sept. 1938, in flower and fruit
(common on Wiener: -grass slopes, often growing in moss-cushions).

9. Hebe Brassii Pennell in Brittonia 2: 185. 1936. PraTeE V, A.

BritisH New Guinea: Mt. Albert Edward, L. J. Brass 4498 (TYPE).
*10. Hebe tenuis sp. nov. PLaTE V, B

Suffrutex repens, ramosus, 1—1.5 an . altus; caulis pilis minutis incurvis
pubescens ; folia See rotundata, crenato-den ntata, glabra, glutinosa,
majora 0.7—0.9 cm. longa, 4-6 mm. lata: racemi 3-7 floribus laxe com-
positi; sepala 4 mm. lon aay oblanceolata vel obovata, glabrata; corolla

longa, alba, campanulata, glabra; capsula 4 mm. longa, obovoidea.

Suffrutescent and extensively repent, distally ascending, 1—1.5 dm. tall,
laxly branched. Stems finely pubescent with minute incurved hairs,
tardily glabrescent below. Leaves shorter than or about equaling the
internodes (which are rather longer than those of the erect bushy species,
and with nodes as clearly marked), the blades elliptic, rounded at apex,

260 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

crenate-dentate, flat, glabrous (with evident midrib), somewhat glutinous,
7-9 mm. long, 4-6 mm. wide, at base cuneate to the short flattened ciliate

NorTHEASTERN NEW GUINEA: Summit of Mt. Sarawaket, alt. 4100 m.,
Clemens 3642 (AA, type), Apr. 8, 1937, in flower (among open grass and shrublets).
11. Hebe Vanderwateri (Wernham) Van Steenis in Bull. Jard. Bot. Buitenzorg II.

13: 252. 1934

Based upon Veronica Vanderwateri Wernham in Trans. Linn. Soc. Il.
Bot. 9: 121. 1916, describing a plant collected by C. Boden Kloss in
1912-13 on Carstensz Peak, Netherlands New Guinea.
12. Hebe diosmoides (Schlechter) Pennell in Brittonia 2: 184. 1936.

Based upon Veronica diosmoides Schlechter in Bot. Jahrb. 59: 111. 1924,
describing a plant collected by Sir William MacGregor in 1889 on Mt.
Victoria in British New Guinea.

11. Veronica Linnaeus

Since Veronica Archboldii was described seven years ago, several more
members of this genus have been noted in collections from New Guinea,
all of them similarly pertaining to the group of V. serpyllifolia L. From
that species and its more cosmopolitan subspecies Aumifusa (Dickson) Vahl
these all differ in the less hairy capsules, and, excepting the diminutive
V. platycarpa, in the relatively wider and more toothed leaf-blades as well.
What may be somewhat arbitrarily accounted the species of Veronica in
New Guinea differ as follows:

A. Capsule widest about midway of its length, often exceeded by the sepals; racemes

becoming many-flowered; leaf-blades elliptic or slightly longer, crenate-dentate ;

stems: -1—4 dm; tall ocd saris dda die ao WSS 20S en Se Oe ns STEMS Hes 1. V. Archboldii.
. Capsule widest near the base, exceeding the sepals; racemes few-flowered; leaf-

blades oblong-elliptic, crenate-serrate with few teeth; stems less than 1 dm. tall,

the whole plant smaller............. 0. cee eee eee eee eee eee 2. V. platycarpa.
1. Veronica Archboldii Pennell in Brittonia 2: 182. 1936,

The following, tentatively considered as subspecies, may prove specifi-
cally distinct. In all the aspect is similar, and there is equal possibility
that further collecting may break down the validity of the characters
adduced. Analogous to the wide distribution of Veronica serpyllifolia and
its subspecies, this is being considered as a single species occurring through-
out the mountain ranges of New Guinea. The subspecies yet known may
be distinguished as follows:

A. Sepals about the length of the capsule (slightly shorter to longer than) ; corolla
pale blue or white; pubescence of stem distally and of pedicels consisting of
minute ascending-incurved hairs; leaf-blades little longer than wide, more evi-
dently crenate-dentate.......... ccc c eee cee eee eee 1b. Subsp. ellipticophylla.

>
>

1943 ] PENNELL. SCROPHULARIACEAE OF NEW GUINEA 261

AA. Sepals much longer than the capsule; leaf-blades with fewer and shallower

dentations.
B. Leaf-blades slightly elongated, 114-2 times as long as wide; pubescence ae
minute ascending-appressed hairs; corolla ‘‘pale purple, streaked with white.’

Beer ees een cect ee eee PT Gea wad 4 Steet ca ee nue aero eg <ereceete la. Subsp. typica.
BB. rine Mae little longer than wide; pubescence of fine spreading hairs; corolla
lc. Subsp. patulifera.

la. Veronica Archboldii subsp. typica.

Veronica Archboldii Pennell in Brittonia 2: 182. 1936.

British New Guinea: Mt. Albert Edward, L. J. Brass 4403 (type), and Murray
Pass, Wharton Range, Brass 4941
*1b. Veronica Archboldii subsp. ellipticophylla subsp. nov.

Caules 2-4 dm. longi, pilis adscendenti-incurvis minute pubescentes; folia
elliptica, 0.7—1.5 cm. longa, an mm. lata; sepala 4-5 mm. longa, oblongo-
lanceolata; rere 3 mm

Stems widely repent, adie distally, 2-4 dm. long, finely pubescent
with minute ascending-incurved hairs. Leaf-blades elliptic, 7-15 mm
long, 5-12 mm. wide, dentate with low teeth. Pedicels slightly longer than
calyces, finely pubescent with incurved-ascending hairs. Sepals oblong-
pumane te becoming 4-5 mm. long. Corolla pale blue or white. Capsule

ong, 4 mm. wide, notched 14 depth. Seeds 1 mm. long."

NETHERLANDS NEW GuinEA: Lake Habbema, north of Mt. Wilhelmina, alt. 3225 m.,

_ J. Brass 9313 (ANSP, type, AA), Aug. 1938, in flower and fruit (on burnt-over
nn 7-11 km. northeast of Mt. Wileleniua, alt. 34CO-3560 m., Brass & Myer-
Drees 9683, 9749, 9861 (all AA, ANSP), in grass and moss.

This appears to be the prevalent member of this group on and near
Mt. Wilhelmina in the Snow Mountains of Netherlands New Guinea.

*ic. Veronica Archboldii subsp. patulifera subsp. nov.

Caules 1-2 dm. longi, pilis patentibus pubescentes; folia elliptica, 0.7-1
cm. longa, 5-7 mm. lata; sepala 4-5 mm. longa, lineari-oblonga; capsula
parva.

Stems widely repent, ascending distally, 1-2 dm. long, distally finely
pubescent with spreading hairs. Leaf-blades elliptic, 7-10 mm. long,
5-7 mm. wide, dentate with low teeth. Pedicels shorter than or the lower
equaling the calyces, ela pubescent with spreading hairs. Sepals linear-

oblong, becoming 4-5 mm. long. Corolla “blue.” Capsule small, not seen
mature (perhaps acne atrophied ).

NETHERLANDS New GuINEA: Northern slope of Mt. Wilhelmina, alt. 4050 m.,
L. J. Brass & E. Myer-Drees 10163 (AA, Type. ANSP), Sept. 1938, in flower (on a wet
tussock-grass slope).

2. Veronica platyearpa Pennell in Notulae Naturae 23: 1. 1939.

Collected by Mary S. Clemens (no. 10120) in 1939 on Mt. Sarawaket,
Northeastern New Guinea.

12. Sopubia Buch. Ham.

One species represented, the Indo-Malayan S. trifida Buch. Ham. , by
Brass 4819, from British New Guinea (already reported in Brittonia

11About twice the length of the seeds of Veronica serpyllifolia L. and V. serpyllifolia
humifusa (Dickson) Vahl. This is the only subspecies of V. Archboldi of which seeds
have been seen

262 JOURNAL OF THE ARNOLD ARBORETUM [vOL, XXIV

2: 187. 1936), and Mary S. Clemens 10736, recently gathered near the
Kajabit Mission in the Morobe District, Northeastern New Guinea

13. Centranthera R. Brown

One species represented, the Indo-Malayan C. cochinchinensis (Lour.)
Merr., reported by Schlechter in Bot. Jahrb. 59: 112. 1924 (as C. hispida
R. Br.), a and by me in Brittonia 2: 187. 1936 (for Brass 3574 & 5703)
and in Jour. Arnold Arb, 20: 81. 1939 (for Brass 7815 & 8270). All the
cited collections are from British New Guinea.

14. Buchnera Linnaeus

Schlechter credited only a single species to New Guinea, but, as might
be anticipated from the number westward in the Malay Archipelago and
southward in northern Australia, there are several. Those now known
may be distinguished as follows:

A. Corolla externally somewhat hairy ; mene TS densely hairy...1. B. tomentosa.
AA. Corolla externally glabrous; inflorescence less hairy.

x and bracts finely pubescent over serra surface, the ari not ridged;

middle and lower leaf-blades lance-oblong, dentate.......... . B. urticifolia.

BB. Calyx and bracts hairy on ribs and margins, or else ae ce middle leaf-
blades lanceolate or nearly linear, the lowest wider and dentate.

C. Bracts ciliate, at least the upper ovate; calyx finely 10-ribbed, the ribs

usually evident by reason of the ascending scabrous hairs, but varying to

glabrous; capsule equaling or slightly exceeding the calyx; flowers more

vate a caudate tip; calva ant BBped:
glabrous except for minute ciliolation of lobes: capsule not seen; flower
less than ten, scattered in an open spike...............-.. 4. B. rariflora.

QO
()
ice]
oa)
ise)
oO
pes
wn
a.
=:
2
i)
_
&
oo
is
oO
c
ao)
no)
5
°

1. Buchnera tomentosa Blume, Bijdr, Fl. Nederl. Ind. 741. 1825.

Originally described from Java. By Schlechter (1. c. 113) this was con-
sidered to be the only species in New Guinea, specimens being cited from
Netherlands and Northeastern New Guinea, and from the Bismarck
Archipelago. None of these collections have been seen and they may not
all fit the concluding note, which characterizes the Papuan plant as differ-
ing from B. urticifolia in its larger flowers and its stronger hairiness, the
hairs being mostly very thick on the rachis of the inflorescence.

NETHERLANDS NEw Guinea: Morobe District, in grassland, alt. less than 1500 m.,
Boana, Mary S. Clemens 41619; Wantoat, Clemens 11208, 40864 (flowers varying from
white to pink
2. Buchnera urticifolia R. Brown, Prodr. Fl. Nov. Holl. 1: 437. 1810.

Originally described from the tropical coastland of northeastern Aus-
tralia, the present Queensland. Occurs in both western and eastern New
Guinea, on open grassland at low altitudes.

NETHERLANDS New Gutnea: Cyclops Mts., K. Gjellerup 500 (GH); Merauke, G.
Versteeg 1837 (GH). NorTHEASTERN NEw Guinea: Morobe District, Kajabit Mission,
Mary S. Clemens, British NEw Guinea: Urunu, Brass 4821; Fly River, Brass 7813,
825312

12Already reported i : ao Arnold Arb. 20: 82. 1939. These are the plants covered
by contrast ‘‘A” of the

1943 J PENNELL. SCROPHULARIACEAE OF NEW GUINEA 263

*3. Buchnera ciliata sp. nov.
Caulis 3-5 dm. altus, laxe pubescens glabratusve, simplex vel pauci-
ramosus; folia scabro- -pubescentia, tricostata, integra dentatave, infima

ora lineari- pects! spica tenuis, 10-15 paribus florum composita;
bracteae 3-4 mm. longae, ovatae, acuminatae, ciliatae; calycis tubus 4 mm.
longus, 10- Se costis scabro- pubescentibus vel glabris, lobis 1 mm.
longis, ovato-lanceolatis, ciliolatis; corolla extus aan. tubo 6-7 mm.
longo ore intus villuloso, lobis oblanceolatis, inferioribus 3—4 mm. longis;
capsula 5 mm. longa; semina 0.5 mm. longa.

m 3-5 dm. tall, loosely pubescent to glabrate, ee or distally
slightly branched. Lowest leaves nearly oval, short, entire or dentate;
middle leaves larger and longer, often nearly oblong, oat 4 cm. long and
10 mm. wide, obtuse, entire or somewhat dentate; upper leaves linear-
lanceolate, 5— 6 cm. long, attenuate, the uppermost slightly smaller and
slightly more remote, but seemingly continuous nearly or quite to the inflo-
rescence; leaves scabrous-pubescent on both surfaces, 3-ribbed beneath.
Spike slender, rather dense (the ascending fruiting calyces nearly or quite
contiguous), of 10 to 15 pairs of flowers. Bracts 3-4 mm. long, ovate,
acuminate or the upper acute, strongly ciliate (hairs 0.3-0.4 mm. long).
Calyx-tube 4 mm. long, finely 10-ribbed, with short ascending scabrous
hairs on ribs or sometimes glabrous; lobes 1-1.2 mm. long, ovate-lanceolate,
attenuate, obscurely ciliolate. Corolla “pink,” externally glabrous, the tube
6-7 m m. long, its orifice finely villose with white hairs, the lobes oblanceo-
late, eae or nearly so, the lower longer, reaching 3—4 mm. long, the
upper 2—3 mm. long. Capsule 5 mm. long, cylindric, equaling or slightly
exceeding the calyx. Seeds about 0.5 mm. long, triangular-conic, brown,
obscurely longitudinally lined.

NETHERLANDS NEW GUINEA: Balim pee northeast of Mt. Wilhelmina, alt. 1600

., L. J. Brass 11725 (ANSP, type, AA), Dec. 1938, in flower and fruit (occasional on
ae deforested slope) ; “Zuid Nieuw Cina pr. O. Kaba,” Branderhorst 89 (GH).
*4. Buchnera rariflora sp. nov.

Caulis 5—6 dm. altus, bifariam Seceiee ae) simplex; folia supra glabra,
subtus scabrella, tricostata, fere integra, infima fere ovalia et brevia, inter-
media etiam parva, oblongo-lanceolata, 2—2. 5 cm, ae superiora lineari-
lanceolata, suprema Breicune et remota; spica tenuissima, circa 3 paribus
florum composita: bracteae 3 mm. lo ongae, lanceolato-ovatae, acuminatae,
parum ciliolatae; calycis tubus 4 mm. longus, ecostatus, glaber, lobis 1 mm.
longis, lanceolato- attenuatis, obscure scabrello- ciliolatis: corolla extus
glabra, tubo 6—7 mm. longo ore ee villoso, lobis oblanceolatis, inferiori-
bus 3—4 mm. longis; eee non

Stem 5-6 dm. tall, aeeneee Pibeealedt simple throughout. Lowest
leaves nearly oval, short, entire or dentate; middle leaves scarcely or not

short and remote, the inflorescence seeming long-peduncled; leaves above
glabrous or slightly scabro-pubescent distally, beneath slightly creel
especially on the three ribs. Spike very slender, lax (the calyces less t

or barely half length of internodes), of about 3 pairs of somewhat ae
flowers. Bracts 3 mm. long, lance-ovate or the upper ovate, acuminate to
slightly caudate, obscurely. ciliolate (hairs less than 0.05 mm. long),

264 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

Calyx-tube 4 mm. long, not evidently ribbed, glabrous; lobes 1 mm. long,
lanceolate-attenuate, obscurely scabrellous-ciliolate. Corolla ‘“‘pink,” ex-
ternally glabrous, the tube 6-7 mm. long, its orifice coarsely villose with
projecting white hairs, its lobes pp einai tl retuse, the lower ea reach-
ing 3—4 mm. long, the upper 2—2.5 m m. long. Capsule not see

British New Guinea: Western Division, Tarara, Wassi Kussa River, L. J. Brass
8571 Spelanel a a AA), Dec. 1936, in or (rare in savanna- forest), ies 8597
(AA, ANSP).1

15. Striga Loureiro

In this Journal for 1939 (20: 83) I discussed the probability of there
being valid color-distinctions between the various species of this genus.
This view seems to be sustained by the further collections now reported.
Also, as there prove to be objections to using either of the specific names
employed by Schlechter for the two species of his enumeration in 1924, it
becomes advisable to present a new summary of the species occurring in
New Guinea.

A. a 5-ribbed; upper lobes over half the length of lower lobes of corolla

rolla white, 12-15 mm. long, externally evidently pubescent, the throat
distinctly wider than the tube and only slightly decurved; posterior calyx-
lobe much shorter than the . ELS; avn ta eh mee eeoues tases as 1. S. alba.
BB. Corolla brightly colored, sma
C. Corolla orange- -yellow ; neat relatively tall and weak, ene branched. .
sic piare’stecs a teres aeiwiewa's sla eisiaereigrears oe bs ay Macon se enters is.s Ny Schlechteri.

CC. ate reddish purple, 7-8 mm. long, ex _ cae pubescent, the
t ill-defined though slightly longer than the tube, rather strongly
rved; posterior calyx-lobe as long as the eo re 3. S. parviflora.

AA. Calyx ois ribbed; upper lobes less than half the length of lower lobes of corolla.
B. Corolla yellow, unlined, the anterior lobes reaching 3 mm. long; oe 5 mm.

long; plant 1-3 dm, tall... 2.2.0... eee ce eee . S. lutea.
BB. Corolla white, venose-lined, the anterior lobes reaching 13 mm. ies calyx
10 mm. long; plant taller... 0.0.0... . cece ce ee eee 5. S. Masuria.

*1. Striga alba sp. nov.

Caulis 7-8 dm. altus, scabro-pilosus, ramosus; folia linearia, obtusa,
scabro-pubescentia, majora 2—4 cm. longa; spicae elongatae, multiflorae:
bracteae lineares, attenuatae, 5—6 mm. longae; bracteolae filiformes, 3.5—5
mm. longae; pe icelli os a 6 mm. longus, tubo paren

superiore 2—3 mm. longo pubescente, labio ee >
glabrato; capsula 2.5 mm. longa, ov voidea : semina mm. lon

Stem 7-8 dm. tall, scabrous-pilose with short broad. based stiff eer
hairs, slightly branched distally. Leaves ascending, linear, obtuse, scabrous-
pubescent, the largest 2—4 cm. long, 1-1.5 mm. wide. ‘Spikes elongated,
scabro-pubescent, that of the main stem of more than 20 opposite or sub-
opposite pairs of flowers. Bracts linear, attenuate, mostly 5-6 mm. long,
about 24 the length of calyx, the lower long an exceeding calyx. Bracte-
oles filiform, 3.5—-5 mm. long. Pedicels scarcely 0.5 mm. long. Calyx 6 mm.

13Already reported in Jour. Arnold Arb, 20: 83. 1939 - a variant of Buchnera
urticifolia. These are the plants covered by contrast KAA” .o e key.

1943 ] PENNELL, SCROPHULARIACEAE OF NEW GUINEA 265

long, its tube with 5 very prominent scabro-pubescent ridges that become
distally Be subulate scabro-pubescent calyx- lobes, four of which are a
2—2.5 mm. long, but the mid-posterior is more slender and short, 1.5 m

long; etna surface of the tube membranous, sparsely minutely pilose.

lobes finely violet-lined, the tube slender and about equaling calyx (5-4
mm. long), glabrous within calyx; throat tubular-campanulate, slightly
enlarged on upper side, 4-5 mm. long, slightly decurved, externally
pubescent, the hairs seemingly slightly gland-tipped; upper lip 2-3 mm
long, arched, its lobes united over half their length, free and laterally up-
curved, externally slightly glandular-pubescent, internally pubescent with
appressed glandless hairs; lower lip 3.5-4.5 mm. long, deflexed-spreading
(its laterally spreading lobes distinct from near base of lip), externally
slightly glandular-puberulent, internally glabrous. Capsule 2.5 mm. long,
ovoid, glabrous, enclosed within calyx-tube. Seeds about 0.5 mm. long,”
irregularly tetrahedral, brown.

British New GurIneA: Gaima, east bank of Lower Fly River, L. J. Brass 8256
(ANSP, type, AA), Nov. 1936, in flower and fruit (common in thick grass, open
savanna-forest, in Jowland).

In 1939 (Jour. Arnold Arb. 20: 83) I identified this plant as Stviga
curviflora (R. Br.) Benth., an opinion I wish now to reconsider. Robert
Brown’s brief description of Buchnera curviflora (Prodr. Fl. Nov. Holl.
438. 1810) should first be quoted: “‘Buchnera curviflora, scaberrima, foliis
elongato-linearibus integerrimis patentibus, calycibus 5-striatis, labio supe-
riore corollae retuso: inferiore 3-partito ter breviore.” Our plant differs
from this in its leaves being ascending rather than spreading, and in its
upper lip being 24 instead of 14 the length of the lower lip of the corolla.
In 1835 (in forming the combination Striga curviflora in Compan. Bot.
Mag. 1: 362) Bentham, who had seen only Brown’s specimens for his
account, called the leaves subspreading (‘‘subpatentibus”’), thus indicating
that they were likely somewhat ascending, and added that the corolla-
tube was pubescent, while he retained in full force the statement about the
relatively small upper lip of the corolla. Years later, with the help of
more collections, Bentham again treated the matter in his Flora Australien-
sis (4: 517) in 1869. There we learn that Brown’s plant, the type of the
species, was gathered on islands in the Gulf of Carpenteria in North
Australia, and that the species is now known both in that province and in
Queensland. Again we have the emphasis placed upon the shortness of
the upper corolla-lip as a main criterion (“less than half as long as the
lower’), while the length of the corolla is now given as “nearly or fully
34 in.,” a size too large for our plant. On Bentham’s key of 1869 our
species fits better the characterization of S. multiflora, which reads:
“Corolla above % in. long, the upper lip more than half as long as the
lower,’ except that the corolla is exactly 4% inch. But from S. multiflora
Benth. of North Australia (as originally defined in Compan. Bot. Mag.

14Imperfect seeds, shaken from a capsule, are only 0.2-0.3 mm. long, but those
lying in folder with specimen and presumably of this species have been described

266 JOURNAL OF THE ARNOLD ARBORETUM [vOL, XXIV

1: 363. 1835) our plant differs in not being much-branched, its leaves not
spreading, its corolla not glabrous, and its habit not semi-decumbent.
Finally, as to S. curviflora again, the scarcely decurved corolla of our plant
must differ from that which suggested this name to Brown. On all these
counts I think that our plant from New Guinea is different from either of
these Australian species. Likely it differs in other characters as well, for
Brown’s description is meagre and Bentham’s far from complete.

2. Striga Schlechteri sp. nov.

Caulis ramosissimus, debilis; folia longa; calyx 5-costatus; corolla lutea.

Stems much-branched, weak and rather tall (relative to S. /utea), the
leaves longer, the calyx 5-ribbed, the corolla bright ee (but dry-
ing dark blue-gray), and both corolla and capsule smaller

NorTHEASTERN NEw GurINeEA: Constantinhafen, . alanis wn (TYPE, presum-
ably in the Berlin Botanical Garden), March 1902, in flower and fru

Other collections cited by Schlechter Sinn this as occurring pelsewuera) in
Northeastern New Guine

This brief characterization is all that is given by Schlechter (in Bot.
Jahrb. 59: 114. 1924), but it seems sufficient to mark this as another new
species. Certainly it can hardly be Striga multiflora Benth., above-
mentioned, to which Schlechter assigned the plant, since Bentham’s
account of this in his Flora Australiensis (4: 517. 1869) gave the corolla
as over 14 inch long, a size much above that of S. /utea. On the other hand
it is just possible that Schlechter’s plant was S. parviflora, but his identi-
fication of it as S. multiflora and the different color assigned to the flowers
lead me to suppose that it is something else. S. multiflora was originally
described as having a glabrous corolla, a detail so modified in the Flora
Australiensis as to lead to the suspicion that several things were included
in the latter account.

3. Striga parviflora (R. Brown) Benth. in Compan. Bot. Mag. 1: 362. 1835.

Based upon Buchnera parviflora R. Brown, Prodr. Fl. Nov. Holl. 1: 438,
1810, of tropical Australia, the type being stated by Bentham, FI. Austral.
4: 517. 1869, to have come from Keppel Bay, Queensland.

As previously reported (in Jour. Arnold Arb. 20: 84. 1939), this was
collected in British New Guinea at Port Moresby (Brass 8780), and now
we may add several collections from near Kajabit Mission in the Morobe
District, Northeastern New Guinea (Mary S. Clemens 10676, 10745,
41170)

4. Striga lutea Loureiro, Fl. Cochinch. 22. 1790.

_ Type actually from Canton, China. Merrill (in Trans. Amer. Philos.
Soc. II. 24(2): 353. 1935) puts this in the synonymy of Striga astatica
(L.) Kuntze, but Linnaeus’ Buchnera asiatica was a composite of what-
ever Strigae he knew, although his allusion to the corolla-limb being purple
will likely prove distinctive. With which component to associate this
earliest name must await more thorough revision.

To collections cited by Schlechter (1. c. 115), all from Northeastern

1943] PENNELL, SCROPHULARIACEAE OF NEW GUINEA 267

New Guinea, may be added a recent one from the Kajabit Mission in
the Morobe District (Mary S. Clemens 40771).
5. Striga Masuria (Buch. Ham.) Benth. in Compan. Bot. Mag. 1: 364. 1835.

In Wallich’s Numerical List of Plants in the Museum of the East India
Company, this oe appears as numbers 3876 Buchnera Wallichii Benth.
and 3877 B. Masuria Ham., under date of 1830. Both names were nomina
subnuda (localities only ae given), but the latter was validated by
publication with description by Bentham in his Scrophularineae Indicae
(p. 41) in 1835. Schlechter in 1924 (1. c. 115) seems to have been the
first to validate the former, but his Striga Wallichii (Benth.) Schlechter
does so only by including Bentham’s account of the other as a synonym.
The correct name for the species therefore remains Masuria, the name
generally used since 1835.

A widespread oriental species that extends eastward to New Guinea,
where it was reported by Schlechter from the northeastern part and has
recently been gathered slightly farther south at the Kajabit Mission,
Morobe, Northeastern New Guinea, by Mary S. Clemens (10547a, 40658,
40772)

16. Euphrasia Linnaeus

This proves to be one of the characteristic genera of alpine heights
throughout New Guinea. Like Hebe in the same situations, the kinship of
the Papuan species is definitely with the flora of New Zealand. Also like
Hebe, the genus will prove to be a large one in New Guinea, with many
species locally endemic to various mountain areas.

The clarity of the taxonomy of Euphrasia in New Guinea has been a
welcome surprise to one who has puzzled over the difficult task of dis-
tinguishing its species in the North Temperate Zone. The boreal species,
all of which are annual, are closely similar in corollas, stamens, capsules,
seeds, calyces, and leaves, so that one must appeal to secondary features of
leaf-dentation, indumentum, and size to distinguish the many local entities.
But in New Guinea the characters are as consequential and well-marked
as in any other genus of this family. The species are perennial, the habit
either erect or repent, the corolla with short or long tube, the calyx-lobes
rounded or acuminate, and the anthers distinct or coherent, uniformly or
unequally awned. Those as yet known from New Guinea may be dis-
tinguished as follows:

A. Corolla externally glabrous, 10-11 mm. long, the lips as long as or longer than

i tube; anthers glabrous, exserted; leaf-blades 3-5-lobulate; stem to 2 dm
emcee Peeh sient e iron eae gk aian et aA incest a sig 20a sa ob koa ear avasse OME Gen EN . E. papuana.

AA. ca Arenas pubescent; anthers included, at least a orifices ciliate; leaf-
blades 3-lobed, or the lower or all entire (all only in E. culminicola).

B. Leaf- blades longer than wide, the lateral lobes less mie Y% the length of the
wide median lobe; lower lip deflexed, usually exceeding the upper lip of the
corolla; stem ascending or erect.

C. Stems shrubby, 1-3 dm. tall; calyx-lobes obtuse; corolla 12-14 mm.
long, white to yellow, externally resi subeoelcnt, the lower (8 mm.)
slightly longer than the upper (6-7 mm.) lip................ 2. E. Lamii.

CC. Stems suffruticose, less than 1 se tall; corolla violet to purple, evidently
hairy externally

268 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

D. Anthers cohering, one awn of anthers of shorter stamens much longer
and thicker than the other anther-awns; corolla 12-15 mm. long, the
campanulately wanes throat and lips as long as or longer than the
Halk tube; leaves glabrous.

. Lower et es exceeding upper lip of corolla, its length about that
roat.

ish.
G. Leat- ane with a pair of short tooth-like Ba lobes ;
ephigee POG caja sc gengreestn hohe 246 tele s . cucullata.
eaf- blades entire, spatulate or obovate... .4. : culminicola.
FF. te lobes acuminate; corolla white or very pale pink........
5.

EE. Lower scarcely exceeding upper lip of corolla, its length about half
- ia t of the decurved corolla-tube and -throat; calyx-lobes
MAC 5 hoses cee eres ees ps esses eases 6. curviflora.
DD. Sees: sia the awns all equal in length, short; ares Aa 11 mm.
long, the slender throat gradually or not widening from the tube, which
together are at least twice as long as the lobes
E. Tube of corolla gradually widening to apex, its orifice not con-
stricted; plant much-branched at base.
F. Calyx-lobes apie corolla pale violet, the galea straight con-
yepantd with th tt capsule little shorter than calyx; leaves
5-6 mm. long, gla en ds Nesta bate ciere raise eis ais age ee 7. E. rectiflora.
FF. Calyx- jokes rounded - apex; corolla lavender to dark violet, the
galea shorter and curved continuously with the tube; capsule
eine than calyx (not iia the ee of its lobes) ;

s 3-4 mm. long, puberulent.............. 8. E. spatulifolia.
EE. Tube “OE corolla ae at mouth, leaves 3.5 mm. na glabrous;
ms simple or sparsely branched............. 9. E. scutellarioides.

wo
ee)

. Leaf-blades as wide as long, the hem oe 1% to Y% the length of te wide
median ware lower lip deflexed-spreading, not or scarcely exceedin
upper lip of the pale corolla; anthers distinct, one awn of anthers of ee
pair slightly longer than the other anther-awns; stem repen
C. Corolla 8-10 mm. long, slightly and mostly sparsely pilose externally ;
orifice of anther-cells abine (the hairs less than 14 the width of the

of the calyx- _— stems extensively repent, bifariously pubescent with
recurved hairs. ....... 000 cece teen eee ener seesaw eens 10. E. humifusa.
. Corolla 4-5 mm. long, hirsute- pubescent externally ; orifice of anther-cells

QO
‘o)

pubescent, the lobes rounded with margin strongly revolute and callose ;

eaves 1-3 mm. long, pubescent, ultimately glabrous, the margin like that

of the calyx-lobes; stems more shortly spreading............ 11. E. callosa.
1. Euphrasia papuana Schlechter in Bot. Jahrb. 59: 116. 1924.

Collected by Sir William MacGregor in 1889 on Mt. Victoria in British
New Guinea.

2. Euphrasia Lamii Diels in Nova Guinea Bot. 14: 538. 1928.

Collected by H. J. Lam in 1920 at Doorman Peak, Netherlands New
Guinea; two specimens cited but neither designated as type. Lam 1587,
very abundant on flat sterile moorland, 3250 meters altitude, is also shown
in his photograph in Die Vegetationsbilder 15: tab. 33. 1924, and evi-
dently reaches the proportions given in the original description of 1928.
Accordingly, I take it as the type. The published field-note states that the

1943 | PENNELL, SCROPHULARIACEAE OF NEW GUINEA 269

corolla was white or bright yellow, sometimes with lilac spots on the lips,
and was yellow on the throat. In contrast, Lam 1644, gathered among
stones on open alpine slopes at 3520 meters altitude, was noted in the field
as with corolla white, occasionally bright yellow, yellow at the throat, and
the tube with violet stripes to the margin of the upper lip; I suspect from
the allusion to striping as well as the different environment that no. 16/4
is a different species and is probably the plant that I am now describing
as Euphrasia cucullata, although unfortunately Mr. Brass has not told us
the color-pattern of the latter. If so, the occasional bright vellow-flowered
individuals may well be interspersed specimens of E. Lamii, such as might
take slightly wetter spots. From Lam’s map in Die Vegetationsbilder,
Doorman Peak is less than a degree northwest of Wilhelmina Peak.
*3. Euphrasia cucullata sp. nov.

Suffruticosa, ramosissima, diffusa, 5-10 cm. alta: caulis retrorse pube-
scens; folia obovato-spathulata, 5-6 mm. longa, glabra, margine apicem
versus valde revoluta (cucullata), utrinque pari loborum triangulari-rotun-
datorum onusta, basi cuneata: flores pauci, axillares; calyx minute pilosus,
6-7 mm. longus, lobis ovatis rotundatis 2 mm. longis: corolla purpurea,

capsula 7 mm. longa, glabra; semina 1.5 mm. longa.
Suffruticose, much-branched below, the stems less than 1 dm. tall,
pubescent with recurved hairs, diffusely ascending or decumbent. Leaves

lute as to appear cucullate: larger leaves 5-6 mm. long, 2-3 mm. wide.
Inflorescence of a few pairs of axillary flowers, these not closely successive.
Pedicels 2 mm. long, pubescent. Calyx 6-7 mm, long, minutely pilose, the
lobes 2 mm. long, ovate, rounded but with margin so recurved as at times
to appear acutish. Corolla “purple,” externally hirsute-pubescent, but
glabrous beneath throat and on lobes distally, internally finely pubescent
below posterior sinus,'* elsewhere glabrous (except perhaps at base of tube),
the tube 7-8 mm. long, narrow, slightly exceeding the calyx: upper lip
5 mm. long, slightly arched and hooded, the lobes free slightly (0.5 mm.)
at apex; lower lip 7-9 mm. long, deflexed-spreading, the mid-anterior obo-
vate, slightly longer than the semirectangular antero-lateral lobes, all retuse.
Filaments glabrous (except probably at base); anthers all coherent, the

ISTf the anthers of Euphrasia Lamii had been described in detail. probably it would
have been possible to associate it with some of the following species. At present it is
impossible to identify its variety Versteegii Diels (in Nova Guinea Bot. 14: 539. 1928),
collected by Versteeg (no. 2473) in 1913 on the Orange Mountains (“Oranje-Gebirge”’)
in Netherlands New Guinea.

16These are shorter and less divaricate than in Euphrasia mirabilis, the lobes of
which are also placed about the middle of the leat’s length (not so distally as is implied
in my original description of that plant). The leaf-blade is distally more strongly
cucullate in £. cucullata than in E, mirabilis.

17Such fine pubescence beneath sinus near apex of upper lip is developed also in
E. mirabilis.

270 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

longitudinal thin white ridges.

NETHERLANDS New GurneA: Mt. Wilhelmina, 2 km. east of summit, alt. 3800 m.,
L. J. Brass & E. Myer-Drees 10118 (ANSP, type, AA), Sept. 1938, in flower (alpine
grassland, plentiful in sterile shallow soil of sandstone crests) ; 3 miles east of summit,

Wilhelmina, alt, 3650 m., Brass 9412 (AA, ANSP) (scattered over black boggy
slopes).
4. Euphrasia culminicola Wernham in Trans. Linn. Soc. II. Bot. 9: 121, 1916.

Collected by C. Boden Kloss in 1912-13 on Carstensz Peak, in Nether-

lands New Guinea.
5. Euphrasia mirabilis Pennell in Brittonia 2: 188. 1936.

Collected by L. J. Brass, no. 4271, in 1933 on Mt. Albert Edward, in
British New Guinea.

*6. Euphrasia curviflora sp. nov.

Caules laxi, 8 cm. et ultra longi, pilis patentibus hirsutuli; folia glabra,
oblanceolata vel obovato-spathulata, 5 mm. longa, margine apicem versus
revoluta, utrinque pari loborum lanceolato-oblongorum brevium onusta,
basi anguste cuneata; flores pauci, axillares; calyx glabratus, 7.5 mm.
longus, lobis ovatis acuminatis 3.5 mm. longis; corolla violaceo-purpurea,
extus pubescens, tubo 5 mm. longo angusto, labio superiore 4 mm. longo,
inferiore vix longiore lobis retusis aucto; antherarum cohaerentium thecae
aristas inaequales ferentes, orificiis ciliolatae; capsula 6 mm. longa, glabra;
semina non visa.

Stems lax, at least 8 cm. long, probably ascending, finely hirsute with
spreading hairs. Leaves glabrous, oblanceolate to obovate-spatulate, nar-
rowly cuneate to base, about midway of length with a pair of lance-oblong
short lobes, the median distal portion rounded-ovate, the margin somewhat
revolute and slightly cucullate, those seen 5 mm. long, 2—2.5 mm. wide. In-
florescence of a few pairs of axillary flowers. Pedicels scarcely 1 mm. long,
apparently glabrate. Calyx 7.5 mm. long, minutely pubescent below
i becoming glabrous throughout, the lobes 3.5 mm. long, ovate,
acuminate, slightly revolute. Corolla externally pubescent, hirsutely so on
tube and minutely villose on galea, the lobes distally glabrate, internally
pubescent below posterior sinus, the tube 5 mm. long, narrow, about equal-
ing the calyx, distally campanulate, widening to a throat 4 mm. long;
upper lip 4 mm. long, arched, the lobes free slightly at apex and laterally
recurved 0.5 mm.: lower lip scarcely longer, deflexed-spreading, the mid-
anterior slightly the widest, all retuse. Filaments glabrous (at least dis-
tally); anthers all coherent, the cells ciliolate at margins of orifice, all
awned (one awn of each posterior anther 0.3 mm. long, the others 0.1 mm.
long, the long awn stouter and projecting farther down into the orifice of
the corolla). Capsule 6 mm. long, 4 mm. wide, obovate-rounded to apex,
flattened, glabrous. Seeds not seen.

NorTHEASTERN NEw GuINnEA: Morobe District, Rawlinson Range, alt. over 3600 m.,
Mary S. Clemens s.n. (ANSP, Type), July 1941, in fawer (open wet hills and dale).

18About twice as long as the corresponding awns of E. mirabilis.

1943 ] PENNELL, SCROPHULARIACEAE OF NEW GUINEA 271

Of this I have seen only a few pieces, including a flower, a nearly ripe
capsule, and another capsule long past dehiscence, all culled from the only
collection of Euphrasia callosa, to be described below. As to color, the
common label states: ‘‘mixture, purple & white, tube yellow of big fl.,”
whence we know that this, the big flower, had a yellow corolla-tube, a
color still evident on the anterior side of the throat. The edge of the lower
corolla-lobes seems still a dark violet, whence I infer that the main part
of the corolla was violet-purple.

*7, Euphrasia rectiflora sp. nov.

Suffruticosa, basi ramosa, erecta vel diffusa, 1-7 cm. alta; caulis retrorse
pubescens vel glabrata; folia glabra, oblanceolato-spathulata, 5-6 m
longa, margine apicem versus revoluta, utrinque pari loborum triangulari-
oblongorum rotundatorum onusta, basi cuneata; flores pauci, axillares;
calyx glaber, 7 mm. longus, lobis lanceolatis acuminatis 3 mm. longis;
corolla violacea, extus minute pubescens, tubo 10 mm. longo angusto
elongato, labiis 3 mm. longis, inferiore lobis rotundato-truncatis aucto;
antherarum distinctarum thecae aristas aequales ferentes, orificiis cilio-
latae; stylus minute pubescens; capsula 5-6 mm. longa, glabra: semina
non visa.

Suffruticose, less than 7 cm. tall, often minute, much-branched at base,

to glabrate. Leaves glabrous, oblanceolate-spatulate, the upper nearly
obovate, cuneate to base, about midway of length with a pair of triangular-
oblong rounded lobes, the mid-blade distally Gee rounded, its margin
revolute (but hardly cucullate) and becoming indurated, the larger leaves
5-6 mm. long, 2-3 mm. wide. Inflorescence of a few (1 or 2) pairs of
axillary flowers, borne contiguously at the summit of the stems. Pedicels
less than 1 mm. long, glabrous. Calyx becoming 7 mm. long, glabrous, the
lobes 3 mm. long, lanceolate, acuminate, the margin strongly revolute.
Corolla ‘“‘pale violet with yellow spot inside,” apparently lined, externally
finely pubescent, the tube anteriorly and the lower lobes glabrous, inter-
nally minutely pubescent below posterior sinus, the tube 10 mm. long,
narrow, gradually widening distally, over twice the length of the calyx:
upper lip 3 mm. long, narrowly arched, scarc cely or not hooded, the lobes
free slightly at apex (0.5 mm.), tending to spread laterally; lower lip 3 mm.
long, widely deflexed- spreading, the lobes 2 mm. long, rounded-truncate,
slightly widening distally. Filaments glabrous (except perhaps at base) ;
anthers distinct, the cells ciliolate at margins of orifice, all equally awned
(awns slender, 0.3 mm. long). Style distally finely pubescent. Capsule
5-6 mm. long, 4 mm. ies poe flattened. Seeds not seen.

awe E Lake Habbema, north of Mt. Wilhelmina, alt.
3225 m.,;.L, J. Brass ees (AA, TYPE, ANSP), Aug. 1938, in flower and fruit (among
dwarf pensece of exposed boggy aineee 7 km. northeast a Mt. Wilhelmina, alt. 3560
m., Brass & Myer-Drees 9794 (AA, ANSP) (on very wet peat in grassy valley).
*8. Euphrasia spatulifolia sp. nov.

Suffruticosa, caespitosa, diffusa, 1-7 cm. alta; caulis De patentibus
pubescens; folia minute pubescentia, oblanceolato- -spathulat mm.
longa, margine apicem versus valde revoluta (cucullata), neiaaie pari

19From label of no. 9794; for no. 9192 stated as “pale purple-pink.”

272 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

loborum oblongorum onusta, basi cuneata; flores pauci, axillares; calyx
hispidulus, 7 mm. longus, lobis oblongis rotundatis 3 mm. longis; corolla
violacea, extus pubescens, tubo 9 mm. longo tenui elongato, labio superiore

mm. , inferiore 4 mm. longo lobis rotundato-truncatis aucto;
antherarum distinctarum thecae aristas aequales ferentes, orificiis ciliatae;

sa.

, forming clumps up to 1 dm. in diameter. Plant less than
7 cm. tall, branched throughout, the stems diffuse from base and ascending,
pubescent with spreading slightly recurved hairs. Leaves minutely
pubescent, oblanceolate-spatulate, cuneate to base, about midway of length
with a pair of oblong rounded or obtuse lobes, the mid-blade distally
broadly rounded, its margin strongly revolute and somewhat hooded, the
larger leaves 3-4 mm. long, 1.5-2 mm. wide. Inflorescence of a few (1
or 2) pairs of axillary flowers. Pedicels 1 mm. long. alyx becoming
7 mm. long, minutely pilose with stiff hairs, the lobes 3 mm. long, oblong,
truncately rounded, the margin narrowly revolute. Corolla “‘lavender”’
(no, 10044) or “dark violet” (no. 9731), with the ‘lowest petal orange at
base,” externally grayish-pubescent, the throat anteriorly and the lobes
distally glabrous, internally minutely pubescent below posterior sinus, the
tube 9 mm. long, narrow, proximally very slender, distally widening to
form throat and somewhat decurved, over twice the length of the calyx;
upper lip 2 mm. long, arched and decurved 1 mm., then the lobes widely
spreading, rounded; lower lip 4 mm. long, deflexed-spreading, the lobes
2-2.5 mm. long, rounded-truncate, slightly widening distally. Filaments
glabrous (at least distally); anthers distinct, the cells ciliate at margins of
orifice, all equally awned (awns slender, 0.2 mm. long). Style distally
finely pubescent. Capsule 4-5 mm. long, 2-3 mm. wide, obovate, flattened.
Seeds not seen.

NETHERLANDS New Guinea: Northern slopes of Mt. Wilhelmina, alt. 3950 m.,
L. J. Brass & E. Myer-Drees 10044 (AA, TypE, ANSP), Sept. 1938, in flower and old
fruit (shallow soil of old grassy scree) ; 11 km. northeast of summit of Mt. Wilhelmina,
alt. 3400 m., Brass & Myer-Drees 9731 (AA, ANSP) (on rather wet sandy slope).

9. Euphrasia scutellarioides Wernham in Trans. Linn. Soc. II. Bot. 9: 122. 1916.

Collected by C. Boden Kloss in 1912-13 on Carstensz Peak, Netherlands
New Guinea.

*10. Euphrasia humifusa sp. nov.

longa, glabra; semina 1 mm. longa.

Perennial, widely effuse. Stems up to 1 dm. long, wholly prostrate, par-
tially repent, much-branched, bifariously pubescent with recurved-
appressed hairs. Leaves glabrous or pilose proximally above on midrib or
with a few hairs at tip of blade or lobes, widely ovate in inclusive outline,

1943 ] PENNELL, SCROPHULARIACEAE OF NEW GUINEA “id

narrowly cuneate to a semi-petiolate or fully petiolate base, about midway
of length with a pair of divergent oblong obtuse lobes, the mid- blade distally
oval or rounded-ovate, its margin relatively slightly revolute; larger leaf-
blades 2-3 mm. long and wide, the petioles sometimes reaching 1 mm. long.
Calyx 3—4 mm. long, pilose with stiff hairs to usually glabrate, the lobes
1.5 mm. long, ovate or ovate-oblong, obtuse, so revolute as at times to
appear acute. Corolla externally pinkish, internally white, the lips (or
perhaps only lower lip) pink,?° externally finely pubescent (usually villosely
so), the throat anteriorly and the lobes distally glabrous, internally
minutely pubescent below posterior sinus, the tube 5 mm. long, narrow,
straight, nearly twice length of calyx, distally abruptly widening to a
campanulate throat 2 mm. long; upper lip 2.5 mm. long, straight or
slightly decurved, one A lobes free only near apex (0.2—0.3 mm. long),
upcurving ; lower li m. long, deflexed-spreading, the lobes 2—2.3 mm.
long, oblong, retuse. "Filaments glabrous (at least distally ) ; anthers dis-

of s
which is about 0.2 mm. long). Style distally nearly glabrous (obscurely
sparsely puberulent). Capsule 4 mm. long, 3 mm. wide, flattened,
obovate-rounded, emarginate, the valves in dehiscence strongly recurved-
spreading. Seeds 1 mm. long, oval with distally flattened end, blackish, with
white longitudinal wing-like ridges.

NETHERLANDS NEW GUINEA: 4 km. northeast of summit of Mt. We arner alt.
3650 m., L. J. Brass & E. Myer-Drees 9971 apd TYPE, AA), Sept. 1938, in flower
and fruit (in shade of tussock grasses) ; 2 km. east of summit of Mt. Wilhelmina, Brass
& Myer-Drees 10187 (AA, ANSP) (in alpine olan
*11. Euphrasia callosa sp. nov.

Herba perennis, effusa, prostrata; caulis 5 cm. longus, ramosus, hirsuto-
pubescens; folia glabra glabratave, late ovata, 1-3 mm. longa, margine
callosa apicem versus valde revoluta (cucullata), aes pari loborum
oblongorum ovatorumve onusta, basi cuneata; flores pauci, irregulariter
axillares; calyx hirsuto- -pubescens, 3—4 mm. longus, lobis oblongis ovatisve
1 mm. longis: corolla albida, extus pubescens, tubo 2.5—4 mm. longo recto,
labiis 1.5—-2.5 mm. longis, inferiore lobis truncatis aucto; antherarum dis-
tinctarum thecae aristas subaequales ferentes, orificiis ciliatae ; a

Perennial, widely effuse. Stems up to 5 cm. long, wholly, ets.
hirsute-pubescent on all sides with spreading hairs. Leaves glabrous, or
slightly pubescent beneath laterally, widely ovate in inclusive outline,
cuneate to a semipetiolate base, about midway of length with a pair of
divergent pee: or ovate rounded lobes, the mid-blade distally ovate-
rounded, the margin strongly revolute, even somewhat hooded, that of the

several Paget! flowers, occurring irregularly at various nodes. Pedicels
1-2 mm. long, hirsute-pubescent. Calyx 3-4 mm. long, hirsute-pubescent
on ribs, the lobes 1-1.2 mm. long, oblong to ovate, distally acutish to
rounded, with revolute strongly callose margin. Corolla probably white

20Field-note to label for no. 9971: “fis. white within, the outside tinged with pink”;
for no. 10187: ‘‘fls. very pale pink, lip darker

274 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

or whitish,2! externally pubescent, the throat anteriorly and the lobes
distally glabrous, internally minutely pubescent below posterior sinus, the
tube (and throat) 2.5-4 mm. long, straight, little longer than calyx; upper
lip 1.5-2.5 mm. long, straight or nearly so, the lobes slightly distinct and
recurved near apex; lower lip about as long, deflexed-spreading, the lobes
1-1.5 mm. long, truncate or shallowly retuse. Filaments glabrous (at
least distally) ; anthers distinct, the cells ciliate on margins of orifice, the
awns slightly unequal (as in E. humifusa). Style distally finely pubescent.
Capsule 3 mm, long and wide, flattened, rounded, glabrous. Seeds 0.7—0.8
mm. long, grayish, with longitudinal white ridges.

NorTHEASTERN New Guinea: Morobe District, Ulap Trail, Mary S. Clemens 41137
(ANSP, type, AA), Apr. 6, 1940, in flower and fruit.

Presumably this locality is in the Rawlinson Range, from which Mrs.
Clemens has sent to the Academy another collection made in July 1941,
on “open wet hills and dale” at over 3600 meters altitude; it was this collec-
tion from which the few pieces of Euphrasia curviflora have been ex-
tracted. Still another collection, her no. 6047, gathered March 29, 1937,
and credited only to the Morobe District at the surprisingly low altitude
of 2400 to 2700 meters (“8—9000 ft.”’) is at the Arnold Arboretum.

EXPLANATION OF PLATES

All plates are photographs of herbarium specimens and are reproduced slightly over
4 actual size.
PraTE I

A. Lindernia papuana; type. B. Torenia crenata; type.

Prate II

A. Hebe rubra; isotype. B. Hebe thymelaeoides; isotype.

PrateE III
A. Hebe albiflora; type. B. Hebe polyphylla; type.

PiaTE IV
A. Hebe ciliata; type. B. Hebe rigida; type.

PLATE V

A. Hebe Brassii; isotype. B. Hebe tenuis; type.

ACADEMY OF NATURAL SCIENCES,
PHILADELPHIA, PA.

21 Type collection labeled ‘“fls. white with purple,” probably meaning white with
purple lines. As to other collection, intermixed with Euphrasia curviflora, I take the
present species to be the pale-flowered component, see discussion under that species.

“IOA “AY GIONAY ‘waof

AIXX

Torte, Crnmote, CPanel,
roa re einyn

L é ine Cate vee

tae DETER MeED He reamcie we peneRLL
itary

Viants of Papua (Hritish New Guinea), collected by the
tos Fly River Expedition of the American Museum
of Natural History, Mr, Richard Archbold, Leader,

Vistriberted: bi thea A =
Ne 7994. J.J. Baw te pteaber, 19a6

car Daviusbu, widdve Tly hiver.
Common om wet rrase plains.

] aLVIg

A. LINDERNIA PAPUANA B. TORENIA CRENATA

Jour. ArNotp ARB. VoL. XXIV

Spence:
Zeb. NSP)

PLATE II

B. HEBE THYMELAEOIDES

A. HEBE RUBRA

A. HEBE ALBIFLORA

POLYPHYLLA

Univers
Ne. 0401 Ld. Brass & HOM yer rere “Sapte 1988
Hale apokyrhayple Pammald, Ty PE
lee northeast of »ilhemt a-top, 2440m,
t rarmiend:oonnen under banks’ of at:

of str
tink ix

‘ayy @IONaYy ‘anof

“IOA

ATXX

If] 4ivid

Jour. AkNoLD ArB. VoL. XXIV

PLATE IV

B. HEBE RIGIDA

A. HEBE CILIATA

A. HEBE Brassu

B. HEBE TENUIS

H ee tennis Paral |

AIXX ‘TOA ‘dy GIONay “yaof

A ALVIg

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 275

A STUDY OF CYTOLOGY AND SPECIATION IN THE GENUS
POPULUS L.

E. CHALMERS SMITH!
With four plates

INTRODUCTION

A stupy of the extent of interspecific hybridization within a genus has
both practical and theoretical importance. It is of value in showing to
what degree hybridization may be utilized within a genus to establish
improved types and, at the same time, it is indicative of the relationships
of the species themselves. In order that the study of the extent of hybrid-
ization within the genus shall prove of maximum value, data on behavior
of the F, and Fy, generations under controlled conditions should be
available. Due to the time factor involved, such data are not available in
most cases when dealing with forest trees.

The study of genetic behavior of interspecific hybrids is of greatest
value when coupled with a comparative study of the cytology of those
hybrids. Studies of this sort have yielded many pertinent facts, particu-
larly as critical evidence for the establishment of probable interspecific
relationships. In addition it has led to a better understanding of the
actual methods by which isolation and speciation have taken place. The
basic assumption underlying this method of approach to the problem of
interrelationships of species is that the pairing of the parental chromo-
somes is a criterion of chromosome homology. This criterion of chromo-
some homology, based on a study of pairing relationships at meiosis in F,
species hybrids, is particularly valuable when supplemented with a study
of chiasmata frequencies, chromosome configurations, and pollen sterility.

The study of chromosome numbers, microsporogenesis, and the de-
velopment of the male gametophyte of species within the genus supple-
ments this program of research, making it more valuable from both points
of view. The study of chromosome behavior and pollen sterility of the
pure species serves as a basis with which to compare the cytological be-
havior of the hybrids. The chromosome numbers will indicate the degree
of polyploidy within the genus and indicate the probable success of any
attempt to induce polyploidy.

The present investigation is a study of the cytology of the genus
Populus L. in general, and as such it is subject to the limitations imposed
by the unavailability of certain species and hybrids for study. With these
limitations in mind it is a study in particular of chromosome numbers and
chromosome behavior in pure species and inter-specific hybrids, especially
as these data are related to speciation within the genus.

1Maria Moors Cabot Foundation Fellow.

276 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIv

MATERIALS AND METHODS

The materials used in this study were obtained from two sources:
(1) the Populus collection of the Arnold Arboretum, and (2) the Populus
plantations of the Northeastern Forest Experiment Station at Frye,
Maine. Cytological material of species and natural hybrids were obtained
from the former location and material of artificial hybrids from the latter.

Branches with flower buds attached were collected during the latter part
of February and early March, placed in the greenhouse, and allowed to
develop. Catkins were collected at appropriate times and fixed in 3:1
alcohol acetic and stored in the fixative at 2° C. until needed. Male
catkins of Populus make excepuonally favorable material for cytological
study, since one catkin contains many successive stages in the develop-
ment of the pollen grain.

Aceto-carmine smear preparations were used entirely in the study of
meiotic chromosomes. Pollen mother cells in prophase stages were diff-
cult to stain, and those having the chromosomes advantageously placed
for study were rare. When favorable cells were found, the chromosomes
were drawn with the aid of a camera ludica. The length of the chromo-
somes was then measured by the use of waxed threads.

For root tip studies, cuttings from the desired species were rooted in
water or sand, the root tips collected and fixed in Navaschin’s solution,
embedded in paraffin, sectioned at 10u and stained with crystal violet.
A similar procedure of fixation, embedding, and sectioning was followed
in the study of the development of the male gametophyte, except in this
case the sections were stained with Haidenhain’s iron-haematoxylin.

In the study of the development of the pollen tube, freshly shed pollen
was sprinkled on slides which had received a thin coating of a sugar, agar,
gelatin, and water mixture. The preparation of these slides has been
described by Newcomber (1938). In this case a mixture of 2 gm. sugar,

5 gm. agar, and 0.5 gm. gelatin to 25 cc. of water was found to be
satisfactory. After the pollen was planted, the slides were placed in a
moist chamber until the pollen had germinated. It was then fixed with
3:1 alcohol acetic and stained with aceto-carmine.

CHROMOSOME NUMBER AND MORPHOLOGY

The first investigation of the chromosomes of Populus was made by
Graf in 1921. He found the reduced chromosome number in P. tremula
and P. canadensis, as determined from reduction divisions in the embryo
sac mother cells, to be four. These counts have since proved to be
erroneous. In 1924 Blackburn and Harrison, on the basis of chromosome
number in seven species of Populus and seventeen species of Salix, estab-
lished the fundamental reduced chromosome number in the Salicaceae
as nineteen. In Salix, a polyploid series based on multiples of two, four
and six was found. Since 1924 various workers have studied chromosome
numbers in Populus. Table I summarizes the chromosome numbers which
have been determined up to this time and lists the authority for each

1943 | SMITH. CYTOLOGY AND SPECIATION IN POPULUS 277

TABLE 1.
PREVIOUSLY DETERMINED CHROMOSOME NUMBERS OF POPULUS
SPECIES, WITH AUTHORITY AND DATE.

Diploid
Species Number Authority

*P. alba 38 Von Wettstein (1933)

*P. alba SS Peto (1938), van Dillewijn (1940)
P. alba var. nivea aureo-intertexta 57 Peto (1938

P. balsamifera! 38 Meurman (1925)

P. canescens 38 Peto (1938)

P. canescens 57 Peto (1938

BE deltode var. missouriensis 38 Van Dillewijn (194

*P. Euge 38 Blackburn (1926), ae (1938)
P. gelrica 38 Van Dillewijn (1940)
*P. gene 38 Blackburn (1926)
bal th eee eae 38 Peto (1938

P. lasiocarpa 38 Von Wettstein (1933)
*P. nigra 38 Blackburn and orn (1924),

van Dillewijn (194

*P. nigra var. italica 38 Van Dillewijn ca

*P. robusta 38 Van Dillewijn (1940)

P. serotina? 38 Blackburn (192

*P. serotina 38 Van Dillewijn (1940)

P. Sieboldii 38 Nakajim

*P. Simonii 38 Meurman (1925)

P. tremula 38 Blackburn and Harrison (1924)

P. tremula 38 Von Wettstein (1933), Muntzing (1936)
P. tremula 57 Muntzing (1936), Tometorp (19-

*P. tremuloides 38 Erlanson and rmann (1927), Peto (1938)
*P. trichocarpa 38 Meurman (1927

* Count confirmed in the present study
1 P. balsamifera L.= P. deltoides Tarek. var. missouriensis Henry.
2 P. Eugenei probably = XP. canadensis Moench, var. Eugenei (Simon-Louis) Schelle
3 P. serotina Hartig= XP. canadensis Moench, var. serotina (Hartig) Rehd.

count. Table II shows the distribution of chromosome numbers among
the sections of the genus. Those determined by the writer are indicated
in this table. In all, some forty-five species, varieties, and natural hybrids
have been investigated.

An examination of these numbers reveals that no polyploid series
similar to that found in Salix exists in Populus. All species exist in the
diploid form with the diploid number of chromosomes equal to thirty-
eight. However, in the section Leuce, three species are found which
possess triploid Sank also, with the unreduced chromosome number of
fifty-seven.

It should be pointed out that, in most cases, the chromosome numbers
for any particular species have been determined from the examination of
but one or, at the most, a few specimens of the species. It is possible that
the examination of a species throughout its entire range might reveal

278 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV
TABLE 2.
A LIST OF CHROMOSOME NUMBERS OF POPULUS SPECIES AND VARIETIES, SHOWING
THE DISTRIBUTION OF THESE NUMBERS AMONG THE SECTIONS OF THE GENUS.
Chromosome
Section Species Number, 2N
Leuce P. alba 38, 57
P. ee var. nivea aureo- intertexta 57
*P. adenopoda 38
P. canescens (Ait.) Sm 38, 57
P. grandidentata Michx 38
P. Sieboldii Mi 38
P. tremula L. 38, 57
P. tremuloides Michx. 38
*P. tomentosa Carr. 38
Leucoides P. lasiocarpa Oliv. 38
Tacamahaca *P. acuminata Rydb 38
**P_ angustifolia James 38
**P_ candicans Ai 38
*P. cathayana Rehd 38
xP. generosa Henry 38
**P. Jackii Sarg 38
*P. koreana Rehd 38
*P. laurifolia Ledeb. 38
ae aximowiczii Henry 38
P. Simonii Carr. 38
**P. Tacamahaca Mill. 38
*P. Tacamahaca var. Michauxii (Dode) Farwell 38
P. trichocarpa Hook. 38
**P_ trichocarpa var. mili Henry 38
*P. Woobstii (Reg.) Do 38
Aegeiros *P. Andrewsii Sarg. 38
**P. angulata Ait. 38
xP. barbantica Houtz. 38
*P. berolinensis Dipp. 38
**P. canadensis Moe 38
*P. canadensis var. erecta (Selys-Longchamps) Rehd. 38
XP. canadensis var. Eugenei (Simon- ars — 38
*XP. canadensis var. marilandica (Poir.) 38
* P. canadensis var. regenerata (Schneid. ; ok 38
xP. canadensis var. serotina (Hartig) Rehd. 38
*P. deltoides Mar: 38
P. deltoides var. missouriensis Henry 38
<P. als Houtz 38
P. nigra 38
P. nigra var. italica Muen 38
rp nigra var. betulifolia (Pursh) 1 38
*P. nigra var. plantierensis reas Louis) Schneid. 38
xP. robusta Schneid. 38
** XP. Rasumowskyana Schneid. 38
** P. Sargentii Dode 38

** Determination made by the writer from root tip preparation.
* Determination made by the writer from meiotic preparations.

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 279

polyploidy within that species. In Sweden, where P. tremula has been
studied most extensively, nine clones of the triploid form have been dis-
covered (Nilsson-Ehle, 1936; Muntzing, 1936; Blomquist, 1937; Tome-
torp, 1937; Johnsson, 1940). One tetraploid form of P. balsamifera
(= P. deltoides?) was reported by Blackburn and Harrison in 1924.
Meurman (1925) thinks it likely that some species other than P. balsami-
fera was examined. Although it is true that workers since 1924 have
found this species to be a diploid, it is possible that it may also exist in
the tetraploid form. Johnson (1939) lists P. Simonii as having the
unreduced number of seventy-six chromosomes. However, Meurman
(1925), who first investigated this species, stated that it is probably a
hybrid and due to this fact frequently shows thirty-eight univalent
chromosomes at meiotic metaphase. Material investigated in the present
study showed a reduced number of nineteen.

Polyploid forms of P. tremula have been produced by different investi-
gators. A tetraploid form of this species resulted from the crossing of two
triploid forms (Nilsson-Ehle, 1938). Similarly, tetraploids and individu-
als possessing all the chromosome numbers from nineteen to thirty-eight
have been obtained from crosses of diploid and triploid forms of
P. tremula (Johnsson, 1940).

The chromosomes of Populus are small and of varying size. Blackburn
and Harrison (1924) made the first observations on size of meiotic
chromosomes in P. tremula. Here they found that the chromosomes were
‘“. . of unequal dimensions; nine were small ones of more or less uni-
Earn size; nine others, larger than these, formed a graded series beginning
with a member of just a little greater coluohe than the individual of the
first group, and ending with one more than four times its volume. Lastly
there was a single chromosome, obviously compound in structure, nearly
always appearing in a flat plate as four-lobed, equalling in volume, if not
exceeding, that of any two of the other eighteen.” These observations
were found to apply to P. nigra also. The studies were made on meiotic
chromosomes at metaphase and anaphase I. Meurman (1925) thought
there were two groups of nine chromosomes, nine smaller a nine larger,
each group varying within itself. In P. balsamifera and P. trichocarpa
one of the chromosomes was noted as being twice the size of any of the
rest. Erlanson and Hermann (1927) saw a similar size classification in
P. tremuloides. Muntzing (1936) and Johnsson (1940) agree that in
P. tremula one of the bivalents is much larger than other chromosomes
of the set. Nakajima (1937) speaks of a particularly long chromosome
in P. Sieboldi, while van Dillewijn finds a ‘‘giant chromosome” in the
meiotic configurations of P. nigra, P. nigra var. italica, P. brabantica,
P. gelrica, P. robusta, P. deltoides var. missouriensis, P. serotina, and
P. alba.

No measurements have been made during this investigation of the size
of mitotic chromosomes of Populus. However, an examination of root tip
chromosomes of P. tremula, as illustrated by Muntzing (1936) and Johns-

280 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

son (1940), would indicate that they range in size from approximately
0.75 to 2.1u. Since the chromosomes are quite small and numerous,
studies of chromosome morphology at mitotic divisions are extremely
difficult. An attempt was therefore made to study the prophase stages
of meiosis. Fig. 1 shows a pachytene stage in the hybrid P. nigra x P.
trichocarpa. Due to the difficulties involved in making accurate measure-
ments, the lengths are to be regarded as approximations. However, it
can be seen from the idiogram (Fig. 2) that there is no “giant chromo-
some” apparent at this stage. Neither can there be any division of the
chromosomes into two groups of nine, one composed of small and one of
large chromosomes. There is, rather, a gradual increase in size from
about 8.5 to about 30.6u in length at this stage. There are one group of
three chromosomes and two groups of two chromosomes which are of
exactly the same length, a fact which may be of some significance. The
discrepancy in reports on the comparative lengths of prophase and meta-
phase meiotic chromosomes may be due to a differential rate of contrac-
tion of the chromosome during the prophase stages. The preparations
did not stain in such a way as to show the spindle insertion point.

The genus Populus is dioecious. For this reason it might be expected
to be the subject of investigation seeking to determine whether or not sex-
chromosomes are present in plants. Blackburn and Harrison (1924)
first described a heteromorphic chromosome pair in the meiotic chromo-
somes of P. tremula as consisting of a medium sized and a small member.
In subsequent investigations the same was found to be true in the case of
P. nigra. Meurman (1925) seems to have been the first to designate the
larger of this chromosome pair as an X chromosome and the smaller as a
Y. Thus the male in Populus is XY, the female XX, making the condi-
tion existing in this genus analogous to that prevailing in most animals.
He was able to demonstrate a heteromorphic chromosome pair in the male
trees of the species P. trichocarpa, P. balsamifera, and P. Simonti. Erlan-
son and Hermann (1927), investigating the cytology of perfect flowers
in a specimen of P. tremuloides, found a heteromorphic pair present at
the meiotic divisions similar to that which they saw in the monoecious
form of the same species. Nakajima (1937) found that a heteromor-
phic chromosome pair exists in P. Sieboldii.

Muntzing (1936) points out that, since the triploid forms which exist
in certain species of Populus must have arisen from the union of two
gametes, one reduced and one unreduced, we might expect intersexual
forms. None, however, has been found. Peto (1938) found no hetero-
morphic chromosome pair which was present consistently. However,
heteromorphic pairs were found in hybrids, and in such instances he thinks
that their presence might be explained on the basis of structural differ-
ences involved in species differentiation.

Although a special study of sex-chromosomes has not been made a part
of this investigation, it may be said that no definitely heteromorphic
chromosome pair has been found consistently in any of the species or

1943 | SMITH, CYTOLOGY AND SPECIATION IN POPULUS 281

hybrids studied. Fig. 10 shows what appears to be a heteromorphic
bivalent similar to those used as illustrations of sex-chromosomes in
Populus by different authors. It would seem that a demonstration of the
presence or absence of a heteromorphic pair at meiosis in a pistillate tree
might offer some proof of the existence of sex-chromosomes in this genus.
Only one investigation of female trees has been made with this point in
mind. Blackburn (1926) states, in reference to P. tremula, “I am able
to state in regard to the female only that all pairs appear to be equal.”
A study of meiotic prophase chromosomes, where conditions are most
favorable for the detection of morphological differences among the
chromosomes, has thus far revealed no strikingly different homologues.

Lawrence (1931) suggests that in general those families with high
chromosome numbers, such as the Salicaceae, will be found to be second-
ary polyploids. Secondary polyploids, according to the definition of
Darlington and Moffett (1930), are ‘“. . . homozygous allopolyploids
in which some chromosomes of the basic set are present more frequently
than others.’ Some examples of such genera are Pyrus (Darlington and
Moffett, 1930), Acer (Meurman, 1933), and Dahlia (Lawrence, 1931).

Cytologically, secondary polyploids may be recognized by the appear-
ance of a secondary association of chromosomes at either pro-metaphase
I or II of the meiotic divisions. The presence of more than two geneti-
cally similar chromosomes will result in the formation of occasional multi-
valent chromosome configurations at the meiotic metaphase I. Further,
if these chromosomes are genetically similar it might be expected that they
would be morphologically similar, also.

Lawrence (1931) found evidence of secondary association of chromo-
somes on examination of illustrations of the meiotic chromosomes of
species of Salix published by Blackburn and Harrison (1924). Van
Dillewijn (1940), working on Populus, found secondary association of
chromosomes in the case of P. nigra and P. nigra var. italica. On this
basis he divides the chromosomes into three groups of three each, four
groups of two each, and an additional group composed of the ‘“‘giant
chromosome” associated with a smaller chromosome. From this latter
association he concludes that the giant chromosome is formed from the
union of two smaller chromosomes and that the group as a whole is a
group of three. There are thus eight groups in all, giving a basic ancestral
number of eight to this genus. Of the original eight chromosomes, four
have been duplicated and four triplicated, while within one group of three
chromosomes, two have become fused.

Additional evidence that Populus is a secondary polyploid is given by the
rare occurrence of trivalent formations in the meiotic metaphase of cer-
tain diploid hybrids (Table VIIT) and the similar occurrence of occasional
quadrivalents in the triploid form of P. alba (Table III). An examination
of the idiogram (Fig. 2) does not reveal the similarity in chromosome size
which one would expect from van Dillewijn’s account of secondary associ-
ation. Certain members of the chromosome complement are of exactly the

282 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

same length. However, on a purely morphological basis it is impossible
to obtain the grouping of chromosomes to give eight as the basic number.

There seems to be evidence that the genus Populus is of polyploid
origin, but obviously there must be a more detailed study before the
actual basic number can be determined with any certainty.

CYTOLOGY OF SPECIES

According to Chamberlain (1897), the stamens of Salix pass the winter
in the spore mother cell stage. He inferred that this was likewise the
case in Populus. The present work confirms this, for catkins collected in
December and early February of 1940 showed the pollen mother cells in
a resting condition.

The rate of the meiotic divisions in all species is quite rapid. All stages
from first meiotic division to the stage where free immature pollen grains
are seen may take place within a space of from twenty-four to thirty-six
hours

The instigation of the meiotic divisions is usually correlated in this

genus with the development of anthocyanin pigment in the anthers.
Further, the sequence of these divisions and consequently of pollen
formation within the catkin seems to be constant within the species and
even within the sections of the genus. Erlanson and Hermann (1927)
noted that in normal male catkins of P. tremuloides the divisions began in
the stamens at the base of the catkin and continued regularly toward
the tip.
In general there seem to be two types of order of flowering. In type I
the meiotic divisions are instigated at either a central or basi-central
position and proceed toward the tip and base of the catkin. In type II
the divisions begin at the tip of the catkin and proceed toward the base.
In the seventeen species and varieties examined, the type was constant
within a section of the genus. Type I was found in the sections Leuce
and Tacamahaca, while type II was typical of the section Aegeiros. The
natural hybrid * P. berolinensis was aberrant, for while belonging to the
section Aegeiros it was of type I.

Observations made at the time of anther dehiscence bear out these
results, for it was noted that those species belonging to type I shed the
pollen at the basal end of the catkin first, while in type II the reverse was
true. Meiotic divisions within a pollen sac are usually at the same stage,
although two successive stages may be present. In general, the same may
be said of all the anthers in one flower. However, in hybrids some irregu-
larities have been observed. Meurman (1925) noted that in P. Simonit,
which is probably a hybrid, stages from prophase I to telophase I in-
clusive were often found in the same anther sac. Erlanson and Hermann
(1927) found a similar condition in the anther sacs of perfect flowers in
P. tremulotdes.

The first and second meiotic divisions, with few exceptions, are regular
in the true species of Populus (Figs. 8-15). The picture is diagrammatic
in its simplicity and regularity.

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 283

TABLE 3.

CHROMOSOME PAIRING AND POLLEN STERILITY IN POPULUS SPECIES.

Species Pairing Relationship Sterility
I IV %
P. acuminata 0.80 18.60 0 0 45
P. adenopoda 0.40 18.80 ) 0 19
P. alba (diploid) 0 19.00 0 0 3
P. alba (triploid) 10.56 13.52 5.56 0.68 23
P. deltoides 0 19.00 0 0 2
P. grandidentata 0 19.00 0 0 2
P. koreana 0.50 18.75 0 0 12
P. laurifolia 0.40 18.80 0 0 40
P. nigra 0 19.00 0 0 3
P. nigra var. italica 0 19.00 0 0 4
P. Sargentii 0 19.00 0 0 7
P. tremuloides 0 19.00 0 0 2

TABLE 4.
CHROMOSOME PAIRING AND POLLEN STERILITY IN A SPECIES OF POPULUS
FROM WHICH FLOWERING BUDS WERE FORCED IN THE GREENHOUSE
R DIFFERENT PERIODS OF T

Date at which Chromosome pairing Pollen

meiosis took at meiosis sterility
collection place I Il Ill %
January 24 February 2 7.92 15.04 0 70
March 9 March 12 7:32 15.28 0.04 72
March 23 March 23 7.80 15.10 0 65

Table III gives the pairing relationships observed in a number of
species. The counts are based on the study of twenty-five metaphase or
late diakinesis configurations for each species, while the sterility counts
were made on the basis of a count of one thousand pollen grains for each
species.

The material for study was obtained from the Populus collection in the
Arnold Arboretum. Flowering branches were brought into the greenhouse
over the period of February to April and forced into flower. The results
might be criticized on the basis that forced material might not behave as
normally developed material. However, Nohara (1924) found in pollen
studies of Salix sp. that results obtained from forced pollen did not differ
either in percentage of perfect grains or in viability from that collected in
the field. That forcing does not affect chromosome pairing and pollen
sterility in Populus may be seen from Table IV, where collections of
Populus sp. (probably a hybrid) made over the period January to March
are compared in regard to chromosome behavior and pollen sterility.

Chromosome pairing was complete in most cases where pure species

284 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

were studied. This pairing was so intimate that the different members of
the pair could be distinguished only with difficulty. Occasional univalent
chromosomes were encountered, as for example in P. laurifolia and P. ade-
nopoda. Apparently this lack of pairing at metaphase I was the result
of a precocious separation due to failure or early terminalization of chias-
mata rather than to lack of homology between the chromosomes concerned.
In this connection Johnsson (1940) reports that in sixteen clones of
diploid P. tremula examined, four clones showed metaphase I plates
having varying numbers of univalents. This is attributed to the influence
of external conditions, such as temperature, on meiosis.

Diakinesis proved to be an advantageous stage for study, for at this

cell. The nucleolus is usually still present at this stage (Fig. 8) and may
in some cases remain visible until the early metaphase. However, in those
cases where the nucleolus persists until late diakinesis, it shows an irregu-
lar outline and a light staining reaction indicative of dissolution. Van
Dillewijn (1940) finds that in P. brabantica, a natural hybrid, the nucle-
olus is still visible at the metaphase in some cells, either on the plate or
near the plate, in the cytoplasm.

One chromosome is invariably associated with the nucleolus (Figs. /,
8, 16) and is easily distinguished from the remaining chromosomes of the
complement, since it is somewhat condensed and darker staining than the
rest. The association of one particular chromosome of the complement
with the nucleolus has been described by several writers, including Heitz
(1931), Sax (1932), and Smith (1933).

In many of the species studied it was noted that the nucleolus possessed a
protuberance or knob. Rarely more than one of these was present per
nucleolus. A similar condition was described in P. nigra by van Dillewijn
(1940). According to this writer no protuberance is visible on the nucle-
olus in the early prophase, but as the prophase progresses a bud develops
until it sometimes reaches the dimensions of a nucleolus itself. It seems
to the present writer that the knob is first visible in very early prophase
and that it reaches its maximum development at approximately the pachy-
tene stage. No difference in size of the knob was noticed between pachy-
tene and early diakinesis. By late diakinesis the knob began to disappear
along with the nucleolus proper. In all cases where the nucleolus pos-
sesses a knob, the associated chromosome is located at the junction of the
knob and nucleolus (Fig. 1).

The percentage of nucleoli which showed knobs differs in different
species. In P. deltoides about seventy-seven per cent, in P. alba (diploid)
about forty per cent, and in P. nigra about eighty percent showed knobs.
In P. alba the knobs are quite small and rarely approach those of P. nigra
and P. deltoides in size.

"The situation where one finds a knobbed protuberance of the nucleolus
has been reported but rarely in the flowering plants. Selim (1930), Nandi
(1936), and Parthasarathy (1938) report a somewhat similar condition

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 285

in Oryza, as do Paul (1937) in Tamarindus and Iyengar (1939) in Cicer.
McClintock (1934) reports a reciprocal interchange in Zea Mays, pro-
duced by x-ray treatment involving the nucleolus organizer, which pro-
duces a condition similar in appearance to that described here.

Following telophase I the nuclei are reorganized and the second meiotic
division follows in a regular manner

As is the case in simultaneous pollen cell formation, the cell wall is
formed by a furrowing process. This is the method most commonly found
in the flowering plants. A condition similar to that which occurs in
Populus has been described in Nicotiana by Farr (1916). The furrows
form along the equator of each spindle, extending from the periphery to
the center of the pollen mother cell, eventually cutting it into four micro-
spores. Previous to the appearance of these furrows, there appears to be
a more or less hyaline area present in the center of the cell, presumably
caused by a migration of the protoplasmatic granules away from this
particular region. This hyaline area extends in four arms from the center
to the periphery, marking the future position of the furrows. This seems
somewhat comparable to the condition found in Melilotus (Castetter,
1925), where similar hyaline areas appeared, caused by a vacuolation of
the cytoplasm. Whether or not the hyaline area in Populus is caused
by a vacuolation must be determined by a more critical examination
aided by differential staining methods.

The furrowing process seems to be easily upset. In many cases in hybrid
material and in the triploid form of P. alba it appears that only one furrow
formed, thereby cutting the pollen mother cell into dyads instead of tetrads,
although the dyads are themselves binucleate. Occasionally the furrowing
process was observed to fail entirely, or it was of such a nature that three
microspores instead of the usual four were formed. The significance of
such aberrant furrowing is indeed great, if such spores are functional.
Numerous examples of such irregularities in microspore formation in
Populus have been described (Peto, 1938; van Dillewijn, 1940).

A fairly wide variation is seen in the pollen sterility of the pure species
studied. In general, these species which showed some univalent chromo-
somes were the most sterile, but the univalent frequency is hardly sufficient
to account for the sterility encountered. Apparently genetic and environ-
mental factors are also involved.

The triploid form of P. alba, with an unreduced chromosome number of
fifty-seven, shows, as one would expect, a varying number of univalents,
bivalents, and trivalents. It also shows quadrivalent formation to some
extent (Table III), the significance of which has already been commented
upon. It is surprising that the pollen sterility of this triploid is less than
that of some of the diploid species studied.

Pollen fertility reported for various triploid forms has been summarized
in Table V. Peto (1938) has commented upon this high fertility. It
is his opinion that this is merely apparent, and he assumes that the
genetically unbalanced pollen grains deteriorate rapidly following their

286 JOURNAL OF THE ARNOLD ARBORETUM [vOL, XXIV

TABLE 5.

POLLEN STERILITIES IN TRIPLOID FORMS OF POPULUS SPECIES.

Triploid Per cent

Species Fertility Authority
P. tremula 58-75 Johnsson, 1940
P. tremula 44 Muntzing, 1936
P. al 63 Peto, 1938
P. alba var. aureo-intertexta 94 Peto, 1938
P. alba 67 Smith, 1942

formation. In the pollen slides which he examined, he observed numerous
tiny specks which he interpreted to be degenerate pollen grains. It has
been shown, however, that in the cross of a triploid & a diploid and a
triploid X a triploid, P. tremula, progeny with intermediate chromosome
numbers varying from thirty-eight to seventy-six can be obtained
(Johnsson, 1940; Bergstrom, 1940). This fact seems to show that un-
balanced pollen grains in at least one species of Populus can survive. It
has been shown that the genus Populus is probably a derived polyploid
(van Dillewijn, 1940, and the present study). If this is the case, then
certain chromosomes and thus certain combinations of genes are duplicated
within the basic set. This condition would be exaggerated in the case of
a triploid form, where the basic set of nineteen chromosomes is present three
times. In a situation of this sort, it might be expected that pollen sterility
due to duplications and deficiencies of whole chromosomes would not be
apparent, for it is conceivable that spores lacking certain duplicated
chromosomes would still have a functional set of genes

On the basis of the pairing relationships of the chromosomes at meiosis,
it is apparent that the triploid forms of diploid species thus far reported
are autotriploids (Muntzing, 1936), which probably arose through the
union of a diploid with a haploid gamete.

Studies of the first and second meiotic divisions and the development of
the immature microspores have been made by different writers. How-
ever, little work has been done on the further development of the male
gametophyte. Chamberlain (1897), working with P. monilifera (prob-
ably P. deltoides), reports that the division of the microspore nucleus into
the tube and generative nuclei takes place in the pollen grain relatively
early, before the tapetal cells of the anther sac-degenerate. He describes
a rather unusual condition in that two divisions of this pollen grain nucleus
occur. On the conclusion of the first division, the smaller of the daughter
nuclei is cut off from the other by a cell wall and degenerates. The remain-
ing nucleus then divides again to form the generative and tube nuclei. He
concludes by stating, ‘Since spores already upon the stigmas showed no
further differentiation, the division of the generative cell which presumably
takes place, although I was not so fortunate as to observe it, must occur
after the pollen tube begins to form.”

1943 | SMITH, CYTOLOGY AND SPECIATION IN POPULUS 287

If the condition described by Chamberlain, in which an extra prothallial
cell is produced in the pollen grain, is correct, then it is unique, as far
as is known in the Angiosperms, and thus of very special interest. Con-
sequently an attempt was made to follow the microspore development up
to the production of the gametes. In this part of the work the species
P, deltoides, P. acuminata, and P. adenopoda were used. It was possible
to demonstrate the first microspore division by aceto-carmine smears
(Fig. 3). However, shortly after this division occurs, the pollen wall be-
comes ‘so opaque that further observation by this method becomes im-
possible. Further study was made by embedding in paraffin, sectioning,
and staining with haematoxylin. A second division within the pollen
grain was not observed. The mature pollen grains examined contained
two nuclei and there was no visible remnant of a disintegrating nucleus.
However, occasional cases, both in P. acuminata and P. adenopoda,
showed a third nucleus within the pollen grain. Since these species show
some irregularity at metaphase I (Table III), it is possible that these
irregularities might account for this third nucleus.

The development of the pollen tube in the species P. deltoides, P.
laurtfolia, P. acuminata, and P. adenopoda was investigated by placing
freshly shed pollen from these species on slides coated with a mixture
of agar-agar, gelatin, and sugar, as described in the section on technique.
The pollen germinates almost immediately. As a rule the tube nucleus
emerges first and the generative nucleus follows. The appearance of the
generative nucleus after its emergence from the pollen grain is such that
the chromosomes may be distinguished within it. In certain cases it
would seem that the term generative cell should be used rather than
generative nucleus. The division of the cell takes place from ten to fif-
teen hours after germination of the pollen grain. Different stages in the
development of the pollen tube are shown in Figs. 4—7.

HYBRIDIZATION IN POPULUS

Natural hybrids are of frequent occurrence within the genus Populus.
Johnson (1939) lists sixteen such interspecific hybrids. Rehder (1940)
lists sixteen hybrids which have been given species or varietal names but
does not include all those listed by Johnson. In all, about twenty-five
natural interspecific hybrids have been reported in the genus.

Among the first interspecific hybrids to be described in this genus were
many which occurred in botanic gardens when an introduced species was
planted near a native species or when two introduced species were planted
together. « Populus berolinensis, the hybrid P. laurifolia « P. nigra var.
italica, which originated in the Botanic Garden of Berlin before 1870,
is an example of this. >< Populus canadensis and its varieties and < P.
robusta are further examples. Peto (1938) reports similar hybrids in
Canada between the European species P. alba and the native species P.
grandidentata and P. tremuloides. It has been noted that female trees
of species which have no male trees in the vicinity set viable seed. P.
Maximowiczii is a good example of this fact, for in the Arnold Arboretum

288 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

it sets seed, some of which produce hybrid seedlings, although there are no
male trees of this species in North America.

At different times natural hybrids have been described from the field.
This has been especially true in eastern Asia and North America, where
the number of species of Populus is most abundant. From China the
‘‘species” P. Simonii is probably a natural hybrid, and & P. Woobstii is
thought to be the cross P. laurifolia P. tristis. In order to determine
the actual extent of natural hybridization among the North American
species, a survey was made of the geographic distribution of the principal
species and their natural hybrids, based upon a study of the herbarium
sheets of Populus in the herbarium of the Arnold Arboretum.

< Populus Andrewsii represents the cross P. Sargentii * P. acuminata.
It has been collected from two stations in Colorado, Welsenberg and Mont-
rose, both within the southern part of the overlap range of the parent
species, which are also represented from these stations. X Populus
Parryi has been collected from three localities in California, Canada de
las Uvas, San Bernando, and Owens Lake. It is the result of the cross
P. Fremontii * P. trichocarpa. K Populus Jacki (P. Tacamahaca X P.
deltoides) has been collected from some twenty localities scattered over
southern Quebec and Ontario, Vermont, and New York. This area
represents but a small part of the overlap range. In many cases both
parents have been collected from the same localities as the hybrid. The
hybrid P. grandidentata *« P. tremuloides is somewhat similar in both
these respects. It has been collected at twelve stations from Quebec
and New England west to Ohio. The probable hybrid P. acuminata
P. Wislizeni has been collected from one locality, Silver City, New Mexico.
The cross P. candicans * P. Tacamahaca is represented by one collection
made in the Arnold Arboretum. Four of these natural hybrids are the
result of intersectional crosses, while two are crosses between species be-
longing to the same section, P. grandidentata *« P. tremuloides and P.
candicans * P. Tacamahaca.

Any conclusion drawn from this study of natural hybridization are
necessarily limited by the fact that only those herbarium sheets contained
in the Arnold Arboretum herbarium were examined. In no case does the
distribution of the hybrid cover the entire range of overlap between the
parent species. In two cases the hybrid is represented by but two collec-
tions. From the numerous artificial hybrids reported one might expect to
find natural ae widespread. Their relative scarcity may be due to
several reasons. 1. These hybrids are capable of reproducing themselves
only through ered means. 2. Hybrids are scarce in those regions
in which relatively little collecting has been done and abound in those
regions where abundant collections have been made. 3. The most probable
reason, however, is that while growing in the same general geographic
region, the parent species may occupy different ecological habitats and
overlap rather infrequently. An attempt will be made in a later part
of this work to account for this lack of hybridization on the basis of

1943 | SMITH, CYTOLOGY AND SPECIATION IN POPULUS 289

TABLE 6.
NATURAL AND ARTIFICIAL INTERSPECIFIC HYBRIDS WITHIN THE
ENUS POPULUS ARRANGED TO SHOW CROSSING
WITHIN AND BETWEEN SECTIONS

Leuce Leuce Leuce Tacamahaca Tacamahaca Aegeiros
xX x x x x

Leuce Tacamahaca Aegeiros Tacamahaca Aegeiros Aegeiros
14 2 4 7 28 28

edaphic isolation. However, it can be shown that, in spite of such isola-
tion, most of the possible hybrids occur between species of the same
geographic region

Artificial interspecific hybrids in this genus have been produced in abun-
dance by different workers. Smith and Nichols (1941) list eighty-one
artificial interspecific hybrids which have been produced and described
by Henry (1914), Heimburger (1936, 1940), von Wettstein (1933),
Schreiner (1934), and others. Table VI shows how these hybrids, along
with the known natural hybrids, are distributed between and within the sec-
tions of the genus. The significance of this distribution will be discussed
in relation to speciation in a later part of this work.

Material for a cytological study of some of the artificial hybrids pro-
duced by Schreiner and his collaborators for the Oxford Pulp and Paper
Company was obtained from Dr. E. J. Schreiner of the Northeastern
Forest Experiment Station. The number of hybrids investigated was
unfortunately limited to those which happened to come in flower over the
period 1939-1941. Most of these were not the wide crosses which were
desired for study. Those hybrids from which collections were made are
described in Table VII, along with the natural hybrids investigated. The
latter were obtained from the collections in the Arnold Arboretum.

pairing and to study the various abnormalities which follow asynapsis.
In each case the pollen sterility was determined by a count of two
thousand pollen grains. In this connection it might be noted that the
sterility was variable, differing somewhat with different collections from
the same tree taken at the same or different times. Where possible, pro-
phase stages of meiosis were studied to obtain some idea of chromosome
pairing at the earlier stages. Table VIII summarizes the data on chromo-
some pairing and pollen sterility obtained from both the natural and
artificial hybrids studied.

CYTOLOGY OF HYBRIDS

In order that chromosomes may pair at meiotic metaphase, three gen-
eral conditions must be realized: first, that genetically similar chromo-
somes are present; second, that asynaptic genes do not influence the pair-

JOURNAL OF THE ARNOLD ARBORETUM (VOL, xxIVv

TABLE 7.

A LIST OF THE NATURAL AND ARTIFICIAL HYBRIDS OF POPULUS
INVESTIGATED, WITH DESCRIPTIONS OF THE CROSSES

WHICH THEY REPRESENT.

Name or number* Cross Description of cross
1. XP. Andrewsii P. Sargentii Between sections Tacamahaca
Aegeiros; both parents native to
P. acuminata North America.
2. XP. berolinensis P. laurifolia Cross within the section Aegeiros
between geographically _ isolated
P. nigra var. species.
italica
3. XP. canadensis Cross within the section Aegeiros
4. XP. canadensis _P.. deltoides between geographically isolated
var. Eugenei species, P. deltoides from North
5. XP. canadensis _P. nigra America, P. nigra from Europe &
var. regenerata western Asia.
6. XP. Jackii P. Tacamahaca Intersectional cross Tacamahaca

Aegeiros; parents native to North
P. deltoides America.

7. P. nigra var P. nigra var. Within the section Aegeiros; both
plantierensis italica parents varieties of the same Euro-
x pean species
nigra var.
betulifolia
8. XP. robusta P. angulata Within the section Aegeiros oo

geographically isolated speci

P. nigra var.
plantierensis

P. Simonii

od

Supposed true species; native to north China.

10. XP. Woobstii

P. laurifolia Cross within the section Tacama-
haca; parents native to western Asia.

P. tristis ?
OP-—64 P. charkoviensis Both parents probably hybrids;
OP-74 x cross might be written thus: (del-
OP-113 P. plantierensis toides XX ni a (nigra var.
italica X nigra var. betulifolia).
OP-96 P. nigra Cross between sections Tacamahac
OP-97 andA jae os eed,
OP-98 P. laurifolia isolated spec
OP-118
OP-99 P. charkoviensis Both parents probably hybrids
within section Aegeiros; (deltoides
P. clone robusta nigra) X (angulata X nigra var.

plantierensis).

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 291

TABLE 7 (continued).

Name or number* Cross Description of cross
OP-102 BP: a X P. bero- Parents within the section Aegeiros;
rare rossica male parent hybrid; deltoides x

nigra var. italica

OP-103 P. nigra X P. nigra Control cross.

OP-104 P. nigra X Intersectional cross between geo-
P. trichocarpa graphically isolated species

OP-105 P. nigra baatanicorum Crosses pega 2 oe related
vitrum X P. volga varieties of P. nig

OP-106 P. nigra baatanicorum
vitrum X P. plantierensis

OP-109 P. Rasumowskyana X Female parent probably the hybrid
P. caudina P. laurifolia x P. nigra; parents

OP-110 P. Rasumowskyana X probably belong to section Aegeiros.
unidentified cotton-wood

OP-111 P. charkoviensis P. charkoviensis probably hybrid P.

OP-114 x deltoides X P. ni gta; male parent
P. incrassata closely related to P. nigra; both be-

long to section Aegeiros.

OP-112 P. deltoides Control cross.

P. deltoides

OP-116 P. charkoviensis Both parents probably hybrids;
x< (nigra x deltoides) & (laurifolia x
P. berolinensis nigra var. italica); both belong to

section Aegeiros

QOP=117 P. charkoviensis (nigra > deltoides) xX gues
cross within the section Aegeiro
P. deltoides

OP-119 P. charkoviensis Female parent ee a
both within section Aegeiro
P. caudina

* Entries numbered 1-10 are natural hybrids. The re yoann entries are artificial
hybrids; clone numbers and parentages supplied by Dr. Schrein

ing; and third, that, after pairing, chiasmata form in the pachytene
chromosomes.

In the present study, while it is recognized that the latter two condi-
tions may play a part in the pairing behavior of the chromosomes at
meiotic metaphase, no attempt has been made to study the chiasmata
formation in the parental species and hybrids due to the extremely small
size of the chromosomes. In general, it was noted that the univalents
tend to be the smaller chromosomes of the complement. At this time
not enough is known of the genetics of these hybrids to determine to

292 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

TABLE 8.
MEIOTIC CHROMOSOME PAIRING, CHROMOSOME NUMBER, INVERSION
BRIDGE FORMATION, AND POLLEN STERILITY IN NATURAL
AND ARTIFICIAL HYBRIDS OF POPULUS.

Pollen
Name or Chro. Metaphase I Analysis Anaphase I Analysis Sterility
Number No. I II III L B er cent
1. 19 0.72 18.64 0 0.72 0.14 1 20
2: 19 5.30 16.20 0.10 2.00 0.12 1 57
3. 19 1.84 18.08 0 0.60 0.12 2 17
4. 19 4.92 16.48 0.04 2.90 0.33 3 63
5. 19 3.92 17.04 0 3.69 0.04 1 56.5
6. 19 2.40 17.80 0 62
7. 19 0 19 0 6
8. 19 6.00 16 0 2.50 0.13 2 75
9. 19 1.76 18.12 0 31
10 19 14.24 11.88 ) 80
OP-64 19 0.28 18.86 0 0.16 0.23 1 5
OP-74 19 0.64 18.68 0 0.20 0.08 1 24
OP-113 19 1.14 18.43 0 0.16 0.23 1 22.4
OP-96 19 0.40 18.80 0 20
OP-97 19 1.00 18.50 0 0.40 0.20 1 16
OP-98 19 0.20 18.90 0 10
OP-118 19 0.80 18.60 0 0.40 0 0 25
— 19 1.00 18.50 0 26
OP-102 19 1.20 18.40 0 20.4
OP-103 19 0.20 18.90 0 5
OP-104 19 0.80 18.60 0 0.66 0.33 2 23
OP-105 19 0.20 18.90 0 0.26 0.06 1 10
OP-106 19 0 19 0 0.06 0 0 6
OP-109 19 0.60 18.70 0 0.86 0.53 3 12
OP-110 19 2.20 17.90 0 1.25 0.30 2 40.5
OP-111 19 0.20 18.90 0 0.40 0 0
OP-114 19 0.40 18.80 0
OP-112 19 0.20 18.90 0 0.46 0 0 6
OP-116 19 3.00 17.50 0
OP-117 19 1.40 18.30 0 1.00 0.26 3
OP-119 19 0.76 18.62 0 0.44 0.12 2 35

I, IJ, and III under metaphase ashes refer to the average number of univalents, bi-
valents, and trivalents per cell. , and M under r anaphase analysis refer to the average
number of lagging chromo cannes: eerie number of inversion bridges and maximum
number of inversion bridges per cell.

what extent asynaptic genes influence pairing relationships. Environmen-
tal factors are also known to cause asynapsis.

Two classes of genetic dissimilarity of the chromosomes are recognized.
The first, which is purely genic and presumably arises through gene
mutation, is usually not assigned a large role in asynapsis. The second
type of dissimilarity is structural and is brought about by rearrange-
ments of genic material within the chromosome. It is usually assumed
that this type of dissimilarity plays the major role in asynapsis

1943 J SMITH, CYTOLOGY AND SPECIATION IN POPULUS 293

Darlington (1937) discusses the classification of hybrids at some length,
dividing them into seven classes: numerical, structural, undefined struc-
tural, complex, polyploid, numerical-structural, and Mendelian hybrids.
Under this classification the interspecific hybrids in Populus considered
here would be placed in the undefined structural hybrid class, these “ . .
resulting from the union of gametes dissimilar as a result of changes which
cannot be defined . . . simply because the structural differences between
their chromosomes are too slight or too numerous to be detected.”

The undefined structural hybrids are further broken down into those
which show potentially complete pairing at metaphase I, those which
show partial pairing, and those which have a potentially complete failure
of pairing. Evidently the interspecific hybrids of Populus investigated
thus far might be placed in the first two groups, although the larger
number belongs to the second, for these hybrids show a partial and always
variable pairing.

Numerous examples may be cited for each of the hybrid classes listed
above. Among the undefined structural hybrids which show potentially
complete pairing are Salix viminalis *« S. caprea (Haakanson, 1929),
Platanus orientalis * P. occidentalis (Sax, 1933), and Catalpa bignonioides
x C. ovata (Smith, 1941). The hybrids Viola arvensis & V. rothmagensis
(Clausen, 1931) and Ribes nigrum R. Grossularia (Meurman, 1928) are
examples of undefined structural hybrids which show partial pairing.

By the cytological examination of the interspecific hybrids from pachy-
tene stage of prophase onward it is possible to demonstrate that a vary-
ing number of bivalents and univalents are present (Figs. 1, 18, 21).
Within a single hybrid the number of univalents present per cell at
metaphase I may vary from none to thirty-eight (Figs. 78, 19, 21). Con-
sidering the hybrids studied as a whole, the number of normal cells,
normal in the sense that they contained nineteen bivalent chromosomes
at metaphase I, varied from four to ninety-six per cent, with an average
of fifty-three per cent.

Univalent chromosomes were present in varying numbers in all the
hybrids examined. Usually these univalents lie on either side of the
metaphase plate (Fig. 18), come onto the plate after the bivalent: chromo-
somes have divided, and then divide (Fig. 21). There is some evidence
that some univalent chromosomes do not divide at anaphase I but go
to the poles without lagging or dividing. On the basis of an examination
of five hundred each of metaphase I and anaphase I figures (twenty-five
each of twenty hybrids), it was found that more univalents were present
at metaphase I than appeared as lagging univalents at anaphase I. The
average was 1.65 univalents at metaphase I as compared to 0.96 univalents
at anaphase I. Since no univalent chromosomes were observed dividing
at anaphase II it seems likely that those univalent chromosomes which
did not divide at anaphase I behave in a normal manner at the following
division.

After dividing, the lagging univalents may or may not reach the poles

294 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV
TABLE 9.

A COMPARISON OF SUPERNUMERARY SPORES AND UNIVALENT
CHROMOSOMES IN SPECIES AND HYBRIDS OF POPULUS.

Species or Univalents Supernumerary
hybrid | spores per cell
P. deltoides 0 0
P. alba (diploid) 0 0
P. nigra 0 0

<P. Andrewsii 0.72 0.55

xP. berolinensis 5.30 0.32

xP. robusta 6.00 0.47

xP. canadensis 1.84 0.30

in time to be incorporated into the daughter nuclei. If they do not reach
the poles they are lost in the cytoplasm (Fig. 23). If they are included
in the dyad nuclei they are distributed at random to the poles at anaphase
II or are lost in the cytoplasm. On the basis of one hundred anaphase II
figures analyzed in two hybrids it was found that three times as many
univalent chromosomes were lost at anaphase I as there were at anaphase
II. The ultimate fate of these univalent chromosomes seems to depend
upon the rapidity with which they progress to the poles at both anaphase
I and II. In both cases if they reach the poles before the daughter
nuclei are formed they are included in these nuclei, otherwise they are
left behind in the cytoplasm, where they form either micronuclei or super-
numerary spores. Just what determines their fate is not clear. It does not
seem to depend upon the number of univalents available, for the nucleus
of the supernumerary spore is as small as the micronucleus. The factor
determining this may be the position of the univalents at the time of the
cell wall formation, those near the microspore nuclei becoming micronuclei
and those farther out becoming supernumerary spores.

That the presence of supernumerary spores is aes with the
presence of univalent chromosomes is shown by Table

An anaphase analysis of these interspecific hybrids ede in nearly
every case a varying number of inversion bridges (Table VIII; Figs. 24—
31). The condition in which a portion of a chromosome is present in the
inverted state is one of the most frequently encountered meiotic aberra-
tions. This condition can be detected in plants when a crossover occurs
within the heterozygous inversion region, for as a result chromatin bridges
are formed at anaphase I. A loop pairing at pachytene is also charac-
teristic. The occurrence of inversions in both plants and animals has
been reported by many writers, among them Muntzing (1934), Richard-
son (1936), and Stebbins (1938). Structural hybridity has not previously
been reported in the genus Populus. Haakanson (1929) has reported a
case of reciprocal translocation in Salix.

Following the occurrence of a crossover within the heterozygous in-
version, a dicentric chromatid and an acentric fragment should be pro-

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 295

duced at anaphase I. The fragment usually lies in the cytoplasm in the
vicinity of the bridge and varies in size with the length and position of
the inverted region. Only five cases were encountered in these hybrids
where the fragment could be seen associated with the bridge (Figs. 25, 30).
This is not an unusual condition, for Swanson (1940) finds that in
Tradescantia fifty per cent of the bridges studied lacked visible fragments.
Sax (1937) and Darlington (1937) report a somewhat similar condition.
These investigators account for the lack of fragments on the basis of the
presence of small subterminal inversions, which result in fragments which
are below the limit of visibility or are obscured by other chromosomes
of the complement.

It is possible that the number of bridges observed was but a fraction
of those which actually occurred, since the bridges formed in the smaller
chromosomes would break in very early anaphase or stretch so thinly
that they could not be seen. Usually the inversion bridge breaks and the
two parts of the chromosome reach the poles, but occasionally the bridge
fails to break and remains in the cytoplasm following the first division
(Fig. 31)

No bridges were seen in material from control crosses or from the
species of Populus examined.

It is of interest that van Dillewijn (1940) notes the presence of chro-
matin strings stretching between the two anaphase plates at late anaphase
I in P. robusta. Meurman (1933) figures the heteromorphic sex-chromo-
some pair in P. Simoni (a probable hybrid) as lagging at anaphase I and
resembling a chromosome bridge. It is possible that what in the past
have been taken to be sex-chromosome pairs at anaphase are in reality
inversion bridges. This seems especially likely since the members of the
heteromorphic pair seem to differ but slightly in size. In the hybrids
studied here, the possibility that what appears to be inversion bridges
are in reality dividing sex-chromosomes seems to be ruled out in those
cases where two or more bridges were seen in a single cell.

In general it may be said that pollen sterility is due to one or more
of three possible factors: purely genetic, structural, and environmental.
It is not possible to separate these three causes in the case of interspecific
hybrids of Populus. From an examination of pure species within this
genus it is apparent that pollen sterility varies from 2—7 per cent in those
species showing complete chromosome pairing and from 19-45 per cent
in those species showing a varying number of univalent chromosomes at
metaphase I. Presumably the sterility in the first group was due largely to
genetic causes, while that of the second group was due to both genetic and
environmental causes. Presumably in the case of the interspecific hybrids
all three factors contribute to sterility. At present it is impossible to de-
termine to what extent each of these exerts its influence. There does seem
to be a correlation between univalent formation and pollen sterility, as
had already been noted in the case of the species studied. In general
those hybrids with the higher number of univalents show the higher sterili-
ties. The correlation is high, since r equals 0.88.

296 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XX1V

TABLE 10.
A COMPARISON OF CERTAIN NATURAL AND ARTIFICIAL INTERSPECIFIC
HYBRIDS OF POPULUS IN REGARD TO DATE OF ORIGIN, UNIVALENT
CHROMOSOME FORMATION, AND POLLEN STERILITY

Probable Univalents
ate of per Percentage
Name or number Cross origin cell of sterility
Natural Hybrids:
XP. berolinensis P. laurifolia < 1870 5.30 57
ae
italica
> P. canadensis var. P. deltoides X 1850 4.92 63
Eugenei P. nigra
Artificial Hybrids:
OP-67 P. nigra X 1925 1.00 16
P. laurifolia
OP-114 P. nigra 1925 0.80 23

P. trichocarpa

The failure of chromosome pairing in these hybrids, followed by a loss
of chromosomes or unequal distribution of chromosomes to the microspores,
results in deficiencies and duplications of entire chromosomes. In the
case of deficiencies, at least, this would lead to the sterility of those
microspores deficient for one or more chromosomes. The loss of frag-
ments of chromosomes through the formation of inversion bridges would
also result in deficiencies for parts of chromosomes.

Peto (1938) notes that the pollen sterility in the hybrids within the
section Leuce compares favorably with the sterility of the parent species.
In the case of the hybrids considered here, it seems that the natural hy-
brids have an average sterility which is considerably higher than that of
the parent species. The artificial hybrids, on the other hand, possess
sterilities corresponding to those of the species, ranging from 5—40 per
cent with an average of 19.7 per cent. The natural hybrids, however, range
from 6-80 per cent with an average of 46.7 per cent pollen sterility.

It is known that the mutation rate increases with age (Cartledge and
Blakeslee, 1934, and 1935). It might be expected that those physiological
changes in the cell which condition this increased mutation rate might
cause an increase in susceptibility to environmental influences, and, in
addition, that structural changes occurring over a long period of time
might accumulate to produce a greater pollen sterility. Certain natural
and artificial interspecific hybrids of Populus are compared in Table X.

Some of the common meiotic abnormalities found in the interspecific
hybrids of Populus have been mentioned. Among these were asynapsis,
with subsequent lagging univalent chromosomes at anaphase I and II,
and irregularities in cytokinesis within the pollen mother cell which result
in dyads, triads, and number of microspores in excess of four. A pre-

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 297

TABLE 11.
A COMPARISON OF INTERSECTIONAL WITH INTRASECTIONAL CROSSES OF fas AND
* CROSS BETWEEN NON-GEOGRAPHICALLY ISOLATED SPECIES H
CROSSES BETWEEN GEOGRAPHICALLY ISOLATED SPECIES.

Average Average

Class of hybrid Number of number number Average

hybrids univalents bridges Per cent

Metaphase | Anaphase I Sterility

Crosses within
the section 13 3.22 0.19 37.9
Crosses between
sections 7 0.90 0.17 25.0

sses between
ally

non- scosraphicaly
isolated spec 3 5.78 0.14 54.0

cocious furrowing was a common abnormality of this sort. It was not
uncommon to find that the furrowing process was well advanced before
the completion of the second meiotic division. Muntzing (1936) and van
Dillewijn (1940) describe cases of spindle fusion during the meiotic
divisions in Populus, which result in microspores with the unreduced
number of chromosomes. This is a possible cause for the occurrence of
autotriploids within the genus. What appears to be a case of a third
division in the pollen mother cell before microspore wall formation is
shown in Fig. 22. Four spindles have formed and the thirty-eight chromo-
somes present have apparently been distributed to the poles at random.

DISCUSSION

Cytological and genetical studies of interspecific hybrids have been
used by many investigators to establish probable interspecific relation-
ships in plants and animals. The cytological study of species hybrids
based upon chromosome configurations and sterility counts has been
particularly useful, since it can be used as a basis for the determination
of the manner in which speciation has taken place.

In the one varietal hybrid of Populus studied, P. nigra var. plantierensis
(P. nigra var. italica * P. nigra var. betulifolia), the chromosome be-
havior and pollen sterility is comparable to that of a pure species.

Since Populus has been subdivided into four sections, and since the
species within a section resemble one another more than they resemble
those of other sections, it might be supposed that intrasectional hybrids
would be more easily obtained and would show a lower percentage
asynapsis and pollen sterility than intersectional hybrids. Table XI com-
pares crosses between and within the sections Tacamahaca and Aegeiros,

298 JOURNAL OF THE ARNOLD ARBORETUM [vOL, XXIV

which are perhaps less distinctly set off from one another than from the
other sections of the genus. From the table it is clear that there is no
significant difference in regard to asynapsis and pollen sterility between
crosses within and those between these sections.

A glance at Table VI will show that the majority of the crosses are
inter- and intrasectional crosses involving species belonging to the two
sections Tacamahaca and Aegeiros. This is to be expected, since these two
sections contain more species than do the others. Further, the species
producing the better timber trees of the genus are placed in these sec-
tions. No hybrids are known either within the section Leucoides or be-
tween species of this section and those of the other sections of the genus.
Interspecific hybrids are known, however, within and between the sec-
tions Leuce and the sections Tacamahaca and Aegeiros. Table VI seems
to indicate that crosses within the section are more easily obtained than
are intersectional crosses (except Tacamahaca and Aegeiros). Heimburger
(1940) sees a definite limitation to species hybridization based on genetic
affinities which cause crossing to follow a series similar to the series aspens
—silver poplars — cottonwoods — balsam poplars. However, Johnson
(1939) states that there appears to be little limitation to species hybridi-
zation within the genus, as far as artificial hybridization is concerned.

A comparison of the cytological behavior of hybrids between geographi-
cally isolated species is of interest, inasmuch as it permits a study of the
effects of isolation upon speciation over long periods of time. Table XI
compares hybrids which result from crossing of geographically isolated
species with hybrids resulting from the crossing of non-geographically
isolated species, in regard to asynapsis, inversion bridge formation, and
pollen sterility. It is apparent that there is no significant difference be-
tween the two groups. Species of Populus native to North America cross
readily with European and Asiatic species to produce hybrids which are
as fertile as those resulting from crosses between native North American
species.

The term ecospecies, defined by Turesson (1922) as uniform types
between which crossing is possible with a relatively high degree of
fertility but which commonly are prevented from doing so by isolating
barriers, either edaphic or geographical, would seem to apply to species
of Populus. An examination of hybrids between certain species reveals
a considerable amount of sterility; however, the Fj and backcross gen-
erations which have been obtained show a segregation indicative of an
exchange of genes between the two species.

Species as discrete units can exist as such only by means of some isolat-
ing mechanism. Various classifications of these mechanisms have been
made (Sax, 1936; Dobzhansky, 1941). In general they may be divided
into five classes: edaphic (adaptation to particular local habitats), geo-
graphic, physiological (probably genetic, but in this case referring to
flowering time), chromosomal (either numerical or structural), and
purely genetic.

1943 | SMITH, CYTOLOGY AND SPECIATION IN POPULUS 299

TABLE 12.

HABITATS AND BLOOMING TIMES OF CERTAIN SPECIES OF PopuLus
WHICH OCCUPY THE SAME GEOGRAPHIC REGION

Physiological Edaphic
Species blooming time in ecological
Arnold Arboretum habitat
P. grandidentata April 20/40 Moist sandy soil,
gravelly hillsides
P. tremuloides April 20/40 Rich moist soil, borders
of streams and swamps
P. deltoides May 5/40 Low, river-bottomlands
P. Tacamahaca May 4/40 Low, often inundated
river-bottomlands,

wamp borders

Considering the three sections of the genus which have been studied,
it is clear that artificial hybrids may be made in any direction. The F,
hybrids have proved to be relatively fertile. However, at this time in-
sufficient data are at hand to determine with certainty the viability of the
second generation. It would seem that genetic isolation itself could not

i

Autotriploid forms of a number of species ‘within the section Leuce
have been described. These, though relatively fertile, are unable to
preserve their identity. No other polyploid forms are at present known in
this genus. Polyploidy (numerical chromosomal isolation) is ruled out
as a factor in speciation within this genus. Evidence has been advanced
to show that Populus is a derived or secondary polyploid, so polyploidy
may have played a part in the ancestral differentiation of the genus. The
discovery of inversion bridges in the F, interspecific hybrids and the
asynapsis present in most hybrids indicate considerable structural dif-
ferentiation of the chromosomes of the different species. But, as in the
case of genetic isolation, it does not appear to prevent the crossing of the
species or the production of a relatively fertile F, generation.

striking example of physiological isolation is found in this genus.
The species included within the section Leuce are earlier in blooming time
by two to three weeks than are those of the other species of the genus.
They are thus definitely set off in nature. This isolation may be over-
come artificially (Smith and Nichols, 1941). However, the major factor
in isolation of the species of Populus seems to be isolation of a geographic
and edaphic nature. Both types are essentially the same, geographic isola-
tion being perhaps more complete and on a larger scale than edaphic
isolation.

The species P. grandidentata, P. tremuloides, P. deltoides, and P.
Tacamahaca occupy a somewhat similar area in northeastern United
States and the adjoining region of Canada. Table XII summarizes their
habitats and blooming times.

300 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Populus grandidentata and P. tremuloides are set off from the remaining
two species by their time of blooming. Presumably they are themselves
isolated by edaphic factors, since their habitats differ somewhat. However,
the hybrid P. grandidentata < P. tremuloides does occur. Similarly
the hybrid & P. Jackii (P. Tacamahaca X P. deltoides) occurs rather
frequently, although presumably the parental species are separated by
edaphic isolation.

A consideration of evidence available would indicate that the first step
in speciation in Populus may have been a physiological isolation of a
group or groups of species from the general population, since the major
geographic groups of Populus species all contain species belonging to the
section Leuce. This physiological isolation was in turn followed by geo-
graphic and edaphic isolation. The structural-chromosomal differentia-
tion which has taken place since that time has been insufficient to pre-
vent interspecific hybridization and the production of reasonably fertile
hybrids.

The slow accumulation of genetic differences through mutation has
probably been the major factor in the differentiation of the species of
Populus, but apparently these species can exist as discrete units only as a
result of physiological isolation, in the case of the section Leuce, and
geographic or edaphic isolation in the case of the individual species.

ACKNOWLEDGMENTS

The writer acknowledges with gratitude the many valuable suggestions
and criticisms of Professor Karl Sax, under whose direction this investiga-
tion has been carried out. The guidance of Professor Alfred Rehder in
the study of the distribution of species of Populus and the occurrence of
natural hybrids, and the kindness of Dr. Ernst J. Schreiner, in supplying
cytological material of the various artificial hybrids, are gratefully
acknowledged.

SUMMARY

The chromosome numbers of thirty-eight species, varieties, and natural
hybrids of Populus have been determined, twelve of which were in con-
firmation of previous work by other investigators. All species exist as
diploids with an unreduced chromosome number of thirty-eight. In the
case of three species, all within the section Leuce, triploid forms with an
unreduced chromosome number of fifty-seven exist.

The chromosomes of the hybrid P. nigra P. trichocarpa at the
pachytene stage of the meiotic prophase vary in length from 30.6 to 8.5
u in length. One group of three and two groups of two chromosomes are
of exactly the same length. No heteromorphic chromosome pair which
might be interpreted as a sex-chromosome pair has been seen consistently.

The occurrence of secondary association of chromosomes has been
noted by other workers. Additional evidence for secondary polyploidy
is found in the fact that morphological similarities exist among certain
groups of chromosomes and that occasional trivalent associations of

1943 ] SMITH, CYTOLOGY AND SPECIATION IN POPULUS 301

chromosomes are found in hybrids and quadrivalent associations in the
triploid form of P. alba.

The instigation and progress of the meiotic divisions within the catkin
are consistent within the species and within the section of the genus. These
divisions are regular as a rule, but in a few species a certain amount of
asynapsis and abnormal pollen sterility is encountered. This is not inter-
preted as due to any genetic dissimilarity between the chromosomes.

One chromosome of the complement is invariably associated with the
nucleolus at meiotic prophase. In the majority of cases the nucleolus
possesses a knob which projects from the point of attachment of the
associated chromosome.

The triploid form of P. alba shows a varying number of univalent,
bivalent, trivalent, and quadrivalent chromosome configurations at meta-
phase I. The fact that the pollen sterility of the triploid forms is somewhat
less than that to be expected on the basis of univalent formation is ac-
credited to the fact that the basic set of chromosomes is triplicated and
thus genetic unbalance, due to the loss of whole chromosomes, is less
likely to occur.

Cytokinesis is of the type usually found in cases of simultaneous pollen
cell formation. The process seems to be easily upset in the case of
triploid forms and interspecific hybrids of Populus.

A study of the development of the male gametophyte from the first
microspore division up to the time of fertilization shows this to be typical
of the process as found in the Angiosperms.

The geographic distribution of the North American species of Populus
has been investigated, especially as related to the occurrence of natural
hybrids. The natural and artificial hybrids which have been reported
are arranged in tabular form to show the extent of crossing within and
between sections of the genus.

The study of chromosome behavior at meiosis in ten natural and twenty-
one artificial interspecific hybrids shows them to be variable in regard
to synapsis. They are probably structurally undefined hybrids under
Darlington’s classification. It has been demonstrated that the chromo-
somes of the different species have been differentiated structurally and
that inversion bridges are of relatively frequent occurrence. Pollen
sterility within the artificial hybrids is of the same order as that within
the parental species, but that of the natural hybrids is somewhat higher.
An attempt is made to explain this fact on the basis of an accumulation
of structural changes within the chromosomes over the relatively longer
period of time these natural hybrids have been in existence as clones. No
significant difference in pollen sterility, univalent formation, or inversion

S Ww
graphically isolated species and non- -geographically isolated species re-
veals no significant difference.

A review of the isolating mechanisms which may operate in this genus
seems to indicate that geographic and edaphic isolation are those which

302 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

separate the species, although one entire section is set off from the other

sections of the genus by a physiological isolation. It seems from the evi-

dence available that differentiation of the species of Populus has been
low

mutation. These species can exist as discrete units, however, only through
the operation of geographic, edaphic and, in a few cases, physiological
isolation. Genetic and structural-chromosomal isolation plays a relatively
minor part in species isolation.

LITERATURE CITED

Bercstrom, I. (1940). On the progeny of diploid x triploid —. tremula with
pie aig to the occurrence of tetraploidy. Hereditas 26: 01.

CKB , K. B. (1926). On the occurrence of sex sheamsannlae in fowert ing plants
with some suggestions as to their origin. Proc. Int. Congr. Plant Science Ithaca
N. Y. - te

d J. W. H. Harrison (1924). A areca sy account of the chromosomes

and aacte behavior in the Salicaceae. Ann. . 38: 361-378.

Ge aesnee S. C. (1937). Ett fynd av jatteasp ia tremula gigas) i Medelpad.

CarTLEDGE, J. L. and A. F. erage Rrcalig Mutation i nae by ageing
seed as shown by pollen abortion. . Nat. Acad. Sci. 20:
1935). Mutation rate a old Datura seeds. seh 81: 482
CasTETTER, E. F. (1925). Studies on — comparative cytology of the a and
biennial varieties of Melilotus alba. Am. Jour. Bot. 12: 270-286.
CavanacH, L. M. (1930). Abnormal eer of Populus grandidenta Michx. Proc.
te) 116.

CHAMBERLAIN, C. J. (1897). Contributions to the life history of Salix. Bot. Gaz.
23: 147-179

CLauseEn, J. (1931). Cyto-genetic and taxonomic investigations in melanium violets.
Hereditas 15: 66.

Darvincton, C. D. (1937). Recent advances in cytology, 2nd edition, P. Blakiston’s
Son and Co. as Philadelphia

37). Chromosome ‘beliavior and structural hybridity in the Tradescantiae

II, biped Genet. 35: —280.

———— and A. Morrer, tS Primary and secondary chromosome balance
in Pyrus. pa Genet. 22:

DiILtewijn, C. VAN narra 1 ake Studien in der Gattung Populus L.

DoszHANSKy, T. (1941). Genetics and the origin of species. New York City:
Columbia University Press.

Ertanson, E. W. and F. J. HermMann (1927). The morphology . cytology of
perfect gees in Populus tremuloides Michx. Papers Mich. Aca i. 8: 97-110.
Farr, C. (1916). bree of the pollen mother cells of stain Dicotyledons.

ma a Gard. 6: 253-317.
GraF, J. (1921). Beitrage zur annie der Gattung Populus. Beih. Bot. Central.
4.

Haakanson, A. (1929). Die pear onesie in der Kreuzung Salix viminalis x Caprea
von Heribert Pe Hereditas 13:

Hastincs, G. T. (1918). Some ania sale flowers. Torreya 18:

HEIMBURGER, C. nen Report on Poplar hybridization. adr "ails

———— (1940). Report on Poplar hybridization II. Forestry Chronicle 16:
149-160.

EITZ, E. (1931). Die Ursache der oe Zahl, Lage, Form und Grosse
pflanzlicher Nukleolen. Planta 12: 775

1943 | SMITH, CYTOLOGY AND SPECIATION IN POPULUS 303

Henry A. (1914). The nen production of vigorous trees.
and Tech. Inst. Ireland 15:

Petre N. M. (1939). ene investigations on the genus Cicer. Ann. Bot. N.S.

271-

JOHNSON, oa V. (1939). A nee list ee paige and aes interspecific
hybrids in ce American forest-tree genera. . Jour : 411-4

Jounsson, H. (1940). Cytological ae — diplid ne eerie: Pips eee and
of crosses beaelies them. Hereditas 26: 32

oe W. J. C. (1931). The aie sain er of chromosomes. Cytologia

Jour. Dept. Agric.

3-384.
Mc one B. (1934). The relation of a particular eae element to the
development se the nucleoli in Zea Mays. Z. Zellforsch 21: 29
Mevurman, O. The chromosome behavior of some ics ene and their
relatives with speci reference to the sex chromosomes. Comm. Biol. Soc. Sci.
Fenn. 2: 1-

a. Cytological studies in the genus Ribes L. Hereditas 11: 289-336.
——— (1933). Chromosome morphology, somatic doubling and secondary asso-
ciation in Acer platanoides L. Hereditas 18: 145-173.
Munrtzinc, A. (1934). Chromosome fragmentation in a Crepis hybrid. Hereditas
19: 2
(1936). The chromosomes of a giant Populus tremula. Hereditas 21:
383-393
NAKASIMA, G. (1937). Sera studies in some dioecious plants. Cytologia:
Fujii. Jub. Vol.: 282-29
Nanor, H. K. (1936). oe ear ee oe secondary association and origin
of cultivated rice. Jour. Genet. 33: 36.
NEwcomeeEr, E. H. (1938). A ie io growing, staining and making permanent
slides of pollen ee es Techn. 13: 89-91.
sinc EHLE, H. (19 a eine in - Nature gefundene Gigasform von Populus
emula. ee 21: 379-
1938). lees av skogstrad meddkat kromosomtal och Okad
ou Svensk. Paperstidning
NouarA, - (1924). Experimental studies on pollen of some Salix. Jap. Jour. Bot.
2: 1-3

ee N. (1938). ae studies in Oryzeae and Phalaridese IJ. Further
studies in Oryza. Cytologia 9: 307-318.

PAUL] RK. (1937): Seana in Tamarindus indica L. Cytologia 8: 48.

Peto, F. H. (1938). Cytology of poplar species and natural hybrids. Canad. Jour.
Res. 16: 445-455.

ReuHp_ER, A. (1940). Manual of cultivated trees and shrubs. 2nd ed. Macmillan Co.,

ew ae rk.

Ricuarpson, M. M. (1936). Structural hybridity in L. Martagon x L. Hansonit.
Jour. Genet. 32: 411-450.

Sax, K. (1932). The cytological mechanism of crossing over. Jour. Arnold Arb.

33). Species hybrids in Platanus and Campsis. Jour. Arnold Arb.
(1936). The cytological analysis of species hybrids. Bot. Review I:

Chromosome behavior and nuclear development in Tradescantia.

Genetics 22: 523-533

SCHREINER, E. J. and ” B. Stout (1934). Descriptions of ten new hybrid poplars.
Bull. Torrey Hot Club 61: 449-460.

Seurm, A. G. (1930). A cytological study of Oryza sativa L. Cytologia 2: 1-

SmitH, E. C. eal a ie romosome behavior in Catalpa hybrida ae Jour.
Arnold Arb. 22:

————- and C. aii (1941). Species hybrids in forest trees. Jour. Arnold Arb.
22: 443-454.

304 JOURNAL OF THE ARNOLD ARBORETUM [vOL, XXIV

Smitu, F. H. (1933). The relation of the satellites to the nucleolus in Galtonia
candicans. Amer. Jour. Bot. 20: 188-195.

Srespins, G. L. Jr. (1938). Cytogenetic — in Paeonia II. The cytology of the
diploid species and hybrids. Genetics 23: 83-110.

Swanson, C. P. (1940). The ieee of inversions in Tradescantia. Genetics

Tesnivnen: G. a 937). The chromosome numbers of two new giant Populus tremula.
Bot. Notiser 1937: 285-290.
Turesson, G. (1922). The genotypic response of the plant species to the habitat.

50.
WetrTstTEIN, W. von (1933). Die Kreuzungsmethode und die Beschreibung von Fy
Bastarden bei Populus. Zeitschr. Zuchtung 18: 597-626

DESCRIPTION OF PLATES
All drawings are at a magnification of x 2090.
PLaTE I

Fic. 1. Pachytene stage of meiosis of the hybrid P. nigra x P. trichocarpa. One
chromosome is associated with the nucleolus at the junction of the knob and the nucle-
olus proper. Failure of synapsis is indicated by arrows. Camera lucida drawing.

os

nd t
Fic. 3. Metaphase of the first microspore division in P. acuminata. Fic. 4. Generative
nucleus in the pollen tube of P. laurifolia twelve hours after germination. Fic.
Metaphase of the division of the generative nucleus in the pollen tube of P. Lometiolis,
Fic. 6. Anaphase of the division of the generative nucleus in the pollen tube of P
laurifolia. Fic. 7. Pollen tube showing tube nucleus (at lower end of tube) and pa
gametes of P. deltoides.

Pate II

Fic. 8. Late diakinesis in P. Sargentii. Nineteen bivalent chromosomes present, one
of which is associated with the nucleolus. Fic. 9. Metaphase I in P. acuminata, show-
F a

f
alba (diploid), with nineteen bivalent oes Fic . 12. Metaphase rj in P, alba
(triploid). Nine univalent, seven bivalent, ten trivalent, me one quadrivalent chromo-
somes are present. Fic. 13. Anaphase I in P. laurifolia, showing nineteen els omosomes
at each pole. Fic. 14. reine II in P. alba (diploid). Fic. 15. Anaphase I in
P. acuminata. Two lagging univalent chromosomes are shown in division

Pirate III

Fic. 16. Late diakinesis in the hybrid x P. Rasumowskyana x P. caudina, with
nineteen bivalent chromosomes present. Fic. 17. Metaphase I in x P. Woobstii show-
ing thirty-eight univalent chromosomes. Fic. 18. Metaphase I in P. nigra * P. lauri-

nineteen bivalent chromosomes on the plate. Fic. 20. Regular ioe nig

P. laurifolia, with nineteen chromosomes at each pole. Fic. ed ee I
in P. nigra X P. laurifolia; thirteen univalent chromosomes pti on the plate;
irregular distribution of chromosomes to the poles. Fic. 22. A third division bere
the pollen mother cell in x P. charkoviensis - P. Piece with a random distribu-
tion of the chromosomes to the poles. Fic. Anaphase II in x P. Rasumowskyana
x P.caudina. Apparently two lagging chromosome have been lost at the first divi-
sion and one is lagging at the second divisio

Jour. ARNoLD ARB. VoL. XXIV

PLATE J

II|

2 CP ¢
wee, | Safe
Se eS
7 6
——— CO

CYTOLOGY AND SPECIATION IN PopuLUS

Jour. Arnotp Ars. VoL. XXIV PraTeE II

CYTOLOGY AND SPECIATION IN POPULUS

Jour. ARNOLD ARB. VoL. XXIV PLATE III

16
18
19 “ a i
©
osctt °
TSS Fh
orgee
2!

CyTOLOGY AND SPECIATION IN POPULUS

Jour. ARNOLD ARB. VoL. XXIV PLATE IV

25
26
28
29
30 3]

CYTOLOGY AND SPECIATION IN POPULUS

1943 | SMITH, CYTOLOGY AND SPECIATION IN POPULUS 305

PiaTE IV

4. Early anaphase I in xP. charkoviensis x XP. berolinensis, showing
an ae bridge without a visible fragment. Fic. 25. Anaphase I in x P. charko-
eae: . . f :

vi erolinensis, with an inversion bridge and fragment I na-
phase I in XP. char ory bu Re of an inversion brid
Fic. 27. Ana ase I in . Rasumowskyana xX P. caudina, showing two inversion

P. laurifolia. Tnversion bridge, no fragment. Fic. 30. Late anaphase I in
charkoviensis X XP. berolinensis. Pade and fragment. Fic. 31. Unbroken inversion
bridge left out in cytoplasm at the first meiotic division in x P. Andrewsii

BIOLOGICAL LABORATORIES,
Harvarp UNIVERSITY.

306 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

PLANTS OF COAHUILA, EASTERN CHIHUAHUA, AND
ADJOINING ZACATECAS AND DURANGO, I

IvaAN M. JOHNSTON

In this, the first of a projected series of papers, I begin an enumeration
and analysis of the vascular flora found on the great arid plateau lying
between the Eastern and Western Sierra Madre of northern Mexico.
Predominantly a region of Cretaceous calcareous rocks, it contains many
broad silty valleys, most of them bolsons, with extensive areas of desert
scrub dominated by Larrea and Flourensia, and many scattered limestone
mountain ranges, some of which support oak-chaparral on the higher ridges
and in their upper canyons, and a few of which are lofty enough for the
development of small coniferous forests. Igneous rocks, scattered and
local in the eastern half of the plateau, increase in abundance and fre-
quency toward the base of the predominantly igneous Western Sierra
Madre and prevail in large areas in eastern Chihuahua, where grasslands
and grassy hills with liveoaks are the characteristic types of vegetation.

The western limit of the region studied extends, roughly, from the area
about Lake Guzman, in northwestern Chihuahua, southeast to near
Chihuahua City and then southward in an irregular line west of Camargo
and Jimenez to include the northern and eastern parts of the State of
Chihuahua. Continuing southward, the limit passes west of Mapimi and
Torreon to include the northeastern portions of the State of Durango.
As the eastern limits of the area, I have followed the eastern boundary of
the State of Coahuila. This includes parts of extreme eastern and north-
eastern Coahuila which are not parts of the plateau. Since, however, the
few collections available from these districts represent a flora which, in
large part, has extended up onto the plateau in the east-central parts of
Coahuila, it has seemed practical and interesting to include in the cata-
logue, at least, records of all the collections available from the entire State
of Coahuila.

The southern boundary is indefinite; in practice it roughly follows the
latitude of the southeastern portion of the State of Coahuila. This includes
in our area the northern portions of the State of Zacatecas. The northern
limit follows the international boundary. Although the eastern and
western limits of the areas studied are roughly natural, the northern and
southern limits are not so. The plateau flora treated continues northward
from our area into southern New Mexico and into trans-Pecos Texas, with
many of its characteristic species reaching southeastern Arizona and the
Edwards Plateau and its escarpments in Texas. The Rio Grande is not
a floristic boundary. Floristically the Big Bend area of Texas and the
mountains along the Rio Grande to the northwest of it have the closest of
relations with the area south of the river. Because of the paucity of col-

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 307

lections made about our southern limits, little can be said regarding the
details of species-distribution in that region. It is clear, however, that a
great many species of our area extend south into the extensive desert tracts
of northern San Luis Potosi. A goodly number reach their southern limits
in the dry valleys of northwestern Hidalgo and some even on the arid
plateau of central Mexico, in western Vera Cruz and adjacent Hidalgo
and Puebla.

In publishing the present enumeration of species I am under no illusion
that it approaches a nearly complete listing of all the species actually
growing in the area. I shall, in fact, be surprised if it includes seventy-five
percent of the total vascular flora. However, it will be vastly more com-
plete and very much more detailed than Watson’s enumeration of Edward
Palmer’s collections, Proc. Am, Acad. 17: 316-361 (1882) and 18:
96-191 (1883), published sixty years ago, the only listing of the flora to
date. The present catalogue is based almost exclusively upon collections
available at the Arnold Arboretum and the Gray Herbarium, particularly
the latter. These include such classical collections as those of Berlandier,
Gregg, Palmer, Pringle, and Purpus, and the less widely distributed recent
collections from the area by L. H. Harvey, Harde LeSueur, E. G. Marsh
Jr., C. H. Muller, Forrest Shreve, Stephen S. White, F. L. Wynd, F. L.
Wynd & C. H. Mueller, and L. R. Stanford, K. L. ‘Retherford & R. D.
Northcraft. Reported upon also are the extensive collections made since
1941 by Mr. Robert Stewart in Coahuila and Chihuahua and those made
in Coahuila and Chihuahua by Dr. C. H. Muller and me in 1940, and
by me alone in 1941. Collections available from the area probably total
well over 15,000 specimens. ‘They do not represent, however, a satis-
factory sampling of the flora of the whole area. There are available few
if any collections from the Laguna District, the valleys of the Rio Florida,
Rio Conchos, and Rio Grande, or from that very promising but unexplored
large tract of desert country west of Ojinaga and east of the El Paso-
Chihuahua City highway. The majority of the species missing from this
catalogue will be discovered in these districts. It is recognized that the
present catalogue can be only an incomplete preliminary one. I believe,
however, that it will be adequate as a basis for a general discussion of the
composition and origins of the flora. These general subjects, as well as
matters regarding the geography and geology of the terrain, floristic areas,
major plant-associations, collectors’ itineraries, etc. will be covered in the
concluding numbers of this series of papers.

My work on the flora of the intermontane plateau of northern Mexico has
been carried on in close cooperation with Dr. Forrest Shreve, of the Desert
Laboratory of the Carnegie Institution. One of the objectives of the
present catalogue is to supply taxonomic and phytogeographical data for
the detailed ecological survey of the intermontane plateau deserts of
northern Mexico and adjacent United States being prepared by Dr. Shreve.
In furtherance of the work I have had three seasons of field work in the
area, in 1938, 1940, and 1941, financed by the Carnegie Institution and
the Arnold Arboretum and by a grant from the William F. Milton Fund.

308 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

POLY PODIACEAE
by C. A. WEATHERBY
Woodsia mexicana Fée, Mém. Fam. Foug. 7: 66. t. 26, f. 3 (1857).
Coanvuita: Sierra del Pino, crevices of limestone cliffs just below crest of high ridge

Rica, Cafion Madera, scarce in shady rock-crevices in canyon, Stewart 2462.
Zacatecas: Concepcion del Oro, in earth in shade of rocks and bushes in elevated
canyon, Aug. 1904, Palmer 257.

Western Texas to Arizona and south to San Luis Potosi and Vera Cruz.
The two Coahuilan collections have the fronds glandular-pubescent be-
neath. They belong, accordingly, to the phase of the species described as
W. pusilla Fourn. Bull. Soc. Bot. France 27: 329 (1880), which outside
our area has been collected in San Luis Potosi and in southern Arizona.
In its glandularity this phase suggests W. Plummerae. It has, however,
the deeply laciniate indusia characteristic of W. mexicana.

Woodsia Plummerae Lemmon, Bot. Gaz. 7: 6

CHIHUAHUA: Wet ledges in the hills northwest Bs i nahie Pringle 455 and 834.

Ranging from trans-Pecos Texas (Davis Mts.) to Arizona and south
into northern Mexico.

Cystopteris fragilis (L.) Bernh. in Schrad. Neu Jour. Bot. 1°: 26. t. 2, f. 9 (1806).

CoanvuiLa: Sierra del Carmen, Canon Sentenela, Wynd & Mueller 521.

The collection cited is young and poorly developed, but it appears to
belong with the Arizona phase of this cosmopolitan species described as
subsp. tenutfolia Clute.

Dryopteris augescens (Link) C. Chr. var. Lindheimeri (C. Chr.) Broun, Ind. N.
2 (1938).
D. normalis C. Fo a Lindheimeri C. Chr. Dansk. Vidensk. Selsk. Skr. Naturv.
Afd. 10: 182 (1913).
Coanvuita: Sierra Hechiceros, Cafion Indio Felipe, along creek in shady canyon,
m

Grande west of Las Delicias, by water, scarce, becoming 1 m. tall, Stewart 2797; El
Tordillo, western margin of Valle de Delicias frequent in shade near water, Stewart
2870

The largest and most luxuriant fern in our area. It is known from
southern Texas (as far east as Houston), Coahuila, Tamaulipas, and
San Luis Potosi.

Phanerophlebia umbonata Underw. Bull. Torr. Bot. Cl. 26: 211 (1899).
Coauutta: Rancho Agua Dulce, wooded canyon on east slope of Sierra San Manuel,

Felipe, shade at base of cliffs in deep canyon, Stewart 523, Johnston & Muller 13062.
Mountains of Nuevo Leon, Tamaulipas, and Coahuila, and Chisos Mts.,
Texas.
Phanerophlebia auriculata Underw. Bull. Torr. Bot. Cl. 26: 212 (1899).
OAHUILA: Sierra Mojada, Cafion Calabasa, fairly common in shade in deep
canyon about 100 m. below crest, Stewart 2198. CHimuAnuA: Canyon in Mapula
Mts., southwest of Mapula Station, Oct. 1886, Pringle 831 (ISOTYPE).

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 309

Ranging a our area northward in western Chihuahua into Arizona
and New Mex
Asplenium Palmeri Maxon, Contr. U. S. Nat. Herb. 13: 39 en

AHUILA: Mountains 6 mi. east of Saltillo, 1880, Palm 435; San Antonio de

los Alamos, under rocks in shaded gulch at base of high ett) ae tuff cliffs, fronds
prostrate, rooting at tip, Johnston 8272. CuutuAHUA: Canyons in hills Rorihwest of
Chihuahua, Oct. 23, 1885, Pringle 444

Ranging from central Texas to Arizona and south to Central America.
Asplenium resiliens Kunze, Linnaea 18: 331 (1844).

CoaHutLa: Rancho Agua Dulce, Sierra San Manuel, Wynd & Mueller 351; Sierra

n
im
pf
Q
oO
joe
ind
io)
oO
a
wn
=]
oO
st)
tal
wn
c
B
=;
>
c
eis
—
=)
aaa
a]
8
~~
3
is:)
s
N
~~
a
ae
°
uo}
oO
» Oo
pas
3
°
co
i=]
o
bo
3
ae
a
'

t 2474.
Ranging from Pennsylvania, Kansas, and Arizona south along the
Andean chain to Argentina (including A. Lealii Alston).

Asplenium exiguum Bedd. Ferns So. India t. 146 (1863).

CoaHvuILaA: Sierra Madera, Canon — Agua, sparse in rock-crevices in moist anaes
wooded canyon, Muller 3251. CHIHUAHUA: Deep shaded damp glen at hea
canyon next south of the large central canyon in the Mapula Mts., southwest of ae
station, Nov. 4, 1886, Pringle 833.

A rare fern known only from scattered stations in Arizona, Sonora,
Chihuahua, Coahuila, and Federal District, and from southern India and
northern China.

Gymnopteris hispida (Mett.) Underw. Our Native apy 6. 84 (1900).

Gymnogramma hispida Mett. ex Kuhn, Linnaea 36: 72

Bommeria hispida Underw. Bull. Torr. Bot. Cl. 29: 633 ee

Coanuita: Sierra Cruces, Canon Tinaja Blanca, shelter of ledge in open canyon,
Johnston & Muller 281. CuraHuanua: Sierra Rica, Cafion adera, common in
shaded places in canyon, Stewart 2466; Piramide, base of rock-masses, Johnston 8146;
8 mi. northwest of Cruces, base of tuff cliff, Johnston 7977; Sierra Encinillas, 7 km
north of Fierro, among rocks on hillside, Gieware 795; near Chihuahua, Privcie 465.

This species ranges from western Texas to Arizona and south into

urango. In our area, at least, the species appears to be confined to areas
of igneous rocks.

Pellaea cardiomorpha Weatherby, as nov.

Pteris cordata Cav. Descr. 267 (1802).

Pellaea cordata (Cav.) J. Sm. Cat. Kew Ferns 4 (1856), non Fée (1850-52).

CoauvuiLa: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 590. Cut-
HUAHUA: Sierra Rica, Canon Madera, scarce in shady rock-crevices in canyon,

310 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Stewart 2525; hills about 8 mi. northeast of Chihuahua, Oct. 17, 1885, Pringle 448.
ZACATECAS: Pico de Teira, southwest of Cedros, Sept. 1908, Lloyd 247 (US).

The new name for this species, P. cardiomorpha, is put forward with
some diffidence. Pellaea cordata (Cav.) J. Sm. is definitely illegitimate,
but, following Hooker, most authors have regarded P. sagittata (Cav.)
Link as synonymous with it. If this be correct, the name P. sagittata,
which is quite clear under Pellaea, should be taken up. But the material
at hand indicates that the two species proposed by Cavanilles really differ
in several characters, most of which, as Hooker long ago pointed out in
regard to some of them, are not as constant as one could wish, but which,
taken together, seem to give adequate grounds for separation. I have
accordingly preferred to coin a new name in place of P. cordata, which has
no indubitable synonym. The two species may be distinguished as follows:

Rhizome- scales bright brown in mass (contrasting with those of bud and stipe), atte

or obtusely triangular in outline, especially so as seen from the commisural face,
averaging about 50 pw in Sincwier bicpergaed eigts a ass Geercisatels sfea'are Pellaea cardiomorpha.
Rhizome-scales pale brown, not contrasting with those of bud and stipe, mostly rather

globose or the larger broadly ellipsoid, varying considerably e size, 65-80 u,
commonly AbOUE 75 Us cers ga cance 54a-9t ous ae naw ies negae eds Pellaea sagittata.

In addition to the above enumerated characters, the pinnae of P. cardio-
morpha tend to stand at a broad angle to the rachis, those of the fertile
fronds of P. sagittata to be more or less strongly ascending. In both, the
lower leaf-surface may be either glabrous or pubescent with white hairs.
In mature and well developed material the spore-characters hold con-
sistently.

As here defined, P. cardiomorpha becomes a species mainly of northern
and central Mexico, extending north to the Davis Mts., Texas, and south
in eastern Mexico to Hidalgo, Michoacan, and Gaznia,. Pellaca sagittata
ranges from San Luis Potosi and Zacatecas south to Guatemala and south-
ward along the Andes to Bolivia.

Pellaea intermedia Mett. ex Kuhn, Linnaea 36: 84 (1869).
CoaHuILA: Sierra del Pino, Cafon Ybarra, arroyo banks, Stewart 1821; Sierra

Alamos, about shaded tuff cliffs, Taknston & Muller 859 and 900; ado 1880, Palmer
1426; Carneros Pass area, 1880, Palmer 1427; 4 km. east of Fraile, Stanford et al.
354; Sierra Negras, 9 km. south of Parras, Stankoed et al. 161. CHIHUAHUA: Sierra
Santa Eulalia, limestone ledges, Pringle 461. ZacaTEcAs: Mountain 18 km. west of
Concepcion del Oro, Stanford et al. 576; Cedros, ravines, Lloyd & Kirkwood 135.

Ranging from trans-Pecos Texas to Arizona and south into our area and
eastward into the mountains of Nuevo Leon. The type of P. intermedia
is the rare glabrous phase of the species. Our plants belong to the common
forma pubescens (Mett. ex Kuhn) Broun.

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 311

Pellaea ovata (Desv.) thera Contr. Gray Herb. nee 34 (1936).

Pellaea flexuosa (Kaulf. ex C. & S.) Link, Fil. Sp. 60 (18

CoaHutiLa: Sierra San Manuel, Rancho Agua Dulce, isis: & Mueller 318; western
base of Sierra Guajes 8 km. east of Buena Vista, igneous hillside, not common, Stewart
1467.

From central and southern Texas south through eastern Coahuila and
the eastern Sierra Madre to Central and South America.
Pellaea atropurpurea (L.) Link, Fil. Sp. 59 (1841).

ILA: Sierra San Manuel, Rancho Agua Dulce, Wynd & Mueller 358; western
base of Sierra del Carmen 8 cae east of Hac. Encantada, shade in canyon, Stewart

cliff- ba vices: Stewart 1111; Sierra Mojada, Canon Hidalgo, in shaded canyon below
crest, Stewart 1065. CHraHuUAHUA: Sierra Rica, Cafion Madera, rock-crevices in
shaded canyon, Stewart 2475.

Ranging from Vermont and western South Dakota south to Guatemala.

Pellaea Wrightiana Hook. Sp. Fil. 2: 142 (1858).

1rA: Sierra Hechiceros, Rancho E] Tule, about rocks on sunny hillside,
Johnston & Muller 1309. CuHrHuAHUA: Sierra Virus one plant on steep rocky east
slope 3 mi. east of Rancho Virulento, Johnston 8090

Ranging from central Texas to Baja Cae Apparently a plant
of igneous rocks.

Pellaea ternifolia (Cav.) Link, Fil. Sp. 59 (1841).

CHIHUAHUA: Grassy summits of the Sierra Santa Eulalia, southeast of Santa
Eulalia, Nov. 5, ae Pringle 446; cool rocky slopes of mountains near Chihuahua,
Oct. 1886, Pringle 9

From es Arizona south along the western Sierra Madre,
extending to Peru, Argentina, and northern Chile. Apparently a plant of
igneous rocks.

Pellaea microphylla Mett. ex Kuhn, Linnaea 36: 86 (1869).

CoaHuizta: Sierra San Manuel, Rancho Agua Dulce, Wynd & Mueller 320; Hill-
coat Mesa, west of Encantada Ranch, Marsh 1454; high mesa 14 km. northwest of
Buena Vista, common on open hillside, Stewart 1435; Sierra Madera, Cafion Charretera,
rocky bed of arroyo, abundant, Johnston 8930; Sierra Gavia, under rocks on canyon
floor, Johnston 7206; Puerto San Lazaro, scattered on open talus slope, Muller 3079;

Sierra del Pino, summit of great western escarpment about 10 mi. north of La Noria,
under rocks, Johnston & Muller 552; Sierra del Pino, crest of eastern ridge about 4 mi.

Mojada, Cafion Hidalgo, shade in canyon below crest, Stewart 1066; east ede of Valle
Acatita, crevices in limestone 2 km. northeast of Parritas, Stewart 2767; margin of

312 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Valle Delicias, 1 km. northwest of mouth of Canon Blanco, frequent in arroyos,

1136; high northwestern end of Sierra Diablo, rocky hillside, Stewart 986; Sierra
Santa Eulalia, limestone hillside, Pringle 440 and 458; pass 19 mi. east of Timeties,
limestone hillside, oe 7854. Zacatecas: valley 15 km. west of Concepcion
del Oro, Stanford et a

Centering in our area and extending eastward into the Sierra Madre of
Nuevo Leon, and northward to the south escarpment of the Edwards
Plateau and trans-Pecos Texas. Confined to limestone.

Notholaena delicatula Maxon & Weatherby, Contr. Gray Herb. 127: 7 (1939).

Coanuita: Sierra Madera, Charretera Cafion, steep north slope in conifer forest,
on are about 8500 ft. alt., Johnston 9046; Lerios, July 1880, Palmer 1387 (TYPE) ;
Carneros Pass area, March 1880, Palmer 1385.

Endemic to the mountains of southeastern Coahuila and northern Nuevo
Leon. The record from Jalisco, Maxon & Weatherby, 1. c., was caused by
an error on one of Pringle’s labels. Though indicated to be from Jalisco,
the collection actually came from near Monterey, Nuevo Leon.

Notholaena limitanea Maxon, var. mexicana (Maxon) Broun, Ind. No. Am. Ferns
119 (1938).
Notholaena limitanea subsp. mexicana Maxon, Amer, Fern Jour. 9: 72 (1919).
Coanuita: Sierra Madera, Canon Charretera, on rocks on steep north slope in

imu
Stanford et al. 107. CHIHUAHUA: Siert ra Santa Eulalia, limestone ledges, Sept. 13,
1885, Pringle 451 (1isoTYPE). ZACATECAS: mountain 18 km. west of Concepcion del
Oro, Stanford et al. 574; Cedros, canyons, Kirkwood 140.

Ranging from our area into Durango and possibly Tamaulipas. Among
the cited collections, Johnston 7269 is exceptionally stout and strict for this
variety. It is, however, approached by some individuals of the type collec-
tion, and in all technical details, especially the strongly reticulate-rugose
spores, it entirely agrees with them
Notholaena aurea (Poir.) Desv. Mém. Soc. Linn. Paris @: 219 (1827).

VERNACULAR NAMES: Canaguala; Canawala.

CoanuiLa: Sierra del Carmen, Canon Sentenela, Wynd & Mueller 511; Mesa

Hechiceros, “Cafion Indio ipe, common at base of ae and on talus, Stewart 39
and 135; Sierra Hechiceros, Rancho El Tule, about rocks on arid hillside, Johnston &

stone cliff, Johnston 7980; 11 mi. northeast of Camargo, Agua cliff, igen 7896;
rocky hills near Chihuahua, Oct. 1885, Pringle 462
Ranging from Texas and Arizona soil to Argentina. Palmer reports

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 313

that this fern is sold in the market at Saltillo, a decoction of the plant being
taken internally for ‘‘pain in the stomach and for coughs.”
Notholaena sinuata (Lag.) Kaulf. Enum. Fil. 135 (1824).
VERNACULAR NAMES: Nacahuela; Lengua de Cervo.
UILA: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 510; Sierra

3 km. southwest of Fraile, in arroyo, Stanford et 336; Sierra del Pino, Canon
Ybarra, arroyo banks, Stewart 1821a; Sierra Madera, Cafion Charretera, moist
ledges, Johnston 9097; Cafion de Jara, east of Socorro, Schroeder 10; ra

Cruces, near Santa Elena, shady arroyo, Stewart 292; Picacho de San José,
crevices of cliffs, Stewart 1110; Sierras Negras, 9 km. so Otte of Parras, Stanford et al.
203; Picacho de Jimulco, summit, Stanford et al. 105. CutHuaHvua: 11 miles northeast
of Beer ee cliff, Johnston 7923. Zacatecas: Cedros, canyons, Lloyd & Kirk-
wood 138

South- sa Texas to Arizona and southward in the Andean region to
northern Argentina.
Notholaena sinuata var. integerrima Hook. Sp. Fil. 5: 108 (1864).

Coanumwa: Rancho Agua Dulce, Sierra San Manuel, Wynd & Mueller 321; Sierra
Azul, Buena Vista Ranch, July 8, 1938, Marsh 1250; Soledad, 1880, Bane 1402;
Saltillo, 1880, Palmer 1401; Buena Vista, 1848, Gregg 297 in pt.; San Antonio de los

side Valle Acatita, Stewart 2731 and 2751; Puerto pirigsiens south of Las Dee.
on slopes, Stewart 2963; mouth of Cafion Blanco, north end of Valle Delicias, Stewart
2904; Lorenzo de Laguna, rey Palmer 1409, Curauanua: Chihuahua, Pringle
464 in pt. Zacatecas: Mounta 8 km. west of Concepcion del Oro, Stanford et al.
575; cone canyons, Lloyd ood 136 in pt. and 137.

Southern Oklahoma (Arbuckle Mts.) and central Texas to southeastern
Arizona and southward, mostly along the eastern Sierra Madre, to
Vera Cruz.

Hooker’s name, NV. sinuata var. integerrima, has very generally been
applied to the plant here classified as V. sinuata var. cochisensis. Tha
application can no longer be maintained. When he proposed var.

Liebmann, Gregg [297] and Seemann [1928]. Of these, only the Lieb-
mann specimen is labelled as belonging to the variety. It is a single small
frond with oblong, quite entire pinnae, the scales of the lower surface like
those of typical NV. sinuata, but those of the upper relatively broad-bladed
and persistent as in var. cochisensis. The Seemann material is similar,
though much more ample, the sheet containing four complete individuals.

herbarium, consists of a detached frond similar to the Liebmann specimen
and an entire plant of var. cochisensis. Had Hooker cited this last Gregg
specimen particularly or accounted for it in his description, it might have
been designated as type and the usual application of his name maintained.

314 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

But one can hardly reconcile his statement “pinnae entire or nearly so’
with var. cochisensis, in which the pinnae, though tiny, show at least one
conspicuous lobe. In view of this and of Hooker’s having labelled only the
Liebmann specimen, it must be taken as type and the epithet integerrima
Pa, accordingly.

eld observation and a restudy of material in the Gray Herbarium indi-
ms that rather numerous specimens with shallowly lobed pinnae, which
have hitherto either been associated with var. cochisensis or regarded as
dwarfed individuals of typical N. sinuata, actually belong with var.
integerrima, as represented by the Liebmann and Seemann collections.
They are lie these collections in their combination of scale-characters, in
their small size, and, except for the lobing of the pinnae, in habit. They
are accordingly here placed in var. integerrima. So understood, that group
becomes a reasonably consistent, if not altogether happily named, variety,
intermediate in characters between typical V. sinuata and var. cochisensis,
grading into both, and with the mainly Texan and northeast-Mexican range
above indicated.

In central Mexico, typical N. sinuata also produces a phase with entire
pinnae (JN. laevis of authors, not Mart. & Gal.; N. crassifolia Moore &
Houlst.; NV. pruinosa Fée).

Notholaena sinuata var. cochisensis (Goodding) beeen comb. nov.

Notholaena cochisensis Goodding, Muhlenbergia 8:

Notholaena sinuata var. crenata Lemmon, Ferns ae Slope 7 (1882), nomen

udum

Vessueuade NAMES: Canelilla; Doradillo.
Between Rancho Santo Domingo and Hac. Piedra Blanca, Wynd ¢

1410; Sierras Negras, 9 km. south of Parras, Stanford et al. 194; Sierra Jimulco, 11 km.
northeast of Timulco, Stanford et al. 42, CuHrauaHua: Sierra Rica, Dec. 1882,
Newberry; sovanamrnd adie Palmer 357: pass 19 mi. east of age limestone,

mi. north of Gee. haere sandstone. rocks, Johnston 7792. ZACATECAS: Valley 15
km. west sd a del Oro, Stanford et al. 548; Cedros, canyons, Lloyd & Kirk-
wood 138 i
This is re plant which has long been called NV. stnuata var. integerrima.
The three recognizable variants of NV. sinmuata may be distinguished as
follows:
Pinnae 1 cm. or more long, ovate, commonly subacute and cut 4-—™% to the midrib into
4—6 pairs of oblong lobes; scales of the upper surface of the lamina with narrow
central portion or reduced to stellate processes, usually soon deciduous, those of the

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 315

ower coe lanceolate, up to 1.5 mm. long; rhizome-scales pectinate-ciliate or
SSELMU ALC arcesiiratenat see eeyaae tele, iat etaeres ase s lel nc apenas see a atialay as eneteyerapeRs N. sinuata (typ:cal).
Pinnae beers less than 1 cm. long, very obtuse, with 1-3 pairs of broadly ovate lobes
or entire; scales of the upper cpa with caer broad central portion, usually
persistent till full maturity of the frond.
Pinnae oblong, entire or with about : pairs of shallow lobes; scales tn lower surface
and rhizome as in the typical variety................0eeceees ar. integerrima.
Pinnae subquadrate, nearly or quite as broad as long, with 1 or 2 eaciy 3) pairs
of lobes; scales of the lower surface ovate, 0.5 mm. long; rhizome-scales entire
OD MEAT VSO erstrcui einen Se. 584s -2y rd 60-8 alae rereeeee peege var. cochisensis.
ood many field observers are of the opinion that var. cochisensts
should be treated as a distinct species, and their contention has been
strengthened by the recent discovery that the variety is poisonous to
stock, the typica] form not. Nevertheless, judged by the usual taxonomic
evidence, var. cochisensis is so connected min the typical variety, through
var. integerrima, that the traditional treatment of it as a variety only is not
unreasonably conservative.
Notholaena Aschenborniana Klotzsch, Linnaea 20: 417 (1847).
Coanvumta: Sierra San Manuel, Wynd & Mueller 337 (US); Sierra Gavia, 5 mi.
)

limestone, Stew 2768 on Blanco, Sierra Margaritas, shade of cliffs, Stewart
2915; Sierra Negras, 9 km. south of Parras, Stanford et al. . CHIHUAHUA:
Santa Eulalia, limestone feces Pringle 466 and 469

Western Texas to Arizona and south to coded Mexico.

ones Schaffneri (Fourn.) Underw. var. Nealleyi (Seaton) Weatherby, comb.

Rey tae pete Beaton ex Coulter, Contr. U.S. Nat. Herb. 1: 61 (1890).

CoaHuiLta: Moun ‘4 mi. west of Cuatro Cienegas, shaded rock-crevices in
small canyon, ae ees Sierra i sheltered on north-facing limestone
ledges at Santa Elena, Johnston & Muller 20

The type of V. Nealleyi came from Limpia Canyon, Jeff Davis Co., Texas
(Nealley 560). These, at least, are the data accompanying the specimen
designated as type in the U. S. Nat. Herb. The poms number has
been changed from 894 to 560. As published, the type was said to come
from the Chinati Mts. and to be numbered 894. Other sectors of this
fern have been made at Goodenough Springs, Val Verde Co., Texas
(Nealley 123), and from Barranco de Santa Maria, Zacuapan, Vera Cruz
(Purpus 6199). They may be distinguished from typical NV. Schaffnert
as follows:
Rhizome-scales narrowly linear-attenuate, joa and ee pectinate-ciliate ;

median pinnae with 4-6 pairs of free pinnules............ Schaffneri (typical).
Rhizome dene linear- subulate, sparsely iv inconspicuously ee -ciliate ; median
innae usually with 1-3 pairs of free pinnules......... N. Schaffneri var. Nealle yi.

The Noaies and Purpus specimens, mentioned above, are small and

316 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

have passed as immature individuals. The two collections from Coahuila,
owever, have fronds up to 22 cm. long and freely soriferous, and are
obviously full grown. The characters of the variety are retained even in
this mature state. Davenport, Bot. Gaz. 16: 54 (1891), observed that
two plants were involved in this species-aggregate, but unfortunately he
supposed spar igiede rather small type material to be the same as Nealley’s
specimens and gave a new name (N. Schaffnerit var. mexicana) to the
typical variety of N. Schafine
Notholaena Grayi Davenp. Bull. Torr. Bot. Cl. 7: 50 (1880).

CoaHutiLa: Soledad, Sept. 1880, Palmer 1388; San Antonio de los Alamos, under
rocks, dry basalt in upper canyon, Johnston & Muller 926. Cuinuanua: Sierra
pene rocky hillside 6 km. north of Fierro, common in crevices, igneous rocks,
Stewart 790; cliffs of volcanic tuff, 8 mi. northwest of Cruces, Johnston 7984; 11 mi.
ail of Camargo, lava cliff, Johnston 7902; 6 mi. west of Piloncillo, lava hillside,
Johnston 7880; rocky hills near Chihuahua, Pringle 463 in pt

Ranging from Texas and Arizona south through Chihuahua and Sonora
to Jalisco.

Notholaena aariie ewes Contr. U.S. Nat. Herb. 17: 605 (1916).

Coa , Sept. 1880, seul 1389 (ISOTYPE). CHIHUAHUA: Rocky
hills near shai rt “il 1885, Palmer 4

Known from Tamaulipas, Ee Chikashia, and Arizona.

Notholaena oo (Mart. & Gal.) Hook. Sp. Fil. 5: 110 (1864).

OAHUI a Juarez, Sept. 1880, Palmer; Muzquiz, Marsh 351 (US); moun-
tains northeast deri Sept. 1880, Palmer 1380; Sierra Gavia, 5 mi. north of
Saucillo, under rocks on terrace in canyon, Johnston 7205; gorge just east of Socorro,
on cliffs, Johnston 8849; western extremity of Sierra Madera, rig narrow canyon
2 km. northeast of Puertecito, ledges on canyon-wall, Johnston 931

Ranging from Texas and New Mexico south to alan
Notholaena neglecta Maxon, Contr. U. S. Nat. Herb. 17: 602 (1916).

Coauvumta: San Lorenzo Canyon, 6 mi. southeast of Saltillo, a few plants in narrow

arrow canyon m. southeast of Puertecito, ges on canyon-wall, one colony,
Johnston 9319a; Sierra pone canyon 5 uthwest of Santa Elena, crevices of
shales on shad canyon-wall, local, Johnston & Muller 822; La ica, Sierra

520 (US). CHrmuanHua: Sierra Santa Eulalia, limestone cliffs, Sept. 9, 1885,
Pringle 452.

Ranging from southeastern Arizona south into Chihuahua and into
eastern Coahuila. The collection from the Sierra de la Fragua is an
unusually narrow-bladed form with relatively small basal pinnae.
Notholaena Standleyi Maxon, Amer. Fern Jour. 5: 1 (1915).

Coauvuita: Near Santo Domingo, igneous hill, Wynd & Mueller 467; Saltillo,
summit of mountain, shaded crevices of sandstone, May 1898, Palmer 184; Saltillo,
1905, Palmer 754; Carricito, north-facing ledge of lava, Johnston é Muller 163;
eastern foothills of Sierra Cruces, vicinity of Santa Elena, shaded rock-crevices,
Stewart 347; Sierra Cruces, Canon Tinaja Blanca, hillsides and along arroyo, Stewart
330 and 624, Johnston & Muller 296; base of tuff cliffs 3 mi. northwest of San Antonio

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 317

de los Alamos, Johnston & Muller 857; La Botica, base of Sierra Margaritas, limestone
cliffs, Stewart 2893; canyon-mouth, Cafion Blanco, Sierra Margaritas, crevices on slope,
Stewart 2907; canyon 6 mi. west of Viesca, Johnston 7744; Picacho de Jimulco, about
summit, Stanford et al. 89; San Lorenzo de la Laguna, 1880, Palmer 1379. Cut-
HUAHUA: Sierra Rica, Dec. 1882, Newberry; Sierra San Carlos, road to mines, base
of limestone cliff at canyon-mouth, Johnston & Muller 37; Serra Encinillas, 7 mi.

—
Sw
—_
—
=
=]
°
a
ec
a
Oo
pf
n
cc
°
—-
@,
a’)
3
p
ad
gg
7)
pa
<
fp
2h
in
~
i=)
a
=
a
~
Ss
=
NI
Co
S
a
Q
p
i |
°
=]
ce
=)
v2)
om
=)
a
=
aa

northwest of Santa Fe, crevices of limestone, common, Stewart 2618.
Ranging from Texas and western Oklahoma to Nevada and Arizona and
south through western Mexico to central Mexico.
ae Greggii (Mett. ex Kuhn) Maxon, Contr. U. S. Nat. Herb. 17: 606
(1916). :

Pellaea Greggii Mett. ex Kuhn, Linnaea 36: 86

Allosorus Greggii (Mett. ex Kuhn) Kuntze, Rev. ie 2: 806 (1891).

Notholaena Pringlei Davenp. Bull. Torr. Bot. Cl. 13: 132. t. 58 (1886).

Cheilanthes Davenportii Domin, Bibl. Bot. 20[Heft 85]: 133 (1915).

Coauutmta: Monclova, Aug. , Palmer 1383; 10 mi. north of Cuatro Cienegas,
Wynd 747; Lomas del Aparejo, ae side of Tine de Guaje, dry limestone ledges
on sunny hillGdes Johnston & Muller 774; Tanque La Luz, south end of Cafada
Oscuro, limestone ledges between gypsum on escarpment, Johnston 8503; Sierra Cruces,
limestone ledges near Santa Elena, Johnston & Muller 206; south base of Picacho de
San José, in arroyo, confined to gypsum, Johnston & Muller 817; San José, under
basaltic rocks on rocky hillside, Johnston & Muller 994; Sierra Planchada, 6 mi. north-
east of Esmeralda, limestone ledges on hillside, Johnston & Muller 833; Sierra Mojada,
Jones 519 (US); Cerro Zapatero, July 1910, Purpus 4633; El Coyote, eastern margin
of Valle Acatita, crevices on slope, Stewart 2743; Rancho Las Uvas, east side Valle

a

CHIHUAHUA: Sierra Rica, Dec. 1882, Newberry; Sierra Santa Eulalia, dry calcareous
ledges and bluffs, April 23, 1885, Pringle 441 (type of N. Pringlei); Sierra a
Eulalia, calcareous bluffs, Nov. 15, 1888, Pringle 857; Cation del Coyote, 20 km. north-
west of Santa Fe, crevices of limestone, common, Stewart 2614. DuRANGo: Rocks
hill Bah als of Mapimi, April 17, 1847, Gregg 467 (1soTYPE of N. Greggii); Lerdo,
Cerro el Raymundo, Chaffey 58 in pt. (US) ; 7 mi. southwest of Chocolate, shaded
slope, Shreve 9113.

This species is practically confined to our area. Outside, it has been
found on the northern side of the Rio Grande at the mouth of Boquillas
Canyon in Texas. While evidently not confined to it, the species is fre-
quently found on or near gypsum. The type of V otholaena Greggit, judg-
ing from the date on the type-specimen, was collected near the Durango-
Chihuahua state-line northwest of Mapimi, between Jaralito and Arroyo de
Cerro Gordo.
Notholaena bey oboe Maxon, Proc. Biol. Soc. Wash. 18: 205 (1905).

CoaHutILa: Western base of Picacho del Fuste, ees ee mountain-side, com-
mon on all gypsum beds and confined to them, Johnston 8354; south end of Canada

clu 8.
Mulle er 243; Rancho del Covote, eastern margin of ‘Valle pons crevices on gypsum,
common, Stewart 2732.

318 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

Outside of our area this fern is known only from the type-collection,
made by Pringle in the mountains of southern Nuevo Leon, 15-20 miles
south of Doctor Arroyo. The plant appears to be a marked gypsophile.
In the Sierra Cruces, near El Fuste, and in Cafiada Oscuro, the plant was
abundant and luxuriating on gypsum and confined to that substratum.
It forms dense clumps which may become as much as a meter in diameter.
Cheilanthes alabamensis (Buckl.) Kunze, Linnaea 20: 4 (1847).

Coanutta: Rancho Agua Dulce, wooded canyon on east slope of Sierra San
Manuel, Wynd & Mueller 377; ravine near Puerto Santa Anna, Hac. Mariposa, Wynd
& Mueller 219; Sierra Guajes, Cafion Milagro, faces of cliffs, Stewart 1714; Caracol
Mts., 1880, Palmer 1419; Soledad, Sept. 1880, Palmer 1420; Saltillo, shaded clay bank
in ia arroyo, 1898, Palmer 366 in pt.; Saltillo, Sierra del Pueblo, moist rock crevices
near ground, 1904, Palmer 433 in pt.; San Lorenzo Canyon, 6 mi. southeast of Saltillo,

shaded places about tuff cliffs, Johnston & Muller 894. Cumauanva: Sierra Almagre,
about rocks in deep shaded canyon, Johnston & Muller 1193; Sierra Santa Eulalia,
shaded places, Nov. 2, 1885, Pringle 449.
This fern ranges from Virginia to Florida, west to Missouri and Arizona,
and south into Nuevo Leon, Coahuila, and Chihuahua.
—. i raaneacaciaaki (Desv.) Maxon ex Weatherby, Contr. Gray Herb. 114:
4 (19

aye NAME: Sanguinaria.
Saltillo, shaded clay bank in deep arroyo, 1898, Palmer 366 in pt

chased in market un name ‘“‘Sanguinaria,”’ Sept. 1898, Pal 368; mountains 6 mi
east of Saltillo, 1880, Palmer 1418; base of mountains southeast of Saltillo, road to
Diamante Pass,, Johnston 7268; San Lorenzo Canyon, southeast of Saltillo

Hidalgo, shade in canyon below crest, common, Stewart 1067; Sierra Negras, 9 km.
south of Parras, Stanford et al. 209. Cuimuanua: Sierra Almagre, under rocks on
shaded canyon-floor, Johnston & Muller 1152; Sierra Santa Eulalia, shaded places,
Nov. 2, 1885, Pringle 449. Zacatecas: Mountain 18 km. west of Concepcion del
Oro, Stanford et al. 573; Cedros, Lloyd 125

Entering our area from the south and southeast, extending north from
Guatemala through eastern Mexico and reaching its northern limit in
Coahuila and Chihuahua, where it grows with the related C. alabamensis
and, at times, is separated with difficulty from that more northerly ranging
species. Palmer reports that small bunches of this fern are sold in the
market at Saltillo under the name “Sanguinaria,” a decoction of the plant
being drunk ‘‘to purify the blood.”
Cheilanthes aemula Maxon, Contr. U.S. Nat. Herb. 10: 495 (1908).

Coanuta: Sierra Guajes, Cafion Milagro, on cliffs, not common, Stewart 1710;

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, I 319

Mt. Caracol, 1880, yess neces Sierra Hechiceros, Canon Indio Felipe, base of talus-
slope, not common, Stewart 155.

Ranging con ae (escarpment of Edwards Plateau) south through
eastern Coahuila and the Sierra Madre of Nuevo Leon and Tamaulipas to
eastern San Luis Potosi.

Cheilanthes horridula Maxon, Amer. Fern. Jour. 8: 94 (1918).

Cheilanthes aspera Hook. Sp. Fil. 2: 111. t. 108a (1852), non Kaulf. (1831).

CoaHuILa: Mountains 24 mi. northeast of Monclova, 1880, Palmer 1422; hillside
2 mi. west of Sacramento, road to Cuatro Cienegas, Johnston 7092; Sierra Gavia, 5 mi.
north of Saucillo, under rocks on slope, Johnston 7207; Saltillo, Sierra del Pueblo,
crevices, 1904, Palm mer 433 in pt.; General Cepeda, high bluff, Palmer 326 in pt.
Picachos Colorados, under rocks below cliffs, Johnston & Muller 112; near Sante
Elena, Sierra Cruces, limestone ledges, Johnston & Muller 204; south hace of Picacho

de San , dry arroyo bank near gypsum exposures, Johnston & Mudler 816; open
limestone can west of Viesca, Johnston 7743; Las Uvas, east side Valle
Acatita, sb acid heed 2697. CHIHUAHUA: Bachimba Canyon, rocky hills, Oct. 31,
1885, Pringle 447 mi. northeast of Camargo, lava cliff, Johnston 7903a. DURANGO:

23 mi. north of ene under sandstone rocks on slope, Johnston 7793; Raymundo
Hill, Lerdo, alt. 1650 m., Nov. 25, 1911, Chaffey 58 in pt.

A local and rather + are ica ranging from central Texas to south-
western New Mexico and Durango.

Cheilanthes moncloviensis Baker, Ann, Bot. 5: 210 (1891).
CoaHvuILaA: Soledad, Sept. 1880, Palmer 1378 (1soTYPE).
Known also from Puebla.

Cheilanthes Wrightii Hook. Sp. Fil. 2: 87. t. 71/0a (1852).

OAHUILA: Near Santo Domingo, igneous hill, Wynd & Mueller 472; Sierra Cruces,
Canon Tinaja Blanca, common among grass on sunny open gravelly terrace in upper
canyon, Johnston & Muller 288; Sierra Cruces, crest north of Puerto Bajito at head
of Canon Tinaja Blanca, common on grassy sunny rocky slope, Stewart 1949.
CHIHUAHUA: Sierra Virulento, about rocks on crest of ridge, Johnston 8073; 8 mi.
northwest of Cruces, sandstone cliff, Johnston 7985; Sierra Encinillas, 6 mi. north of
Fierro, among rocks on hillside, errs common, Stewart 794; rocky hills northwest of
Chihuahua, Oct. 7, 1885, Pringle 445; 11 mi. northeast of Camargo, about lava cliff,
Johnston 7903.

Arizona to Texas and south to Durango. In our area the species is con-
fined to areas with igneous rocks and frequently grows with short grass on
sunny gravelly terraces and slopes.

Cheilanthes meifolia D. C. Eaton, Proc. Am. Acad. 18: 185 (1883).

CoaHumLa: Rancho Agua Dulce, wooded canyon on eastern slope of Sierra San
Manuel, Wynd & Mueller 350.

Known also from Nuevo Leon, Tamaulipas, and San Luis Potosi.
Cheilanthes Feei Moore, Ind. Fil. p. xxxviii (1857).

CoauuiLa: Canon Chojo Grande, 27 mi. southeast of Saltillo, growing out of
small hole in an exposed rock at base of canyon, 1904, Palmer 374; Saltillo, exposed
rocks, 1904, Palmer 432; Puerto Colorado, faces of sandstone eae ee 8693;
San José, crevices of north-facing basalt crags on slope, fronds flat against rock,

on dry cliff of volcanic rock, fronds geeky appressed seanict rock, Johnston & Muller
1203:

Widely distributed in western United States and extending south into
adjacent Mexico.

320 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Cheilanthes tomentosa Link, Hort Berol. 2: 42 (1833).

Coauuita: Sierra Hechiceros, Cafon Indio Felipe, about rocks in deep wooded
eanyon, Stewart 130 and 137, Johnston & Muller 1363; Rancho Agua Dulce, Sierra San

, dry arroyos, Wynd & Mueller 337. CurauAnvua: Chihuahua, northwestern

hills, Oct. 23, 1885, Pringle.

Ranging from Virginia and Georgia west to Arkansas and Arizona, and
south into northern Mexico.
Cheilanthes castanea Maxon, Proc. Biol. Soc. Wash. 32: 111 (1919).

CoanutLa: Sierra Guajes, Cafon Milagro, on os not common, Stewart 1711;
Soledad, 1880, Palmer 1390 in pt.; Caracol Mt., 1880, _Palmer 1391 in pt.; Saltillo,
r.

Carneros Pass, ledges, Piste 2777; Carneros Pass area, 1880, Palmer 1390 (ISOTYPE) ;
Sierra del Pino, La Noria, shady bushy arroyo-bank, Johnston & Muller 477; Sierra

Jimulco, Stanford et al. 43. ZACATECAS: Concepcion del Oro, high up in canyon in
shady moist places, 1904, Palmer 260 in pt.; Concepcion ~ ia 1902, Palmer 388;
mountain 18 km. west of Concepcion del Oro, Stanford et a

Ranging from Texas to Arizona and south to ene and Hidalgo.
Cheilanthes Eatoni Baker, Syn. Fil. 140 (1867).

Coanutta: Soledad, 1880, Palmer 1394, 1395, 1396; east of La Rosa, Wynd &
Mueller 43; hills near Teen shaded clay bank in deep arroyo, 1898, Palmer 367 ;

e of mountains southeast of ans road to Diamante Pass, limestone ledge,

John nite n & Muller 1308; Sierra Pechicerns. Cafion Indio Felipe, base of cliffs, ee
Muller 162; Sierra del Pino, Cafion Ybarra, dry hillside, Stewart 1875; Sierra del Pino,

under rocks on ridge- crest 4 mi. northeast of La Noria, Johnston & Muller 653; Sierra

Picacho de San José, crevices of cliffs, Stewart 1112;*San José, about basalt crags on
paaritige Johnston & Muller 980; San Antonio de los Alamos, under pe lava rocks in

pen upper re Johnston & Muller 903; Sierra Negras, 9 km. south of Parras,
pr arie et al, in pt.; ag ra de Parras, July 1910, a ae CHIHUAHUA:
Sierra Santa Sie Oct. 27, 1885, Pringle; rocky hills near Chihuahua, Oct. 1885,

Pringle ee in raed Meoqui, LeSueur 1147. Zacatecas: Cedros, canyons, Lloyd &
Kirkwood 1

ice gore Oklahoma and Texas west to Arizona and south to
Durango and San Luis Potosi.
Cheilanthes jamaicensis Maxon, Contr. U. S. Nat. Herb. 24: 51 (1922).

Coanutta: Sierra de la Gloria, March 4, 1939, Marsh 1964; Sierra Madera, Canon
del Agua, abundant in rock-crevices in moist densely wooded canyon, Muller 3252;
mountains 6 mi. east of Saltillo, 1880, Palmer 1418 in pt

Here first reported from Mexico; previously known only from Jamaica
and from Santo Domingo (var. domingensis C. Chr.). The geographical
distribution of the species is unusual; one would not, off hand, expect a local
West Indian species to appear, apparently just as locally, in the mountains

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 321

of northeastern Mexico. However, this is not unprecedented; C. notho-
laenoides, though much more common than C. jamaicensis in the Mexican
highlands, similarly occurs in Jamaica and Hispaniola. Since its nearest
relatives are Cordilleran, C. jamaicensis, in spite of its rarity on the main-
land, may reasonably be classed with those continental species which have
outlying stations in the West Indies.

Cheilanthes myriophylla Desv. Berl. Mag. 5: 328 (1811).

CuIHvuAHUA: Hills northwest of Chihuahua, cool cliffs, Oct. 16, 1886, Pringle 829
ZACATECAS: Concepcion del Oro, shaded moist places among rocks and bushes high
in canyon, Aug. 1904, Palmer 260 in pt. and 25

Ranging from our area to Durango and San Luis Potosi and along the
Andes to Chile and Argentina. Differing from C. villosa in having the
upper surface of the fronds glabrous (rather than bearing coarse hairs)
and scales of the lower surface fibrillose
Cheilanthes villosa Davenp. Cat. Davenport Herb. Suppl. 45 (1883).

Cc A: Sierra Guajes, Canon Milagro, on cliffs, not common, Tider 1709;

basalt hill, Johnston & Muller 979; San Antonio de los Alamos, under basalt rocks in
open canyon, Johnston & Muller 904; La Botica, limestone cliffs, scarce, Stewart
2895; Bieras Negras, 9 km. south of Parras, Stanford et al. 202 in pt.; Picacho de
Finules: summit, Stanford et al. 106 in pt.; Jimulco, April 28, 1885, Pringle.
CHIHUAHUA: Sierra Rica, Dec. 1882, Newberry; 8 mi. northwest of Garces: about
tuff cliff, dicate 7978, Sierra Saiite Eulalia, April 6, 1885, Pringle; Sierra Santa
Eulalia, Nov. 2, 1885, Pringle 459; hills west of Chihuahua, about head of aqueduct,
May 8, 1885, Sas

Ranging from trans-Pecos Texas (Davis Mts.) to southern Arizona, and
south into our area.

Cheilanthes Lindheimeri Hook. Sp. Fil. 2: 101. t. 07a (1858).

CHIHUAHUA: 1 mi. west of Poza de Villa, under rocks on small igneous hill,
Johnston & Muller 1386; 3 mi. south of Piramide, terrace along rocky arroyo, under
rocks, Johnston 8113; 11 mi. northeast of Camargo, lava cliff, Johnston 7924;
Chihuahua, rock-crevices, shady riverbank, 1908, Palmer 358

Texas to Arizona and south to Sonora, Durango, and San Luis Potosi.
In our region found only in areas of igneous rock.

Cheilanthes lendigera (Cav.) Sw. Syn. Fil. 128, 328 (1806).

CHIHUAHUA: Mapula Mts., southwest of Mapula station, central canyon on shaded

— a cool cliffs, Oct. 21, 1886, Pringle 828 and 835.

exas (Chisos Mts.); Arizona — along the western Sierra Madre,
Penver northwestern South Ameri
Cheilanthes mexicana mee Bull. Torr. Bot. Cl. 15: 227 (1888).

CH :-. Portrer aM Sierra gee Eulalia, northeast of eae station,
verge ce a ae cliff near hie ummit, Oct. 12, 1886, Pringle 827 (Typ

Only the type epee seen. re gh, in eeanne ie species,
Davenport compared it with C. viscida and C. Parishii, it is much more

322 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

closely related to C. lendigera and C. Schaffneri Moore (Myriopteris rufa
Fée, non C. rufa Don; C. cinnamomea D. C. Eaton). From the former
it differs in its compact habit, smaller pinnules, and nearly rudimentary
indusium. There seem to be no very satisfactory characters whereby to
separate it from the latter as represented by Schaffner 911 and 914, from
San Luis Potosi, referred to C. cinnamomea by Eaton; more material may
show that C. mexicana should be reduced to synonymy under C. Schaffneri.
Cheilanthes pyramidalis Fée, Mém. Foug. 7: 38. t. 25, f. 3 (1857).

IHUAHUA: Mapula Mts., large central canyon southwest of Mapula station,
cool rocky slopes, Oct. 1886, Pringle 832.

Previous students of ferns have usually treated C. pyramidalis as a syno-
nym of C. marginata; even Fournier treated it as only varietally distinct.
It is not a very strong species. True C. marginata of South America, how-
ever, has a broadly deltoid lamina, ovate to short-linear ultimate segments,
and ‘strongly ciliate indusia rey are usually decurrent on the rachillae.
Cheilanthes pyramidalis of Mexico is distinguishable by its narrowly
deltoid to deltoid-lanceolate lamina and its strong tendency to develop
elongate-linear ultimate segments. There is much variation in the degree
of ciliation and decurrence of the indusium (in Fée’s type, as he figures it,
it is strongly ciliate but not at all decurrent), but it is always somewhat
ciliate. The species has been collected in the western states of Mexico and
in Vera Cruz and Guatemala, and apparently it reappears in Venezuela
(Fendler 90). The geographic relationship between C. pyramidalis and
C. marginata is not unlike that between Notholaena nivea and N. incana.
At its extreme northern limit, C. pyramidalis passes into var. arizonica
(Maxon) Broun, characterized by its slender habit, deltoid-ovate lamina,
elliptical to oblanceolate ial segments, and merely papillate-denticu-
late, non-decurrent indusi

Pringle’s collection (832) from the Mapula Mts. does not have the
elongate segments of typical C. pyramidalis, but it does possess relatively
narrow fronds and ciliate, more or less decurrent indusia, and therefore
it is referred to the typical variety rather than to var. arizonica. Maxon,
Amer. Fern Jour. 8: 117 (1918), cites Pringle 1442 as intermediate a
characters. There are two sheets of this collection in the Gray Herbarium,
containing three individuals, two of which are very good C. pyramidalis.
The third, though suggesting the variety, has the relatively narrow frond
of the typical form, and the indusia, though only weakly decurrent, are
definitely ciliate.

Cheilanthes Kaulfussii Kunze, Linnaea 13: 145 (1839).

Cuinvuanvua: Rocky hills northwest of Chihuahua, at base of cliffs in shade,
Pringle 457 and 826

Ranging from Central America north to Nuevo Leon, Durango, and
Chihuahua; trans-Pecos Texas (Davis and Chisos Mts.).

Cheilanthes leucopoda Link, Fil. Sp. 66 (1841).

Coanurta: Sierra Jimulco, 11 km. northeast of Jimulco, Stanford ¢ et al, 82; steep

open north canyon, 6 mi. west of Viesca, Johnston 7745, CHIHUAHUA Sierra Santa

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, I 323

Eulalia, in soil about pian a Pringle 442. DurRaNco: 7 mi. southwest of Choco-
late, shaded slope, Shreve

From our area ranging hut in Durango, reaching San Luis Potosi.
It extends north to the southern escarpment of the Edwards Plateau in
Texas.

Adiantum Capillus-Veneris L. Sp. Pl. 1096 (1753).

VERNACULAR NAMES: Culantrillo; Silantrillo.

Coauutta: Muzquiz, Marsh 1138; Sierra Madera, Canon del Agua, along edge of
water in upper canyon, Muller 3249; Cuatro Cienegas, Marsh 2021; Saltillo, 1880,
Palmer 1430; Saltillo, shady narrow arroyo, abundant on wet rocks, 1898, Palmer 71 ;
Chojo Grande, 27 mi. southeast of Saltillo, common on wet canyon-wall and about
waterfall, 1904, Palmer 360; Sierra Hechiceros, Cafion Indio Felipe, along creek and

bank by water, Stewart 2801; San Lorenzo de la Laguna, 1880, Palmer 1431. Cut-
HUAHUA: Sierra Almagre, Ojo del Almagre, locally common about spring, Johnston &
Muller 1210; Chihuahua, moist crevices on shaded river bank, 1908, Palmer 331.
ZacaTEcAS: Cedros, Lloyd & Kirkwood 114.

Widely distributed in the warmer parts of both hemispheres. Palmer
reported that this fern was sold in the market at Saltillo under the name
“Silantrillo” and notes that it was ‘‘used to assist menstruations in fe-
males.”’

Adiantum tricholepis Fée, Mém. Foug. 8: 72 (1857).

Coauuimta: Hac. La Rosita, Wynd & Mueller 296; La Mariposa, Wynd 691.
CHIHUAHUA: Sierra Santa Eulalia, age — Pringle; rocky ledges in the hills north-
east of Chihuahua, Oct. 10, 1885, Pringle 4:

Known from Texas (south pee of the Edwards Plateau),
Tamaulipas, Coahuila, Nuevo Leon, Vera Cruz, Yucatan, Morelos,
Guerrero, Jalisco, Sinaloa, and Chihuahua.

Polypodium peltatum Cav. Descr. 244 (1802

Polypodium polylepis Roem. ex Kunze, rena 13: 131 (1839).

Coauuita: Mountain 24 km. northwest of Fraile, on a log, Stanford et al. 413.

Ranging northward along the eastern Sierra Madre from central and
southern Mexico.

Polypodium erythrolepis Weatherby, Contr. Gray Herb. 65: 11 (1922).

11a: Canon Sentenela, Sierra del Carmen, Wynd & Mueller 599 and 610;
Sierra del Carmen, Aug. 26, 1936, Marsh 626 (US). CuHiHUAHUA: Portrero Peak,
northeast of Mapula station, cold cliffs, Sept. 10, 1886, Pringle 825 (TYPE).

Ot ise known from western Chihuahua, adjacent Sonora, and
Durango. e above-cited collections from Coahuila, together with one
from western Chihuahua (LeSueur 1128), go very far to break down the
differences between P. erythrolepis and P. peltatum. In them, the abun-
dant, ovate, deeply lacerate-margined scales of the former, which seemed so
distinctive when it was proposed, nearly disappear and are replaced by
suborbicular ones. The surviving distinctions are: P. erythrolepis, stipe
nearly as long as the blade, costa green on the lower surface; P. peltatum,
stipe conspicuously shorter than the blade, costa black on lower surface.

324 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

In addition, P. erythrolepis tends to have narrower rhizome-scales with
narrower, more definitely erose-serrulate hyaline margins; but this is only
a tendency. Furthermore, the collection here cited under P. peltatum
(Stanford et al. 413) is also transitional in that the costa, though some-
what darker than the leaf-tissue, is green beneath and the orbicular scales
of the under surface of the lamina are more or less erose-serrulate. In all
probability, P. erythrolepis would best be treated as a variety of P.
peltatum

Polypodium guttatum Maxon, Contr. U. S. Nat. Herb. 17: 575 (1916).

VERNACULAR NAME: Canahuala.

CoauuILa: Sierra del Carmen, Canon Sentenela, Wynd & Mueller 553; Sierra
Madera, Cafion Charretera, on rocks in moist shaded canyon under oaks at lower edge
of pine-belt, Johnston 8985; shady canyon near Saltillo, abundant, 1898, Palmer 65
(ISOTYPE) ; medicinal herb bought in Saltillo market, 1898, Palmer 651%; Carneros Pass
area, 1880, Palmer 1373; mountain 25 km. northwest of Fraile, Stanford et al. 371.

Ranging from Hidalgo and Guanajuato northward along the eastern
Sierra Madre into eastern Coahuila; Oaxaca; Baja California. Palm
ports this plant as sold in the market at Saltillo. Infusions are fae as
tea and used externally as a remedy for pain in the joints and particularly
those of the shoulder.

Polypodium plesiosorum Kunze var. Bakeri Davenp. Garden and Forest 4: 556
(1891).

CoanuiLa: Sierra de la Gloria, Marsh 1926.
Known from Michoacan, Jalisco, and Nuevo Leon.

a rage | Facer (L.) Watt, var. Michauxianum Weatherby, Contr.
erb. 124:

Des HUILA: Sierr on a ae el, Rancho Agua Dulce, Wynd & Mueller 368; Caracol
Mts., 1880, Palmer . Saltillo, baggies del Puebla, shaded crevices of detached rocks
cit summit, Nov. 3, 1904, Palm

Ranging from Mersin. tinue and Missouri southward and through
eastern Mexico to Guatemala
Polypodium thyssanolepis A. Br. ex Klotzsch, Linnaea 20: 392 (1847).

CutHuAHUA: Cold cliffs in rocky hills northeast of Chihuahua, Oct. 26, 1885,
Pringle 443

Ranging from southern Arizona to western Texas (Chisos Mts.) and
southward to Costa Rica and Andean South America.

SCHIZEACEAE
by C. A. WEATHERBY
Anemia mexicana Klotzsch, Linnaea 18: 526 (1844).
CoanvuttaA: Rancho Agua Dulce, Sierra San Manuel, Wynd & Mueller 319; Hac.

eiatiey dae near Puerto Santa Anna, Wynd & Mueller 229; Caracol Mt., 1880,
Palmer

eae from central Texas south to Hidalgo and Morelos.
MARSILIACEAE
by C. A. WEATHERBY

Marsilea Fournieri C. Chr. Ind. Fil. 418 (1906).
Marsilea minuta Fourn. Bull. Soc. Bot. France 27: 329 (1880), non L. (1771).

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 325

CoanurLa: Cerro de Cypriano, July 1910, Purpus 4525. Curauanua: Wet places
near Chihiehus Pringle 1121.

The species is also known from San Luis Potosi and Jalisco. The Mexi-
can material of Marsilea at hand is scanty and often without fruit or other-
wise unsatisfactory; determinations in the genus are therefore tentative
and subject to correction.

Marsilea mucronata A. Br. Am. Jour. Sci. I. 3:55, f.2 (1847).

CoanuiLa: Torreon, 1898, Palmer 467.

As Braun has suggested, M. mucronata may be no more than a variety
of M. vestita. Baker so treated it, but without making the proper nomen-
clatural combination. The two are geographically separated, M. vestitu
on the Pacific Slope, 1. mucronata in the high plains and eastern Rockies
with outlying stations in the Great Basin. Within these areas the char-
acters of pubescence given by Braun, the relatively abundant, long, slender
and somewhat spreading hairs of M. vestita, the sparse, short, broad and
appressed hairs of M. mucronata, hold so consistently that it seems much
more natural to give M. mucronata some recognition than to reduce it out-
right to M. vestita, as has commonly been done in recent years.

The species ranges from southern Saskatchewan and Alberta south to
Texas, New Mexico, and Arizona. Wright (2112) collected the species in
low eround near oe Elizario, Texas. It is to be expected elsewhere in
the low ground along the Rio Grande at our northern boundary
Marsilea sp.

Cuinuanua: Pond just east of Organos, growing in water up to a foot deep,
blades floating on ey of pond, common, Stewart & Johnston 2048; Rio Conchos
near Camargo, White 2244.

The two above-cited specimens are sterile. They suggest both M.
uncinata and M. mexicana but are not definitely determinable.

PSILOTACEAE
by C. A. WEATHERBY
Psilotum nudum (L.) Griseb. Abh. Ges. Wiss. Gottingen 7: 278 (repr. 130) (1857).
CHIHUAHUA: Hills about 8 mi. northeast of Chihuahua, growing from seams of
rock in canyon, Oct. 1885, Pringle 450

EQUISETACEAE
by C. A. WEATHERBY
aie maint A. Br. Am. Jour. Sci. 46: 87 (1844).
CoAHUIL uzquiz, Marsh 229 and va Sierra Hechiceros, Cafion del Indio
Felipe, aa at edge of creek, scarce, Stewart 3
Widely ranging in the United States ue extending south through Mexico
to Guatemala.
SELAGINELLACEAE
by C. A. WEATHERBY
sit aan rupincola Underw. Bull. Torr. Bot. Cl. 25: 129 (1898).
Sierra Cruces, Cafion de Tinaja Blanca, ledges of igneous rock, st

bf

ems
Meahine: Towasion & Muller 307. Curavanua: 20 km. north of Chihuahua, vol-

326 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

canic hills, rocky talus at ere = cliff, more or less erect, Stewart & Johnston 2122;
Chihuahua, 1908, Palmer 38 in pt.; Sierra Azul, southwest of Mapula, crevices of
igneous rock, Pennell 18646 (US) Meoqui, 1936, LeSueur 1146.

Arizona and western New Mexico south along the western Sierra Madre
to Durango and Guanajuato. A species apparently confined to igneous
rocks. Its stems are assurgent to nearly erect and are ascendingly branched.
The shoots are symmetrical and equally clothed on all sides by appressed
leaves. The leaves are terminated by elongate white setae which form a
conspicuous tuft at the end of sterile shoots.

Selaginella viridissima Weatherby, sp. nov.

Caules graciles, foliis inclusis circa 1 mm. diametro, elongati (ad 15 cm.

longi), prostrati tegetem magnam intricatam laxam formantes, parce

ramulis plerumque brevibus (1 cm. vel minus longis). Folia uniformia,
arcte adpressa, saturate viridia, plana vel leviter convexa, oblongo- me
earia, acuta vel obtusiuscula, pler mque 1.6—2 mm. longa, 0.3—0.4 m
lata, dorso coins sulcata, seat margine ciliis Soaks 0.1 mm.
vel minus longis folii apicem versus ad denticulos reductis praedita. Seta
terminalis nulla. Spicae apice caulis ramorumque gestae usque ad 1 cm
longae. Sporophylla ovato- deltoidea, acuminata, e basi leviter dilatata
subsagittata subabrupte in acuminem longam contracta, convexa, utroque
margine crebre minuteque serrulato-ciliolata, 1.8-2 mm. longa, 0.3—0.4
mm. lata, sulcae medianae utroque latere vitta pallida ornata, sine seta
terminali 7 pao microsporangia intermixta. Macrospori 0.4—0.45

. diametro, ies dense leviterque reticulato-rugosi. Microspori auran-
tect circa 40 uw

COAHUILA: er Hes los Osos, west end of Sierra Fragua, 2-3 km. north of Puerto
Colorado, forming mats in shaded canyon, Sept. 1, 1941, Johnston 8683 ; Sierra Mojada,
Cafion Calabasa, fairly common on shaded cliffs 100 m. below the crest, hanging in
mats 1 m. in diameter, Oct. 27, 1941, Stewart 2204 (type, Gray Herb

A plant with slender elongate much branched aie a stems forming
loose mats. The minute dark green leaves are acute, devoid of setae, and
closely appressed to the rather wiry elongate stems. It grows on limestone
at the two stations where it has been collected. A pretty species, related
to S. extensa and S. Sartorii, from both of which it may be distinguished
by its muticous leaves. From S. mutica and its immediate allies, S. viri-
dissima differs in its much longer, relatively narrower, and plane leaves.
Selaginella macrathera Weatherby, sp. nov.

Caules repentes, ad 8 cm. longi, usque ad apicem parce radicantes,
foliis inclusis circa 1.5 mm. diametro, bipinnatim crebreque ramosi (spatiis
inter ramos ca. 5 mm.). Folia uniformia, laxe adpressa, hang beaten
oblongo-linearia, seta exclusa plerumque 1.8-2.2 mm. longa, 30.4
lata, acuta, ventro plana, dorso leviter convexa conspicue nimeedicre
sulcata, basi fasciculum pilorum brevium ciliis marginalibus similium
margine utroque 10-12 ciliis brevissimis 0.1 mm. vel minus longis apicem
folii versus ad denticulos hyalinos reductis ae apice in setam gracilem

m. longam scabriusculam desinentia. Spicae ad 1 cm. longae apice
caulis Cami ie superiorum gestae. Sporophylla anguste deltoidea,
1.8-2 mm. longa, basi leviter sagittata 0.6—-0.8 mm. lata, valde convexa

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 327

vix Carinata, dorso leviter sulcata, oes breviter crebreque serrulato-
ciliolata, seta ut in foliis praedit Megasporangia absentia vel pauca
unicum visum apicem versus nee gestum. Megaspori visi immaturi

leviter, latere altero valde crasseque reticulato-rugosi. Microsporangia
multa; microspori aurantiaci, ca. 40 u diametro, irregulariter tuberculati.

CuIHvuAHUA: Sierra del Virulento, 2-3 mi. east of Rancho Virulento, ledges on
north-facing lava cliffs, common and forming i Aug. 11, 1941, Johnston 8067
(TYPE, Gray Herb.).

A plant with creeping stems. The shoots are symmetrical and equally
clothed on all sides with appressed leaves bearing a very long white ter-
minal seta. In spite of its repent habit, the species apparently belongs
to the group of S. rupincola, from all members of which it is distinguished
by its combination of very short cilia and very long terminal seta
Selaginella habia Hieron. Hedwigia 39: 298 (1900).

CoaHuILA: El ens ee 13, 1939, spec 1173 (US); Sierra Gavia, 5 mi.
north of Gauci: rock So ohnsion 7208; Sierra San Vicente, Cafion Espantosa,
Schroeder 72; Saltillo, ae ea, (US); aera del Pino, 4 mi. northeast of La Noria,

ane de ‘cliffs, Sohiaeh 8854; Sierra Mojada, April 19, 1892, Jones 485 (US); 5 km.
south of Sierra Mojada, Harvey 1265a. Cuimuafiua: Sierra ‘Almagre, moist limestone
ledges in shaded canyon, Johnston & Muller 1192; east slopes of Sierra Santa Eulalia
2 km. north of San Antonio, Harvey 1507. Zacatecas: Cedros, stony hills, Lloyd &
Kirkwood 142; Lloyd 20 (US).

Ranging from the Edwards Plateau and its escarpments, in Texas, west
to southeastern New Mexico, and south through our area and the moun-
tains of northeastern Mexico to San Luis Potosi, and possibly to Puebla.
The type specimen, Wright 829, was collected June 25, 1849, “on hills near
Turkey Creek, on flat rocks slightly covered with earth.” This locality is
near the present town of Cline, in western Uvalde County, Texas. The
species appears to grow only on limestone. It is a creeping plant carpeting
the ground under sheltering rocks or on ledges on north-facing cliffs. The
erect fruiting spikes, 1-3 cm. long, are commonly produced in great
abundance. The rather firm leaves have a short slightly tawny terminal
seta. The leaves tend to be laterally arranged and the shoot is hence
somewhat dorsi-ventral.

Selaginella Sheldoni Maxon, Proc. Biol. Soc. Wash. 31: 171 (1918).

CoanurLa: Picacho de Jimulco, summit, 13 ssi east of Jimulco, Stanford et al.
118, CHtHuaHuA: Chihuahua, 1908, Palmer 38 in

Southwestern Oklahoma, central and western Texas, and New Mexico.
A creeping species with somewhat dorsi-ventral shoots. The leaves tend
to be laterally spreading, and are terminated by a slender elongate white
seta. ost of the known stations for the species are in areas of igneous
rock.

Selaginella Parishii Underw. Bull. Torr. Bot. Cl. 33: 202 (1906).
Coauutta: Saltillo, Mil 105 (US); Cerro Vega, west of Saltillo, crevices of sand-

328 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

stone, Pennell 17272 (US); mountains near nae June 1909, Nil (US). ZacaTEcas:

Near Concepcion del Oro, sheltered rocky ledges, plant very dark green, 1904, Palmer

306 (IsoTYPE); Tarey Canyon, near Cedros, ie of slate rock, Feb. 7, 1911, Chaffey
(US).

An endemic species with relatives in sasha Mexico and southwestern
United States (cf. Maxon, Smithsonian Misc. Coll. 72: no. 5, p. 4. 1920).
A prostrate repent plant ‘with strongly dorsi-ventral shoots. The rather
broad and thin acute leaves, without setae, are laterally widely spreading
under favorable nani but curve upward and become more or less
connivent when

The specimens ae associated with the type collection of S. Parishi
differ from it somewhat in gross appearance, and at one time they were
annotated by Dr. Maxon as constituting a possible new species. In details,
however, they are very close to the Palmer isotype; the more conservative
course is to leave them in S. Parishii pending a thorough revision of the
Mexican species of this group.
Selaginella lepidophylla (Hook. & Grev.) Spring, Monog. Lycopod. 2: 72 (1849).

VERNACULAR NAMES: Flor de la Pefia; Siempre Viva.

Coanutta: Muzquiz-La Mariposa, Dec. 5, 1936, Marsh 1041; Sierra Encantada
west of Buena Vista, July 14, 1938, Marsh 1414; 6 mi. north of Hipolito, limestone
outcrop on slope, Johnston 7236; La Rosa, dry mountain slope, Wynd & Mueller 45;

Picacho del Fuste, north-facing bank of cemented gravels, common, Johnston 8442;
one kr of Cafiada Oscuro near Tanque La Luz, steep slopes of escarpment, common
on and off gypsum beds, Johnston 8494; west end of Sierra Fragua, Aguaje del Pajarito,
anor hie limestone eae Johnston 8806; Cahon Blanco, Sierra Margaritas, open
slopes, common, Stewart 2914. CuHimuaHuA: Sierra San Carlos, road to mine, on
cliff near parece mouth, Johnston & Muller 41. Zacatecas: Cedros, rocky hills,
Kirkwood 134.

Western Texas and New Mexico south to southern Mexico. A common
plant on north-facing dry rocky slopes and ledges in limestone areas. The
plant avoids the direct sun but grows in open situations in which it can
only have water available during and for a short time after desert showers.
Because it avoids direct sunlight and commonly occurs in abundance only
on north-facing situations, it serves as a handy and rather reliable indicator
of direction to a traveller in the desert mountains where it flourishes.
During most of the year the plant is an inconspicuous brownish ball of
brittle inrolled leaves as big as one’s fist. Only after a rain, when the
fronds unroll and reveal their green upper surfaces, forming flat bright
green rosettes and magically bringing unexpected verdure to gray cliffs and
banks, does one realize how common and abundant it is in a region. The
plant is a slow growing perennial and probably grows for a good many
years. Some old plants have their rosettes lifted as much as 5 cm. above
the substratum by the accumulation of half decayed fronds of seasons past.
A surprising amount of dirt and gravel collects within the rosette about
the base of the old fronds.

Selaginella pilifera A. Br. Ind. Sem. Hort. Berol. App. 20 (1857).

Selaginella Pringlei Baker, Handb. Fern Allies 88 (1887).

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 329

Selaginella pilifera var. Pringleit (Baker) Morton, Amer. Fern Jour. 29: 15 (1939).
Coanuira: Yerda Spring, near Muzquiz, Marsh 270; Muzquiz-Mariposa, Marsh
1042; Saltillo, Arséne 10677 and Palmer 321 (US, fide Morton) ; western end of Sierra

8748. CHIHUAHUA: Sierra Santa Eulalia, March 30, 1885, Pringle 211 (isotype of
DE en Sierra Almagre, moist shaded limestone cliffs in deep canyon, Johnston &
Muller 1

ae Texas and adjacent southeastern New Mexico south to northern
Sonora, San Luis Potosi, Nuevo Leon, and northern Tamaulipas. Appar-
ently confined to limestones. A plant with habit similar to S. lepidophylla,
but with more slender, less rigid stems and paler green, bristle-tipped
leaves. In western Coahuila and adjacent Chihuahua it is uncommon and
found on sheltered moderately moist cliffs in the oak-belt, and not with
Yucca, Dasylirion, Hechtia, Euphorbia antisyphilitica, Notholaena sinuata,
etc., the associates of Selaginalla lepidophylla, on the lower and open
slopes of the mountains.

The type of S. pilifera is given as based on “‘Specimina Texana in mon-
tosis ad fluvium Rio Grande infra El Paso uno cum S. lepidophylla a cl.
Wright anno 1849 collecta comm. Dr. G. Engelmann.” In the Gray
Herbarium there is only a single collection of S. lepidophylla made by
Charles Wright during 1849. This is his no. 827, collected from “high
rocky bluffs of Devils River, July 22, 1849” in southern Val Verde County,
Texas. His field-notes for 1849 have no entry which can be identified as
pertaining to another collection of this species. It is possible, therefore,
that the type of S. pilifera actually was collected in Val Verde County,
Texas, rather than near the Rio Grande (presumably in the Quitman Mts.)
below El Paso, as originally stated.

Morton has discussed the relationship of S. pilifera and S. Pringlei and
has concluded that these two species differ only in trivial details, the former
having entire, the latter having minutely serrulate margins on the lateral
leaves. Except for the type, all the material he cites as belonging to
typical S. pilifera comes from eastern Coahuila, Tamaulipas, and Nuevo

eon. The material from trans-Pecos Texas, New Mexico, Chihuahua,
and San Luis Potosi he places in the var. Pringlei. Recently, however,
Stephen White (522) has collected material in the valley of the Rio
Bavispe, in northeastern Sonora, which has distinctly serrulate leaves.
Since the characters of leaf-margin are weak at best and are not geographi-
cally correlated, it seems best to permit S. Pringlei to subside into
synonymy.

Z
o

PINACEAE

Pinus cembroides Zucc. Abh. Akad. Wiss. Miinchen 1: 392 (1832).

Pinus osteosperma Engelm., in Wislizenus, Mem. Tour Mex. 89 (1848).

VERNACULAR NAME: Pifion.

Coanvuma: Sierra Encantada, Stewart 1434, Marsh 1358; Sierra Madera, Canon
del Agua, Muller 3229; Sierra del P ino, Johnston & Muller 523, Stewart 1243; Sierra
Gavia, Wynd & Afiielicy 165, Muller 3064; Sierra San Vicente, Cafion Espantosa,
Schroeder 95; 6 mi. east of Saltillo, 1880, Palmer; Chojo Grande near Saltillo, 1905,
Palmer 768; Buena Vista, Gregg; Carneros Pass, Pringte 2659 and 4018, Pilner: near

330 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXIV

General pcan on mesa, Pringle 13664; summit of Picacho de Jimulco, oo et al.
10. Curmuanua: Sierra Rica, Stewart 2506; Sierra Diablo, Stewart

From aan, ee Mexico, and trans-Pecos Texas pers : Hidalgo.
A small tree, usually 5-10 m. tall, commonly growing along arroyo-banks,
on ridges, and on steep open slopes, usually associated with Juniperus.
In western Coahuila and eastern Chihuahua the tree is seldom abundant
in any locality.

Pinus sleriores Gordon, Pinetum 204 greet Shaw, Gard. Chron. III. 38: 122.
fig. (1905), Pines of Mexico 7. tab. 2 (1909).

Pinus latisquama Engelm. Gard. Chron. II. 18: 712. fig. (1882).

TUILA: West end of Sierra Fragua just ‘north of Puerto Colorado, abundant,
Johnston 8735; Sierras Negras, 9 km. south of Parras, Stanford et al. 148; General
Cepeda, Nelson 6140; Carneros Pass, Palmer 1299 in 1880 (type of P. ee
Pringle 2293 ‘i 13207A, Shaw. Zacatecas: Pico de Teira, southwest of Cedro
1908, Lloyd 3

This ite swell marked pinyon-pine is known only from scattered sta-
tions in our area, and from Hidalgo and the peak of Orizaba in east- central
Mexico. In our region it was first collected in March 1880, in the Carneros
Pass area by Palmer. His material became the type of Pinus latisquama
Engelm. The trees, which rarely become more than 7 m. tall, have a broad
rounded crown whose silhouette from a distance is more suggestive of an
oak tree than a pine. The ellipsoidal cones, russet when fully ripe, are
borne on stalks and at the ends of the long supple brittle pendulous
branchlets. The trunk becomes 2—6 dm. thick and commonly branches
less than 2 m. above the ground. The bark is grayish, somewhat furrowed
on the trunk and smooth on the branches. Where it has been found, the
pine grows with scrub oaks and is confined to sheltered slopes and canyons.
Pinus Ayacahuite Ehrenb. ex Schlechtend. Linnaea 12: 492 (1838).

VERNACULAR NAMES: Acanita; Pinaveta.

OAHUILA: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 630; Sierra del
Carmen, Sept. 12, 1936, Marsh 821; Sierra Madera, Muller 3210, Johnston 8998;
Carneros area, March 1880, Palmer; sierra 26 km. northwest of Fraile, Stanford et al.
456; General Cepeda, Nelson 6136.

I have seen cones for only one of the cited collections, Johnston 8998.
This has seeds with the wing at least 10 mm. long. The form of the
species growing in the western Sierra Madre, from Arizona to Durango,
has the wing on the seeds only a few millimeters in length and has been
distinguished from the typical plants under the name var. brachyptera
Shaw (= P. strobiformis Engelm.). Some plants from the Sierra Madre
of Nuevo Leon (Muller 1244, 2283) also have seeds with very short wings.
The variety is probably also represented in Coahuila.

The species, with its varieties, ranges from Central America northward
along the eastern Sierra Madre into Coahuila and along the western Sierra
Madre into Arizona. It commonly attains a height of 15 m. and in favor-
able situations may approach 30 m. In Coahuila it associates with
Pseudotsuga to form the forests on cool shaded north-slopes in the higher
mountains. Along canyons and on open slopes its lower altitudinal limit
is several hundred meters above that of Pinus arizonica and about 100 m.
below that of Pseudotsuga.

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, I 331

Shaw, Pines of Mexico 12 (1909), reports Pinus flexilis from the moun-
tains south of General Cepeda upon the basis of Nelson 6136. That speci-
men, in my judgment, appears referable to P. Ayacahuite. It has seeds
with a broad wing about 8 mm. long, according to Shaw’s manuscript notes.
Pinus Greggii Engelm. ex Parlatore in DC. Prodr. 162: 396 (1868) ; Shaw in Sargent,

s and Shrubs 2: 53. tab. 124 (1907); Shaw, Pines of Mexico 28. tab. 21

(1909 ‘
CoaHUIL San Antonio de los Alanzanes, abundant, 30-50 ft., ue 31, 1848,
Gregg 402 nea. Canon de las Iglesias, near Saltillo, Shaw, Pringle 1 0142.

A species known only from the Sierra Madre of Nuevo Leon and adja-
cent Coahuila, and perhaps Hidalgo. It is a 3-needle pine with sessile re-
flexed long-persistent tardily opening cones 6-12 cm. long. The cones are
light colored (usually café au lait) and lustrous. The needles are very
slender and 7-10 cm. long.

Pinus arizonica Engelm. in Wheeler, Rep. U. S. Geol. Surv. W. of 100th Merid. 6:
260 (1878).

VERNACULAR NAMES: Pino; Pino Real.

Coanuma: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 650; Sierra del
Carmen, Sept. 12, 1936, Marsh 830; west of Buena Vista [ ?Sierra Encantada], Marsh
2290; Sierra del Pino, Johnston & Muller 446; Sierra Madera, Muller 3208, Johnston
8935; Sierra Gloria, Marsh 1931; Sierra Caracol, 1880, Palmer; Chojo Grande near
Saltillo, 1905, Palmer 769; Carneros area, Palmer, Pringle 2826; mountains south of
General Cepeda, Shaw, Pringle 10139.

Widely distributed and frequently the dominant pine in the eastern
Sierra Madre of Nuevo Leon and Tamaulipas, south at least to Miqui-
huana and Doctor Arroyo. Also in the Chisos Mts. of Texas, southern
New Mexico and Arizona, and northeastern Sonora. In northeastern parts
of Mexico this pine has passed mostly as P. Montezumae, P. pseudostrobus,
and P. ponderosa. Watson, Proc. Am. Acad. 18: 158 (1883), reported it
from Coahuila as P. Montezumae and P. teocote. I am unable to separate
the pine of Coahuila and the eastern Sierra Madre from typical P. arizonica
of Arizona. Sudworth, Pine Trees of the Rocky Mts. Region, U. S. Dept.
Agr. Bull. 460: tab. 16, 17 (1917), gives an excellent illustration of the
Arizonan plant. It agrees perfectly with Coahuilan material. In general
appearance the Coahuilan plant much resembles forms of Pinus ponderosa
growing in the Rocky Mountains. Its bark is the same. It differs from
Pinus ponderosa in its somber brownish (rather than russet), more or less
asymmetric, frequently stalked cones, weak, non-pungent umbo on the
cone-scales, 3—5 needles, usually glaucescent branchlets, and more southern
distribution. From P. Montezumae our tree differs in its smaller and pro-
portionately broader cones, usually glaucescent branchlets, and northern
range. From P. pseudostrobus it differs in its coarser more rigid non-
pendulous foliage, more rigid and woody short-stalked or sessile cones, and
northern distribution. In northeastern Mexico P. arizonica is to be con-
fused only with P. Hartwegii, a tree of high altitudes in the Sierra Madre,
which has very coarse loose needle-sheathes, conspicuous long-persistent
bud-scales, non-glaucescent branchlets, somewhat thinner less rigid cone-
scales, and cones averaging slightly smaller.

332 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XXIV

In Coahuila P. arizonica is probably the most common pine, forming
open forests in the open valleys and on the drier slopes and ridges in the
higher mountains. From its selections of habitats it appears to be in-
tolerant of shade. It commonly grows 10-20 m. tall, with a clear trunk
4—10 dm. thick for a quarter or third of its total height.

Pseudotsuga taxifolia (Lam.) Britt. Trans. N. Y. Acad. Sci. 8: 74 (1889).

VERNACULAR NAMES: Guayamé; Hallarin.

Coanvutta: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 632; Sierra del
Carmen, Sept. 12, 1936, Marsh 822; Sierra Madera, Muller 3221, Johnston 8995; Sierra
sloria, Marsh 1885; Carneros area, March 1880, Palmer; mountains 26 km. northwest

Growing on slopes and along canyons in cool shaded places in the higher
mountains, forming trees 10-30 m. tall. From Hidalgo extending north
in the eastern Sierra Madre into our area, and north into the Chisos Mts.
of Texas. Widely distributed in western United States.

Abies sp.
VERNACULAR NAME: Huallame.
Vicinity of Carneros Pass, tree 40 ft. tall, 18 inches diameter,

et al. 45

The two collections are unaccompanied by cones. They appear to be
identical, however, with a very distinct Abies collected, in 1938, by Prof.
Maximino Martinez in the Sierra Madre near Santa Catarina, between
Monterey and Saltillo. The new species involved will soon be published in
Mexico City. The foliage of the present species is not distichous. Its
numerous crowded short rigid leaves ascend from all sides of the coarse
branchlets in a manner more suggestive of a Picea than an Abies. Its
short leaves bear numerous stomates on the flattened or broadly convex
upper surface, and their vascular bundles, though clearly juxtaposed,
remain distinct for most of their length. The hypoderm seems unusually
well thickened under the middle third of the lower leaf-surface. These
are all characters which permit the species to be quickly distinguished
from A. religiosa, of central Mexico, and from the other, unnamed, Coa-
huilan species of the genus, the only Mexican plants to be confused with it.
Among the species found in the United States, the present fir most sug-
gests A. lasiocarpa. hat northern high altitude species, however, has
more pointed less regularly arranged leaves, whose resin-canals are large
and distant from the lateral margins of the leaf. I doubt if it has any close
relationship with the present Coahuilan species. It may be noted that
Palmer’s collection from “the Sierra Madre 40 miles south of Saltillo,”
cited as Abies religiosa by Watson, Proc. Am. Acad. 18: 158 (1883), and
by Rehder, Jour. Arnold Arb. 20: 283 (1939), is identical with the Palmer
collection which I have cited with more explicit geographical data above.
Abies coahuilensis sp. nov.

Arbor ad 30 m. alta habitum Pseudotsugae taxifoliae simulans; trunco
ad 9 dm. crasso in parte inferiore cortice fusco rugoso crasso praedito, in

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, | 333

parte superiore pallido sublevi; ramulis brunneis hirtellis eos A. religiosae
simulantibus; foliis subdistichis e ramulis sub angulo 60—90° abeuntibus

tosis, subtus conspicue bisulcatis costa prominente margine recurvo serie-
bus stomatum 4—5 congestis, intus canalibus resiniferis solitariis subepi-
dermalibus ad utrumque marginem faciei inferioris donatis, fascias fibro-
vasculares conjunctas gerentibus, hypodermate sub facie superiore folii
interrupta solum sub epidermate partium marginum et partis mediae Sa
inferioris continua donatis; strobilis subsessilibus subcylindricis ca. 10 c
longis supra basim ca. 4 cm. diametro; squamis 24-28 mm. latis 14-19 mm.
longis, margine exteriore hirtellis sursum curvatis, alis minute eroso-denticu-
latis, margine interiore fere recto utrisque lateribus basi unguis 4-6 mm.
longi et sinibus rotundis 1-2 mm. profundis 2~3 mm. latis donatis; bractea
squamae haud vel vix exserta quam squama 2/5—4/5 longa, apicem versus
6-7 mm. lata deinde basim versus gradatim attenuata, apice truncata erosa
mucronata; seminibus 7—8 mm. longis, alis 12-13 mm. ‘latis ca. 1 cm. longis.

AHUILA: Corte Branco fork of Charretera Canyon, Sierra Madera, frequent
above 7500 it. alt., Sept. 14, 1941, Johnston 9050; head of La Pipa fork of Charretera
Canyon, Sierra Madera, dense cool shady conifer forest on steep north slope, Sept. 13,
1941, Paknsion 9010 (rvpE, Arn. Arb.)

This fir is frequent in the dense conifer forests on the northern slopes
of the main ridge of the Sierra Madera in the drainage of Charretera
Canyon. The tree grows mixed with Pinus Ayacahuite, Cupressus, and
Pseudotsuga, but it is very much less common than these other trees. In
appearance it simulates the Pseudotsuga so closely that I was unable to
distinguish them at any distance and was able to make positive identifica-
tion in the field only after examining the terminal buds on the branchlets,
observing cones or cone-axes on the trees, or discovering cones or cone-
scales beneath the trees. Timber has been cut and dragged out of the
forests of the Sierra Madera for many years. Questioning men who know
the forests and have cut timber there, I could find no evidence that this
Abies had ever been distinguished by the local people from the more
common and very similar appearing Pseudotsuga, well known to them as
‘““Guayamé.”’

This species, and the Abies previously listed, were recognized as unnamed
species and were described before I learned that Prof. Maximino Martinez
was at work on a monograph of the Mexican species of the genus. Material
of the two species of Abies was sent Prof. Martinez, and from the notes,
specimens, and photographs he so obligingly sent in return I was readily
able to identify the species I report from the Carneros Pass area and from
northwest of Fraile with his material from Santa Catarina which he will
soon publish as a new species. With this identification Prof. Martinez
agrees. We are in disagreement, however, regarding the identification of
the Abies of the Sierra Madera. Prof. Martinez identifies it with material
collected by J. H. Faull near El Salto, in southwestern Durango, which
will be described as a new species in the near future.

Through the courtesy of Prof. Faull I have been able to make a detailed

334 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

study of his collections from El Salto which Prof. Martinez identifies with
the present plant of the Sierra Madera. The Durango collections obviously
represent a good undescribed species. Its vegetative characters suggest a
relationship with A. religiosa. It differs from that species, however, in
proportionately broader cone-scales and very short non-exserted bracts.
In shape and size of the scales, bracts, and seeds, the Durango plant is
very much like that from Coahuila. The vegetative characters, however,
differ in a number of striking details. The Coahuilan plant has twigs
which are dusky and duller brown in color, and which are not glabrous
but evidently hispidulous. The leaves have a much thicker epidermis,
are heavier and firmer in texture, and are green. They are not glaucous
when young. Their petiolular base, better developed than in the Durango
tree, is conspicuously erect, appressed to the stem below the middle, and
above departing from the stem in an abrupt curve. The leaves of the
Durango species are usually straight or nearly so and spread from their
point of attachment. The Coahuilan tree has the lower surface of the
leaf with very prominently thickened midrib and margins, and accordingly
very deeply and narrowly bisulcate. The lower surface of the leaf of the
Durango tree has a very much less thickened and prominent midrib and
margins, and the intervening grooves are shallow and broad, bearing 4-10
rows of stomates. The upper leaf-surface in the Coahuilan plant bears few
if any stomates, whereas that from Durango has several rows of them down
the middle.

In gross appearance the plants also differ. Specimens of the Durango
plant are suggestive of A. religiosa. Those of the Coahuilan tree suggest
Pseudotsuga or some the Abies of the United States. The habit of the
trees appear to differ also. Professor Faull tells me that the tree at El
Salto has a distinctive local name, is well known and distinguished by the
local people, and has the characteristic habit permitting it to be instantly
recognized as an Abies, The Coahuilan tree grows intermixed with
Pseudotsuga and simulates that tree to a truly remarkable degree. The
two have not been distinguished by local people knowing the forests, and
I must confess that I should not have been aware than an Abies was
growing with Pseudotsuga in the Sierra Madera had I not chanced upon
Abies cones cut down by squirrels.

Geographical distribution also suggests that distinct species of Abies
are present in the Sierra Madera and about El Salto. The flora of the
coniferous forests of the Sierra Madera is made up largely of species that
have migrated along the Sierra Madre Oriental or have extended south
from the United States by other routes. There is some evidence that a
small proportion of the species may have reached the Sierra Madera by
migrations from the Sierra Madre Occidental, but this group of species 1s
made up of plants otherwise found in the forests of northern Chihuahua
and adjoining Arizona. Since other species show absolutely no evidence
of direct floristic connections between the Sierra Madera and the El Salto
area, far away in the Sierra Madre Occidental in southwestern Durango,

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO. I 335

this fact lends additional support to the belief that the Abies of Coahuila
and Durango are different.

The precise relationships of A. coahuilensis are not certain. Among
the Mexican species it is probably most closely related to the undescribed
plant of Durango. It is readily distinguished from A. religiosa of central
Mexico by its non-exserted cone-bracts, proportionately shorter cone-scales,
smaller seeds, and markedly bisulcate leaves. In many ways, however, it
shows closer relationships with A. concolor and A. grandis, of the western
United States. From the former it differs in its slender brownish his-
pidulous twigs, more slender green bisulcate leaves, and few or absent
stomates on the upper leaf-surface. Its hispidulous twigs, more slender
acutish leaves, not markedly glaucous beneath, and proportionately broader
cone-scales readily separate 4. coahuilensis from A. grandis.

TAXODIACEAE

Taxodium mucronatum Ten. Ann. Sci. Nat. IIT. 9: 355 (1853).

VERNACULAR NAME abin

CoaHvUIL ac. Wea riposa, Wynd 694; Muzquiz, Sabinas pee Marsh 406;
Sabinas, along Sabinas River, Johnston 7040; Villa Juarez, Marsh 2

Reported in wet places near Parras, Fedde Repert. oe oo 14: 100
(1915). To be expected also along the Rio Nazas west of Torreon.

CUPRESSACEAE

Cupressus arizonica Greene, Bull. Torr. Bot. Cl. 9: 64 (1882).

VERNACULAR NAMES: Cedro; Pinabeta; Tasco

Coanvuma: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 502; Sierra
Madera, Muller 3193, Johnston 8925; mountains near Saltillo, Pringle 13612; Carneros
area, 1880, Palmer 1293; Sierra Encarnacion, Nelson 3894a; sierras 24 km. northwest
of Fraile, Stanford et al. 383; Sierra Parras, Purpus 4987. CHIHUAHUA: Sierra Rica,
Cafion Madera, Stewart 2464 and 2479; Sierra Santa Eulalia, Pringle 178.

This species ranges from Texas (Chisos Mts.) and Arizona south to
Zacatecas and San Luis Potosi. It appears to differ from C. Benthami,
of central Mexico, in its larger and more leathery cones and somewhat
coarser usually paler foliage. In the Sierra Madera I found this cypress
common in the luxuriant coniferous forests on north slopes, where it grew
with Pinus Ayacahuite and Pseudotsuga taxifolia and formed straight single
trunks, commonly attaining 20-30 m. in height. About the lower edge of
the conifer-belt it was occasional along arroyos associated with Pinus
arizonica. In such situations it rarely reached 10 m. in height and was
broadly conical in form. The younger branches are smooth and brown.
The trunks of younger trees have irregular furrows in the thick persistent
laminated old bark. The tall trees in the dense forest have a tight check-
ered bark. Mr. Stewart reports that the tree was common on Sierra Rica
along the shady bottom of Cafon de la Madera and there formed a tree
rarely up to 20 m. tall with a trunk 12 dm. thick. In that locality it was
locally called ‘‘Pinabeta.”. Two of Mr. Marsh’s collections, nos. 803 and
1999, from the Sierra del Carmen and Sierra Gloria, are sterile but seem to
represent this species.

336 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Juniperus pachyphloea Torr. U. S. Rep. Explor. Miss. Pacif. 4: 142 (1857).

VERNACULAR NAMES: Cedro; Tascate.

CoaH Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 535; west ot
Buena Vista Ranch { ?Sierra Encantada], Marsh 1359 and seen ; Sierra del Pino , com-
mon in pine forest, Johnston & Muller 528, Stewart 2294, > Carneros Pass area
March 1880, Palmer 1296. CHIHUAHUA: Sierra Rica, a ae rl scarce on open

slopes, Stewart 2549.

The above-cited specimens, agreeing with most Texan collections, have
resin exuding from only a few scattered leaves or from none at all. Other-
wise the plant agrees well with the “Alligator Juniper” of Arizona and New
Mexico, the type of which came from the Zuni Mts. in western New Mexico.
It is a large tree with heavy trunk covered with characteristic checkered
bark. It is closely related to J. Deppeana Steud. (based on J. mexicana
C. & S., not Spreng.), a Mexican tree, also with checkered bark, growing
in the states of Vera Cruz, Puebla, and Hidalgo, and apparently also in
Zacatecas, Durango, Chihuahua, and Sonora, which differs slightly in the
less conspicuous resin-glands on its smoother less prominently keeled
leaves. In J. pachyphloea the leaves are usually sulcate on either side of
the gland and consequently appear to be three-ridged. Only in Sonora and
Chihuahua do the ranges of J. Deppeana and J. pachyphloca appear to
approach one another. The type of J. Deppeana came from the margin of
the plateau in western parts of Vera Cruz, between Las Vigas and Perote,
northwest of Jalapa and north of Cofre de Perote. Curiously it does not
appear to have extended its range northward into the Sierra Madre of
Nuevo Leon or Tamaulipas.

Juniperus flaccida Schlechtend. Linnaea 12: 492 (1938).

VERNACULAR NAMES: ‘Tascate; Cedro.

Coauurta: Sierra del Carmen, Sept. 7, 1936, Marsh 794; Sierra San Manuel, Rancho
Agua Dulce, Wynd & Mueller 359; Sierra del Carmen, 8 km. east of Hac. Encantada,
Stewart 1585; Hillcoat Mesa, west of Hac. Encantada, Marsh 1425; Mesa Grande,
40 km. northwest of Hac. Encantada, Stewart 1617; Sierra Eavantada, Stewart 1424;
west of Rancho Buena Vista [ ?Sierra Encantada], Marsh 1360; Sierra del Pino,

Muller 3209, Johnston 8936; Sierra Mojada, Stewart 1058; mountains near Saltillo,
northwest of Fraile, Stanford et al. 390; General Cepeda, Nelson 6122; Sierra Pata

Jimulco, 11 km. northeast of Jimulco, Stanford et al. 133. Cuimuanua: Sierra
Almagre, Johnston & Muller 1176; Sierra Diablo, Stewart 942.

Ranging from central Mexico (the type came from the mountains of
Hidalgo), this tree extends northward into the Sierra Madre of Chihuahua
and through our area into the Chisos and Davis Mountains of Texas. The
leaves are decussate and the branchlets are distichous, making the leafy
branchlets flat and more or less fan-like. The conspicuously two-ranked
smooth elongate acute cuspidate leaves, the pendulous (‘‘weeping’’) leafy
branchlets, and the large resinous multi-seeded non-baccate fruits all unite
in permitting the ready recognition of the species. It is a tree with reddish
fibrous bark. In our area it commonly becomes 4-8 m. tall and appears

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, I 337

to favor limestone. It grows scattered on slopes or more commonly along
arroyo-banks, in sheltered places in the oak and lower pine belt.

Juniperus Ashei Buchholz, Bot. Gaz. 90: 329 30).
Juniperus occidentalis var. conjungens nee oe Acad. Sci. St. Louis 3: 590
877).

Juniperus tetragona var. oligosperma Engelm. 1. c. 591.

CoanumLa: Saltillo, frequent in highlands, shrub 5-10 ft., Gregg 106 (isotype of
var. conjungens) ; escarpment near mines, Potrero de la Mula, tree 12 ft., Johnston
9195; Sierra San Manuel near Rancho Agua Dulce, Wynd & Mueller 360; near Puerto
Santa Anna, Hac. Mariposa, Wynd & Mueller 284.

This is the well known “cedar” of the Edwards Plateau of central Texas,
which has passed as J. sabinoides, J. mexicana, and J. tetragona. These
names, however, properly apply to very different species of southern and
central Mexico. From our area the species extends north into central
Texas and from thence to Missouri. Under the name J. mexicana, its dis-
tribution in the United States has been discussed by Hopkins, Rhodora
40: 425 (1938). It forms a small tree and has dark blue globose berries
about 8 mm. in diameter.

Juniperus erythrocarpa Cory, Rhodora 38: 186 (1936).

VERNACULAR NAME: Tascate.

CoanuiLa: Sierra Hechiceros, common along canyons, Johnston & Muller 1290,
Stewart 168 and 169; Castillon, one tree on gypsum flat by corrals, wees & Muller

r
Rancho El Pino, southeast of Sierra Rica, frequent, Stewart 2426 and 2574; low hills
13 mi. west of San Carlos, Johnston & Mu ler 28.

Forming a dense rounded bush 10-15 dm. tall or a tree up to 6 m. high.
The fruit is not blue, but at maturity it is brownish or reddish, juicy, and
commonly about 8 mm. in diameter. Berries with the seed more or less
exposed are very common. Submature fruits are usually ovoid. This
plant occurs in the Chisos, Chinati, and Davis Mountains of trans-Pecos
Texas. I believe the type of J. erythrocarpa, from the Chisos Mts., is a
form of this species collected late in the season, with the berries large,
perhaps because they are fully developed. Mr. Cory, who formerly treated
this species as J. gymnocarpa, is not satisfied that I am correct in identify-
ing it with the plants he described as J. erythrocarpa. If his suspicions are
correct our plant is without a name, for the name J. gymnocarpa (Lemmon)
Cory is based upon a phase of true J. monosperma from the Sandia Mts.,
New Mexico.

Juniperus monosperma (Engelm.) Sargent, Silva No. Am. 10: 89 (1896).

CoauurLa: Carneros Pass, Pringle 2305; Sierra Parras, Purpus 1104; Sierra Negras,
9 km. south of sda Stanford et al. 150. CuHutHuAHuA: Sierra Santa Eulalia,
Pringle 710. ZACATECA Near — de las Bocas, 20 mi. S.W. of Concepcion del
Oro, Shreve 9374; Cedros, Kirkwood 1

In its typical form J. eel ranges in New Mexico and into adjoin-
ing Arizona and Colorado. It has coppery bluish immature fruits, which

338 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

on maturity form a coppery-blue distinctly fleshy berry with a single plump

seed. The old dried fruits are raisin-like. They are usually glaucous and,
ea blue, have a reddish tone that is very conspicuous if comp ared
with the blackish blue mature berries of J. Asheit. The Mexican rater
I have cited is not typical J. monosperma and belongs to a group of un-
named forms, obviously closely related to J. monosperma, which occur in
western Oklahoma, trans-Pecos Texas, Sierra Madre of western Chihuahua
south into Durango, and in our area. The collections from Santa Eulalia
Mts., from near Parras, and from Picachos de las Bocas have glaucescent
foliage. The material from near Carneros has yellowish green non-glaucous
foliage very suggestive of J. erythrocarpa. Possibly it was subjected to
excessive heat in drying and the waxy bloom destroyed. Its fruit seems
to agree well enough with the other forms of J. monosperma that I have
here associated with it.

EPHEDRACEAE

Ephedra trifurea Torr. ex Wats. Bot. King Exped. 329 (1871).

VERNACULAR NAME: Hitomorial.

Coauvutta: Sierra del Carmen, Aug. 29, 1936, Marsh 694; gypsum ridge east of
Laguna Jaco, Johnston & Muller 1076; 21 mi. west of El Oro, road to Guimbalete,
shen 2010; Laguna del Rey, gypsum on _ Stewart 3015. CHIHUAHUA: 10 mi.
south of Ojinaga, clays, Johnston & Muller 13; wed of Chihuahua, 1908, Palmer
68, Pringle 868; 12 mi. south of Camargo, citi 233.

A bush 10-15 dm. tall with a loose eons habit, ranging from trans-
Pecos Texas to Arizona and south into adjoining Mexico. In Coahuila
the plant is frequent on gypsum and on gypseous clays. During my several
visits in Mexico, between July and October, I have never found a fruiting
plant. The plant is easily distinguished in our area by having its leaves
3 at a node, elongate, persistent, and becoming shredded in age, by its
coarse branches, and by its pungent terminal bud. This species and
E. Torreyana have the scales of the cones dry and papery.

Ephedra Torreyana Wats. Proc. Am. Acad. 14: 299 (1879).

CHIHUAHUA: 10 mi. south of Ojinaga, clays, a gray-green bush 1-2 ft. tall, Johnston
& Muller 14.

This species probably occurs in the valley of the Rio Grande between
Ojinaga and El Paso. It has been collected from western Texas to Arizona
and Utah, frequently in gypseous soils. It is a small bush with slender
widely divergent branchlets and short spreading ternate leaves.

Ephedra aspera Engelm. ex Wats. Proc. Am. Acad. 18: 157 (1883).
VERNACULAR NAMES: Popotillo; Cafiutilla; Pitamoreal; Hintimoreal.
CoaHurma: El Berrendo, White 1800; Puerto San Lazaro, Wynd & Mueller 143;

April 11, 1847, Gregg 414; Sierra Parras, March 1905, Purpus 1102; Sierras Negras,
9 km. south of Parras, Stanford et al. 168. CHiHuAHUA: Sierra San Carlos near the
mines, Johnston & Muller 60; Santa Eulalia plain, Wilkinson 118; crest of Santa

1943 J JOHNSTON, PLANTS OF NORTHERN MEXICO, I 339

Eulalia Mts., Pringle 38. Zacatecas: 15 km. west of Concepcion del Oro, Stanford
et al. 521; Cedros, Kirkwood 24, Kirkwood & Lloyd 86

This is the common Ephedra on rocky limestone slopes in Coahuila.
Commonly a bush 8-10 dm. tall, with leaves opposite and cone-scales thick
but not fleshy. The species ranges in southern New Mexico and trans-
Pecos Texas south on the plateau to San Luis Potosi.

Ephedra pedunculata Engelm. ex Wats. Proc. Am. Acad. 18: 157 (1881).
CULAR NAMES: Comida del Vibora; oe Tepopote; Canatilla; Sangui-
naria; Retamo Real; Hintimoreal; Itamoreal; Pit eal.

CoaHuU Villa Juarez, 1880, Palmer 1290; San de Jara, west of meee
Cienegas, pa 8837, 8841, 8847 ; Cuatro Ciensees Marsh 2057, Sierra Gavia, 5 mi.
north of Saucillo, Fohnston 7220; Saltillo, 1880, Palmer 1289; Saltillo, 1898, “Palmer
283; Valle de los Guajes, 8 km. east of Puerto del Aire, Stewart 1319; Santa Elena,

& Muller 310; La Botica, limestone slope, Stewart 2939. Curnuanua: Santa Eulalia
plain, Wilkinson 117 in pt.; Bachimba Canyon, Pringle 134. Zacatecas: Cedros,
Lloyd 75 and 214; near Concepcion del Oro, 1902, Palmer 372. Duranco: Between
Mapimi and Teviener, Apr. 18, 1867, Gregg 484 (Mo).

Usually scrambling up through bushes and frequently reaching a height
of 3 or 4 meters. Its branches cascade over the top of the supporting
vegetation and become very conspicuous when covered with an abundance
of red juicy fruits. At Ocampo it is very common in the hedge rows about
the town. The plant is not always lofty. In Canon de Jara, on very
arid cemented gravels, I observed plants of this species forming depressed
mats a meter broad and scarcely a decimeter high. Even with this unusual
habit the species was readily recognized by its pedunculate fleshy red cones.
Ephedra compacta Rose, Contr. U.S. Nat. Herb, 12: 261 (1909).

VERNACULAR NAME: Hitamo Real.

vegan Battlefield near Buena Vista, rocky soil, May 19, 1848, ier aa
Sierras Negras, 9 km. south of i Stanford et al. 1694; Sierra Jimulco, km
northeast of onc. Stanford et al. 3

A depressed spreading aon bush usually less than 5 dm. tall. The
species is known from the States of Coahuila, San Luis Potosi, Puebla,
and Oaxaca. It has opposite leaves and a sessile cone that produces two
seeds and becomes red and fleshy at maturity.

Ephedra antisyphilitica Berl. ex C. A. Mey. Mém. Acad. Sci. St. Pétersb. VI. Sci. Nat.
5: 291 (1846).

To be expected in northeastern Coahuila. Widely distributed in Texas
and collected near the Rio Grande at various points between Laredo and
the Big Bend. A bush becoming a meter tall, with opposite leaves and a
sessile, single-seeded fleshy red cone.

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

340 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXIV

THE COMPARATIVE MORPHOLOGY OF THE WINTERACEAE
I. POLLEN AND STAMENS

I. W. BaILEy AND CHARLOTTE G. Nast
With three plates

INTRODUCTION

THE genera Drimys, Pseudowintera, Belliolum, Bubbia, Exospermum,
Zygogynum, Tetracentron, and Trochodendron are the only known living
representatives of the dicotyledons that have retained a primitive vessel-
less type of secondary xylem. In fact, it was upon the basis of their ex-
ceptional wood structure that van Tieghem (6) segregated them in three
families of a distinct order, the Homoxylées. Tetracentron and Trochoden-
dron are monotypic, whereas the six winteraceous genera are now repre-
sented by approximately 88 species. Thus, the Winteraceae may no longer
be regarded as a few bizarre relics, since they are a flourishing family,
having representatives in Mexico, Central and South America, New Zea-
land, Australia, New Caledonia, New Guinea, the Solomons, the Philip-
pines, and adjacent regions.

There has been much uncertainty in the past concerning the relation-
ships and the morphological significance of the Winteraceae. Now that
much new material has been accumulated, it seems desirable to institute a
detailed re-investigation of the family. Our colleague, Dr. A. C. Smith
(4, 5), has published two extensive taxonomic revisions of the family. In
so doing, he assembled a large amount of material from American herbaria,
which has provided us, in turn, with a wide range of accurately determined
specimens, upon which to base a comprehensive study of the comparative
morphology of the flowers, leaves, and stems of the Winteraceae. The
first paper of our series deals with the pollen and stamens of the family.
The succeeding paper will discuss the remarkable carpels of the Win-
teraceae, which rival their vesselless wood in morphological significance.
The herbarium specimens upon which our studies are based have been
listed in Dr. Smith’s papers and need not be relisted here.

POLLEN

Wodehouse (7, 8) has advanced the intriguing hypothesis that there are
two basically different types of pollen morphology. The single-grooved or
monocolpate type (variously modified by phylogenetic changes) charac-
terizes the gymnosperms from the Bennettitales to the Coniferales and is
of common occurrence in monocotyledons, but is confined among dicoty-
ledons to certain representatives of the Ranales. On the contrary, the
3-grooved or tricolpate type (and its derived forms) is characteristic of
most dicotyledons. In recent comprehensive surveys of the pollen of

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, I 341

various dicotyledonous families, we have accumulated considerable evi-
dence in support of certain aspects of this hypothesis. It is evident, for
example, that plants of ranalian affinities may be divided upon the basis
of their pollen morphology into two distinct categories:

I. WITH MONOCOLPATE OR DERIVED TYPES OF POLLEN

Winteraceae Annonaceae Hernandiaceae Lactoridaceae
Degeneriaceae Myristicaceae Austrobaileya Piperaceae
Magnoliaceae Monimiaceae Canellaceae Saururaceae
Himantandraceae Lauraceae Calycanthaceae Cabomboideaze
Eupomatiaceae Gomortegaceae Chloranthaceae Nymphaeoideae

Il. WITH TRICOLPATE OR DERIVED TYPES OF POLLEN

Schizandraceae Trochodendraceae Ranunculaceae Menispermaceae
Tllictum Eupteleaceae Lardizabalaceae Nelumbonoideae
Tetracentron Berberidaceae

All of the plants in the first category, with the exception of the aquatic
Cabomboideae and Nymphaeoideae, have secretory cells of the well known
ranalian type. In the second category, such cells occur in the Schizan-
draceae, Jilicium, and Tetracentron only.

Wodehouse (7) homologizes the monocolpate pollen of angiosperms
with similar one-furrowed pollen of the lower gymnosperms. In so doing,
he assumes that the single germinal furrow is on the distal face of the
pollen grains, i.e. the exposed outer surface of the pollen grains during the
tetrad stage of development. The pollen of the Winteraceae is shed in
tetrahedral tetrads, Figs. 1-13, and each of the four grains is provided
with a circular germ pore in its distal (outer) face. Wodehouse inter-
prets the pollen of Drimys as of a modified monocolpate type, and infers
from this that the single germinal furrow of other ranalian plants is on
the distal face of the grains. It is significant in this connection, however,
that one of our co-workers, Mr. S. J. Golub, finds that the pollen of the
Annonaceae not infrequently tends to be more or less coherent in tetrag-
onal tetrads at the time of shedding. These tetrads, Fig. 14, demonstrate
that the germinal furrow of annonaceous pollen, Fig. 15, is on the proximal
(inner) face of the pollen grains; this raises the question whether such is
not likewise the case in the Myristicaceae, Magnoliaceae, and other related
ranalian families. The tetrads of the Winteraceae are firmly coherent,
and only in one instance (Drimys brasiliensis var. campestris (St. Hil.)
Miers, Hochne 28700) have we encountered a few dissociated tetrads. As
indicated in Fig. 13, the isolated pollen grains have, in addition to the dis-
tally located circular germ pore, proximal facets that closely resemble those
of annonaceous pollen, compare Figs. 13 and 15. In other words, it cannot
be assumed a priori that the distal germ pore of Drimys developed by a
simple reduction in the size of the germinal furrow of Magnoliaceae, since
the germinal furrow of annonaceous pollen is not a direct homologue of
the distally oriented circular germ pore of winteraceous pollen, but rather
of its unthickened proximal facets, compare Figs. 12, 13 with Figs. 14, 15.

The morphological specializations of monocolpate types of pollen are
numerous and diverse, apparently leading to the formation of dicolpate

342 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

(Calycanthaceae, Monimiaceae, monocotyledons), belted (Monimiaceae,
Eupomatiaceae, Nymphaeaceae, monocotyledons), polycolpate (Chloran-
thaceae, monocotyledons), and acolpate (Chloranthaceae, Monimiaceae,
Lauraceae, Hernandiaceae, monocotyledons) types. Such phylogenetic
trends in the specialization of angiospermic pollen can be clarified only by
intensive investigations of a wide range of Ranales and monocotyledons.
Furthermore, it is essential to determine the effects of contacts not only
within tetrads, but also within groups of contiguous tetrads.

In the case of the Winteraceae, the problems of phylogeny are compli-
cated by the fact that the pollen of all investigated genera and species is
shed in permanent tetrads. Since the structure of the individual grains of
such tetrads frequently is much modified through excessive specialization,
there are no reliable means of determining what the morphological charac-
teristics of the ancestral free grains may have been. Nor is it possible to
solve the difficulty by comparisons with the permanent tetrads of Lac-
toridaceae, Hedycarya (Monimiaceae), or Victoria (Nymphaeaceae ),
for in these tetrads specialization has progressed along different lines.
The tetrads of the Winteraceae are morphologically unique among plants
of ranalian affinities and cannot justifiably be cited as evidence of closer
relationship to the Magnoliaceae than to other ranalian families.

e most comprehensive previous investigation of the pollen of the
Winteraceae is that of van Tieghem (6), who studied representatives of
all six genera of the family. Although providing no illustrations or de-
tailed. descriptions of the pollen, he noted certain significant morphological
differences within the family. Thus, he emphasized the fact that the
pollen of Drimys Winteri, D. brasiliensis, and D. granadensis forms pro-
tuberant papillae when moistened, whereas that of D. membranea, D. piper-
ita, Pseudowintera,' Bubbia, Belliolum, Exospermum, and Zygogynum
does not. In addition, he states that the pollen of Belliolum, Exospermum,
and Zygogynum has a granular exine, whereas that of Drimys, Pseudv-
wintera, and Bubbia exhibits a verrucose sculpture.

Our own investigations of numerous species of the genus Drimys indicate
that in general the tetrads of the Old World Section (Tasmannia), Figs.
2-5, are conspicuously smaller than the tetrads of the New World Section
(Wintera) of the genus, Fig. 1. In the case of Sect. Tasmannia, Figs, 2,
3, and 5, as in Pseudowintera, Bubbia, Belliolum, Exospermum, and
Zygogynum, the entire floor of the circular pore bulges outward more or
less uniformly during re-expansion of the pollen, whereas in Sect. Wintera
the central part of this floor bulges rapidly and precociously, leaving a con-
Sstricting rim or collar of presumably thicker or less elastic material,
Figs. 1, 12, and 13. The individual pollen grains of Drimys, Figs. 1-5, 9,
II, and 13, Pseudowintera, Fig. 8, and Bubbia, Fig. 7, are provided (be-
tween their distal germ pore and their adnate proximal facets) with a
broad zone or belt, Fig. 13, of coarsely reticulate thickening, whereas
those of Exospermum and Zygogynum, Figs. 6 and 10, have minutely re-

1Pseudowintera Dandy, i. e. Wintera sensu v. Tiegh., non Murray.

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, |} 343

ticulate thickening. The pollen of Belliolum crassifolium ( Baill.) v. Tiegh.,
B. haplopus (Burtt) A. C. Sm., Fig. 11, and B. Burttianum A. C. Sm. is
coarsely reticulate and resembles that of the former category of genera.
Therefore, we are unable to follow van Tieghem (6) in describing the
pollen of Belliolum as granular rather than as verrucose. Since he pro-
vides no illustrations or detailed descriptions, it is difficult to determine
just what he had in mind in using the unqualified general terms ‘‘verru-
queuse” and ‘‘granuleuse”. The more or less conspicuously buttressed
reticulate thickenings, Figs. 1-5, 7-9, 11, and 13, of the tetrads of Drimys,
Pseudowintera, Bubbia, and Belliolum are composed of rows of more or
less coalesced rods, appearing linear or granular at different focal levels,
Fig. 16. Furthermore, the reticulate thickening appears more or less gran-
ie in surface view, depending upon the degree of coalescence in its con-
stitutent rods and upon the amount of buttressing. In certain species of
Bubbia, e.g. B. Clemensiae A. C. Sm., B. longifolia A. C.Sm., and B. mono-
carpa A. C. Sm., Fig. 16, the pollen has a finer mesh of more numerous,
ee partly coalesced rods. The reticulate thickenings of such tetrads
are conspicuously granular appearing at lower focal levels. Similarly, in
the case of Exospermum and Zygogynum, the exine appears finely re-
ticulate, Fig. 10, or minutely granular, Fig. 6, at different focal levels.
There is, however, in the material that we have studied, a wide structural
gap between the minutely granular-reticulate exines of Exospermum and
Zygogynum and the coarsely granular-reticulate exines of the other four
genera of the Winteraceae.

There are families of dicotyledons in which the pollen is of very con-
siderable taxonomic significance, not only in the differentiation of sub-
families and tribes, but also of genera and species. Our investigations of
numerous collections of all four species of the New World Section (Win-
tera) of Drimys and of ten of the thirteen taxonomic entities recognized by
Smith (4) indicate that, although the pollen of these plants may be
easily differentiated from that of other representatives of the family, it is
difficult to distinguish species and varieties within the Section Wintera.
In the case of the Old World Section (Tasmannia) of Drimys, Pseudowin-
tera, Belliolum, and Bubbia, the size and form of the tetrads, Figs. 2—5,
the diameter of the germ pore, Figs. 9 and 11, the distribution of wart-like
thickenings on the floor of the germ pore, Figs. 9, 11, and 16, the detailed
structure of the reticulate thickenings, Figs. 9, 11, and 16, and other
characters fluctuate more or less from species to species. More material
must be examined, however, before attempting to construct a key for dif-
ferentiating species and genera. Our investigations do suggest that the
pollen of Sect. Wintera of Drimys, on the one hand, and of Exospermum
and Zygogynum, on the other hand, represent two stabilized specializations
from the more varied and generalized types of pollen encountered in Bellio-
lum, Bubbia, Pseudowintera, and Sect. Tasmannia of Drimys.

STAMENS
In the Degeneriaceae (Bailey and Smith 1) and the Himantandraceae

344 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

(Diels 3, Bailey, Nast, and Smith 2), the stamens are not differentiated
into filament, anther, and connective, but are essentially 3-veined micro-
sporophylls of comparatively unmodified form. The two pairs of slender,
vertically elongated sporangia are immersed beneath the abaxial surface
of the sporophylls. They are situated between the median and lateral
veins, and neither these veins nor their branches are directed toward the
sporangia, ‘Thus, the narrow shields of endothecia are not in contact with
vascular tissue. Dehiscence is longitudinal and extrorse. The closely
allied Magnoliaceae exhibit various modifications of such primitive 3-
veined microsporophylls, leading to the formation of more typical stamineal
organs. The much enlarged, conspicuously protuberant, paired sporangia
(thecae) are lateral on a much constricted part (connective) above the
broad base (filament) of the microsporophyll. The thecae may be turned
slightly outward or inward, and the longitudinal dehiscence, therefore,
fluctuates between extrorse and introrse. In certain representatives of
the family, the lateral veins of the modified microsporophylls are much
reduced and may at times be eliminated, leaving a single-veined stamen,
such as characterizes so many of the dicotyledons,

Although the stamens of the Winteraceae vary considerably in external
form, Figs. 17-24, they are throughout the family of the single, dorsally-
veined type. Four of the genera, Bubbia, Fig. 22, Pseudowintera, Fig. 23,
Exospermum, Fig. 24, and Zygogynum have short, comparatively broad,
more or less truncated and apically flaring microsporophylls. The pro-
tuberant sporangia are attached to the broad apex of these sporophylls and
are oriented either at right angles to the dorsal vein, Figs. 22 and 23, or
in various diagonal positions, Fig. 24. Dehiscence is, therefore, apical
and transverse or obliquely apical. On the contrary, the stamens of Belli-
olum, Fig. 21, are generically characterized by their more elongate form
and particularly by having their laterally attached sporangia oriented
parallel to the long axis of the sporophyll. Dehiscence is longitudinal and
more or less conspicuously extrorse.

The microsporophylls of these five genera of the Winteraceae have no
constricted part that may be designated as connective, and the sporangia
are not excessively protuberant beyond the outlines of the sporophylls.
The stamen illustrated in Fig. 21 resembles in form the microsporophylls
of Himantandra and Austrobaileya. It differs from those of the former
genus in its more protuberant sporangia and in having no lateral veins,
from those of the latter genus in its unbranched median vein, which does
not extend beyond the level of the sporangia. In such species of Belliolum
as B, crassifolium (Baill.) v. Tiegh. and B. Burttianum A. C. Sm. the un-
vascularized upper part of the sporophyll is considerably reduced in length.
This suggests that the types of stamens illustrated in Figs. 22-24 may
have developed phylogenetically by elimination of this part of the micro-
sporophyll, with concomitant shifting of the sporangia from longitudinal
lateral orientations to obliquely apical and transversely apical ones.

The microsporophylls of Drimys, Figs. 17-20, fluctuate considerably in
length, not only in different species, but also within the same flower. They

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, I 345

are characterized, however, by having markedly protuberant thecae that
are attached to the much constricted upper part of the sporophyll. The
subapical thecae are oriented parallel to the long axis of the sporophyll or
at acute angles to it. Dehiscence is, therefore, approximately longitudinal-
lateral and more or less conspicuously extrorse. The microsporophylls of the
New World Section (Wintera) of the genus are relatively broad, Figs.
17-19, but those of the Old World Section (Tasmannia) occasionally are
much elongated and narrow, Fig. 20. The latter are more typically stamin-
eal, Ni exhibiting differentiation into filament, connective, and
anthe

The median (dorsal) vein of winteraceous stamens may extend through-
out the sporophyll without branching, Figs. 18, 21, 22, and 24, or it may
give off one, Fig. 19, or more, Figs. 17, 20, and 23, chart branches that are
directed toward the thecae. Branching of the dorsal vein is in general more
extensive and conspicuous and occurs at a lower level in Pseudowintera and
certain species of Bubbia, e.g. B. pachyantha A. C. Sm., than in other repre-
sentatives of the Winteraceae. Spherical secretory cells, Fig. 18, are of
common occurrence in the microsporophyll, as in other organs, of the
Winteraceae. Their contents usually are dissolved during the clearing and
mounting of the stamens and, therefore, are invisible in most figures of
Plate III. The coriaceous floral appendages of certain species of Bubbia,
Exosperum, and Zygogynum contain very numerous sclereids or clusters of
sclereids. As indicated in Fig. 24, the stamens of such flowers may contain
more or less numerous sclerenchymatous cells.

It should be emphasized in conclusion that there appear to be two distinct
trends of specialization in the microsporophylls of the Winteraceae, leading
in Pseudowintera, Bubbia, Exospermum, and Zygogynum to the formation
of broadly truncated sporophylls bearing transversely oriented apical
sporangia, and in Drimys to apically constricted sporophylls bearing later-
ally attached subapical sporangia.

LITERATURE CITED

—

. Barrey, I. W., . C. SmirH. Degeneriaceae, a new gaa of flowering plants
from fees jou. eee Arb. 23: 356-365. pl. 1-5.
—_—— Nast, and A. C. Smitn. The i ied tencieanas Jour.
Arnold rea ov 190-206. pl. 1-6. 1943.

3 Ls, L eber oe cee haar, ihre Verbreitung und ihre systematische
Stellung. Bot. Jahrb. 55: el O177

4. SmitrH, A.C. The aa. species of - Jour. Arnold Arb. 24: 1-33. f. 1-3.
1943

————.. Taxonomic ne on the Old World species of Winteraceae. Jour. Arnold

Arb. 24: 119-164. f. 1-6. 1943.

6. TIEGHEM, P. van. Sur ee dicotylédones du groupe des Homoxylées. Jour. de
Bot. 14: 259-297, 330-361. 1900.

7. WopenouseE, R. P. Pollen grains. McGraw-Hill, N. Y. 1935.

8 Evolution of pollen grains. Bot. Review 2: 67-84. f. 1-8. 1936.

346 JOURNAL OF THE ARNOLD ARBORETUM (VoL. XXIV
EXPLANATION OF PLATES

Pirate I

Unstained tetrads mounted in lactic acid and photographed at a magnification of
1180. 1c. 1. Drimys brasiliensis var. campestris (St. Hil.) Miers, Hoehne 28700.
Fic. 2. Drimys Beccariana Gibbs, Brass 11294. Fic. 3. Drimys oligandra A. C. Sm.,
Brass 12975. Fic. 4. Drimys sige Hook. f., Williams 754. Fic. 5. Drimys obovata
A. C. Sm., Brass 11295. Fic. é: Zovoesnum Vieillardi Baill., Fra nc 1740. Fic. 7.
Bubbia megacorpa A .C, Sm., Bhs 10249. Fic. 8. per rare axillaris var. colorata
(Raoul) A. C. Sm., Rival in 1843.

onl

Piate IT

rawn with camera lucida from pollen mounted unstained in lactic acid. Fic
Drimys lnceolat (Poir.) Baill., Baker in 1890. Tetrad showing detail of one secre
grain. x 1180. Fic. 10. Zygogynum Vieillardi Baill., Franc 1740. Drawn at a
higher focal level than Fig. 6, showing finely reticulate exine. X 1180. Fic. 11.

Belliolum haplopus (Burtt) A. C. Sm., Kajewski 1994. ™% 1180. Fic. 12. Drimys
brasiliensis var. campestris (St. ) Miers. Hoehne 5 Outline of tetrad showing
protuberances. x 617. Fic. The same. Single grain from dissociated tetrad,

showing apeia i from aa germ pore and the structure of proximal facet.

. 14. Asimina pirate a A. Gray, a 1143. Outline of tetragonal
tetrad, ara proximal p on of germinal furrows. xX 187. Fic. 15. ame
Detail of single grain rtd pene tetrad, eae proximal furrow, xX 480.
Fic. 16. Bubbia monocarpa A. C. Sm., Kanehira & Hatusima 12105. Detail of one
grain of tetrad; (a) granular appearance at lower focal level. Xx 1180

Priate III

Stamens cleared in hot dilute NaOH and mounted unstained in diaphane. Magni-
fication * 24. Fic. 17. Drimys So goon Bock 49. Fic. 18. Drimys grana-
densis var. grandiflora Hieron., Arc 1202, Fic. 19. Drimys Winteri var. chilensis
(DC.) A. Gray, Werdermann 73. Fic. 20. Drimys stipitata Vickery, White 7572.
Fic. 21. Belliolum haplopus (Burtt) A. C “Sa. Brass 2959, Fic. 22. Bubbia Clemensiae
A. C. Sm., Clemens 4596. Fic. 23. Picidowitee axillaris var. oe C. Sm., Kirk.
Fic. 24. Exostermun stipitatum (Baill.) v. Tiegh., Vieidlard 2281

BIOLOGICAL LABORATORIES,
HARVARD UNIVERSITY.

Jour. ArNotp ARB. Vor. XXIV PLATE I

COMPARATIVE MORPHOLOGY OF THE WINTERACEAE

PLaTE II

Jour. Arnotp Ars. VoL. XXIV

15

Recs
3
()

8%
20
@

CE
j
¢

OM
oe
o
ee
Oe
eee
vd

0 o 4 we

fs °.
esate gore? Me eeeceeseeneeneeeettt
<ecelSecae?

“eee

°
{ S
ry
Nig;

Cy
>
a

ee wid
te eeeneee®

INTERACEAE

y

THE W

COMPARATIVE MORPHOLOGY OF

Prare III

ARNOLD ARB. VoL. XXIV

JR.

Jou

—

Ty keephoas Sys

Sen oe

—

ie Ones

tS

WINTERACEAE

THE

COMPARATIVE MORPHOLOGY OF

|
| :
.

|

; .
| 7
. :

1943] SMITH, PACIFIC ISLAND PLANTS, II 347

STUDIES OF PACIFIC ISLAND PLANTS, IP!
NOTES ON THE PACIFIC SPECIES OF PIPER

A. C. SMITH

In attempting to name a series of specimens of Piper L. from Fiji, kindly
forwarded by the Curator of the Bernice P. Bishop Museum, it was found
desirable to prepare a consideration of the known species of the genus in
Fiji. In the course of this study, the species from adjacent Pacific groups
were examined, and notes on two of the difficult complexes are here included.
A revision of all the Pacific species is highly desirable, but this can scarcely
be undertaken without examination of extensive collections and of types
deposited in European herbaria. I am indebted to the authorities of the
following institutions for the privilege of studying herbarium material, the
place of deposit being indicated by the parenthetical letters: Arnold Arbore-
tum (A), Bishop Museum (Bish), Gray Herbarium (GH), New York
Botanical Garden (NY), University of California (UC), U. S. National
Herbarium (US).

PIPER PUBERULUM (BENTH.) BENTH. AND ITS VARIETIES

The most common species of § Macropiper in Fiji, Samoa, and Tonga is
the shrub with the following essential characters: petioles 1-4 cm. long,
vaginate from one-half to nearly their entire length; leaf-blades ovate, of
moderate size, generally up to 15 by 10 cm. but sometimes up to 22 by
17 cm., either puberulent beneath or glabrous on both surfaces, obtuse to
rounded or subcordate at base, and with 5 or 7 (sometimes 9) nerves diver-
gent from the petiole; spikes axillary, usually solitary but sometimes (espe-
cially in distal axils) paired, comparatively long, usually 7-17 cm. long at
anthesis excluding peduncle (both staminate and pistillate), rarely 4-19 cm.
long (scarcely mature when less than 7 em.)

This is the plant commonly passing in herbaria and literature as P. Mac-
gdlivrayi C. DC. An examination of the various treatments of this species
demonstrates that de Candolle’s binomial must be replaced by P. puberylum
(Benth.) Benth. ex Seem.

The first description referable to this species was that of Bentham in
1843, of Macropiper puberulum, based on a Fijian collection of Hinds and
Barclay. De Candolle, in his later considerations of the species, obviously
should have made use of this specific epithet, but instead he proposed the
name Piper Macgillivrayi, which has been associated with the species
through practically every consideration up to the present. That de Can-
dolle was aware of Bentham’s earlier name is obvious from his citation
of Macropiper puberulum in synonymy, both in Seemann’s Flora Vitiensis
(1868) and in the Prodromus (1869); the manuscript for Seemann’s work

1See Bull. Torrey Bot. Club 68: 397-406. 1941.

348 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

may possibly not have been seen by de Candolle before publication. In
this work we find the binomial Piper puberulum Benth. occurring twice, once
parenthetically on page 262 in the “explanation of plate 75,” and again on
the plate itself. This mention of Piper puberulum, since it is accompanied
by a detailed plate, is here accepted as an authentic transfer of Bentham’s
Macropiper puberulum. 1 have no doubt that Seemann used the binomial
Piper puberulum in good faith on his plate, but was deterred from taking it
up in his text because of his discovery of de Candolle’s manuscript name.”

Numerous varieties have been proposed within the comprehensive con-
cept of P. Macgillivrayi, based upon Fijian and Samoan specimens. For
the most part these varieties appear properly placed, but I believe that var.
fasciculare Warb. (10:609) is best removed from the species to P. Timothi-
anum, as stated below in my detailed consideration of the Fijian species.
Var. glabrum Warb. (10:609) appears to be both a homonym and a syno-
nym of var. glabrum C. DC.; cotype material of Warburg’s variety (Rein-
ecke 75 [US|) does not differ from the glabrous Fijian form upon which de
Candolle’s varietal name is based. I am unable to pass upon the proper
position of the following Samoan varieties: abbreviatum Warb., scandens
Warb. (for these see 10: 609), subrotundifolium C. DC. (4: 264), and
upoluanum C. DC. (nomen?, see 5: 258). Students of the Samoan flora
should consider whether these are worth retaining as trinomials under
P. puberulum.

The following varieties of P. Macgillivrayi have been based upon Fijian
collections: parvifolium C. DC. and glabrum C. DC. (2: 335). The first of
these probably represents a depauperate form of the common glabrous-
leaved variety of P. puberulum, but the second is difficult to interpret. De
Candolle’s description of var. glabrum reads merely: ‘‘foliis utrinque
glabris,” but the only specimen he cites is “Seemann 567 ! in h. DC.”
Macropiper puberulum Seem. in Bonplandia 1861, p. 259 is cited as a
synonym. On the basis of Seemann 567 in the Kew Herbarium, Bentham’s
description of Macropiper puberulum, and Seemann’s description and plate
in Flora Vitiensis (1868), one might suppose that var. glabrum is actually
typified by the puberulent-leaved form. However, it is obvious from the
varietal name and from the phrase “foliis utrinque glabris” that de Can-

2Qne might suspect that de Candolle avoided Bentham’s specific epithet because of
an earlier Piper puberulum, but I cannot find that this is the case. No such bi-

1861. Because Maximowicz’ binomial in 1886 is a later homonym of P. puberulum
Benth. ex Seem. (1868), the Hongkong plant should be known as Piper hongkongense
C. DC. (2: 347). This latter binomial is accounted for in de Candolle’s key (S: 201);

tions were deposited in Seemann’s herbarium, it seems very likely that Piper hong-
kongense and Chavica puberula are based upon parts of the same collection.

1943 | SMITH, PACIFIC ISLAND PLANTS, II 349

dolle did not intend this interpretation. The specimen of Seemann 567 in
the Gray Herbarium is a mixture, having large puberulent leaves from one
plant and smaller glabrous leaves from another. I believe, therefore, that
de Candolle’s specimen of this collection was the glabrous form, and
propose to interpret var. glabrum according to his obvious intent and his
description, that is, excluding the synonym Macropiper puberulum and the
puberulent-leaved portion of Seemann 567.

In Fiji, P. puberulum is divisible into two varieties, which I designate
below as var. typicum and var. glabrum.

PIPER LATIFOLIUM L. F. AND ITS ALLIES

The nomenclatural confusion which has been attached to this binomial is
due to the fact that it first appeared in the Emendanda to the younger
Linnaeus’ Supplementum Plantarum (1781) and_ thus replaced that
author’s Piper methysticum as described on page 91 of the same work.
One is therefore justified in considering P. methysticum L. f. as a name
published in synonymy and in taking P. latifolium L. f. as the correct name
for the Tahitian plant described on page 91 of Linnaeus’ work. This plant
is characterized by its several axillary spikes and cannot be confused with
the widely cultivated “kava” or “yanggona,” which was first botanically
described by G. Forster (Pl. Esc. Ins. Oc. Austr. 76. 1786) as Piper methy-
sticum. It does not appear necessary to take Forster’s name as a later
homonym of P. methysticum L. {., which, having been corrected by the
author in the same original work, has no nomenclatural status. Practically
all modern taxonomists who have considered the matter are in agreement
with Moore (6), whose lucid discussion of the problem indicates that Piper
methysticum Forst. f. is the correct name for the common cultivated
“kava.”

In a consideration of the Pacific species of Piper, one of the most diffi-
cult problems is to fix the geographic limits of P. latifolium. Although the
species was originally based on a single collection from Tahiti, numerous
writers have taken the species to include plants from as far west as Tonga
and the New Hebrides. This extension of the range was probably first
indicated by G. Forster (FI. Ins. Austr. Prodr. 5. 1786). C. de Candolle,
in 1869 (2: 335) notes the range as “in Ins. Tahiti, ins. Societatis, Ami-
corum, Novarum Hebridum, Timor,” but subsequently (5: 172) there is an
indication of uncertainty, as he states the range merely as “Tahiti, etc.”

The only Tahitian specimens I have seen which match the original
description and de Candolle’s interpretation in his key (5: 172) are U. S.
Expl. Exped. 3, in part (GH) and Setchell & Parks 274 (UC), the latter
being cited as such by Setchell (9: 163). Setchell implies that the species is
endemic to Tahiti. However, I believe that F. Brown’s reference of
Marquesan specimens to P. latifolium (1: 17) is correct; the several speci-
mens which Brown cites from the Austral Islands are not now available to
me, and they may possibly be similar to the Raratongan plants discussed
below.

Another species which must be considered in connection with P. Jati-

350 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

folium is P. tristachyon C, DC. (2: 335), at least as regards its Tahitian
components, the species having been based on material from both Tahiti
and the Hawaiian group. According to de Candolle’s key (5), P. trista-
chyon differs from P. latifolium primarily in its leaf-blades being pubescent
beneath. The Tahitian form of P. tristachyon appears to be represented
by Setchell & Parks 341 (UC) (see Setchell, 9: 163), Tilden 429 (GH),
and U. S. Expl. Exped. 1 § (GH). While P. tristachyon is superficially
distinct from P. latifolium on the basis of its crispate-pilose lower leaf-blades
and petioles, it is perilously similar in its other characters, such as leaf-shape
and petioles almost completely vaginate. In general, its leaf-blades are
slightly broader in proportion than those of P. latifolium and with less pro-
nounced apices. The probability that these two entities are not specifically
distinct is strengthened by the occurrence in Tahiti of a plant precisely re-
sembling P. tristachyon in all details except its completely glabrous habit.
This is represented by U. S. Expl. Exped. 1 @ (GH, US), and Setchell &
Parks 212 (UC) and 340 (UC). The latter two specimens have been re-
ferred, and probably correctly, by Setchell (9: 163) to P. excelsum var.
tahitianum C. DC. This variety appears not to be conspecific with P. excel-
sum Forst. £. of New Zealand, which has quite different leaves and has fruits
immersed in the rachis (C. DC., 5: 171). I believe that P. excelsum var.
tahitianum should definitely be separated from the New Zealand species and
placed in the synonymy of P. tristachyon, of which it is no more than a gla-
brous form. One may consider the advisability of referring all the Tahitian
plants here discussed to P. latifolium, which would then be characterized
chiefly by its long and nearly completely vaginate petioles, broad many-
nerved leaves, and several (3-5) axillary spikes.

The three Tahitian entities here discussed (P. latifolium, P. tristachyon,
and P. excelsum var. tahitianum), on the basis of material now available
and without consultation of the types, are kept apart only with difficulty.
An essentially similar conclusion has already been expressed by Nadeaud
(Enum. Pl. Indig. Tahiti 41. 1873).

The occurrence of P. latifolium on Raratonga in the Cook Islands has
been noted by Cheeseman (in Trans. Linn. Soc. I. Bot. 6: 293. 1903) and
Wilder (in Bishop Mus. Bull. 86: 38. 1931). The plant observed by them
is apparently common on Raratonga and is represented by: H. E. & S. T.
Parks 22036 (GH, UC, US), 22211 (UC), and 22301 (A, UC, US), and
Wilder 3 (A, NY, UC). Superficially it differs from the Tahitian forms of
P. latifolium in having its petioles only one-third to one-half vaginate; its
staminate flowers have consistently 3 or 4 stamens, while those of the
Tahitian specimens have either 2 or 3 stamens. For the present I have not
seen enough material to decide whether a reasonable concept of P. lati-
folium may be extended to include this Raratongan form or whether the
latter should be separated as a subspecific unit.

The occurrence of P. latifolium in the New Hebrides, noted by Forster in
1786, has been further recorded by several writers, including Miquel (Syst.
Piper. 219. 1843), whose concept of the species included even such forms
as P. guahamense C. DC. (2: 336). Guillaumin has more recently men-

1943 ] SMITH, PACIFIC ISLAND PLANTS, II 351

tioned P. haan from the New Hebrides (in Bull. Soc. Bot. Fr. 66:
275. 1919, op. cit. 74: 703. 1927, in Jour. Arnold Arb. 13: 82. 1932).
Among fhe eer upon which Guillaumin’s concept was based, Kajew-
ski 3, 436, and 718 (all A) are available to me, and the most detailed
examination fails to reveal any consequential characters by which this New
Hebrides plant can be distinguished from a concept of P. latifolium which
includes the various Tahitian and Raratongan forms discussed above.

While my conclusions are admittedly based upon insufficient material,
for the time being I follow earlier students in thus accrediting P. latifolium
with a range extending from the Marquesas to the New Hebrides, although
its absence from Fiji, among the collections now available from that group,
must remain surprising. One may anticipate that examination of abundant
Pacific material and comparison with the historic collections will reveal
lines upon which this present concept of P. /atifolium may be intelligibly
ivided.

Tue FIJIAN SPECIES OF PIPER

The only previous treatment of the Fijian species of Piper is that of de
Candolle in 1909 (3). In this paper five new Fijian species were described,
two of which I cannot accurately place, the types not being available. I
am able to recognize ten indigenous Fijian species, the total thus being
probably twelve. An additional three species, which occur in Fiji either in
cultivation or as weeds, are included in this consideration, being the first
three species in my key. The indigenous Fijian species fall into the Sections
Eupiper and Macropiper, as outlined by de Candolle (5). In the present
treatment I describe three new species, propose a new name for P. polysta-
chyum C. DC., and take up P. puberulum as an older name than the well-
known P. Macgillivrayi C. DC
Spikes leaf-opposed, solitary; leaf-blades plinerved or pinnate-nerved, at least the

inner nerves partially pacients: stipules sometimes free and then the petioles
unwinged.
Inflorescence-scales copiously pilose; leaf-blades narrowly oblong-elliptic, mmequi-
rally rotund-subcordate at base, scabrid above, puberulent beneath (at least
on nerves), pinnate-nerved, the lateral nerves 4-7 per side, ascending; weed,
Native Lo 7Amenica. (SSL fens1) os ico 54% oes ose wie ono rome 1. P. aduncum.
Inflorescence-scales glabrous; leaf-blades smooth above, not scabrid, papi pie
incipal nerves ve nieve for less than half the length of the costa (pinnate-
nerved in no. 7) (§ Eupiper
Erect shrub; ae blades large, 136 25 x 10-20 cm. at maturity, santas cordate at
base, minutely puberulent on nerves beneath, the principal nerves 9-13, freely
spreading from petiole except the 3 innermost, these loosely aan for
5-15 soe mature spikes (excl. peduncle) 2-6 cm. long; in Fiji only in
cultivation 202m
Scandent ohne. (except no. 7); leaf-blades not exceeding 14 x 10 cm. (rarely
t

—
\)
ct
oOo
—

co
>5
oO
=
ee
=]
2
U
Pp
[=]
oO
Lan 1
<
oO
wn
un
|
+
a
oO
5
=)
om
~~
io)
=)
oO
n
=>
ia}
3.
ta
w)
°
=]
ei
=
ia}
—
oO
3
o
o
°
=
pS
i

Spikes at pie 2 cm. long and usually much lane on peduncles at ae 6 mm.
long; leaf-blades plinerved; scandent plan
Fruits costing fully embedded in pulp oe concrescent with He rachis;
filaments as broad as anthers; in Fiji only in cultivation...... . P. Betle.
Fruits no more than semi-immersed in the se
much narrower than anthers; indigenous in Fij

5, not a filaments

352 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

Stipules free, inconspicuous, 5-12 mm. long; leaf- —o with the inner nerves
usually long-concurrent, the costa with obvio peoeetiog nerves, i vein-
lets usually prominulous on both surfac ova and fruits semi-
immersed in the rachis; stamens pane 4, sre : per Fake the
meas small, 0.15-0.25 x 0.3-0.4 mm. (42 inflorescence not known

Plant nee throughout, or inflorescence-rachis sparsely pilose.......
SN ee ee nee 4. P. insectifugum.
Branchlets distally, petioles, peduncles, and leaf-blades on both surfaces
crispate-hispid, the upper ne surface eventually se apiloenens
inflorescence- ieee densely pilose................4. 5. P. crispatum

— free or adnate to petiole, ae ae , 15-22 mm. long; leat. blades
early plinerved, oe inner 3 nerves jc hvunenk for only 7-20 mm., the

ar without important lateral nerves, the veinlets obscure or slightly
impressed above; ovaries apparently frees stamens apparently 2 per

flower, the anthers comparatively large, about 0.25 x 0.7 mm.......

5 tea aN Goeas's wine tara whe ai oS i BAG ap iw Seal eal ach tana eG oie ea P.s

— ae least 2 ) 1-1.5 cm. long, on peduncles 2-4 mm. long; ovaries free, not
sed in the rachis; leaf-blades essentially pinnate-veined, with 2—4 pairs
eT SBCUD:. fo 454e ens eee acasawdeh au saaae ae generi.

one axillary: leaf- blades with nerves freely spreading from the petiole; stipules
ate to petioles, these at least partially vaginate; indigenous in Fiji (§ Macro-

fi per).
Spikes solitary, sometimes paired in distal leaf-axils.

‘Leaf-blades hispid-pilose on both surfaces; inflorescence-scales ciliate-setose; car-
pels conspicuously narrowed toward apex and with obscure stigmas..........
pind Niet a We weak SS a wa Ea kee wee Gee a ewes 8. P. oxycarpum.

— eos glabrous at least above; inflorescence-scales not ciliate-setose; carpels

sentially rounded at apex, Bee obvious stigmas.
peduncle, rarely only 4 cm. but vie aaees
mature; floral parts meen: large, the scales ae mm, in diamet
leaf-blades (7—)8-15(—22) by (3—)4-10(-17) cm., 5—9-nerved.
Leaf-blades puberulent beneath.............. 9a. P. pub sa i var. hate
Leaf-blades glabrous on both surfaces........ 9b. P. puberulum var. glabru
Spikes 2.5-5.5 cm. long at maturity, excluding peduncle; floral parts smaller, a.
scales 0.35-0.7 mm. in diameter; leaf-blades 6-11 - 2-5.5 cm., 3- or 5-
d . P. melanostachyum.
Spikes 3 or more per leaf-axil, rarely only 2
Petioles of mature leaves 8-17 cm. long, usually vaginate only in the lower quarter;
leaf-blades up to 25 b ae cm., 11-13-nerved, cordate at base veep so on
larger blades) ; — (at least ¢ ) 10-22 per leaf-axil.......... 11. P. vitiense.
Petioles shorter, 1-4.5 c
least more than half their length; leaf-blades up to 17 by 15 cm., (5—)7—9-
nerved, obtuse to subcordate at base; spikes (both ¢@ and 9) usually 3-5
per leaf-axil, fae
Spikes (both and 2) 0.8-3. 5 cm. long (excl. peduncle) ; stigmas glabrous or
sparsely a ee ee eer ra erie 12. P. hianum.
Spikes (at least 2) 5-8 cm. long (excl. peduncle) ; a densely and obvi-
lOUSIY: PHOSEse-.5.o:5-tse se Seis Hesse Ruka eles eee as oi . P. kandavuense.
1. go shee ng: aduncum L. Sp. Pl. 29. 1753; C. DC. in ig Prodr, 16(1):
9; B. E. Parham in Agr. Jour. Dept. Agr. Fiji 9(3): 1938
DistRIBUTION: Common throughout a large part of tropical a Of recent
introduction into Fiji; no to Parham, in 1938: “During the past five years this
species of Piper has been noted as an aggressive weed plant in the south-eastern part of
Viti Levu, it has apparently spread rapidly with Suva as the cent
Fiyt. Vitt Levu: Rewa: Lami, Parks 20060 (Bish); between Suva and Lami,

1943 ] SMITH, PACIFIC ISLAND PLANTS, II 353

Gilles pie 2079 (A, Bish); 4 mi. west of Suva, sien be (Bish); Rewa or
Kombalevu, alt. 400 m., Paks 2 (Bish

NATIVE NAME: Yanggona ni Onolulu Biceetite to ‘Parham: e. “Honolulu Piper,”
a misnomer, as the species is probably not found in Hawaii).

As it occurs in Fiji, the plant is a shrub or ainda tree up to 8 m. high,
occurring on roadsides or in bush-land, most often near cultivation. It has
not been reported from any other Pacific group.

2. Piper (§ Eupiper) methysticum Forst. f. Pl. Esc. Ins. Oc. eas 76. 1786, FI.
Austr. Prodr. 5. 1786; Seem. Fl. Vit. 260. 1868; C. DC. DC. Prodr.
16(1): 354. 1869, in Candollea 1: 1923; non L. f. Suppl. 91 i synonym of

L. latifolium L. f. in Emendanda]. 1781.
aia rmethssticum Miq. Comm. ee 6. pl. 4, D. ae Syst. Piper. 217.

E. Parham in Agr. Jour. Dept. Agr. a 8(1): 2.

ISTRIBUTION: Throughout the Pacific Islands from New ae and Micronesia
eastward. The species is found only in cultivation, at least in Fiji, and its source is

men, giving the localities of Tahiti and the Tongan and Hawaiian groups. There

appear to be no Fijian specimens in American herbaria, but the plant is found in every

Fijian village where conditions are suitable.

I have cited above only the basic literature referring to P. methysticum

and treatments specifically discussing the plant in Fiji. The nomencla-

tural problems pertaining to the binomial have been competently discussed
by Moore (6). F. Brown’s account (1: 18-19) discusses the numerous
varieties cultivated in the Marquesas. Seemann (8: 260-261) has
described the use of the plant in Fiji, while Parham (7) has recently
written an interesting account of the species in Fiji, considering its varieties,
methods of cultivation, disease-control, etc.

3. Piper (§ Eupiper) Betle L. Sp. Pl. 28. 1753; C. DC. in DC. Prodr. 16(1): 359.
1869, in Candollea 1: 189. 1923; Ouisae hing in Philip. Jour. Sci. 43: 85. 1930;
Jan in Agr. Jour. Dept. Agr. Fiji 8(4): 49. 1937.

DistrR1BUuTION: Malaya to India, widely elevate throughout the tropics. The
common betel pepper is used and cultivated to a certain extent by the Indian popula-
tion of Fiji, as indicated by the discussion of Jan cited above. I have seen no herbarium
specimens of the plant from Fiji.

4. Piper (§ Eupiper) insectifugum C. DC. ex Seem. FI,, Vit. 262. 1868, in DC. Prodr.

6(1): 354. 1869; Billow in Gartenflora 45: 575. 1896; C. DC. in Jour. Linn.
Soc. Bot. 39: 164. 1909, in Candollea 1: 178. 1923
TION: Fs common in Fiji, where it occurs in forest or open woods a

The occurrence of the species in Samoa is noted only by Billow, as , indicated above,
and possibly it is endemic to Fiji.

Fiyt. Virr Levu: Seemann 569 8 (type cotyt.,GH); Tholo North: Vicinity
of Nandarivatu, Degener & Ordonez a ster. (A, NY), ane es fr. (A, NY,
WCAUS)s (Grecainoce 867 ster. (A); Rewa: Korombamba Mt., Gillespie 2315 ster.
(A, Bish). Kanpavu: Above anne and Ngaloa Bays, pa 75 8 (Bish, GH,
NY, UC, US). Vanua Levu: Mbua: Seatovo Range, Smith 1545 3 (Bish, GH,
NY, UC, US); Thakaundrove: Savu Savu Bay region, Degener & Ordonez

354 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXIV

13907 8 (A, NY, UC, US). Taveunr: Western slope between Somosomo and
Wairiki, Smith 841 ¢& (Bish, GH, NY, UC, US). Wuirxovut pDEFINITE LOCALITY:
Gillespie 2209 & (A, Bish), 2225 ster. (Bish), 3876 ¢@ (A, Bish).

This distinctive climbing species of § Eupiper is not correctly placed in
de Candolle’s key (5: 178). The stigmas are 3 and sessile, rather than 2
and on a style, while the stamens are 3 or 4 per flower. The presence of 4
stamens in flowers of § Eupiper is not indicated by de Candolle (5: 176),
but there can be no doubt of the proper place of P. imsectifugum in this
section. Its actual alliance is difficult to ascertain, but its leaves are some-
times remarkably similar to those of P. Betle; in inflorescence characters
the two species are quite different, and P. Betle is not indigenous in Fiji.
5. Piper (§ Eupiper) crispatum sp. nov.

Frutex scandens, ramulis apicem versus, petiolis, laminis utrinque,
et pedunculis pilis stramineis vel pallide brunneis crispatis multicellulatis

m. longis debiliter hispidis; ramulis subteretibus gracilibus nodis
conspicue incrassatis et mox disarticulatis, internodiis apicem ramulorum
versus 1—3.5 cm. longis; stipulis apice ramulorum lanceolatis ad 1 cm.
longis hirtellis mox caducis; foliis alternatis, petiolis paullo canaliculatis

0-16 mm. longis, laminis chartaceis in sicco brunneis ovatis, 9-13 cm.
longis, 6-8 cm. latis, basi inaequilateraliter rotundatis, apice gradatim
acuminatis (acumine ipso 1-2 cm. longo ad apicem nervis marginalibus
, Margine integris, supra demum subglabrescentibus, plerumque
5-pli-nerviis, nervis paullo supra basim orientibus, costa utrinque valida
superne nervos secundarios laterales conspicuos utrinsec cus 2-4 utrinque

culis 10-17 mm. longis, spicis sub anthesi gracilibus 4—5 cm. longis, rhachi
pilis multicellulatis circiter 0.5 mm. lo ongis dense pilosa; bracteis primo
imbricatis membranaceis peltatis inconspicue glanduloso-punctatis 1—1.2
mm, diametro breviter stipitatis; ovario in rhachi semi-immerso rotundato-
conico sub anthesi circiter 0.4 mm. diametro, stigmatibus ut videtur 3
minutis.

DISTRIBUTION: Known only from the type collection.

Fiji. WutTHOUT DEFINITE LOCALITY: Gillespie 3092 @ (A, Bish, TYPE), 1927-28
(woody vine, scrambling on tree

Although doubtless a close relative of P. insectifugum, this plant seems
specifically eigen by virtue of its crispate-hispid parts, as mentioned in

my key. The numerous specimens of P. insectifugum examined bear no
trace of the desea pubescence of the new species.

6. Piper (§ Eupiper) stipulare sp. nov.

Frutex dioecus scandens ubique praeter rhachem inflorescentiae inter-
dum obscure pilosam glaber, ramulis gracilibus subteretibus nodis incras-
satis, internodiis apicem ramulorum versus 1.5—6 cm. longis; stipulis
liberis vel raro petiolo adnatis conspicuis oblongis 15-22 mm. longis inter-
dum subpersistentibus; foliis alternatis, petiolis gracilibus 12-17 mm.
longis, laminis chartaceis in sicco brunneo-olivaceis ovatis vel elliptico-
ovatis, (7—)10—-14 cm. longis, (4—)6-10.5 cm. latis, basi inaequilateraliter
obtusis vel rotundatis, apice cuspidato- -acuminatis (acumine ipso 5-10

1943] SMITH, PACIFIC ISLAND PLANTS, II 355

mm. longo), margine integris, 7-pli-nerviis, nervis cum costa supra paullo
subtus valde elevatis, interioribus 7-20 mm. concurrentibus apicem attin-
gentibus, aliis antea evanescentibus, rete venularum obscuro vel subtus
leviter prominulo; inflorescentiis 4 et 2 apicem ramulorum versus oppo-
sitifoliis, pedunculis 7-10 mm. longis, spicis sub anthesi gracilibus 4-5 cm.
longis; bracteis membranaceis peltatis obscure pellucido-punctatis 0.7—0.8
mm. diametro breviter (ad 0.2 mm.) stipitatis; staminibus ut videtur 2,
antheris subsessilibus subreniformibus vel transverse ellipsoideis, circiter
0.25 0.7 mm.; ovario ut videtur libero obscure luteo-glanduloso
ovoideo- subgloboso sub anthesi circiter 0.7 mm. diametro, stigmatibus 3
sessilibus 0.3 mm. longis.
DISTRIBUTION: ee only from Viti tees

Fiji. has Leyu: Parks 20879 & (Bish); Namosi: 2 miles from Namuamua,
in woods near Namosi trail, oe 300 m., pes se, ae decomposed spikes (A,
Bish) (thick vine, climbing on tree); N athe iri: 7.5 miles from Suva, near road

ast ner village, in oo alt. 150 m., Gillespie 2423 Q (A, Bish, TYPE),
aie, 21 LOD Tes

bee om is probably most closely allied to P. Graeffei Warb. and
the several other Samoan species of § Eupiper proposed by C. de Candolle.
However, it is distinguished from all of these by its large and often per-
sistent stipules and its much shorter spikes.
7. Piper (§ Eupiper) Degeneri A. C. Sm. in Sargentia 1: 10. 1942.

DISTRIBUTION: Known only from the type collection.

Fijt. Vanua Levu: Thakaundrove: Eastern drainage of Yanawai River,
alt. 150 m., Degener & Ordonez 14096 2 (A, TYPE, NY).

As remarked in the original consideration, this very distinct and appar-
ently rare species has no close relatives in the Pacific.
8. Piper (§ Macropiper) oxycarpum C. DC. in Jour. Linn. Soc. Bot. 39: 164. 1909,

in Candollea 1: 1

DIstTRIBUTION: Endemic to Fiji and probably limited to the higher art in the
interior of Viti Levu. The type is Gibbs 604, from Nandarivatu, Tholo N

Fijt. Vitr Levu: In forest at 1250 m., Parks 20738 9 (A, aed Tho
North: Nandarivatu, in open bush, alt. 1000 m., Parks 20572 2 (Bish). Wirmour
DEFINITE LOCALITY: Gillespie 3839 2 (Bish)

This very distinct species is at once distinguished from all other members

of § Macropiper by having its branchlets distally, petioles, peduncles, and
leaf-blades on both surfaces conspicuously hispid-pilose, its inflorescence-
scales densely ciliate-setose, and its carpels conspicuously narrowed toward
the apex and bearing 2 or 3 very obscure stigmas. Other essential charac-
ters of the species are as follows: petioles 1-3 cm. long, one-half to three-
quarters vaginate; leaf-blades ovate, 8-17 cm. long, 5-12 cm. broad, sub-
cordate at base, gradually acuminate at apex, 7- or sometime 9-nerved;
spikes (at least @) solitary, 5-11 cm. long excluding peduncles, these
slender, 2.5—5 cm. long. According to de Candolle the type collection is
from a shrub 2 m. high.
9. Piper (§ Macropiper) puberulum (Benth.) Benth. ex Seem. Fl. Vit. 262, as syno-
nym, and fl. 75. 1868; non P. puberulum Maxim. (1886).
Macropiper puberulum Benth, in Hook. Lond. Jour. Bot. 2: 235. 1843.

356 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Piper Macgillivrayi C. DC. ex Seem. Fl. Vit. 262. 1868; C. DC.8 in DC. Prodr.
16(1): 335. 1869; Engl. in Bot. Jahrb. 7: 450. 1886, in Forschung. “Gazelle”

4: Siphon, 25. 1889; C. DC. in Denkschr. Akad. Wiss. Wien 85: 264. 1910;

Turrill in Jour. Linn. Soc. Bot. 43: 35, 1915

DistripuTiIon: Fiji, Samoa, Tonga, and probably some of the adjacent groups.
In Fiji the species is reported as a low erect shrub up to 3 m. high, occurring at alti-
peat from sea-level up to 900 m. or possibly higher. Var. typicum occurs in thickets
n the edges of forest, apparently only below 500 m., being especially common in
ee thickets. Var. glabrum usually occurs in the forest a nd is not reported from

spelled “yaqoyaqona”’). The two varieties recognize Fiji are not very sharply
differentiated and their value is dubious. Both varieties occur in Samoa, but I have
seen only var. glabrum from Tong

The necessity of replacing ite well-known binomial P. Macgillivrayi is
pointed out in my discussion above, where I also consider the typification
af the species and the two varieties which are known from Fiji.

9a. Piper puberulum var. typicum nom. nov
Macropiper puberulum Ad a in Hook. Lond. Jour. Bot. 2: 235. 1843; nee Syst.
Piper. 221. 1843; Seem. in Bonplandia 9: 259. 1861, in Jour. Bot. 2: 1864.
Piper dale Benth. ex Seem. FI. Vit. 262, as synonym, and fl. eS 1868;
low in Gartenflora 45: 575, nomen. 1896.

Piper Maceillivrayi C. DC. ex Seem. FI. Vit. 262. 1868, in Candollea 1: 172. 1923.

ae laminae subtus puberulae.

Viti Levu: Parks 20173, in part, 9 (Bish), rt 2 (Bish); Lautoka:
Reed pe Natalau, Degener 15005 @ (A, NY, UC, US); olo West: Mbulu, De-
gener 15042 9° (A, NY, UC, US). Kanpavu: Above Lhieaet and Ngaloa Bays, ‘Smith
56 9 (Bish, GH, NY, ‘UC, US), 115 : (Bish, GH, NY, UC, US); Mt. Mbuke Levu,
er 210 2 (Bish, GH, NY, UC, US). Ovarau: U.S. Expl. Exped. 1, in part,

9 (GH). Koro: East coast, Smith 1105 2 (Bish, GH, NY, UC, US). Vanua Levu:
U.S. Expl. Exped. 1,in part 9? (GH); Thakaundrove: Savu Savu Bay region,
Degener & Ordonez 13871 fr. (A, NY, UC, US). Vanua MBaravu: Smith 1408 9
(Bish, NY). FurLanca: Smith 1137 9 (Bish, NY). WurHouTt DEFINITE LOCALITY:
Seemann 567, in part, 2 (cotype coll. of Piper Macgillivrayi, GH), Prince (GH).

It should be noted that some specimens here cited (e.g. Smith 115 and
Degener & Ordonez 13871) bear essentially glabrous leaves on the same
branches with puberulent leaves. The possibility that the degree of
pubescence is merely a concomitant of shade conditions is thus indicated,
and it may be questioned whether the two varieties have any genetic
foundation.
9b. Piper puberulum var. pci (C. DC.) comb.

Piper Macgillivrayi var. glabrum C. DC. in DC. Prod 16(1): 335, excl. pe
Macropiper puberulum. 1869; Warb. (as var. nov.) in Bot. Jahrb. 25:
iy C. DC. in Jour. Linn. Soc. Bot. 39: - 1909, in Denkschr. Akad. —

n 85: 264. 1910, in Candollea 1: 172, 19
Piper “Mailers sensu aie in oe ce ‘Soc. Bot. 30: 189. 1894; Burkill
n Jour. Linn. Soc. Bot. 35: 52.

’The cited references to P. Macgillivrayi, with the exception of the first, cannot be
referred to a definite variety and therefore are listed under the species as an inclusive
concept.

4The numerous Tongan specimens I have seen belong to var. glabrum, and it seems
likely that only this variety occurs in Tonga; hence I have cited Niterabure referring to
Tongan collections here.

1943 ] SMITH, PACIFIC ISLAND PLANTS, II 357

and cpa glabrous on both surfaces.

Fiy EVU: Gillespie 2687 2 (Bish), Parks 20173, in part, fr. (Bish),
ae 9 ae 20452 fr. (A, Bish), 20731 fr. (Bish), 20735 9, fr. (A, Bish); Tholo
North: Vicinity of Nandarivatu, Parks 20606 fr. (Bish), sos aaa & Ordonez 13569
fr. (A, NY, UC, US), Degener 14361 fr. (A), 14659 fr. (A, NY); Namosi: Voma
Mt., as 2927 ? (Bish);Serua: Vicinity of Ngaloa, Degener 15140 8 (A, NY),
15179 $, 3% (A, NY, UC, US); Rewa: Mt. Korombamba, Gillespie 2235 fr.
ay, Kanpavu: Above Namalata and Ngaloa Bays, Smith 167 9 (Bish, GH, NY,
UC, US). Ovarau: U. S. Expl. Exped. 2 (GH). Vanua Levu: Thakaun-
drove: Savu Savu Bay region, Smith 331 9 (Bish, GH, NY, UC, US), 395 i “i
(Bish, GH, NY, UC, US), Degener & Ordonez 13829 fr. (A, NY), 13908 9, fr. (A
UC, US), 13967 fr. (A, NY); Maravu, near Salt Lake, Degener & Ordonez on ‘
(A, NY, UC, US). Wirnout peFrnite Locatity: Seemann 567, in part, Q (TYPE

of Piper Macgillivrayi var. glabrum, GH), Gillespie 2697 9 (A, ae 2698 Q
(Bish), 2974 fr. (Bish), 3004 2 (Bish), 3307.4 9 (A, Bish), 4658 9 (A, B
10. Piper (§ Macropiper) oe C. DC. in Jour. Linn. Soc. Bot. 39: 162.
1909, in Candollea 1: 172.

DIstRIBUTION: Endemic to a and a limited to Viti Levu. The type is
Gibbs 703, from Nandarivatu, Tholo Nor

Fijt. Vitt Levu: Tholo No ae : Nandarivatu, alt. 1000 m., Parks 20546 fr.
(Bish); Tholo ba aiden nes of Mbelo, near Mpc Tabualewa
15556 & (A, NY US); Rewa: Korombamba Mt., alt. 400-550 m., Gillespie
2217 6 (A, sa) ee. e) (A. Bish). WiTHOUT DEFINITE LOCALITY: U.S S. Expl.
Exped. 3 $ (G

According . ae Candolle, this is a slender shrub 1.5 m. high; the above-
cited collections have no habit data. The plant probably occurs in woods
or forests at middle elevations, although the Tabualewa and U. S. Exploring
Expedition collections may have been obtained near sea-level. The essen-
tial characters of the species are as follows: petioles 0.8—2 cm. long, vaginate
one-half or nearly all their length; leaf-blades ovate-oblong, 6-11 cm. long,
2—5.5 cm. broad, acute to obtuse at base, gradually acuminate at apex,
3- or S-nerved from the petiole; spikes solitary, axillary, slender, short
(2.5—5.5 cm. long excluding peduncle, even at anthesis or in fruit). In
floral characters, P. melanostachyum differs from P. puberulum only in its
slightly smaller parts, the peltate scales being 0.35-0.7 mm. in diameter;
the minute stamens and the stigmas are always three.

Although this plant may be only a montane derivative from the common
P. puberulum, I am inclined to agree with de Candolle in granting it specific
status, at least until more adequate material establishes a complete series
of forms between the two entities.

11. Piper (§ Macropiper) vitiense nom. nov
Piper latifolium sensu Seem. FI. Vit. 261, quoad spec. vit. 1868; no
Piper polystachyum C. DC. in Jour. Linn. Soc. Bot. ee 162. io in Candollea
923; A. C. Sm. in Bishop Mus. Bull. 141: 25. 1936; non Piper poly-
stachyon Ait. Hort. Kew. 1: 49. 1789 (= papi pe ey

5Degener 15179 is remarkable for its polygamo-monoecious reat some spikes
having only staminate flowers while others on the same plant have hermaphrodite

normal ovary. s is y
specimen of § Macropiper I have seen with hermaphrodite flowers, but it scarcely
weakens the characters of the section as defined by de Candolle (5: 171).

358 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

DistrisuTION: Endemic to Fiji, or possibly also in the New Hebrides. Occurring
in Fiji at elevations of 600 to 1100 m. in forest, often common locally. The type is
Gibbs 794, from Nandarivatu, Tholo North, Viti Levu

Fiyt. Vitt Levu: Namosi: Naitarandamu Mt., Gillespie 3360 92 (A, Bish).
Vanua Levu: T : akaun - rove: Mt. Mariko, Smith 458 6 (Bish, GH, NY, US).
TAVEUNI: Seemann 566 ster. (GH); Mt. Manuka, Smith 791 @ (Bish, NY). WirH-
OUT DEFINITE LOCALITY: need 3123 g (A, Bish). See also Smith in 193

ew name is needed for de Candolle’s species because of the earlier
a. a Ait. The epithets polystachyon and polystachyum
must be considered orthographic variants, according to Article 70 of the
International Rules of Botanical Nomenclature, 1935.

Piper vitiense, a member of the general alliance of P. latifolium L. f.,
differs from that species and its other relatives in its long-petiolate large-
bladed leaves and its numerous long-pedunculate staminate inflorescences.
Mature leaves of our specimens have the petioles up to 17 cm. long and the
sheaths 2.5—3.5 cm. long; in general the petioles are vaginate only in the
lower quarter, while P. latifolium has sheaths usually nearly as long as the
petioles. The largest leaf-blades of P. vitiense now available are up to
25 by 28 cm. and 13-nerved. The number of staminate spikes in the leaf-
axils is somewhat more variable than the 14 described by de Candolle.
Our material shows these spikes to be about 10-22 in number, giving the
species its most distinctive character. The staminate spikes are up to 6 cm
long and are borne on slender peduncles up to 5 cm. long. The only avail-
able pistillate specimen, Gillespie 3360, has broken detached spikes, which
offer no unusual character except as to number (which one may anticipate
approaches the number of staminate spikes) ; the stigmas are 3, as usual in
this section of the genus.

12. Piper (§ Macropiper) Timothianum A. C. _ in Sargentia I: 1942.

Piper Macgillivrayi var. fasciculare Warb. in aa Jahrb. 25: 609, as Ci edie
1898; C. DC. in Denkschr. Akad. bees Wien 85: 264, as “fascicularis.’”’ 1910,
in Ann. Cons. Jard. Bot. Genéve 15: as “fascicularis. ” 1012; Turrill in Jour.
Linn. Soc. Bot. 43: 35, as ae " ers C. DC. in Candolles 1: 172. 1923;
Christoph. in Bishop Mus. Bull. 154: 5. 1938

Piper fascicularis (sic) vel a. Rechinge er in Karsten & Schenck, Vegeta-
tionsbilder 6: pl. 5. ee n P. fasciculare Rudge, Pl. Guian. Rar. 1: 9. pl. 4.

O05 (= Lacistema sp P. fasciculatum Ruiz & Pav. Syst. Veg. 1: 362. 1798.
Piper Macgillivrayi var. eases (sic) forma 6 C. DC. in Jour. Linn. Soc. Bot.
62. 19

DistRIBUTION: Fiji and Samoa. In Fiji the species is common locally in pa forest
and ridge-thickets of Viti Levu at elevations of 550-1200 m.; it is a shrub 2-5 m. high
In Samoa (as P. Macgillivrayi var. fasciculare) it is said to occur in some abundance on
Savaii and Upolu, in essentially similar habitats at elevations up to 1500

Fiyt. Vitr Levu: Tholo North: Nandarivatu, Degener & a 13570 @
(A, TypE, NY, UC, US), Parks 20777 & (Bish), ee fr. (Bish), Gillespie 4214 i
(A, Bish) ; Nauwanga, Degener 14360 fr. (A, NY, UC, US), 14620 fr. Y
Nandrau, Degener 14891 fr. (A, NY); Namosi: Vicin ity of Namosi, Gillespie 2688
@ (A, Bish), Parks 20238 fr. (Bish), 20257 fr. (Bish) ; Korombasambasanga Mt., B. E.
Parham 2212 fr. Sate Vanua Levu: Thakaundrove-Mathuata
boundary: Korotini Range, Smith 548 Q (Bish, GH, = UC, US). WitrHout
DEFINITE LOCALITY: Gillespie 2782 9 (Bish), 3/24 fr. (Bish

In proposing this entity as a new species in oe 1 considered its simi-

1943 ] SMITH, PACIFIC ISLAND PLANTS, II 359

larity to the Samoan plant which has been passing as P. Macgillivrayi var.
fasciculare, but it seemed to me at that time, as at present, that the plant
cannot be placed in “P. Macgillivrayt” (i.e. P. puberulum, as defined in the
present treatment) without undue expansion of that concept. The other
varieties of P. puberulum have the spikes usually solitary, but sometimes
those at the upper nodes are paired. Piper Timothianum, on the other
hand, has the spikes normally 3—7 per axil, very rarely 2. Several students
of the Samoan flora, including Christophersen in 1938, have remarked that
the number of spikes may vary from 1 to 4, but I have not observed fewer
than 2 (and this very rarely) in the cited Fijian material. Furthermore
the spikes (excluding peduncles), both staminate and pistillate, are only
0.8-3.5 cm. long. Christophersen finds that Samoan plants may have the
pistillate spikes up to 5 cm. long and the staminate up to 10 cm. Piper
puberulum, in the sense adopted by me, has the spikes between 4 and 19 cm.
long, but the spikes of either sex are rarely less than about 7 cm. long.
Although the two species are doubtless closely related and quite possibly
interfertile, I fail to see how the present entity can be included in P. puber-
ulum without expanding that concept to an unwarranted degree, perhaps
even submerging it in P. latifolium L. f. Few students of Piper will wish io
combine species to this extent, in which case the whole Section Macro piper
would scarcely be divisible into species.

Apparently only Rechinger, in 1908, has thought P. Macgillivrayi var.
fasciculare worthy of specific rank; he used the specific epithets “fascicu-
laris” and ‘“‘fasciculatum” indiscriminately, but neither is available for use
in Piper. The plant was collected at Nandarivatu by both Gibbs and im
Thurn, whose specimens were referred to Warburg’s variety by de Candolle
and Turrill.

13. Piper (§ Macropiper) kandavuense sp. nov.

Frutex 3 m. altus ubique inflorescentia excepta glaber, ee teretibus
nodis valde incrassatis, internodiis apicem ramulorum_ versu cm.
longis; foliis alternatis, petiolis 2-3 cm. longis fere ad ai conspicue
vaginantibus (alis 3-4 mm. latis superne ad petiolum abrupte decurrenti-

bus), laminis chartaceis in sicco olivaceis late ovatis, 10-15 cm. longis,
8-14 cm. latis, basi truncato-subcordatis, apice cuspidato- acuminatis,
margine integris, 7(vel inconspicue 9-)-nerviis, nervis e basi divergentibus
utrinque conspicue elevatis, rete venularum utrinque haud_ prominulo;
inflorescentiis @ solis visis in axillis foliorum apicem ramulorum versus
4 vel 5 aggregatis, pedunculis validis glabris 1-2 cm. longis, spicis paullo
post anthesin 2-3 mm. diametro 5—8 cm. longis; rhachi pilis pallidis
0.2—0.4 mm. longis sparse pubescente; bracteis liberis peltatis membrana-
ceis circiter 0.8 mm. diametro breviter stipitatis; ovario globoso-ellipsoideo
circiter 1 mm. diametro (immaturo), apice rotundato, stigmatibus 3
patentibus circiter 0.3 mm. longis dense et conspicue brunneo-pilosis et
ciliatis coronato.

DISTRIBUTION: Known only from the type collection.

Kanpavu: Mt. Mbuke Levu, alt. 200-500 m., Smith 219 9 (Bish, GH,
TYPE, NY. UC, US), Oct. 23, 1933 (shrub 3 m, high, in Hie pee
The specimen above described belongs among the allies of P. latifolium

360 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XXIV

L. f., but in several details it differs from my concept of that species. On
the whole, it has shorter petioles than P. latifolium, at least on leaves of
comparable size and maturity, and its petiolar sheaths are slighly broader
in proportion and more abrupt distally. In the material of P. latifolium
which I have seen from the eastern Pacific, the spikes, both staminate and
pistillate, are never more than 3, but the original description states that the
spikes are 5 or more; whether the original specimen was staminate or
pistillate is not stated. Although the stigmas of P. latifolium are glandular-
puberulent, they are never as conspicuously pilose as those of Smith 219.

In view of these differences, and especially the difference pertaining to
the stigmatic character, I doubt if Smith 219 can be referred to P. latifolium.
No Fijian specimen which has yet come to my attention seems to agree
precisely with material of P. /atifolium from the eastern Pacific, but Smith
219 seems closest, among Fijian plants, to Linnaeus’ species, the limits of
which are not vet entirely understood, as stated above.

INSUFFICIENTLY KNOWN ENTITIES FROM FIJI

PIPER a maeor awe MacciLiivrayi C. DC. var. pARvIFOLIUM C, DC, in DC. Prodr.
(1):

fi ae S a treatment of this is as follows: “ . . . limbis 0,06
longis, 0,03 latis 5—7-nerviis . . . In ins. Fijee (Barclay ! in h. Kew.).”
Leaves with these small dimensions have been observed among the available
Fijian collections only on plants referred to P. melanostachyum C. DC., in
which the leaf-blades are not 7-nerved. It seems probable that P. M acgilli-
vrayi var. parvifolium is a very depauperate individual of P. puberulum var.
glabrum.

PIPER (§ reno pig — C. DC. in Jour. Linn. Soc. Bot. 39: 163. 1909, in
andollea 1: 17

According to de “Candolle, this species is characterized by its small leaf-
blades (8 3.5 cm.), which are densely hirtellous beneath and 7-nerved,
its hirtellous petioles and peduncles, its short pistillate spikes (3.5 cm. long),
and especially its hirsute ovaries. On the basis of the original description
this appears to be a distinct species, suggestive of P. oxycarpum and possibly
P. puberulum var. typicum. No Fijian material available to me can be
referred to P. Gibbsiae, which I hesitate to place without seeing the type,
Gibbs 722, from Nandarivatu, Tholo North, Viti Levu.

PIPER Pe rica et ries C. DC. in Jour. Linn. Soc. Bot. 39: 163. 1909, in
andollea 1:

aos the prea ae amie and de Candolle’s key (5: 173), one may
assume that this species is a close relative of P. Gibbsiae, differing chiefly
in its slightly larger (13 5.8 cm.) and more obviously acuminate leaf-
blades, longer pistillate spikes (6 cm. long), and more sparsely pilose
ovaries. Without examining the type, it is inadvisable to draw conclusions
as to the value of the species, which is based on Gibbs 599, from Nanda-
rivatu, Tholo North, Viti Levu.

1943 ] SMITH, PACIFIC ISLAND PLANTS, II 361

PRINCIPAL LITERATURE CITED

1, Brown, F.B.H. Flora of of Polynesia, III. Dicotyledons. Bishop Mus.
Bull. 130: Peter 17-20.
. CANDOLLE, . DE. a a Prodr. 16(1): 23565_471. 1869.

2

3, ———__ peraceae (in Gibbs, A oo to the montane flora of Fiji).
Jour, tee eae Bot. 39: 162-167.

4, ————. Pi eraceae Ae Snr ota he und Zoologische Ergebnisse . . .).
Denkschr. Akad. Wiss. Wein 85: 264-269 [reprint 3: 90-95]. 1910

5, ———___., a eieia clavis ie Candollea 1: 65-415, 1923.

6. Moore, J. W. Taxonomic studies of Raiatean plants. Bishop Mus. Occ. Pap.
10(19): 1-8. 1934.

ParHAM, B. E. V. Wilt disease of ““Yangona.” Agr. Jour. Dept. Agr. Fiji 8(1):
2-8. 1935

sJ

8. SEEMANN, B. Flora Vitiensis: Piperaceae, 259-262, 1868.

9. SeTcHELL, W. A. Tahitian spermatophytes ner by W. A. eee CyB:
Setchell, and H. E. Parks. Univ. Cal. Publ. Bot. 12: Piper, 163.

10. Warsurc, O. Piperaceae (in Reinecke, Die nee der eae ae Jahrb.

25: 609-612. 1898.

ARNOLD ARBORETUM,
HArvARD UNIVERSITY.

362 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

NOTES ON THE FLORA OF INDO-CHINA
Hut-Lin Li

Tuts paper is based almost entirely on a part of the very excellent and
extensive collections of botanical material from the extreme northeastern
part of Tonkin, adjacent to the Kwangtung border, assembled by Mr. W. T.
Tsang on the second, third, and fourth Lingnan Indo-China Expeditions.
These expeditions were sponsored by the Botanical Survey of Lingnan
University in co-operation with the Arnold Arboretum and were supported
by grants from the latter institution. The second expedition was in the
fall of 1936, the third in the spring and early summer of 1939, and the
fourth from May to November, 1940. All the material of the first two
expeditions, when received from the field, was stored in Canton in order
that the necessary herbarium labels could be prepared. The fourth expedi-
tion was actually in the field when the Japanese occupied Indo-China, but
Mr. Tsang succeeded in returning to Hong Kong with all of his material.
Because of the rapidly deteriorating conditions in the Orient, all of the
material from the three expeditions was assembled in Hong Kong, and,
through the active interest of Prof. F. A. McClure, the Arnold Arboretum
share of the collections was shipped to the United States in January 1941.
Because of the critical situation in Canton and the increasingly critical one
in Hong Kong, time did not permit the actual transcription of the field notes
for the numbers involved on the fourth and last expedition; we do, however,
have the localities and the inclusive dates of collection for each locality, and
we were thus in a position to have the necessary printed herbarium labels
prepared.

In the three collections there is a total of about 2000 numbers. Of the
second and third expedition collections we normally have four sets of
specimens for each number; but the number of duplicates of the fourth
collection is very much larger, often running from ten to fifteen specimens
for each number.

Mr. Tsang, on all three expeditions, operated in the coastal regions in
the vicinity of Ha Coi and the mountainous region inland from Ha Coi and
Tien Yen, much of the material being from the mountain ranges immedi-
ately south of the juncture of the Kwangsi-Kwangtung-Tonkin boundaries.
Naturally, in the collection as a whole, a great many Chinese elements are
represented in the form of genera and species originally described from
Kwangtung, Kwangsi, and neighboring provinces, as well as the Island of
Hainan. When the entire collection is studied, group by group, many addi-
tions to the Indo-Chinese flora will result. All specimens cited are depos-
ited in the herbarium of the Arnold Arboretum.

This study was made possible through a grant made from the Milton
Fund of Harvard University to Dr. E. D. Merrill, Director of the Arnold

1943 J LI, NOTES ON THE FLORA OF INDO-CHINA 363

Arboretum, to be utilized in completing the identification work on the
recently received botanical collections from various parts of China and
neighboring regions.

PROTEACEAE

Helicia Loureiro
Helicia Henryi Diels, Repert. Sp. Nov. 13: 528

Helicia pallidiflora W. W. Smith, Notes Bot. ne) ea 10: 179. 1918.

Inpo-Cu1na: Tonkin, northwest of Mon-cay, Pac-si and vicinity, W. T. Tsang
26961, Oct. 1-8, 1936, 7 ft. high, fairly common, growing in thickets, on dry clayey soil,
fruits yellow or black; Taai Wong Mo Shan, Chuk-phai, W. 7. Tsang 27078, Oct. 23-31,
1936, 27252, Nov. 10-17, 1936, 29299, July 1-13, 1939, 7-9 ft. high, fairly common,
growing in thickets, on dry clayey or sandy soil, fruits vellow or black. Yunnan,
Kwangsi, Hainan. New to Indo-China.

Helicia vestita W. W. Smith, Notes Bot. Gard. Edinb. 10: 181. 1918.

HINA: Tonkin, Dam-ha, Sai Wong Mo Shan, Lung Wan Village, W. T.
ee 30083, May 18 — July 5, 1940. Yunnan. New to Indo-China.
Helicia Tsangii sp. nov.

Arbor parva circiter 3-5 m. alta, ramulis ee dense ferrugineo-
tomentosis serius glabrescentibus; foliis chartaceis vel subcoriaceis,
breviter petiolatis (S-10 mm.), anguste ablonee ee vel oblongo-
oblanceolatis, 17-26 cm. longis, 5.5-9.5 cm. latis, apice plus minusve
abrupte acuminatis, basi angustatis, acutis vel cuneatis, margine crasse et
remote serratis, supra glabris subnitidis, in sicco olivaceis, subtus glabris

tiis axillaribus erectis ad 16 cm. longis, dense subcastaneo-tomentosis,
multifloris, bracteis bracteolisque ovato-lanceolatis, ad 2—3 mm. longis,

scentibus, solitariis vel binis vel plerum mque 2-4-fasciculatis; floribus
circiter 15 mm. longis, albis, extus leviter tomentosis; antheris circiter
1.5 mm. longis; ovario circiter 1 mm. longo glabro, stylo gracili glabro;
squamis hypogynis oblongo-quadratis, basi leviter cohaerentibus; fructibus
sit eines subglobosis, circiter 1 cm. diametro, brunneis.

NDO-CHIN Tonkin, northeast of Mon-cay, Pac-si and vicinity, W. T. Tsang
en Sept. 7 30, 1936, 10 ft. high, abundant, alas in thickets, fruits brown:
Ha-coi, Taai Wong Mo Shan, Chan Uk Village near Chuk-phai, W. T. Tsang 28974,
May 3-10, 1939, 29174 (tTypE), June 1-9, 1939, 10-15 ft. high, fairly common, growing
in thickets, on dry clayey soil, flowers white, fragrant.

A species allied to Helicia silvicola W. W. Smith, but distinguished by the
much broader, narrowly obovate to oblong-oblanceolate, coarsely toothed,
short-petioled leaves, and by the longer flowers, which are hairy on the
outsid

LORANTHACEAE
Elytranthe Blume
Elytranthe cochinchinensis (Lour.) G. Don var. tonkinensis (Lecomte) comb. nov.

Elytranthe ampullacea G. Don var. peyenions pear FI. Gén. Indo-Chine 5: 205.

ee ener Tonkin, Tien-yen, Kau Nga Shan and vicinity, W. T. Tsang 27491,

364 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIN

Jan. 1-9, 1937, 2 it. high, fairly common, on trees in thickets, fruits purplish-red, flowers
fragrant.

Three varieties are recognized by Lecomte in Indo-China. The other two
are Elytranthe cochinchinensis (Lour.) G. Don var. puberula (Le-
comte) comb. nov. (E. ampullacea G. Don var. puberula Lecomte, FI. Gen.
Indo-Chine 5: 205. 1915), from Cambodia, and var. Harmandii (Le-
comte) comb, nov. (EZ. ampullacea G. Don var. Harmandii Lecomte, |. c.
from Laos. Danser (Bull. Jard. Bot. Buitenz. III. 16: 5. 1938) refers all
three to the type of Macrosolen cochinchinensis (Lour.) van Tieghem =
Elytranthe cochinchinensis (Lour.) G. Don, as representing different forms
of a polymorphic species. But the specimens from Indo-China show enough
differences to prove that Lecomte is not unjustified in proposing these
varieties.

Loranthus Linnaeus
Loranthus cordifolius Wall. in Roxb. Fl. Ind. 2: 222. 1824.

Inpo-Cutna: Tonkin, Dam-ha, Sai Wong Mo Shan, Lomg Ngong Village, W. T.
Tsang 30430, July 18 — Sept. 9, 1940; Tien-yen, Kau Nga Shan and vicinity, W. T
Tsang 30570, Sept. 23 — Oct. 7, 1940. Eastern india. New to Indo-Chin

This belongs in Scurrula Linn. in accordance with Danser’s achease of
classification.

Loranthus tienyensis sp. nov

Frutex parasiticus, ramis ramulisque teretibus dense castaneo-tomen-
tosis; foliis firmiter chartaceis, petiolatis, juventute utrinque breviter
dense ferrugineo-tomentosis vel subtus secus costam tomentos sis, maturitate
mox glabris ellipticis, 2.5-4.5 cm. longis, 1.5—2.3 cm. latis, utrinque late
rotundatis, in sicco supra olivaceis, subtus paullo pallidioribus, costa supra
obscura subtus elevata, nervis lateralibus obsoletis vel subobsoletis; peti-
olo circiter 7 mm. longo ferrugineo-pubescente; inflorescentiis axillaribus
aah plerumque bifloris, pedunculis teretibus circiter 5 mm. longis

nse breviter subcastaneo- tomentosis: pedicellis teretibus vatataae 7 mm
ane bracteis minutis; calyce subellipsoideo 2-3 mm. longo, margine
rac corolla dense breviter subplumoso-tomentosa, ad 4.5 mm. longa,
3 mn gsi apice 4 4-lobata, lobis obtusis 1 mm. latis; filamentis
ee 1m _lon ngis, antheris oblongis 1.5 mm. longis; stylis glabris fili-
formibus, alae lobis aequalibus vel paullo longioribus, stigmate capitato.

Inpo-Cutna: Tonkin, Tien-yen, Ho Yung Shan and vicinity, W. 7. Tsang
30689 (TYPE), Oct. 13 — Nov. 22, 1940.

A very distinct species, characterized by its small elliptic leaves, which
are rounded at both ends, without distinct lateral nerves, and with a dense
tomentum along the lower portion of the midrib beneath, the mature leaves
otherwise being wholly glabrous. It is remotely allied to Loranthus noto-
thixoides Hance, differing, among other characters, in the larger leaves and
very much longer flowers. In Danser’s classification it falls into Scurrula
Linn.

ANNONACEAE

Goniothalamus Hooker f. & Thomson
Goniothalamus chartaceus sp. nov.

Frutex vel arbor parva, ramis ramulisque glabris nigris; foliis sub-

1943 | LI. NOTES ON THE FLORA OF INDO-CHINA 365

chartaceis glabris breviter petiolatis be es oblongo- pr venti vel lan-
ceolatis, 11-17 cm. longis, 1.7—2.8 cm. latis, breviter obtuse acuminatis,
basi acutis, margine leviter revolutis, in sicco olivaceis a eee concolori-
bus, minute puncticulatis, costa supra leviter impressa subtus distincte
elevata, nervis lateralibus utrinsecus 10-12, utrinque subconspicuis, tenui-
bus, fere ad medium inter costam atque marginem curvato-anastomo-
santibus, venis tertiariis inconspicuis; petiolo circiter 5 mm. longo;
floribus axillaribus solitariis, pedicellis circiter 1.2 mm. longis, glabris,
prope basim 3-bracteatis bracteis 1-1.5 mm. longis acuminatis; ee

mm.
4 mm. latis, fere liberis, oieas petalis exterioribus liberis pee
longe acuminatis, circiter 1.8 cm. longis, petalis ae oe
coalitis, ovato- triangularibus, breviter acuminatis, circiter 1 ongis;
staminibus numerosis circiter 1.5 mm. longis; carpellis numerosis minutis
brunneo-tomentosis
Inpo-CuHina: Ton ‘kin, Dam-ha, Tins Wong Mo Shan, Lung Wan Village, W. 7.
Tsang 30097 (TYPE), May 18 - July 5
A species characterized by its pe narrow, puncticulate leaves and long
and slenderly acuminate outer petals, being quite unlike any previously
described species from this region, perhaps most closely allied to Gonio-
thalamus elegans Ast.
HAMAMELIDACEAE
Embolanthera Merrill
Embolanthera glabrescens sp. nov.
utex vel arbor parva, glabra, ramis brunneo-cinereis teretibus,
ramulis gracilibus obscure brunneis; foliis firmiter chartaceis glabris,
petiolatis, lanceolatis vel oblongo-lanceolatis, 7-12 cm. longis, 1.5—4 cm.
latis, longe acuminatis, basi plerumque perspicue inaequilateralibus, uno
latere acutis, altero rotundatis vel late acutis, margine integris subcarti-

supra subconspicuis, subtus conspicuis; petiolis canaliculatis,

longis; inflorescentiis terminalibus solitariis spicatis multifloris, 8-12 cm.
longis, glabris, pedunculis 1—-1.5 cm. longis; floribus her maphroditis sessili-
bus, calyce toto glabro, circa 3.8 mm. longo, infra cum ovario connato, ad
basim 2-partito, lobo uno apice breviter fisso, lobis apice ciliatis; petalis
perigynis saepissime 5 raro 4—6, in alabastro corrugato-involutis, maturis
luteis linearibus, circa 2 cm. longis, 1-1.5 mm. latis, breviter acuminatis,

yngo, connectivo breviter producto, thecis . Esp rubro- efinner
basifixis, locellis 4; staminodiis nullis; ovario semi-infero 2-loculari supra
pube escente: stylis 2 glabris 6 mm. longis cylindricis acuminatis, ovulo in
loculo quove singulo pendulo; pees capsulari, basi annulato, ovoideo, in
valvis 2 dehiscente, 8-10 mm. lon 1 mm. crasso, epicarpio levi corneo
crasso, endocarpio corneo see calycis limbo circumscisso-deciduo,
semine ignoto.
Inpo-CH1InA: Tonkin, Tien-yen, Ho Yung Shan and vicinity, W. 7. Tsang 30709
(TYPE), Oct. 13 — Nov. 22, 1940
This is an interesting nadibion to the flora of Indo-China. The genus

366 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

Embolanthera was described from a single Philippine collection from
Palawan, its type species being E. spicata Merr. Asa genus, it is strongly
characterized by the spicate inflorescences, manifestly auriculate bases of
the petals, membranaceous irregularly 2- or 3-lobed calyces, and the
absence of staminodes. The present species differs from E. spicata Merr.
in the narrower, more slenderly acuminate, and generally fewer-nerved
leaves, shorter petioles, glabrous spikes which are mostly terminal, and
glabrous calyces. The fruits were previously unknown, but unfortunately
I am not able to describe the seeds, for in the abundant material now avail-
able the seeds are all fallen.

This new species considerably extends the range of the genus. The type
was collected at Tien-yen in northeastern Tonkin, near the Kwangtung-
Kwangsi border. It is highly probable that the same species, or related
ones, may eventually be found also in southern Kwangsi, when the latter
area is more extensively explored. The flora of Kwangsi is particularly
close to that of Tonkin, as is shown by numerous species in recent
collections being present in both areas.

STERCULIACEAE
Reevesia Lindley
Reevesia macrocarpa sp. nov.
Arbor, ramis glabris teretibus, ramulis glabris; foliis firmiter chartaceis,
utrinque Pras pallidis, nitidis, oblongo-ellipticis, 13-18 cm. longis,
4.5—-6 cm. latis ie basi ae aneraa | leviter race costa

tertiariis utrinque elevatis; aera 2.5-4.5 cm. longo, pa floribus
ignotis; a pg pedicellatis lignosis obovoideo-oblongis, 5.5—6 cm.
longis , 2.5-2.7 latis, 5-lobatis, apice rotundatis, basi acutis, extus
griseo- Parsee ETI circiter 4.5 cm. longo; seminibus (cum alis )
circiter 3 cm. longis, alis brunneis circiter 2.2 cm. longis, 0.7 cm. latis,
oblongis, apice rotundatis

Inpo-Cuina: Tonkin, Da m-ha, Nae Mo Shan, Lomg Ngong Village, W. T.
Tsang 30473 (TYPE), July 18 — Sept. 9,

A species distinguished in ee al genus by its rather large entirely
glabrous leaves and the large fruits. It is nearest to Reevesia longipetiolata
Merr. & Chun of Hainan, but it differs in its vegetative characters as well
as in its larger fruits. This i is the second actual record of Reevesia as occur-
ring in Indo-China. Gagnepain, in Lecomte, Fl. Gén. Indo-Chine 1: 486.
1910, includes the description of Reevesia thyrsoidea Lindl. on the basis of
a Bon specimen which perhaps came from Tonkin. He states: ae Py origin
tonkinoise de cette espéce n’est pas certaine, le P. Bon ayant herborisé a
Hong-Kong.” Merrill, Jour. Arnold Arb. 19: 53. 1938, has credited
Reevesia pubescens Nast to Tonkin.

DILLENIACEAE
Actinidia Lindley
Actinidia tonkinensis sp. nov.
Frutex scandens, ramis glabris, ramulis junioribus subdecidue breviter

1943] LI, NOTES ON THE FLORA OF INDO-CHINA 367

puberulis; foliis chartaceis, longe petiolatis, ae vel elliptico-ovatis,
7-12.5 cm. longis, 4.5—6.8 cm. latis, apice acutis vel breviter acuminatis,

basi angustatis, margine infra medium eee sursum distanter serrulatis,
in sicco olivaceo-viridibus, utrinque subconcoloribus vel subtus paullo
pallidioribus, supra glabris subnitidis, subtus minute consperse stellato-
puberulis, nervis lateralibus utrinsecus 6-8 gracilibus utrinque manifestis
subtus elevatis, venis tertiariis plus minusve parallelis, supra subconspic-
uis, subtus distinctis petiolo 2.2—4.8 cm. longo, primo puberulo demum
su ygla ro; inflores centiis dense subferrugineo- pubescentibus, cymosis,
4—5 cm. longis, 5—20-floris; pedunculis 2—3 cm. longis; floribus ¢ 5-meris,
circiter 1 cm. diametro, pedicellis 0.7—1 cm. longis; sepalis ovatis obtusis
extus dense pubescentibus, circiter 4 mm. longis; petalis oblongis glabris,
circiter 5 mm. longis et 2 mm. latis, apice rotundatis; staminibus circiter
40, l-seriatis, filamentis 2-3 mm. longis gracilibus, antheris oblongo-
ovatis 1 mm. ongis ; ovario subgloboso dense pubescente; floribus herma-
phroditis non visis.

Inpo-Cuina: Tonkin, Dam-ha, ie Wong Mo Shan, Lung Wan Village, W. T.
Tsang 29907 (TYPE), May is July 5,

A species apparently in the ae of A. latifolia (Champ.) Merr., but
distinguished, among other characters, by the leaf base being acute or
broadly acute instead of cordate or broadly rounded, the lower surface
less pubescent, with scattered stellate hairs, and the more delicate and lax
inflorescences with longer and more slender peduncles and pedicels and
somewhat smaller flowers.

FLACOURTIACEAE

Bennettiodendron Merrill
Bennettiodendron cordatum Merr. Jour. Arnold Arb. 20: 352. 1939.

npo-CHINA: Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsan
27043, Oct. 16-22, 1936, 27122, Oct. 23-31, 1936, 29261, June 23-30, 1939, 5-30 ft. high,
fairly common, growing in thickets, in sandy soil ; ane ha, Sai Wong Mo Shan, Lung
Wan bed W. i i pis 30058, May 18 — July 5, 1940.

The type, a flowering specimen, was from Mount Bavi, Tonkin; the
above- cited Pitecaons are all in fruit. Infructescences narrowly panicu-
late, 9 to 12 cm. long, sparingly pubescent, ultimately glabrous or nearly
so, the longest branches not exceeding 2 cm. in length. Fruits globose, dark
brown, glabrous, about 7 mm. in diameter.

THYMELAEACEAE
Wikstroemia Sprengel
Wikstroemia nutans Champ. in Hook. Jour. Bot. Kew Gard. Misc. 5: 195. 1853.
Inpo-Cu1na: Tonkin, Tien-yen, Kau Nga Shan, W. T. Tsang 27340, Dec.

13-22, 1936, 4 ft. high, fairly common, in thickets, flowers yellow. Kwangsi, Kwang-
tung, Hainan. New to Indo-Chi

ARALIACEAE
Dendropanax Decaisne & Planchon

Dendropanax parviflorus (Champ.) ae Fl. Hongk. 137. 1861; Merr. Brittonia
4: 132. 1941; Li, Sargentia 2: 45.

368 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV
Hedera parviflora Champ. ex Benth. in Hook. Jour. Bot. Kew Gard. Misc. 4: 122.
1852.
ae parviflora Harms in Engl. & Prantl, os ee 3(8): 41. 1894.

Textoria parviflora Nakai, Jour. Jap. Bot. 1
Inpo-CHINA: bbe onkin, northwest of Mon-cay, es si and oe W. T. Tsang

26981, Oct. 1-8, 1936; Ha- col, Chuk-phai, Taai Wang Mo Sha . T. Tsang 27241,
Nov. 10-17, ery Kwangtung and Kwangsi. New to Indo- eis
CORNACEAE
Cornus ee
Cornus hongkongensis Hemsl. Jour. Linn. Soc . 23: 345,

1888.
Benthamia japonica Sieb. & Zucc. var. sinensis ey in Hook. Jour. Bot. Kew Gard.
Misc. 4: 165. 1852.
sioehisagiy a Hutchinson, Ann. Bot. 6: 93, 1942.
npo-CHinA: Tonkin, Mount Bavi, A. Pételot 2147, June 24, 1939; Dam-ha,
Sai ae Mo Shan, Lung Wan eee W.T. Tsang 29942, May—July, 1940. Southern
and eastern China. New to Indo-Chin
This species has been much hee with Cornus capitata Wall., a
species of wide distribution, which also occurs in Tonkin. It differs from
Wallich’s species chiefly in having a truncate, entire calyx. In addition,
it can be distinguished by its leaves being generally broader, more coria-
ceous, slightly lustrous above and pale green, not grayish nor whitish
beneath. In leaf form, size, and length of acumen, the two species are
similar.
SAPOTACEAE

Madhuea J. F. Gmelin

Madhuea Tsangii sp. nov.

Arbor ramulis atrobrunneis glabris; foliis ad ramulorum apices sub-
verticillatim oo chartaceis vel subcoriaceis, petiolatis, obovatis,

4.5—-6.5 cm. longis, a5. 3.5 cm. latis, apice obtusis vel late rotundatis,
deorsum angustatis, basi cuneatis, utrinque glabris, subtus minutissime
puncticulatis, costa subtus valde prominente, nervis lateralibus supra
inconspicuis, subtus gracilibus, utrinsecus circa 12- 1S rectis prope mar-

lobis 4 biseriatis ovatis, 5 mm. latis, toto pubescentibus; corollis exsertis,
8 mm. longis, glabris, tubo 2-3 mm. longo, petalis 8 oblongis, 5 mm. longis,
2.5 mm. latis, subrotundatis; staminibus 16 fauce insertis, filamentis
1 ongis, antheris 2.5 mm. longis acuminatis; ovario ovoideo pubes-
cente 8- incalan: stylis 1 cm. longis, basi pubescentibus; fructu ignoto

Inpo-Cu1na: Tonkin, Dam-ha, Sai Wong Mo Shan, Lomg Ngong Village, W. 7.
Tsang 30271 (TYPE), July 18 ~ Sept. 9, 1940.

In the pubescent ovary and obovate leaves, this new species resembles
Madhuca Thorelii (Merr.) H. J. Lam and Madhuea cambodiana (H. Le-
comte) comb. nov. (Payena cambodiana H. Lecomte, Fl. Gén. Indo-Chine
3: 912. 1930). The former is an incompletely known species which, on
the basis of the original description, differs from the present species in the
smaller leaves, shorter pedicels, and smaller flowers. The latter has larger

1943 | LI, NOTES ON THE FLORA OF INDO-CHINA 369
leaves with loosely arranged veinlets, as well as shorter pedicels and smaller
S

Sideroxylon Linnaeus
Sideroxylon Wightianum Hook. & Arn. var. tonkinense var. nov.

A typo speciei differt foliis lanceolatis, ad 13-16 cm. longis, 3-4 cm.

sg longe acuminatis, basi valde attenuatis, petiolis 1.5—2 cm. longis.

npo-Cuina: Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsang
ea a ae 16-22, 1936, 10 ft. high, fairly common, in thickets, on dry sandy
soil, flower

The ae - the species was from Kwangtung. This variety differs an
the typical form in the much more lanceolate leaves. Lecomte, FI.
Indo-Chine 3: 887. 1930, gives an extensive description of Cee
Wightianum Hook. & Arn. var. Balansae Lec. but his description applies
to a form distinctly different from the one above characterized, its leaves
being but 10-12 cm. long, obtusely acuminate, and its petioles ae 8-12
mm. in length.

SARCOSPERMATACEAE
Sarcosperma Hooker f.
Sarcosperma laurinum (Benth.) Hook. f. in Benth. & Hook. f. Gen. Pl. 2: 655.
1876; Lam & Varos, Blumea 3: 195. 1938

Reptonia laurina Benth. Fl. Hongk. 208. 1861.

Inpe-Cuina: Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsang
27162, Nov. 1-9, 1936, .. ft. high, fairly common, in thickets, een pale yellow,
fragrant; Tien-yen, enn Nga Shan, W. T. Tsang 27415, Dec. 23-29, 1936, 10 ft. high,
fairly common, in thickets, es light yellow, fragrant; Tien-yen, Ho Yung Shan,

W. T. Tsang 30726, Oct. 13 — Nov. 22, 1940. Kwangtung, Kwangsi, Hainan, and
southern Yunnan. New to i on
EBENACEAE

Diospyros Linnaeus
Diospyros Morrisiana Hance, Walp. Ann. 3: 14. 1852-53, Jour. Bot. 18: 299. 1880.

Inpo-Cuina: Tonkin, northeast of Mon-cay, Pac-si, W. T. Tsang 26973, Oct.
1-8, 1936, 10 ft. high, fairly common, in thickets on dry clayey oe fruit yellow;
Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsang 27313, Nov — Dec. 2, nase
10 ft. high, fairly common, in thickets, fruit yellow; Tien-yen, Kau Ne Shan, W.
Tsang 27385, Dec. 13-22, 1936, 15 ft. high, fairly common, in thickets, on steep ee
fruits yellow: ; same locality, W. T. Tsang ne Sept. 23 — Oct. 7, 1940; Dam-ha, Sai
Wong Mo Shan, Lung Wan Village, W. Tsang 29809, May 18 — oe 5, 1940.
Kwangtung, Kwangsi, Fukien, Chekiang, ee New to Indo-Chi
Diospyros potingensis Merr. & Chun, Sunyatsenia 5: 164. 1940.

Inpo-Cuina: Tonkin, northeast of Mon-cay, Pac-si, W. T. Tsang 26925, Oct.
1-8, 1936, 10 ft. high, fairly common, in thickets, on dry sandy soil, fruits yellow:
Ha-coi, Chuk-phai, Taai Wong Mo Shan , W. T. Tsang 27159, Nov. 1-9, 1936, 28996,
May 3-10, 1939, woody, 20 ft. high, fairly common, in thickets, on dry sandy soil, fruits
yellow; Dam- ha, Sai Wong Mo Shan, Lomg Ngong Village. W. T. Tsang 30191, July
18 — Sept. 9, 1940. Hainan. New to Indo-China

STYRACACEAE
Rehderodendron Hu
Rehderodendron indochinense sp. nov.
Arbor circiter 13 m. alta, ramulis glabris subcinereis; foliis paren
glabris utrinque concoloribus elliptico-oblongis, 7-10 cm. longis, 3-3.

370 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXIV

latis, falcato-acuminatis, basi anguste acutis, margine minute distanter
glanduloso-denticulatis, nervis lateralibus utrinsecus 6-8 curvatis circiter
ad marginem arcuato-anastomosantibus, cum venulis reticulatis eyo
perspicuis petiolis 5-7 mm. longis glabris: paniculis axillaribus ad 5 c
ongis, dense substellatim cinereo-tomentosis, pedicellis 6-10 mm, a
bracteolis ad basim pedicelli lanceolatis, cinereo-tomentulosis, m.
longis, deciduis; calyce campanulato 4 mm. alto, 5-dentato, dentibus tri-
angularibus acuminatis; corolla 5-partita, lobis elliptico- oblongis obtusis,

.4 cm. longis mm, latis, utrinque substellatim cinereo-tomentulosis ;
staminibus 10, corollae tubo adnatis, exsertis, alternis longioribus, stylis
cinereo- -tomentulosis, 1.5 mm. longis, ‘stigmate ‘capitato: ovario 5-loculari;
fructibus glabris magnis Rete eid oblongis, 7.5 cm. longis, 2.8 cm. crassis,
apice planis vel leviter depressis, in sicco rubro- brunnescentibus, 10-cos-
tatis, exocarpio duro 1 mm. cCrasso, mesocarpio 8-10 mm. crasso inter
processos endocarpii fibroso-spongioso, endocarpio lignoso 8-10 mm
crasso, processis circiter 10 longitudinalibus 8-10 mm. latis et 1 mm.
crassis ornato; seminibus oblongo-linearibus, 5 cm. longis.

Inpo-Cu1na: Tonkin, Chapa, A. Pételot 6258 (TypE, flowering specimen), in
forests, alt. about 1500 m. , February 1931; Ha-coi, Chuk-phai, Taai Wong Mo Shan,
W. sang 27094, Oct. 23-31, 1936 (fruiting specimen), a tree 40 ft. ‘high, fairly
common, growing in thickets, on dry loamy soil, fruit brownish yellow

This species is allied to Rehderodendron kwangtungense Chen: R. Hui
Chun, and R. praeteritum Sleumer, all from Kwangtung, but it may be
distinguished by its entirely glabrous, much smaller leaves and the larger,
longer, and relatively narrower fruits. This is the second species of the
genus known from Indo-China. The other one is R. macrocarpum Hu, a
species extending from Szechuan through Kweichow to Kwangsi, which: § is
represented by A. Pételot 4767 from Tonkin.

3

Huodendron Rehder
Huodendron parvifolium sp. nov.
arva circiter 7 m. alta inflorescentiis exceptis eae ramulis
wpe teretibus; foliis chartaceis oblongo-ellipticis, 5.5—7.5 cm. m. longis,
cm. latis, utrinque subopacis subconcoloribusque, distincte acumi-

distinctis, venis tertiariis transversis, utrinque subconspicuis; petiolis
1-1.3 cm. longis glabris in sicco nigris; floribus ignotis; infructescentiis
axillaribus puberulis, racemosis, a cm. longis; fructibus ovoideis
cinereo-puberulis, 7.5 mm. longis, 5 mm. latis, loculicide dehiscentibus,
pedicellis 0.7-1 cm. longis, manifeste recurvis; calycibus persistentibus
puberulis, aan ak 4 mm. longis, minute 5-dentatis; seminibus
brunneis, 2—3 mm. longi

INDO-CHINA: nkin, ie of Mon-cay, Pac-si and vicinity, W. T. Tsang
26886 (TYPE), tae a 30, 1936, a small tree. 20 ft. high, fairly common, in thickets, on
dry clayey soil, fruit light gray.

This is the second species of this characteristic Chinese genus to be
known from Indo-China. It differs from the other Indo-Chinese species
of the genus in its smaller leaves, which are dull on both sides, in the rela-
tively larger fruits, which are racemosely arranged in short, simple infruc-
tescences, and in the very small calyx teeth.

1943 ] LI, NOTES ON THE FLORA OF INDO-CHINA 371

Styrax Linnaeus
Styrax argentifolius sp. nov.

Frutex circiter 3-5 m. altus, ramulis novellis dense adpresse furfuraceo-
lepidotis, indumento cinereo- brunn neis; ramis gracilibus glabris; foliis
firmiter chartaceis, supra glabris viridibus opacis, subtus dense minute
adpresse lepidulotis ee avn ento cinereo-albido nitidis, oblongis, 7-17
cm. longis, 2—5 cm. latis, longe acuminatis, basi acutis, margine integris,
nervis lateralibus ae 6-8, supra subconspicuis, subtus_ elevatis
prominentibus, curvatis, arcuato- anastomosantibus, venis tertiariis dense
reticulatis, utrinque subconspicuis; petiolis 0.7-1 cm. longis, adpresse
furfuraceo-lepidotis; inflorescentiis ignotis; infructescentiis axillaribus vel
terminalibus, ad 6 cm. longis, dense adpresse furfuraceo- lepidotis; ee
subovoideis distincte rostratis, ad 2.5 cm. longis et 1.8 cm. crassis, den
adpresse lepidulotis, calyce deciduo minuto, a6 mm. longo, te eee
a eae ae extus sae lepidoto.

NDO-CH Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsang
i Oct. pu 31, 1936, ooo July 1-13, 1939, 10-16 ft. re fairly common, in
thickets, on moist sandy soil, ea dirty white to gray; Dam a oe Mo Shan,
Lomg Ngong Village, W. T. Tsang 30238 (tyre), July 18 — 3

This species ee tines. as suberifolius Hook. & i st southern
China, but may be readily distinguished by its appressed furfuraceous-
lepidote indumentum never being stellate, and by its larger, distinctly
rostrate fruits.

SYMPLOCACEAE
Symplocos Jacquin
Symplocos Delavayi Brand, Repert. Nov. Sp. 3: 218. 1906.

Replies eaae Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. 7. Tsang

206, Nov. 10-17, 1936, Saniall tree, 20 ft. high, fairly common, growing in thickets,
rine black. Yunnan, sarneneete Tibet, and northern Burma. New to Indo-China.
Symplocos indochinensis sp. nov. Subgen. Hopea, § Bobua, Lodhra.

Arbor 5-7 m. alta, ramis cinereo-brunneis, ramulis novellis ferrugineis,
longe patule villosis, ramis vetustioribus saree foliis chartaceis breviter
petiolatis oblongo- ellipticis, 7-12 cm. lon 2—4 cm. latis, acuminatis,

acutis, margine minute Efe denticulatis ‘vel integris,
utrinque eticbae levibed supra parce pilosis vel glabrescentibus haud
nitidis, subtus consperse villosis, costa supra impressa, nervis lateralibus
utrinsecus circa 4 vel 5 arcuato-anastomosantibus, supra impressis, subtus

subtus elevatis; petiolis brevibus ad 5 mm. longis, dense patule
ferrugineo-villosis; inflorescentiis axillaribus glomeratis paucifloris sessi-
libus, axillaribus et in axillis defoliatis in ramulis annotinis; floribus sessili-
bus, bracteis late ovatis dense pubescentibus, 1 mm. rie calycis tubo
crasso circa 1 mm. longo, lobis late ovatis pubescentibus, 1 mm. longis:
petalis 5 oblongis, 3 mm. longis, utrinque glabris; staminibus circa 25,
filamentis ume liberis glabris, circa 4 mm. longis; disco annulari
cinereo-pubescente, ovario 3-loculari, stylo 4 mm. longo; fructibus glo-
bosis (ee ee 6 mm. diametro, calyce persistente.

Inpo-Cuina: Tonkin, northeast of Mon-cay, Pac-si, W. T. Tsang 26902 (TYPE),
Sept. 27-30, 1936, a small a 17 ft. high, fairly common, in thickets on ae clayey
soil, flowers white, fragrant; Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsang

372 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

27027, Nov. 10-17, 1936, a tree 20 ft. high, fairly common, in thickets, fruits heron
ak Chan Uk Village near Chuk-phai, Taai Wong Mo Shan, W. 7. Tsang 29130,
21-31, 1939, a tree 15 ft. high, fairly common, in thickets, on dry cea soil,
ets aie ae ro Sai Wong Mo Shan, Lung Wan Village, W. 7. Tsang 30047,
Mz: — July 5, 1940.
aa ee bee of the branches and leaves, and in the sessile glomeru-
late inflorescences, this species is apparently close to S. yunnanensis Brand.
It differs from the ‘latter by its thinner, much shorter, and relatively broader
leaves. Most of the inflorescences are in the ails of fallen leaves, and
the smooth, globose sessile fruits are all borne on the older leafless branches,

OLEACEAE
Linociera Swartz
Linociera verticillata Gagnep. Bull. Soc. Bot. France 79: 788. 1932 [1933], et in
Lecomte, Fl. Gén. Indo-Chine 3: 1072. f. 122, 1. 1933

In HINA: Tonki Tien-yen, Ho Yung Shan and vicinity, W. T. Tsang
ie, a 13-22, 1940. pee Newt o Tonkin

Olea Linnaeus
Olea cordatula sp. nov,

Arbor parva circiter 10 m. alta, ramis ramulisque dense breviter pube-
scentibus, consperse lenticellatis; foliis amplis rigidis coriaceis subsessili-
bus vel brevissime petiolatis late oblongo-oblanceolatis, 14-29 cm. longis,
4.5-9.5 cm. latis, acuminatis, basi anguste cordatis, cone remote
dentato-serratis, supra costa nervisque valde impressis exceptis glabris,
in sicco olivaceo- brunneis, subtus paullo pallidioribus perspicue breviter
molliter pubescentibus, costa supra impressa subtus distincte elevata,

)
dense pubescente, supra canaliculato; inflorescentiis paniculatis axillari-
bus, ad 6 cm. longis et 2.5 cm. latis, molliter pubescentibus, floribus (im-
maturis) minutis, pedicellis 1 mm. longis; calycibus 0.5 mm. altis ad
medium 4- lobatis, lobis ovatis acutis extus pubescentibus; corolla pallide
lutea, 1 mm. longa, 1.5 mm. diametro, ad medium lobata, lobis acutis
valvatis margine revolutis; antheris subsessilibus ad 0.5 mm. longis;
ovario ovoideo, stylo brevi, stigmate inconspicuo ; fructibus ovoideis nigris
glabris, 1-1.2 cm. longis, 5—6 cm. crassis.

Inpo-Cutna: Tonkin, Ha-coi, Chuk-Phai, Taai Wong Mo Shan, Cha
Village, W. T. Tsang 29241 (rvPE), on 10-22, 1939, a tree pi - high, fairly common,
in thickets, on dry clayey soil, flowers pale yellow; — ha, Sai Wong Mo Shan,
Lung Wan Village, W. T. Tsang 30081, May 18 — July 5, 1940, with young fruits;
Dam-ha, Sai Wong Mo Shan, Lomg Ngong Village, W. T. ae 30170, July 18 — Sept.
9, 1940, with mature fruits.

This species is characterized by its large broadly oblong-oblanceolate
rigid coriaceous distinctly toothed leaves, which are narrowly cordate at
the base, nearly sessile, and softly pubescent beneath. The flowers, prob-
ably not fully developed, show the corolla divided to about the middle.
It seems to be more appropriate to refer this species to Olea rather than to
Linociera, although it suggests no close relationships to any of the described
species of either genus.

1943] LI, NOTES ON THE FLORA OF INDO-CHINA 373

Ligustrum Linnaeus
Ligustrum retusum Merr. Lingnan Sci. Jour. 14: 49. 1935.

ndo-Cuina: Tonkin, Tien-yen, Ho Yung Shan ie vicinity, W. T. Tsang
30644, Oct. 13 — Nov. 22, 1940. Hainan. New to Indo-Chin
RUBIACEAE
Xanthophytopsis Pitard
ear Balansae Pitard in Lecomte, Fl. Gén. Indo-Chine 3: 90. fig. 10,
1922; Chun & How, Sunyatsenia 4: 13. pl. 4. 1939
INDO- oe Tonkin, Tien-yen, Kau Nga Shan and vicinity, W. T. Tsang
30583, Sept. 23 — Oct. 7, 1940; Tien-yen, Ho Yung Shan and vicinity, W. T. Tsang
30704, Oct. 13 — Nov. 22, 1940.

The second collection of this new genus from Tonkin. This species has
also been recorded from Kwangtung, near the Indo-Chinese border.
Xanthophytopsis kwangtungensis Chun & How, Sunyatsenia 4: 14. pl. 5. 1939.

Inpo-Cuina: Tonkin, northeast of Mon-cay, Pac-si and vicinity, W. T. Tsang
26293, Oct. 1-8, 1936, a shea 2 ft. high, abundant, growing in thickets on dry clayey
soil, Wowes wie: Gileless: Ha-coi, Chuk-phai, Taai Wong Mo Shan and vicinity,
W. T. Tsang 27037, Oct. 16-22, 1936, 29015, May 3-10, 1939, Sours Aug. 1-31, 1939,
a shrub 2 ft. high, fairly common, growing in thickets, on dry clayey soil, flowers ‘white,
fruits brown; Dam- A Sai Wong Mo Shan, Lung Wan Village, W. T. Tsang 29880,
May 18 - Fas. 194

This species, ae second one of the genus, was originally nao from
western Kwangtung, near the Tonkin border. New for Indo-Chin

Randia Linnaeus

Randia Pamseaia Champ. ex Benth. in Hook. Jour. Bot. Kew Gard. Misc. 4: 194.

INDo- in iwA: Tonkin, Dam-ha, Sai Wong Mo Shan, Lomg Ngong Village,

. T. Tsang 20420, . ly 18 — Sept. 9, 1940. This is a fruiting specimen; its leaves
are larger than are eos of the Chinese form. Kwangtung. New to Indo-Chi
artes acuminatissima Merr. Philip. Jour. Sci. 15: 259. 1919.

NbDO-CuHINA: Ton kin, Dam-ha, Sai Wong Mo Shan, Lung Wan Village, W. T.

ran 29904, May 18 — fie 5, 1940. China: Kwangtung, Hainan. New to Indo-
Chin

Mussaenda Linnaeus

Mussaenda erosa a in Hook. Jour. Bot. Kew Gard. Misc. 4: 193. 1852.

kin, Tien-yen, Kau Nga Shan, W. T. Tsang 27461, Jan. 1-9,
1937, climber, 7 ft. high, fairly common, growing among scattered shrubs, fruits black;
same locality, W. T. Tsang 30540, Sept. 23 — Oct. 7, 1940; Ha-coi, Pani Wong Mo
Shan, W. T. Tsang 29071, May 11-20, 1939, 29550, Sept. 11-23, 1939, climber, 7-10 ft.
high, fairly common, growing in thickets, feults yellow; Damc-ha, eal Wong Mo Shan,
Lung Wan Village, W. T. Tsang 29990, May 18 — July 5, 1940. Southern China.
New to Indo-China
Mussaenda longipetala sp. nov.

Frutex scandens, ramulis teretibus perspicue subpatule ciliato-villosis ;
foliis NET Se ste vel submembranaceis, in paribus aequalibus, oblongo-
ovatis vel elliptico-ovatis, 8-12 cm. longis, 3.5—6 cm. latis, i
tis, basi late acutis vel subrotundatis, utrinque perspicue molliter villosis,
nervis lateralibus utrinsecus 6-8, curvato- adscendentibus, utrinque con-
spicuis, venis tertiariis subconspicuis vel inconspicuis; petiolis 0.5-1 cm

374 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

longis dense villosis; stipulis linearibus 6-8 mm. longis caducis; inflore-
scentiis terminalibus cymosis compactis circiter 5 cm. longis, breviter ramo-
sis, perspicue patule albido-villosis vel ciliatis, bracteis bracteolisque lineari-
bus 5 mm. longis, floribus sat numerosis subsessilibus confertis; calycis tubo
crasso, 2 mm. longo, dense ciliato, lobis normalibus 4 vel 5 linearibus, 6-8
mm. longis, 1-2 mm. latis, dense 'villosis, longe acuminatis, uno interdum
iedak ovato, 4.5-5.5 cm. longo, acuminato, utrinque longe consperse
villoso, stipitato; stipite perspicue ciliato ad 2 cm. longo, nervis 5—
corollae tubo circiter 3 cm. longo et 2 mm. lato, su i
extus perspicue villoso, intus superne leviter pubescente, lobis 5 lanceolatis,
1.2 cm. longis, 1.5 mm. latis, longe acuminatis, extus villosis: staminibus
medio tubi affixis, antheris 4 mm. longis, basi bifidis; stylo tubi longitudi-
nem aequante, glabro, apice vix lobato

Inpo-Cu1na: Tonkin, Dam-ha, -. Ww ong Mo Shan, Lung Wan Village, W. T.
Tsang 29811 (TYPE), May 18 - July 5,

This species is near M ee ell Pierre, differing in the leaves
being mostly rounded at the base, the much smaller petaloid sepals, the
much narrower and longer corolla-lobes, and the stamens inserted near the
middle of the corolla-tube, which is slender throughout and only slightly
enlarged from the insertion of the anthers upward.

Urophyllum Wallich
Urophyllum chinense Merr. & Chun, Sunyatsenia 2: 19. pl. 10. 1934.

Inpo-Cutna: Tonkin, Ha-Coi, Taai Wong Mo Shan and vicinity, W. T. Tsang
27220, Nov. 10-17, 1936, 29219, June 10-22, 1939, 29516, Sept. 1-10, 1939, a shrub
5-7 ft. high, fairly common, in thickets, flowers white, fragrant, fruits yellow or black;
Tien-yen, Kau Nga Shan, W. T. Tsang 27434, Dec. 23-29, 1936, a shrub 5 ft. high,
abundant, in thickets, feuiits yellow; Dam- Sai Wong Mo Shan, Lung Wan Village,
W. T. Tsang 30094, May 18 — June 5, 1940; Tien-yen, Ho Yung Shan, W. T. Tsang
30673, Oct. 13 — Nov. 22, 1940, Also hae by the following collections from
Kwangsi Province: W.T. Tsang 23879, 24055, 24690. It was originally described from
Kwangtung specimens. New to Indo-China

Gardenia Ellis
Gardenia stenophylla Merr. Philip. Jour. Sci. 19: 678. 1922.

Inpo-Cutwa: Tonkin, Ha-coi, Chuk-phai, Taai Wong Mo Shan, W. T. Tsang
27063, Oct. 23-31, 1936, nn May 3-10, 1939, 29019, May 11-20, 1939, 29598, Sept.
11—23, 1939, a shrub 1.2-2 ft. high, fairly common, growing in thickets, in sandy soil,
flowers white, fragrant, fruits yellow; Dam-ha, Sai Wong Mo Shan, W. T. Tsang
29820, May 18 — July 5, 1940, 30402, July 18 — Sept. 9, 1940. Hainan. New to
Indo-China.

Psychotria Linnaeus
Psychotria rubra (Lour.) Poir. var. lanceolata var. nov

A typo speciei differt foliis oblongo- aneeolats = 16 cm. longis, 2-3
cm. latis, apice longe acuminatis, basi longe a

Inpo-Cu1na: Tonkin, Ha-coi, Taai Wong Mo pel Chan Uk Village near
Chuk-phai, W. T. Tsang 28959 (type), May 3-10, 1939, a shrub 4 ft. high, abundant,
growing in thickets on sandy soil, flowers pale yellow, fragrant.

ARNOLD ARBORETUM

HarvARD UNIVERSITY.

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXIV OCTOBER, 1943 NuMBER 4

PLANTS OF COAHUILA, EASTERN CHIHUAHUA, AND
ADJOINING ZACATECAS AND DURANGO, II

IvAN M. JOHNSTON

TYPHACEAE
Typha truxillensis H.B.K. Nov. Gen. et Sp. 1: 68 (1815).
CHIHUAHUA: 3 mi. west of Camargo, White.
Widely distributed in our area along the rivers and elsewhere about per-
manent water.
NAIADACEAE
Naias guadalupensis (Spreng.) Morong, Mem. Torr. Bot. Cl. 3: 60 (1893).
CoaHuiLa: Monclova, in the river, White 1768. CHIHUAHUA: 3 mi. west of
Camargo, White 2279,
A submerged aquatic, widely distributed in America.

POTAMOGETONACEAE

Zannichellia palustris L. Sp. Pl. 969 (1753).

Coanutta: Monclova, Marsh 1688.

An aquatic of saline waters, world-wide in distribution.
Ruppia maritima L. Sp. Pl. 127 (1753).

CoauuILa: Laguna de Jaco, washed up on the beach, Stewart & Johnston 1978.

A nearly cosmopolitan species, usually in saline waters.
Potamogeton pectinatus L. Sp. Pl. 127 (1753).

Collected in ponds in the Rio Grande bottoms, near San Elizario (Wright 1895).

Widely distributed over the world in brackish waters.
Potamogeton illinoensis Morong, Bot. Gaz. 5: 50 (1880).

CoanuiLta: Monclova, Marsh 1672, det. by E. C. Ogden.

Widely distributed in the United States and ranging south to Central
America.
Potamogeton nodhaue Poir. in Lam. Encyc. Suppl. 4: 535 (1816).

a Hechiceros, Posegay Encampanada, along creek, Stewart 196;

Ojo Caliente, Oct. ic or Thurber

An aquatic, nearly aoe in Paiste eueion:

376 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

ALISMACEAE
oo cordifolius (L.) Gris. Abh. K. Ges. Wiss. Goettingen 7: 257 (1857).
Torreon, periodically flooded ono 1898, Palmer 466. CHIHUAHUA:

Saas a pond, Oct. 11, 1852, Thurber 8
Lophotocarpus calycinus (Engelm.) J. G. Smith, Lophot. U.S. 3 (Sept. 1899).

CoauuiILA: Hermanas, Marsh 2260.

Eastern United States west to South Dakota and New Mexico, and south
in Coahuila
Lophotocarpus fluitans (Engelm.) J. G. Smith, Lophot. U.S. 1 (Sept. 1899).

The type of this species, of southern New Mexico and trans-Pecos Texas,
was collected by Wright (nos. 1899 and 679) in ponds in the bottom-land
of the Rio Grande near San Elizario, Texas.

Sagittaria longiloba Engelm. ex Torr. Bot. Mex. Bound. 212 (1859).

This species has been repeatedly collected in the wet bottom-lands along
the Rio Grande between El Paso and old Fort Quitman, Texas. It ranges
from central United States south into Mexico. The type came from near
San Elizario, Texas.

GRAMINEAE
— page ona (Fourn.) Scribn. & Merr. U. S. Dept. Agric., Div. Agrost. Bull.
24:30 (Jan. 1901); Scribn. Bull. Torr. Bot. Cl. 28: 246 (April, 1901).
AHUILA: Saltillo, 1898, Palmer 5, 266; Parras, May 15, 1847, Gregg. ZACATECAS:
Valley 15 km. west of Concepcion del Oro, Stanford et al. 556

A common, chiefly ruderal species of central Mexico. Collections have
been generally identified as B. laciniatus Beal, but that is a montane plant
obviously distinct from the weedy species concerned here.

Bromus sp.

CHIHUAHUA: Sierra Organos, 1937, LeSueur 211 in pt.

“The collection is similar to the plant of Arizona and New Mexico referred
to B. carinatus.

Bromus sp.

CoAHUIL Mesa Grande, 40 km. northwest of Hac. Encantada, common in

eee Stewart 1633; Hillcoat Mesa; west of Encantada Ranch, July 25, 1938,
Marsh 1

Lae to the plants of the Chisos and Davis Mountains of Texas pass-
ing as B. marginatus and B. polyanthus.

Bromus anomalus Rupr. ex Fourn. Mex. Pl. 2: 126 (1886).

CoanvuILa: Sierra del Carmen, Aug. 26, 1936, Marsh 628; Hillcoat Canyon, west
of Buena Vista Ranch, July 13, 1938, Marsh 1309; trail between Encantada Mesa and
Fresno Mesa, July 20, 1938, Wh 1399; Sierra Glens Marsh 1945, 1947; Carneros
Pass area, July 1880, Palmer 1372 ene del Pino, La Noria, in shaded arroyo and

tera, rocky arroyo in oak belt, Johnston 8926. CHTHUAHUA: Sierra Organos, 1937,
LeSueur 211 in

Widely distributed in western United States and south to southern
Mexico. A very variable species, particularly in the amounts and distri-
bution of indument. The material from the Sierra del Pino and the Sierra

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 377

Madera represents a hairy robust form with broad leaves and a large
drooping panicle.
Brachypodium mexicanum (R. & S.) Linx, Hort. Berol. 1: 41 (1833).

CoaHvutta: Sierra del Carmen, Aug. 26, 1936, Marsh 613; Sierras Negras, 9 km.
south of Parras, Stanford et al. 230

Coahuila and Nuevo Leon south to Central America. A very variable
species with forms differing greatly in appearance, apparently even in a
single locality. The cited collections have very narrow leaves and short-

awned lemmas.

Festuca ligulata Swallen, Am. Jour. Bot. 19: 436 (1932).

Coanurma: Mountains 24 km. northwest of Fraile, Stanford et al. 405.

A coarser, densely tufted plant, with the spikelets larger than in the
type of the species from the Guadalupe Mts., Texas, but a aes better
referred to F. ligulata than to the more northern F. Thurbe

Poa sp.
COAHUILA: at ma 1905, Purpus 1112; Sierra de Parras, April 1905, oC
1146; Sierras Negras, 9 km. south of Parras, eee et al. 167. Zacatecas: 15 km

west of Concepcion naa Oro, Seaford et al. 477

I am unable to name this species. The collection from the Sierras Negras

has been identified as P. involuta Hitchc.
Poa sp.

CoAHUILA: 6 mi. east of Saltillo, 1880, Palmer 1366.

The cited specimen has been identified as P. Ruprechtti Peyr.
Poa Bigelovii Vasey & Scribn. Descr. Cat. Grasses U.S. 81 (1885).

Coanuita: Saltillo, 1905, Palmer 532.

Oklahoma and western Texas to southern California and south into
northern Mexico. The species has been repeatedly collected about El Paso,
Texas, and is to be expected in adjacent northern Chihuahua.

Poa annua L. Sp. Pl. 68 (1753).

CoaHuILaA: Saltillo, 1905, ine a Saltillo, 1898, Palmer 6; Saltillo, Gregg.
CuinvuAHuaA: Chihuahua, 1908, Palm

A European grass, widely ae in the United States and Mexico.
Eragrostis obtusiflora Scribn. U. S. Dept. Agric., Div. Agrost. Bull. 8: 10. t. 5 (1897).

CHIHUAHUA: Margin of Laguna de Santa erie April 20, 1852, Wright 193
(ISOTYPE) ; near Laguna Santa Maria, 1899, Nelson 6

Known only from saline soils in eon Chihuahua, southwestern
New Mexico, and el Arizona. It much resembles Déstichlis
stricta in boner ha
Eragrostis eee Buckl. Proc. Acad. Nat. Sci. Phila. 1862: 97 (1863).

CoaHvuiLta: Don Martin Dam, Harvey 932.

Kansas and New Mexico to Texas and northeastern Coahuila.

Eragrostis reptans (Michx.) Nees, Agrost. Bras. 514 (1829).
Coanuita: Don Martin Dam, mud flats, Harvey 927, 928.
Kentucky to South Dakota and Texas, south into northeastern Mexico.

378 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

pita ieee? (Koel.) Link, Hort. Berol. 1: 187 (1827).

CoaH Near Don Martin Dam, Harvey 948; Sabinas, Nelson 6823 (US);
Saltillo, bene’ Palmer 389; valley-floor east of Puerto Caballo, near bushes by
ephemeral charco, Johnston 8330; Sierra Cruces, 4 mi. west of Santa Elena, black Joam
on flats, Stewart 828; north of Sierra Cruces, west of San Rafael, on sabaneta, Johnston
te Muller 1038; 7 mi. south of Jaco, about mogote, Johnston & Muller 1110. Cut-

AHUA: Rancho El Pino, southeast of Sierra Rica, slopes, Stewart 2399; Piramide,

south of Camargo, arroyo, Harvey 1377. Duranco: Tlahualilo, barren hills, Pittier
475 (US)

A European weed, widely introduced in America. Where I have seen
this plant in Coahuila and Chihuahua, however, it behaves like an indige-
nous species, associating with indutable native species and growing with
them in remote undisturbed places where a single introduced species is cer-
tainly not to be expected.

Eragrostis diffusa Buckl. Proc. Acad. Nat. Sci. Phila. 1862: 97 (1863).

Coanutira: Saltillo, 1898, Palmer 811, 812; 7 mi. south of Jaco, about a mogote,
Johnston & Muller 1109; Torreon, low ground along the Rio Nazas, 1898, Palmer
5 ,

A

Almagre, wet sand in canyon, Johnston & Muller 1214; Chihuahua, Pringle 416,
LeSueur 78; 3 mi. north of Charca Piedra, under bushes on silty plain, Johnston 7929;
Camargo, banks of the Rio Conchos, Harvey 1404; 50 km. west of Camargo, Harvey
1414; Jimenez, banks of the Rio Florido, Harvey 1313

Central Texas to Arizona and south into aarthers Mexico.
pases sp.

COAHUILA ere 1898, Palmer 376; Sierra del Pino, La Noria, meadows and open
hillsides, Stewart 1205. HIHUAHUA: Rancho El Pino, southeast of Sierra Rica,
slopes, Stewart 2400; canyon west of Organos, damp gravelly arroyo, Stewart &
Johnston 2081.

Closely related to E. diffusa and perhaps only a form of it, differing in
its dense elongate inflorescence. The branches of the panicle are very short
and strict and bear crowded strictly ascending spikelets. In typical F.
diffusa the panicle is open, with elongate spreading branches. I have seen
the plant from Coahuila, Chihuahua, trans-Pecos Texas, and New Mexico.
Eragrostis mexicana (Hornem.) Link, Hort. Berol. 1: 190 (1827).

: Monclova, 1939, Marsh 1834; ape 1898, Palmer 409—412; Saltillo,
1905, Palmer 710; San Lorenzo Canyon, 6 mi. southeast of Saltillo, about old zoat
pen in canyon, 1904, Palmer 398; Choice Grande, es mi. southeast of Saltillo, 1904,
Palmer 334, 335; Parras, 1880, Palmer 1367.

Texas to Arizona and south through Mexico to South America.
Eragrostis neomexicana Vasey, Contr. U. S. Nat. Herb. 2: 542 (1894).

CoanvutILa: Sierra del Carmen, El Jardin del Sur, Sept. 3, 1936, Marsh 766.

Western Texas to Arizona.

Eragrostis caudata Fourn. Mex. Pl. 2: 115 (1881).

Eragrostis Palmeri Wats. Proc. Am. Acad. 18: 182 (1883).

Eragrostis erosa Scribn. ex Beal, Grasses No. Am. 2: 483 (1896).

U Villa Juarez, 1880, Palmer 1368 (type of FE. Palmeri); Sierra del Pino,
La Noria, arroyo banks, Johnston é Muller 664; Sierra Cruces, 5 mi. south of Santa

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, II 379

Elena, rocky flat among bushes, Johnston & Muller 1378; western base of Picacho del
Fuste, brushy rocky slope, Johnston 8413; Sierra Madera, Cahon Charretera, edge of
thickets on rocky flat, Johnston 9062; west end of Sierra Fragua, Aguaje Pajarito,
rocky ridge, Johnston 8676; high rocky west ridge of Sierra Fragua, north of Puerto
Colorado, Johnston 8760; Sierras Negras, 9 km. south of Parras, Stanford et al. 135.
CHIHUAHUA: Sierra Santa Eulalia, Oct. 1885, Pringle 415 (isotype of EF. erosa)

Southern and western Texas south into Chihuahua, Coahuila, and north-
ern Tamaulipas. An isotype of E. caudata (from Matamores) at the Gray
Herbarium is evidently conspecific with the type of E. Palmert. From iso-
typic material of E. erosa they differ only in their slightly smaller spikelets.
Eragrostis intermedia Hitchc. Jour. Wash. Acad. 23: 450 (1933).

CoauuiLa: Allende, Marsh 2237; Sierra del Carmen, Sept. 2, 1936, Marsh 800;
Santo Domingo, igneous hill, Wynd & Mueller 476; Palm Canyon near Muzquiz,

Marsh 1510; Rancho Santa Teresa, south of Castahos, Wynd & Mueller 200; Saltillo,
1909, Arséne; Saltillo, 1898, Palmer 408; Sierra del Pino, La Noria, hillsides and along
arroyo, Tounste n & Muller 498, Stewart 1203.

Central Texas west to Arizona and south into northern Mexico. A
densely tufted perennial with tall erect culms. Habitally very similar to
E. caudata but differing in having hairy nodes in the panicle, spreading
spikelets, and a more open usually proportionately broader inflorescence.
Some of the collections from northeastern Coahuila have rather small
spikelets and approach FE. lugens Nees, a widely distributed species in
tropical America, which extends northward through Mexico to eastern

Monanthochloé littoralis Engelm. Trans. Acad. Sci. St. Louis 1: 437 (1859).

COAHUILA: 4 mi. west of Cuatro Cienegas, common in strongly saline and gypseous
soil on flat, Johnston 7129.

This species has been known only from salt marshes about the Gulf of
Mexico and along the Pacific coast of Mexico and adjacent United States.
The material from Cuatro Cienegas is sterile, but in all vegetative details
it agrees with material from coastal salt marshes. The plant grows only a
few centimeters high, from well-developed scaly rhizomes, and covers small
areas of very saline gypsiferous soil.

Distichlis texama (Vasey) Scribn. U. S. Dept. Agric., Div. Agrost. Cir. 16: 2 (1899).

Coanutta: Torreon, large masses on sandy banks of Rio Nazas, 18-24 inches high,
with runners rooting at the nodes, Oct. 1898, Palmer 507.

A coarse grass with long trailing ae growing in sandy places. The
species has been collected near the R @ Grande at Presidio and Castolon,
Texas, and hence it may be oe in northern Coahuila and nor:
eastern Chihuahua.

Distichlis stricta (Torr.) Rydb. Bull. Torr. Bot. Cl. 32: 602 (1905).

Coanumra: Cuatro Cienegas, saline soil, Johnston 7125, Harvey 1234, Marsh 2080;
Laguna de Jaco, salt flats at south end of lake, Johnston & Muller 1089. CHIHUAHUA:
Laguna de Santa Maria, 1899, Nelson 6416

A salt grass widely distributed over western United States and Mexico.
It has been collected at many stations along the Rio Grande.

380 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Arundo Donax L. Sp. Pl. 81 (1753).

VERNACULAR NAME: Carrizo.

Coan Palm Canyon, near Muzquiz, Marsh 975; Monclova, 1880, Palmer 13-'5 ;
Monclova, Harvey 1132.

Texas to California and southward. A plant of the Old World, now
widely established in the warmer parts of America. Well established along
the Rio Grande and elsewhere about ponds and streams in our area.
Phragmites communis Trin. Fund. Agrost. 134 (1820).

Collected along the Texan bank of the Rio Grande and hence to be ex-
pected in northern Coahuila and Chihuahua. Widely distributed in the
warmer parts of the world.

Melica montezumae Piper, Proc. Biol. Soc. Wash. 18: 144 (1905).

Melica alba Hitchc. Contr. U. S. Nat. Herb. 17: 367 (1913).

AHUILA: San Lorenzo Canyon, 6 mi. southeast of Saltillo, 1905, Palmer 551;
Sierra Mojada, April 19, 1892, Jonas: 482 (US). Curuuanua: Sierra Santa Eulalia,
shaded places, April 6, 1885, Pringle 430 (IsoTYPE) ; Chihuahua, Wilkinson (US).

Edwards Plateau and Big Bend, Texas, south into our area. Piper and
Hitchcock independently named this species, basing their names on the
same group of specimens and indicating the same collections as the type.
The species has been taken to be endemic to our area, but Mr. W. S. Boyle,
who is monographing the genus, has properly indicated, through his identi-
fication of the Gray Herbarium material, that the species is also present in
Texas

Melica nitens Nutt. ex Piper, Bull. Torr. Bot. Cl. 32: 387 (1905).
JILA: — arroyo south and west of Sierra Azul, Buena Vista Ranch, July 8,
1938, ‘Marsh 1230, 1260.
Eastern sisi States west to Kansas and Arizona and south through
eastern Coahuila to Nuevo Leon.
Triodia pulchella H.B.K. Nov. Gen. et Sp. 1: 155 (1816).
VERNACULAR NAMES: Zacate borreguerro; Zacate pelillo; Zacatito.
uILA: Sierra del Carmen, Aug. 29, 1936, Marsh 692; between Santo Domingo
and Piedra Blanca, open country, Wynd & Mueller 496; Cuatro Cienegas, Marsh 2050;

a
Stewart 840; Puerto Ventanillas, south of Las Delicias, limestone slope, Stewart 2967 ;
Parras, 1880, Palmer 1359. CuHimuaHuA: Laguna Santa Maria, Nelson 6414; i-
huahua, LeSueur 11; 20 mi. south of Camargo, Harvey 1399. DuRAnco: Torr
rocky hills, Hiickeock 7547 (US); Tlahualilo, barren hills, Pittzer 480; nie San
Ignacio, Purpus 4616 (US). Zacatecas: Concepcion del Oro, very commo
Palmer 263; valley 15 km. west of Concepcion del Oro, Stanford et al. 560; na
footslopes and hills, Lloyd 89 (US).

Western Texas to southern Nevada and southern California, south to
central Mexico. A common but unobtrusive widely distributed, almost
ubiquitous, grass among the desert scrub on dry slopes and in desert valleys.
A capable xerophyte that flowers throughout the summer and appears to
remain unaffected by long droughts. Although it is frequently locally
abundant, horses and cattle ignore it-if any other food is available. During

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 381

long dry spells the plants appear to become shaggy-white- villous. The
shaggy indument disappears after a rain and is apparently composed of
myriads of hair-like water-soluble crystals which are products of tran-
spiration.
Triodia pilosa (Buckl.) Merr. U. S. Dept. Agric., Div. Agrost. Cir. 32: 9 (1901).
Coanumra: Don Martin Dam, Harvey 933; calcareous soil near Piedras Negras,
Pringle 8306; Puerto Santa Anna, Marsh 942; Yerda Spring, Marsh 296; Hillcoat
Mesa, west ef aeaee ae Fily 25, 1938, Marsh 1448; Canon Espantosa, Sierra
San Vicente, Schroeder
Kansas and sane to Nevada and Arizona, south into northeastern
Mexico.
Triodia avenacea H.B.K. Nov. Gen. et Sp. 1: 156 (1816).

OAHUILA: 3 km. southwest of Fraile, in arroyo, Stanford et al. 331. ZACATECAS:
Concepcion del Oro, ie mesa, 1904, Palmer 280; valley 15 km. west of Concepcion
del Oro, Stanford et al. 5

Northern oe ad Nuevo Leon southward to central Mexico. A
low spreading plant, with stolons.
Triodia grandiflora Vasey, Contr. U. S. Nat. Herb. 1: 59 :
Triodia Nealleyi Vasey, U.S. Dept. Agric., Div. Bot. Bull. 122: ¢. 36 (1891).
COAHUILA: Saltillo, 1898, stony hillside, Palmer 414, 813; Saltillo, 1905, Palmer

Muller 297; a foothills of Sierra Cruces, north of Santa Elena, rocky flats,
Johnston & Muller 1384; west of Santa Elena, sunny hillside, Stewart 827. Cut-
HUAHUA: Sierra Viena rocky slope, Johnston 8093A ; Sierra ue era Aug. 12,
1885, Pringle 406. Zacatecas: Concepcion del Oro, 1904, Palmer

Although I am pees Hitchcock in delimiting this species, I am con-
vinced that it is an aggregate containing several undifferentiated species.
Typical T. grandiflora is a plant with the habit of T. pilosa, having usually
pale well-developed spikelets with acute or subulate lemma-lobes. Triodia
Nealleyi is a species more closely related to T. avenacea and has a spicate
crowded inflorescence, in which the individual spikelets are less obvious,
smaller, usually reddish, and the lemma-lobes elongate-spreading and
obtuse. Both of these re are present in our area. Plants from Chi-
huahua and Arizona are distinguishable, but are unnamed. Certain plants
from the Sierra Guajes, Sierra del Pino, and Sierra Madera are also
separable, but without a name. The aggregate of forms here included
ranges from western Texas to Arizona and south in eastern Mexico to
Hidalgo and Oaxaca.
Triodia texana Wats. Proc. Am. Acad. 18: 180 (1883).

CoaHuiLta: Rio Grande Valley near Piedras a lies 9019; Allende, Marsh
shel 11 mi. south of Allende, ee lined arroyo, Johnston 7017; Verda Spring, Marsh

6; Monclova, 1880, Palmer 1371; near Rancho sel South of Castafios, moist
nn in desert, Wynd & Mueller 178.

382 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Central and southern Texas and northeastern Mexico.
Triodia albescens Vasey, U.S. Dept. Agric., Div. Bot. Bull. 122: t. 33 (1891).
umLa: Rio Paaaal Valley near Piedras Negras, Pringle 9023; 21 mi. south of
Sabinas, pore 704
Texas and ia to Colorado and New Mexico, south into northeastern
Coahuila.
Triodia mutica (Torr.) Scribn. Bull. Torr. Bot. Cl. 10: 30 (1883). .
Sierra del Carmen, Sept. 13, 1936, Marsh 908; Hillcoat Mesa, west of

Encantada Ranch, July 25, 1938, Marsh 2270, northwest of Fronteras, road to
Natadores, silty Larrea rap Johnston 7174; near iano silty bajada, Johnston
7091; Santa Teresa, south of Castahos, Wynd & Mueller 196; 3 mi. east of Cuatro

Cienegas, rocky bajada, psa 7110; between Hac. La Rosa and Hac. Lechuguilla,
dry desert, Wynd & Mueller 62; 14 mi. east of Paila, Shreve & Tinkham 9891; hills
20 mi. west of Saltillo, Shreve & Tinkham 9820; mountains west of Saltillo, 1880,

rocky flat, Johnston 8414. Cuinuanva: Sierra Santa Eulalia, dry calcareous soil,
Pringle 405; arroyo 20 km. south of Camargo, poe 1376. Duranco: Tlahualilo,
barren hills, Pittier 468 (US)

Texas to Arizona and ers into northern Mexico.

Blepharidachne Bigelovii (Wats.) Hack. in DC. Monogr. Phan. 6: 261 (1889).

CoanvuiLa: Several miles west of Buenavista, on road between San Antonio de los
Alamos and Puerto Caballo, banks of gypsiferous shales, small clumps 1-3 inches
broad, pale, frequent, leaves very pungent, Johnston 8309; south of Laguna de Leche,
confined to gypsiferous shales, rigid, erect, local, leaves pungent, Johnston 8615; north-
ern foothills of the Sierra Cruces, about 10 mi. north of Santa Elena, gypsiferous
shales, local, Johnston 9411; east base of Picacho de San José, gypsiferous shales,
Johnston 9401. Texas: Rustler Springs, Culberson Co., 1928, Cory 1238; rocky hills
near Frontera, north of El Paso, in small compact bunches, May 4, 1852, Wright
2028 (tyPE); rocky hills near Frontera, May 4, 1852, Bigelow; Frontera, July 1852,
Parry.

Known only from the collections cited. Where I have seen the plant it
was confined to thin gypsum seams in Upper Cretaceous shales. Since
gypsum is present in the areas in Texas where it has been collected, the
probabilities are that the species is gypsophilous. The plant forms very
strict stiff tufts less than a decimeter tall. The pale rigid leaves have a
pungent tip. Because of its distinctive appearance it is readily recogniz-
able, even in the sterile state.

Cottea pappophoroides Kunth, Rév. Gram. 1: 84 (1829).

Coauutta: Sierra Hechiceros, Cafion Indio Felipe, sides of dry arroyo, Stewart
175; gypsum beds in the valley between La Vibora and Matrimonio Viejo, Johnston
0340; 16 mi. south of Laguna del Rey, gypsum plain, Johnston 7816; Rancho Las Uvas,
Valle Acatita, frequent on gypsum flats, Stewart 2695. pains Chihuahua,
rocky hills near Millers Dam, Sept. 12, 1885, Pringle 420; 20 km. south of Camargo,
Harvey 1395. Duranco: Torreon, rocky hill, Hitchcock Oe (US).

Southern and western Texas to Arizona and south to southern Mexico;
also in South America. The type came from Peru. Although obviously not
confined to gypsum, the species frequently occurs on gypsum in Coahuila,
and where I have seen it, it is locally confined to that substratum. On
gypsum the plants form a very dense caudex of stem-bases and cleisto-

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 383

genes that becomes 3-7 cm. in diameter, although supporting only a rela-
tively few leafy stems. Plants from other substrata develop very loose
caudices or none at all. I can detect no other differences between these
plants.
Pappophorum Wrightii Wats. Proc. Am. Acad. 18: 178 (1883).

Coanuita: Mt. Carmel Canyon, Rio Grande, Parry; Saltillo, in graveyard, 1898,
Palmer 395; near Rosario, about mogote, Jonnie 8823; 42 mi. west of Saltillo,

flat, Johnston 8256; a Berueey La Vibora and sini set Viejo, margin of

ez, banks of the Rio Florido,
Harvey 1329; 6 mi. west of Piloncillo, lava hillside, pi 7876

Central ewe to Arizona and south to Oaxaca; western South America.
As with Cottea, this grass, although not restricted to gypsum, is encoun-
tered most frequently about gypsum beds and usually in greater abundance
there than on other substrata.

oO
oO
aa
i
co
6S
=
=
=
o>
is)
Ss
Co
Do
on ac)
@
4
a
-c
&
=m
co
“as
en!
io)
x
=

Pappophorum mucronulatum Nees, Agrost. Bras. 412 (1829).

CoanuILa: Cuatro Cienegas, Marsh 2048; west of Puerto de las Monjas, mouth o
sae aad ore Saltillo, dry ground near ieee field, Hitchcock 450; Saltillo,

ommon praveyard, 1898, Palmer 377; near La Rosa, Shreve & Tinbham 9904;
eee (ee. Palmer 1

Texas to Arizona oar northern Mexico.

Pappophorum bicolor Fourn. Mex. Pl. 2: 133 (1886).

Coanuita: Don Martin Dam, Harvey 939; 21 mi. south of Sabinas, bank of arroyo,
Johnston 7047; Hermanas, Marsh 1591; Monclova, 1880, Palmer 1362; near Rancho
Santa Teresa, south of Castafios, Wynd & Mueller 203; hills 20 mi. west of Saltillo,
Shreve & Tinkham 9834; on plain 1 mi. south of Ocampo, in mogote, Johnston 8885 ;
7 mi. south of Jaco, near mogote, Johnston & Muller 1107.

Texas to Arizona and south into northeastern Mexico.

Scleropogon brevifolius Phil. Ann. Univ. Chile 36: 206 (1870).

VERNACULAR NAMES: Colo del Zorro; Zacate del Burro.

OAHU Sierra . ae Sept. 13, 1936, Marsh 898; Muzquiz, Marsh 1118;
Saltillo, colonies in depressions on mesas, 1898, Palmer 386; Chojo Grande, 27 m
Souehease of Saltillo, fr places on mesas, a Palmer 340; Parras, Oct.
Palmer 5005. CuiHuauHua: Candelaria, Stearns 254 (US); mae Ahumada, eur
72; Santa Eulalia Plains, 1885, Wilkinson (US); 15 km. savin of Camargo, Harvey
1401, 1402. ZACATECAS: Ccdras, Lloyd 169 (US)

Western Texas to Colorado and Arizona and south to southern Mexico;
Argentina. A common and characteristic grass of the silty valley soils in
Coahuila, especially of the flat area where some run-off temporarily accumu-
lates during rains. Frequently associated with tobosa (Hilaria mutica)
and common about the margin of well-developed tobosa flats. The saba-
netas of Coahuila, level grassy places of varying size scattered in the desert
scrub on the gentle slopes of the major valleys or in the broad open canyons,
are usually dominated by this grass. The plant spreads by rhizomes and

384 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

large areas may be covered by a pure colony of the plant. The awns are
usually stramineous but occasional plants have the awns pink and so permit
an observer to determine the extent of an individual plant in the colony. I
have noted plants covering three or four square meters and so dominating
its particular area that adjoining plants of the species scarcely transgress
upon it. The plants are extremely prolific and during the summer the
female plants are a mass of fruiting inflorescences; although useless as
animal food, they are an attractive element in the desert scene.
ee sp.

UIL Hillcoat ~ west of Buena Vista Ranch, July 13, 1938, Marsh
vies "Sierra Gloria, Marsh

The cited specimens ne to the group of A. arizonicum and A. spica-
tum, but differ from them in their large, elongate, attenuate, awn-tipped
glumes.

Elymus canadensis L. Sp. Pl. 83 (1753).

CoauutLa: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 529; Sierra del
Carmen, Aug. 9, 1936, Marsh 635; Hillcoat Mesa, west of Encantada Ranch, July 25,
1938, Marsh 1440; Saltillo, in orchard, 1898, Palmer 260; — a July 16,
1848, Gregg 263. CurnHuanua: 5 km. west of Camargo, Harvey

Widely distributed in the United States, extending ber into Chihuahua,
Coahuila, and Nuevo Leon.

Sitanion Hystrix (Nutt.) J. G. Smith, U. S. Dept. Agric., Div. Agrost. Bull. 18: 15
(1899).

ILA: Sierra del Carmen, Sept. 8, 1936, Marsh ie Hillcoat Canyon, west of

cacti, spikes nodding, Johnston oe —— 15 km. west £ Concepcion del
Oro, 18-30 inches tall, Stanford et al. 5

Western United States se adjacent Mexico; south along the eastern
Sierra Madre to central Mexic
Koeleria cristata (L.) Pers. Syn. Pl. 1: 97 (1905).

CoauurILa: Sierra del Carmen, Cafon Sentenela, Wynd & Mueller 539; Sierra del
Carmen, Aug. 26, 1936, Marsh 631; Hillcoat Canyon, west of Buena Vista Ranch,
July 13, 1938, Marsh 1315; Hillcoat Mesa, west of Encantada Ranch, July 25, 1938,
Marsh 1445; Sierra Gloria, Marsh 1902, 1944

Widely distributed in the United States; in Mexico extending south in
the mountains to Central America.

Sphenopholis obtusata (Michx.) Scribn. Rhodora 8: 144 (1906).

Coauvutta: Rio Grande Valley at Piedras Negras, April 17, 1900, Pringle 8285.

Widely distributed in the United States and south to southern Mexico.
Trisetum deyeuxioides (H.B.K.) Kunth, Rév. Gram. 102 (1829).

Hitchcock, Contr. U. S. Nat. Herb. 17: 325 (1913), reports a collection
of Pringle (no. 1432) from wet banks of a stream in the Mapula Mts. In
his published diaries Pringle mentions collecting the species in the Mapula
Mts. on Oct. 30, 1886. The species is known from the mountains of
Chihuahua and south to southern Mexico.

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 385

Avena fatua L. Sp. Pl. 80 (1753).

CoanvulILa: Saltillo, old field, April 1898, Palmer 8.

A European weed, widely introduced in the United States and Mexico.
Danthonia mexicana Scribn. Proc. Acad. Nat. Sci. Phila. 1891: 301 (1891).

Coanuita: West base of Picacho del Fuste, coarse tufts on limestone ledges on
north slope, Johnston 8382; Carneros Pass, limestone ledes Sept. 20, 1890, Pringle
3279 (ISOTYPE) ; Sierra de Barras, April 1905, Purpus 1125 in pt.

A rare species, known only from the stations ae above, and from
Tehuacan, Puebla.
Agrostis semiverticillata (Forsk.) C. Chr. Dansk. Bot. Arkiv 43: 12 (1922).

Coanumwa: Muzquiz, Sabinas River, 1936, Marsh 410; Muzquiz, 1938, Marsh L157,
1169; Hermanas, 1939, Marsh 1621; Monclova, 1939, Marsh 1697; Saltillo, 1905,
Palmer 527 ; Saltillo, 1898, Palmer 8006; San Juan de la Vaqueria, May 25, 1847, Gregg
fil, Paras: 1880, Palmer 2019; Pam ras, Feb. 1905, Purpus 1111. Cee:
Chihuahua, low meadows, 1908, Paley 29; Chihuahua, wet river bank in shade, 1908,
Palmer 160; 2 km. west of Jimenez, Harvey 1315.

A species of the Old World, now widely distributed in wet soils in the
warmer parts of America.
Agrostis exarata Trin. Gram. Unifl. 207 (1824).

Coanuita: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 547.

From Alaska south through western United States into northern Mexico.
Agrostis hiemalis (Walp.) B.S.P. Prelim. Cat. N. Y. 68 (1888).

CoanuvuiLa: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 544.

Boreal North America south into the mountains of northern Mexico.
Polypogon monspeliensis (L.) Desf. Fl. Atlant. 1: 67 (1798).

CHIHUAHUA: 2 km. west of Jimenez, Harvey 1317.

A European weed, widely distributed in the United States and northern
Mexico
Polypogon elongatus H.B.K. Nov. Gen. et Sp. 1: 134 (1815).

Coanuita: Saltillo, along water in — in garden, 1898, Palmer 2. CuHrmuAHua:
Presa de Chihuahua, 1936, LeSueur 15

Wet soils from southern ae south through Mexico, reaching South
America.
Lycurus phleoides H.B.K. Nov. Gen. et Sp. 1: 142. t. 45 (1815).

Lycurus phleoides var. glaucifolius Beal, Grasses No. Am. 2: 271 (1896).

CoaHuILa: Sierra del Carmen » Aug. 14, 1936, Marsh 658; limestone hill near Santo

of

13, 1938, Marsh 1274, 1285; Mesa Grande, 40 km. northwest of Hac. oe

HIHUAHUA: rocky hills near Chihuahua, May 28, 1885, Pringle 426 (isotype of var.
glaucifolius); Chihuahua, 1935, og 76. cae Valley 15 km. west of
Concepcion del Oro, Stanford et al.

Ranging from Oklahoma mae western Texas west to Arizona and south

386 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

in Mexico to Guanajuato (the type locality), Hidalgo, and Puebla. The
species is frequent on rocky soils in our area.

Hitchcock, Contr. U. S. Nat. Herb. 17: 305 (1913), reports L. phala-
roides H.B. K. from Cedros, Zac. (Lloyd 179). The report needs verifica-
tion. Perhaps a slender specimen of L. phleoides is involved.
Muhlenbergia biloba Hitchc. Contr. U.S. Nat. Herb. he _ (1913).

Bealia mexicana Scribn. in Hack. True Grasses 103 (189

CurHvuanvua: Hills northwest of Chihuahua, Oct. 7, i Pringle 819 (ISOTYPE).

Known from a few stations in Baja California, Chihuahua, and Durango.
Muhlenbergia texana Buckl. Proc. Acad. Nat. Sci. Phila. 1862: 91 (1863).

CuinvuAHvA: Volcanic hills 20 km. north of Chihuahua, open canyon, in gravel
of stream-bed, Stewart & Johnston 2139; hills northeast of Chihuahua, wet ledges,
Oct. 7, 1885, Pringle 399; rocky hills northwest of Chihuahua, gravel bars of stream,
Oct. 21, 1885, Pringle 400.

Trans-Pecos Texas to Arizona and south along the Sierra Madre, in
Chihuahua and Sonora, to Durango.
Muhlenbergia crispiseta Hitchc. No. Am. Fl. 17: 440 (1935).

CurnuAHuA: Mapula Mts., thin soil of summits, Nov. 11, 1886, Pringle 824.

Known from a few collections in San Luis Potosi and the mountains of
Chihuahua.
Muhlenbergia implicata (H.B.K.) Kunth, Rév. Gram. 1: 63 (1829).

IUAHUA: Portrero Peak, east of Mapula station, rocky banks of stream, Oct. 12,
18.

North through Mexico to Hidalgo, Durango, and Chihuahua.
Muhlenbergia i Scribn. Bot. Gaz. 9: 187 (1884).

Coanutma: Along seepage on limestone ledges at top of escarpment at west side of
Potrero de la Mula, ‘locally abundant, spree Johnston 9250. Curmuanua: Pira-
mide, shaded crevices at base of large rock-masses, Johnston 8122; Sierra Santa
Eulalia, thin dry soil of ledges, 1885, Prats 404.

Trans-Pecos Texas to Arizona and south to Central Mexico.

Muhlenbergia arenacea (Buckl.) Hitche. Proc. Biol. Soc. Wash. 41: 161 (1928).

Coanutta: Sierra del Carmen, Sept. 13, 1936, Marsh 907; northeastern foothills of
the Sierra Cruces, 5 mi. south of San se silty flat in arroyo, shins ston & Muller
1032; 3 mi. east of San José, silty flats, Johnston 8 HIHUAH 50 km. n

of Jimenez, in arroyo, Harvey 1375. Zacatecas: Valley 15 km. went of Concepcion
del Oro, Stanford et al. 517.

Western Texas and adjacent New Mexico south into our area.
Muhlenbergia asperifolia (Nees & Meyen) Parodi, Rev. Fac. Agron. Buenos Aires

6: 117 (1928)

DurANnco: Mapimi, dense masses in bottom of damp arroyo, 1898, Palmer 554.

Western United States south into northern Mexico; also in southern
South America. The species has been collected along the Rio Grande near
E] Paso and is to be expected in northern Chihuahua.
~ glauca (Nees) pe Rep. Sp. Nov. 17: 214 (1921).

OA Sierra de la Paila, Oct. 1910, Purpus 5006; Sierra Madera, Canon

Pe anc gravelly bed of arroyo, one plant, Johnston 8906; Slerta del Pino, La Noria,

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 387
shaded arroyo-bank, gravelly soil, Johnston & Muller 466. Cutnuanua: Rocky
hills northeast of Chihuahua, cold wet ledges, Sept. 28, 1885, Pringle 395.
Trans-Pecos Texas to Arizona and south to central Mexico.
Muhlenbergia Emersleyi Vasey, Contr. U. S. Nat. Herb. 3: 66 (1892).

HUILA: Sierra del Carmen, Aug. 9-26, 6, Marsh 624, 655; trail from southern
extremity of Hillcoat Mesa to Buena Vista ale ane July 27, 1938, Marsh 1511;
Sierra Madera, Cafion Charretera, bed of arro nd y flats, Johnston 8950,

high on canyon side, 1904, Palmer 401; Sierra del Pino, La Noria, Stewart 1208,
Johnston & Muller 462, 587. CutHuanua: Encampanada, Sierra Hechiceros, sunny
open slopes, Stewart 202; Piramide, sheltered crevices about base of rock-masses,
Johnston 8118; Organos, rocky open canyon, Stewart & Johnston 2066.

Trans-Pecos Texas to Arizona and south to Durango and Hidalgo. A
species of the oak-belt, forming coarse clumps in rocky soil, commonly at
the edges of thickets or on sheltered arroyo-banks. The material from
eastern Coahuila has a denser stiffer plumbeous, rather than pinkish,
panicle, and shorter awns than typical M. Emersleyi. This aberrant eastern
material, well exemplified by Palmer 401, may deserve nomenclatural
recognition.

Muhlenbergia lanata (H.B.K.) Hitchc. No. Am. Fl. 17: 459 (1935).

CuIHvuAHUA: Rocky hills northeast of Chihuahua, cool slopes, Oct. 10, 1885,
Pringle 391.

Known from scattered stations in Chihuahua, and from San Luis Potosi
to Puebla.

Muhlenbergia abata sp. nov.

Planta perennis gracilis humilis e rhizomatibus gracilibus oriens; cauli-
bus numerosissimis gracilibus stricte ramosis saepe 1-3 dm. longis rigidulis
laxe decumbentibus vel procumbentibus; vaginis quam internodiis ™%4
brevioribus vel eis non raro subaequilongis, maturitate solum partem infra

medium internodii culmi amplectantibus, margine scabridulis; ligula ad
1 mm. longa apice rotundata basi decurrente; lamina rigidula plus minusve
curvata, 3-6 cm. longa, 1-2 mm. lata | saepe arcte involuta, subtus glabra
supra minute scabr ae paniculis scabridulis 2-4 cm. ongis paucifloris
eee angustis subspicatis infra medium brevissime stricteque pauci-

mosis; spiculis ca. 3 mm. longis strictis acutis elongatis 0.3-5 mm. longe
pedicclintis: glumis subaequilongis hyalinis pallidis obscure uninervatis
acutis 1.8-2 mm. longis; lemmatibus plumbeis obscurissime nervatis
elongatis, supra medium sparse minuteque scabridulis, alibi glabris, apice
acutis vel breviter sed distincte rostratis.

Texas: Big Springs, 1902, Tracy 8218; San Elizario, in field, Sept. 26, 1849,
Wright 746. New Mexico: Cook’s Spring, northern Luna Co., Nov. 3, 1887, Bigelow;
Ft. Bayard Watershed, Grant Co., 1905, Blumer 1781; Mangas ae 18 mi. north-
west of Silver City, Grant Co., Metcalfe 774; valley: of the Rio Grande 10-100 mi.

Herb.)

above El Paso, Wright 1982 (tTypE, Gray . CxsravuaHua: Chihuahua, Oct. 1935,
LeSueur 50 El B , northeast of El re, Bavispe Area, Santos 2134
San Luis Potosr: 14 mi. northw edral, dense pure col depression near

road on desert plain, 1938, Johnston 7609; valley of San Luis Potosi, 1876, Schaffner
1025 in pt.

This is the species accepted as M. repens by Hitchcock, No. Am. FI. 17:
451 (1935) and Man. Grasses U. S. 362. fig. 737 (1935). As discussed

388 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

under the following species, the name ‘‘M. repens” properly applies to the
plant which Hitchcock called M. utilis. From the true M. repens (that
is, M. utilis), the present plant differs in its distinctly larger spikelets,
scabrid lemmas and pedicels, looser paler acute glumes more than half as
long as the lemma, somewhat larger ligule, and coarser stems and leaves.
It ranges in western Texas and southern New Mexico south through
Chihuahua to San Luis Potosi.

Muhlenbergia repens (Presl) Hitchc. in Jepson, Fl. Calif. 1: 111 (1912).

Muhlenbergia utilis hia Hitchce. ose _ Acad. 23: 453 (1933).

ie sacatilla Fourn. Mex. P]. 2: 101 86).

: Parras, May 16, 1847, eames Parras, thick masses on alkali bottom,

‘ton, "Sta 452. CurmuaHua: Valley of the Sacramento near Chihuahua, by
stream, Nov. 6, 1885, Pringle 418

Central Texas, oeiieen California (where probably introduced),
Durango, San Luis Potosi, and central Mexico. Hitchcock has treated this
plant as M. utilis, cf. No. Am. Fl. 17: 451 (1935) and Man. Grasses U.S.
362. fig. 738 (1935), and applied the name M. repens to the plant I have
called M. abata. The original ae gre . Sporobolus repens Presl, and
Scribner’s illustration, Ann. Mo. Bot. . 10: 53. t. 30 (1899), of the
isotype of Presl’s species at St. Louis . little doubt as to the precise
identity of the species. It is obviously one of the forms of M. utilis found
in central and southern Mexico.

Muhlenbergia montana (Nutt.) Hitchc. Bull. U. S. Dept. Agric. 772: 145, 147
(1920).

CHIHUAHUA: High summits of the Sierra Santa Eulalia, 1885, Pringle 392.
Western United States south into Chihuahua and Sonora.

areal Porteri spats in Beal, Grasses No. Am. 2: 259 (1896).
Dam, Harvey 946; 2 mi rthwest of Frontera, road to

eastern aren of Llano de Guage near La Pistola, forming tangled masses supported
by bushes in mogote, Johnston & Muller 357, 764; San Antonio de los Alamos, one
colony at base of tuff cliffs, Johnston 8265; 4 mi. west of Lag. de Leche, sprawling or
vining in protection of shrubs, scattered on shrubby desert hillside, Muller 3285;
Torreon, large masses at base of bushes, 1898, Palmer 511. Curauanua: Presidio del
Norte Dinaca).. July 1852, Parry; Juarez, dry mesa, Sept. 26, 1902, Pringle 11233;
Chihuahua, hills and aed Pringle 478 (US) ; Paral: Chihuahua road, 10 km. north
of Rio San Pedro, Harvey

Texas and ANG to California and south into northern Mexico. A
common grass in silty valley soils, usually growing in the shelter of bushes.
The globose entangled masses of stems, 3—4 dm. in diameter and supported

features of the mogotes in Coahuila. The plant, when covered with its
very abundant open airy purplish panicles, is conspicuous and attractive.
Muhlenbergia arizonica Scribn. Bull. Torr. Bot. Cl. 15: 8 (1888).

Cuinvuanua: Rocky hills northeast of Chihuahua, thin dry soil, Sept. 16, 1885,
Pringle 402.

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 389

Arizona southward in the mountains of Sonora and Chihuahua to Sinaloa
and Durango.

i nia slaeaites Buckl. Proc. Acad. Nat. Sci. Phila. 1862: 91 (1863).

CoAHUI Western base of Picacho del Fuste, silty places on pe tufted, not
common, Foheion ee tableland north of Canon del Cuervo Chico, common on wide
grassy valley, Johnston 8538; between Palos Blancos and San Paes east of Cuesta
Zozaya, common on grassy valley on tableland, Johnston 9273, 9275; foot slopes at

Laguna Jaco, fairly common, Stewart & Johnston 1962; eastern foothills of Sierra
Cruces, 8 mi. north of Santa Elena, one flat, Johnsto ne Muller 1024; 3 mi. east of

de Villa on Coahuilan boundary, silty plain, Johnston 8180; 2 mi. south of San Fer-
pear silty plain, Johnston a Chihuahua, plains, Pringle 479 (US); arroyo 50 km.
orth of Jimenez, Harv
Kansas to Texas es ate and south into Zacatecas.
Muhlenbergia setifolia Vasey, Bot. Gaz. 7: 92 (1882).
CoAHUILA: Sierra del Pino, La Noria, banks of arroyo, Johnston & Muller 662, 665;

Western Texas south into Coahuila. The oe is closely related to
M. rigida and apparently separable from it by no single cter. It
range is mostly just beyond the northern limit of u. rigida, but it grows
with that species, along the northeastern limits of the latter, in eastern
Coahuila. It is a smaller more slender plant than M. rigida, with looser
tufts of stems and leaves, filiform involute leaf-blades, and narrower fewer-
flowered green or brownish (not purple) panicles.

Muhlenbergia rigida (H.B.K.) Kunth, Rév. Gram. 1: 63 (1829).

Coauutmta: Sierra del Carmen, Sept. 8, 1936, Marsh 719; ie Grande, 40 km.
northwest of Hac. Encantada, meadows and se Het common, Stewart 1628, 1630;
Sierra del Pino, La Noria, gravelly flats among clumps of ae oaks, Johnston &
Muller 659; Puerto San Lazaro, dominant grass on Bete slopes, Muller 3095; north
end of Carneros a among cacti, not common, Johnston 7289. CHIHUAHUA: Sierra
Organos, south of Organos at base of oak-clad slope, coarse tufts, rocky places, Stewart
& Johnston 2065; Sierra Santa Eulalia, Sept. 17, 1885, Pringle

Trans-Pecos Texas to Arizona and saute to Central America. The
species appears to be absent in northeastern Mexico and to reach its eastern
limit along the western base of the Sierra Madre in Coahuila and Nuevo

e€ It has been collected in Hidalgo. In eastern Coahuila it is con-
nected by intergrades with the closely related and generally more northerly
ranging M. setifolia.
Muhlenbergia dubia Fourn. ex Hemsl. Biol. Centr. Am. Bats 3: 540 (1885).
Sierra Madera, Cafion Charretera, coarse tufts in rocky arroyo-bottom
just below pine- belt, 3-4 ft. tall, Johnston 8975, 9069; Saltillo, among large rocks on

i)
Ln]
rt)

390 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

outer rim of treeless mountain, 1898, Palmer 416; Saltillo, deep ravines, 1898, Palmer
79; San Lorenzo Canyon, 6 mi. southeast of Saltillo, high up canyon, large bunch

canyon side, 1904, Palmer 341; north end of Carneros Pass, coarse tufts between
bushes, Johnston 7288. HIHUAHUA: Rocky hills northeast of Chihuahua, cool
slopes, Oct. 20, 1885, Pringle 403.

Trans-Pecos Texas to New Mexico and adjoining Mexico, south through
Nuevo Leon and eastern Coahuila to eastern San Luis Potosi. The type of
M. dubia came from the Chinantla, Puebla. I have seen no authentic
material and no collections from south of San Luis Potosi. Our plant is
the same as the Texan material described as M. acuminata Vasey. Perhaps
that name is the proper one for our present plant.

Muhlenbergia peaespaie Hitchc. No. Am. Fl. 17: 470 (1935) ; Johnston, Jour. Arnold
Arb, 22: 155 (1941).

utta: Locally common on the gypsum ridges east of Laguna de Jaco, John-
ston & Muller 1074, Stewart & Johnston 1963; 10 mi. east of Fraile, abundant on
valley floor, local, Johnston 7305; 6 mi. north of La Ventura, common on gypsum

plain, local, Johnston 7642, Shreve & Tinkham 9607.

Known only from scattered stations in Coahuila, San Luis Potosi, Nuevo
Leon, and southern Tamaulipas, apparently confined to gypsum.
Muhlenbergia parviglumis Vasey, Contr. U. S. Nat. Herb. 3: 71 (1892).

CoanuILa: Sierra del Carmen, Sept. 9, , Marsh 717; Sierra Madera, Canon
Charretera, banks of arroyo in oak-belt, mes gare Folnsion 9076.

Known from Uvalde, Val Verde, and Jeff Davis Counties, Texas, and
from Nuevo Leon and eastern Coahuila.

Muhlenbergia polycaulis Scribn. Bull. Torr. Bot. Cl. 38: 327 (1911).

Cuimuanua: Hills northeast of Chihuahua, cool wet ledges, 1885, Pringle 394.

Trans-Pecos Texas to Arizona and south in the mountains of Baja Cali-
fornia, Sonora, and Chihuahua to Durango.

Muhlenbergia pauciflora Buckl. Proc. Acad. Nat. Sci. Phila. 1862: 91 (1863).
Coanuma: Sierra Mojada, Canon San Salvador, abundant on slopes, Muller 3312.
Trans-Pecos Texas to Arizona and south in the mountains of Baja

California, Sonora, and Chihuahua. The type of the species (Wright 732)

was collected “‘in declivities in the mountains near El Paso, Sept. 12, 1849.”

eee monticola Buckl. Proc. Acad. Nat. Sci. Phila. 1862: 91 (1863).
CoanvuILA: Camp near Mt. Carmel canyon, Oct. 1852, Parry; Sierra del Carmen,

Sept. 1, 1936, Marsh 872; limestone hill near Santo Domingo, Wynd & Mueller 453;

Muraquls Marsh 549; Sierra Madera, Canon del Agua, among rocks in oak-pinyon

belt, in lower canyon, Muller en 3259; Cuatro pie Marsh 2053; Puerto San

cliffs, Johnston & Muller is Sierra Cruces, ester foothills near Santa Elena,

among bushes along arroyo, Johnston & Mu 238; Sierra Cruces, Cano aja
Blanca, about rock ledges on crest at head of canyon, Johnston & Muller 298; Sierra
Mojada, Cafon San Salvador, abundant on slopes, aad 3312, CHIHUAHUA rra

1937, LeSueur; Sierra Santa Eulalia, 2 km. north of San Antonio, Harvey 1513; Sierra
Santa Eulalia, dry limestone ledges, Aug. 1885, Pringle 396

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 391

Trans-Pecos Texas to Arizona and south into northern Mexico. In its
extreme form, typical M. monticola is separated from typical M. tenuifolia,
of central Mexico, by having narrow inflorescences with strict branches,
Strict spikelets, and green acute lanceolate glumes. Muhlenbergia tenuifolia
has purple inflorescences, which have spreading or ascending branches on
which the spikelets are divaricate or even reflexed. Its glumes are deep
purple, smaller than in M. monticola, and commonly denticulate and
obtusish and abruptly apiculate at the apex. In our area, the ranges of

. monticola and M. tenuifolia meet and the species intergrade very badly.
I have arbitrarily referred to M. monticola those plants with narrow green
or weakly purplish inflorescences (i.e., those with strict panicle-branches
and strict spikelets), and to M. tenuifolia those plants having a usually
darkly colored panicle with spreading branches and spikelets.

e of M. monticola (Wright 731) was collected in Limpia Canyon
in the Davis Mts., Texas. This typical form has been illustrated by Hitch-
cock, Man. Grasses U. S. fig. 788 (1935). Very similar plants have been
collected elsewhere in trans-Pecos Texas. Material from Arizona, how-
ever, differs in having a distinctly looser panicle. Vasey, U.S. Dept. Apric.,
Div. Bot. Bull. 131: ¢. 79 (1892), has an illustration (sub M. calamagro-
stidea) of the Arizonan form. From Arizona this aberrant form extends
south into the Sierra Madre of Sonora and Chihuahua, where the inflo-
rescence becomes larger and more open and the spikelets divaricate or even
reflexed. These large plants, except for the green, not purple, panicles, are
remarkably similar in gross aspect to some forms of M. tenuifolia ae
about Mexico City, the type locality of that species. It is clear that the
two species, M. monticola and M. tenuifolia, intergrade in the northern
states of Mexico and that, if they are both to be recognized, this can be
justified only for convenience and performed in an arbitrary manner.
Muhlenbergia tenuifolia (H.B.K.) Kunth, Rév. Gram. 1: 63 (1829).

VERNACULAR NAME: Zacate espumilla.

CoanuiLa: Mouth of Cafion La Cruz, 20 km. south of Ocampo, gravelly bed of
large arroyo, Johnston 9187; Cafion Bocatoche, o open arroyo, Muller 3119; La Rosita,
Shreve & Tinkham 9591; 2 mi. west of Saltillo, Harvey 1097; Saltillo, ee of garden
under trees, 1898, Palmer 393,; Sierra Cruces, near Santa Elena, hillsides, Stewart 284;
San Antonio de los Alamos, eae flat above tuff cliffs, Johnston 8255; aoe 1880,
Palmer 1348; Sierra Parras, Oct. 1910, Purpus 5007; Sierra Negras, 9 km. south of
Parras, Stanford et al. 165. Cutmuanvua: Rocky hills near Chihuahua, May e 1885,
Pringle 428; hills and plains near Chihuahua, Oct. 23 ae pissy 397. ZACATECAS:
Valley 15 km. west of Concepcion del Oro, Stanford et al, :

Ranging from central and southern Mexico ate into our area. Over
most of its range a well marked and readily recognized species, but in our
area completely intergrading with the more northern M. monticola. The
species is a weak perennial and, like M. monticola, ee sheltered places
at bases of cliffs, along rocky arroyo banks, or on slopes i in deep canyons.
Muhlenbergia elongata Scribn. in Beal, Grasses No. Am. 2: 251 (1896).

CHIHUAHUA: Rocky hills east of Chihuahua, ledges, 1885, Pringle 398 (1soTYPE).

Known only from near Chihuahua and from extreme southwestern parts
of the state (Palmer 159).

392 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Muhlenbergia Marshii sp. nov.
Planta perennis; culmis ca. 1 m. altis erectis teretibus glabris dense
caespitosis simplicibus, basi ad 3 mm. crassis internodiis 1-1.5 dm. longis;

rigidulis utrinque scabridulis; panicula spiciformi 2.5—-4 dm. longa 3-6
mm. crassa, haud vel vix interrupta, basi e vagina superiore saepe haud
exserta, ramis infra medium paniculae 2—4 cm. longis strictis multifloris;
spiculis strictis 0-0.5 mm. longe pedicellatis (aristis glumarum exclusis)
ca. 4 mm. longis; glumis subaequalibus 3 mm. longis pallidis subhyalinis
obscure medio-nervatis (nervis prominentulis scabridis) lanceolatis, paullo
supra basim latioribus deinde sursum in aristam rectam 0.3-0.7 mm
longam gradatim contractis, vel non raro Summum ad apicem basim aristae

CoanvuiLa: Sierra del Carmen, Sept. 8, 1936, E. G. Marsh Jr. 746 (TYPE, Gray
Herb.). Texas: Davis Mts., H. O. Canyon above Sawtooth, soil in cracks of rocks
in stream-bed, tough tightly rooted clumps, July 1936, Hinckley.

A member of the species-complex passing as M. rigens, among the mem-
bers of which it is readily distinguished by its short-awned glumes and
lemmas and extreme eastern occurrence. The basal portion of the inflo-
rescence bears elongate strict branches and is not exserted from the upper-
most leaf-sheaths. In these characters it agrees with true M. rigens of
central California and closely related forms from southern California and
southern Arizona and adjacent Mexico. True M. rigens has a somewhat
interrupted inflorescence composed of rather elongate strict branchlets,
and it appears to be confined to western middle California.

Muhlenbergia mundula sp. nov.

OAHUILA: Sierra Hechiceros, Cafion Indio Felipe, side of dry arroyo, common,
Stewart 174. Curmauanua: Rancho El Pino, 10 km. southeast of Sierra Rica, rocky
slope, frequent, Stewart 2407; by streams near Chihuahua, Oct. 13, 1885, Pringle 417
(type, Gray Herb.); Quicorichi, Rio Mayo, Gentry 1931; southwestern Chihuahua,
1885, Palmer 21. Sonora: Canon Bellota, Sierra Cabellera, Bavispe Area, Santos
2096. New Mexico: Berendo Creek, Black Range, Sierra Co., Metcalfe 1391. ARI-
zona: Chiricahua Mts., Rigg’s Ranch, Blumer 1491; Pinal Creek, Miami, Harrison &

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 393

Kearney 6342; Mule es Goodding 907; Rincon Mts., Manning Camp, Blumer 3397;

White Mts., Griffiths 5
This species ania most of the plants of Arizona, New Mexico, and

northern Mexico which have passed as M. rigens. The plant illustrated by

Hitchcock, Man. Grasses U.S., as M. rigens belongs to M. mundula. It is

readily distinguished from true M. rigens of central California, and from

closely related unnamed forms from southern California and southern

Arizona, by having the spike well exserted from the upper leaf-sheaths and

by having the lower branches of the panicle only 5-15 mm. rather than

20-30 mm. long. In appearance the species much resembles M. leptoura,

of northwestern Chihuahua, but it differs from that species in having the

glumes distinctly shorter than rather than equalling or surpassing the

Jemma.

Blepharoneuron tricholepis (Torr.) Nash, Bull. Torr. Bot. Cl. 25: 88 (1898).
CHIHUAHUA: Mapula Mts., thin soil on summits, Oct. 26, 1886, Pringle 822
Colorado and Utah south through Arizona, New Mexico, and trans-

Pecos Texas into the mountains of northern Mexico. The species is known

from the Davis, Chinati, and Chisos Mountains of Texas and is, accord-

ingly, to be expected in the mountains of northern Coahuila and north-
eastern Chihuahua

Sporobolus microspermus (Lag.) Hitchc. Jour. Wash. Acad. 23: 453 (1933).

San Antonio de los Alamos, colony in loose gravelly soil on flats above
tuff cliffs, Johnston 8244. CHIHUAHUA: Near Rancho El Pino, 10 km. southeast of
Sierra Rica, rocky slope, Stewart 2411; Los Medanos, 1935, LeSubuy 74.

As currently accepted, this species ranges in western United States and
south to Costa Rica. <A critical study will probably show it to be an aggre-
gate of several well-marked species of limited distribution. Hitchcock,
Contr. U. S. Nat. Herb. 17: 308 (1913), sub S. ramulosus, reports ne
plant from the Sierra Santa Eulalia and from near Chihuahua.

Sporobolus pyramidatus (Lam.) Hitchc. U. S. Dept. Agric., Misc. Pub. 243: 84

1936).

Cee as pulvinatus Swallen, Jour. Wash. Acad. 31: 351 (1941).

CoanuiILa: South of Laguna de Leche, saline and perhaps gypseous silt in mogote,
Johnston 8626; Saltillo, along ditch, Hiichoast 5580 (US). CHinUAHUA: 5% mi.
south of Ojinaga, ou eh from saline and gypseous shales, Johnston 8005; sandy
plains near Chihuahua, Sept. 22, 1886, ees 816. Duranco: Flats on plains 3 mi.
northeast of Bermejillo, Weiss 7787.

Texas and Oklahoma to Arizona and south into Mexico; South America.
Sporobolus airoides Torr. Pac. R. R. Rep. 73: 21 (1858).

COAHUILA: 4 mi. west 9 Cuatro Cienegas, coarse tufted grass on alkaline flat,
Johnston 7133; Laguna de Jaco, saline meadow south of the lake, abundant, coarse
tufts, Johnston & Muller a CuIHUAHUA: Villa Ahumada flat, 1935, EeSueus 7a

Widely distributed in saline soils in western United States and extending
south into northern Mexico.

Sporobolus regis sp. nov.

Planta robusta ca. 12 dm. alta e rhizomate elongato nodoso (internodiis
10-12 mm. longis) 3-4 mm. crasso oriens; culmis simplicibus erectis con-

394 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

fertis foliosis; vaginis quam internodiis saepe subaequilongis vel usque ad
cm. brevioribus, extus pilis numerosis gracilibus 1-2 mm. longis adpressis
pallide flavescentibus donatis, vetustioribus glabrescentibus; ligula subnulla
mm. longa vel breviore fimbriata vel dense et minute ciliolata; laminis
1-3 dm. longis 3—4 mm. latis saepe laxe involutis, subtus levibus, supra pal-
lidioribus minute scabridis; paniculis apertis 3-4 dm. longis ad 15 cm.
crassis, basi e vagina superiore haud exsertis, ramis alternis 2-12 mm. dis-
tantibus ramulosis adscendentibus vel laxe adscendentibus 5—10 cm. longis,
axillis villosis, ramulis laxe adscendentibus saepe purpureis; spiculis 2—2.5
mm. longis glaberrimis unifloris, pedicellis strictis 1-5 mm. longis; gluma
exteriore 1.2-1.5 mm. longa hyalina late lanceolata acuta, supra basim in
medio purpurea; gluma interiore 1.5-1.8 mm. longa hyalina lanceolata
enervata acuta; lemmate hyalino lanceolato-elliptico obscure uninervato,
apice late acuto, longitudinem paleae aequante.

Coanuua: Salt flat 4 km. southeast of Laguna del Rey, abundant, Sept. 18, 1942,
Stewart 2653 (type, Gray Herb.).

A very well marked species, probably most closely related to S. airozdes
Torr. and S. Wrightii Munro. From these and most other members of the
genus, S. regis differs in the very hairy leaf-sheaths, the tufts of hairs in the
axils of the panicle-branches, and the very coarse rhizomes. The bases of
the culms and the younger nodes of the rhizomes bear shredded remnants of
old leaves.

Sporobolus Wrightii Munro ex Scribn. Bull. Torr. Bot. Cl. 9: 103 (1882).

VERNACULAR NAME: Zacaton.

CoaHUILA: 5S mi. north of Allende, oak thicket on plain, coarse tufts, culms 3-6 ft.
tall, Johnston 7010; open country between Santo Domingo and Piedra Blanca, Wynd

northeast of Jimulco, Stanford et 72. a: Presidio del Norte, 6-8 ft.
tall, Parry; 11 mi. south of Ojinaga, along small arroyo in low hills, 3-5 ft. tall, John-
ston 8034; Piramide, low gr , avy soil a of cornfield, 3-6 ft. tall,

Johnston 8139; Carrizal, Aug. 18 or 19, 1846, Wislizenus 103; Meoqui, 1938, LeSueur

; 20 km. south of Camargo, Harvey 1396. ZACATECAS: 15 km. west of Concepcion
del Oro, Stanford et al. 528; 7 mi. north of San Tiburcio, heavy, probably alkaline
soil among mesquites, Johnston 7362.

Trans-Pecos Texas to southern California and south to central Mexico.
Hitchcock, Contr. U. S. Nat. Herb. 17: 309 (1913), reports the species
from Chihuahua, Torreon, and Saltillo. It appears to be present in most
parts of our area. Although I have accepted this species in the broad tra-
ditional sense, I suspect that it is an aggregate of several critical species.

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 395

The plants I have included in S. Wrightii vary greatly in size, appearance,
and selection of habitats. The leaves are broad to narrow, dark or pale
green, flat to involute. The plants may form large very coarse tussocks
over a meter broad, and with culms nearly 2 m. tall, or low clumps 1-2 dm.
tall, with the culms much less than a meter in height. The larger plants
are mostly from permanently wet soils; the small plants come from the
plains where water stands and evaporates after storms. The soils support-
ing the plant vary from those with no appreciable amount of salt to those
with large amounts of alkali and even gypsum.

Sporobolus cryptandrus (Torr.) Gray, Man. 576 (1848).

CoaHurta: Saltillo, sandy field, Hitchcock 5625 (US); Tanque Colorado, among
bushes on red dunes, Johnston 8663; Los Medanos, 1935, LeSueur 60, 80; valley near
Chihuahua, Oct. 4, 1885, Pringle 419

Widely distributed in sandy glace over the United States and south into
northern Mexico. The two following species are probably no more than
varieties.

Sporobolus flexuosus (Thurb.) Rydb. Bull. Torr. Bot. Cl. 32: 601 (1905)

sinha eisai var. pin Thurb. in Wats. Bot. Calif. 2: 269 (1880).

CHIHUAE Medanos, 1935, LeSueur 80a; between Los Medanos and
enna sand nin Sept. 1886, Pringle 815; Colonia Diaz, 1899, Nelson 6458.

Southwestern United States and northern Chihuahua. Differing from
S. cryptandrus only in having a more open panicle, with the spikelets and
ultimate branchlets spreading rather than appressed along the primary
branches of the panicle.

Sporobolus giganteus Nash, Bull. Torr. Bot. Cl. 25: 88 (1898).

Sporobolus cryptandrus var. strictus Scribn. Bull. Torr. Bot. Cl. 9: 103 (1882).

Sporobolus contractus Hitchce. Am. Jour. Bot. 2: 303 (1915).

HuILA: North end of Canada Oscuro, gravelly slopes among brush, common,
Johnston 8460; south of Laguna de Leche, shaly bank among brush, frequent, John-
ston 8625; Sierra Cruces, 8 mi. north of Santa Elena, stony flat among bushes, Johnston
& Muller 1027. CHrauaHuaA: Los Medanos, 1935, LeSueur 13, 58; 20 km. south of
Camargo, Harvey 1397.

Southwestern United States and northern Mexico. Probably only a
phase of S. cryptandrus, with the inflorescence bearing short strict crowded
branches and branchlets and accordingly spike-like in form. Hitchcock has
distinguished the robust plants of this form as S. giganteus and the more
slender forms as S. contractus.

Sporobolus Nealleyi Vasey, Contr. U. S. Nat. Herb. 1: 57 (1890); Johnston, Jour.
d Arb, 22: 155 (1941)

Coanuita: Castillon, confined to gypsum flat, oS Johnston & Muller 1268 ;
east of Laguna de Jaco, frequent, confined to gyp , Johnston & Muller 1073,
Stewart & Johnston 1954; Sierra Cruces, gypsu sh near Santa Elena, Johnston &
Muller 247; gypsum bed west of Buena Vista, oe road between San Antonio de los
Alamos and Puerto Caballo, rare, Johnston 8314.

Western Texas and eastern New Mexico and south into Coahuila.
Apparently confined to gypsum. The type-collection of the species was
originally given as collected at “Brazos Santiago, Texas,” that is, near the
mouth of the Rio Grande, where the species is neither known nor to be

396 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

expected. This is apparently the result of a clerical error. Nealley’s
specimens came from Screw Bean, a locality near the Pecos River, in Reeves
County, Texas, where a large variety of gypsophiles has been collected.
Sporobolus spiciformis Swallen, Proc. Biol. Soc. Wash. 56: 78 (1943).

Coanvuita: Puerto del Norte, Cuatro Cienegas, July 1939, Harvey 1225 (TYPE,
US.) ; 4 mi. west of Cuatro Cienegas, common and conspicuous on saline and zypseous
flats, 1938, Johnston 7132; 4 km. southeast of Laguna del Rey, abundant on salt flats,
1942, Stewart 2654; Noria de San Juan, southeast of Laguna del Rey, common on
saline plain, 1942, Stewart 3008.

Endemic to our area, and apparently confined to saline gypseous soils.
The species is most closely related to S. phleoides Hack., of saline soils in
the deserts of western and northern Argentina. It differs from the southern
plant in its paler color, involute leaves, more rigid tighter leaf-sheaths,
broader erose or dentate glumes, non-rostrate palea, and much larger
anthers.

Oryzopsis hymenoides (R. & S.) Ricker ex Piper, Contr. U. S. Nat. Herb. 11: 109
(1906).

A widely distributed species of sandy soils in western United States.
Hitchcock, Contr. U. S. Nat. Herb. 17: 285 (1913), reports a collection
from “sandhills near Paso del Norte, Pringle 1053.”

Piptochaetium fimbriatum (H.B.K.) Hitchc. var. confine var. nov.

A forma typica austro-Mexicana glumis viridibus evidenter nervatis haud
purpureis differt.

Coanuita: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 643; Hillcoat
Canyon, west of Buena Vista Ranch, July 13, 1938, Marsh 1284; Sierra Madera, Canon
Charretera, in shade in oak thickets, Johnston 9077; Sierra del Pino, La Noria, in
thickets on arroyo-bank, Johnston & Muller 486 (typr, Gray Herb.).

Ranging in the mountains along the Mexican Boundary, from trans-
Pecos Texas (Chisos and Davis Mts.) west to Arizona, and from Coahuila
and northern Nuevo Leon to northeastern Sonora. Typical P. fimbriatum
comes from central Mexico and differs from our northern plant in having
firmer obscurely veined purple glumes and usually less slender and flaccid
leaves. Where I have seen the var. confine in Coahuila, it has always
grown in very sheltered shaded places, usually under bushes in dense oak
thickets, and it is seldom common.

Stipa neomexicana (Thurb.) Scribn. U. S. Dept. Agric., Div. Agrost. Bull. 17: 132
(1899).

Coanutta: Along trail from southern extremity of Hillcoat Mesa to Buena Vista
Ranch headquarters, July 27, 1938, Marsh 1492; tableland north of Cafion del Cuervo
Chico, basal slopes of low rounded limestone hills, Johnston 8564; north end of
Carneros Pass, conspicuous grass among low shrubbery, Johnston 7297.

Western Texas and Colorado to Utah and Arizona and south through
Coahuila to the mountains of Nuevo Leon.

Stipa leucotricha Trin. & Rupr. Mém. Acad. St. Pétersb. VI. Sci. Nat. 51: 54 (1842).

Coanvuita: Rio Grande Valley near Piedras Negras, - a 1900, Pringle 8292;
Muzquiz, Dec. 5, 1936, Marsh 1077; Monclova, Marsh 1691

Oklahoma south through central Texas into code se Coahuila. The

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 397

species appears to differ from S. mucronata H.B.K., of central and eastern
Mexico, only in the large size of the spikelet and fruit and in the usually
green rather than purpurescent glumes. The material from Coahuila and
adjacent Texas is distinctly smaller than the typical form of S. leucotricha,
and some of the specimens have colored glumes.

Stipa eminens Cav. Icon. Pl. 5: 42. t. 467 (1799).
Coanurta: Limestone hill near Santo Domingo, Wynd & Mueller 460; Hillcoat

of Canon Charretera, open flat, Johnston 9161; near Santa Rosa, limestone hits, Shreve

Noria, common grass on flats and meadows, Johnston & Muller 454, 663, Stewart
1207; tableland north of Canon del Cuervo Chico, gravelly places on low limestone
hills, Johnston 8562; Sierra Parras, April 1905, Purpus 1125 in pt.; Sierra Cruces, lime-
stone foothills 8 mi, north of Santa Elena, Johnston & Muller 1026; Sierra Cruces,
about volcanic ledges on ridge at head of Cafion Tinaja Blanca, Pi aacis & Muller
305. CHIHUAHUA: Sierra Santa Eulalia, Aug. 12, 1885, Pringle

Trans-Pecos Texas to Arizona and south to central ace Hitchcock,
Contr. U.S. Nat. Herb. 24: 238 (1925), reports the species from Cedros,
Zacatecas.

Stipa angustifolia Hitchc. Contr. U.S. Nat. Herb. 24: 246 (1925).

CoanulILa: Saltillo, July 25, 1905, Palmer 526.

This species was based upon a specimen collected by Palmer July 25,
1905, ‘‘among rocks on summit of Sierra de la Puebla, near Saltillo.” The
collection number is given as ‘Palmer 726.” The specimen in the Gray
Herbarium, agreeing with the original description, and collected near
Saltillo on the same date as the type, bears the number Palmer 626. This
species has very slender involute leaves. It has been collected in Nuevo
Leon (near Pablillo, Mueller 522) and in southern Tamaulipas (Miqui-
huana, Stanford et a
Stipa clandestina Hackel, Rep. Sp. Nov. 8: 516 (1910).

CoaHUILA: Saltillo, alt. 1650 m., March 1908, Arséne 3441 (1soTYPE); Saltillo,
alt. 1600 m., March 1908, Arséne 3467; Saltillo, in large bunches, banks of irrigation
ditches, 1898, Palmer 3; Saltillo, ne ground, 1910, Hitchcock 815. ZACATECAS:
Valley 15 km. west of ei sceicion del Oro, valley floor, 18 in. tall, Stanford et al. 553.

Known only from our area.

Stipa editorum Fourn. Mex. Pl. 2: 75 (1886).

CoaHuILa: 10 mi. east of Fraile, abundant and most conspicuous species on silty
floor of alles: ee Johnston 7303. ZAacaTECAS: 7 mi. north of San Tiburcio, heavy
soil among mesquites, Johnston 7361

Known pat ann Pientecds Nuevo Leon, southern Tamaulipas,
and Puebla. The species may possibly be gypsophilous.

Stipa multinodis Scribn. ex Beal, Grasses No. Am. 2: 222 (1896).

CHIHUAHUA: Sierra Santa Eulalia, Aug. 1885, Pringle 385 (1soTYPE).

Closely related to S. editorum, with which it agrees in the many-noded
strict culms 7—12 dm. tall and somewhat fruticulose at the base. It differs
in a slightly smaller lemma and nearly absent ligule. The species is known
only from the type collection.

398 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Stipa robusta (Vasey) Scribn. U. S. Dept. Agric., Div. Agron. Bull. 5: 23 (1897).

Coanurta: Sierra Madera, Cafion Charretera, common in rocky bed of arroyo in
oak belt, becoming 4 ft. tall, Johnston 9073; Saltillo, 1902, Palmer 317; Lirios, 1880,
Palmer 1249; ~ Antonio de las Alanzanas, near mountain border, 3 ft. tall, Aug. 31,
1848, Gregg 3

Colorado if Arizona and trans-Pecos Texas, and south through Coahuila
to the mountains of Nuevo Leon. A large coarse grass apparently restricted
to the oak and lower pine belts. The type came from the Chinati Mts.,
Texas, and accordingly the species is to be expected in the mountains -
northeastern Chihuahua. In the United States this species has been re-
ported to have narcotic effects on animals eating it. In the notes accom-
panying his collections from the mountains of southeastern Coahuila, cited
above, Gregg states that the grass is “very injurious to animals, intoxicat-
ing and often killing them. Animals acquainted with it will not eat it.’
Stipa tenuissima Trin. Mém. Acad. St. Pétersb. VI. Sci. Nat. 21: 36 (1836).

CoaHUILA: Sierra del Carmen, Sept. 1, 1936, Marsh 878; Sierra del Carmen, Canon
Sentenela, Wynd & Mueller 538; Hillcoat Canyon, west of Buena Vista Ranch, July
13, 1938, Marsh 1314; Saltillo, bunches on shady slope of hill, 1904, Palmer 455;
Saltillo, 1909, rare, Arséne 3469; Chojo Grande, 27 mi. southeast of Saltillo, 1904,
Palmer 341; north end of Carneros Pass, among low bushes, Johnston 7296; Carneros
Pass, limestone hills, Sept. 20, 1890, Pringle 3274.

New Mexico and trans-Pecos Texas south through Coahuila and Nuevo
Leon to Puebla; Argentina.
Aristida adscensionis L. Sp. Pl. 82 (1753).

1UILA: Sierra del Carmen, July 29, 1936, Marsh 675; 21 mi. south of Sabinas,

heavy soil, Johnston 7043; Santa Anna Canyon, Marsh 433; Cuatro Cienegas, Marsh
2059 in pt.; Saltillo, 1898, Palmer 388; Chojo Grande, 27 mi. southeast of Saltillo, in
ravine, 1904, Palmer 333; 42 mi. west of Saltillo, Shreve & Tinkham 9839; west base
of Picacho del Fuste, cemented gravels, Johnston 8415; San Antonio de los Alamos,
on canyon-wall and on gravelly flats at top of cliffs, Johnston & Muller 874, Johnston

LeSueur 36; Chihuahua, 1935, LeSueur 3; rocky hills near Chihuahua, Aug. 1885,
Pringle 390; Sierra Organos, 1937, LeSueur 191; 60 km. north of Escalon, Harvey 1306.

A variable annual species widely distributed in the warmer parts of
America and the Old World. This species has been consistently described
as always having the lemma with three awns. Three of the collections cited
above (Pringle 390, Johnston 8252 and 8415) have the lateral awns minute
or nearly wanting, but otherwise they agree closely with the specimens asso-
ciated with them
Aristida ternipes Cav. Icon. Pl. 5: 46 (1799).

Cuinvuanva: Rocky hills northeast of Chihuahua, Aug. 13, 1885, Pringle 387;
Meoqui, 1936, LeSueur 120 in pt.; Meoqui, 1935, LeSueur 38.

Ranging from trans-Pecos Texas to Arizona and south through Sonora
and Chihuahua; reaching northern South America.
Aristida Schiedeana Trin. & Rupr. Mém. Acad. St. Pétersb. VI. Sci. Nat. 51: 120

1842).
Aristida Orcuttiana Vasey, Bull. Torr. Bot. Cl. 13: 27 (1886).

Coanurta: Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 607. Cut-
HUAHUA: Sierra Santi Eulalia, Sept. 10, 1885, Pringle 386.

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 399

Trans-Pecos Texas to southern California and south to Guatemala.
Very closely related to A. ternipes and differing from that species chiefly in
having the neck of the fruit twisted and bent. Our plants represent the
northern phase of the species (A. Orcuttiana), which is weakly distin-
guished from the typical southern phase by usually having glabrous and
somewhat firmer, paler glumes
Aristida divaricata H. & B. ex Willd. Enum. Pl. 1: 99 (1809).

CuriHuAHvA: On the Coahuila boundary 1 mi. east of Poza de Villa, silty a
pe 8177; Rancho El Pino, 10 km. southeast of Sierra Rica, rocky slope, Stew

- 7 mi. south of Piramide, silty flat, Johnston 8108; 4 mi. southeast of pseitl
ae at base of grassy slope, Stewart & Johnston 2041.

Western Kansas and western Texas to California and south through
Chihuahua and Sonora, reaching Guatemala. Much resembling the two
previous species and having a similar very lax panicle with long slender
abruptly and widely spreading branches, but differing in having three well-
developed awns on the lemma. The collection from near Piramide has the
fruiting lemma with a stout untwisted neck and accordingly belongs to the
form distinguished as A. Aamulosa Henr.

Aristida longiseta Steud. Syn. Pl. Glum. 1: 420 (1855).

CoAHUILA: me mesas near Piedras a ae Apr. 23, 1900, Pringle 9037; Sierra del
Carmen, Aug. 14, 1936, Marsh 665; 20 mi. northwest of Hac. La Babia, valley
floor, Wynd & Mueller 445. CHIEUAnUA: Los Medanos, 1935, LeSueur

Widely distributed in the western United States and eres i into
our area. The Coahuilan collections, representing the var. rariflora Hitchc.,
have smooth lemmas with a stout neck. LeSueur’s collection has more
attenuate minutely tuberculate lemmas. Hitchcock, Contr. U. S. Nat.
Herb. 22: 563 eae reports a collection of the species (Pringle 473)
from near Chihuahua
Aristida Roemeriana Scheele, Linnaea 32: 343 (1849).

Aristida purpurea var. micrantha Vasey, Contr. U. S. Nat. Herb. 3: 47 (1892).

Coanutra: Allende, Marsh 1799; igneous a near Santo Domingo, Wynd &
Mueller 468; Zacate, July 14, 1936, Marsh 501; Yerda Spring, July 6, 1936, Marsh
Hermanas, Marsh 1586, 1617; Monclova, ee 1694; Canon Bocatoche, grassy
valley floor, Muller 3111; near Rancho Santa Teresa, Wynd & Mueller 173, 202, 206;
La Rosita, Shreve & Tinkham 9593; east of Hac. La Rosa, mountain slope, Wynd &
Mueller 44; mountains west of Saltillo, 1880, Palmer 1351; Saltillo, stony hills, 1898,
Palmer 392; Sierra Madera, Cafion Charretera, lower canyon, open flat, Johnston 9160;
mouth of Cafon La Cruz, 20 km. south of Ocampo, base of mountains, Johnston 9176,
9178; Cuatro Cienegas, 1939, Marsh oe in pt.; Sierra Cruces, 8 km. north of Santa
Elena, stony flat, Johnston & Muller 1

A variable and ill-defined species ranging from Texas south through
northeastern Mexico to Hidalgo. It is a plant with small spikelets with
strongly unequal, usually purpurescent, glumes in a rather loose panicle
with more or less nodding branches. It intergrades completely with A.
purpurea, A. longiseta, A. Reverchoni, A. curvifolia, A. dissita, and A.
glauca, all intergrading and variable species, and like them incapable of
precise definition. The condition is probably the result of free and much
repeated interspecific hybridization and subsequent segregation and re-

400 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

combination of characters. Since technical characters of the spikelet are
as variable and erratic as those found in the form of inflorescence and the
gross aspect of the plant, I have preferred to use these latter in defining the
“species” in this most difficult genus of grasses.

Aristida Reverchoni Vasey, Bull. Torr. Bot. Cl. 13: 52 (1886).

Aristida Nealleyi Vasey, Contr. U.S. Nat. Herb. 3: 45 (1892).

CHIHUAHUA: Sierra Santa Eulalia, dry ledges, Sept. 8, 1885, Pringle 389.

Texas to southern California and south into Chihuahua. A plant with
very strict subsessile clusters of spikelets forming a spike about 15 cm. long.
The awns are rather slender and usually pinkish. The plant intergrades
with A. Roemeriana, A. curvifolia, A. longiseta, A. dissita, and A, glauca.
Hitchcock and Henrard treated this plant as a synonym of A. glauca, but
that is a plant of eastern Mexico with looser spikelet-clusters and a more
interrupted spike, more suggestive of a slender form of A. curvifolia than of
the present species.

Aristida curvifolia Fourn. Mex. Pl. 2: 78 (18

Aristida Wrightii Nash in Small, Fl. S. E. U. 4 ae 1327 (1903).

OAHUILA: Sierra del Carmen, Sept. 12, 1936, Marsh 841; Rancho Agua Dulce,
lower slopes of Sierra San Manuel, Wynd & Mueller 328; limestone hill near Santo
Domingo, Wynd & Mueller 452; Santa Anna Canyon, Marsh 498; Palm Canyon, near
Muzquiz, Marsh 322; Flores Gustave, near Muzquiz, Marsh 315; Hillcoat Canyon,
west of Buena Vista Ranch, July 13, 1938, Marsh 1330; Hillcoat Mesa, west of Encan-
tada Ranch, July 25, 1938, Marsh 1437; Sierra del Pino, La Noria, flats, Johnston &
Muller 452, 695, 696; Sierra Madera, Cafion Charretera, bed of arroyo in oak belt,
Johnston 9068; south of Laguna Leche, slightly saline and gypseous soil near mogote,
Johnston 8621

Ranging from Texas to Arizona and south to southern Mexico. A rather
coarse and stiff plant with strict subsessile clusters of spikelets forming a
stiff interrupted spike 15-30 cm. long. Differing from A. Reverchoni, with
which it intergrades, in the stiffer brownish awns, firmer non-purpurescent
glumes, somewhat interrupted longer spike, and frequently loosely appressed
spikelet-clusters. It intergrades with A. pansa, A. arizonica, A. Reverchoni,
and A. Roemeriana.

To this species I have referred most of the larger plants of Texas and
northeastern Mexico which Hitchcock placed in A. Wrightt and in A.
glauca. Hitchcock attempted to distinguish A. glauca and A. Wrightti by
attributing to the former a more slender and elongate, usually somewhat
twisted, beak to the lemma, but J find these characters too indefinite and
variable to be of any use, if not actually illusionary. Henrard placed great
emphasis on the obtuse somewhat dentate tips of the glumes found in the
type of A. curvifolia, but this character is variable and, furthermore,
appears sporadically in various species of northern Mexico and the western
United States. Otherwise the species seems nearly the same as the Texan
plant described as A. Wrightii.

Aristida pansa Woot. & Standl. Contr. U.S. Nat. Herb. 16: 112 (1913).

Cc umLA: Along trail between southern end of Hillcoat Mesa and Buena Vista
Sega July 27, 1938, Marsh 1493; western base of Picacho del Fuste, gypsum

beds on north-facing mountain side, Johnston 8399; western base of Picacho del

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 401
ales rocky flats, Johnston 8419; tableland north of Cafion del Cuervo Chico, base

Bate flats, Johnston 8714; north of La Ventura, on gypsum beds, Shreve & Tinkham
9608; Picacho de San José, about ledges on high northwestern slope, Johnston 8209.

Western Texas and New Mexico south into our area. A plant with
branched inflorescence, the branches short, stiffly ascending, and bearing
crowded appressed spikelets in dense spicate clusters. The species inter-
grades with A. divaricata, A. dissita, and A. curvifolia. Typically it has
three subequal awns. Among the collections above cited, however, John-
ston 8209, 8399, and 8414 are plants otherwise agreeing with A. pansa but
having only a well-developed middle awn, the lateral ones being reduced
and nearly wanting. Another collection, Johnston 8419, consists of plants
with three subequal awns and was mixed with plants, otherwise similar,
having the lateral awns reduced and nearly wanting. The plants with
single awns suggest A. Schiedeana, but they are obviously variations of
A. pansa. Curiously they come from the same area in Coahuila where I
found a homologous form of A. adscensionis.

Aristida dissita sp. nov.

Planta perennis caespitosa 3—7 dm. alta perinconspicue minuteque Ppube-
scens et scabridula; culmis numerosis confertis simplicibus teretibus incon-
spicue striatis; vaginis striatis quam internodiis longioribus, margine ad
apicem villosis: ligula subnulla dense breviterque villoso- ciliata: laminis
1-2 dm. longis rigidulis rectis vel saepe plus minusve curvatis glauco-
viridibus, saepe involutis et 0.5-1 mm. crassis, raro subplanis et ad 1.5 mm.
latis, supra minute hispidulis et scabridulis, subtus sublevibus; panicula
exserta elongata 1-3 dm. longa saepe 6-8 cm. crassa aperta dissitiflora
adscendente ramosa; rhachi tereti superne subangulata, ramis in nodis
solitariis vel binis vel raro trinis gracilibus angulo ca. 45° ab rhachi ee
gentibus, infra medium paniculae 3-5 cm. distantibus, supra mediu
cm. distantibus; spiculis 0-5 (raro ad 10) mm. longe pedicellatis aris
ca 45° a ramis divergentibus, supra medium ramis 2-8 laxe dispositis ;
glumis violaceis vel purpureo-viridibus angustis acuminatis, non raro in

arista inconspicua gracili terminatis, glabris uninervatis carinatis, gluma
exteriore (7—)8—9 mm. longa; gluma interiore (8—)9—-10 mm. longa, qua
exteriore 1-2 mm. longiore; fructibus angustissimis graciliter attenuatis
nigrescentibus non tortis praesertim supra medium non raro minutissime
scabridis; aristis subaequalibus 12-18 mm. longis rectis tenuibus; callo ca.
0.9 mm longo acuto bar

CoanuiLa: Several miles west of Buena Vista, along road from San Antonio de los
Alamos, banks of gypsiferous shales, Johnston 8302; 3 mi. east of San José, silty flat,
openings in mogote, Johnston 8218. CH HIHUARUA: 4 mi. southeast of Mestefas,
grassy rocky slope, Stewart & Johnston 2016; Chihuahua, 1935, LeSueur 19; near

fall 4161; Niehes Ranch, Glass Mts., Brewster Co., Warnock T7532. New MEeExico:
15 mi. or of Las Lunas, Valencio Co., sandy alluvium, Bacigalupi 572.

This species keys to A. pansa in Hitchcock’s treatment of the genus in
the Manual of the Grasses of the United States, 440 (1935), and in North

402 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

American Flora, 17: 376 (1935), and is probably closely related to that
species. It agrees with A. pansa in having an inflorescence with mod-
erately elongate ascending branches, but it differs in having the spikelets
spreading from the branch and hence very loosely disposed, rather than
closely appressed to the branch and forming a spicate cluster. The pro-

sed species frequently suggests A. barbata, but that plant has a propor-
tionately broader panicle with longer and more widely spreading branches,
as well as larger and more widely spreading spikelets on more elongate
pedicels. Some of the cited material of A. dissita has been identified as
A. hamulosa, but that species is readily distinguished by its elongate widely
spreading or even reflexed panicle-branches and its larger very strict spike-
lets in spicate clusters. The same characters readily separate it from
A. divaricata, the species to which the type collection was referred by
Hitchcock, Contr. U. S. Nat. Herb. 22: 549 (1924). In trans-Pecos Texas
the panicle-branches of A. dissita shorten and become more strict and it
passes into A. Reverchoni, A. pansa, A. Roemeriana, and even A, curvifolia.
The loose inflorescence of A. dissita, with its ascending panicle-branches, up
to 7 cm. long, and loosely disposed ascending spikelets, gives the species a
distinctive aspect and permits its ready recognition. It is difficult to under-
stand why the plant should have remained so long without a name.
Tragus Berteronianus Schult. Mant. 2: 205 (1824).

CoanuttaA: Palm Canyon, near Muzquiz, Marsh 1000; Saltillo, 1898, Palmer 396;
north end of Sierra Cruces, sabaneta west of San Rafael, Johnston & Muller 1035.

slope, local, Stewart & Johnston 2036; hills and plains near Chihuahua, Aug. 1885,
Pringle 421; Meoqui, 1935, LeSueur 33. Zacatecas: Near Concepcion del Oro, many
iain mee among bushes and rocks on stony mesas, 1904, Palmer 279.

s to Arizona and south to South America and in the warmer parts of
the Pe World. Although stated by some authors to be introduced into our
area, it behaves like an indigenous species. It associates with indubitably
native species in scattered localities far from human habitations, and it was
collected in Texas, New Mexico, and Arizona far from settlements by the
early botanists exploring that then frontier area a hundred years ago.
Hitchcock, Contr. U. S. Nat. Herb. 17: 216 (1913), reports collections of
the species from Tlahualilo, Durango, and from Sabinas and Soledad,
Coahuila
Hilaria mutica (Buckl.) Benth. Jour. Linn. Soc. Bot. 19: 62 (1881).

VERNACULAR NAME: Tobosa.

Coanuita: Sierra del Carmen, Sept. 6, 1936, Marsh 853; Rancho Agua Dulce, east-
ern slope of Sierra San Manuel, Wynd & Mueller 481; Hillcoat Mesa, west of Encan-
tada Ranch, July 25, 1938, Marsh 1434; El Berrendo, Harvey 1175; near La Rosa,
Shreve & Tinkham 9905; between Hac. La Rosa and Hac. Lechuguilla, dry desert,
Wynd & Mueller 61; 6 ak north of Castillon, dominant on large flat, Johnston &
Muller 191; northeastern foothills of Sierra Cruces, 5 mi. south of San Rafael, silty
flat in arroyo, Johnston & Muller 1031; 7 mi. south of Jaco, small flat among
mesquites, Johnston & Muller 1115. Curavuanua: Chihuahua, plains, iy 485;
Meoqui, 1935, LeSueur 40; 10 km. east of Jimenez, Harvey 1348; 5 mi. east of Catrilio,
extensive clumps on lower slopes of dunes, Muller 3321. Duranco: 3 mi. northeast

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, II 403

of Bermejillo, flats ie mesquites, Johnston 7788; Torreon, 3-4 ft. tall, protection
of mesquites, 1898, Palmer 500.

Ranging from western Oklahoma and central Texas west to Arizona and
south into our area. A common and widespread grass characteristic of flats
where water collects after storms and stands for some time before evapo-
rating. The tobosa flats are most common on the clay soils of the calcare-
ous areas, where they vary from a few square meters in extent up to con-
siderable size. In the igneous areas in eastern Chihuahua, tobosa flats are
found chiefly on the bottoms of the larger basins. In Coahuila, in calcare-
ous Clays, tobosa flats may be encountered in varying abundance from the
bottom of the broad valleys up the long slopes to the bases of the moun-
tains and even on flat places in open canyons in the foothills. The soils
of tobosa flats may be moderately gypsiferous, but they seldom if ever are
saline. Tobosa growing in favorable situations with continually renewed
soil-moisture is considered a good feed for stock. The tobosa developing
on flats where it is subjected to irregular wettings and long periods of
drought is said to become excessively siliceous and rapidly wears down the
teeth of stock subsisting on it. Because of this fact, stock-men distinguish
the ecological forms of tobosa and have various names for them. As a
botanist I have even been severely criticized for insisting they were all one
species
Hilaria Belangeri (Steud.) Nash, No. Am. Fl. 17: 135 (1912).

CoaHu Base of the mountains 3 mi. southeast of Saltillo, Johnston 7251.
CuinvuaHvuA: Chihuahua, LeSueur 18; Chihuahua, hills and plains, Pringle 493;
Parral-Chihuahua road, 19 mi. north of Rio San Pedro, ies 1432

Texas to Arizona and south into northern Mexico. The arecent northern
plant is very closely related to H. cenchroides H.B.K., of central Mexico,
and is perhaps not specifically distinct.

Aegopogon cenchroides H. & B. ex Willd. Sp. Pl. 4: 899 (1806).

CuIHUAHUA: Mapula Mts., thin soil on ledges, Oct. 25, 1886, Pringle 823.

Ranging along the western Sierra Madre, from northwestern Chihuahua
and adjacent Sonora south to South America.

Microchloa Kunthii Desv. Opusc. 75 (1831).

CHIHUAHUA: Piramide, sheltered places about the bases of large rock masse
Prien 8120; dry gravelly soil on mesa west of Chihuahua, Aug. 6, 1885, Pringle js.

Baja California and northern Chihuahua south to Renta.
Leptochloa dubia (H.B.K.) Nees, Syll. Pl. Ratisb. 1: 4 (1824).

umLa: Sierra del Carmen, Sept. 8, 1936, Marsh 723; Jardin del Sur, Sept. 3,
1936, Marsh 767; Hillcoat Mesa, west of Encantada Ranch, July 24, 1938, Marsh 1449;
Sierra Guajes, Canon Madera, hillsides, Stewart 1506; ee Madera, Canon
Charretera, ledges among bushes on sunny hillside in oak belt, Johnston 9103;
Saltillo, sandy bottomland, Hitchcock 677; Saltillo, 1898, Palmer 381, 382; 2
mi. west of Saltillo, ies 1092A; Sierra del Pino, La Noria, arroyo banks in
oak belt, Johnston & Muller 493, Stewart 1201; Sierra Cruces, 8 mi. north of Santa
Elena, thickets along arroyo, Johnston & Muller 1029; Sierra Negras, 9 km. south of
Parras, Stanford et al. 189. CHiNUAHUA: Ojo Aliagre, Sierra Almagre, wet sand,
Johnston & Muller 1212; hills and ii near Chihuahua, Pringle 422. ZACATECAS:
Concepcion del Oro, 1904, Palmer 268

404. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Texas and Oklahoma to Arizona and southward through Mexico; Argen-
tina. Hitchcock, Contr. U. S. Nat. Herb. 17: 350 (1913), reports the
species from Cedros and Pico de Tiera, Zacatecas. A perennial species
usually growing among brush on rocky hillsides and flats. .

Leptochloa filiformis (Lam.) Beauv. Ess. Agrost. 71, 166 (1812).

CoanuILa: Monclova, 1880, Palmer 1364; Monclova, 1939, Marsh 1843.

A weedy plant, widely distributed in the warmer parts of America.
Hitchcock, Contr. U. S. Nat. Herb. 17: 349 (1913), reports a collection
by Pringle (no. 1761) from ‘‘Paso del Norte,” Chihuahua.

Leptochloa viscida (Scribn.) Beal, Grasses No. Am. 2: 434 (1896).
CHIHUAHUA: Plains near Chihuahua, wet places, 1886, Pringle 814.
Wet places, Texas to California and northern Mexico.

Leptochloa fascicularis (Lam.) Gray, Man. 588 (1848).

Coauutra: Road to Don Martin Dam, Harvey 947. CHrHuAHUA: Plains near
Chihuahua, shallow water, 1886, Pringle 813; Camargo, along the Rio Conchos,
Harvey 1403. Duranco: Torreon, in soft mud in overflowed land, 1898, Palmer 503.

Wet, oe brackish soils. Widely distributed in the warmer parts
of America

Leptochloa uninervia (Presl) Hitchc. & Chase, Contr. U. S. Nat. Herb. 18: 383

Collected on the Texas bank of the river (Boquillas, Mariscal Canyon,
and mouth of Tornillo Creek) in the Big Bend of the Rio Grande and pre-
sumably on the Coahuilan bank of the river also. Widely distributed in
wet places in the warmer parts of America.

Eleusine indica (L.) Gaertn. Fruct. et Sem. 1: 8 (1788).

COAHUILA: eri 1939, Marsh 1840. CuHrHuaHuA: Presa de Chihuahua,
1936, LeSueur 1

Widely disteibitted weedy plant, introduced from the Old World.

Cynodon Dactylon (L.) Pers. Syn. 1: 85 (1805).

Coauuita: Sierra del Carmen, Sept. 8, 1936, Marsh 740; Yerda Spring, Marsh 265;

eg hel Marsh 1611; Monclova, Marsh 1697, Harvey 1131; Saltillo, 1898, Palmer
ras, Nov. 1910, Purpus 5087; 11 km. northeast of Jimulco, Stanford et al. 3.
Ciaran Rancho El Pino, southeast of Sierra Rica, sandy arroyo, Stewart 2404;

k west of Camargo, Harvey 1410. Duranco: Torreon, low places, 1898,
an 814.

Introduced from the Old World, now widely distributed in the warmer
parts of America. A common plant in moist soils, along river bottoms,
about fenced tanques, along irrigation ditches, and frequently a tenacious
weed in cultivated soils

Spartina Spartinae (Trin.) Merrill ex Hitchc. Contr. U. S. Nat. Herb. 17: 329

CoauuILa: Cuatro Cienegas, 1939, Marsh 2037, 2039; Viesca, 1938, Shreve 8772.

Wet saline soils from Florida to Texas and south to Central America;
Argentina. In eastern Mexico it has been collected inland in Coahuila and
about Hacienda Angostura, east of San Luis Potosi.

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 405

Chloris virgata Sw. FI. Ind. Occ. 203 (1797).

CoaHutLa: Sierra del Carmen, Sept. 8, 1936, Marsh 749; 20 mi. northwest of La
Babia, open valley floor, Wynd & Mueller 443; Santa Anna Canyon, Marsh 429;
Monclova, Marsh 1848; west of Puerto de las Monjas, 1-4 ft. tall, low ground, John-

El Pino, southeast of Sierra Rica, Stewart 2397; Los Medanos, 1935, LeSueur 63;
Meoqui, 1935, LeSueur 31; Jimenez, banks of the Rio Florido, Harvey 1318.

Texas to southern California, and south to Argentina. Hitchcock,
Contr. U. S. Nat. Herb. 17: 332-333 (1913), reports specimens from
Chihuahua City, Torreon, Tlahualilo, a Saltillo. This annual grass is
most commonly found in and about mogotes, growing through low bushes
on bajillos, in fenced areas about tanques, or in other areas where silty soil
is frequently well moistened. In well watered situations it may grow over a
meter high. In some unfavored places depauperate plants less than a
decimeter high may be found
Chloris submutica H.B.K. Nov. Gen. et Sp. 1: 167 (1816).

COAHUILA: rar along ditch at edge of corn field, 1898, Palmer 390. Cut-
HUAHUA: Northeast of cae valley near Nombre de Dios, Aug. 20, 1885,
Habs 424. eee 5 km. west of Concepcion del Oro, valley floor, Stanford
et al. 558

setts our area south to central Mexico.

Chloris latisquamea Nash, Bull. Torr. Bot. Cl. 25: 439 (1898).

CoaHvuILa: 11 mi. south of Allende, along tree-lined arroyo, Johnston 7018 ; Sabinas
River, Muzquiz, Marsh 399

Texas and northeastern Coahuila.

Chloris aristata (Cav.) Swallen, No. Am. Fl. 17: 596 (1939).
CoaHuILA: Parras, among weeds on bank of ditch, 1898, Palmer 450.
From our area south to Costa Rica.

Chloris andropogonoides Fourn. Pl. Mex. 2: 143 (1886).

CoaHvuILaA: Palm Canyon, near Muzquiz, Marsh 984.

Texas and northeastern Mexico.

Trichloris mendocina (Phil.) Kurtz, Mem. Fac. Cien. Univ. Cérdoba 1896: 37

(1897)

CoauHvuILa: Perros Bravos, Sept. 20, 1848, Gregg 473; near gone 1937, Wynd
774; valley of the Nazas, May 10, 1847, Gregg 611. CmHuanva Near El Carmen,
1935, LeSueur 70; valley near Chihuahua: 1885, Pringle 475. a ee Torreon,
aone dry ditch among shrubs and cacti, Hitchcock 658.

Arizona to trans-Pecos Texas, along the Rio Grande Valley to Laredo,
and south into our area; southern South America
Trichloris pluriflora Fourn. Mex. Pl. 2: 142 (1886).

Hitchcock, Contr. U. S. Nat. Herb. 17: 335 (1913), reports a collection
(Nelson 6827) from Sabinas, Coahuila. Southern Texas, west to Val Verde
County, and south in eastern Mexico; also in Andean South America.
Bouteloua simplex Lag. Var. Cien. 4: 141 (1805).

CoanuiLa: Saltillo, corn field, 1898, Palmer 397, 398; Chojo Grande, 27 mi. south-

406 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

east of Saltillo, appearing after rains in level places near water-course, 1904, Palmer 332.
Western Texas and Colorado to Arizona, south to central Mexico.
Bouteloua Parryi (Fourn.) Griffiths, Contr. U.S. Nat. Herb. 14: 381 (1912).
Bouteloua a var. vestita Wats. Proc. Am. Acad. 18: 177 (1883).
Coanuita: Carneros Pass area, 1880, Palmer 1357 (type of var. vestita). Cui-
HUAHUA: Pirdmide, gravelly plain near rock masses, Johnston 8132; center of large
grassy plain 7 mi. northeast of La Morita, common, Johnston 7971; base of Sierra
Santa Eulalia, sandy alluvium of streams in rocky hills, Pringle 413.
New Mexico to Arizona and south to San Luis Potosi.
Bouteloua barbata Lag. Var. Cien. 4: 141 (1805).
VERNACULAR NAME: Pata del Cuervo.

CoaHUILA: Sacr inne gravelly arroyo, Johnston 7086; Cuatro migra Marsh
2065; 4 mi. east - Cuatro Cienegas, ditch by road, Johns ton = 17; near Rosario,
shout edge of mogote, eon spreading, Johnston 8824; 20 km rth i. junction of

Monclova and Torreon roads, Harvey 1109; 42 mi. east of Saltillo, Sowue & Tinkham
9837; Saltillo, near ditch, prostrate, 1898, Palmer 400, 401; south base of Picacho San
José, gravelly flat, erect, Johnston & Muller 804; 3 mi. east of San José, silty plain,

HiEOw, Parry; ay mi. south of Oinae. outwash from saline shales, Johnston
8006: 4% mi. northwest of San Francisco, grassy flat, ascending, Stewart &
Seton 2010; Los Medanos, LeSueur 59; Villa Ahumada, flats, LeSueur 65;
Chihuahua, hills and plains, Pringle 490

Texas to southern California and south to southern Mexico. This small
rapidly growing annual grass is widely distributed in our area, especially
about mogotes, ditches, and similar sites where run-off collects after storms
and the soil is moistened at frequent intervals during the summer.
Bouteloua hirsuta Lag. Var. Cien. 4: 141 (1805).

VERNACULAR NAMES: Grama; Navaijitas.

Coanumta: Sierra del Carmen, Aug. 14, 1936 and Sept. 1, er Banh 657, 887;
trail foe Encantada Mesa to Fresno Mesa, July 14, 1938, Marsh 1397; 3 mi. southeast

of Picacho del Fuste, rocky flats, Johnston 8425; Sierra Madera, Cafion Charretera,
openings in brush on rocky flats, Johnston 9058; eastern foothills of Sierra Cruces
near Santa mee rocky flats and slopes, tee & Muller 217, 1385, Stewart 831.
oop ania Sierra Virulento, east of Rancho Virulento, rocky terrace, Johnston
4; Sierra ‘Orsanos, 1937, LeSueur 165; hills and plains near Chihuahua, Pringle 409.
ies distributed in the western United States and south to Guatemala.
Well-drained soils on plains and hillsides. Usually growing with the more
common B. gracilis.
Bouteloua scorpioides Lag. Gen. et Sp. Nov. 5 (1816).
Reported by Griffiths and Hitchcock from Cedros, Zacatecas (Lloyd
105), where it is said to make turf on the plains. Otherwise the species is
known only from central Mexico, from San Luis Potosi to Puebla.

Bouteloua ramosa Scribn. ex Vasey, U. S. Dept. Agric., Div. Bot. Bull. 12!: t. 44
(1890

VERNACULAR NAMES: Chino; Zacate Chino.
CoauvutLa: 20 km. south of Ocampo, gravelly flats at base of mountains, Johnston

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, II 407

9177; Monclova, 1880, Palmer spe Cuatro Cienegas, Puerto del Norte, Harvey 1207;
Cuatro Cienegas, Marsh 2054; near Rancho Santa Teresa, south of Castanos, Wynd &
Mueller 172; 25 mi. south of Mo nclova, limy hillside, Johnston 7200; hills 20 mi. west
of Saltillo, Shreve & Tinkham 9826; Saltillo, rough stony mountain-side, 1898, Palmer
404; west base of Picacho del Fuste, rocky flats, Johnston 8418; west end of Sierra

CHIHU an Sierra Santa Eulalia, a calcareous banks, Fae 14, 1885, Pringle 414;
13 mi. s of Jimenez, Harvey 1339; 31 mi. southeast of Jimenez, Muller 3329.
es gee back of Apizalaya, Hac. Cedros, He 254 (US).

Southern trans-Pecos Texas south into our area. A common and char-
acteristic grass on stony slopes and flats along the base and in the foothills

the limestone mountains of Coahuila, particularly in the Palma Belt,
where it is usually abundant and a major forage grass

In recent treatments of Bouteloua, the present species, B. ramosa, Las
been treated as a synonym of B. breviseta Vasey. These two species, how-
ever, although obviously closely related, differ in distribution, soil prefer-
ence, habit of growth, and in some minor morphological characters. Boute-
loua breviseta is a plant of highly gypsiferous soils in the northern half of
trans-Pecos Texas and in southeastern New Mexico. Its leaves are strongly
involute and its spikes are pale and erect or stiffly ascending. Bouteloua
ramosa, ranging south of B. breviseta, is a plant of stony, prevailingly lime-
stone, flats and slopes. Its leaf-blades are usually nearly flat, with the
upper surface much less hairy than in B. breviseta. The dark-colored
spikes are stouter and much more eee The stems become more
fruticulose, are strictly erect, and form denser tufts
Bouteloua gracilis (H.B.K.) Lag. ex Steud. Nom. Bot. ed. 2, 1: 219 (1840).

VERNACULAR NAMES: Grama; Navajitas.

vita: Sierra del Carmen, Aug. 21 and 29, 1936, Marsh 570, 693; 20 mi.
cere of Hac. La Babia, open valley floor, Wynd & Mueller 435; trail from
southern end of Hillcoat Mesa to Buena Vista headquarters, July 27, 1938, Marsh 1506;
Mesa Grande, 40 km. northwest of Hac. Encantada, meadows, abundant, Stewart 1632;
Saltillo, stony mountain-side, 1898, Palmer 399, 406; Saltillo, along ditches, 1898,
almer 4 SRY rae i

2-3% ft. tall, Johnston 8424; Sierra Ma dera, Canon Charretera, open ‘rocky canyon
floor, 2-3 ft. tall, Johnston 9159; tableland north of Cafion del Cuervo Chico, rocky
slopes of low limestone hills, Johnston 8560; between Palos Blancos and San Pedro,
: ;

Laguna Jaco, common, Stewart & Johnston 1964, 6. CurmuAHuaA: Rancho El
terrace, Johnston 9082; 2 mi. west of Pozo de Villa, sabaneta under low bushes,
2011; Sierra Organos, 1937, LeSueur 166; White Water, 1893, Mearns 2301; Villa
Ahumada, flats, LeSueur 66; near Chihuahua, Pringle 407 ; 6 mi. west of Piloncillo, with

tobosa in grassland, Jobin 7878. Zacatecas: Valley 15 km. west of Concepcion
del Oro, Stanford et al. 522; Pico de Teira, Lloyd 242 (US).

408 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

Widely distributed in the western United States and south to central
Mexico. A generally distributed grass in our area. It is abundant in some
of the larger valleys and on the igneous oak-clad hills of eastern Chihuahua,
where it may become the dominant plant over large areas. In the limestone
areas of Coahuila it is rather common on the tablelands and larger valleys
in the oak and lower pine belts. On the lower slopes of the limestone
mountains it is frequent with other grasses among the bushes on rocky flats
and slopes in the Palma Belt. It avoids clay, and along the foot of lime-
stone mountains it is found only where the soil is stony.

Bouteloua eriopoda Torr. Pac. R. R. Rep. 4: 155 (1857).
ILA: Sierra del Carmen, Sept. 13, 1936, Marsh 896 ; west base of Picacho del

5 e
grayish, erect or ascending, Johnston 8718; eastern foothills of Sierra Cruces near
Santa Elena, gypsum flat, Johnston & Muller 245; Sierra Cruces, 8 mi. north of Santa
Elena, stony flat among bushes, stems laxly ascending, Johnston & Muller 1021, 1025;
south base of Picacho San José, rocky slope, stems wiry, ascending, Johnston & Muller
802. CHIHUAHUA: 1 mi. east of Pozo de Villa, silty plain, among bushes, Johnston
8176; Presidio del Norte, Sept. 1, 1852, Bigelow; east base of Sierra Virulento, rocky
bench, Johnston 8085; 41% mi. northwest of San Francisco, fairly abundant on grassy
flat, Stewart & Johnston 2012; Sierra Organos, 1937, LeSueur 162; Sierra Santa Eulalia,
Aug. 31, 1885, Pringle 411; south of San Fernando, silty soil on plain, common,
Johnston 7937 ; 6 mi. west of Piloncillo, lava slope, Johnston 7875.

Western Texas to Arizona and northern Mexico. A generally distributed
but not abundant grass in northern Coahuila and eastern Chihuahua. Its
pallid lax loosely ascending frequently somewhat sprawling stems and
woolly leaf-sheaths give the plant a very distinctive appearance.

Bouteloua Karwinskii (Fourn.) Griffiths, Contr. U. S. Nat. Herb. 14: 394 ene

CoAHUILA: hss ey floor 3-4 km. east of Puerto Caballo, with tobosa, common,
Johnston 8319; south of Laguna de ae flats formed by planed- off Upper /evenette
beds, slightly fee and gypsiferous soil, Johnston 8618; 1-2 mi. west of Matrimonio
Viejo, about mogote at base of slightly saline and gypsiferous slope, Johnston 9370.
Zacatecas: Cedros, Lloyd 170, 183 (US); 7 mi. north of San Tiburcio, heavy slightly
saline soil on flats with mesquites, abundant, Johnston 7358.

Ranging from our area south and east to southern Tamaulipas (Canon
de las Minas et Victoria, Karwinski 1479, type) and San Luis Potosi.
Griffiths, l.c., collected the species at Alonzo, east of San Luis Potosi, and
gives an illustration, pl. 76, of the habitat. The species appears to be con-
fined to slightly saline and gypsiferous clays. In habit it suggests small
slender plants of B. gracilis, but it has smaller pale spikes.

Bouteloua trifida Thurb. in Wats. Proc. Am. Acad. 18: 177 (1883).

Coanutta: Road to Don Martin Dam, 9 km. from Nuevo Leon border, Harvey
925; Allende, set 1788; Palm Canyon, near Muzquiz, Marsh 977; Yerda Spring,
Marsh 290; Santa Anna Canyon, Marsh 467; Zacate, Marsh 503; Hermanas, Marsh
7 2 mi. ca of Fronteras, road to Natadores, silty desert plain, Johnston

- Monclova, 1880, Palmer 1355 (Type); Cafion Bocatoche, dominant grass of
pit valley-floor, clumps 3 in. in diameter, Muller 3110; on desert near Rancho Santa
Teresa, south of Castafios, Wynd & Mueller 205; dry desert between Hac. La Rosa

tain, 1898, Palmer 402; Saltillo, 1905, Palmer 522; base of mountains 20 km. south of

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 409

Ocampo, one plant in mogote, Johnston 9179; south of Laguna de Leche, slightly saline
and gypseous silty flat, Johnston 8619; Sierra Cruces, north of Santa Elena, rocky
flat among bushes, Johnston & Mudler 1020, 1379. CHIHUAHUA: Base of Sierra Santa
Eulalia, dry gravelly soil, Pringle 412.

Western Texas to southern Nevada and Arizona, south to San Luis
Potosi.

Bouteloua aristidoides (H.B.K.) Griseb. Fl. Brit. W. Ind. 537 (1864).

COAHUILA: eae 1880, Palmer 1354; Monclova, 1880, Palmer 1353; Torreon,
abundant in deep cut along railroad, 1898, Palmer 513. CHIHUAHUA: ancho
‘Pino, southeast of Sierra Rica, rocky slope, Stewart 2393; Piramide, gravelly flats
under large oaks, Johnston ae low ridge southwest of Mestenas » rocky slope,
ascending, Stewart & Johnston 2022; Sierra Organos, 1937, LeSueur 39; 11 mi. north-
east of Camargo, gravelly benches, common, Johnston 7904; Jimenez, banks of Rio
Florido, Harvey 1323; Cafion La Renga, 15 km. northwest of Santa Fe, dry arroyos,
ascending, Stewart 2620

Western Texas to southern California and south to central Mexico;
South America

Bouteloua chondrosioides (H.B.K.) Benth. ex Wats. Proc. Am. Acad. 18: 179
(1883)

CHIHUAHUA: Volcanic hills 20 km. north of Chihuahua, locally common along base
of rocky slope, Stewart & Johnston 2129; rocky hills northeast of Chihuahua, forming
close sod on small patches, Pringle 410

Trans-Pecos Texas to Arizona and south to southern Mexico. Appar-
ently favoring igneous rocks
Bouteloua rigidiseta (Steud.) Hitchc. Jour. Wash. Acad. 23: 453 (1933).

Coauurta: Calcareous mesa near Piedras Negras, April 20, 1900, Pringle 8018.

Oklahoma south through central and eastern Texas into adjacent Mexico.
Bouteloua radicosa (Fourn.) Griffiths, Contr. U. S. Nat. Herb. 14: 411 (1912).

AHU Don Martin Dam, Harvey 934; Caracol pe 1880, Palmer 1354;
Puerto San Le common on rocky arroyo banks, Muller

Coahuila to Arizona and south to southern ae
Bouteloua Johnstoni Swallen, Proc. Biol. Soc. Wash. 56: 79 (1943).

CoanHuita: South end of Cafiada Oscuro, confined to gypsum beds on the escarp-
ment near Tanque La Luz, locally very common, Johnston 8491 (1soTyPE) ; high west
end of the Sierra Fragua, north on Puerto aaa ~ large colony on east slope just
below high crest, Johnston 8751; 1 km. northeast of Parritas, east side of Valle
Acatita, common on gypsum oe Stewart 2763.

A very distinct species, of which only the three collections cited above
are known. Near Tanque La Luz in Canada Oscuro, and a mile or so to
the southeast, on the steep north-facing slopes at the mouth of Canon del
Cuervo Chico, the grass was common on all the gypsum beds exposed on the
escarpment. Its behavior was that of a marked gypsophile. On the steep
west-facing slope up which I climbed from near Aguaje Pajarito to the
high western crest of the Sierra Fragua, I found the grass again common.
Here it was confined to a sharply delimited belt over a hundred feet wide.
It was associated with some gypsum indicators, but the soil on which it grew
was not pure gypsum, only very moderately gypseous at most. My collec-
tion no. 8751 came from a small colony near the ridge crest, and the only

410 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

one observed away from the belt of the plant just mentioned. The soil
gave no indications of being gypsiferous and no recognized gypsophiles were
growing with it. Mr. Stewart’s collection from near Rancho Parritas came
from gypsum.

The plant has a very distinctive habit of growth. The stems, 15—45 cm.
long, are ascending or decumbent. The clump appears to die in the middle
and soon forms loose spongy rings of growth 5-10 dm. in diameter.
Bouteloua curtipendula (Michx.) Torr. in Emory, Notes Mil. Recon. 154 (1848).

Coauuma: Desert 25 mi. southwest of Sabinas, Wynd o Mueller 217; Sierra del

v
=]
~
is)
n
N
w
~
5B
wn
°
oc
pe
a
=
oO
wn
pe
o
al
fe]
=
a
oO
<4
a
=
~~
=)
po
a
s:)
~
Q
~~
w
Do
NI
Nn
oO
=
—
p
Q
Ln}
Cc
ie)
oO
hg
=
Oo
sot)
ad
op)
pP
=]
ie)

Elena, abundant on open hillsides, up to 10 dm. tall, Stewart 832; San Antonio de los
Alamos, gravelly flat on top of cliffs, Johnston 8254; west end of Sierra Fragua near
Aguaje Pajarito, rocky flats, Johnston 8792; Sierra. Negras, 9 km. south of Parras,
Stanford et al. 195; 11 km. northeast of Jimulco, ier A et al, 51. CHIHUAHUA:
Rancho El Pino, southeast of Sierra Rica, rocky slopes, Stewart 2406; Chihuahua,
rocky hillsides, 1908, Palmer 114; Chihuahua, hills and plains near Chihuahua, Aug. 27,
1885, Pringle 408. Zacatecas: Concepcion del Oro, 1904, Palmer 264; Cedros, ioot-
slopes, 1908, Lloyd 201.

Widely distributed in central and eastern United States and south to
Central America; South America. A common and widely distributed grass
in our area, frequent in the grasslands of eastern Chihuahua. In the lime-
stone mountains of Coahuila, with B. ramosa, forming the best pasturage
for horses and cattle on the rocky slopes and flats in the foothills, and with
B. gracilis the good pasturage on the tablelands and mountain valleys. In
the eastern parts of our area the plants tend to have few spikelets in each
spike and to intergrade with B. uniflora.

Bouteloua uniflora Vasey, Bot. Gaz. 16: 26 (1891).

CoAHUI Trail from Encantada Mesa to Fresno Mesa, July 20, 1938, Marsh
1376; Sierra del Pino, La Noria, meadow at lower edge of pine belt, Johnston & Muller
450, Stewart 1217; Sierra Madera, Cafion Charretera, openings in oak thickets on rocky
flat, sopgitt 9060; Saltillo, 1906, Griffiths 8408; Carneros Pass, tufts among bushes,
Johns 7299. oo Valley 15 km. west of Concepcion del Oro, plant 18 inches
tall, pitt et al. 5

Known from Crockett and Val Verde Counties, Texas, south through
eastern Coahuila to southwestern Nuevo Leon. is species is very
closely related to B. curtipendula and is perhaps no more than an extreme
form of that species. The material cited above agrees closely with the type
collection. In all the specimens, only a single spikelet is borne at the base
of the spike-rachis, which is prolonged above and distinctly overtops the
lower glume of the spikelet. Material of B. curtipendula from eastern
Coahuila and Nuevo Leon frequently has reduced spikes and some speci-
mens have occasional spikes bearing only a single spikelet. In these spikes
the rachis is slightly shorter than or about equal to the first glume, and
it does not distinctly surpass it, as in typical B. uniflora. Since the eastern

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 411

material of B. curtipendula tends to have the dark-colored spikelets usually
present in B. uniflora and to simulate that species in habit, the differences
seem more technical than important.

Cathestecum erectum Vasey & Hack. Bull. Torr. Bot. Cl. 11: 37. t. 45 (1884).

CHIHUAHUA: Presidio del Norte, Bigelow, Parry; 10 mi. south of Ojinaga, one
colony in ravine in low hills, Johnston 8018.

In Texas known from the Big Bend area, near Presidio, and near
Porvenir. The plant in Texas and adjoining Mexico seems confined to
areas of gypseous, frequently saline, Upper Cretaceous clays and shales.
The type came from Presidio, Texas. The species is also reported from
the Pacific slope of Mexico, from Sonora to El Salvador, cf. Swallen, Jour.
Wash. Acad. 27: 500 (1937).

Munroa squarrosa (Nutt.) Torr. Pac. R. R. Rep. 4: 158 (1857).

CuiHuAHUA: Sandy soil near Juarez, Sept. 26, 1902, Pringle; near Colonia Diaz,
Nelson 6457.

Central United States south into Chihuahua.

Buchloé dactyloides (Nutt.) Engelm. Trans. Acad. Sci. St. Louis 1: 432 (1859).

Coanuita: Saltillo, banks of ravine in compact mat, 1898, Palmer 7; 2 mi. west
of Saltillo on road to Torreon, Harvey 1087; valley near Fraile, Stanford et al. 271, 288.
CHIHUAHUA: Rancho E] Pino, southeast of Sierra Rica, wet sandy arroyo, common,
Stewart are 10 km. east of Jimenez, Harvey 1346. Zacatecas: Concepcion del Oro,
cemetery, 1904, Palmer 267; valley 15 km. west of Concepcion del Oro, Stanford et al.
552; Cedros, sinks and flats, Lloyd 211

Central United States and south in eastern Mexico to Puebla.

Phalaris canariensis L. Sp. Pl. 54 (1753).

CoauurILa: Monclova, 1939, Marsh 1693.

A Mediterranean species, introduced in various parts of America.
Phalaris caroliniana Walt. Fl]. Carol. 74 (1788).

OAHUI Muzquiz, 1936, Marsh 1078. CHIHUAHUA: Common, forming large

pure stands in low meadows and along ditches, 1908, Palmer 32.

Widely distributed in the southern half of the United States and in
adjoining Mexico.
Trichachne insularis (L.) Nees, Agrost. Bras. 86 (1829).

CoanuiLa: Santa Anna Canyon, Marsh 430; hills 20 mi. west of Saltillo, Shreve &
Tinkham 9829. CuHinuAHUA: Rocky hills near Chihuahua, Aug. 1885, Pringle 378.

Florida to New Mexico and south to Argentina. A coarse plant in the
tropics but becoming small in stature in our area and frequently rather
similar in aspect to 7. californica, but readily distinguished from it by its
proportionately narrower, lanceolate spikelets, bearing sordid or tawny,
rather than pure white or purplish, hairs.
Trichachne Hitchcockii Chase, Jour. Wash. Acad. 23: 454 (1933).

Coauuita: Sierra Cruces, limestone hillside just west of Santa Elena, rare, John-
ston 8195

Known from a few collections from Texas (San Antonio west to Sander-
son) and south to San Luis Potosi.

412 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Trichachne californica (Benth.) Chase, Jour. Wash. Acad. 23: 455 (1933).
Martin Dam, Harvey 935; Hermanas, Marsh 1626; Monclova,

ston 8570; Rosario, among bushes in mogote, Johnston 8825; 14 mi. east of Paila,
Shreve & Tinkham fone north of Sierra Cruces, about mogote west of San Rafael,
Johnston & Muller 1040. CurmuanuaA: Rancho El Pino, southeast of Sierra Rica,
rocky slope, Stewart 2409; Coahuilan boundary 1 mi. east of Poza de Villa, among
bushes on silty plain, Johnston 8173; Chihuahua, 1935, LeSueur 77.

Texas to Colorado and Arizona, and south to central Mexico.
Digitaria sanguinalis (L.) Scop. Fl. Carn. ed. 2, 1: 52 (1772).

CoaHuILa: Muzquiz-La Mariposa, 1936, Marsh 1049; Monclova, 1939, Harvey
1165; Saltillo, 1898, Palmer 387. CHIHUAHUA: Presa de Chihuahua, 1936, LeSueur
142; 5 km..west of Camargo, 1939, Harvey 1415.

Widely introduced European garden weed.

Leptoloma cognatum (Schult.) Chase, Proc. Biol. Soc. Wash. 19: 192 (1906).

Coauuita: Sabinas, his 6822 (US); Santo Domingo, open s slopes of igneous an

Sie

in low bushes on grassy bate — common, Stewart & Joknston 2047 ; Chihuahua,
hills and plains, Pringle 4

Eastern United Sates to Minnesota and Texas and west along the
boundary to Arizona; south through eastern Mexico to San Luis Potosi.
Eriochloa gracilis (Fourn.) Hitchc. Jour. Wash. Acad. 23: 455 (1933).

Coauutta: Torreon, under bushes on banks of Rio Nazas, 1898, Palmer 509.
CHIHUAHUA: Cieneguita, wet arroyo bottom, Johnston & Muller 1417; Los Medanos,
1935, LeSueur 64; near the Sacramento, Chihuahua, Sept. 16, 1886, Pringle 812;
20 km. south of Camargo, Harvey 1385.

Texas to Arizona and south to Central America.

Eriochloa punctata (L.) Desv. ex Hamilt. Prodr. Ind. Occ. 5 (1825).

Coanvuita: Monclova, Harvey 1147.

Louisiana and eastern Texas south in eastern Mexico to Vera Cruz; South
America.

Brachiaria Meziana Hitchc. Contr. U.S. Nat. Herb. 12: 140 (1908).

CoaHUILA: Saltillo a, irrigation ditch, cease Hitchcock 402. CHIHUAHUA:
Valley near Chihuahua, ‘Sink , 1885, Pringle :

Coahuila and ete nuk to oa.

Paspalum distichum L. Syst. Nat. ed. 10, 2: 855 (1759).

AHUILA: Saltillo, pide Palmer 259, 391. CHIHUAHUA: Rancho El Pino, ie
east of Sierra Rica, wet sandy arroyo, manele 2401; 5 km. west of Camargo, Harve
1407. DuRANco: salary 1898, Palm

Along streams and ditches ak other wet places, widely distributed in
America.
ae paca Rupr. ex Fourn. Mex. P]. 2: 11 (1886).

Coa : erda Spring, Marsh 292; Santa Anna Canyon, Marsh 435; Palm
7a "Marsh pore Muzquiz, Marsh 1156; Mesa Grande, ee of Hac. Encan-
tada, wet arroyo, fairly common, Stewart 1619; Hermanas, Marsh 2259; Monclova,
Marsh 1721; 50 km. south of Monclova, Harvey 1126; Sierra Hechiceros, Cafion Indio
Felipe, creek bank, Stewart 91; Sierra Cruces, Cafion Tinaja Blanca, arroyo bank,

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 413

erect, Stewart 1133; Torreon, 1898, Palmer 515; Jimulco Springs, May 13, ae

Pringle 427, CurmuaHvuA: Sierra Almagre, Ojo Almagre, about spring, becom

6 ft. tall, Johnston & Muller 1201; near Chihuahua, by streams, Sept. 1885, pee

374; 5 km. west of Camargo, Harvey 1405a.

Louisiana and Texas south to southern Mexico.

Paspalum crinitum Chase in Hitchc. Contr. U. S. Nat. Herb. 17: 237 (1913).
CoaHuita: Chojo Grande, 27 mi. southeast of Saltillo, about summit of waterfall

in moist place, 1904, Palmer 338.

Reported by Chase, Contr. U. S. Nat. Herb. 28: 61 (1929) from
Coahuila, San Luis Potosi, Jalisco, and Puebla.

Paspalum Hartwegianum Fourn. Mex. Pl. 2: 12 (1886).

Coauvuita: 24 km. east of Don Martin Dam, Harvey 949.

Texas south to southern Mexico.

Paspalum mutabile Chase, Contr. U. S. Nat. Herb. 28: 61 (1929).

CoauvuiILa: Palm Canyon, Muzquiz area, Marsh 978

Northeastern Mexico south to Hidalgo.

Paspalum ciliatifolium Michx. Fl. Bor. Am. 1: 44 (1803).

CHIHUAHUA: Los Medanos, 1935, LeSueur 62.

A very variable species of sandy soils, widely distributed in the eastern
half of the United States; reported from the West Indies and Central
America. The cited collection is an unusually glabrous plant and keys to
P. propinquum in Chase’s monograph. I am, however, perfectly content
to refer it to typical P. ciliatifolium, cf. Rhodora 36: 21 (1934). Chase
cites a collection of P. stramineum Nash from near Juarez, Chihuahua
(“Paso del Norte, Pringle 1123”). This is presumably a form of the pres-
ent species with puberulent foliage and somewhat hairy fruit, i.e. the var.
stramineum (Nash) Fernald.

Panicum ramisetum Scribn. U.S. Dept. Agric., Div. Agrost. Cir. 27:9 (1910).
Coauuita: Rio Grande Valley near Piedras Negras, April 23, 1900, Pringle 8323.
Texas south into adjoining Coahuila. The present species probably

should be united with P. Reverchoni Vasey (1889), an older species having

practically the same geographical distribution.

Panicum lanuginosum Ell. var. Lindheimeri (Nash) Fernald, Rhodora 36: 77

(1934).

Coanutra: Mu Lease oe 1171; Canon Agua Grande, west of Las Delicias, by
water, erect, Stewart

A phase of a ae species widely distributed in the eastern United
Sales Although ranging with the other variants of the species farther
northward and eastward, the present glabrous form is the most common,
if not the only, phase of the species in south central and trans-Pecos Texas.

Panicum oligosanthes Schultes, var. Scribnerianum (Nash) Fernald, Rhodora 36:
0) a 4).

OAH Sierra del Carmen, Cafion Sentenela, Wynd & Mueller 519; Sierra
mae Canon Indio Felipe, common on creek bank, Stewart 74.
Widely distributed in the United States; in Mexico known only from
Coahuila

414 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Panicum —_ Vasey, U. S. Dept. Agric., Div. Bot. Bull. 8: 28 (1889).
Coa Sierra Madera, Cafion Charretera, common in gravelly bed of arroyo in
oak or Johan 8919.
Central Texas and eastern Coahuila.
Panicum fasciculatum Sw. Prodr. Veg. Ind. Occ. 22 (178
Panicum fasciculatum var. reticulatum Beal, Grasses No. an 2: 117 (1896).
CuinvuaAnua: Plains near Chihuahua, oe 2, 1885, Pringle 379, 380; 6 mi. west of
Piloncillo, low place in grassland, Johnston
Florida; Texas to Arizona and Hen into South America.

Panicum arizonicum Scribn. & Merr. U. S. Dept. Agric., Div. Agrost. Cir. 32: 2
(1901)

Coanvutta: San Antonio de los Alamos, summit of tuff cliffs, gravelly flat, two
plants only, Johnston 8250. Cuimuanua: 7% mi. south of Piramide, silty flat,

abundant on gentle grassy slope, Stewart & Johnston 2034; north of El Carmen, 1935,
ueur 69; Chihuahua, Pringle 487 (US); Meoqui, 1935, LeSueur 37; 20 km. sou uth
of Camargo, Harvey 1391

Trans-Pecos Texas to southern California and south in western Mexico
to Oaxaca.

Panicum hirticaule Presl, Rel. Haenk. 1: 308 (1830).

Coauuma: Rancho Las Uvas, east side Valle Acatita, shale on slope, Stewart 2701.
Curmvuanua: Llano Chilicote, 7 mi. east of ers Station, grassy flat, Johnston
7992; Chihuahua, 1935, LeSueur 12; “Meo oqui, 1935, LeSueur 32; 20 km. south of
Camargo, Harvey 1378; 6 mi. west at Piloncillo, he hillside, Johnston 7866.

Texas to southern California and south to South America. An annual
species, becoming large and coarse in the tropics. Our reduced northern
form is 1-3 dm. tall.

Panicum filipes Scribn. ex Heller, Contr. Herb. Frankl. Marsh. College 1: 13 (1895).

CoauurLa: El Berrendo, Harvey 1186.

Texas and eastern Coahuila.

Panicum Hallii Vasey, Bull. Torr. Bot. Cl. 11: 64 (1884).
Coanutta: Yerda Spring, Marsh 288; Santa Anna Canyon, Marsh 427; Hillcoat

Palmer 1338; hills 20 mi. west of Saltillo, Shreve & Tinkham 9821; Sierra del Pino,
Noria, meadows and arroyo-bank, Johnston & Muller 467, 694, Stewart 1211;
Sierra Madera, Cafion Charretera, ledges on sunny slope in oak belt, Johnston 9104;

on top of tuff cliffs, Johnston 8250; Sierra Cruces, 8 mi. north of Santa Elena, stony
flat, Johnston & Muller 1017; 7 mi. south of Jaco, about mogote, Johnston & Muller
1111. CHrnuAHuA: Rancho El Pino, southeast of Sierra Rica, rocky slope, Stewart
2395; Coahuilan boundary a mile east of Pozo de Villa, silty plain, Johnston 8175;
Sierra Santa Eulalia, Aug. 1885, Pringle 376

Texas to Arizona and south to Hidalgo and Durango.
Panicum lepidulum Hitchc. & Chase, Contr. U. S. Nat. Herb. 15: 75 (1910).

Based upon Pringle 487, collected Sept. 22, 1885, by stream in rocky
hills near (west of) Chihuahua, the type, and on material from Durango
and the Federal District. I have not seen the type collection. The other

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 415

specimens originally cited, however, suggest that it may possibly be only
a form of P. Ghiesbreghti Fourn.
Panicum bulbosum H.B.K. Nov. Gen. et Sp. 1: 99 (1815).

CoaHvuILA: Sierra del Carmen, Aug. 26, 1936, Marsh 630; Sierra del Carmen, Cafion
Sentenela, Wynd & Mueller 514; Sierra Gloria, Marsh 1875. ee River
canyon west of Chihuahua, moist places, Aug. 7, 1885, ee

Trans-Pecos Texas to Arizona and soit) to Oaxaca.

Panicum virgatum L. Sp. Pl. 59 (1753).
Panicum plenum Hitchc. & Chase, Contr. U. S. Nat. Herb. 15: 80 (1910).
ILA: Sierra del Pino, head of Cahon Ybarra, dry hillside, Stewart 1248;
Sierra del Pino, La Noria, arroyo banks, Johnston & Muller 465. CHIHUAHUA:
Piramide, low ground, coarse clumps 2—4 ft. tall, Johnston 8140
United States, except the Pacific States, south to Central America.

Panicum Havardii Vasey, Bull. Torr. Bot. Cl. 14: 95 (1887).

CHIHUAHUA: Los Medanos, 1935, LeSueur 82.

Extreme western Texas, eastern New Mexico, and northern Chihuahua,
in sandy places. Hitchcock & Chase, Contr. U. S. Nat. Herb. 15: 94
(1910), report a collection from ‘Paso del Norte, Pringle 1124.”
Panicum agrostoides Spreng. Pl. Pugil. 2: 4 (1815).

Panicum condensum Nash in Small. FI. S. E. U.S. 93 (1903).

CoaHutLa: Muzquiz Swamp, 1936, Marsh 928.

Eastern United States and northeastern Mexico.

Panicum obtusum H.B.K. Nov. Gen. et Sp. 1: 98 (1816).

VERNACULAR NAME: Zacate Gramilla.

CoaHuILA: Open country between Rancho Santo Domingo and Hac. Piedra Blanca,
Wynd & Mueller 486; Santa Anna Canyon, Marsh 463; Yerda Spring, Marsh 964;

Saltillo, in graveyard, 1898, Palmer 394; 24 km. northwest of Fraile, south slope of
mountain, Stanford et al. 397a; La Ventura, Nelson 3908 (US); Torreon, overflowed

huahua, Pringle 476 (US); south ae Sage 1936, LeSueur 131; 10 km. east of
Jimenez, Harvey 1349. Zacatecas: Concepcion del Oro, ape 1904, Palmer 266;
mountain 18 km. west of Concepcion del Oro, Seated et al. 5

Missouri and Texas to Colorado and Arizona, ae south to central
Mexico. Growing in wet soil or low places where storm water frequently
collects.
Oplismenus hirtellus (L.) Beauv. Ess. Agrost. 54, 168 (1812).

CoauuiILa: Palm Canyon near Muzquiz, Sept. 19, 1936, Marsh 985.

Northern Mexico to Argentina.
Echinochloa colonum (L.) Link, Hort. Berol. 2: 209 (1833).

Coauuira: Sierra del Carmen, Aug. 29, 1936, Marsh 699; Santa Anna Canyon,

416 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Marsh 428, 431; re River, Muzquiz, Marsh 405; El Berrendo, Harvey 1179;

bottom-land, 1898, Palmer 454; Torreon, under bushes along Rio Nazas, 1898, Palmer
508. Curnuanua: Presidio del Norte, Bigelow; Presa de Chihuahua, LeSueur 134;
Sierra Organos, LeSueur 176; Jimenez, banks of Rio Florido, Harvey 1330.

A European weed, widely distributed in gardens and bottom-lands.
Echinochloa Walteri (Pursh) Heller, Cat. No. Am. PI. ed. 2, 21 (1900).

Coanuita: Muzquiz Swamp, Marsh 891.

Eastern United States south to Texas and northern Coahuila.
Echinochloa crus-galli (L.) Beauv. Ess. sine 53 (1812).

CoanvurLa: Sierra del Carmen, Sept. 8, 1936, Marsh 751; Hermanas, Marsh 1576;
Cuatro Cienegas, 1939, Marsh 2023; Sin 1898, Palmer 380, 418. CHIHUAHUA:
Presa de Chihuahua, LeSueur 135.

Widely distributed weedy plant in gardens and wet soils.

Chaetium bromoides (Pres!) Benth. ex Hemsl. Biol. Centr. Am. Bot. 3: 503 (1885).

Hitchcock, Contr. U. S. Nat. Herb. 17: 259 (1913), reports this species
from “CHIHUAHUA: Chihuahua, Palmer in 1886.” The data on the speci-
men is possibly erroneous. Except for this specimen, the species is known
only from central Mexico south into Central America.

Setaria lutescens (Weigel) Hubbard, Rhodora 18: 232 (1916).

CuHIHvuAHUA: 5 km. west of Camargo, 1939, Harvey 1413.

A European weed, widely distributed in temperate North America.
Setaria geniculata (Lam.) Beauv. Ess. Agrost. 51, 178 (1812).

CoanvuILa: 20 mi. northwest of La Babia, open valley floor, Wynd & Mueller 431;
Santa Anna Canyon, Marsh 434; Sabinas River, Muzquiz, Marsh 396; Muzquiz Swamp,
Marsh 936; valley below Saltillo, frequent, 2 ft. tall, Sept. 23, 1848, Gregg 536; Saltillo,
1898, Palmer 383; Parras, 1898, Palmer pats Jimulco Springs, May 13, 1885, Pringle
431. CHimuAHUA: Chihuahua, LeSueur

Southeastern United States west to as and south to Argentina. The
type of Chaetochloa gibbosa Scribn. & Merr., referred to C. macrostachya
by Hitchcock, Contr. U. S. Nat. Herb. 22: 204 (1920), properly belongs in
the synonymy of the present species.

Setaria verticillata (L.) Beauv. Ess. Agrost. 51, 178 (1810).

COAHUILA: viene 1939, Marsh 1841; ees shaded garden, 1898, Palmer 453.
CHIHUAHUA: Presa de Chihuahua, 1936, LeSueur I

A European pee
Setaria Grisebachii Fourn. Mex. Pl. 2: 45 (1886).

Coanvuita: Sierra del Carmen, Sept. 8, 1936, Marsh 762; Jardin del Sur, Sept. 3,
1936, Marsh 776; Sierra Guajes, Cafion Madera, east of Buena Vista, hillside, Stewart

oO

shade of rocky ledges, 1904, Palmer 336, 337; Sierra del Pino, La Noria, hillsides and
meadows, Stewart 1206; San Antonio de los Alamos, talus at base of cliffs, Johnston
— CuiHuAHUA: 7 mi. northwest of Temporales de Honorato, abundant in
mogote, Stewart & Johnston 1989; Chihuahua, LeSueur 25; rocky hills northeast of
Caboshia. baded places, Sept. 16, 1885, Pringle 381.

1943] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 417

Texas to Arizona and south to southern Mexico. A native annual, very
suggestive of S. verticillata in general habit.

Setaria macrostachya H.B.K. Nov. Gen. et Sp. Bs 110 (1816).
Chaetochloa leucopila Scribn. & Mer. U. S. Dept. Agric., Div. Agrost. Bull. 21: 26
900).

(1
Setaria leucopila Schum. in Just’s Bot. Jahresb. 281: 417 (1902).
Coanuita: Allende, Marsh 1791; Sierra del Carmen, Aug. 8, Sept. 6 and 13, 1936,
Marsh 687, 852, 903; northwest of Hac. La Babia, open valley floor, Wynd & Mueller
441; Cafion Milagro, Sierra Guajes, shaded places, Stewart 1707; trail from south end

1363 (isotype of S. lence play: Parras, 1898, Palmer 449. CHIHUAHUA: Silty plain on
Coahuilan boundary 1 mi. east of Pozo de Villa, Johnston 8174; Rancho El Pino, south-
east of Sierra Rica, rocky slope, Stewart 2396; Los Medanos, 1935, LeSueur 55; 20 km.
south of Camargo, Harvey 1383. Zacatecas: Concepcion del Oro, 1904, Palmer 261.

Central and southern Texas to Arizona and south to Central America. A
very variable species in height, leaf-width, and size and form of the spike,
apparently in response to varying ecological conditions. Our common form
has the leaves less than 8 mm. broad, the spike cylindrical, and the plant
usually 4-8 dm. tall. It has been described as S. leucopila. More vigorous
plants have leaves 10-15 mm. wide, the plant over a meter tall, and the
spike cylindrical or broadest below the middle and more or less attenuate
above. Typical S. macrostachya is this large form with cylindrical spike.
From the large forms with attenuate inflorescence there are numerous
transitions to S. Scheelei, a more robust plant with more or less branched
open inflorescence.
Setaria Scheelei (Steud.) Hitchc. Proc. Biol. Soc. Wash. 41: 163 (1928).

LA: Torreon, 1898, Palmer 505. CutHuaHua: Ojo Almagre, Sierra Almagre,

about spring, 6 ft. tall, Johnston & Muller 1202

Texas and northeastern Mexico. Setaria villosissima (Scribn. & Merr.)
Schum. is probably a synonym of this species.
Cenchrus myosuroides H.B.K. Nov. Gen. et Sp. 1: 115 (1816).

uiLA: Santa Anna Canyon, Marsh 437; Saltillo, Hitchcock 5647 pen
i km

t=
°
“
au
a od
=
Oo
aw
pat
°
pow
2)
p
i=)
A
»
&
oO
i=}
=
p
3
jo)
3
Pie]
=
xe)
(a)
oe
ta
pat
=]
wn
a
i)
[o*
o
Qa
fe)
it)
=
i
°
5
=)
°
=
ra)
°
°
Ze
o
&
8
2
>
bo
a
Wa
ery

CuHinuaAHua: Wet places near Chihuahua, ware 25% 1885, Pringle 429; Meoqui, 1936,
LeSueur 140; 60 km. north of Escalon, Harvey
Texas (Uvalde to Brewster pees south through Mexico to South
America.
Cenchrus echinatus L. Sp. Pl. 1050 (1753).
schahaaiae aera 1880, Palmer 1343; Monclova, 1939, Marsh 1831.
Chase, Contr. U. S. Nat. Herb. 22: 61 (1920), reports the species from
Torreon ae 7558). Texas to Arizona and south into tropical
merica
Cenchrus eewellioens Benth. Bot. Voy. Sulphur 56 (1840).
VERNACULAR NAME: Chancaquilla.

418 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Coanurta: On desert 25 mi. southwest of Sabinas, Wynd & Mueller 218; Rancho
Agua Tiles: valley floor, Wynd & Mueller 404; 20 mi. northwest of Hac. La Babia,
valley floor, Wynd & Mueller 444; Yerda Spring, 1936, Marsh 287; Hac. Encantada,
abundant on flats, 1941, Stewart 1733; Monclova, 1939, Marsh 1823; near Esmeralda,
fairly common along arroyo, Stewart 2179; Cafion Agua Grande, west of Las Delicias,
on flats, common, Stewart 2825. Cuimuanua: 1935, LeSueur 7; 20 km. south of
Camargo, 1939, Harvey 1393.

Chase, Contr. U. S. Nat. Herb. 22: 71 (1920), reports the species from
Saltillo. This species is probably indigenous to Mexico and Texas and
was formerly much less generally distributed than at present. It has be-
come a widely distributed and obnoxious weed along roads and about towns
in waste ground. Chase refers the common, apparently indigenous
Cenchrus of central United States to C. pauciflorus, but that seems doubt-
fully correct, for the broader darker green leaves and the shape and arma-
ture of the burs of that plant are more suggestive of C. echinatus.
eo brevifolia Vasey, Bull. Torr. Bot. Cl. 13: 26 (1886).

Imperata Hookeri (Anderss.) Hack. in DC. Monogr. Phan. 6: 97 (1889).

calmed in the bottoms of the Rio Grande on the Texan side of the
river between El Paso and old Fort Quitman, and almost certainly to be
found on the Chihuahuan side also.

Andropogon hirtiflorus (Nees) Kunth, Rév. Gram. 1: Suppl. xxxix (1830).

Coanumta: Sierra del Carmen, Canon Sentenela, Wynd & Mueller 548. Cut-
HUAHUA: Sierra Organos, common on oak-clad slopes south of Organos, Stewart &
Johnston 2069; rocky hills northeast of Chihuahua, Aug. 29, 1885, Pringle 383.

Trans-Pecos Texas to Arizona and south into tropical America. Our
plants represent the Mexican var. feensis (Fourn.) Hitchc. The plant
closely resembles A. scoparius and A. cirratus, from which it differs chiefly
in the scabrid strigose hairy glumes.

Andropogon cirratus Hack. Flora 68: 119 (1885).

Curmuanva: Rocky hills northeast of Chihuahua, Oct. 17, 1885, Pringle 382.

Southeastern Arizona to trans-Pecos Texas (east to the Davis Mts.) and
south into Chihuahua and eastern Sonora. Closely related to A. scoparius
and probably only a well-marked geographical variety, differing in having
the hairs on the spikelet and pedicels scanty or nearly absent.

Andropogon scoparius Michx. Fl. Bor. Am. 1: 57 (1803).
CoaHUILA: Sierra del Carmen, Aug. 26, 1936, Marsh 607; Hillcoat Canyon, west

bottom and on rocky flats among scrub-oaks, Johnston & Muller 449, 661; tableland
north of Canon Cuervo Chico, ae slopes of low rounded limestone hills, Johnston
58,

Eastern United States west to Idaho and northern Arizona, apparently
entering Mexico only in northern Coahuila. Our material falls into the var.
neomexicana (Nash) Hitchcock, cf. Rhodora 37: 143 (1935).
Andropogon virginicus L. var. tenuispatheus (Nash) Fern. & Grisc. Rhodora 37:

142
uma: Cafion Agua Grande, west of Las Delicias, near water, erect, 3 m. tall,
scarce, Stewart 2818.

1943 ] JOHNSTON, PLANTS OF NORTHERN MEXICO, II 419

Wet ground from southeastern United States to California and south
into tropical America.

Andropogon ternarius Michx. Fl. Bor. Am. 1: 57 (1803).

CoanuiLa: Sierra de los Guajes, Cafion Madera, fairly abundant on hillsides,
Stewart 1504; Sierra Madera, Canon Charretera, rocky bed of open arroyo in oak belt,
common, becoming 4 ft. tall, Johnston 9074.

Delaware to Missouri and south to Florida and central Texas, entering
Mexico only in northern Coahuila

Andropogon Hallii Hack. Sitzungsb. Akad. Wiss. Wien 891: 127 (1884).

CurHuAHuA: Los Medanos, 1935, LeSueur 61.

Sandy places from North Dakota to Utah, and south to Arizona and
trans-Pecos Texas and northern Chihuahua
Andropogon Gerardi Vitman, Summa PI. 6: 16 (1792).

Andropogon furcatus Muhl. ex Willd. Sp. Pl. 4: 919 (1806).

CoanuILa: Sierra del Pino, meadows in the pine forests north of La Noria, com-
mon and conspicuous, Johnston & Muller 542, Stewart 1226

Widely distributed in eastern and central United States. Known in
Mexico only in Coahuila
Andropogon saccharoides Sw. Prodr. 26 (1788).

Andropogon barbinodis Lag. Gen. et Sp. Nov. 3 (1816).

Andropogon perforatus Trin. ex Fourn. Mex. Pl. 2: 59 (1886).

VERNACULAR NAME: Zacate aceite.

CoaHuILaA: Sierra del Carmen, July 29, 1936, Marsh 638; Hillcoat Canyon, west of
Buena Vista Ranch, July 13, 1938, Marsh 1313; Hillcoat Mesa, lying west of Encantada
Ranch, July 25, 1938, Marsh 1435; Hermanas, Marsh 1628 and 2254; Monclova, 1880,
Palmer 1347; Monclova, Marsh 1692; desert near Rancho Santa Teresa, Wynd &
Mueller 207; 2 mi. west of Saltillo, Harvey 1096; Saltillo, 1898, Palmer 4, 261, 810;
Sierra del Pino, La Noria, Stewart 1212, Johnston & Muller 667 ; Sierra Madera, Cafion
Charretera, bed of arroyo, Johnston 9075; 10 km. southwest of El Oro, on bajillo,
Stewart 3028; Sierra Jimulco, 5 km. northeast of Jimulco, Stanford et al. 132. Cut-
HUAHUA: Vicinity of Rancho El Pino, 10 km. southeast of Sierra Rica, rocky slopes,
Stewart 2410; near Juarez, pet 30, 1888, oe 1994; Agua Caliente, 1935, LeSueur
ae Chihuahua, 1935, LeSueur 2; arroyo 20 km. south of Camargo, Harvey 1382,

; 6 mi. west of Piloncillo, grassland, ARE 7879, ZACATECAS: Concepcion del
Ge. ‘1904, Palmer 202.

Alabama and Missouri west to southern California and south to Argen-
tina. A variable widely spread species which I am here accepting in the
broad sense used by Hackel in his monumental treatment of the genus.
Attempts to segregate out certain forms, such as A. barbinodis, by stress-
ing length of peduncle, hairiness of the nodes, shape of panicle, size of
spikelets, etc., have been singularly unsuccessful but persistent. The char-
acters used show little tendency to vary together and some of them are
probably associated with the vigor and rapidity of growth. Significantly,
these segregates and the restricted species have practically the same geo-
graphical distribution in Mexico and the United States and commonly may
be detected in the same locality. Some plants of A. saccharoides have a
conspicuous pore developed on the glumes and have been distinguished as

. perforatus. The distribution of the form is sporadic and may occur in

420 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

plants referred either to A. saccharoides or to A. barbinodis. Pitted glumes
are known in other species of Andropogon. I see no reason why the de-
velopment should be given specific recognition in the present case.
Sorghum halepense (L.) Pers. Syn. Pl. 1: 101 (1805).

CoanHvuILa: Sierra del Carmen, Sept. 8, 1936, Marsh 760; trail from the southern
end of Hillcoat Mesa to Buena Vista iia deatastnnn: July 27, 1938, Marsh 1508; Santa
Anna Canyon, Marsh 436; Monclova, Marsh 1663; Saltillo, weed in field, Hitchcock
5649 (US). CuHrmuanua: Arroyo 20 km. south at Camargo, Harvey 1398

A Mediterranean grass, now widely introduced into the warmer parts of
America. A serious weed in irrigated lands, particularly in the Laguna
District.

Sorghastrum nutans (L.) Nash in Small, Fl. S. E. U.S. 66 (1903).

Coaunurta: Palm Canyon near Muzquiz, Marsh 986.

Eastern United States and south into Mexico.

Heteropogon contortus (L.) Beauv. ex R. & S. Syst. Veg. 2: 836 (1817).

Coanuita: Sierra del Carmen, Aug. 22, 1936, Marsh 583; igneous hill near Santo
Domingo, Wynd & Mueller 470; Palm Canyon, 3 near Mascut Marsh 993; Puerto del
Norte, Cuatro Cienegas, Raves 3 206; Monclova, 1880, Palmer 1346; Puerto Santo
Lazaro, Sierra Gavia, Muller 3070; limestone hills near Santa Rosa, Shreve & Tinkham

Sierra Organos, 1937, LeSueur 161; Sierra Santa rte Pringle 480; Meoqui, 1936,
LeSueur 141; 13 km. southwest of Jimenez, Harvey 13

Florida i Arizona and southward; ey aan in the warmer
parts of the world. Although growing on limestone, this species is more
abundant on igneous rocks, particularly basalt. Usually growing on rocky
slopes and at times dominating large areas.

Heteropogon melanocarpus (Ell.) Benth. Jour. Linn. Soc. Bot. 14: 71 (1882).

Hitchcock, Contr. U. S. Nat. Herb. 17: 212 (1913), cites a collection
from the Mapula Mts., Chihuahua, Pringle 820. The species is widely dis-
tributed in the warmer parts of the world and extends north through west-
ern Mexico to Arizona.

Trachypogon greninngs (H.B.K.) Nees, Agrost. Bras. 342 (1829).
AHUA: ocky flat just _ ee Organos, locally common on ledges, Stewart
é& Johnston 2059; eae LeSue
Southern and eastern Texas; oe Mexico to Argentina.
Elyonurus barbiculmis Hack. in DC. Monogr. Phan. 6: 339 (1889).

CurHuAHUA: Rocky flats just east of Organos, locally common about ledges, leaf-
tufts bright green, Stewart & Johnston 2058; west of Chihuahua, 1935, LeSueur 14;
Cerro Coronel, Chihuahua, rocky hills, Aug. 5, 1885, Pringle 423

Trans-Pecos Texas to Arizona and south to atte
Manisuris altissimus (Poir.) Hitchc. Jour. Wash. Acad. 24: 292 (1934).

Coauuita: Sabinas River, near Muzquiz, 1936, Marsh 404

Wet places in the warmer parts of the world; introduced into America.
The species has been collected repeatedly on the Texan bank of the river
in the Big Bend of the Rio Grande.

1943 | JOHNSTON, PLANTS OF NORTHERN MEXICO, II 421

Hackelochloa speae (L.) Kuntze, Rev. Gen. 2: 776 (1891).

CHIHUAHUA: Open canyon in igneous hills 20 km. north of Chihuahua, sprawling
in moist gravel, rare, Sean & Johnston 2128; hills near Chihuahua, Pringle 1057
(US).

Arizona south through Chihuahua to Central America. A weedy grass
widely distributed in the warmer parts of the world. Said to be introduced
in our area, but, if so, introduced at a very early date, for it was collected
at unfrequented places in southeastern Arizona by Charles Wright as early
as 1851
Coix lacryma-jobi L. Sp. Pl. 972 (1853).

Coanuita: Saltillo, July 1880, Palmer 1337.

A species of the Old World tropics, widely cultivated in America for its
bead-like fruits and frequently spontaneous
Tripsacum dactyloides L. Syst. Nat. ed. 10, 1261 (1759).

Tripsacum dactyloides var. occidentale Cutler & Anderson, Ann. Mo. Bot. Gard. 28:

258 (1941).

Coanuita: Sierra del Carmen, Canon ahaipeliee Wynd & Mueller 536; Sierra En-
cantada, 7 km. west of Buena Vista, fairly common in wet canyon, Stewart 1450;
Santa Anna Canyon, Marsh 438; Palm Canyon, near Muzquiz, Marsh 9

Connecticut to Iowa and south to Florida and through ae onl north-
eastern Mexico to San Luis Potosi. Hitchcock, and recently Cutler &
Anderson, have placed the Tripsacum of northeastern Mexico in T. lanceo-
latum Rupr. I am, however, unable to separate Texan specimens from
those collected in Coahuila, Nuevo Leon, Tamaulipas, and eastern San Luis
Potosi. Characters in the size, shape, and surface of the segments of the
female inflorescence readily separate these specimens from the more south-
erly and westerly 7. anceolatum.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

422 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

PLANTAE PAPUANAE ARCHBOLDIANAE, XIII*
E. D. MERRILL AND L. M. PERRY

MELASTOMATACEAE

Our study of the Melastomataceae as represented by the material from
the Richard Archbold Expeditions to New Guinea and the specimens col-
lected by Brass and Kajewski in the Solomon Islands follows the basic
work of Mansfeld, Bot. Jahrb. 60: 105-143. 1925. We have found nothing
new in the Osbeckieae, Oxysporeae, or Sonerileae. In the Dissochaeteae
the only new records we have are of Medinilla Gaudichaud. The genus is
so diverse or variable in character that we have accepted Mansfeld’s inter-
pretation, although we are not wholly convinced that Hederella Stapf
rightfully belongs here. In the Papuan material the new species may be
readily placed in Mansfeld’s key, and for the convenience of future workers
we have inserted a running key patterned after that of Mansfeld. For the
consideration of the Solomon Islands material, it may be helpful to note
one particular point. Towards the end of Mansfeld’s key, p. 118, one finds
the caption, ‘‘Flores bracteis persistentibus stipati” (flowers surrounded by
persistent bracts). The only authentic material available for comparison
in this group was a very fragmentary specimen of M. Schlechteri Mansfeld,
an isotype, and in the unnamed collections a specimen which we determined
as representing M/. Pulleana Mansf. In both of these species the bracts are
always at the nodes of the inflorescences, the ultimate pair being at the base
of the pedicel. On the other hand, in much of the Solomon Islands material
the inflorescences are characterized not only by persistent bracts similarly
placed, but in addition have a pair of persistent bracteoles (usually more
showy and larger than the bracts) at the base of the calyx. This same
feature is found in a number of Philippine species and also in those of
Polynesia. Taken together, the species probably form a definite section
of Medinilla, with a geographical range south from the Philippines, includ-
ing the Solomon Islands, and eastward to western Polynesia (Fiji and
Samoa). Another distinctive group in the Philippines, represented by
Cephalomedinilla Merr., which we now believe ought to be considered as a
section of Medinilla, also occurs in the Solomon Islands.

In most of the species with setose nodes, it has been somewhat difficult
to characterize the pubescence satisfactorily. Mansfeld has used the term
plumose-pilose, or, if the trichomes were shorter, furfuraceous. In most
cases in our work this has been designated as subplumulose-pilose, for the
projections forming the “feathery” part are mostly very short, sometimes
hardly even barbellate, and are found either all along the main hair, or, in
some instances, only at the base of it.

*Botanical Results of the Richard Archbold Expeditions. See Jour. Arnold Arb.
24: 207-217. 1943.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 423

In the Astronieae there are several new species of Astronidium A. Gray,
and in the Memecyleae one new species of Memecylon Linnaeus.

Medinilla Gaudichaud

A. Leaves verticillate, opposite, or often appearing alternate, mostly equal
in size, or if unequal, similar in shape ; inflorescence bracteate or not,
the Hower. not bracteolate

B. Plants glabrous, or 1f oon the nodes not setose.
aves verticillate or opposite
D. Leaves verticillate.

Medinilla cauliflora Hemsl. Kew Bull. 1895: 135. 1895.

SoLomon Istanps: Bougainville: Kupei Gold Field, Kajewski 1671, 1718,
April 1930, alt. 950 m. and 1000 m., common on rain-forest trees; Koniguru, Bui
Kajewski 2142, August 1930, alt. mmon on taller trees a rain-for ae
Guadalc al: witho ae ot "Kajewski 2644, May 19

The ere may be summarized as follows: plant up to on m. long;
petals white; calyx light green; fruit dark red to black, up to 9 mm. long
and 8 mm. diameter. This appears to be the first record of any collection
of this species since the original description.
biameerna! ae ai Flora 509. 1831; Cogn. Monog. ays 574. 1891;

. Nov. Guin. aes 203. 1924, Bot. Jahrb. 60: 119.

Soran IsLtanps: Bou ville Kugumaru, Buin, sae 1986, July
1930, alt. 150 m., rain-forest me seadents fruit white when ripe, almost globose,
1 cm. diameter). "Malay sia and New Guinea.

D. Leaves always opposite.
E. Leaves sessile or subsessile.
gaara a ae Bak. . Trans. Linn. Soc. II. Bot. 9: 55. 1916; Manstf. Bot. Jahrb.
60 1925; vela
han New Guinea: Balim River, Brass 11754, causes 1938, alt. 2100 m.,

and pedicel red); 15 km. southwest of Bernhard Camp, Idenburg River, Brass 11899,
January 1939, alt. 1800 m., mossy forest, frequent in open situations (large shrub 2 m
high; calyx and pedicels oa petals white).

With no material for comparison, we find it difficult to distinguish
Medinilla Forbesii Bak. {., M. novo-guineensis Bak. {., and M. Bakeriana
Mansf. Both specimens cited above have cymose sh Rn OR in lateral
fascicles at the nodes. The other two species, according to the descrip-
tions, have fascicled flowers. The leaves of Brass 11754 are smaller (11 X
6 cm.) than those of Brass 11899 (19 & 10 cm.), but the structure of the
flowers is the same.

Medinilla Peekelii Mansf. Notizbl. Bot. Gart. Berl. 10: 282. 1928.

SoLoMon IsLtanps: Ysabel: Maruto, Brass 3395, December 1932, alt. 300 m.
(epiphytic shrub with pale fleshy leaves; flowers pale purple; fruit smooth, ie red).
The type was described from a tee tion made in the Bismarck Archipelag

E. Leaves petiola
. Flowers ate ee
Medinilla Hollrungiana Mansf. Bot. Jahrb. 60: 120. 1925.
NETHERLANDS NEW GuINEA: Boemi, 40 km. from Nabire, Kanehira & Hatusima

424 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

12728, March 1940, alt. 300 m., in forest; Bele River, 18 km. northeast of Lake
Habbema, Brass 11355, November 1938, alt. 2200 m., common in grassy second growths
on river banks (upright serra - . high; leaves feshiy and brittle, the average size
+ 14 xX 6 cm.; flowers pink, h red calyx and pedicel). British New E

creek in ie ower forest, rare (sparsely foliaged shrub 1 m. tall; leaves fleshy ; numerous
lateral aie of reddish pink flowers); Palmer River, 2 miles below junction of
Black River, Brass 7114 (det. es aig June 1936, alt. 100 m., occasional in a special
swamp forest community in the ridges (semiscandent epiphytic shrub; leaves fleshy ;
flowers dark pink; fruit ae Described from Northeastern New Guinea

Medinilla tenuipedicellata Bak. f. Trans. Linn. Soc. I]. Bot. 9: 53. t. 3, f. 44-47.

NETHERLANDS NEW GuINEA: 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 11878, 12066, January 1939, alt. 1800 m., mossy forest, gregarious in semi-shade
neve protruding + 20 cm. above the thick ground moss; flowers pale purple-

ink; —the second collection a shrub 60 cm. tall; calyx white, rimmed with red;
sahil pink).

Part of the material of these collections is an exact match for the plate,
but the rest shows considerable variation in the size of the leaves, some of
which are as large as 11 cm. long and 4 cm. broad.

Medinilla Versteegii Mansf. Bot. Jahrb. 60: 125. 1925.

NETHERLANDS NEw GuINneEA: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13714, March ane alt. 850 m., frequent on trees along river in rain-forest
(epiphytic shirals about 1 m. high; flowers white; fruit red) ; Bernhard Camp, Idenburg
River, Brass 13996, — ere alt. 50 m., frequent in flooded rain-forest of river plain
(epiphytic shrub 2 m. high; flowers pinkish white; fruit red). British New GuINeEA:
Palmer River, 2 rs below junction of Black Sivan Brass 7198, 7254 (det. Markgraf),
July 1936, alt. 100 common on river bank trees (large epiphytic shrub or small
tree with brittle eaves terminal panicles of waxy white flowers; fruit red, 6-7 mm.
diameter). In Brass 7254 the inflorescences are lateral and axillary. The species has
been reported babel for both Northeastern New Guinea and Netherlands New
Guinea.

Medinilla exigua sp. nov.
Arbuscula epiphytica 2 m. alta glaberrima multiramosa; ramis teretibus
cinereis; ramulis angulatis vel sulcatis sara: foliis oppositis repent
ak ellipticis, 1.2-3 cm. longis, 0.7—1.5 cm. latis, basi anguste c
is, apice obtusis, trinerviis, nervis supra eed subtus xr ee
petiolo 4—9 mm. longo; floribus in cymas axillares vel terminales pauci-
floras (1-3) dispositis; pedunculo communi 3-6 mm. longo, bracteis

minutis; gto + 4 mm. longis; calycis tubo anguste obconico vix 3
mm. longo, limbo truncato 1 mm. longo; petalis 4, pint circiter
1 cm. lon aa hes eins staminibus 8, antheris 3.5 mm. lon S postice
calcaratis, calcare 1 . longo, antice inappendiculatis ; ache im-

maturis a re

NETHERLANDS New GUINEA: 18 km. southwest of Bernhard Camp, Idenburg River,
Brass 12696 (type), February 1939, alt. 2050 m., rain-forest of a ravine (profusely
branched epiphytic shrub 2 m. high, with fragile white flowers and fleshy red fruit).

This species is readily ai Saal by the small leaves, the angular
branchlets, and the reduced cym

F. Flowers es,
Medinilla Erpetina ie eau Linn. Soc. 28: 87. t. 7, f. 94 d. 1871-73; Cogn.
Monog. Phan. 7: 589.
Erpetina radicans Naud. Pile Se Nat. III. 15: 299. t. 14. 1851.

1943 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 425
SoLoMon IsLtANnps: Ysabel: Mount Marescott, Brass 3260, December 1932, alt.

leaves dark green, thick and somewhat fleshy; flowers purple; fruit red, fleshy) ;
Tiratona, Brass 3531, December bi alt. 600 m., mountain Teanebs, common (small
root-climber; flowers reddish); Bougainville: Lake Luralu, Koniguru, Buin,
Kajewski 2066, August 1930, alt. 1500 m., on rain-forest ee common (vine; petals
pink; stamens light yellow; style light red).

We suspect this is the species represented by Medinilla nodosa Fosberg,
the main difference being in the more obtuse posticous appendage of the
anther of the latter. However, since neither type is available for examina-
tion at present, it seems best only to call attention to the very strong
resemblance between the two. Fosberg suggests that his species is closest
to M. acutifolia Hemsl. from the Solomon Islands. From the context, the
latter name is undoubtedly an error for M. cauliflora Hemsl.

C. Leaves, although opposite, often appearing alternate; inflo-
rescence mostly borne on verrucae (Hederella Stapf).

Medinilla longistylis Mansf. Bot. Jahrb. 60: 124. 1925.

British New GuINeEA: Palmer River, 2 miles below junction of Black River,
Brass 7044 (det. Me eae), June 1936, alt. 100 m., on a ridge crest (climbing to the
crown of a very tall tree, the slender branches pendent 6-8 m. below the limbs of the
tree; upper surface of leaves shining, the nerves deeply impressed above, prominent
below; flowers pink). The s tig is recorded previously only from the type, col-
lected in Northeastern New Gui
Medinilla lysipetala (F. v. Muell.) Mansf. Bot. Jahrb. 60: 124. 1925; vel. aff.

Catanthera lysipetala F. v. Muell. Jour. Bot. 24: 289. 1886.

Medinilla anomala Cogn. Monog. Phan. 7: 1185. 1891.

Hederella Forbesii Stapf in Hook. Icon. 25: t, 2415. 1895.

NETHERLANDS NEw Guinea: 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 12302, January 1939, alt. 1800 m., frequent in mossy forest (root-climbing
epiphyte; leaves convex; unopened flowers a dark reddish pink); 8 km. southwest of
Bernhard Camp, Idenberg River, Brass 12726, February 1939, alt. 1600 m., common in
open situations in mossy forest (leaves stiff, convex; flowers dark rose, not opening
widely

The material cited above and other material in our herbarium, from
Northeastern New Guinea, show a considerable amount of variation, a
we have been unable to distinguish more than one species in it. We are not
at all sure that it belongs to Medinilla lysipetala (F. v. Muell.) Pere
but it seems best to place it here provisionally. It is to be noted that the
type was collected near the base of the Owen Stanley Range, whereas the
material above cited shows a higher altitude. Mueller describes the leaves
as strongly 5-nerved from near the base, but these collections have obvi-
ously 3-nerved leaves. The inflorescences are axillary as well as lateral.

B. At least the nodes setose or hairy.
G. Leaves sessile or subsessile.
Medinilla arfakensis Bak. f. in Gibbs, Phyt. & Fl. Arfak Mts. 158. 1917.

NETHERLANDS NEw GuINEA: 15 km. southwest of Bernhard Camp, Idenburg River,
ah Sagi January 1939, alt. 1500 m., rain-forest (epiphytic shrub 1.5 m. high;

red; flowers white). Noxrunasiean New Guinea: Kani Mountains,
as 17037 (isotype of M. Brassii Markgr.), December 1907, alt. 1000 m.

426 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

tisH New Guinea: Bella Vista, Brass aaihes November 1933, alt. 1450 m., oak
forest fringe (shrub 1.5 m. high; flowers pale p

These collections appear to suit the ie aati of Medinilla arfakensis
Bak. f. reasonably well. All have sessile leaves, small 5-merous flowers,
and anthers with a posticous upwardly recurved spur. We have examined
exact duplicates of Schlechter 17037 (the type-number of M. Bras ssi
Markgr.) in the herbarium of the New York Botanical Garden and in our
own herbarium, and also the single specimen of Brass 5114 cited in the
original description. It would seem that Markgraf used the latter collection
for most of his description of M. Brassii, but, wishing to retain the type at
Berlin, designated Schlechter’s specimen as the type. The two collections
do not belong to the same species. Unfortunately the specific name must

o with the type designated, rather than with the collection to which it
more logically belongs.
Medinilla Lorentziana Mansf. Nov. Guin. Bot. 14: 206. 1924, Bot. Jahrb. 60: 128.
1925

B u New GuINEA: Palmer River, 2 miles below junction of Black River, Brass
6934, a (det. Markgr.), July 1936, alt. 100 m., epiphytic in ridge forests, fairly
common (large loosely branched shrub; bark niberoe: deeply furrowed; leaves con-
cave, recurved, pale underneath; panicles terminal, conspicuous ; peduncle and _ pedi-
cels red; flowers pale waxy pink).

Type from Netherlands New Guinea, with a variety occurring also in
Northeastern New Guinea.

Medinilla leucantha sp. nov.

Frutex epiphyticus; ramulis valde compresso-tetragonis, quadrialatis
(alis approximatis), nodis dense setosis; foliis oppositis similibus sessilibus,
basi pulvino persistente reflexo circumdatis, lamina anguste elliptica, 18- 40
cm. longa, 8-13.5 cm. lata, utrinque angustata, basi subcordata, apice acu-
minata, acumine 1 cm longo, novella subtus in costa, nervis ac acumine
minute furfuracea, cito glabrata, 9-13-plinervia, nervis supra manifestis
subtus perspicuis; inflorescentiis terminalibus vel axillaribus paniculatis
6-12 cm. = fere glabris, interdum parce furfuraceis, minute pustulatis ;

pedunculo communi 2-5 cm. longo, ramis 1-1.5 cm. longis, saepissime
quaternis, in peer 3-5 dispositis; bracteis minutis; pedicellis + 5
mm. longis; calycis tubo cupuliformi 3 mm. longo, imbo 2 mm. longo

truncato; ae 5 oblanceolatis, circiter 7 mm. longis; foc ah 10,
antheris 3 mm. longis, antice appendices duas subulatas gerentibus, postice
calcaratis, calcare uncinato gracili 1 mm. longo; fructibus subglobosis + 6
mm. diametro.

NETHERLANDS NEw GuINEA: Bernhard Camp, Idenburg River, Brass 13770 (TYPE),
April 1939, alt. 60 m., rain-forest (large sar pr shrub with greenish white flowers

and red fruit) : 4 km. southwest of Bernhard Camp, Idenburg River, Brass 13288,
March 1939, alt. 850 m., Cae epiphyte on ae trees along river, and high on trees
of forest (flowers translucent white); Dalman, 45 mi om ire, Kanehira &

Hatusima 12261, March 1940, alt. 500 m., in Agathis forest (epiphyte 1 m. tall; flower
white)

The leaves of Medinilla leucantha closely resemble those of M. Teys-
mannii Miq., according to the original description of the latter species, but
the first may be readily distinguished by the much shorter inflorescence, the

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 427

considerably smaller white flowers, and the anticous subulate appendages
of the anthers.
G. Leaves petiolate.

Medinilla albida sp. nov.

Frutex 1-2.5 m. altus; ramulis tetragonis vel subteretibus interdum angus-
tissime alatis, novellis subplumuloso- pilosis cito glabratis, maturis cinereis,
nodis dense setosis: foliis chartaceis oppositis in quoque pari aequalibus vel
subaequalibus; petiolo 7-12 mm. longo, novello piloso cito glabrato, basi
pulvino angusto persistente reflexo circumdato; lamina oblongo-1 anceolata
vel lanceolato-elliptica, 9-23 cm. longa, 4-9 cm. lata, basi late cuneata,
apice breviter acuminata, acumine 0.5—1 cm. longo, supra minute furfuracea
vel glabra, subtus minute furfuracea et nervis pilosa, 7-plinervia, nervis
supra manifestis subtus perspicuis; inflorescentiis terminalibus + 7.5 cm.
longis, paniculatis, axi pedicellisque ++ furfuraceis, nodis inferioribus seto-
sis: bracteis minutis; pedicellis + 4 mm. longis; calyce 3 mm. longo, ae
cyathiformi, limbo truncato; petalis 5, obovato-ellipticis, 7 mm. longis;
staminibus 10, antheris 2.5 mm. longis, antice mange sae ne a
calcaratis, calcare leviter uncinato, 0.7 mm. longo; stylo 4.5 mm. longo;
fructibus ‘subglobosis + 6mm _ diamet tro.

NETHERLANDS NEw GUINEA: 6 km. southwest of Bernhard Camp, Idenburg River,
Brass 12944, 12993 (type), February 1939, alt. 1200 m. and 1050 m., banks of rain-
forest stream (shrub 1-2.5 m. high; flowers white; fruiting panicle wholly red).

The species suggests Medinilla plumosa Mansf., but the latter has defi-
nitely sessile leaves. Then again there is some resemblance between this
and the description of M. schraderbergensis Mansf., but the pubescence of
the latter is of simple hairs.

Medinilla Mansfeldiana sp. nov.

Verisimiliter arbor parva; ramulis dense hirsuto-setosis; nodis longe
barbatis; foliis ellipticis, 10-20 cm. longis,.4.5—10 cm. latis, utrinque paullo
angustatis, basi obtusiusculis, apice breviter acuminatis, S-nerviis vel sub-
quintuplinerviis, supra glabris, nae praecipue costa nervisque parce
gee venis obscuris; petiolo 1-1. Hie ee parce setoso; susie eatin

pedicellis brevibus, 2-5 mm. longis, dense setulosis; calyce cyathiformi,
ongo, longe an petalis non visis; staminibus 10, filamentis

3.5 mm. longis, antheris 4 mm. lo ongis postice vix ‘calcaratis breviter obtusis,

antice breviter biauriculatis; stylo 9 mm. longo; fructibus subglobosis.

NorTHEASTERN NEw GuINEA: Goridjoa, Schlechter 19744 (type), June 1909, alt.
1200 m., mountain woods.

The species apparently is closely related to Medinilla sogeriensis Bak. f.
and M. Schlechteri Mansf. It may be distinguished from the first by the
larger inflorescence on the older branches, and from the second by the
petiolate 5-nerved leaves. M. Schlechteri Mansf. has 7—9-plinerved leaves.
Medinilla Markgrafii sp.

Medinilla Brassi Ma paloue 2: 142. 1936, quoad Brass 5114, excl. spec. typ.

Frutex interdum scandens 1-2 m. altus; ramulis cinereis vel fuscescenti-

428 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

bus obtuse angulatis vel novellis valde compressis glabris nodosis, nodis

setosis; gs subcoriaceis gol oppositis ellipticis vel oblongo- ellipticis,
12-22 cm. longis, 4-10 cm. latis, basi cuneatis vel obtuse cuneatis, apice
breviter acuminatis vel meee) a 7-plinerviis, bia subtransversis supra
interdum manifestis subtus obscuris; petiolo + 1.5 cm. longo, basi pulvino
angusto persistente reflexo circumdato: stipulis interpetiolaribus setas ad
nodos tegentibus; inflorescentiis paniculatis + 10 cm. longis, axillaribus
vel terminalibus, bracteatis, ad nodos + setulosis; bracteis inferioribus
+ 1.5 cm. longis sursum minoribus, subtus apicem versus interdum + pube-

tetrameris; calyce 2.5-3 mm. longo, novello consperse a pilosulo,
maturo glabro vel subfurfuraceo, truncato; petalis + 5 m m. longis, albis;
staminibus 8, antheris lineari-oblongis, 2.5-3 mm. nee apice paullo

angustatis, postice calcaratis, calcare brevi obtuso non recurvo; stylo
glabro; fructibus globosis + 5 mm. diametro.

British New oe Mount Tafa, Brass 4018 (Type in herb. New — Bot.
Gard., isotype in herb. A. A.), May 1933, alt. 2310 m., common on edge of rest house
clearing (weak reer ees 1—2 m. tall; upper ie of leaves iridescent a the
lower side pale; flowers white; pedicels reddish, unripe fruit dark ame
ange Brass 5114, September 1933, alt. 2400 m., in a landslip shrubbery, rare ‘(shrub
1 ; peduncle red; petals white; fruit green).

T hese two collections belong to a single species, which, according to
Mansfeld’s key for Medinilla, falls near M. warica Mansf., and, Sees ng
to that of Baker f., near M. rubiginosa Cogn. However, M. kgrafit

differs from both of these in pubescence, but we cannot elles a ae
affinity at present. Brass 5114 appears to have been used largely in the
description of M. Brassii Markgr.; unfortunately it is not conspecific with
the type designated for that species.

Medinilla rubiginosa Cogn. Monog. Phan. 7: 598. 1891; Mansf. Bot. Jahrb. 60:

NETHERLANDS NEW GUINEA: 9 ae rrgungped of Lake Habbema, Brass 10786,
10870, October 1938, alt. 2750 m. and 2650 m., common in moist open spots of forest
undergrowth (tree 2.5-3 m. high; eres red; flowers white); Bele River, 18 km.
northeast of Lake Habbema, Brass 11535, November 1 1938, alt. 2200 m., abundant in
brushy second growths (shrub or tree 2-4 m. high; panicles red; flowers white; fruit

en); Angi, Arfak Mountains, Kanehira & aia 13609, 13767, 13926, April

1940, alt. 1900

Previously known only from the type-collection.

A. Leaves heey) each pair dimorphic (exceedingly unequal in size and
unlike in shape), except in M. tulagiensis; inflorescence with both
bracts and bracteoles.

The remaining species of Medinilla in this treatment are all from the

representatives of this section have thus far been reported from New
Guinea.
KEY TO THE SOLOMON ISLANDS SPECIES WITH BRACTEOLATE FLOWERS

Inflorescence open, usually racemose; calyx-limb tru
Leaves of each pair somewhat unequal in size, none. similar...... M. tulagiensis.

1943] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 429

Leaves of each pair exceedingly unequal in size, the smaller usually ovate or ovate-
orbicular and sessile or subsessile.
Larger leaf of each pair sessile or subses
Branchlets and inflorescence fone ue leaf of each pair = 25 cm. long..
Bes eevee acca eR eae 26S eb n6 a x6 4 TORE Eee, ee PO Pe PSEC acct aO On M. Aajeweki,
Branchlets and inflorescence pilose; larger leaf of each pair 6.5-13 cm. Jong....
A BAMA clo afetag SRE a ev NCUANR ova es boa Fa) oes a8 Sr cada epee am eae ae eee RSE M. luraluensis.
Larger leaf of each pair obviously petiolat
Plant apparently glabrous; inflorescence oe Mm. longs his. ease ¢ M. sae
At ce the nodes and the calyces pubescent ; pea ee smaller, up to 10 ¢

lon
fae Beackecies large, 1.5—2.3 cm. long.
Bracteoles pink; leaves with a long narrow acumen, almost glabrous be-
Meath) wen Mature: <5 ¢ 4 4s% sesso eee eee ee ee oe M. calliantha.
Bracteoles dirty cream-color or greenish white; leaves short-acuminate,
pubescence + persisting on the lower surface.
se elliptic or lance-elliptic, rounded at base; floral bracteoles 1.5 cm.

TD ese re ee cere ee ease 3 0°b.0o0'G dad eae er eae eee ae M. oe
ieee ovate-elliptic, cordate at base; floral bracteoles 2—2.3 cm. long.
Pee Pee CT neato aa ce «ice eaves aiodca a: td Ga gO ORD RE ara ee te) chee oe M. vagans.

Floral bracteoles small, 5-6 mm. long.
Larger leaf of each pair lanceolate; inflorescence 1.5 cm. long; plant very
Sparsely -pubescent.5 v::sAcsuinioeaeseorer eaten eiatwua se ai araiales « M. lancifolia.
Larger leaf of each pair lance-elliptic; inflorescence + 10 cm. long; aa
lets, lower surface of leaves, and inflorescence obviously pubesce
M

EE Oe ee re eo en ee IO ae ere ee : Hee.
ata capitate (very compact) ; calyx-limb 4-lobed.
Larger leaf of each pair petiolate; flowers large, the calyx 1.5 cm. Pris Be See eke
syoustdcca oils. nt DIB d ages ies 95 han bia os coe RRO aN eae ed eae Rete tat eh Shere 5 ates . fais
Larger leaf of each pair sessile or subsessile; flowers smaller, the calyx 7 mm. long
ne er ane REPEC 7) a 9 ce) CCT ET OC a RCI M. ee
Medinilla tulagiensis sp. nov
Frutex scandens; eal novellis brunnescentibus barbellato-pilosis,
vetustioribus: cinereis value nodis barbellatis demum glabratis; foliis

cm. latis, petiolo 0.5—-1.5 cm. longo, caeterum ut majoribus; aap an
axillaribus racemosis solitariis vel interdum duobus in axillis, +
longis; axi ed eee ea novellis subplumuloso- -pilosulis deinde Sat ralie:
bracteis oblanceolatis, + 1 cm. longis; pedicellis + 1 cm. longis; bracteolis
1-1.3 cm. longis, lpi, basi elongato-cuneatis, parce pubescentibus;
floribus 4-meris; calyce 4 mm. longo, minute su plumuloso- piloso, limbo
a 4- dentato; RY in alabastro 5 mm. longis; staminibus 8, antheris
m. longis, postice calcaratis antice inappendiculatis; ovario 4- loculato;
fructus subglobosis + 7 mm. diametro.

SoLomon Istanps: Tulagi: Brass 3521 (type), January 1933, alt. 50 m., rain-
forests, common aad climber with ar rather fleshy leaves; bracteoles white; aren
pink ; fruit whit
Medinilla oe sp. nov.

Frutex scandens; ramulis teretibus glabris nodis barbellatis demum
glabris; foliis sessilibus vel subsessilibus glabris chartaceis valde disparibus:
majoribus ellipticis, + 25 cm. longis, 11-12 cm. latis, utrinque paullo

430 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

angustatis, basi rotundatis vel subcordatis, apice acutis vel breviter acumi-
natis, subseptuplinerviis, venis subtransversis supra inconspicuis subtus

rotundatis, mucronatis, quintuplinerviis; inflorescentiis axillaribus brevi-
bus: bracteis obovatis obtusis, 6 mm. longis, 5 mm. latis; pedicellis + 1
cm. longis; bracteolis ut bracteis; fructibus subglobosis, 7 mm. diametro;
seminibus semiconicis, complanatis, apice rotundatis, vix 1 mm. longis,
0.6-0.8 mm. latis

arene Bou inville: Kupei Gold Field, Kajewski 1633 (TYPE),
April a6. alt. 900 m. aes climber ascending 4 m.; bracts white; pedicels and
petals purple).

This species is readily recognizable by its strongly unequal and sessile
leaves, its short open inflorescence, and its rather large somewhat hood-
shaped seeds
Medinilla luraluensis sp. nov.

Frutex scandens; ramulis teretibus nodosis glabratis, novellis pilosis,
pilis crassis subplumulosis, nodis novellis barbellatis demum glabris; foliis
subcoriaceis valde disparibus: majoribus subsessilibus vel breviter petio-
latis (petiolo 3-5 mm. longo), 6.5-13 cm. longis, 2.5—6 cm. latis, basi
cordatis paullo inaequalibus, apice acutis vel breviter acuminatis, supra
glabris, subtus basim versus costa pubescentibus, quintuplinerviis ; minori-
bus sessilibus subrotundatis vel ovatis, + 2 cm. longis et 2.5 cm. latis,
cordatis, hase supra glabris subtus nervis + + pilosis; inflorescentiis
axillaribus + 10 cm. longis; axi pedicellisque parce pilosulis; bracteis
ovato-orbicularibus basi breviter cuneatis, 8 mm. longis, 6 mm. latis, subtus
consperse pilosulis; pedicellis 1-1.3 cm. longis: bracteolis ut bracteis;
calyce in alabastro pilosulo, 2.5 mm. longo; petalis 3.5 mm. longis; stamini-
bus 8, antheris postice calcaratis; ovario 4-loculato; fructibus subglobosis
5 mm. diametro.

Sotomon Istanps: Bougainville: Lake Luralu, Peet ee 2061 sik
Pavan 1930, alt. 1500 m., rain-forest, common (climbing shrub or vine; leaves with
purple veins on lower siace: bracts showy pink; fruit white with patie ak n
calyx-

The bracts and bracteoles of this species are smaller than in most of the
group of species with dimorphic leaves.

Medinilla anisophylla sp. nov.

Frutex scandens glaber; ramulis ae ean ey nodosis; foliis
chartaceis, valde disparibus: majoribus 12-1 ongis —8.5 cm. latis,
basi emarginatis vel rotundatis, apice ia. acumine fe 1 cm.
longo, quintuplinerviis vel subseptuplinerviis, venis vix manifestis; petiolo

cm. longo; minoribus ovatis, + 2 cm. longis latisque, basi subcordatis,
apice acutis, subsessilibus vel brevissime petiolatis; inflorescentiis axillari-
bus racemosis, 10-17 cm. longis, bracteis + 1 cm. longis, lanceolatis; pedi-
cellis + 1 cm. longis; bracteolis eile agrees circiter 1.5 cm.
diametro; calycis limbo 4-denticulato; antheris (tantum uno viso
longo, postice crasse calcaratis, antice heaton pas fructibus 5- 7 mm.
iat

Sotomon Istanps: San Cristobal: Puepue River, Brass 2793 (TYPE),
Ga 1932, alt. 50 m., riverine rain-forest, common (profusely flowering loosely

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 431

branched shrub, scandent; leaves fleshy, pale green; each flower between two large
fleshy greenish white persistent bracts; corolla pink; filaments purple and red).
Medinilla calliantha sp. nov.

utex scandens; ramulis novellis dense et grosse ee -pilosis, ste
oe sub ae subplumulosis, cito glabratis teretibus nodos
setulosis; foliis valde disparibus: majoribus chartaceis ellipticis, 13- 20. cm.
longis, 6.5-12 cm. latis, pee Po hindatic vel obtusis, apice obtusis deinde
abrupte acuminatis, acumine 1.5—2 cm. longo sublineari, novellis dense
patenti-pilosis cito ‘glabratis, maturis utrinque glabris vel subtus costa
nervisque parce pilosis, 5—7- plinerviis; petiolo + 1 cm. longo, glabrato;
minoribus ovatis, 4.5 cm. longis, 3 cm. latis, sessilibus caeterum ut majori-

spathulatis, roseis; axi pedicellisque dense et grosse patenti-pilosis; pedi-
cellis circiter 1 cm. longis, floribus 4-meris decussatim oppositis; bracteolis
ovato-orbicularibus, circiter 2 cm. longis latisque, basi rotundatis, apice
obtusiusculis, utrinque consperse subplumuloso-pubescentibus vel interdum
conspersissime pilosis; calyce cyathiformi, + 5 mm. longo, dense piloso;
petalis 1 cm. longis, 6 mm. latis, ellipticis; antheris 6 mm. longis, postice
1.5 mm. calcaratis, antice minute biappendiculatis; stylo + 8 mm. |] ,
apice ee. stigmate minuto

SoLoMON Istanps: Ysabel: Tiratona Brass 3220 (TYPE), 3328, 3540, Novem-
ber and December 1932, alt. 600 m., mountain forests, common (large stiffly branched
climber; bark corky, ie hick: young ee covered with brown hairs; fruit in
3220 yellow, in pe! white).

Medinilla calliantha var. bella var. nov.
forma typica differt ramulis novellis parce subplumuloso-pilosulis;
foliis glabris; ‘inflorescentiis + pilosulis.

Sotomon IsLtanps: Bo nee aera Without field label, Kajewski 1746A;

uadalcanal: ulolo, Tutuve Mountain, Kajewski 2504 (type of var.), April
1933, alt. 1200 m., rain-forest, common (climbing shrub, very showy; veins of leaves
pink ; flowers pin),

This variety closely resembles the species but is much more nearly gla-
brous, with slightly smaller flowers and fruit; the leaves tend to be nar-
rower, the innermost pair of nerves arising from the midrib 2.5 cm. above
the base, whereas in M. calliantha they are not more than 2 cm. above the
base.

Medinilla pubiflora sp. nov.

Frutex scandens; ramulis teretibus nodosis cinereis, novellis brunneis
dense piloso-hirsutis, pilis sub lente subplumulosis, deinde glabratis; foliis
valde oo majoribus ee ellipticis vel lanceolato-ellipticis,
14-17(—27) cm. longis, 7.5(—12 cm. latis, basi rotundatis interdum

sissime subtus (costa nervisque + dense) brunneo-pilosis, 5—7-plinerviis,
nervis supra eens subtus prominulis, venis subtus manifeste clathra-
tis; petiolo 1-1. ‘lo ongo, dense ee minoribus cet be sub-
sessilibus late a carbicuian yee 2.5(—4) cm. longis, 2.5(—3) cm. latis,
5-nerviis, basi cordatis, apice obtusis apiculatis; inflorescentiis Seance
racemosis, + 4 cm. longis: axi dense, bracteis et bracteolis utrinque
pilosis; bracteis vix 1 cm. longis, cuneatis interdum foliiformibus; pedi-

432 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

cellis 4 mm. longis; bracteolis ovatis, 1.5 cm. longis latisque; floribus tan-
tum in alabastro visis; calyce 5 mm. longo, dense piloso, tubo cyathiformi,
limbo truncato; petalis 4; staminibus 8, antheris postice longiuscule crasse
calcaratis, antice non visis

SoLoMON IsLANpDS: Boug shat ville: Koniguru, Buin, Kajewski 2153 (TYPE),
August 1930, alt. 900 m., rain-forest, common (vine or semi-scandent shrub; flowers
dirty cream-color, covered with fine hairs); Kugumaru, Buin, Kajewski 1958, July
1930, alt. 150 m., rain-forest, common (scandent shrub; bracts dirty cream-color;
petals pe purple) Kieta, pe pee 1564, March 1930, alt. 100 m., in gullies in rain-
forest, common (s up to 2 m. high; bracts white, covered with brown hair, giving
them a ales appearance; petals ae blue; anthers dark blue).

In general habit, Medinilla pubiflora calls to mind M. vagans, but the
pubescence is more evenly distributed and denser, the base of the leaf is
rounded rather than cordate, the inflorescence is shorter, and the bracteoles
are smaller than in the latter species.

Medinilla vagans sp. nov.

Frutex vagans; ramulis teretibus, novellis obtuse angulatis sulcatis, parce
patenti- pore cito glabratis, nodis barbatis; foliis valde disparibus: majori-
bus ovato-ellipticis usque 11 cm. longis, 5.5-8 cm. latis, basi cordatis, apice
abrupte breviter acuminatis, acumine 5-10 mm. longo, supra glabris vel

distincte subclathratis; petiolo + cm. heen subtus glabrato supra
patenti-piloso ; minoribus 3 cm. longis, 2.5 cm, latis, subsessilibus vel brevi-
ter petiolatis, caeterum ut majoribus; inflorescentiis cymosis axillaribus,
circiter 5 cm. longis, paucifloris (verisimiliter 2), pedunculo ad 1 cm. supra
basim bibracteato, bracteis caeterum 4 basi pedicellorum, subovato-orbicu-
laribus, + 7 mm. longis latisque basi interdum anguste cuneatis

minibus 9 (in duobus floribus dissectis), antheris (in alabastro) 4 mm.
longis, postice obtuse breviter calcaratis, calcare verruculoso, crasso, antice
biauriculatis; stylo 5 mm. longo.

SoLtoMon IsLtAnps: Ysabel: Tataba, Brass 3438 (Type), January 1933, alt.
50 m., amongst regrowth trees on a rain-forest clearing (large rambling shrub; leaves
dull; bracteoles white; petals and filaments pale pink; anthers blue; fruit greenish
white; a very showy gaat, with brown ae.

This species is fairly easy to recognize by the ovate-elliptic cordate leaves
in unequal pairs and the pubescent cymose (subracemose) inflorescence
with rather large white bracts.

Medinilla lancifolia sp. nov.

Frutex scandens; ramulis subteretibus cinereis nodosis glabris, nodis
minute stellato-pubescentibus deinde glabratis; foliis valde disparibus:

angustatis, basi cuneatis, apice obtuse acuminatis, novellis consperse
es a cito glabratis, vel consperse minute papillatis, tripli-
nerviis vel interdum subquintuplinerviis; petiolo usque 5 mm. longo glabro;
minoribus sessilibus ovatis usque 1.7 cm. longis, 1 cm. latis; cent

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 433

oper cadae + 1.5 cm. longis, racemosis, novellis consperse eau pube-

entibus; bracteis albido-viridescentibus, usque 6 mm. longis et 3 mm. latis
cp camer semen rte consperse minute pubescentibus: pedicellis 2 mm. oie
bracteolis ut bracteis; calyce minute pubescente, tubo cyathiformi, + 3
mm. longo, limbo vix 2 mm. longo; petalis 4, obovatis, 6 mm. longis, apice
abrupte acutis; staminibus 8, neal 4 mm. longis; stylo + 7 mm. longo,
apice angustato: stigmate mir

SoLoMON IsL_anps: Bougali wilie : Kupei Gold Field, Kajewski 1699 (TYPE),
April 1930, alt. 950 m., growing aoe crevices in tall rain-forest trees (petals mauve;
anthers blue; bracts cream-green)

In the short bracteate inflorescence the species somewhat suggests
Medinilla involucrata Merr., but it is amply distinct from that species in
the size and venation of the leaves as well as in the lack of pubescence.
Medinilla rubescens sp. nov.

Frutex scandens; ramulis novellis dense subplumoso-pilosis deinde gla-
bratis nodosis; foliis valde disparibus: majoribus chartaceis lanceolato-
ellipticis, 9-14 cm. longis, 3.5—6 cm. latis, basi rotundatis vel emarginatis,
apice acutis vel breviter acuminatis, supra costa nervisque praecipue ad
basim pilosis, subtus (costa nervisque dense) ferrugineo-pilosis, 5—7-pli-

n
dense pilosulo; minoribus sessilibus vel breviter petiolatis (petiolis 2-3 mm.
longis), ovatis, 1.5-3 cm. longis, 1—2 cm. latis, peti ats ut
majoribus; inflorescentiis axillaribus, paniculatis vel cymosis, 5—10 cm.
longis, ramis paucis racemiformibus; axi pedicellisque niloedte vel pube-
nodis bracteatis: bracteis oblanceolatis basi cuneatis, 5 mm.
longis, 2 mm. latis, glabratis; bracteolis late ovatis obtusis, circiter 5 mm.
longis et 4 mm. latis, stellato-pubescentibus; calyce 2 mm. longo, cyathi-
tormi, dense pubescente; petalis 4, ellipticis, obtusis, 4-5 mm. longis;
staminibus 8, antheris 3 mm. longis, antice biauriculatis, postice breviter
calcaratis, calcare crasso verruculoso; fructibus subglobosis, + m

longis

SOLOMON ISLANDs: - uadalcanal: Uulolo, Tutuve Mountain, Kajewski 2514
(type), April 1931, alt. 1200 m., rain-forest, common (scandent shrub; flowers and
bracts pink; small green ales 6 mm. diameter).

Medinilla cephalantha sp. nov.
Frutex scandens; ramulis teretibus, novellis dense hirsutis cito glabratis,
brunneis deinde cinereis nodosis _ nodis setosis; foliis valde disparibus:

acuminatis, novellis dense hirsutis, maturis glabris, 7- plinerviis, nervis saepe
ad 1-2 cm. inter se distantibus, ‘utrinque perspicuis, venis oblique trans-
versis manifestis; petiolo 1.5—2. 5c m. longo, hirsuto; minoribus sessilibus
vel subsessilibus, Cato abel ae 3 cm. longis, 2.5 cm. latis, basi
cordatis, apice rotundatis subinde brevissime acuminatis; inflorescentiis
sessilibus e ramulis vetustis defoliatis orientibus vel interdum axillaribus,
capitatis, interdum oppositis, 2.5—-6.5 cm. longis, dense multifloris; floribus
subsessilibus vel breviter pedicellatis, pedicello + 4 mm. longo, hirsuto;
bracteis ut bracteolis, 1.5-1.7 cm. longis, late ese vel ellipticis, apice
obtusis, basi sensim anguste cuneatis, subtus hirsutis; calyce 1.5 cm. longo,
dense hirsuto, limbo 4-lobato vel 4-fido, intus piloso, lobis rotundatis retusis,

434. JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXIV

petalis in swage tantum visis glabris; staminibus 8 aequalibus, antheris
sub anthesi 6 mm. longis apice poro aperientibus, postice obtuse inconspicue
calcaratis ie 0.5 mm. longo, crasso), antice biauriculato, auriculis
minute verruculosis; stylo + 1.5 cm. longo.

Sotomon Istanps: Ysabel: Kakatio, Brass 3255 (type), mi i 1932, alt.
900 m., common in rain-forests (large scandent shrub; bracteoles pink; flowers white) ;
Guadalcanal: Sorvorhio Basin, Kajewski 2705, January 1932, af 300 m., in wet
gullies of rain-forest, common (small tree or large shrub; inflorescence pink); San
Cristobal: Hinuahaoro, Brass 3022, September rors alt. 900 m., mountain rain-
forests, common (scandent shrub with long drooping branches; flower pink

On account of the variability within the genus Medinilla Gaudichaud, we
are now inclined to believe that the Philippine Cephalomedinilla Merr. is
probably best considered as a section of Medinilla. This species and M
sessilis from the Solomon Islands clearly belong to the same section, but

may be readily distinguished from the Philippine species by the
stronger dimorphism of the pairs of opposite leaves, the smaller of these
being ovate-orbicular rather than similar in shape to ‘the larger ones.

Medinilla sessilis sp. nov.

Frutex scandens; ramulis teretibus + hirsutis cito glabratis, nodis dense
barbatis; ioliis valde epi berries vel subsessilibus, basi pulvino
persistente reflexo circumdat majoribus ellipticis utrinque angustatis,

basi subcordatis, apice pean ese acumine + 1 cm. longo, supra glabris,
subtus costa oe elias dense) consperse pilosis, 5—7-plinerviis,
nervis inter + 1 cm. distantibus, venis clathratis subtus prominulis;

minoribus suborbicularibus vel ovato- orbicularibus fere citer nme pose

m
, limbo 3—4 mm. longo, 4-lobato, intus hirtello; petalis 4; stamini-
bus 8, antheris postice calcaratis, antice verisimiliter ee (sta-
minibus immaturis, appendicibus non visis) ; ovario pagel piloso
MON IstaAnDs: Bougainville: Koniguru, Buin, Kajewski 2023 (type),
August 1930, alt. 850 m., rain-forest, common (scandent; aa purple).

This species is readily distinguished from Medinilla cephalantha Merr. &
Perry by the sessile or subsessile larger leaf of each pair; also, the retlexed
narrow base ets the leaf-attachment is much more Hinton here
than in the other specie

Astronia Blume
Astronia papetaria Bl. Rumphia 1: 20. t. 6. 1835, var. novo-guineensis var. nov.

A forma typica differt foliis minoribus, 8-15 cm. longis, 1.5—4 cm. latis,
in quoque pari vix aequalibus; petiolo 1.5-2 cm. longo ad ramuli inser-
tionem non calloso-verrucoso; calycis lobis Sietoe ban, late triangularibus,
acutiusculis.

NETHERLANDS New Guinea: 4 km. southwest of Bernhard Camp, Idenburg River,
Brass 13292, 13454 (Type of var.), Mar. 1939, alt. 900 m., frequent in Agathis forest
undergrowth, and occasional in rain-forest of ridges (tree 3 —4 m. high; leaves brown
beneath; flowers pink) ; cm. southwest of Bernhard Camp, Idenbure River, Brass
13671, March 1939, alt. 750 m., common in rain-forest of slopes (tree 5-6 m. high;
leaves brown underneath; flowers pink).

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 435

The collections cited above show a strong similarity to Blume’s plate of
Astronia papetaria from the Moluccas, but differ as indicated above.
Astronia atro-viridis Mansf. Bot. Jahrb. 60: 131. 1925.

NETHERLANDS NEw GUINEA: 9 km. northeast of Lake Habbema, Brass & Aisi
10466, October 1938, alt. 2750 m., frequent in primary forest; Bele River, 18 km. north-
east of Lake Ha bbema, Brass & Versizevh 11151, November 1938, alt. 2300 m. winded
substage tree of primary forest; 15 km. southwest of Bernhard Camp, Idenburg River,

white; fruit yellow-green). British New Guinea: Mount Tafa, Brass 4953

In these collections the leaves vary greatly in size from those of the
original material, in some cases being as small as 4 X 2 cm.; again, occa-
sionally the leaves are almost densely lepidote, and only slightly, if at all,
furfuraceous, but the characters are too inconstant to be of value. At the
apex for a very short distance the margins are rolled to meet, causing the
apex to appear almost like an appendage attached to the rest of the leaf
by a constriction.

Astronidium A. Gray

Astronidium nigrescens (Mansf.) Markgr. Notizbl. Bot. Gart. Berl. 12: 48, 49. 1934.
Everettia nigrescens Mansi. Bot. Jahrb
NETHERLANDS NEw GUINEA: 15 km. southwest of Bernhard Camp, Idenburg River,
Brass 12091, January 1939, alt. 1800 m., frequent in rather open rain-forest ravines
(tree 4-5 m. high; inner surface of balals white, the outer surface pink; stamens and
pistil white). Described from Northeastern New Guinea.
Astronidium novo-guineense sp. nov.
Arbor usque 10 m. alta; ramulis teretibus vel infra nodos subangulatis
atro- cinereis glabris; foliis subcoriaceis ellipticis vel leviter obovato- ellipti-
]

furaceo-lepidotis, triplinerviis, nervis in laminae parte centrali 1.5—2 cm. a
costa dispositis, supra insculptis, rel be perspicuis, venis clathratis =e |

remotis, supra inconspicuis subtus prominulis, vena marginali 1—2
mm. intra marginem disposita, reticulo conferto; inflorescentiis ae a
cymoso-paniculatis, + 10 cm. longis latisque, glabris, pedicellis 6-7 mm.
longis; calyce urceolato, a 2.5-3 mm. longo, limbo 1 mm. longo truncato;
petalis 5, oblongis, 6 mm. longis; staminibus 10, in alabastro filamentis
4 mm. longis eee seule complanatis, antheris + 5 mm. longis (in ala-
bastro apice reflexis) , postice calcaratis, calcare crasso, 1 mm, longo; stylo
5 mm. longo; ovario 4- vel 5-loculato.

NETHERLANDS NEw GUINEA: 18 km. southwest of Bernhard Camp, Idenburg River,
Brass 12695 (tTypE), February 1939, alt. 2150 m., mossy forest, common in gullies (up
to 8-10 m. high; flower-buds white); 4 km. southwest of Bernhard Camp, Idenburg
River, Brass 13312, March 1939, alt. 900 m., abundant in Agathis forest and more open
parts of mossy forest (tree 6—7 m. high)

This species is very close to Astronidium palauense (Kaneh.) Markgr.
(including A. carolinense (Kaneh.) Markgr., which, from a comparison of
the isotypes, appears to be the same species). It differs in having an
urceolate, not cupular, calyx, and a much closer reticulum in the leaves than
is characteristic of the Micronesian material.

436 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Astronidium insulare sp. nov.

Arbor usque 17 m. alta; ramulis teretibus, novellis inter nodos com-
pressis leviter sulcatis et minute lepidotis; foliis oppositis ellipticis, 9-16
cm. longis, 4.5-7.5 cm. latis, basi obtusis subinde cuneatis + 5 mm. de-
currentibus, apice acutis vel breviter acuminatis, apiculatis, supra glabris,
subtus minute lepidotis, triplinerviis, nervis in laminae parte centrali 2—2.5
cm. a costa dispositis, supra distinctis, subtus prominulis, venis clathratis

+ 1 cm. remotis, supra manifestis subtus prominulis, venulis inconspicuis;
petiolo 2-2.5 cm. longo; inflorescentiis immaturis 5 cm. longis, terminali-
bus, dense lepidotis; pedicellis 2 cm. longis; calyce 4 mm. longo obpyri-
formi, apice 5-lobato, lobis 0.4 mm. longis, + 1.5 mm. latis, mucronulatis;
ovario 5-loculato.

SoLtomon Istanps: Bougainville: Koniguru, Buin, Kajewski 1999 (tTyPE),
August 1930, alt. 800 m., rain-forest, common (small tree up to 17 m. high; buds
green).

The species is perhaps most like the description of Astronidium novae-
hannoverae (Engl.) comb. nov. (Astronia novae-hannoverae Engl. Bot.
Jahrb. 7: 468. 1886), from the Bismarck Archipelago. However, the calyx
of the flower-buds, although still not approaching anthesis, appears to be
about twice as large as that described for the latter species. We have not
found any reference to Pages species either in Mansfeld’s “Die Melasto-

mataceen von Papuasien,” Bot. Jahrb. l.c., or in Markgraf’s ee hasan
Astronidium A. Gray,” Notizbl. Bot. Gart. Berl. 12: 47—50.

Astronidium montanum sp. nov.

Arbor usque 7—8 m. alta; ramulis tetragonis vix Seri minute lepidotis;
foliis tenuiter chartaceis, 10-15 cm. longis, 3.5—6 cm. latis, minute pellucido-
punctatis, subtus minute papillatis parce minute oe basi cuneatis,
apice obtuse acuminatis, trinerviis, nervis in laminae parte centrali 1-
cm. a costa dispositis, venis marginalibus (2 vel 1) circiter 2—5 mm. intra
marginem manifestis, venis clathratis + 1 cm. remotis; petiolo 2—2.5 cm
longo, gracili; inflorescentiis terminalibus cymoso-paniculatis, + 8 cm.
longis latisque, glabris, ramis gracilibus; pedicellis circiter 7 mm. longis;
bracteis non visis; calyce cyathiformi, 5-6 mm. longo, 5 mm. diametro,
limbo subirregulariter 4-lobato, petalis 6, oblanceolato- oblongis, 12 mm.
longis, apice obtusis; staminibus 12, filamentis complanatis, 9 mm. longis,
antheris 7 mm. longis, linearibus, apice reflexis, postice calcaratis, calcare
1 mm. longo; stylo 10 mm. longo; ovario 5- vel 6-lo oculato.

SoLoMoN Istanps: Guadalcanal: Uulolo, Tutuve Mountain, Kajewski 2515
(TYPE), January 1931, alt. 1200 m., rain-forest, common (small tree 7-8 m. high;
flowers light green).

In habit, Astronidium montanum suggests A. victoriae (Gillespie) A. C.
Sm., of the Fiji Islands, but the leaves of the former are not rounded or
obtuse at the base, and the flowers are about twice as large and lack the
lepidote character of those of the Fijian species.

Astronidium muscosum sp. nov.

Arbuscula 3 m. alta; nodis perspicue setuloso-pilosis, internodiis, basi
setuloso-pilosa excepta, glabris, valde compressis sulcatis; foliis coriaceis
late ellipticis vel obovato-ellipticis, 18-29 cm. longis, 10-16 cm. latis, basi
rotundatis vel obtusis, apice acutis vel breviter acuminatis, supra glabris

1943] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 437

subtus nervis venisque pilosis deinde glabratis, quintunerviis vel quintupli-
nerviis (nervo marginali non incluso), nervis supra insculptis subtus per-

prominulis, venulis subtus distinctis; petiolo 4—7 cm. longo, basi (1 cm.)
ee ee caeterum glabrato; inflorescentiis paniculatis terminalibus,
‘10-12 cm. longis latisque, ramis oppositis + crispe pilosis; floribus sub-
sessilibus ad ramulorum hirtellorum apicem dense glomeratis bracteatis,
bracteis caducis; calyce clavato,.5 mm. longo, tubo 3.5 mm. longo basim
versus minute consperse setuloso, lobis 5 obtusiusculis; petalis 5 in ala-
bastris 34 connatis; staminibus 10, filamentis 3.5 mm. longis complanatis,
antheris laesis eae calcaratis, calcare 1.2 mm. longo, obvio reflexo apice
expanso; stylo 4

OLOMON ISLANDS: San edstabet : Hinuahaoro, Brass 3035 (TYPE),
September 1932, alt. 900 m., mountain rain-forest, rare (large shrub or small tree 3 m.
high, with stiff ascending branches). Probably also belonging here is Kajewski 2507
from Uulolo, Tutuve Mountain, Guadalcanal. The specimen has almost glabrous
leaves and flower-buds, and the receptacle is densely pilose around the base of the
calyx

In leaf-outline and in contour of the flower-buds, Astronitdium muscosum
suggests A, Brassii Markgr. of New Guinea, but the latter has chartaceous
triplinerved leaves, a smaller and much more open inflorescence, and lacks
the obviously coarsely hairy petiole-bases and nodes which are so character-
istic of our species. The specific name is chosen to indicate the last men-
tioned character.

Astronidium salomonense sp. nov.

Arbor usque 15 m. alta; ramulis novellis valde compressis tetragonis
anguste alatis, minute lepidotis; foliis subcoriaceis ellipticis, 8-14 cm.
longis, 3.5-7 cm. latis, basi cuneatis, apice abrupte breviter acuminatis,
supra glabris vel ‘utrinque minute lepidotis, triplinerviis, nervis in laminae
parte centrali 1.5-2.5 a costa remotis, venis subtransversis clathratis utrin-

et ramulis valde compressis tetragonis; pede circiter 1 cm. longis;

alabastris tantum visis; calyce ovali, 9 mm. longo, 6-7 mm. diametro, apice

oe brevibus fissuris aperiente; petalis 6; staminibus 12; ovario
culato

Sotomon IsLtanps: Bougainville: Kieta Gold Field, Kajewski 1703 (TYPE),
April 1930, alt. 1000 m., rain-forest, common (small tree up to 15 m. high; flower-buds
green).

The general habit of Astronidium salomonense is similar to that of 4.
aneityense (Guillaumin) A. C. Sm., of the New Hebrides, but the leaves
of the former are more nearly coriaceous and opposite, the flower-buds are
larger, and the branchlets are definitely tetragonous.

Astronidium sessilifolium sp. nov.

Arbor usque 10 m. alta, glabra; ramulis novellis valde compressis, leviter
sulcatis, parce minute lepidotis: foliis chartaceis sessilibus obovato-ellipticis,
15-25 cm. longis, 5-10 cm. latis, basim versus angustatis, apice obtusiusculis
vel forsan acutiusculis, subtus minute granulosis, triplinerviis, nervis a
costa supra basim 2.5— 6 cm. ortis, in laminae parte centrali 3 cm. a costa

438 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

distantibus; venis ascendenti- patentibus, circiter 1 cm. remotis, subtus
prominulis, nervo marginali 4-7 mm. intra marginem disposito ; inflore-
scentiis terminalibus, circiter 15 cm. longis et 10 cm. latis; axi, ramis ramu-
lisque subangulatis valde compressis, sulcatis; pedicellis 5-8 mm. longis;
alabastris globosis, + 6 mm. diametro, calyce apice vix aperto; geen
6- vel 7-loculatis, depresso-globosis, 4— 5 mm. altis, 7-8 mm. diametro, apice
calycis limbo recto (3 mm. longo) margine leviter 6-lobato ue
seminibus 0.8 mm. longis, lineari-angulato-clavatis

Sotomon Istanps: Bougainville: Siwai, a tae 193 (ser. no. 22983)
(rypPE), January 1933, shrub or small tree near water; Koniguru, Buin, Kajewski 2163,
August 1930, alt. 900 m., rain-forest, common (small tree up to 10 m. high; the flower-
buds have a disagreeable odor when crushed)

This species, in the shape of the leaves and the lack of a petiole, suggests
the Fijian Astronidium sessile (A. C. Sm.) A. C. Sm., but the main nerves
in the latter lie 0.8-1.5 cm. within the margin, whereas in the former they
are almost half way between the midrib and the margin; the flowers too
are smaller in the Solomon Islands material and the bracts of the young
inflorescence are oblong rather than orbicular.

Astronidium anomalum sp. nov.

Arbor parva 6-7 m. alta glabra; ramulis cae gs brunneis; foliis
tenuiter coriaceis lanceolato- ellipticis, 5-8 cm. S, 3 cm. latis,
utrinque angustatis, basi obtuse cuneatis, apice Berd 7. breviter acumi-

1 gine is, in sicco supra olivaceo-viridescentibus,
subtus brunnescentibus, nervis primariis utrinsecus 11-14 late patentibus
fere subtransversis, supra subobscuris, subtus manifestis, reticulo subtus
vix manifesto, con nferto, vena margin nali 1-1.5 mm. intra marginem dis-
posita ; petiolo 0.8—1.7 cm. longo, gracili, supra canaliculato; inflorescentiis
4.5—7 cm. longis, 4-6 cm. latis, terminalibus cymoso- -paniculatis, axi, ramis
ramulisque valde compressis, ‘subangulatis : pedice 1.5-2 mm. longis;
floribus non visis; fructibus minute lepidotis vel glandulosis, depresso-

cuneatis, — — vel
suboblique truncatis, + angulatis; placentis 3, circiter 1 m ongis.

Sotomon’ ISLAND Guadalcanal: Uulolo, Tutuve aN Kajewski
2623 (TYPE), May 733 1, alt. 1500 m., rain-forest (small tree up to 6-7 m. high, with
es a twisted habit; fruit cream-green, about 34 mature)

Although the fruit is small, in structure it is like that of Astronidium
ie Gray. The leaves, however, are not triplinerved, as one usually finds in

his genus, but have only the midrib obvious; the lateral pinnately ride
erves are easily seen on the lower surface but are not at all raise

Memecylon Linnaeus
Memecylon papuanum sp. nov.

Arbor parva circiter 9 m. alta, glabra; ramulis teretibus, novellis inter-
dum inconspicue angulatis brunnescentibus: foliis coriaceis late ellipticis,
12-20 cm. longis, 5.7-11 cm. latis, basi cuneato-obtusis, apice abrupte
acuminatis, costa supra impressa, subtus conspicua, venis primariis utrinque
subobscuris vel vix manifestis, utrinsecus -+- 12; petiolo circiter 5 mm.

1943 ] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XIII 439

ve crasso; Het axillaribus brevissimis, 5-7 mm. longis, bracteis
1.5m

ovatis acutis; pedicellis 1 mm. longis: ee ovoideis,
} mm. ton is, vix "4 mm. Gis alyce obconico, 3 mm. longo, truncato;
disco stamina circumdante costato; petalis 4 late ovatis, 2.5 mm. longis,

3 mm. latis, obtusiusculis; staminibus 8; fructibus pedicellatis, a globosis
5 mm. longis, 6 mm. ‘diametro) calycis limbo (vix 1.5 mm. longo)
coronatis.

NORTHEASTERN NEW GurtneA: Mountains near Yaduma, Schlechter 19291, April
1909, alt. 300 m. British New Gutvea: Ihu, Vailala River, Brass 977 (TYPE),
February 1926, rain-forest (small slender tree 30 feet high, with thick pale leaves).

Memecylon papuanum seems to be most like the descriptions of the
Javanese M. excelsum Bl. and M. floribundum Bl., but these have larger
fruits, less dense inflorescence, and longer pediceled flowers.

ARNOLD ARBORETUM,
Harvarp UNIvERSITY.

440 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

FERNS OF THE SECOND ARCHBOLD EXPEDITION TO
NEW GUINEA*

E. B. CopELAND

Tue ferns of the First Archbold Expedition to New Guinea were dis-
cussed by Dr. Carl Christensen, in Brittonia 2: 265-317. 1937. These were
from the high mountains of British New Guinea, and, from the 343 numbers
collected, Christensen described 40 new species.

The Third Archbold Expedition collected in Netherlands New Guinea,
mostly at high altitudes. From an incomparable wealth of more than 900
field numbers, I described 109 species as new. Publication of these began
in the Philippine Journal of Science but was interrupted by war. To record
and guard the names, I published brief diagnoses of 87 species, still awaiting
publication and illustration in Manila, in the University of California Pub-
lications in Botany (18: 217-226. 1942).

The Second Archbold Expedition collected at low altitudes, although far
from the coast, in the plains and foothills far up the Fly River. Such country
is poor in ferns as compared with the mountains, and the ferns which do
occur in the lowlands are mostly species of wide range. It is therefore not
surprising that the 171 collection numbers of this expedition include no
more than five definitely new species. Descriptions of these and comment
on a few other species follow. All the numbers cited are represented in the
Gray Herbarium, and the types of new species, unless otherwise indicated,
are deposited in the author’s herbarium
Cephalomanes Ledermanni (Brause) comb.

Trichomanes Ledermanni Brause in Bot. Jahrb. ‘36: 35. 1920.

British New Guinea: Fly River, 528-mile Camp, alt. 80 m., in ridge forests,
Brass 6663.

Related to C. atrovirens Presl (Trichomanes rhomboideum J. Sm.), but
distinguished, as correctly indicated by Brause, by smaller size, lax vena-
tion, and small, obconic involucres. The sori of our specimen are more
abundant than on Brause’s type, and therefore they occupy a larger part
of the frond and occupy the ends as well as the acroscopic sides of the
pinnae; some occur even below the ends, on the basiscopic sides.

Trichomanes maluense Brause is distinguished in part by just such a
more ample production of sori. Whether or not its rather feeble other
differences are more significant, I do not venture to guess.

Cyclosorus gregarius sp. nov.

C. rhizomate adscendente, sicco 6 mm. crasso, apice paleis fuscis lineari-
lanceolatis glabris debilibus 3-4 mm. longis vestito; stipitibus approxi-
matis, usque ad pinnas reductas 25 cm., ad pinnas normales 70 cm. altis,

*Botanical Results of the Richard Archbold Expeditions.

1943 ] COPELAND, FERNS OF NEW GUINEA 441

rhachibusque glabris; fronde, pinnis basalibus remotis abrupte reductis
exclusis, ee . cm, alta et 25 cm. lata, pinnata, pinna apicali caeteris con-
forme ore; pinnis normalibus remotis, alternantibus, sessilibus,
inferioribus re 15 cm. lon ngis et 15 mm. latis, ’gradatim acuminatis, basi
late cuneatis, leviter crenato-lobatis lobis crenulatis, herbaceis, ‘costa
superne setulis inconspicuis inflexis praedita, aliter glabrescentibus, venulis
acroscopicis 3, basicopicis 2 anastomosantibus; soris medialibus vel infra-
medialibus, indusio parce et breviter ciliato, caduco.

riTISH NEw GuINEA: Fly River, 528-mile Camp, alt. 80 m., ae 6759 (TYPE),
“oregarious in clumps over 1 m. high, on mud in shaded creek bott tom

Indusia can be detected only on the youngest sori.

Lindsaea subtripjnnata sp. nov.

L. gregis L. heterophyllae Dry. et. orbiculatae (Lam. ) Mett., fronde

apice rotundatis, venis em, soro continuo ‘vel rarius interrupto, indusio
cum margine conterm

British New Gui ree , Was parse ra Western Division, Brass 8491
(TYPE), common on cae of pallies in rain-for

The stipe is about 30 cm. and the fit aeont 25 cm. long.

Great as is the variety of fronds referred to L. orbiculata and L. hetero-
phylla, L. subtripinnata seems sufficiently distinguished by its small pin-
nules and the absence of larger undivided pinnae. The free venation is
correlated with the fine dissection of the frond. I have no New Guinean
specimen of either L. heterophylla or L. orbiculata, but the former has
been reported from the island. The group runs riot in New Caledonia.

Oleandra subdimorpha sp. nov.
Epiphytica, caudice gracili, paleis supra basim peltatis, nigris marginem
lacerum versus pallescentibus ad ramos laterales breves imbricatis ad

frondium fertilium paullo longioribus; fronde sterili ca. 18 cm. longa et
4 cm. lata, basi subinaequaliter rotundata, apice abrupte in caudam angus-
tissima 3 cm. longam contracta, coriacea, glabra, venis arcte approxi-

bus, indusiis late reniformibus, oblique insertis, atrocastaneis, Coriaceis.
BritisH NEw GurINneEa: Palmer River, 2 miles below junction with Black River,
alt. 100 m., Brass 6886 (TYPE, in Gray Herb.), “stiff climbing epiphyte.”
A relative of O. Werneri Ros., but less dimorphic, the base of the sterile
frond broader, the sori farther from the margin.

Humata papuana sp. nov.

H. gregis H. repentis, rhizomate gracili late repente, paleis atrofuscis
lanceolatis 3 mm. longis tum demum deciduis vestito, deinde glauco; frondi-
bus remotis, dimorphis, i i
usque ad 5.5 cm. longis, sparse et decidue squamiferis, laminis deltoideis
usque ad 5 cm. longis, pinnatis pinnis infimis tantum pinnatifidis sessilibus
apice rotundatis coriaceis glabris, segmentis sequentibus lobatis, superiori-

442 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXIV

bus integris; frondium fertilium stipitibus 9 cm. altis, gracilibus, laminis
7 cm. longis late deltoideis basi tripinnatifidis, soris et axialibus et ad bases
dentium brevium nao or ila indusiis quam longis multo latioribus,
etenim marginem ee

British New Guinea: Palmer River, 2 miles below junction with Black River,
alt. 100 m., Brass 6087 (TYPE), “matted on branches of tall as trees.” Brass
6593, Fly River, 528-mile Camp, alt. 80 m., “creeping in moss mats high on canopy
trees, common,” is a depauperate form of the same species, the sterile aa at most
2 cm. long, on stipes 7 mm. lon

Related to H. kinabaluensis, which has the teeth subtending the sori more
completely suppressed, and to H. pusilloides, which has them much more
conspicuous. Humata alpina var. edentula Ros. is like H. kinabaluensis in
suppression of the teeth; judging by a single specimen, the sterile frond is
more contracted, but with longer pinnae.

? Polypodium neglectum Blume, Enum. Pl. Jav. 121. 1828, Fl. Jav. Fil. 133. pl. 54,
f. 1. 1828.

British New Guinea: Palmer River, alt. 100 m., common on branches of tall
trees, Brass 6881.

The identification is by description and doubtful. The rhizome is
“repens, filiforme, tenue, ramosum, paleis lineari-lanceolatis acutissimis
... imbricatis albo-scariosis . . . vestitum,” quoting, with omissions, Presl,
Epim. 124; Presl’s description is likely to have been based on a “Manila”
plant of Meyen. Van Alderwerelt’s description, Malayan Ferns 435, fits
Brass’ plant in most respects. Backer & Posthumus, Varenflora voor Java
195, reduce P. neglectum to P. stenophyllum Blume, which seems most
unlikely to be correct. Brass’ plant is certainly not P. stenophyllum.,

P. redimiens Brause is also known to me from description only. This
description fits Brass’ plant as to the fronds, but the rhizome is said to be
“pallidum,” “auffallend bleich,” while that of Brass’ plant is partly fuscous,
mostly black; and the paleae are said to be “clathratis, deltoideis margine
spinuloso-dentato,” all of which is inappropriate. However, I suspect its
identity with our plant. There is near affinity to Polypodium pyrolaefolium
Bergsmann, the type of Crypsinus; and, still nearer, to P. Whitfordi
Copel., of Luzon.

Selliguea Archboldii sp. nov.
S. gregis S. Feet, rhizomate late repente, paleis nigris, basi peltatis fusco-

)
ceis, basi cuneatis, stipitibus 6 cm. longis, venis haud occultis; frondibus
fertilibus 10-12 cm. longis et 12—15 mm. latis, basi attenuatis; soris super-
ficialibus, e costa ad eg a protensis.

British New Guinea: Fly River, 528-mile Camp, alt. m., Brass 6836 (TYPE,
in Gray Herb.), “creeping epiphyte, common on branches of eh canopy trees; fronds
very stiff.”

Most like S. feezoides Copel., of Fiji, Samoa, and Tahiti, but smaller,
with smaller, darker and narrower paleae.

1943 ] COPELAND, FERNS OF NEW GUINEA 443

a ane dispar ee in Nova Guinea 8: 155. 1909; v. A. v. R. in Bull. Jard.
t. Buitenz. II. 1: 4. pl. 2, f. 2, 3.1911
H NEw Ee Palmer River, alt. 100 m., Brass 6872, ey (2935; 7361;
epiphy tie, ee on mossy branches of canopy trees; very unifor
The fronds are short-stipitate rather than subsessile, es the conformity
with Christ’s description is reasonably close. I find nothing like the lid over
the soral cavity, nor the great tufts of hairs described and figured by van
Alderwerelt. The indument of the nether surface of the sterile frond is
persistent. The sori are in an irregular row, or in two or three hardly dis-
tinguishable rows. On fully fruiting fronds, the sori come into contact as
they expand.

UNIVERSITY OF CALIFORNIA,
BERKELEY, CALIFORNIA.

444 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

NOTES ON THE FLORA OF KWANGSI PROVINCE, CHINA
Hut-Lin Li

Kwancsi Province, in southern China, bordering on Kwangtung
to the east, Yunnan to the west, and Tonkin to the south, received
comparatively little attention from botanical collectors in the early
years of the botanical exploration of China. In recent years, however,
extensive and important collections have been assembled from Kwangsi,
particularly through the initiative and interest of Prof. W. Y. Chun,
Director of the Botanical Institute, Sun Yatsen University, Canton, Dr.
F. P. Metcalf of Lingnan University, Canton, and Dr. A. N. Steward of
the University of Nanking, Nanking, China. Various expeditions organized
by these men, and supported, in part, by grants made from the Arnold
Arboretum of Harvard University, have operated in most parts of
Kwangsi. Yet it is clear that the area has not been thoroughly covered
from a botanical standpoint, and the desirability of additional future
explorations is indicated. The material assembled at the Arnold Arboretum,
on which this study is based, represents that institution’s share of the
collections made under codperative arrangements with the several Chinese
institutions.

This study is based on representatives of a few selected families of plants
and clearly indicates what may be expected in the form of additions to our
knowledge of the flora of China as the work of identification progresses.
Twenty-six new species and two new varieties are described, nine of the
species having been so designated by Dr. E. D. Merrill in his preliminary
work on the collections. This study was made possible by a grant from
the Milton Fund of Harvard University to Dr. Merrill, to assist him in
preparing data for publication on the very extensive collections of Chinese
botanical material assembled at the Arnold Arboretum in recent years.

PROTEACEAE
Helicia Loureiro
Helicia vestita W. W. Smith var. mixta var, nov.

A typo differt foliis brevioribus latioribus integris vel sursum irregulariter
pauce dentatis.

‘AN: Man-ning, S. K. Lau 28252, Nov. 26, 1936, a tree 10 m. high, in forest,
fruit green. Kwanctunc: Shih Wan Tai Shan, H. Y. Liang 69841, July 23, 1937, a
tree 12 m. high, in mixed forests. Kwanost: Shang-sze District, Shih Wan Tai Shan,
near Iu Shan Village, W. 7. Tsang 22430, June 2-7, 1933, a tree 22 ft. high, fairly
common in thickets, flowers yellow; Tseung-yuen, Liow Shiang, C. Wang 39624
(TYPE), June 30, 1936, a tree 20 m. high, in thin woods, flowers white.

These specimens closely resemble each other and represent a form differ-
ing from the Yunnan type in that the leaves are shorter and relatively
broader (about 10-16 cm. long and 5.5—6 cm. broad), entire or with only
a few shallow distant teeth near the apex.

1943 ] LI, NOTES ON THE FLORA OF KWANGSI 445

ANNONACEAE
Orophea Blume
aii anceps Pierre, Fl. For. Sapeet 1: t. 46. 1881; Finet & Gagnep. in
mte, Fl. Gén. Indo-Chine 1: 117.
ANGSI: Pin-lam, S. P. Ko 55657, ne . 1935, a shrub on forested slopes, fruits
d.

Although three species of Orophea are known from Hainan, this is the
first record of the genus in continental China. Indo-China.

HAMAMELIDACEAE
Corylopsis Siebold & Zuccarini
Corylopsis cordata Merrill in herb. sp. nov.

Frutex 2-3 m. altus, ramulis glabris rubro-brunneis parce lenticellatis;
foliis subchartaceis petiolatis glabris utrinque subconcoloribus oblongo-

acute acuminatis, basi distincte cordatis, margine sinuato-dentatis, dentibus
longe mucronulatis, nervis lateralibus utrinsecus 7—9 laxis, inferioribus
parce ramosis, cum costa supra leviter impressis, subtus elevatis perspicuis,
venulis dense reticulatis tenuibus, supra leviter elevatis, subtus perspicuis;
petiolis 2—2.5 cm. longis glabris: floribus ignotis; infructescentiis sub-
spicatis vel racemosis, 3—3.5 cm. longis, ee: circa 1 cm. ane
pubescentibus vel glabrescentibus: capsulis 1 cm. longis, 8 mm. crassis,
glabris, brunneis subsessilibus vel breviter crasseque pedicellatis: pee ra
nigris, 8 mm. longis, laevibus, nitidis.

Kwancsi: Shang-sze ene Shih Wan Tai Shan, near Iu Shan Village, W. 7.
Tsang 22261 (TYPE), May 1 » 1933, a fairly common shrub 2 m. high in thickets; Shih

high, fairly common in thickets; Shih Wan Tai Shan, Nam She Village, W. 7. Tsang
24747, Nov. 26, 1934, like the preceding number

A species close to Corylopsis Wind Hemsley, differing in the broader
leaves, which are glabrous on both surfaces and more strongly toothed, the
shorter infructescences, and the smaller fruits.

Eustigma Gardner & Champion
Eustigma jogger! Oliv. in se ie Pl. 20: ¢. 1954, 1891; Guillaum. in Lecomte,
Fl. . Indo-Chine 2: 710

ee Lin Yuin District, a ive N. Steward & H.C. Cheo 696, June 14; 1933,
a shrub 6 m. high, in valley, alt. 1000 m., fruits (immature) green.

This species was originally described from Tonkin, in Indo-China. New
to China.

SIMARUBACEAE

Brucea J. F. Miller
Brucea acuminata sp. nov.

Frutex, ramis rubro-brunneis subdense albo- ae aide foliis 40-45 cm.
longis, thachibus teretibus puberulis, petiolis 7—8 ‘lon ngis, teretibus
puberulis; foliolis circa 15, oppositis, breviter paicliiaes chartaceis,
oblongo-lanceolatis, 5-8 cm. longis, 1.5-2.5 cm. latis, longe graciliter
acuminatis, basi obtusis vel late acutis, aequalibus vel subobliquis, margine

446 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

integris, supra atro-viridibus, subtus viridibus, utrinque minute consperse
pubescentibus, nervis lateralibus utrinsecu s 8-10, supra subconspicuis,
subtus elevatis distinctis, venis tertiariis ea inconspicuis, subtus con-
spicuis vel obscuris, petiolulis ad 3 mm. longis; floribus ignotis; infruc-
tescentiis axillaribus (ut videtur, plerumque in axillis defoliatis) andes
elongatis, ad 15 cm. longis, rhachibus parce puberulis vel glabratis, pedicellis
4-5 mm. longis, puberulis, calyce persistente minuto 4-partito, fructibus
ovoideis, 8-9 mm. lon gis, 6-7 mm. crassis, in sicco brunneis (ex collectore
rubris), extus glabris in sicco subreticulatis.

Kwancsr: Ching Hsi District, S. i _ 56114 (Type), Dec. 10, 1935, a shrub along
roads in the margins of thickets or fore

This species is apparently pr to Brucea mollis Wall., originally de-
scribed from India, its variety tonkinensis Lecomte being “recorded from
Indo-China and from southern China. The new species is distinguished
by its much smaller, narrower, longer acuminate, and shorter petiolulate
leatlets and its slender infructescences.

STERCULIACEAE

Reevesia Lindley
Reevesia tomentosa sp. nov.

Arbor circa 12 m. alta, ramis teretibus stellato-tomentosis, ramulis
dense stellato-tomentosis, indumento ferrugineo; foliis subcoriaceis oblongo-
ovatis, 8-14 cm. longis, 3-6 cm. latis, acutis vel obtusis, basi rotundatis
vel obscure subcordatis, supra olivaceis, disperse stellato-tomentosis, sub-
tus pallidioribus dense brunneo- tomentosis, nervis lateralibus utrinsecus

basi acutis, extus dense stellato-tomentosis, indumento ferrugineo; pedi-
cellis 2.5-3 cm. longis; seminibus circa 2.6 cm. longis, alis brunneis circa
2.2 cm. longis, basim versus 0.8 cm. latis, oblongis, apice oblique rotundatis.

Kwanocsi: Yung District, Ta Tseh Tsuen, A. N. Steward & H.C. Cheo 922 (TYPE),
Sept. 3, 1933, a tree 12 m. high, valley roadside, alt. 350 m.

A species allied to Reevesia pubescens Mast., but the leaves are more
densely tomentose beneath and also scattered stellate-tomentose above, and
the mature fruits are densely covered by brownish stellate hairs.

FLACOURTIACEAE
Hydnocarpus Gaertner
Hydnocarpus Merrillianus sp. nov.

Arbor, ramulis teretibus dense fulvo- Selgaeieigate foliis chartaceis
oblongo- -ilipticis, 18-25 cm. longis, 6.5—11.5 latis, abrupte acutis,
basi late acutis, margine integris leviter ee wi supra atro- viridibus
glabris, subtus viridibus parce pubescentibus, venis lateralibus utrinsecus
7 vel 8, supra conspicuis, subtus elevatis distinctis, valde arcuatim adscen-
dentibus, venulis reticulatis, utrinque perspicuis; petiolis 1.5-3.5 ¢
longis, dense fulvo- pubescentibus: floribus ignotis; fructibus magnis axil-
laribus solitariis globosis, junioribus dense fulvo-pubescentibus, maturis

1943 ] LI, NOTES ON THE FLORA OF KWANGSI 447

castaneo-velutinis, 8 cm. diametro, pericarpio 5 mm. crasso; seminibus
numerosis compresso-ovoideis, 2.5 cm. longis, 1.7 cm. latis; pedicellis 1 cm.
longis.

WANGcsI: Tai Chin Shan, S. P. Ko 55311 (Tyree), June 14, 1935, 55421, July 4,
1935, a tree, in woods or borders of woods on slopes.

This is the second species of the genus known from China and the first
one from continental China, ce aes H ydnocar pus hainanensis (Merr.)
Sleumer, being known from Hainan. Thi is new species is characterized by
the large, entire leaves with ieee ne eee petioles, and the large globose
fruits.

PASSIFLORACEAE

Passiflora Linnaeus
Passiflora Papilio sp. nov.

Suffruticosa scandens glabra, ramis subligneis gracilibus 2.5 mm. dia-
metro, ramulis ultimis 1 mm. diametro; foliis subchartaceis, petiolatis, supra
in sicco olivaceis, subtus pallide glaucescentibus minute consperse puberulis,
3-5 cm. longis, 9.5-12 cm. atis, basi rotundato-truncatis, apice ee
retusis in lobos 2 ovatos magnos divergentes rotundato-acuminatos abeunti-
bus, quove 5—7 cm. longo, 2.5—4 cm. lato, margine integris, costa ae in
mucronulum brevissimum abeunte, nervis lateralibus circa 3, jugo maximo

bus glandulas binas ferentibus; cirrhis gracilibus glabris ad 6 cm. longis;
floribus ignotis; fructibus globos sis, 1-1.2 cm. diametro, pedicellis gracilibus,
1 cm. longis, medium versus articulatis: aa a 5 mm. longis, cinereo-
nigris foveolatis.

Kwancsi: Tai Chin Shan, S. P. Ko 55426 (type), July 7, 1935, scandent on trees
on slopes.

A very distinct species, characterized by the broad, divergently 2-lobed
leaves, which suggest a butterfly in shape, and the small fruits. The field
label has notes on the flowers, but the specimen studied has only detached
fruits, some of them with remnants of the calyx. The available flowering
parts are too fragmentary for description, although, judging from the
unique shape of the leaves as compared with other Chinese species, the
flower is apt to be of interesting structure.

Passiflora kwangsiensis sp.

Passiflora cupiformis sensu Chun, Sunyatsenia 4: 184. 1940, pro parte; non Masters.

Suffruticosa scandens glabra, ramis subligneis. 3 mm. diametro; foliis
chartaceis glabris late ellipticis vel leviter obovato-ellipticis, 8-11 cm.
longis, 7-9.5 cm. latis, latissime rotundatis vel truncato-rotundatis, obscure
undulatis, basi rotundatis ad modum angustissime peltatis, 3— 5-nerviis,
integris, supra in sicco olivaceis, subtus pallidis, costa nervisque utrinque
elevatis distinctis, nervis inferioribus ramosis, venulis dense reticulatis,
utrinque elevatis conspicuis; petiolis 4-5 cm longis, in partibus inferiori-
bus glandulas binas ferentibus; cirrhis gracilibus glabris usque ad 7 cm.

S
ad

ca . diametro, glabris; pedicellis 1.5 cm. longis, medium versus
articulatis: seminibus 3.5 mm. longis, pallidis, foveolatis.

448 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Kwan Ling Yuin District, S. K. Lau 28577 (type), July 10, 1937, scandent in
light woods, traits green

A species apparently close to Passiflora cupiformis Mast., but differing
in the vegetative details, the leaves being broadly elliptic to obovate-elliptic,
broadly rounded or truncate-rounded at their apices and sometimes even
undulate, their bases being broadly rounded and sometimes very narrowly
peltate.

THY MELAEACEAE
Wikstroemia Endlicher

Wikstroemia paniculata sp. nov.

_Frutex, ramis ramulisque teretibus glabris gracilibus, ramulis ultimis

metro; foliis oppositis chartaceis breviter petiolatis oblongo-

ellipticis vel eset AE ad 5.5 cm. longis et 1.6 cm. latis, plerumque acutis,
basi acutis vel rotundatis, margine leviter revolutis, utrinque glabris, supra
viridibus, subtus paullo pallidioribus, nervis lateralibus circa 20 utrinque
subconspicuis in venam a margine ipso circa 1 mm. remotam cum margine
parallelam anastomosantibus, venulis obscuris; petiolis ad 3 mm. longis;
inflorescentiis terminalibus vel lateralibus paniculatis valde ramosis, ad

persistentibus ; floribus luteis parvis, ad 4 mm. longis, 4-meris, perianthii
tubo circa 4 mm. longo et 1.5 mm. crasso, ae ‘pubescente, lobis 4 minutis
0.5 mm. longis; antheris 8, 2-seriatis, linearibus, 0.5 mm. longis; ovario
glabro; fructu immaturo ovoideo, 6 mm. ongo.

K WANG Ching Sai Village, S. P. Ko 55710 (type), Sept. 7, 1935, a shrub in open
places near slopes, flowers yellow

A distinct species, strongly characterized by its leaves with lateral veins
united into a single vein parallel with and close to the leaf-margins, and by
the much-branched panicles bearing rather small flowers.

ALANGIACEAE
Alangium Lamarck
Alangium Chungii sp. nov.
Frutex vel arbor parva, ramulis ultimis atro-brunneis gracilibus teretibus
leviter adpresse tomentosis vel glabrescentibus; foliis chartaceis longe

valde inaequilateralibus, 3—6-nerviis, margine integris, supra glabris costa
nervisque interdum parce tomentosis exceptis, nervis lateralibus utrinque
4—6 prope marginem anastomosantibus, utrinsecus conspicuis, venulis dense
reticulatis, utrinsecus perspicuis; petiolis 5-10 cm. longis, teretibus minute
adpresse tome ntosis; inflorescentiis axillaribus adpresse tomentosis vel
subtomentosis, ad 8 cm. longis, 1- vel 2-ramosis, 4—7-floris, pedunculis
2-3.5 cm. longis, pedicellis 1—2 cm. longis; floribus 2—2.5 cm. longis; calycis
tubo infundibuliformi, 2 mm. longo, leviter tomentoso, margine minute
5—7-lobato; petalis 6 vel 7 basi leviter cohaerentibus, lanceolatis, 2—2.5 cm
longis, 1.5 mm . latis, extus tomentosis, intus glabris; staminibus 6 vel 7,
circa 1.8 cm. longis, filamentis 6 mm. longis, dilatatis dense villosis, antheris
linearibus, 1.2 cm. longis, connectivo villoso; disco su bgloboso: ovario
1-loculari, stigmate capitato 4-partito; fructu ignoto
KwancsI: no data, Z. S. Chung 82038 (TYPE).

1943] LI, NOTES ON THE FLORA OF KWANGSI 449

A species related to Alangium platanifolium (Sieb. & Zucc.) Hance and
A. barbatum (C. B. Clarke) Harms, differing from the former in the strictly
entire leaves and the smaller Beer and from the latter in the longer
petiolate leaves, larger flowers, and the densely villose filaments.

CLETHRACEAE

Clethra Linnaeus
Clethra Liangii sp. nov.

Frutex 2—2.5 m. altus, ramis brunneis, ramulis glabris vel novellis plus
minusve tomentosis; foliis chartaceis oblongis vel late oblongo-lanceolatis,
6-12 cm. longis, 1.8- 3.5 cm. latis, acuminatis, deorsum plus minusve
angustatis, basi acutis, distanter serrulatis, deorsum integris, supra atro-
viridibus primo minute stellatis, subtus pallide viridibus primo tomentosis,
utrinque mox glabris, costa supra impressa, subtus valde elevata, nervis
lateralibus utrinsecus 8-12, supra leviter impressis, subtus perspicuis,
rare anastomosantibus, venulis subconspicuis vel obscuris; petiolis
6-10 mm. longis; inflorescentiis 1—7 terminalibus racemosis, a cm.

lineari-obovatis, 5 mm. longis, glabris; filamentis 3.5 mm. longis, inferne
dilatatis, antheris 1.5 mm. longis; ovario dense pubescente; stylis ad 6 mm.
longis, glabris, stigmate minute 3-lobato; fructu immaturo 3 mm. diametro,
pubescente, stylo persistente, 5 mm. lon ngo.

Kwanctunc: Ta Mien Shan, Shih Wan Tai Shan, H. Y. por. 69645 (TYPE),
July 14, 1937, a shrub 2 m. high, in dense forest, flowers white. Kwancsr: Pin- lam,
Ching Sai, S. P. Ko 55527, Aug. 22, 1935, a shrub 2.5 m. high, on forested slopes, alt.
900 m., flowers white.

A species possibly most closely allied to Clethra Bodinieri H. Lév., differ-
ing, among other characters, in the numerous racemes with dense brown
non-appressed indumentum and stouter pedicels.

Clethra Ea sp. nov.

Frutex, ramis robustis brunneis junioribus dense breviter substellatim
fulvo- pipe taba foliis subchartaceis lanceolatis, circa 15 cm. longis et
4 cm. latis, longe acuminatis, basi subrotundatis vel late acutis, margine

mosantibus, supra eabennepue: subtus perspicuis, venulis sab conepicnis
petiolis i 2 cm. longis, pubescentibus; racemis circa 7, subfasciculatis,

calyce 2.5 mm. longo, 5-dentato toto subalbido-pubescente; petalis 5
obovatis, 5 mm. longis, apice valde truncatis; filamentis gracilibus 4-5 mm.
longis, antheris 1 mm. longis; ovario pubescente: stylis 3 mm. longis,
glabris, stigmate integro.

450 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Kwancst: Ling Wan District, S. K. Lau 28767 (Type), in 1937, no field notes
available.

This species is very close to Clethra kaipoensis H. Lev. , differing chiefly
in the more narrowly lanceolate leaves, with more eins lateral veins
and closer denticulations, and the elongated lanceolate bracteoles.

COMBRETACEAE
Combretum Linnaeus

Combretum kwangsiense sp. nov.
_ Frutex scandens, ramis cinereo-albidis, glabris, cortice longitudinaliter
moso, ramulis dense lepidotis, lepidibus minutis: foliis chartaceis —
antes ellipticis vel late oblongo- ellipticis, 14-17 cm. longis, 7-10 c
latis, rotundatis, basi late acutis, margine leviter revolutis, supra olvaceo-

floribus ignotis; sued anaes racemosis simplicibus axillaribus, 6-8 cm
longis, polycarpicis, pedunculis 1.5—2 cm. longis, minute lepidotis, fructibus
in stipitibus circa 1 mm. longis, luteo-brunneis, e collectore luteis, sparse
minute glandulosis (glandulis flavidis), nitidis, in ambitu ellipticis vel
elliptico-ovoideis, ad 15-18 mm. longis et latis, 4- alatis, utrinque rotundatis.

Kwancsr: Yang District, Ch’ang ne. oo & Cheo 1194 (type), Oct. 23, 1933,
vine, on rocks in forest, alt. 200 m., fruit y

A species allied to Combretum Aljredit Hance, but distinguished by the
larger, broader leaves, smaller infructescences, and much smaller fruits.

EBENACEAE
Diospyros Linnaeus

Diospyros siderophyllus sp. nov.
Arbor parva 5 m. alta, partibus junioribus fructibusque exceptis glabris
nica ee ramis ‘teretibus nigris glabris, ramulis ultimis circa 1.5
brunneis glabris vel breviter uae hirsutis; foliis sub-
coriaceis, See 8-14 cm. longis, 2-4 cm. latis, breviter obtuse acumi-
natis, basi acutis, supra in sicco atro-olivaceis, glabris, subtus paullo
pallidioribus glabris vel consperse breviter adpresse hirsutis, costa supra
impressa, subtus valde elevata, nervis lateralibus utrinsecus 10-12, gracili-
bus curvato-adscendentibus anastomosantibus, supra subconspicuis vel in-

sutis; fructibus axillaribus vel in axillis defoliatis solitariis sessilibus glo-
bosis, circa 2 cm. diametro, dense breviter adpresse brunneo-hirsu - 5- vel
6-locellatis; seminibus circa 8 mm. longis et 5 mm. latis, compressis,
albumine aequabili; sepalis 4 sub fructu patulis valde coriaceis tee tri-
angularibus, circa 8 mm. longis et 6 mm. latis, acutis hir

Kwancsi: Pin-lam, S. P. Ko 55679 (type), Sept. 1, 1935, a oats tree 5 m. high, in
forests on or near slopes; On Tak, S. P. Ko 55773, Sept. 19, 1935, a small tree 5 m.
high, in similar habitats.

In the vegetative characters, this species resembles Diospyros Roi H.

1943 ] LI, NOTES ON THE FLORA OF KWANGSI 451

Lecomte, differing in the leaves being generally narrower, olivaceous when
dry, and with more numerous lateral veins. It differs further in the fruit,
with its persistent calyx, which is densely brownish-appressed-hirsute, while
the seeds are fewer in number.

STYRACACEAE

Styrax Linnaeus
Styrax oligophlebius Merrill in herb. sp. nov.

Frutex 2.5 m. altus, ramulis novellis dense breviter stellatim brunneo-
tomentosis; foliis chartaceis vel subcoriaceis, supra glabris viridibus niti-
dis, subtus dense stellatim subfulvo- tomentosis oblongis, 4-6 cm. longis,
1.5—2.5 cm. latis, breviter acuminatis, basi acutis, integris, nervis laterali-
bus utrinsecus 5 vel 6, cum costa supra subconspicuis leviter impressis,
subtus elevatis prominulis, venis tertiariis utrinque obscuris; petiolis 0.8—
1.3 cm. longis, dense breviter stellatim tomentosis; inflorescentiis infructe-
scentiisque ignotis; fructibus globosis, circa 1.2 cm metro, globosis vel
subovoideis, rotundatis, extus dense cinereo-tom ie eens persistente
cupulato, 6 mm. alto, cinereo-tomentello leviter Gene lobato; pedi-
cello circa 3 mm. longo.

Shang-sze District, Shih Wan Tai Shan, Tang Lung Village, W. T.
Tsang 24489 (TYPE), Oct. 1-16, 1934, scattered shrubs, about 2 m. high, fairly common.

A distinct species, strongly characterized by its rather small, coriaceous,
few-nerved leaves, which are green, glabrous, and glossy above and densely
and brightly brownish-tomentose beneath, with short crowded stellate hairs.

SYMPLOCACEAE
Symplocos Jacquin
Symplocos kwangsiensis Merrill in herb. sp. nov. Subgen. Hopea, § Bobua, Lodhra.
Frutex circiter 2 m. altus, ramis nigris, primo plus minusve pubescentibus,
vetustioribus glabris, ramulis nove is dense brunneo-pilosis atque
ciliatis; foliis chartaceis utrinque subconcoloribus, supra nitidis, subtus
longe ciliatis, oblongo-ovatis vel late ovato- lanceolatis, 2.5-4 cm. longis,
1-1.5 cm. latis, longe acute acuminatis, basi late acutis vel rotundatis, mar-
gine glanduloso-serrulatis, costa supra leviter impressa pilosa vel glabrata,
nervis lateralibus utrinsecus 3 vel 4 utrinque subconspicuis arcuato-anasto-

mosant
scentiis axillaribus in ramulis hornotinis fasciculatis subsessilibus pauci-
floris; floribus sessilibus, bracteis late ovatis, dense pubescentibus, 1 mm.
longis: calycis tubo crasso, circa 1 mm. longo, lobis 5 late ovatis pubescenti-
bus, 1 mm. longis; petalis 5 albis oblongis, 3.5 mm. longis, utrinque glabris;
ei circa 20-25, filamentis liberis, glabris, circa 4 mm. longis; disco
annulari cinereo- pubescente; ovario 3- loculari, a 5 mm. longo; fructibus
ovoideis, 6 mm. longis, 4.5 mm. crassis, viridibu

Shang-sze District, Shih Wan Tai ae Tang Lung Village, W. T.

H.Y. Liang 70044, Aug. 7, 1937, shrub, 2 m. high, in dense woods, fruits green to black.
A distinct species, apparently allied to Symplocos glandulifera Brand
and S. yunnanensis Brand, both of Yunnan, the latter extending also to

452 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Kwangsi and Indo-China. It may be readily distinguished from both of
these, among other characters, by its unusually small leaves.
Symplocos mollipila sp. nov. Subgen. Hopea, § Bobua, Lodhra.

Frutex vel arbor parva, ramulis novellis teretibus gracilibus dense fulvo-
villosis; ramis vetustioribus plus minusve pubescentibus, ata ia pa

molliter pilosis, costa supra leviter impressa dense pubescente, subtus
elevata, nervis lateralibus utrinsecus circa 10, utrinque subconspicuis, prope
marginem anastomosantibus; petiolis 2-3:mm. longis, dense villosis; flori-
bus ignotis; infructescentiis axillaribus fasciculatis subsessilibus 1—4-fruc-
tigeris, bracteis minutis, late ovatis, 1.5 mm . longis, pubescentibus, fructi-
bus ellipsoideis, 1 cm. longis, 7.5 mm. crassis, parce pubescentibus vel
glabrescentibus, leviter longitudinaliter sulcatis, 3-locularibus, lobis calyci-
nis persistentibus rotundatis, 1.5 mm. longis, pubescentibus.

wancsi: Ling Wan District, S. K. Lau 28722, 28725 (TYPE), 1937, no field notes
available.

A species apparently close to Symplocos glandulifera Brand, differing in
the shorter leaves, which are not glandular on the margins and which have
broader, sometimes rounded bases, as well as in the shorter fruits.
Symplocos punctato-marginata A. Chev. ex Guillaum. Bull. Soc. Bot. France 79:

932; Lecomte, Fl. Gén. Indo-Chine 3: 1004. 1933; Merr. Lingnan Sci.
Jour. 15: 424. 1936.

Symplocos Stewardii Sleumer, Repert. Sp. Nov. 42: 266. 1937, syn. nov.

Kwancsi: Yung District, Ta Tseh Tseun, A. N. Steward & H.C. Cheo 771 (iso-
syntype of S. Stewardit Sleumer), Aug. 8, 1933, a shrub 5 m hia te valley forests,
alt. 540 m., flowers white, fragrant, 865 (isosynty pe of S. Steward Sleumer), ion
1933, a tree 12 m. high, in forests, alt. 540 m., flowers white, fragrant; Shang
District, Shih Wan Tai Shan, near Hoh Lung Village: W. T. Tsang 22569 gece
of S. Stewardii Sleumer), June 26, 1933, a shrub 10 ft. high, fairly common, in thickets,
flowers white; Shih Wan Tai Shan, Tang Lung Village, W. T. Tsang 24189, Sept. 4,
1934, a shrub 15 ft. high, fairly common in thickets, flowers white, fragrant, fruits
black.

Originally described from Indo-China. In recording the species from
Kwangtung, Merrill has also noted the occurrence of Chevalier’s species in
Kwangsi, mentioning Steward & Cheo 771 and 865. This was apparently
overlooked by Sleumer. An additional specimen, besides those noted by
Merrill and by Sleumer, Tsang 24189, shows that, together with the other
specimens, the Kwangsi plant is virtually the same as the Indo-Chinese
species.

Symplocos myriadenia Merr. Univ. Calif. Publ. Bot. 10: 428. 1924.
Shang-sze District, Shih Wan Tai Shan, near Hoh Lung Village, W. T.
Tsang 2 22568, June 26, 1933, 22648, July 6, 1933, a small tree 9-15 ft. high, fairly com-
mon in thickets; Ling Wan District, S. K. Lau 28627, 1937. Indo-China. New to
gsi.

crn anomala Brand, var. nitida var. nov.
o differt foliis 8-12 cm. longis, 3-5 cm. latis, supra valde nitidis,
ace calycibusque glabris.

1943 | LI, NOTES ON THE FLORA OF KWANGSI 453

Kwancsi: Waitsap District, Tong Shan, near Sap-luk Po Village, W. T. Tsang
22738, Sept. 8, 1933, 22752, Sept. 10, 1933, a shrub 5-9 ft. high, in thickets, flowers
white, ee tru black; Shang-sze District, Shih Wan Tai Shan, Tang Lung
Village, W. ang 24389, Sept. 30, 1934, 24423 (type), Oct. 1-16, 1934, a shrub
9-15 ft. Meee ah common in thickets, flowers white, fragrant.

This differs from the typical form of the species, which occurs in
Szechuan, Kweichow, Hupeh, Hunan, Kiangsu, Anhwei, Chekiang, Fukien,
Kwangsi, and Hainan, in the larger strongly shining leaves and the
glabrous inflorescences and calyces.

RUBIACEAE
Xanthophytopsis Pitard
Xanthophytopsis kwangtungensis Chun & How, Sunyatsenia 4: 14. pl. 5. 1939.

Kwancsi: Shang-sze District, Shih Wan Tai Shan, W. T. Tsang 22306, May 16,
1933, 22485, June 17, 1933, 23966, July 11-30, 1934, 23975, Aug. 8, — 24532, Oct.
22-31, 1934. A woody plant 1-1/% ft. high, fairly common in thicket

This species, originally described from Kwangtung, is the ae species
of the genus, which was originally described from Indo-China material.
New to Kwangsi. The specimens above cited are all from various localities
in the Shang-sze District, close to the Kwangtung border.

Ophiorrhiza Linnaeus

Ophiorrhiza succirubra King ex Hook. f. Fl. Brit. Ind. 3: 82. 1880.
Kwancsi: Waitsap District, Sze Tze Shan, near Tung Chung Village, W. T. Tsang

ce oa fed, fractant’ Yao Shan, C. Wane 40467, Dec. 14, 1936, near streams, flowers
shit

These specimens seem to represent the same entity as Henry 11345 from
Yunnan, which Dunn, Jour. Linn. Soc. Bot. 39: 471. 1911, has identified
as representing King’s species. Himalayan region to Yunnan, and, if our
identification is correct, also in Kwangsi.

Ophiorrhiza kwangsiensis Merrill in herb. sp. nov.

Herba prostrata crepe 15-18 cm. alta ramosa, caulibus circiter 1
mm. diametro teretibus glabris, ramulis ultimis circa 0.5 mm. diametro
plus minusve em curvato-hirsutis; foliis in oe subaequalibus
chartaceis vel submembranaceis cordato- ovatis, 1.5-1.75 cm. longis, 1—1.5
cm. latis, acutis, basi cordatis, utrinque glabris, in sicco supra subolivaceis,

inconspicuis; petiolis 0.5—1.5 cm. longis, glabris gracilibus; stipulis caducis;
inflorescentiis terminalibus breviter pedunculatis (pedunculo sub fructu ad
12 mm. longo) cymosis paucifloris breviter pubescentibus, floribus imma-
turis circiter 2.5 mm. longis, pedicellis puberulis ad 1 mm. longis; bracteis
bracteolisque linearibus vel lineari-lanceolatis acuminatis circiter 5 mm
yes obscure consperse pubescentibus; sepalis 5 oblongo-lanceolatis acutis
ongis; corolla immatura 3 mm. longa; capsula pedic ellata, circa

6 mm. lata et 2 mm. alta, breviter puberula, pedicellis ad 5 mm.
Kwancsi: Shang-sze District, Shih Wan Tai Shan, Hoh Lung Village, W. T. Tsang

454 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

22425 (typPE), June 4, 1933, 5 in. high, fairly common in thickets on steep slopes ;
Shih Wan Tai Shan, Tang Lung Village, W. T. Tsang 24388, Sept. 30, 1934, fairly
common on steep slopes, flowers pale yellow.

A species characterized by its habit and by its small leaves and flowers.

Luculia Sweet

Luculia intermedia Hutchinson in Sargent, Pl. Wils. 3: 408. 1916.

cst: Pin-lam, S. P. Ko 55584, Aug. 23, 1935, a shrub 1 m. high, in woods on
slopes; Ching Sai Village, S. P. Ko 55718, Sept. 7, 1935, a shrub on slopes, flowers white.
Burma and Yunnan.

Adina Salisbury
Adina Metcalfii Merrill in herb. sp. nov.
Frutex circa 2 m. altus, ramulis novellis castaneis, obscure tetragonis,
glabris vel subglabris; foliis chartaceis petiolatis utrinque glabris, oblongis

costa supra leviter impressa, subtus elevata, nervis lateralibus utrinsecus
8-10, adscendentibus prope marginem anastomosantibus, supra subcon-
spicuis, subtus elevatis, venis tertiariis utrinque subconspicuis; petiolis
1.5-2.5 cm. longis, tenuibus; inflorescentiis terminalibus, ad 8 cm. ongis,
glabris, capitulis circa 1.5 cm. diametro, 5—7-racemosim dispositis, interdum
foliis reductis subtentis, pedunculis gracilibus, 2.5-5.5 cm. longis; calyce
hirsuto, lobis oblongis, circa 1 mm. longis, obtusis; corollae tubo circa 4mm.
longo, glabro, lobis 5 ovatis glabris, circa 1 mm. longis; staminibus fauce
corollae insertis, antheris apiculatis inclusis; stylis exsertis, circa 8 mm.
longis, stigmate subgloboso.

K "e Ch’uan District, Pai-yun-an and vicinity, W. T. Tsang 27683 (TYPE),
June 18, 1937, a shrub about 2 m. high, fairly common in thickets, flowers yellow,
fragrant.

In the racemosely arranged heads, this new species is allied to Adina
racemosa Miq., differing, however, in the much narrower, long attenuate
leaves and the glabrous corollas.

Mussaenda Linnaeus

Mussaenda anomala sp. nov.

pallidioribus, nervis lateralibus utrinsecus 8-10, supra conspicuis, subtus
elevatis curvato-subadscendentibus, venulis reticulatis, supra subconspicuis,

bracteis caducis, bracteolis acuminatis, lanceolatis pubescentibus ad | cm.
longis, deciduis; pedicellis 2-3 mm. longis; calycis tubo oblongo, circa 5

m. longo, adpresse subhirsuto, lobis plerumque 5, omnibus petaloideis, in
lamina ampla petiolatis productis albis ovato-ellipticis, 2-4 cm. longis,
1.5—2.5 cm. latis, acutis, basi acutis vel cuneatis 5-nerviis, margine ner-
visque plus minusve pubescentibus, stipite 1.5—2.5 cm. longo; corollae tubo

1943 | LI, NOTES ON THE FLORA OF KWANGSI 455

(alabastro) circa 1.2 cm. longo et 4 mm. crasso, superne ampliato, extus
dense adpresse pubescente, intus superne dense ‘sulphureo- -piloso, inferne
glabro, lobis plerumque 5, ovatis, breviter acuminatis, circa 3 mm. longis
(immaturis), extus pubescentibus: staminibus 4 vel 5, in tubo insertis,
antheris elongatis, 3 mm. longis, inclusis, subsessilibus; ovario 2-loculari,
disco annulari, stylis circa 6 mm. longis, ‘inclusis, stigmatibus 2 linearibus
compressis, 4 mm. longis; fructu immaturo 6 mm. longo, 4 mm. crasso,
seminibus ignoti

KWANGSI: Thy: -yuen, Yao Shan, C. Wang 39567 (TYPE), June 27, 1936, climbing
on trees.

In the large, broad, thin leaves, this species superficially resembles
Mussaenda Esquiroli H. Lév. (M. Wilsonii Hutchinson). Its one out-
standing character, in which it differs from all other known representatives
of the genus from China and Indo-China, and for that matter from most
previously known representatives of the genus, is that not one but all of
the calyx-lobes are accrescent and petaloid. This character has been noted
in the Philippine Mussaenda philippica Rich. var. aurorae Sulit, Philip.
Jour. Forestry 2: 39. ¢. 3. f. 1. 1939
Mussaenda kwangsiensis sp. nov.

Frutex scandens, ramulis teretibus dense adpresse pubescentibus; foliis
tenuitet chartaceis in paribus aequalibus oblongo-lanceolatis vel lanceolatis,
8-11 cm. longis, 2.5—4 cm. latis, longe acute acuminatis, basi attenuatis
vel cuneatis, supra sparse adpresse pubescentibus, subtus consperse molliter
pubescentibus, nervis lateralibus utrinsecus 6-8 curvato-adscendentibus,
utrinque conspicuis, venis tertiariis a petiolis 5-8 mm. longis,
adpresse pubescen Lone, axillis plerumque folia 2—4 valde reducta gerenti-
bus; stipulis linearibus, 6 mm. longis, ‘Gee inflorescentiis terminalibus
cymosis compactis, circa 4 cm. longis, vix ramosis, perspicue adpresse pube-
scentibus, bracteis bracteolisque linearibus, 0.5-1.5 cm. ongis; floribu
sessilibus confertis; calycis tubo oblongo, circa 5 mm. longo, dense pube-
scente, 5-lobato, lobis normalibus linearibus, 2.5—3 cm. longis, 1-2 mm. latis,
see ee longe acuminatis, lobis petaloideis paucis ovatis, saltem

ngis, 5—7-nerviis, acuminatis, utrinque consperse pubescentibus,
stipitatis: corollae tubo 2—2.5 cm. longo, 1.5 mm. lato, sursum breviter
ampliato, extus perspicue cinereo-pubescente, intus superne villoso, lobis 5
ovatis, 3 mm. longis, 1.5 mm. latis, acuminatis vel mucronatis; staminibus
inclusis, antheris 4-5 mm. longis, filamentis glabris; stylo brevi, 5 mm
longo, gla ro.

Kwancsi: Tseung- iim Yao Shan, C. Wang 40448 (TYPE), Nov. 9, 1936, scandent
in thickets, flowers yello

A species sie cciliys in the group with Mussaenda hirsutula Miq., differ-
ing in the longer and narrower leaves and in the very slender, extremely
long normal calyx-lobes, which mostly exceed the corollas in length.

Mussaenda densiflora sp. nov.

Frutex scandens saltem 2 m. altus, ramis teretibus perspicue villosis;
foliis membranaceis vel tenuiter chartaceis in paribus aequalibus, late
oblongo-lanceolatis vel anguste oblongo-ellipticis, 8-13 cm. longis, 3-5 cm.

latis, longe acute acuminatis " basi acutis, plerumque late acutis, raro sub-
rotundatis, supra conspersissime breviter ciliatis, subtus praesertim secus

456 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

costam nervosque molliter pubescentibus, nervis lateralibus utrinsecus 8-10,
curvato-adscendentibus, utrinque conspicuis, venis tertiariis inconspicuis
vel subconspicuis; petiolis 1-2 cm. longis, pubescentibus; stipulis linearibus,

ongis, caducis; inflorescentiis terminalibus cymosis compactis,
subcapitatis, circiter 6 cm. longis, breviter ramosis, perspicue molliter pube-
scentibus, bracteis bracteolisque linearibus, 1-1.5 cm. longis; floribus

Kwanost: Tai Ching Shan, S. P. Ko 55121 (type), May 25, 1935, a small scandent
shrub, in woods, alt. 1250 ft., flowers yellow; Ching Sai Village, S. P. Ko 55711, a
climber, fruit green.

This species is allied to Mussaenda subsessilis Pierre of Indo-China,
differing chiefly in the much smaller petaloid sepals and the much: longer
and narrower corolla-lobes. A collection from Tonkin, Indo-China, W. T.
Tsang 29049, is referable to this same species. The fruits have been de-
scribed from S. P. Ko 55711, of which the leaf-bases are distinctly rounded
rather than broadly acute.

Randia Linnaeus
Randia salicifolia sp. nov.

Frutex erectus glaber, ramis ramulisque gracilibus teretibus, ramulis
circa 1.5 mm. diametro; foliis chartaceis glabris breviter petiolatis lanceo-
latis, 15-19 cm. longis, 1.5—3 cm. latis, sursum longe angustatis, apice acute
acuminatis, basi acutis, supra atro-olivaceis, subtus pallidioribus, nervis
lateralibus utrinsecus 9-12 gracilibus, utrinque subdistinctis, prope mar-
ginem arcuato-adscendentibus, reticulis obsoletis; petiolo 5-8 mm. longo;
stipulis triangularibus, longe acuminatis, circa 6 mm. longis; floribus igno-
tis: infructescentiis oppositifoliis depauperato-cymosis, circa 1.5 cm. longis,
fructibus junioribus globosis, circa 8 mm. diametro, in siccitate nigris,
nitidis; pedicellis 4 mm. longis, bracteis minoribus oblongis circa 1 mm.
longis.

Kwanost: Ping-nan District, C. Wang 40398 (type), Nov. 2, 1936, a shrub, in
dense forests, fruits green.

This species is manifestly close to Randia Henryi Pritzel, differing par-
ticularly in its elongated narrow leaves and reduced infructescences.

Ixora Linnaeus
Ixora Tsangii Merrill in herb. sp. nov.

Frutex parvus 1 m. altus, inflorescentiis leviter pubescentibus exceptis
glaber, ramis teretibus 3 mm. diametro, internodiis 2.5-3.5 cm. ongis; foliis
chartaceis oblongo-lanceolatis, 13-20 cm. longis, 3-6 cm. latis, longe acumi-
natis, basi late acutis, subolivaceis subopacis, nervis lateralibus utrinsecus

1943 | LI, NOTES ON THE FLORA OF KWANGSI 457

14-16, supra distinctis, subtus elevatis, prope marginem arcuato-anastomo-
santibus: petiolis 1—1. 5 cm. longis; stipulis circa 8 mm. longis deorsum
oblongo- ovatis subabrupte caudato-acuminatis caducis; cymis terminalibus
sessilibus e Higgs ramosis, circa 3.5 cm. longis, leviter pubescentibus tricho-
tome ramosis S primariis haud 1 cm. longis, floribus Sia aie in
ramulis ote in triadibus dispositis, pedicellatis, pedicellis 3-5 m
longis, bracteis lineari-lanceolatis acuminatis circa 3 mm. longis, aaa
minoribus; calycibus glabris circa 1.5 mm. longis, lobis ovatis acuminatis
quam tubo brevioribus; corolla alba (ex collectore) , tubo gracili 1.5--1.8
cm. longo haud barbato, lobis reflexis circa 5 mm. longis subrotundatis:
antheris exsertis lanceolatis circa 3.5 mm. longis, stylis exsertis, ramis 1.5
mm. longis

Kwancsi: Shang-sze District, Shih Wan Tai Shan, Tang Lung Village, W. 7.
Tsang 24240 (TYPE), Sept. 8, 1934, a fairly common shrub in thickets, about 1 m. high,
flowers white, fragrant

This species is allied to Jxora hainanensis Merr. & Chun of Hainan, differ-
ing in the leaves having more numerous, prominent, and more oblique
lateral nerves, and in the much shorter flowers with subrounded corolla-
lobes. Moreover, the inflorescences are sessile, while in J. Aainanensis they
are distinctly pedinediate. H.Y. Lia z 70086, from Shih Wan Tai Shan,
Kwangtung Province, unfortunately with immature flowers, may represent
the same species.

Psychotria Linnaeus
Psychotria kwangsiensis sp. nov.

Frutex parvus erectus ubique glaber, ramis ultimis crassis, 4 mm. dia-
metro; foliis chartaceis Sages longe petiolatis ellipticis, circa 26 cm.
longis et 12 cm. latis, apice acutis vel acuminatis, basi longe attenuatis,
supra olivaceo-viridibus, subtus paullo pallidioribus, nervis lateralibus
utrinsecus circa 16, utrin ue conspicuis, prope marginem tenuiter curvato-
arcuatis, venulis utrinque obscuris; petiolis 4-5 cm. longis; stipulis de-
ciduis; floribus ignotis ; infructescentiis terminalibus oe sublaxe
cymoso- paniculatis, circa 7.5 cm. longis, pedunculis circa 4.5 cm. longis,
bracteis lanceolatis acuminatis, circa 2.5 mm. longis, m margine (enter cili-

pedicellatis, 1 cm. longis, 4-5 mm. crassis, longitudinaliter sulcatis, glabris,

calycis lobis lanceolatis circa 1 mm. longis plus minusve persistentibus
coronatis; seminibus circa 1 cm. longis, plano-convexis, dorso leviter 4-
costato, albumine haud ruminato.

KWANGSI: en Pien District, S. os Ko 56011 (type), Nov. 5, 1935, on slopes in
forested ravines, probably small in siz

A species strongly pes by its large leaves and its characteristic
fruits

Lasianthus Jack
Lasianthus Tsangii Merrill in herb. sp. nov.

Frutex erectus circa 1 m, altus, ramulis inflorescentiis et foliis subtus
villosis, indumento interdum sordide brunneo vel in venulis pallido, ramulis
ultimis teretibus 3 mm. diametro; foliis subcoriaceis petiolatis oblongo-
lanceolatis, 9-12.5 cm. longis, 3— 4.2 cm. latis, acuminatis, basi acutis,
supra glabris viridibus subnitidis, subtus paullo pallidioribus praesertim

458 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

secus costam nervosque molliter villosis, nervis lateralibus utrinsecus 5—7,
supra leviter impressis, subtus elevatis, perspicuis, curvato-adscendentibus,
venis tertiariis parallelis, supra indistinctis, subtus elevatis; petiolo 8-10
mm. longo dense subadpresse hirsuto; inflorescentiis axillaribus sessilibus
paucifloris, bracteis lanceolatis, dense villosis, circa 1 cm, longis; floribus
sessilibus; calycis tubo circa 4 mm. longo piloso, lobis 5, anguste lanceo-
latis acuminatis crasse Ciliatis circa 3 mm. longis, persistentibus; fructibus
subovoideis glabris circa 4 mm. diametro.

Kwanost: Shang-sze District, Shih Wan Tai Shan, Na Wai Village, W. T. Tsang
23940 (type), July 11-30, 1934, a shrub about 1 m. high, fairly common in thickets,
fruits blackish blue.

This species is characterized by its prominently nerved leaves, which
are glabrous above and prominently villose beneath, with sharply ascend-
ing lateral nerves, its axillary, few-flowered inflorescences with elongated,
persistent, lanceolate, and densely villose bracts, and its persistent pubescent
calyx-lobes. It resembles L. Koi Merr. & Chun in vegetative characters, but
the leaves of the latter are narrowly caudate-acuminate, while the inflo-
rescences are more crowded, more numerously flowered, and with numerous
narrow and villous bracts.

Lasianthus kwangsiensis Merrill in herb. sp. nov.

tibus arcuato-anastomosantibus plerumque dense ciliatis, venis tertiariis

oblongo-ovatis circa 1.5 mm. longis; staminibus tubo insertis, antheris
subexsertis; stylis tubo corollae subaequilongis, stigmatibus ciliatis S-
lobatis.

Kwanosr: Shang-sze District, Shih Wan Tai Shan, Nam She Village, W. T. Tsang
24679 (tvpE), Nov. 18, 1934, a shrub 3 ft. high, in thickets, fairly common, flowers
white, fragrant.

This species is near Lasianthus Fordii Hance, but it may be distinguished,
among other characters, by its corolla being prominently hirsute.

Paederia Linnaeus
Paederia pertomentosa Merrill in herb. sp. nov.

Suffruticosa vel herbacea scandens, circa 3.5 m. alta, caulibus et ramis
et foliis subtus dense breviter tomentosis, caulibus teretibus 2 mm. diametro,
indumento sordide substramineo, ramulis circa 1 mm. diametro; foliis char-
taceis, ovato-ellipticis vel oblongo-ellipticis, acute acuminatis, basi plerum-
que rotundatis, haud cordatis, interdum leviter decurrentibus, 6-11
2.5—5 cm., supra subolivaceis consperse puberulis ad costam dense puberulis,

1943 ] LI, NOTES ON THE FLORA OF KWANGSI 459

subtus densissime breviter subalbido- tomentosis, nervis primariis utrinsecus

subglomeratim dispositis, glomerulis inter se 1-3 cm. distantibus at
minusve confertis breviter pedicellatis; floribus rosaceis; calycibus dense
tomentosis, dentibus triangularibus, acutis, intus glabris; corollae tubo
extus dense puberulo, 5 mm. longo, lobis ovatis acutis, 1-1.2 mm. longis.

SOUTHERN Kiancsi: Hong San, J. L. Gressitt 1471 (tyPE), June 23, 1936, in thick-
ets, alt. 840 m.; Lung-nan District, Oo Chi Shan, near Lam Uk Village, S. K. Lau 2
Oct. 1-25, 1934, a semi-woody climber, rare on dry steep slopes in forests. KWAN
Nam Tan-yuen, C. Wang 40918, June 26, 1937, scandent, in forests; Ling-chuan District,
Hai-yan-shan, Lian-chai-miao, W. T. T ane 27825, July 13-19, 1937, climber, 3 ft.
high, fairly common in thickets on steep slopes.

This is one of the very tomentose forms formerly referred to Paederia
tomentosa Bl. = P. scandens (Lour.) Merr. and to P. foetida Linn., but
manifestly it is so different from these that there seems to be little justi-
fication in extending the limits of either species to take this extreme form.
It is allied to P. Cavaleriet H. Léev., differing among other characters in the
leaves being densely short-white-tomentose beneath.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

460 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

STUDIES OF SOUTH AMERICAN PLANTS, X
NOTEWORTHY MYRISTICACEAE AND VACCINIACEAE

A. C. SMITH

Turoucu the kindness of the authorities of the U.S. National Herbarium
and the Instituto Botanico of Bogota, a series of Colombian specimens rep-
resenting the Myristicaceae and Vacciniaceae has been made available to
me for study. Most of these specimens were collected in the Comisaria of
Putumayo by J. Cuatrecasas and in the Department of Antioquia by
R. D. Metcalf and J. Cuatrecasas. One new species is based upon a collec-
tion from Antioquia by Brother Daniel, while an Ecuadorean plant collected
by A. Rimbach is also described as new. Eleven new species are here
described; of interest is the discovery in Colombia for the first time of the
genera Semiramisia and Ceratostema (sensu vero). In the following pages
the place of deposit of specimens is indicated by parenthetical letters as
follows: (A), Arnold Arboretum; (Col), Instituto Botanico, Universidad
de Colombia, Bogota; (US), U. S. National Herbarium.

MYRISTICACEAE

ee parvifolia Markegraf in Notizbl. Bot. Gart. Berlin 9: 964. 1926;
A.C. in Brittonia 2: 417. 1938, 3: 339. 1939
easels Putumayo: Selva wot del rio San Miguel, en el afluente

izquierda Quebrada de la Hormiga, alt. , Cuatrecasas 11139 (Col, US); Mocoa,
bosque higrofilo en la Quebrada del rio Malo, alt. 570-600 m., Cuatrecasas 11314
(A, Col, US). Borrvia: La Paz: 7. Larecaja, Tuiri, near Mapiri, alt. 490-

750 m., Krukoff 10734 (A, etc.).

In my monograph of 1938 I listed this species only from Peru and adja-
cent Brazil; the above-cited collections from the Amazonian portions of
Colombia and Bolivia demonstrate that the species has a fairly wide range.
In 1939 I reported the plant from Venezuela, and another collection from
that country (Wdaliams 10210 [A], from National Park, alt. 850-1100 m.)
is now available. Dialyanthera parvifolia is therefore now known to have
a fairly extensive distribution along the western edge of the Amazon basin
from Bolivia to Colombia and continuing northward to Aragua in Venezuela.
—— ae higese (A. DC.) phone in Aled Acta Acad. Leop.-Carol. 68: 188. 1897;

.in Brittonia 2: 472.

Panes Putumayo: ee higréfila del rio Putumayo; Puerto Porvenir,
arriba de Puerto Ospina, hacia la Loma, alt. 230-250 m., Cuatrecasas 10746 (Col, US).

The cited collection extends to Colombia the range of V. peruviana, pre-
viously recorded from Amazonian Peru and Brazil, but only from the
southern tributaries of the Amazon. The Cuatrecasas specimen has leaf-
blades up to 40 by 12.5 cm. (larger than those previously known), and its
tomentum is perhaps somewhat more persistent throughout; nevertheless
it agrees with earlier collections of the species in all essential details.

1943 | SMITH, SOUTH AMERICAN PLANTS, X 461

haa lear pels in os Acta Acad. Leop.-Carol. 68: 231. 1897; A. C. Sm.
ittonia 2: 474.
MBIA: Putumayo: Selva higrofila del rio Putumayo; Puerto Porvenir,
arriba de Puerto Oipine, hacia la Loma, alt. 230-250 m., Cuatrecasas 10653 (Col, US).
The cited specimen is of especial interest as being the first Colombian
collection which can be referred to V. calophylla without doubt. Other-
wise the species has a range in Amazonian Brazil, Venezuela, and Peru.
In 1938 I referred to V. calophylla the type of V. incolor Warb., a sterile
juvenile specimen from Villavicencio, and thus the species has already been
reported from Colombia, but only inadequately so. The Cuatrecasas collec-
tion, being in fruit, permits a correction of my earlier description. The
pubescence of the fruit is much coarser than previously indicated, the indi-
vidual hairs being somewhat more than 1 mm. in length, copiously jointed,
and with numerous minute lateral branches, somewhat like the pubescence
illustrated for V. loretensis A. C. Sm. (Brittonia 2: f. 7, g. 1938). Appar-
ently these hairs are at length deciduous, leaving the fruit closely tomen-
tellous.
VACCINIACEAE

Semiramisia pulcherrima sp. nov

Frutex ad 50 cm. altus dinate filamentis exceptis glaber, ramulis teretibus
gracilibus ut videtur elongatis; petiolis subteretibus rugulosis 3-4 mm.
longis; laminis subcoriaceis ovatis, 4.5-6 cm. longis, 2.5—3 cm. latis, basi
rotundatis vel late obtusis, apice radatim. acuminatis, margine leviter

folii versus costa conjunctis, costa et nervis secundariis supra leviter im-
pressis subtus elevatis, nervis marginalibus indistinctis et rete venularum
immersis vel subtus paullo prominulis; inflorescentiis apicem ramulorum
versus axillaribus racemosis 5—8-floris, basi bracteis paucis ovatis acutis
circiter 1.5 mm. longis subtentis, rhachi subtereti 1-1.5 mm. diametro 1—2
cm. longa; floribus alternatis basi bracteatis, bracteis papyraceis elongato-
deltoideis 1.5-2 mm. longis acutis parce glanduloso- -marginatis; pedicellis
teretibus 8-15 mm, longis apicem versus bibracteolatis (bracteolis ut brac-
teis sed minoribus), basi gracilibus, distaliter in calycem gradatim incrassa-

diametro, limbo papyraceo erecto-patente lobis inclusis 5-6 mm. longo
inconspicue nervato, lobis inconspicuis in cuspidem 0.5—0.8 mm. longam
apiculatis margine biglandulosis, glandulis linearibus 0.5—-2 mm. longis;
corolla carnosa campanulato-cylindrica sub anthesi 40-50 mm. longa, basim
versus 10-15 mm. ut videtur apice ad 20 mm. diametro, lobis deltoideis
acutis 6-10 mm. longis et latis; staminibus corollam fere subaequantibus;
filamentis liberis 6-8 mm. longis, 1—1.5 mm. latis, ubique praeter basim pilis
circiter 0.2 mm. longis puberulis; thecis crassis circiter n mm. longis basi
leviter incurvatis; tubulis gracillimis (basi circiter 0.4 mm. apicem versus
circiter 0.25 m m. diametro) 25-30 mm. longis, poris eminalibie vel leviter
obliquis Aeron stylo corollam subaequante gracili truncato

CotomsiA: Putumayo: Alta cuenca del rio Putumayo, filo de la Cordillera
entre El] Encano y Sibundoy; paramo de San Antonio del Bordoncillo, alt. 3250 m.,
4 Enero 1941, Cuatrecasas 11771 (Col, US no. 1798516, Type), fruitex de 0.50 m.;
corola crasa rosado-anaranjada.

In its distinct and pilose filaments, S. pulcherrima resembles S. Kar-

462 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

steniana K\., from which it differs in its smaller and fewer-nerved leaf-
blades, shorter pedicels, and much larger flowers, especially the larger calyx
and broader corolla. From the remaining species of Semiramisia, the new
species differs in its separate and pilose filaments, as well as in details of
foliage, its more complex inflorescence, and the proportions of its calyx
and corolla, etc. Semiramisia pulcherrima is the first recorded specimen of
the genus from Colombia, it having previously been known from the Andes
of Venezuela, Ecuador, and Peru.

Ceratostema amplexicaule sp. nov.

Frutex, ramulis gracilibus teretibus molliter et breviter albo-pilosis;
foliis sessilibus vel minutissime siactering laminis subcoriaceis ovato-
oblongis, 5—7.5 cm. longis, 3—4.5 cm. latis, basi profunde cordatis et am-
plexicaulibus, apice obtusis, eae leviter recurvatis, utrinque molliter
pilosis vel puberulis ut videtur demum glabrescentibus, nervis secundariis
utrinsecus circiter 4 basim versus orientibus cum costa supra subimmersis

bracteis bracteolisque, pedicellis, calyce et corolla extus) pilis 0.5—0.8 mm
longis molliter albo-pilosis; rhachi leviter angulata 2.5—4.5 cm, longa 7— ‘7,
flora; bracteis sub pedicellis papyraceis deltoideis 2-3 mm. longis et latis
acutis intus glabris; pedicellis subteretibus 10-17 mm. longis, cum calyce
obscure articulatis, paullo infra medium vel basim versus bibracteolatis,
bracteolis suboppositis ut bracteis sed minoribus et obscure pauciglandu-
loso- ee calycis tubo cupuliformi obscure 10-costato sub anthesi
circiter 3 mm. longo et 5 mm. diametro, limbo subpatente lobis inclusis
circiter 3 mm. longo intus glabro, lobis 5 deltoideis acutis, circiter 2 mm.
longis et 4 mm. latis, secus margines praeter apicem versus glandulis
linearibus praeditis, sinibus acutis; corolla crasse carnosa urceolato-cylin-
drica vel juventute leviter angulata, sub anthesi 25-30 mm. longa et basim
versus 6—7 mm. diametro, distaliter gradatim angustata, intus glabra,
demum profunde 5-lobata, lobis subulatis ad 15 mm. longis et basi 3 m
latis, apice subacutis; staminibus corollam subaequantibus_ vel pane
exsertis, filamentis membranaceis glabris mox liberis ligulatis 6-8 mm.
longis, thecis valde granulosis 9-10 mm, longis 1-1.5 mm. crassis, tubulis
gracillimis circiter 0.2 mm. diametro 13-14 mm. longis basim versus saepe
tuberculato-granulosis, poris ovalibus oblique terminalibus circiter 0.7 mm
longis dehiscentibus; stylo corollam subaequante gracili (circiter 0.5 mm.
diametro) truncato.

CotomBiA: Putumayo: Vertiente oes de la Cordillera, bosques higrofilos
entre Mocoa y Sac heraies, alt. 600-700 m., 29 Diciembre 1940, Cuatrecasas 11407
(Col, US no. 7798807, TYPE), frutex; corola oe -anaranjada.

In its subsessile leaf-blades, which are deeply cordate and amplexicaul
at base, the new species resembles only C. peruvianum Gmel., the remaining
species of the genus having leaf-blades cuneate to attenuate at base. In
their pubescence, C. amplexicaule and C. peruvianum are essentially similar.
However, the new species differs sharply from C, peruvianum in its very
small calyx-lobes, those of Gmelin’s species being 10-15 mm. long and
6-9 mm. broad. In other respects, also, the flowers of C. peruvianum are

1943 | SMITH, SOUTH AMERICAN PLANTS, X 463

substantially larger than those of the new species, the corolla and stamens
being about 40 mm. long. Furthermore, the leaf-blades of the new species
are more nearly sessile and more obtuse at apex than those of C. peruvt-
anum,

The genus Ceratostema Juss. (sensu Sleumer in Notizbl. Bot. Gart.
Berlin 12: 278-282. 1935, et A. C. Sm. in Bull. Torrey Bot. Cl. 63: 307-
308. 1936; non sensu falso A. C. Sm. in Contr. U. S. Nat. Herb. 28: 335-
348. 1932) has previously been known only from Ecuador (this being
probably true even for C. peruvianum, the genotype). It should he noted
that the present species falls into the genus Englerodoxa Hoer. as treated
by me in 1932 (op. cit. 350-352), a synonym of Ceratostema Juss., which
is now known to contain six species. The numerous other species which
have been referred to Ceratostema belong to Pellegrinia, Demosthenesia,
Plutarchia, etc. (see Sleumer, loc. cit. 1935, and A. C. Sm., loc. cit. 1936).
Psammisia flaviflora sp. nov.

Frutex scandens forsan epiphyticus ubique praeter ain giaber,
ramulis subtereiibus gracilibus (apicem versus 2-3 mm. diametro) ; petiolis
crassis (2-3 mm. diametro) semiteretibus 6-12 mm. longis; laminis charta-
ceis oblongo-ellipticis, 13-20 cm. longis, 4.5-8 cm. latis, basi acutis et in
petiolum decurrentibus, apice abrupte caudato- acuminatis (acumine gracili
1—1.5 cm. longo acuto), margine leviter recurvatis, 5- vel 7-pli-nerviis, ner-

S

nervisque supra paullo elevatis et saepe insculptis subtus prominentibus,
rete venularum copioso utrinque prominulo; inflorescentiis ut videtur api-
cem ramulorum versus axillaribus subfasciculatis vel obscure breviracemosis,
thachi ad 4 mm. longa Seah breviore; floribus paucis bracteis papyra-
ceis deltoideis circiter 1.5 & 1.5 mm. subacutis pauciglanduloso- -marginatis
subtentis; pedicellis rugulosis crassis (1-2 mm. diametro) subcurvatis
15-20 mm. longis paullo supra basim eee erate oppositis
bracteis similibus; calyce coriaceo sub anthesi 8-10 mm. longo, tubo cupuli-
formi 4—6 mm. longo et circiter 7 mm. secs fen erecto- -patente lobis
inclusis 3-4 mm. longo, lobis deltoideis acutis 2-3 mm. longis et 3-4 mm.
latis, sinibus acutis; corolla carnosa subgloboso-urceolata, sub anthesi 7—8

longa et circiter 6 mm. diametro, basim et apicem versus contracta,
lobis 5 deltoideis circiter 1.5 1.5 mm. subacutis; staminibus 10, fila-
mentis liberis submembranaceis ligulatis circiter 3. 5 mm. longis margine
superne ciliolato-puberulis antheras apice thecarum affixis, connectivis
brevibus omnibus bicalcaratis (calcaribus subacutis conspicuis saepe
antrorsis), antheris circiter 4 mm. longis, thecis valde granulosis circiter

mm. longis et 1.3-1.5 mm. crassis, tubulis distinctis gracillimis brevibus
(circiter 1 mm. longis) acutis per rimas ovales dehiscentibus; stylo crasso
corollam subaequante truncato.

Ecuapor: Guayas(?): Western Cordillera, valley of Rio Chimbo, alt. 800 m.,
Rimbach 67 (A, TYPE), shrub, climbing high among forest trees; ts ot ae
tube bright red, the calyx-lobes greenish, the corolla and lobes yel

The closest relative of this very distinct new species is the recently de-
scribed P. occidentalis A. C. Sm. (in Am. Jour. Bot. 27: 542. 1940), from
western Colombia. However, P. flaviflora has leaf-blades with the sec-
ondary nerves oriented nearer the base and more prominent on the upper

464 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

surface, while its inflorescence has fewer flowers and a shorter rachis. The
flowers of the new species have longer pedicels and are substantially larger
throughout, the calyx-lobes being entirely different in shape. Other small-
flowered species of Psammisia with subfasciculate inflorescences — P.
Pennellii A. C. Sm. and P. caudatula Sleumer — have quite distinct foliage
as well as minor differences in floral proportions.

Psammisia ferruginea A. C. Sm. in Contr. U. S. Nat. Herb. 28: 391. pl. 10. 1932.
CorompBiA: Putumayo: Vertiente oriental de la Cordillera, entre Sachamates
San Francisco de Sibundoy, Quebrada de Susunga, alt. 1600-1800 m., Cuatrecasas

11447 (Col, US), arbusto de ramas scandens; pedunculos y caliz ‘oaadniehpdenos: corola

rosada.

This very distinct species has previously been known only from the De-
partments of El Cauca and Narifo, and therefore the new record is of
unusual interest; the type is a Lobb specimen of uncertain origin, but
probably from the Pacific slope of Colombia.

Psammisia columbiensis Hoer. in Bot. Jahrb. 42: 303. 1909; A. C. Sm. in Contr.

U.S. Nat. Herb. 28: 398. 1932.

CoromsBiA: Putumayo: Vertiente oriental de la Cordillera, entre Sachamates

San Francisco de Sibundoy, Quebrada de Susunga, alt. 1600-1800 m., Cwuatrecasas
11446 (A, Col), arbustito de ramas péndulas; pedunculos, calices y corolas rojos, apice
blanco.

This species has previously been recorded only from the type, collected
by Lehmann in the “Central Andes of Popayan,” Dept. El Cauca, alt.
2500-2800 m. The present collection agrees with the type in all essential
characters, including the connate filaments, but has its leaf-blades slightly
broader, 2.5—4.8 cm. broad.

Psammisia Cuatrecasasii sp. nov
Frutex ramosus ubique abi filamenta bracteas bracteolasque glaber,

ramulis crassis (apicem versus circiter 8 mm. diametro) subteretibus;

petiolis crassis (4-5 mm. diametro) circiter 2 cm. longis; laminis coriaceis
siccitate olivaceis late ovato-oblongis, 22-27 cm. longis, 12—14 cm. latis,
basi rotundatis vel obtusis et in petiolum paullo decurrentibus, apice ut
videtur obtuse cuspidatis, margine anguste recurvatis, 7—9-pli-nerviis, nervis
infimis e basi orientibus summis costa ad 5 cm. concurrentibus, costa
ale supra subplanis subtus valde prominentibus, rete venularum sub-
immerso subtus leviter prominulo; inflorescentiis axillaribus racemosis
6—9-floris, rhachi crassa subtereti circiter 4 cm. longa, pedicellis sub anthesi

25-30 mm. longis, bracteis subcoriaceis oblongis circiter 6 mm. longis et

3 mm. latis apice rotundatis margine obscure ciliolato-puberulis subtentis,

paullo supra medium bibracteolatis, bracteolis suboppositis papyraceis

deltoideo-oblongis circiter 3 2 mm. acutis margine ciliolato-puberulis et

pauciglandulosis; calyce coriaceo late cupuliformi, tubo circiter 2 mm

iametro, limbo patente lobis inclusis 3-4 mm. longo, lobis

praeter apicem versus glanduloso-incrassatis; corolla carnosa cylindrico-
urceolata sub anthesi 21-27 mm. longa, basim versus circiter 7 mm. dia-
metro, supra ad 4 mm. angustata, lobis 5 incrassatis deltoideis subacutis
circiter 3 3 mm.; staminibus 10, filamentis liberis subcarnosis ligulatis
6—7 mm. longis margine superne puberulis, connectivis superne alternatim

1943 ] SMITH, SOUTH AMERICAN PLANTS, X 465

conspicue bicalcaratis Sooty 0.7-1 mm. longis acutis antrorsis) , thecis
valde granulosis 8—9 mm. longis basi i inconspicue tuberculatis, tubulis leviter
cohaerentibus circiter "4 mm, longis per rimas ovales =; mm. spas
dehiscentibus; stylo filiformi sub anthesi conspicue exserto 25-30 m
longo truncato.

CoLoMBIA: Putumayo: Alta cuenca del rio Putumayo en el Valle de Sibundoy
extremo E., junto a San Francisco, alt. 2200 m., 1 Enero 1941, oe ante 11562
(Col, Type), arbusto ramoso; pédunculo, caliz y corola carmin, extremo

Psammisia Cuatrecasasii is at once distinguished by its large ee pro-
portionately broad 7—9-pli-nerved leaf-blades. In foliage it most suggests

Ulbrichiana Hoer., but it has even broader leaves, larger bracts, and
larger and much more conspicuously spurred stamens. Perhaps a closer
relative of the new species is P. falcata (H. B. K.) KI., a species with leaf-
blades only 3.5—7 cm. broad and 5-—7-pli-nerved, usually longer inflo-
rescences, and smaller bracts.

Psammisia aestuans sp. nov.
anes (?) ubique praeter filamenta glaber, ramulis gracilibus (apicem
s | mm. vel minus diametro ) subteretibus; petiolis gracilibus leviter
carelicula fe 2—5 mm. longis; laminis subcoriaceis in sicco olivaceis vel
saepe metallicis ovatis, 3-5 cm. longis, 1.2-2.2 cm. latis, basi rotundatis vel
late obtusis, apice acumine gracili attenuato 1-2 cm. lon ngo conspicue prae-
ditis, margine anguste recurvatis, 3- vel 5-pli-nerviis (nervis exteriori-

S

basim versus 5-7 mm. diametro ae Se eee. lobis 5
elongato-deltoideis acutis circiter 2 & 1.5 mm.; staminibus 10 corollam
fere aequantibus, filamentis submembranaceis liberis ligulatis circiter 2 mm.
longis et 1—-1.5 mm. latis margine superne minute ciliolatis, connectivis
angustis omnibus superne bicalcaratis (calcaribus acutis vel subacutis

granulosis 5.5-6 mm. longis crassis (circiter 1.5 mm. diametro ) basi in-
curvatis, tubulis liberis gracilibus 7—7.5 mm. longis per rimas ovales 0.7—1

mm. longas dehiscentibus; stylo filiformi corollam subaequante truncato;
fructibus immaturis coriaceis subglobosis 7-8 mm. diametro, calycis limbo
persistente coronatis.

Coromsia: Antioquia: a Ceja, Diciembre 1939, Hermano Daniel 2178 (US
no. 1778630, TYPE), corola en el o roja, en el apice verde.

Psammisia aestuans is so distinct from the other species of Psammisia
that one places it here with hesitation; yet, on the basis of floral characters,
there can be no doubt of its place in ffs genus. The small, long-acuminate,

466 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

few-nerved leaves suggest those of no other species of Psammisia, but the
flowers are typical for the genus, although the inconspicuous calyx-lobes
and the long anther-tubules are unusual.

Plutarchia angulata sp. nov.

rutex, ramulis subteretibus apicem versus pilis 0.4-0.7 mm. longis
dense cinereo-pilosis demum glabrescentibus; petiolis 1-3 mm. longis in-
crassatis (1.5-2 mm. diametro) ut ramulis pilosis; laminis subcoriaceis in
5 cm. longis, 2.2-4 cm. latis, basi
leviter cordatis, apice obtusis vel subrotundatis, margine anguste recurvatis,
supra minute puberulis mox glabris, subtus praecipue costa ut petiolis
hispidulo-pilosis demum glabrescentibus, 7- vel 9-pli-nerviis, costa supra
leviter impressa subtus elevata, nervis secundariis basim versus orientibus
curvatis supra subimmersis subtus prominulis, venulis obscuris vel subtus
paullo prominulis; inflorescentiis apicem ramulorum versus axillaribus vel
subterminalibus 2—5-floris breviter racemosis, basi bracteis paucis papyra-
ceis ovatis acutis pilosis 2-3 mm. longis circumdatis, rhachi subtereti rugu-
losa 5—25 mm. longa ut ramulis dense pilosa vel puberula; floribus bracteis
papyraceis anguste oblongis acuminatis 3-4 mm. longis extus puberulis
subtentis; pedicellis subteretibus dense pilosis 8-20 mm. longis basim
versus bibracteolatis, bracteolis suboppositis bracteis similibus sed minori-
bus mox caducis; calyce cum pedicello articulato 12-15 mm. longo extus
pilis albidis mollibus circiter 0.3 mm. longis dense piloso demum sub-
glabrescente, tubo 4—6 mm. longo basi rotundato alis 4 conspicuis 2-3 mm.
latis sinibus oppositis ornato, limbo 8-9 mm. longo erecto intus glabro fere
ad basim 4-lobato, lobis subcoriaceis elongato-deltoideis basi 5-8 mm
latis forsan accrescentibus ad apicem acutum gradatim angustatis, sinibus
acutis: corolla tenuiter carnosa tubulosa 4-angulata, sub anthesi 17-21 mm.
longa et 5-6 mm. diametro, extus pilis 0.5—1 mm. longis densissime villoso-
velutina, intus glabra, lobis 4 oblongo-deltoideis subacutis circiter 2 >< 2
mm.: staminibus 8 similibus quam corolla paullo brevioribus, filamentis
liberis carnosis glabris ligulatis 2-3 mm. longis, antheris 12-17 mm. iongis,
thecis minute granulosis 4-7 mm. longis et circiter 1 mm. diametro, tubulis
latis flexilibus plerumque quam thecis circiter duplo longioribus per rimas
elongatas dehiscentibus; disco inconspicuo glabro; stylo filiformi corollam
subaequante, stigmate truncato vel minute subpeltato.

CoromsBiA: Putumayo: Alta cuenca del rio Putumayo, filo de la Cordillera
entre El Encano y Sibundoy; paramo de San Antonio del Bordoncillo, alt. 3250 m.,
4 Enero 1941, Cuatrecasas 11733 (Col, US no. 1798515, TYPE), frutex; caliz y corola
rosados.

Although P. angulata does not bear a close superficial resemblance to the
other species of Plutarchia, its staminal characters indicate that it can
belong to no other genus. It differs from the other species of the genus in
its 4-merous flowers, its winged calyx-tube, and its angled densely pubescent
corolla, while its proportionately broad cordate subsessile leaf-blades, which
are soft-pilose beneath, are also characteristic. Plutarchia pubiflora
(Wedd.) A. C. Sm. is perhaps the closest ally of P. angulata, but there are
numerous obvious differences between the two species.

Cavendishia Cuatrecasasii A. C. Sm. in Rev. Acad. Colomb. Ci. Ex. Fis.-Quim. Nat.
5: 38

1943 ] SMITH, SOUTH AMERICAN PLANTS, X 467

Since the publication of this well-marked species, two additional collec-
tions have come to my attention:

CoLomsBiA: Putumayo: Vertiente Ria easy de la Cordillera, entre Sachamates

y San Francisco de Siandey, alt. 1600-1750 m., Cuatrecasas 11464 (A, Col, US),
hee de ramas divaricadas; bracteas y flores rosado cardenas; Huila-
Caqueta: Cordillera Oriental sobre el filo divisorio, en Gabinete, alt. 2300-2450 m.,
Cuatrecasas 8479 (Col, US), gran frutex; bracteas rosadas; caliz muy largo, rosado;
corola, base y apice blanco rosado, tercio superior roja.

These specimens make desirable a slight amplification of the original
description, as follows: petioles up to 15 mm. long; leaf-blades up to 18 cm.
long and 9 cm. broad; corolla often only 22 mm. long, the stamens pro-
See shorter than those originally described. These slight varia-
tions are no more than individual.

Cavendishia speciosa sp. nov.

Frutex ad 5 m. altus ubique praeter filamenta glaber, ramulis teretibus
gracilibus (apicem versus 2—3 mm. diametro); petiolis subteretibus rugosis
incrassatis (2-3 mm. diametro) 5-8 mm. lene laminis coriaceis in sicco
fuscis anguste oblongis, 8-14 cm. longis, 2—3.6 cm. latis, basi obtusis, apice
caudato-acuminatis (acumine gracili 10-15 mm. lon 20), margine anguste
et basim versus conspicue revolutis, 3-pli-nerviis, ie paullo supra basim
orientibus adscendentibus costa apicem versus inconspicue conjunctis, costa
nervisque supra impressis subtus valde elevatis, nervis marginalibus inter-
dum visis sed immersis, rete venularum immerso; inflorescentia axillari vel
subterminali multiflora, rhachi crassa (4-6 mm. diametro) basi florum
30-40 incrassata; bracteis sub floribus membranaceis copiose reticulato-
venosis obovato-oblongis, 30-40 mm. longis, 18-30 mm. latis, apice rotun-
datis vel conspicue bilobatis, dorso glandulas minutas subglobosas sessiles
saepe copiose gerentibus; pedicellis teretibus crassis 6-7 mm. longis ut
videtur ebracteolatis (bracteo lis non visis forsan mox caducis); calyce
11-12 mm. longo, tubo breviter cylindrico 3—4 mm. longo et circiter 4.5 mm.
diametro, limbo submembranaceo erecto lobis inclusis 7-8 mm. longo, lobis
deltoideis acutis circiter 2 mm. longis et 3.5 mm. latis glandulas plures albas
sessiles margine gerentibus, sinibus rotundatis; corolla tenuiter carnosa
cylindrica, sub anthesi circiter 30 mm. longa et basim versus
dia metro, See angustata, lobis 5 oblongis subacutis circiter 2
1.5 mm.; staminibus sae corolla multo tee filamentis ligulatis
aes yete 4 mm. et 7-8 mm. longis superne angustatis intus et
margine distaliter Sigg, antheris alternatim circiter 11 mm. et 10 mm.
longis, thecis 4-5 mm. longis, tubulis quam thecis paullo longioribus per
rimas elongatas ovales dehiscentibus; stylo filiformi corollam subaequante,
stigmate minute peltato

CotomsBiA: Antioquia: Between Valdivia and Yarumal, alt. 2000 m., moist
canyon, Feb. 20, 1942, Metcalf & Cuatrecasas 30101 (A, TYPE, US), shrub 4-5 m. high;
bracts vivid red, with brown glands; calyx white; corolla white and rose.

This beautiful and spectacular species is one of the most distinct in
Cavendishia, being characterized by its thick 3-nerved leaf-blades, its
membranaceous reticulate-veined bracts, its elongate glandular-margined
calyx-limb, and its large showy corollas. Its closest relative is doubtless
C. Kalbreyeri Mansf., also from Antioquia, from which it differs in its
more elongate inflorescence with more numerous flowers, its larger and

468 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

thinner bracts, its calyx with glandular rather than thick-margined lobes,
its larger corollas, and its shorter and differently proportioned anthers, those
of C. Kalbreyeri having tubules much longer than the thecae.

Cavendishia rosea sp. nov.

Frutex ad 5 m. altus ubique praeter filamenta glaber, ramulis eo
apicem versus 3-4 mm. diametro; petiolis inconspicuis incrassatis 2—5 m
longis; oer eens siccitate olivaceis ovato-oblongis, 6-9 cm. longis,
2.5—4.3 cm. latis, basi conspicue cordatis, apice obtuse cuspidatis, margine
leviter eae 5—7-pli-nerviis, nervis infimis patentibus debilibus, nervis
superioribus costa ad 2 cm. concurrentibus vel e basi orientibus, costa
nervisque supra prominulis vel subplanis subtus prominentibus, rete venu-
larum utrinque paullo prominulo, venulis brevibus e costa saepe patentibus;
inflorescentia apicem ramulorum versus axillari racemosa 20—30-flora,
rhachi angulata 1.5-2.5 mm. diametro 8-9 cm. longa basi pedicellorum
incrassata; floribus bracteis papyraceis ius 25- 32 mm. longis et 8-12

mm. latis acutis conspicue punctatis subtentis; pedicellis teretibus 10-14
mm. longis basim_ versus igen steamer bracteolis papyraceis
elliptico-oblongis, 4-6.5 mm. longis, 2-3 mm. latis, apice rotundatis et
minute apiculatis; calyce seers ‘cylindrico leviter 5-angulato sub anthesi
circiter 5 mm. longo, tubo 4-5 mm. diametro, limbo erecto papyraceo lobis
inclusis 2-3 mm. ees lobis saepe incurvatis late deltoideis subacutis

angustata, lobis deltoideis subacutis circiter 0.7 1.5 mm.; staminibus
corollam fere aequantibus alternatim leviter inaequalibus, filamentis =
membranaceis ligulatis intus parce puberulis alternatim 2-2.5 mm.
5-6 mm. longis, antheris alternatim 11-12 mm. et 10-11 mm. ne
thecis 3-4 mm. longis, tubulis quam thecis plus minusve duplo longioribus
per rimas elongatas dehiscentibus; stylo filiformi corollam subaequante,
—_ minute subpeltato

Cou Antioquia: Between Valdivia and Yarumal, alt. 2200 m., shaded
hillside, Feb. 20, 1942, Metcalf & Cuatrecasas 30123 (A, type, US), shrub 4-5 m. high ;
stems red-streaked; leaves green with red margins and reddish tinge above; bracts red;
corolla rose-red ; maturing calyx greenish yellow.

The new species is related only to C. subamplexicaulis A. C. Sm., also
from northwestern Colombia, with which it has in common an elongate
inflorescence and cordate-based leaf-blades. However, C. rosea differs
from its ally in its proportionately narrower leaf-blades, which are less
obviously clasping at base, have fewer and more prominent secondary
nerves, and are cuspidate rather than obtuse at apex. The rachis of the
new species is more slender, and the bracts and corollas are slightly larger;
the anther-proportions are different from those of C. subamplexicaulis, in
which the thecae and tubules are nearly equal in length. These two closely
related species appear to have no other immediate allies.

Cavendishia Dugandiana sp. nov.

d 6 m. altus, ramulis fusco-cinereis apicem versus gracilibus et
cinereo- -puberulis; petiolis subteretibus puberulis 2-4 mm. longis; laminis
parvis coriaceis in sicco olivaceis anguste oblongis, 2—4 cm. longis, 0.8-1.4
cm. latis, basi rotundatis vel obscure subcordatis, apice obtusis et saepe

1943 ] SMITH, SOUTH AMERICAN PLANTS, X 469

minute mucronulatis, margine subplanis, supra inconspicue scabridis et
parce brunneo- glandulosis, subtus pilos minutos brunneos dispersos gerenti-
bus et interdum secus nervos basim versus puberulis, 3- vel obscure 5- -pli-
nerviis, nervis secundariis prope basim orientibus et costa supra leviter im-
pressis subtus elevatis, rete venularum immerso; inflorescentia apicem
ramulorum versus axillari breviter racemosa 2—4-flora , basi bracteis pluri-

oribus maximis oblongis ‘ad 2 cm. longis et 0.8 cm. latis apice rotundatis
vel leviter emarginatis; rhachi glabra leviter angulata sub anthesi ad 8 mm.
longa, floribus bracteis eis basi rhachis similibus subtentis; pedicellis
puberulis et parce albo-glandulosis, 2—3.5 mm. longis, basim versus uni-
vel bibracteolatis, bracteolis linearibus circiter 3 mm. longis et 0.4 m

latis conspicue albo- glanduloso-marginatis; floribus extus albo- ibaalis:

calyce 5—6 mm. longo, tubo cupuliformi sub anthesi circiter 3 « 3 mm

limbo suberecto intus glabro lobis inclusis 2—3 mm. ongo, lobis 5 deltoideis
acutis, 1-1.5 mm. longis, circiter 2 mm. latis, margine glandulas albas
lineares patenter gerentibus, sinibus rotundatis: corolla tenuiter carnosa
intus glabra cylindrica, sub anthesi 10-12 mm. longa et 4—5 mm. lata,
apice paullo contracta, lobis deltoideis acutis circiter 1 & 1.5 mm. stamini-
bus 10 quam ee paullo brevioribus, filamentis membranaceis ect
circiter 1.5 mm. et 4 mm. longis margine puberulo-ciliolatis apice angustatis,
antheris aeaee circiter 9 mm. et 8 mm. longis, tubulis quam thecis
paullo longioribus per ar ious dehiscentibus : stylo filiformi corol-
lam subaequante subtrunca

CoLomsra: Antioquia: oe ween Yarumal and Medellin, alt. 2700 m., near
stream-bed on paramo, Feb. 20, 1942, Metcalf & Cuatrecasas 30159 (A, TypPE, US),
shrub 4-6 m. high; bracts light rose-colored; calyx green, with silvery hairs; corolla
red, green-tipped.

Cavendishia Dugandiana is characterized by its small leaves, pilose
flowers, and linear glandular-margined bracteoles. Its closest relatives
appear to be C. Killipit A. C. Sm. and C. scabriuscula (H. B. K.) Hoer.,
from both of which it is distinguished by its smaller and fewer-nerved eae
blades, fewer-flowered inflorescences, shorter pedicels, smaller flowers,
calyx-lobes with linear glands, etc

It is a pleasure to dedicate this species to Dr. Armando Dugand G.,
Director of the Instituto Botanico, Universidad Nacional de Colombia, in
recognition of his valuable work on the flora of Colombia.

Satyria arborea sp. nov.

Arbor ad 12 m. alta ubique praeter inflorescentiam glabra, ramulis crassis
subteretibus cinereis; petiolis rugosis valde incrassatis (2.5-3 mm. dia-
metro) 3-5 mm. longis: laminis coriaceis elliptico- oblongis, 11-12 cm.
longis, 3.3—4 cm. latis, basi gradatim acutis, apice obtusis, margine valde

(pilis 0.1-0.25 mm. longis), copiosis, in axillis foliorum mox de pany
dispositis, racemosis, 10—25-floris; rhachi leviter angulata gracili 5-20 m
longa basim versus interdum ramosa, basi bracteis pluribus imbricatis

470 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXIV

papyraceis ovatis acutis pauciglanduloso-marginatis 1-2 mm. longis et latis
ornata: floribus bracteis oblongis acuminatis circiter 1.5 X 0.7 mm. sub-
tentis: pedicellis gracilibus sub anthesi 13-18 mm. longis basim versus
bibracteolatis; calycis tubo cupuliformi 1.5—2 mm. longo et circiter 3 mm.
diametro, limbo papyraceo intus glabro subpatente lobis inclusis 1-1.5 mm.
longo, lobis 5 inconspicuis apiculatis, sinibus complanatis ; corolla tenuiter
carnosa intus glabra cylindrico-urceolata, sub anthesi 8-11 mm. longa et
basim versus 2-3 mm. diametro, lobis 5 deltoideis subacutis circiter 0.5
1 mm.: staminibus quam corolla multo brevioribus, filamentis glabris in
tubo membranaceo 2.5—-3 mm. longo connatis, antheris alternatim circiter
3 mm. et 3.5-4 mm. longis, tubulis thecas subaequantibus per rimas latas
ovales dehiscentibus; stylo filiformi corollam subaequante, stigmate minute
peltato.

COLOMBIA: tioquia: Between Valdivia and Yarumal, alt. 2200 m., along
roadside in partial shade, Feb. 20, 1942, Metcalf & Cuatrecasas 30131 (A, TyPE, US),

e 10-12 m. high; calyx and corolla rose-colored, finely white-pilose, the corolla
white distally.

Satyria arborea is apparently most closely allied to the recently de-
scribed S. Allenii A. C. Sm. from Panama (in Ann. Mo. Bot. Gard. 28:
451. 1941), from which it differs in its arborescent rather than epiphytic
habit, shorter petioles, slightly larger leaf-blades with more highly connate
secondary nerves, longer and more copiously flowered inflorescences, longer
pedicels, and longer anthers (these being about twice as long in S. arborea
than in S. Allenii). Another relative of the new species is S. breviflora
Hoer., but the two plants differ in many details of foliage and inflorescence ;
the flowers of S. breviflora are larger throughout and are glabrous, at least
at anthesis, while those of S. arborea are persistently puberulent.

Themistoclesia epiphytica sp. nov.

ter racemosa 4—7-flora (floribus raro ad 1 reductis), rhachi pedicellisque
lus minusve hispidulis; rhachi gracili plerumque 3-5 mm. longa, basi
bracteis pluribus papyraceis oblongis acutis circiter 2 mm. longis circum-

pidulis subtentis; pedicellis gracilibus 6-13 mm. longis medium versus
bibracteolatis, bracteolis ut bracteis sed minoribus margine hispidulis;

lyce 5-6 mm. longo extus pilis 0.15-1 mm. longis plus minusve dense
hispidulo etiam interdum sparse brunneo-glanduloso, tubo obconico leviter
5-angulato 3.5—4.5 mm. longo 2.5—3 mm. diametro, limbo papyraceo erecto-
patente lobis inclusis circiter 1.5 mm. longo, lobis 5 apiculatis 0.3—0.7 mm.
longis, sinibus complanatis; disco pulvinato minute vel conspicue hispido
(pilis ad 0.5 mm, longis); corolla tenuiter carnosa praeter lobos parce

1943 ] SMITH, SOUTH AMERICAN PLANTS, X 471

hispidulos glabra, 9-10 mm. longa, 4-5 mm. diametro, lobis 5 deltoideis
subacutis circiter 1 & 1.5 mm.; staminibus quam corolla paullo breviori-
bus, filamentis glabris membranaceis filiformi-ligulatis alternatim circiter
3 mm. et 4 mm. longis, antheris 5—5.5 mm. longis, thecis 1.5—2 mm. longis,
tubulis quam thecis fere duplo longioribus per rimas ovales 0.7-1 mm.
longas dehiscentibus; stylo filiformi corollam subaequante truncato,

CotomspiA: Narifio: Paramo del Tabano, alto de la Cordillera, entre Pasto y
El Encano, vertiente occidental, alt. 3200 m., 11 Enero 1941, Cuatrecasas 11899 (A,
type, Col), friitex epifito, ramoso péndulo; corola vermell6n; Putumayo: Alta
cuenca del rio Putumayo, filo de la Cordillera entre El Encano y Sibundoy ; paramo
de San Antonio del Bordoncillo, alt. 3250 m., Cuatrecasas 11725 (Col, US), frutex epi-
fito scandens; corola vermellon.

The genus Themistoclesia, which in 1932 (in Contr. U.S. Nat. Herb. 28:
439-444) I supposed to consist of only six species, has since been greatly
expanded; in the most recent treatment (Sleumer in Bot. Jahrb. 71: 389-
393. 1941) 17 species are recognized. Themistoclesia epiphytica is char-
acterized by its epiphytic habit, proportionately long calyx-tube, and long
filaments. It is probably most closely related to T. Pennellit (A. C. Sm.)
Sleumer, from which it differs in its larger and acuminate leaf-blades,
longer pedicels, slightly larger calyx and corolla, unequal filaments, and
somewhat smaller and differently proportioned anthers. Themistoclesia
peruviana A. C. Sm., another ally of the new species, differs in its more per-
sistently hispid habit, short pedicels, short filaments, differently propor-
tioned anthers, etc.

ARNOLD ARBORETUM,
HARVARD UNIVERSITY.

472 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

THE COMPARATIVE MORPHOLOGY OF THE WINTERACEAE
II. CARPELS

I. W. BAILEY AND CHARLOTTE G. NAST

With six plates

INTRODUCTION

IN a previous paper, Bailey and Smith (1) called attention to the
remarkable stamens and carpel of Degeneria, suggesting that they might
prove to be of considerable significance in interpreting the floral morphology
of the Ranales. In this genus of the monotypic family Degeneriaceae, both
the stamens and the carpel appear to be primitive, palmately 3-veined
sporophylls of but slightly modified form. The lamina of the megasporo-
phyll is adaxially folded or conduplicate and bears numerous ovules on its
morphological upper surface. In other words, the ovules are not attached
to the margins of a classical, involute, sealed sporophyll, but to the ventral
surface of the megasporophyll as in certain of the Pteridospermae.

Carpels of a fundamentally similar type occur in the Tasmannia section
of Drimys and in certan species of Bubbia. Other representatives of the
Winteraceae exhibit various trends of specialization in the modification of
these primitive ranalian megasporophylls. The numerous species of the six
genera of the family provide abundant material for comparative studies and
demonstrate that the salient trends of specialization in the megasporophylls
of the Winteraceae are unlike those which characterize the Magnoliaceae
and the Himantandraceae (Bailey, Nast, and Smith, 2).

MATERIAL AND METHODS

The herbarium specimens upon which our floral studies are based have
been listed in Dr. Smith’s papers (3, 4, 5) and need not be relisted here.
The vascularization of carpels and other floral organs is commonly recon-
structed by the laborious study of serial sections. We have found, how-
ever, that such investigations may be facilitated and much accelerated by
carefully controlled clearing of flowers or parts of flowers. In the case of
material from herbarium specimens, the following procedure is helpful and
widely applicable to both flowers and leaves. The dry flowers or leaves
are first heated in water until thoroughly saturated and freed of air. They
are then transferred to a dilute aqueous solution of NaOH and placed in
sealed bottles in an incubator at 55° C. until cleared to the desired degree.
They are next washed free of NaOH and transferred to vials of 95% alcohol
for visual study. Permanent mounts can be made by passing the material
through absolute alcohol, diaphane solvent to diaphane or through absolute
alcohol, toluene to clarite. The clarity of the vascularization depends upon
the refractive index of the medium and the consistency of the tissues.

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, II 473

Commonly the venation shows best in alcohol, and the unmounted speci-
mens may be turned for examination at all angles. Staining of the material
is unnecessary either for visual examination or for photomicrography. The
method is adequate except in cases where organs are excessively hairy or
where they contain numerous clusters of sclereids, as in the floral parts of
certain species of Bubbia, Exospermum, and Zygogynum. In dealing with
such material, one is forced to rely largely upon serial sections.

CARPELS OF THE TASMANNIA SECTION OF DRIMYS

There is a conspicuous tendency in many ranalian families for the leaves
to be adaxially folded or conduplicate during the earlier stages of their
ontogenetic development. In such species of the Tasmannia section of
Drimys as D. piperita Hook. f., the carpels resemble in external form these
conduplicate young leaves. There is a similar differentiation of the mega-
sporophylls into stipe (petiole), Fig. 7, and conduplicate lamina, compare
Figs. 1 and 12. e more or less closely approximated ventral ‘surfaces!
of the conduplicate megasporophyll frequently are not extensively and
firmly concrescent prior to and during anthesis. Thus, the conduplicate
lamina may be spread open without serious or extensive rupture and dis-
tortion of the carpel or of its constituent tissues, Fig. 17.

The carpels are vascularized by three veins, a median vein and two lateral
veins. The median vein frequently bifurcates and has numerous conspicu-
ous branches of considerable length, Fig. 17. The two lateral veins com-
monly have short branches that are directed both outward toward the
margins of the megasporophyll and inward toward the branches of the
median vein, Figs. 1 and 3 and left half of Fig. 17. In certain cases, the
lateral veins have extensive branches that run parallel to them, right half
of Fig. 17. The ovules are remote from the margins of the megasporophyll
and are invariably attached between the median and the lateral veins, i.e.
in the parts of the carpel where the branches of the median and lateral
vascular systems tend to overlap and anastomose. The ovules are vascu-
larized in part by extensions of the veinlets of the lateral systems (a in
Figs. 17 and 18), in part by extensions of the veinlets of the median system
(6 in Figs. 17 and 18), and in part by strands arising from anastomosed

1In order to avoid confusion in morphological descriptions, it is essential to recog-
nize that the hoes ventral and dorsal are used in two distinct ways, (1) in referring to
the upper and lower surfaces of flat, cladified appendages, and (2) in designating parts
or sides of folded meeasporophyls viz. carpels. There are ontogenetic and phylogenetic
implications in both usa

We shall refer to he eae or morphologically adaxial surface of leaves and sporo-
phylls as ventral, to the lower or morphologically abaxial surface as dorsal. In con
duplicate ranalian carpels, the exposed outer surfaces are, therefore, dorsal and the
internal ones ventral.

In dealing with conduplicate cua carpels, we shall refer to the primitively

orsal

whereas in certain speci eit forms, Figs. 20-23, their actual orientation may be in
part terminal or even abaxia

474 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXIV

branches of the median and lateral systems (c in Figs. 17 and 18). The
details of the vascularization, both of the carpels and of the ovules, fluc-
tuate considerably from carpel to carpel of the same species and of the
same flower. Thus, the ovules may be vascularized at times largely by the
median system, by the lateral systems, or by varying combinations of
these systems.

As shown in Fig. 12, the free margins of the conduplicate megasporophylls
are provided with conspicuous, glandular-appearing, papillate cells or hairs,
which extend backward from the margins for varying distances over the
exposed dorsal surfaces of the conduplicate megasporophylls. That the
papillae are glandular and function as a stigmatic surface is demonstrated
by adhering pollen at anthesis, Fig. 2, and by pollen-tubes which penetrate
the mat of papillae. Thus, the carpels are provided with extensive stig-
matic crests (actually double), which extend from the region of the stipe
along the i aS adaxial parts of the sporophyll and slightly overtop
its apex, Figs. 1 and 2.

The fertile oe have a fundamentally similar conduplicate form,
placentation, and vascularization throughout the 15 species of the Tas-
mannia section of Drimys that we have studied. There are variations in the
length of the stipe, in the size and form of the conduplicate lamina, in the
extension of the outer stigmatic surfaces, in the concrescence of the approxi-
mated ventral surfaces, in the number and form of the ovules, and in the
details of the vascularization, but only in two of the investigated species
are the deviations of considerable magnitude. The carpels of D. stipitata
Vickery are characterized by their excessively elongated stipe; those of
D. lanceolata (Poir.) Baill. by their unusually fleshy conduplicate lamina
of nearly globular form, Fig. 27.

The sterile megasporophylls of staminate flowers usually are smaller, but
they have a similar conduplicate, palmately 3-veined lamina, Fig. 2.
Although no ovules are formed, the stigmatic surfaces are conspicuously
developed and are encrusted with firmly adherent pollen at anthesis, Fig. 2.
In the more rudimentary forms of sterile carpels, the branches of the median
and lateral veins may be feebly developed or absent, thus resembling the
venation of fertile carpels during the earlier stages of their ontogeny.

CARPELS OF THE WINTERA SECTION OF DRIMYS

In the Tasmannia section of Drimys, the carpels are adaxially folded or
conduplicate and may be spread open into megasporophylls of but slightly
modified form. When unfolded, Fig. 17, their venation resembles that of a
palmately 3-veined appendage. The elongated locule of the folded carpel
is oriented approximately parallel to the long axis of the megasporophyll,
and the numerous ovules are attached to two placental ridges that are situ-
ated between the median and lateral veins.

The carpels throughout the Wintera section of Drimys have a funda-
mentally similar vascularization and placentation, but the external stigmatic
surfaces are restricted to the adaxially projecting, subapical, ventral part
of the carpels, Fig. 6. The approximated ventral surfaces of the condupli-

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, II 475

cate carpels are firmly concrescent, Fig. 14, except at the level of the stig-
matic projection, Fig. 13, and therefore the megasporophylls cannot be un-
folded as in the case of the more primitive sporophylls of the Tasmannia
type. At the level of the stigmatic projection, Fig. 13, there is a cleft-like
opening or a loose suture which extends outward from the locule as in
Drimys piperita, Fig. 12, of the Tasmannia section of Drimys. Serial trans-
verse sections indicate that the closure of the carpels progressed upward
from the base and downward from the apex of the conduplicate lamina,
and commonly also centripetally, since vestiges of the cleft-like opening tend
to persist internally, Fig. 14, after they have been completely eliminated
externally. These conclusions regarding closure may be verified by the
serial sectioning of Tasmannia type carpels, many of which exhibit incipient
stages of concrescence. With the closure of the cleft-like opening, the car-
pels retract, and eventually eliminate, the stigmatic crests from the sealed
parts of the megasporophylls. Thus, the subapical projection of the
Wintera type of carpel is not to be interpreted as a style-like outgrowth,
but rather as a persistent remnant of the extensive adaxial stigmatic crests
of Tasmannia type megasporophylls, Fig. 19.

CARPELS OF BUBBIA, BELLIOLUM, EXOSPERMUM, AND ZYGOGYNUM

Various transitional stages in the closure of conduplicate carpels and in
the restriction of their stigmatic crests occur in the genus Bubbia, but in
this genus, as in Belliolum, Exospermum, and Zygogynum, there is in
addition a more or less pronounced abaxially directed deformation of the
conduplicate megasporophylls.

The carpels of Bubbia Archboldiana A. C. Sm. (Brass 12712) resemble
those of the Tasmannia section of Drimys in their vascularization, placen-
tation, and in having extensive stigmatic surfaces, Fig. 20. They differ in
their angular external form and in their conspicuously broadened and
flattened apices. There are, however, no significant modifications in the
longitudinal orientation of the locule, the placental ridges, or the median
and lateral veins. The carpel of Bubbia megacarpa A. C. Sm. (Brass
10249) likewise has extensive stigmatic margins, Fig. 27, but it exhibits
a profoundly modified form, due to the overtopping of the shortened dorsal
side of the sporophyll by its over-extended, conduplicate, ventral side. In
other words, there is an abaxially directed deformation of the megasporo-
phyll which produces a short, apically much broadened carpel. The locule,
the placental ridges, the lateral veins, and the stigmatic crests all show
pronounced abaxial curvatures. Owing to these concomitant deformations
from longitudinal to approximately transverse orientations, it is evident
that the apparently terminal parts of the carpel, Fig. 21, actually are
homologues of the ventral parts of the primitive carpels illustrated in Fig. 1.
The true apex of the carpel is curved around onto the dorsal side of the
megasporophyll. The massive median trace dissociates in the base of the
carpel, Fig. 21, into numerous vascular strands, the majority of which
extend upward toward the transversely oriented parts of the placental
ridges. Their recurved ends either terminate in the placental ridges or

476 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

anastomose with the short downwardly directed branches of the lateral
veins. Comparatively few branches of the median vein are directed
diagonally toward the longitudinally oriented basal parts of the placental
ridges. Most of the ovules are vascularized by veinlets of the lateral
systems.

The carpels of other species of Bubbia, as of Belliolum and Zygogynum,
exhibit more or less conspicuous abaxially directed deformations and, in
addition, much restricted stigmatic crests which tend to assume a trans-
Heed terminal orientation. Although the carpels of Bubbia longifolia
A. C. Sm. (Brass 13868) have less exaggerated abaxial deformation, Figs.
10 and 22, and therefore a less modified median vein than that of B. mega-
carpa, Fig. 21, the stigmatic crests and the ovules are restricted to the
diagonal upper part of the distorted megasporophylls. The approximated
ventral surfaces in the lower ventral part of the conduplicate carpel are
concrescent and there is no cleft-like opening extending outward from the
locule except in the upper part of the megasporophyll which subtends the
stigmatic crests. The placental ridges and ovules are likewise restricted to
the upper part of the sporophyll. The ovules are vascularized in part by
veinlets of the lateral systems and in part by extensions of the median
system, Fig. 22. The single, terminal megasporophyll of Bubbia mono-
carpa A. C. Sm. (Kanehira & Hatusima 12105), Fig. 9, resembles the
carpels of B. longifolia both in its abaxial deformation and its internal
structure. It demonstrates, as does the single terminal carpel of B. mega-
carpa, that the distorted forms of the megasporophylls in polycarpellate
species are not due solely to excessive compression of adjacent organs
during ontogenetic development. The immature carpels, illustrated in
Figs. 7 and 8, indicate furthermore that the abaxial deformations are of
phylogenetic rather than purely ontogenetic development.

In Bubbia Clemensiae A. C. Sm. (Clemens 5157 and 4596), abaxial
deformation coupled with concomitant modifications of the lower part of
the megasporophylls have produced a more nearly symmetrical carpel,
Fig. 23. Externally it is difficult to distinguish the sealed, crestless, adaxial
side of the conduplicate sporophyll from its dorsal side. Furthermore, the
lateral veins not infrequently are fused in the lower closed part of the
carpel and separate at a higher level, thus simulating the bifurcation of the
median vein. The transversely oriented placental ridges extend downward
into the locule for a considerable distance, Fig. 15. Therefore, the attach-
ment of the ovules is more remote from the stigmatic margins and the
ovules are vascularized by more downwardly extended veinlets of the
lateral and median systems.

The carpels of Bubbia Whiteana A. C. Sm. (Brass 2278), Figs. 11 and
24, resemble the megasporophylls of B. Clemensiae in their concealed
abaxial deformation and in their much modified median and lateral vascular
systems, but differ from them in having less extensive terminal stigmatic
crests and ovules that are attached at a higher level of the locule. The
ovules, as in B. Clemensiae, are vascularized by extensions of both the
lateral and median systems.

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, II 477

Restriction of the stigmatic crests is carried to an extreme in the short,
terminally broadened, fleshy carpels of Bubbia auriculata v. Tiegh.
(Vieillard 2280) and B. semecarpoides (F. v. Muell.) Burtt (Kajewski
1216). There is a pronounced abaxially directed deformation in the
ventral part of the conduplicate carpel, as indicated by the curvature of the
lateral veins and the diagonal orientation of the placental ridges, Fig. 25,
but the conduplicate ventral part of the carpel does not extend across the
broad terminal face of the megasporophyll and does not overtop a shortened
dorsal side as in B. megacarpa, Fig. 21, or B. Clemensiae, Fig. 23.

In most species of Bubbia, as in the Tasmannia section of Drimys, the
placental ridges are closely correlated in orientation and extension with the
stigmatic crests. In certain carpels, however, e.g. those of B. pachyantha
A.C. Sm. (Brass 4371) and B. isoneura v. Tiegh. (Vteillard 17), there are
more or less conspicuous unconformities, as in the megasporophylls of the
Wintera section of Drimys, Fig. 19. In other words, the placental ridges
persist for varying distances in the sealed, crestless, ventral part of the
megasporophylls. Such unconformities between the orientation and exten-
sion of the placental ridges and the stigmatic crests occur at times in the
carpels of Belliolum, Fig. 26. Although the carpels of Bedliolum in general
resemble those of the more highly specialized species of Bubbia (viz. those
having more or less restricted, terminally oriented crests), the attachment
of the ovules tends to be at lower levels and the branches of the lateral
veins are more downwardly extended, as in Bubbia Clemensiae, Fig. 23.

The coriaceous megasporophylls of certain polycarpellate species of
Bubbia, e.g. B. pachyantha, are closely crowded and firmly coherent both
preceding and during anthesis. Thus, as stated by Smith (3), “the gynae-
cium has the appearance of a compound ovary with a 3- or 4-parted stellate
stigma.” Such gynaecia closely resemble that of Exospermum stipitatum
(Baill.) v. Tiegh. (Viedlard 2281). There is, accordingly, no such sharply
defined morphological distinction between coherent (Exospermum) and
free (Bubbia) carpels as hypothesized by van Tieghem (6). The indi-
vidual carpels may be readily separated after clearing treatments and each
is provided with an independent epidermal layer. Only in the gynaecia of
Zygogynum are the carpels fused into a concrescent mass without internal
evidences of sutures, Fig. 16.

The ovules of Exospermum stipitatum are not restricted in their attach-
ment to conspicuous stigmatic ridges, but are scattered over the walls of
the locule, a modified type of placentation that is suggestive of certain
Nymphaeaceae and Lardizabalaceae. The cleft-like opening at anthesis
is partly closed externally and does not extend outward from the locule to
the stigmatic surface as in Zygogynum spathulatum v. Tiegh. (Vieillard
2266), Fig. 16. In this, as in other species of Zygogynum, the shortened
placental ridges, Fig. 16, are situated on the abaxial side of the locule.
The ovules are not attached to the dorsal part of the carpel, however, as
hypothesized by van Tieghem (6), but to a morphologically ventral part
of the conduplicate megasporophylls that has been deflected into an

478 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

abaxial orientation and thus overtops the much shortened morphologically
dorsal part of the carpel.

The carpels of Pseudowintera fluctuate considerably in form. In general,
those of P. axillaris var. colorata (Raoul) A. C. Sm., Fig. 4, tend to resemble
the megasporophylls of the Wintera section of Drimys, whereas those of
P. axillaris var. typica A. C. Sm., Fig. 5, exhibit more pronounced abaxially
directed deformation as in certain species of Bubbia.

SIGNIFICANCE OF INTERNAL PAPILLATE SURFACES

The conduplicate megasporophylls of Degeneria (Bailey and Smith, 1)
have more or less conspicuously flaring free margins and are characterized
by having short, glandular-appearing hairs that are distributed inwardly
from the margins along the approximated ventral surfaces as far as the
flanks of the placental ridges. Thus, the cleft-like opening that extends
outward from the locule is partly occluded by interlocking papillae, and
pollen does not have direct access to the locule in most cases. The pollen
grains become attached to the outer glandular projections and the pollen-
tubes penetrate apparently through the mat of interlocking papillae.

In the Winteraceae, the free margins of the conduplicate lamina (in
unsealed parts of the carpels) are more closely approximated and the
stigmatic papillae extend backward from the margins over the exposed
dorsal surfaces of the sporophylls, Figs. 12 and 13. There are, in addition,
more or less numerous papillate cells along the ventral surfaces of the
conduplicate megasporophylls, Figs. 12 and 13. These projecting cells
commonly jacket both flanks of the placental ridges even in sealed parts
of the carpels, Fig. 14. In regions of incipient closure (phylogenetic, not
ontogenetic) the papillae not infrequently appear to enlarge and to inter-
lock and possibly at times to play an initial role in the developing suture.

Unfortunately, herbarium specimens do not provide adequate material
for studying the finer cytological and histological details of the closure of
carpels or of the penetration of pollen-tubes. Such details can be clarified
only by the study of living and adequately killed and fixed material. Fur-
thermore, it is essential that the carpels of Degeneria and the Winteraceae
be studied in all stages of their ontogenetic development and during the
changes that they undergo subsequent to anthesis.

It should be emphasized in this connection that there is considerable
variation in the form of winteraceous carpels, in the details of their vascu-
larization, in the extent of their closure, etc., not only in material from
different collections of the same species, but also in different carpels from
the same flower. Therefore, our descriptions and illustrations represent
average or typical conditions. Numerous variations in the structural de-
tails of particular genera and species may be anticipated when more
abundant and complete collections of these remarkable plants become
available.

DISCUSSION AND CONCLUSIONS

The carpels of the Tasmannia section of Drimys are conduplicate mega-
sporophylls of but slightly modified form and closely resemble the mega-

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, II 479

sporophyll of Degeneria. In both cases, the megasporophyll is clearly
differentiated into stipe and adaxially folded lamina, Figs. 1 and 2. When
spread open, the lamina exhibits a palmately 3-veined vascularization such
as characterizes both the microsporophylls and the sterile sporophylls
(staminodes) of the Degeneriaceae and Himantandraceae. The numerous
anatropous ovules, Fig. 18, are attached to more or less conspicuous placen-
tal ridges that are situated between the median and the lateral veins, Fig.
17. The ovules are vascularized in part by short branches of the two lateral
veins, in part by branches of the median vein, and in part by strands origi-
nating near anastomoses of the lateral and median vascular systems, the
ratios of the three types of vascularization fluctuating from carpel to
carpel. The conduplicate form, placentation, and vascularization of the
megasporophyll do not conform with the classical interpretation of the
angiosperm carpel as an involute megasporophyll bearing marginally
attached ovules. In the Winteraceae and Degeneriaceae, the ovules are
borne on the morphological upper surface of the megasporophyll, between
the lateral and median veins, Fig. 17, and remote from the margins of the
sporophyll, Figs. 12, 13, 15, and 16. It should be noted in this connection
that the broad, palmately 3-veined microsporophylls of the Degeneriaceae
and Himantandraceae are not differentiated into filament, anther, and con-
nective, and that they bear four slender elongated sporangia that are im-
mersed beneath the dorsal surface of the sporophyll, midway between the
median and the lateral veins. Thus, in these primitive ranalian carpels and
stamens, neither the megasporangia nor the microsporangia are borne upon
the margins of the sporophylls. It should be noted, in addition, that there is
no conclusive evidence at present for inferring marginal attachments in
ancestral angiosperms rather than ventral and dorsal ones as in certain of
the Pteridospermae.

In the Winteraceae, the chief trends of specialization of the primitive
ranalian megasporophyll lead toward closure of the conduplicate sporo-
phyll (by concrescence of its approximated ventral surfaces) and con-
comitant restriction of its external stigmatic surfaces. In Bubbdia,
Belliolum, Exospermum, and Zygogynum these trends of specialization
are complicated by more or less pronounced abaxially directed deformation
which results in an apically much broadened carpel bearing more or less
terminally or even abaxially (Zygogynum) oriented stigmatic crests. In
the more specialized forms, the ovules tend to be attached to transversely
or even abaxially oriented placental ridges in the upper part of the carpel.

On the contrary, in the Himantandraceae and Magnoliaceae (the closest
relatives of the Degeneriaceae) the specializations of the primitive ranalian
megasporophyll lead toward a pronounced constriction of the upper part
of the conduplicate lamina. This constricted, sterilized, upper part, viz.
style, has more or less extensively “decurrent” stigmatic surfaces and still
exhibits a conduplicate structure. The few remaining ovules are thus con-
fined to the fertile, lower, broader part of the conduplicate carpels, which
may remain partly open (Himantandraceae) or be firmly sealed (Mag-
noliaceae).

480 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Our detailed investigations of the numerous representatives of the
Winteraceae support the suggestion (Bailey and Smith, 1) that the re-
markable megasporophyll of Degeneria may afford significant clues for
interpreting the diverse carpellary structures of the Ranales. The occur-
rence of fundamentally similar types of conduplicate megasporophylls
throughout the Tasmannia section of Drimys provides a broad basis for
comparative investigations of the various ranalian families. If the dico-
tyledons are monophyletic, the megasporophylls of the Degeneriaceae and
Winteraceae should prove to be equally significant in studying the car-
pellary specializations of other orders.

LITERATURE CITED

1. Battey, I. W., and A. C. Smitu. Degeneriaceae, a new family of flowering plants
from Fiji. jue Arnold Arb. 23: 356-365. pl. 1-5. 1942.
————, C. ast, and A. C. Smitn. The family Himantandraceae. Jour.
Arnold Arb. “ba, 190-206. pl. 1-6. 1943.
3. Smiru, A. C. Studies of Papuasian plants, V. Jour. Arnold Arb. 23: 417-443. 1942.
The American species of Drimys. Jour. Arnold Arb. 24: 1-33. f. 1-3.
1943.
—--- xonomic notes on the Old World species of Winteraceae. Jour.
Aen Pats 24: 119-164. f. 1-6. 1943
6. TiEGHEM, P. vAN. Sur les dicotylédones is groupe des Homoxylées. Jour. de Bot.
14: 259-297, 330-361. 1900.

EXPLANATION OF PLATES

PiaTE I

Carpels cleared in dilute NaOH and apg unstained in 95% alcohol.
Fic. 1. Drimys piperita Hook. f., Ramos and Edano 38897. Two attached carpels,
showing stipe and venation of conduplicate lamina, x 20. Fic. 2. Drimys macrantha
AC.

stigmatic surface 24. Fic. 3. Drimys insipida (R. Br.) Pilger, Caley. Detached
young fruit, rn at left outwardly projecting branches of lateral vein, x 17.

PLATE II

Carpels cleared in dilute NaOH and mounted unstained in diaphane. 4.
phialdoge cen axillaris var. colorata (Raoul) A. C. Sm., Kirk. Two attached carpels,

typica A. C. ‘Sm., bgpea Two attached carpels, showing lateral view of vasculariza-
tion, X 20. Fic. 6. Drimys confertifolia Phil., Moseley. Two attached carpels, show-
ing vascularization aad apex of torus, x 20.

Puiate III

Carpels cleared in dilute NaOH and photographed unstained in 95% alcohol. Fic. 7.
Bubbia monocar pa A. C. Sm., Kanehira and Hatusima 12105. Young carpel, showing
early stage in the deodineuent of median and lateral vein 24. or 8. The same.
Somewhat older carpel, showing extensions of the vasculaneation, x
same. Carpel showing vascularization at anthesis, x 24. Fic. 10. Bubble longifolia

C. Sm., Brass 13868. ature detached carpel, showing vascularization,

Fic. 11. Bubbia Whiteana A. C. Sm., Brass 2278. Two somewhat coherent carpels,
showing vascularization, x 17

Jour. Arnotp Ars. Vor. XXIV Prate |

CoMPARATIVE MorRPHOLOGY OF THE WINTERACEAE

Pirate I]

Jour. Arnotp Ars. VoL. XXIV

THE WINTERACEAE

F

C

MorPHOLOGY

ARATIVE

Comp

Prate III

XXIV

Jour. ArNotp Ars. Vou.

THE WINTERACEAE

COMPARATIVE MorpHOLOGY OF

Piare IV

Jour. Arnoip Ars. Vor. XXIV

a,

es

‘ eae = Nate, :
Oye ey as : ©
is “ POOLE a Eee: oS bie’

as,
aT

, a ae = NG “4 my Sere Fans ne ic rae Sais ELLE Re Se]
(tye ok a So Rose Saas nee & eee Bee e Aa Sere
s = SS Sa: S i y NS Bets Os Bs CK os

v7 ate FS
Cee ts : e i oy pve ve Se, Poe ages e
Sacre eS aN: ii Reet eee lies; " Re is = ae
em . sii sao SSN SS » 2 a » : 2; Een RS BE IRE Lo
ea SIN ~ ge

chs 3: WN S Se ee
we : x

SS “
S EET ww
¥: cf

Ta.
a Zee

pa

Lt oe a Cage

~ > J oe hy ti ars OSH, ee Testes . A
Ke BS Uf, Pe spies Sekagee SSS LS eal
EST ee Jie san ae a Cue, je see Se Ste

KS es = iS, ee
Se

ie
aS Ve
: Pa
AE
3. Sgr po —

Se

cere oe

THE WINTERACEAE

CoMPARATIVE MORPHOLOGY OF

Jour. ARNotp Ars. VoLt. XXIV Piate V

-——_—
~~
_—
—

—

ere ERE EE RY
Roast Se :

WINTERACEAE

CoMPARATIVE MORPHOLOGY OF THE

Jour. Arnotp Ars. Vou. XXIV Piate VI

CoMPARATIVE MorpHo.tocy OF THE WINTERACEAE

1943] BAILEY & NAST, MORPHOLOGY OF THE WINTERACEAE, II 481

Pirate IV

Sections of rag re-expanded carpels, stained in Haidenhain’s haematoxylin and
safranin. Fic. 12. Drimys piperita Hook. f., Ramos and Edano 38897. Transverse
section of pres carpel, showing stigmatic surfaces and ovules, x 50. Fic. 13.
Drimys granadensis var. mexicana (DC.) A. C. Sm., Tonduz 7342. eee section
of conduplicate, open, Be part of carpel, showing stigmatic surfaces and attachment
of ovules, x 50. Fic. 14. The same. Lower sealed part of carpel, ate internal
vestige of cleft and aia papillae, x 50. Fic. 15. Bubbia Clemensia m.,
Clemens 4596. Longitudinal section, showing cleft-like opening, downwardly project-
ing stigmatic ridges, and ovules, x 24. Fic. 16. Zygogynum spathulatum v. Tiegh.,
Vieillard 2266. Transverse pete of gynaecium, showing parts of three ra
carpels, x 40

PLATE V

17. Composite diagram of opened Tasmannia type carpels, showing typical
i

placentation, and vascularization of a cleared Wintera type carpel, showing restriction
of the stigmatic crests to a subapical projection. Broken lines indicate the extent of
the stigmatic crests in primitive Tasmannia type carpels.

PiaTE VI

ams of cleared carpels, showing the extent and orientation of the stigmatic
et

in placental tissue. and by broken lines as they enter the funicles. A. adaxial side of
carpel. Magnification x 18. Fic. 20. Bubbia Archboldiana ‘i C. Sm., Brass 12712.
Fic. 21. Bubbia megacarpa A. C. Sm., Brass 10249. Fic. 22. Bubbia longifolia A. C.
Sm., Brass 13868. Fic. 23. Bubbia Clemensiae A. C. Sm., Clemens 4596. Fic. 24.
Bubbia sags? A. C. Sm., Brass 2278. Fic. = Bubbia aupieulat v. Tiegh., Vieillard
228 26. Belliolum haplopas (Burtt) A. C. Sm., Brass 2959. Immature carpel
wih eee vascularization. Fic. 27. inte anceolatn (Poir.) Baill., Boorman.

BIoLocIcAL LABORATORIES,
HarvarpD UNIVERSITY.

482 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

FORSYTHIA VAHL, NOMEN GENERICUM CONSERVANDUM
ALFRED REHDER

Tue fact that Forsythia Vahl needs conservation, on account of the older
homonym Forsythia Walter of 1788, seems so far to have been overlooked ;
at least, no proposal for its conservation has been put forward, which is
probably not strange, since up to 1930 there was no necessity to conserve
the name, because Walter’s name is a clear synonym of the older Decumaria
L. (1762) and the so-called homonym rule (Internat. Rules Bot. Nomencl.
ed. 3, p. 19, Art. 61. 1935) was not adopted until 1930.

There can be no doubt that the name should be conserved, for the genus
contains some of the best known ornamental shrubs brightening in early
spring the gardens in temperate America and Europe, and the name is well
known in botanical and horticultural and even general literature. Though
not a large genus, containing only six or seven species, it includes a con-
siderable number of named varieties of spontaneous as well as garden
origin, the latter partly hybrids. The name will certainly be accepted by
a large majority as a nomen conservandum, when proposed for conserva-
tion at the next International Botanical Congress. It seems, therefore,
advisable to make herewith a formal proposal which should prevent the
taking up the next oldest name, Rangium Jussieu, by some author, and the
creation of a number of new combinations which can with certainty be
expected to be relegated to synonymy by the next Botanical Congress. So
far, only one author, namely Ohwi in 1932, seems to have taken up
Jussieu’s name and made a number of combinations.

Forsythia Vahl, Enum. Pl. 1: 39 (1805)

versus
Forsythia Walter, Fl. Carol. 154 (178
Rangium Jussieu in Dict. Sci. Nat. o4, 200 (1822

Type species: F. suspensa (Thunb. ) Vahl bDiessiviie suspensum Thunberg).
Forsythia Vahl has been accepted by all later authors up to 1932, when

Ohwi (in Act. Phytotax. Geobot. 1: 140) took up Rangium.
Forsythia Walter has not been accepted by any author and has always been
treated as a synonym of Decumaria Linnaeus (Sp. Pl. ed. 2, 1663.
762)

Rangium Jussieu, based on the monotypic Forsythia Vahl, remained with-
out specific epithet until Ohwi (1. c.), in 1932, took up the name and
made the following combinations:

Rangium suspensum (Thunb.) Ohwi, |. c. = alba hin suspensa Vahl, |. c

Rangium viridissimum (Lindl.) Ohwi, 1. c. = Forsythia viridissima Lindl. in Jour.
Hort. Soc. Lond. 1: 226 (1846).
Rangium Pidgeon {Rehd.] Ohwi, 1. c. = Forsythia viridissima var. koreana

in Jour. Arnold Arb. 5: 134 (1924).—Syn.: F. koreana (Rehd.) Nakai
in a Mag. Tokyo, 40: 471 (1926).

1943] REHDER, FORSYTHIA VAHL 483
Rangium ovatum (Nakai) Ohwi, 1. c. = Forsythia ovata Nakai in Bot. Mag. Tokyo,
1917).

Rangium japonicum (Mak.) Ohwi, |. c. = Forsythia japonica Makino in Bot. Mag.
Tokyo, 28: 105, fig. 4 (1914).

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

484 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

ROYLE’S “ILLUSTRATIONS OF THE BOTANY OF THE
HIMALAYAN MOUNTAINS”

WILLIAM T. STEARN

Tue “Illustrations of the Botany and other Branches of the Natural
History of the Himalayan Mountains, and of the Flora of Cashmere” (2
vols. quarto; London) by John Forbes Royle (1799-1858) stands with
Roxburgh’s “Plants of the Coast of Coromandel” (1795-1819), Wallich’s
“Plantae Asiaticae rariores” (1830-1832), Wight’s “Icones Plantarum
Indiae Orientalis” (1838-1853),1 and Wight’s “Illustrations of Indian
Botany” (4840-1850) as one of the most important illustrated works on
the flora of India. Royle was a pioneer economic botanist. His work is
not so much a descriptive systematic flora as an attempt to reveal “the
immense resources of British India, both as regards whatever is necessary
for the Agriculture, Manufactures, and Internal trade of the people, as
for the supply of a much extended External Commerce.” In it “the Geo-
graphical Distribution of Plants, as connected with Climate, is considered,
their Useful Properties detailed, and the principles which should guide their
culture in new situations deduced.” Medicinal plants receive special
attention. Many new species came to light during the preparation of the
work and are concisely described in its pages. It was issued in eleven parts,
costing £1. each, between 1833 and 1840. An unnumbered page of the
Introduction states the text-content of each part and its date of issue but
gives no information about the plates. This is an omission of some im-
portance. The plates illustrate in colour many of the new species de-
scribed in the text; they sometimes appeared before the corresponding
descriptions and, because they contain figures of floral dissections, the valid
publication of certain botanical names dates not from the text but from
the earlier issued plates.2, As Dr. T. A. Sprague remarks, the most satis-
factory way to determine the details of issue of a work of this kind is to
examine a copy still in the original wrappers. No such copy of Royle’s
work being known, he endeavoured to ascertain the issue of the plates by
a survey of contemporary journals. Notices in Loudon’s Gardeners’ Maga-
zine, vols. 9-16 (London, 1833-1840), enabled him to fix with precision
the issue of 30 out of a total 100 plates. He assumed that the remaining
70 plates were issued in numerical sequence. Since the publication of

1For dates of publication of Wight’s “Icones” see Merrill in Jour. Arnold Arb. 22:
222-224. 1941.

2In this respect Royle’s “Illustrations” is by no means unique. Webb and Berthelot’s
“Histoire naturelle des Iles Canaries” (1835-1850) is another important work in which
a number of names were first published on the plates; for fuller details see Stearn in
Jour. Soc. Bibl. Nat. Hist. 1: 58-59. 1937.

1943 ] STEARN, BIBLIOGRAPHICAL NOTES 485

Sprague’s paper (“‘The Dates of Publication of Royle’s Illustrations,” in
Kew Bull. 1933: 378-390. 1933), parts 1-10 in wrappers as issued have
come into the present writer’s hands. They show that the plates were not
issued in numerical sequence. ‘This find necessitates some modification,
fortunately not extensive, of Sprague’s account

The contents and dates of publication of the parts are as follows:
Part I. (Sept. 1833): pp. 1-40; pls. 4, 11-18, 22.
Part II. (March, 1834): pp. v—xii, 41-72; pls. 1, 19-21, 23-28.
Part III. (June, 1834): pp. xili-xx, 73-104; pls. 2, 5, 29, 31-35, 37, 38.
Part IV. (Sept. 1834): pp. 105-136; pls. 30, 39, 40, 42, 44-46, 64, 76 as 75

Part V. (Jan. 1835): pp. 137-176; pls. 3, 41, 48-51, 57, 62, 63, 74.

Part VI. (April, 1835): pp. 177-216; pls. 7, 36, 43, 55, 56, 58, 60, 61, 75 as 75a
(Phlomis, Salvia), and View of the Himalayan Moonta ar ti
to Vol. 1)

Part VII. (Aug. 1835): pp. 217-248; pls. 8, 9, 47, 52, 59, 65, 67-69, 71, 77.

Part VIII. (Dec. 1835): pp. 249-288; pls. 53, 54, 66, 70, 72, 73, 79 (63a), 80, 87, 88.

Part IX. (May, 1836): pp. 289-336; pls. 10, 81, 82, 83 (Procris), 84-86, 90, 100
(83, Putranjiva).

Part 2x. (Feb. 1839): pp. 337-384; pls. 89, 91-96, 98 (84a), 99 (78a), and Plan of
the H.E.I.C. Botanic Garden at Saharinpore (Frontispiece to Vol. 2

Part XI. (1840, before July): pp. xxi-IV’nxx, 385-472, title-pa san dedications,
preface, synoptic table of contents, list of plates, list of plants figured;
pls. 6, 97 [details of this part obtained from eg oe Gard. Mag. 16: 348.
1840 (July), Sprague in Kew Bull. 1933: 382.

Hence the dates of the ae of the plates and the parts in which
they were contained are as follo
Plate 1 (II, March, 1834), 2 i on 1834), 3 (V, Jan. 1835), 4 (I, Sept. 1833),

5 (III, June, eee 6 (XI, 1840), 7 (VI, April, 1835), 8-9 (VII, Aug. 1835), 10 (IX,

May, 1836), 11-18 in Sept. 1833), 19-21 (II, March, 1834), 22 (I, Sept. 1833), 23-28

(II, March, 1834), 29 (III, June, 1834), 30 (IV, Sept. 1834), 31-35 (III, June, 1834),

36 (VI, April, 1835), 37-38 (III, June, 1834), 39-40 (IV, Sept. 1834), 41 (V, Jan

1835), 42 (IV, Sept. 1834), 43 (VI, April, 1835), 44-46 (IV, Sept. 1834), 47 (VII, Aug.

1835), 48-51 (V, Jan. 1835), 52 (VII, Aug. 1835), 53-54 (VIII, Dec. 1835), 55-56 (VI,

April, 1835), 57 (V, Jan. 1835), 58 (VI, April, 1835), 59 (VII, Aug. 1835), 60-61 (VI,

eke 1835), 62-63 (V, Jan. 1835), 64 (IV, Sept. 1834), 65 (VII, Aug. 1835), 66 (VIII,

mt Dec. 1835), 74 (V, Jan. 1835), 75 as 75a (VI, April, 1835), 76 as 75 (IV, Sept.
1834), 77 (VII, Aug. 1835), 78 (IV, Sept. 1834), 79-80 (VIII, Dec. 1835), 81-86 (IX,
May, 1836), 87-88 (VIII, Dec. 1835), 89 (X, Feb. 1839), 90 (IX, May, 1836), 91-96
(X, Feb. 1839), 97 (XI, 1840), 98 as 84a, 99 (X, Feb. 1839), 100 as 83 (IX, May, 1836).

The dates of publication of Royle’s new species can be readily ascertained
from the above. As Sprague points out, the new species figured by Royle
“fall into two categories: (1) those with figures accompanied by analyses,
which date from the publication of the plate, where this is earlier than
the corresponding text; (2) those without analyses, which date from the
publication of the description (if any) in the text.’ Of the dates which
Sprague gives for the 134 new species figured by Royle, only 24 need
amendment. The following are corrected citations for these:

486 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

A plotaxis pecehincn DC. ex Royle, t. 59, sine anal. (Aug. 1835), p. 251, nomen
Dec. 1835); DC. Prodr. 6: 542 (Jan. 1837),° as A. gnaphalodes.

Astragalus fsieetle Royle, p. 199, t. 36 (April, 1835).

Campanula cashmeriana Royle, t. 62 (Jan. 1835), pp. 253-254 (Dec. 1835).

Cerasus cornuta Wall. ex Royle, t. 38 (June, 1834), pp. 205, 207 (April, 1835).

Chaptalia gossypina Royle, p. 18, nomen (Sept. 1833), p. 246, t. 59 (Aug. 1835), pp.
250, 251 (Dec. 1835).

Circaea cordata Royle, p. 211, t. 43 (April, 1835).

Codonopsis rotundifolia Royle, t. 62 (Jan. 1835), pp. 253-254 stig 1835).

Cucumis pseudo-Colocynthis Royle, pp. 218, 220, t. 47 _

Cyanathus lobatus Royle, t. 69 (Aug. 1835), p. 309 (May, 1

Dendrobium alpestre Lindley ex Royle, pp. 362, 365, t. bn ae 1835), p. 370 (Feb.
1839), non Swartz (1799).

Deutzia corymbosa R. Brown ex Royle, t. 46 (Sept. 1834), p. 216 are 1835).

Gaultheria trichophylla Royle, t. 63 (Jan. 1835), pp. 257-260 (D 835).

Holostemma Brunonianum Royle, p. 276, nomen, t. 66, sine anal. yon 1835), Decaisne

DC. Prodr. 8: 533 (March, 1844

Kohautia coccinea Royle, p. 241 (Aug. 1835), t. 53 (Dec. 1835).

Lonicera bracteata Royle, pp. 236-237 i t. 53 (Dec. 1835).

Osmorrhiza laxa Royle, p. 233, t. 52 (Aug. 1835).

Picrorhiza Kurroa Bentham in ey t. 71 eres 1835), p. 291 (May, 1836); Bentham,
Scroph. Ind. p. 47 (? Aug. 1835).

Primula elliptica Royle, t. 76 as t. 75 (Sept. 1834), pp. 310, 311 org 1836).

Primula rosea Royle, t. 76 as t. 75 (Sept. 1834), p. 311 (May, 1

eae utilis Royle, t. 38 (June, 1834), pp. 202, 206 (April, 1835).

Rheum spiciforme Royle, p. 37, nomen (Sept. 1833), t. 78 ries 1834), pp. 315, 316,
ae (May, 1836).

Roscoea alpina Royle, p. 19, nomen (Sept. 1833), pp. 357, 361, t. 89 (Feb. 1839).

Roscoea lutea Royle, p. 361, t. 89, “named R. spicata in plate, by csharneeevitaee (Feb.
1839).

“Roscoea purpurea” Royle, pp. 357, 361, t. 89 (Feb. 1839), non Smith (1806).

Plate 92 exists in three states. One has the name Lilium Thomsonianum
in lithographed lettering (e.g. at Royal Botanic Gardens, Kew; Arnold
Arboretum). Another has the name Fritillaria Thomsoniana instead, the
word Fritillaria and the final a@ of Thomsoniana being handwritten but
Thomsonian lithographed (e.g. at Lindley Library of Royal Horticultural
Society; Linnean Society of London; British Museum, Bloomsbury ;
British Museum [Natural History|, S. Kensington; University Library,
Cambridge, England; Royal Botanic Garden, Edinburgh; Bodleian Library,
Oxford). <A third state has the name Fritillaria Thomsoniana litho-
graphed throughout and the letters more evenly spaced than in the
second state (e.g. at British Museum [Natural History|; Botany School,
Cambridge, England; Cornell University). This plate was issued in
February, 1839, in part 10, on the back wrapper of which it is listed as
Fritillaria Thomsoniana,; the corresponding letter-press (under the name
Fritillaria Thomsoniana) was issued in 1840 in part 11. It would appear
that the plate was originally lettered Lilium Thomsonianum, then, follow-
ing David Don’s decision that the plant depicted belonged to the genus
Fritillaria, the word Lilium was erased from the plates already printed and

3For dates of publication of De Candolle’s “Prodromus,” see Stearn in Candollea
8: 1-4. 1939, and in Jour. Bot. 79: 27. 1941

1943 | STEARN, BIBLIOGRAPHICAL NOTES 487

coloured and the word Fritillaria was written in its place, the final m a
Thomsonianum being likewise erased and the u converted into an

later issues the name Fritidlaria Thomsoniana was completely tee che
anew. The copies at Kew and the Arnold Arboretum seem to have escaped
correction. The plant is now referred to Notholirion, a genus intermediate
between Lilium and Fritillaria; see Grove and Cotton, Suppl. to Elwes,
Mon. Lilium, p. 129. 1940.

Royle was born at Cawnpore in 1799 and educated at Edinburgh. He
was appointed assistant surgeon on the Hon. East India Company’s Bengal
establishment in March, 1820, became curator of the Saharunpur Botanic
Garden in 1823, and retired to England in 1831. From 1837 to 1856 he
was professor of materia medica at King’s College, London. Economic
Botany, as the pages of his “Illustrations” abundantly testify, was his
ruling interest, and he played a part in extending the cultivation of tea and
cotton in India. Bentham, de Candolle, David Don, Hooker, and Lindley
helped him with the classification and naming of his material. From 1851
until his death in 1858 he was secretary of the Horticultural Society of
London.

Linpiey Liprary,
RoyaL HorTICULTURAL SOCIETY,
Lonpon, ENGLAND.

488 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXIV

THE ARNOLD ARBORETUM DURING THE FISCAL YEAR
ENDED JUNE 30, 1943

As in the preceding two years, because of war conditions and the con-
comitant unsettled economic situation, no special appeal was made for extra-
budgetary support; yet the total gifts to the Arboretum were impressive, in-
cluding $925.00 for publication, $2346.27 for general unrestricted purposes,
$1050.00 from the Committee for Inter-American Artistic and Intellec-
tual Relations, to cover the expenses and emolument of Dr. Armando
Dugand, Director of the Instituto Biologico of Bogota, and a grant of
$500.00 from the American Philosophical Society for my use in connection
with the study of our accumulated collections of Chinese material. In
connection with the botanical survey of the Alcan Highway, discussed
below, Dr. H. M. Raup received a grant of $1500.00 from the Milton
Fund of Harvard University, supplemented by a grant of $500.00 from
the Bache Fund of the National Academy of Sciences and one of $600.00
from the General Purpose Fund of the American Academy of Arts and
Sciences. The latter organization also granted $150.00 to Professor
Rehder for his use in connection with the completion of his bibliographic
index. Five hundred dollars was received from the War Department to
enable us to prepare the necessary illustrations for a treatise on emergency
food plants for the Old World Tropics. An important gift to endowment
was the receipt of $50,000.00 in December from Miss Louisa W. Case of
Weston, together with her estate in Weston, consisting of 59 acres of land
with the buildings thereon, assessed at $84,000.00, but actually valued in
excess of that figure, for the buildings alone are insured on their appraised
value of $114,450.00. This gift is a memorial to her father, Mr. James B.
Case. While under the terms of gift the Weston property may be sold after
a period of three years and the proceeds added to the James B. Case fund,
it is our hope and desire that the Case estate be developed and maintained
as an adjunct to the Arnold Arboretum. The annual accretions to capital
under the terms of gift of the James Arnold and Charles Sprague Sargent
funds were credited to these funds as usual. The James R. Jewett and the
Vieno T. Johnson prizes were awarded in August in accordance with the
terms of gift.

In passing, the badly overcrowded condition of the library, and especially
of the herbarium, is again mentioned, although the library situation has
been somewhat alleviated through the transfer of certain forestry periodicals
to the Harvard Forest at Petersham on deposit. The overcrowded herba-
rium situation can be alleviated only by additional construction, and even if
funds were available for this purpose, which is not the case, an addition to
the Administration Building could not be accomplished at this time because
of the present restrictions on building material.

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 489

Staff. — The staff remains about the same as in the preceding year,
only one member of the technical staff having been drafted for military
service, this being Dr. C. E. Kobuski, who was granted leave of absence
when he was inducted into the service in October. Dr. F. P. Metcalf
resigned in April, 1942, when he was commissioned in the United States
Army. Dr. Armando Dugand, Director of the Instituto Biologico of
Bogota, Colombia, was appointed Research Associate during the period
that he was in the United States under the auspices of the Committee for
Inter- eres Artistic and Intellectual Relations, September 1, 1942 to
March 1,

Instruction. — Several staff members continue to codperate with the
Division of Biology of Harvard University in offering undergraduate and
graduate courses and in supervising the research work of candidates for
advanced degrees. The number of graduate students has decreased because
of war conditions, but the demand for certain types of undergraduate in-
struction has increased to provide for the needs of special groups of students
in residence at Harvard under the auspices of the Army and the Navy.
To meet this situation we have waived the condition of a half-unit course
every other year on the part of our staff members, and for the duration of
the present emergency our staff members may be called upon for more
course work supervision than would normally be the case.

Buildings and grounds, including horticulture. — Normal main-
tenance of all buildings has been provided for, the most important items
being essential furnace repairs and the installation of a new pipe line from
the water main on the Arborway to the Administration Building.

In assimilating large collections of living plants from various parts of the
world, it becomes necessary from time to time to re-check the living plants
already in cultivation to detect duplications, as well as those which are
incorrectly named. This was done with the lilacs last year, and this year
the difficult genera Weigela, Philadelphus, Deutzia, and Rosa were care-
fully examined. In many cases it was found that we were growing far too
many duplicates, and in an equally large number of cases it was found that,
on examining and comparing the horticultural varieties in flower, many
supposedly different varieties were identical, even though they had been
received from widely separated sources and had been growing here for
many years under different names. The checking and re-identification of
these groups took considerable time, but it is necessarily one of the
important functions of an arboretum. At present the number of species
and varieties of these groups represented in the living collections is Deutzia
63, Philadelphus 103, Rosa 241, and Weigela 56.

During the past year, 576 different kinds of plants were transferred from
the nursery to the living collections, many of these being entirely new
accessions. Including desirable duplicates, a total of 826 living plants was
added during the current year. In addition, approximately 300 crab apple

490 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

and oriental cherry seedlings were planted in the field near the Palmer
house and in the Walter Street tract.

Included among the many new plants added to the collections this year
was a collection of Clematis hybrids, the gift of Mr. Louis Vasseur of Milton,
Massachusetts. Mr. Vasseur has specialized in the growing of Clematis
hybrids for years and gave the Arboretum sixty of these in the fall of 1942.
These were planted on the six-foot woven wire fence at the rear of the
Aesculus collection. Another large collection of named varieties of Clematis
was given by the James I. George Company of Fairport, New York, special-
ists in Clematis. These were planted in the nursery in the fall of 1942
and came through the cold winter remarkably well, but since that time
many of them have unfortunately become seriously infected with disease.

The past winter was unusually severe, with temperatures at times well
below zero. While there were only six days when the temperature fell
below zero, as registered at the Arboretum greenhouses, more winter injury
occurred to plants than at any other time since the severe winter of
1933-34. A detailed discussion of the winter injury and the species and
varieties injured is included in Arnoldia 3: 25-36, 1943.

A survey was made of all the crab apples being grown in North America
at the present time, in conjunction with a committee of the American Asso-
ciation of Botanical Gardens and Arboretums, of which Dr. Wyman was
chairman. The study had as its objective the collecting of all available
information about the crab apples being grown today and listing this
information in convenient form together with complete bibliographical and
source data. A greater part of the work was done in the collections of the
Arnold Arboretum, and the report will be published in July, 1943.

During the year a total of 1713 living plants was received, chiefly from
various parts of the United States, but including a few from Canada, and
even three from England. In addition, 171 lots of scions and 59 packets
of seeds were received. Distribution of material to other institutions and
to individuals totaled 1542 living plants, 94 lots of scions, and an un-
recorded number of seeds.

Under a policy inaugurated last year, approximately twenty of the larger
nurseries in the United States and Canada were selected which were inter-
ested in new and rare plants and which were listed to receive living speci-
mens of new and rare varieties grown by the Arboretum. These plants were
not to be sold by the nurseries but were to be used as stock plants. This
arrangement was enthusiastically accepted by the nurseries, thus insuring
a properly controlled outlet for new or rare shrubs and trees of ornamental
value. The first year of operation proved to be highly satisfactory to
everyone concerned.

Due to the seriousness of the food situation, some of the experimental
land and some of the nursery space was assigned for garden work to twenty-
seven staff members and to individuals working in neighboring institutions,
and one area to a Boy Scout troop.

At the James B. Case estate in Weston, mentioned above, one hundred

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 491

and thirty crab apples and oriental cherries were planted late in 1942 in
one of the large fields and over 200 trees and shrubs were established in a
nursery. The grounds are being maintained in good condition, and it is
hoped that this estate may be maintained and developed as a permanent
adjunct to the Arboretum, near the city, yet remote enough to enable us to
accomplish various types of work without interruptions entailed because
of the urban location of the Arboretum proper

The Arboretum has had its share of gar troubles,” but the staff is
trying to carry on as well as it can under the circumstances. The curtail-
ment in gasolene and labor and inability to obtain new mechanical equip-
ment and repair parts for old machines are the chief causes for conditions
noted by the public. We are trying to maintain the grounds and the collec-
tions in good condition with the equipment and help available. At the
present time there is no one in the mapping and labeling department, both
the young men formerly employed for this work having left, one to engage
in war work, and the other to join the Army. Because the actual mapping
work was completed, it is possible to let some of the routine remain
dormant a year, but it does create many difficulties, some of which are
unforeseen. However, the difficulties encountered in the maintenance of the
living collections, though very real to us, are of the general type encoun-
tered everywhere during these unsettled times.

The War Effort. — This is not a discussion of war problems that the
institution faces because of shortages in labor and materials due to war
conditions. At first sight it would seem that a botanical institution could
contribute little to war purposes, and yet what we have been able to
accomplish is of considerable significance. Staff members have been at the
service of both State and Federal governments in supplying horticultural
and botanical information on camouflage problems, and a joint Camouflage
Research Committee was set up consisting of staff members of the
Arboretum, the Maria Moors Cabot Foundation for Botanical Research,
the Harvard Forest, and the Biological Laboratories of Harvard University.
The investigations undertaken by this group, in association with the United
States Army Engineer Board at Fort Belvoir, were not competitive, but
were cooperative, in that our investigations were supplementary to those
prosecuted elsewhere. In addition to supplying special lists of plants suit-
able for camouflage purposes to the Army officials, a series of experimental
studies on methods of prolonging the life of cut branches was initiated, and
this was done with both native and exotic (European and Asiatic) species.
Dr. Wyman has been a member of the Camouflage Committee of the
Massachusetts Committee on Public Safety since its inception.

In addition to the camouflage investigation work, much time has been
given to various emergency matters. Data have been freely supplied to
officials in various branches of the armed services regarding poisonous
plants and emergency food plants. In September, through the National
Research Council, because so many conflicting agencies were becoming

492 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XXIV

interested in the problem, I was drafted to prepare for the Quartermaster’s
Department, United States Army, a treatise on the potential food and
poisonous plants of the Old World Tropics. Work was commenced on this
about September 15, 1942, and the completed copy, with illustrations, was
sent to Washington on January 15, 1943. It was issued April 15, 1943,
in a very large edition, as Technical Manual 10-420, under the title
“Emergency Food Plants and Poisonous Plants of the Islands of the
Pacific,” pp. 1-149. fig. 1-113. It covers all of Polynesia, Melanesia,
Malaysia, and the Philippines, and for all practical purposes all of tropical
Asia. In addition to special work in this field I have had to go to Wash-
ington every two months to lecture on the same subject to each incoming
group of trainees in the intensive course on tropical medicine at the Army
Medical School.

Botanical Survey of the Alean Highway. — In the early part of
1943, with the announcement of the opening of the Alcan Highway, it
occurred to me that here was a real opportunity to accomplish some pro-
ductive field work in a hitherto little known area. I accordingly suggested
to Dr. H. M. Raup, who had conducted eight botanical field trips in
northern Canada, that it would be a good idea to plan for a trip along the
Alcan Highway perhaps in 1944 or 1945. In preparation for such a trip
Dr. Raup applied for a grant from the Milton Fund of Harvard University.
After the application was made it developed that the Joint Economic Com-
mittee, Canada-United States, was much interested in having the botanical
survey made at once, because certain data were needed by the local authori-
ties now. The Committee took up the matter with the military authorities
and secured not only the necessary permission but also their cooperation.
The National Museum of Canada is also codperating. As noted earlier
in this report, the Milton Fund grant of $1500.00 was made and $1100.00
was received from other sources. It was then decided to add a glacial
geologist to the group, and Dr. Charles S. Denny of Wesleyan University
was selected. He secured the necessary permit to be absent from the
University for the summer and further secured a grant of $900.00 from the
Penrose Fund of the American Geological Society to cover his traveling
expenses. Still later the services of Dr, Donovan S. Correll were secured
as assistant botanist, by providing funds to reimburse the Botanical
Museum for his salary during the time he would be in the field. The party,
consisting of Dr. and Mrs. Raup, their two sons, Dr. Correll, and Dr.
Denny, left Boston on May 31, and is expected to return about the middle
of September. Meagre reports received from the field indicate a most
successful summer campaign.

Cytogenetics. — The plant breeding work has resulted in new and
interesting types of Forsythia, lilacs, roses, and ornamental apples and
cherries. Several dwarf or compact types of Forsythia appear to be of
particular interest. Among the hybrid cherries, one of the segregates is a

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 493

semi-double-flowered form of the Prunus subhirtella type which is very
hardy and which remains in flower for two weeks. Of the several hundred
apple hybrids which have flowered, six have been selected for propagation
and further testing. Polyploid forms of Forsythia and Philadelphus have
been produced. One of the polyploid forsythias bloomed freely after the
severe winter, which destroyed the flowers of most species. This plant
has large flowers which are darker than those of the diploid species. The
polyploid Philadelphus has large flowers but the petals are thick and they
fall quickly. Seedlings of this plant may prove to be of value. Cytological
work has been limited to the continued study of X-ray effects on chromo-
somes and on the viability of seeds and seedlings.

Wood Anatomy. — Professor Bailey and Dr. Nast have continued their
collaboration with Dr. Smith in the study of woody ranalian families.
Intensive investigations of the floral and vegetative organs of the Degeneria-
ceae and Himantandraceae have shown that these families are closely re-
lated to the Magnoliaceae. The three families form a compact group
within the Ranales, being more closely related to each other, on the basis
of important morphological details, than any one of them is to other fam-
ilies. On the contrary, the Winteraceae exhibit no close relationship to the
Magnoliaceae either florally or vegetatively. Nor do they exhibit close
affinities to the Schizandraceae, Trochodendraceae, or other specific ranalian
families. The remarkable carpels of the Winteraceae rival their vesselless
wood in morphological significance. The palmately 3-veined mega-
sporophylls are adaxially folded or conduplicate and bear numerous ovules
on their morphological upper surface. In other words, the ovules are not
attached to the margins of a classical, involute, sealed sporophyll. The
conduplicate, open carpels of Degeneria and of the Section Tasmannia of
Drimys afford significant clues for re-interpreting the carpellary structures
of the Ranales, and in all probability of the angiosperms as a whole.

The Herbarium. — A total of 20,050 specimens was mounted during
the year, and of these 16,476 were inserted into the herbarium; the re-
maining were herbaceous specimens not kept in the Arboretum collections.
The herbarium now includes a total of 608,732 specimens.

The number of specimens received by exchange, gift, subsidy, purchase,
or for identification was 22,585. The greater part of these — 17,519 speci-
mens — was from North and South America, while the remainder may be
broken down geographically as follows: from Polynesia, 3,477; from India,
976; from Australia, 314; from Africa, eastern Asia, and Europe, 299.
Important acquisitions include about 3,000 specimens, mostly from Hawaii,
collected and given by Mr. Otto Degener, 2,807 specimens collected in
Cuba by Dr. Richard A. Howard, about 1,300 specimens collected in Idaho
by Mr. Arthur Cronquist, 1,047 specimens collected in Mexico by Dr. C. H.
Muller, 750 numbers, with duplicates, collected in Mexico by Mr. Robert
Stewart, and 2,734 specimens obtained over a period of several years by

494 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Mr. E. J. Palmer, representing cultivated plants growing in the Arboretum.

To other institutions the Arboretum distributed 36,152 specimens; these
were necessarily all sent to American institutions this year. Of this number,
26,925 specimens were sent in exchange, while 8,896 specimens were trans-
ferred to the Gray Herbarium; the remaining specimens were sent out either
as gifts or for identification by specialists. A total of 410 mounted illustra-
tions was transferred to the Gray Herbarium and the Ames Orchid
Herbarium at the Botanical Museum. Microfilm to the equivalent value
of 1,916 specimens was distributed under a special exchange arrangement.
The total number of specimens or their equivalent in mounted illustrations
and microfilm distributed by the Arboretum, therefore, was 38,478. Addi-
tional thousands of specimens were set aside for shipment to European
herbaria after the war.

Twenty-three loans, totaling 2,003 specimens, were made for study by
specialists in 15 American institutions. For study by members of the
Arboretum staff, 31 loans consisting of 2,280 specimens were borrowed
from 12 institutions.

A total of 2,037 cards was added to the catalogue of references to new
species and other important literature appertaining to woody plants, this
catalogue now consisting of 133,732 cards. The collection of negatives rep-
resenting types and other critical specimens now totals 4,211, as 73 nega-
tives were added during the year.

Routine herbarium work has been continued under crowded conditions,
only the most essential specimens being added to the general herbarium,
and the less necessary material being stored in generic order in cardboard
cases. Although this material is thus available to students, the need for
additional steel cases and space to place them becomes more acute each year.
Our accessions show a decrease from the figures of normal years, as expected
under the present international conditions. Because of this decrease, the
mounting department is now practically up to date. Herbarium work
has included routine incorporation of clippings, typed descriptions, and
illustrations.

Members of the herbarium staff continued their special studies, with the
result that numerous technical papers were prepared for publication, while
many identifications were made and various parts of the herbarium were
better organized. Professor Rehder brought the bibliographical supple-
ment to his Manual of Cultivated Trees and Shrubs nearly to completion,
this comprehensive work so far based on the library resources of the Arnold
Arboretum. There remain to be checked a number of references to litera-
ture not available here, to be searched for in other libraries. Dr. Smith, in
collaboration with Professor Bailey and Dr. Nast, continued his study of
ranalian families, also working on special groups of Papuasian, Polynesian,
and tropical American plants. Dr. Johnston has continued his work on
the flora of the intermontane plateau of northern Mexico, the first part of
his report being published, the second part in press, and the third part now
being prepared for the printer. Dr. Raup devoted much time to the com-

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 495

pletion of a report on his Mackenzie Mountain Expedition collections of
1939, which is nearly ready for publication, and to a study of Salix from
the Hudson Bay and Labrador Peninsula regions. The latter is in press,
but, because it became necessary for him to devote much time late in the
year to preparation for the Alcan Highway trip, mentioned above, com-
pletion of the former must await his return from the field. Dr. Kobuski
brought to completion his study of the tropical American members of the
genus Ternstroemia, and was granted leave of absence for military service
in October. Mr. Palmer continued to collect specimens of plants cultivated
in the Arboretum, also carrying on his studies of Quercus and Crataegus.
Dr. Allen’s studies of the American Lauraceae were extended, especially of
Mexico and Central America, and she continued her work on the genus
Halenia. Dr. Perry has further studied the Papuasian collections assem-
bled by the Richard Archbold Expeditions, and the greater part of this
valuable material has now been determined and reported on in this Journal,
although certain important groups are still under study. Dr. Croizat con-
tinued his studies of the families Cactaceae and Euphorbiaceae, giving
special attention to the genus Croton in North and South America. Dr. Li
studied and identified the material of many families of the large Chinese
and Indo-Chinese collections accumulated at the Arboretum, preparing
several papers for publication. My own work has been largely in connec-
tion with Rafinesque problems, especially the preparation of a comprehen-
sive Index Rafinesquianus, now in rough draft stage and to be checked
before the final copy is prepared. I have also supplied data on economic
plants to various representatives of our armed forces and have worked with
Dr. Li and Dr. Perry on Chinese and Papuasian botanical problems.

Linnaean microfilms. — A most important accession received in 1942
was a complete microfilm record of the Linnaean herbarium specimens and
other natural history collections of Linnaeus, together with records of his
manuscripts and of all books, including his own volumes, in which he had
made marginal annotations. This important record, in the form of a posi-
tive microfilm, came as a gift from the Linnaean Society of London. I
took some part in the negotiations with the Carnegie Corporation which
resulted in a special grant to the Linnaean Society for the purpose of
defraying the cost of making this record. The Council of the Linnaean
Society, in applying for the grant, offered to deposit a complete microfilm
record in some American institution. Asa matter of fact, it generously sent
two complete sets and later directed me to transmit one set to the Smith-
sonian Institution in Washington. There are about 160,000 exposures in
each set.

those parts appertaining to the herbarium material, we have had a
new negative film prepared from the positive and have arranged to have
enlarged prints made so as to have a graphic representation of each
herbarium specimen. When the task is completed we will then be able to
provide prints at cost to workers in other American institutions who may

496 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

have problems to solve in reference to the identity of Linnaean types. The
films appertaining to the Linnaean types of shells, insects, fishes, and other
animals have been deposited in the Museum of Comparative Zoology in
Cambridge.

Bibliography. — Dr. Verdoorn edited volumes 8, 9, 10, and 11 of his
new series of plant science books and vol. 7 of Chronica Botanica. As he
has been in touch with the Botanical Garden at Buitenzorg, Java, since
1930, he has been giving part time service to the Board for the Netherlands
East Indies, Surinam and Curagao in Washington as botanical advisor, and
in that capacity he organized the Central Depositary Library for the
Netherlands East Indies in New York. The objective here is to assemble
all foreign publications that normally would have been received by sci-
entific and technical libraries in the Netherlands East Indies, the plan
being to ship these to Java when conditions permit. In connection with
the preparation of the Index Botanicorum, card indices to all literature
dealing with botanical and horticultural history, bibliography, general
biology, and the history of botanical gardens are being prepared. Some
ten thousand references have been added to the standard forms on which
information regarding individuals is being compiled, and these data have
been carefully arranged for ready reference. Chiefly with the objective
of gaining more time for historico-botanical investigations, Chronica
Botanica will be discontinued as a serial and beginning with volume 8 will
appear in book form. The first issue of a new series devoted to the history
and methodology of botany and zoology is in press.

The Library. — At the end of the fiscal year the library contained
45,313 bound volumes, 13,322 pamphlets, and approximately 18,900 photo-
graphs. Accessions amounted to 191 volumes and 139 pamphlets. The
cards added to the periodical and author catalogue numbered 550, of which
150 contained bibliographical information, and 700 slips were added to the
files which supplement the printed author and subject catalogues of the
library. About 250 volumes have been loaned to other libraries and many
have been borrowed for use here, the University messenger service helping
greatly in such exchanges. The demand for photostats and microfilms
continued to be large, and prints of two of E. H. Wilson’s collections of
photographs numbering about 300 were made to order and sold. Exchanges
of periodicals with foreign countries were even further curtailed due to the
risks of shipping.

Atkins Institution of the Arnold Arboretum, Soledad, Cienfuegos,
Cuba. — Because of war conditions and restrictions on travel this unit has
been operated on a routine basis. It was not possible to assign graduate
students to the Atkins Institution for tropical experience even on a fellow-
ship or scholarship basis, partly because of the reduction in the number of
graduate students at Harvard University, partly because of restrictions on

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 497

travel. During the summer and autumn of 1942, further transplantings
were made in the palm section, and the temporary foot bridge across the
stream was replaced by a more permanent causeway. Because of the
unusually dry autumn and early winter, the reservoirs became dry, and
advantage was taken of this to remove the accumulated silt. There was
at this season an acute shortage of water, the small stream from which
water is pumped becoming almost dry. Two springs in the newer parts of
the garden were investigated, and this resulted in the development of
excellent wells of clear water sufficient for the garden and house require-
ments. This, however, made it necessary to reorganize the pipe lines
throughout the garden, which was in part accomplished. It is indeed
fortunate that this additional source of water could be developed within the
garden area. In the early spring some of the larger trees were transferred
from the nursery to their permanent sites in the garden. Maintenance has
been hampered because of lack of gasolene for the power mower. At the
request of the United States Department of Agriculture, a nursery plot for
Hevea brasiliensis was prepared and one shipment of 200 budded stumps
was received from Costa Rica. This shipment was unfortunately delayed
in transit, the resulting growth being poor. Over 400 pounds of Crypto-
stegia grandiflora seeds were supplied to the same organization and to the
Bureau of Economic Warfare. Conditions being what they are, only 174
packets of seeds were shipped in exchange and 79 packets were received.
The rainfall for the year was 49.29 inches, and the lowest temperature
recorded was 45°, on February 24, 194

Publications, — The usual numbers of the Journal were published, the
new and more compact format adopted at the beginning of 1942 permitting
the publication of more material per number. The new publication
Sargentia, the name honoring Dr. Charles Sprague Sargent, continuing the
Contributions from the Arnold Arboretum, received an auspicious begin-
ning with three numbers. The first of these: published in July, included
Dr. Smith’s study of the important Fijian collections assembled in 1940-41
by Mr. Otto Degener, on the Pacific cruise of the “Cheng Ho,” sponsored
by Mrs. Anne Archbold. In October Dr. Li’s comprehensive monograph of
the family Araliaceae in China was published. Number three, appearing
in January, contained a revision of the genus Sabia, by Dr. Luetta Chen,
and an extended discussion of the genus Ormosia in China and Indo-China,
by Dr. Chen and myself. A fourth number of Sargentia, with articles by
Dr. Raup and Dr. A. E. Porsild, is now in press. Arnoldia was issued as
usual, and its mailing list was revised. A bibliography of the published
papers by staff members and students follows.

498 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Bibliography of the Published Writings of the Staff and Students
July 1, 1942 — June 30, 1943

ALLEN, C. K. Halenia. In Cuatrecasas, J. eee 7 os flora de Colombia, V. Rev.
Kead: Colomb. Cienc. Exact. Fis. Nat. 5: 37-38.

———— Lauraceae. In Smith, A. C. Fijian shes Cae II. Sargentia 1: 34-35.
1942.

——— Studies in the Lauraceae, V. Some eve ern species of Beilschmiedia
and isa genera. Jour. Arnold Arb. 23: 444-4 942.

Asmous, V. C. Aleutian Islands: their military mies and _ historico-geographical
sketch. Rossiya 10(2494) : 3-4. 1942 (In Russian).

———— Eduard von ei (1815-1892) and his contributions to botany and horti-
culture in barnes Chron. Bot. 7: 199-200, portr. 1942.

——_— t Russian mission to Japan under the command of Lieutenant Adam
fcesen omens 10(2419) : 3-4. 1942 (In Russian

————— Inseparable friends: life, travels and c contemporatanin death of Lieutenant
Khvostov and Ensign Davydov. Rossiya 11(2621): 3-4. 1943 (In Russian).

———— K. A. Timiriazev (1843-1920): an appreciation. Chron. Bot. 7: 310-311.
1943.

————— K. A. Timiriazev, a great Russian physiological botanist. Rossiya 11(2641):
3-4. 1943 (In Russian).

——— Karelin (1801-1872) and Kirilov beet 1842), explorers of Siberia and
Middle Asi, Jour. Arnold Arb, 24: 107-1 1943.

a oe R.N.” Rossiya 112546) 33. 1943 (A jie in Rule
————— A new-comer to America from Soviet Union — kok-saghyz een ditilie.
lion). cans 10(2431): 3. 1942 (In Russian).
——— inine, malaria bie war. Rossiya 11(2576): 3. 1943 (In Russian).
ian explorers, I. Vladimir Atlasov, the conqueror of Kamchatka.
Rossiya 10(2477): 3-4. re (In Russian).
———— War activity of Soviet botanists. Rossiya mea _ 6. 1943 (In Russian).
— Willibald von Besser (1784-1842). ccna 151: 1943.

Atkins, I. M. & MancersporF, P. C. The isolation of isogenic lines as a means of
measuring the effects awns and other ere in small grains. Jour. Am. Soc.
Agron. 34: 667-668. 1942.

BatLey, I. W., Nast, . G. & a A. C. The family Himantandraceae. Jour.
Acoold tae — 190-206, 6 pl. 194:

— egeneriaceae, a new family of flowering plants from Fiji.
Jour. Acacia Arb. 23: 356- 365, 5 pl. 1942.

BRUMFIELD, R. T. Effect of colchicine aaa on the frequency of chromosomal
aberrations induced by x-radiation. Proc. Nat. Acad. Sci. 29: 190-193. 1943.

CroizaT, L. Amoenitates cactologicae. Desert 14: 114-117, 134-136. 1942.

— e cactus and its body. Cactus & Succ. Jour. 15: 37-42, 12 fig. 1943

——_—— Las especies sudamericanas del genero Croton, I-II. Comunic. a Mus.
Hist. Nat. Montevideo 1(2): 1-4, 2 pl. 1942.

— Euphorbiaceae. In Woodson, R. E., & Schery, R. W. an
toward a flora of Panama, VI. Ann. Missouri a, Gard 4. 29: 353-357.

Euphorbiaceae. In Smith, A. C. Fijian plant studies, II. ae 1:

46-53. 1942.

——— Grafting in and out. Desert 15: 37-39. 1943.

nal Porfiria and Obregonia, which and when? Desert 14: 85-87. 1942.
a eocereus peers (syn.: Binghamia Britton and Rose p.p.). Cactus
& Succ. ee 14: 145-148. 1942

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 499

Identification of Japanese maples. Am. Nurseryman 76(9): 5-7, 12 fig.

1942.
— Identifying Viburnums in winter. Am. Nurseryman arse 5-7, 8 fig. 1943.
s Binghamia valid? Cactus & Siice. Jour. 14: 126-1
Navajoa, a new genus in Cactaceae. Cactus & Succ. ae is: 88-89, 1 fig.
1943,

— New combinations and notes under Echinofossulocactus Lawrence. Cactus
& Succ. Jour. 14: 111-112. 1942.

—_— New and ae Euphorbiaceae from the tropical Far East. Jour. Arnold
Arb. 23: ears aes

——— milies. Phos ice Jour. 15: 64.

New et of Croton L. from New re ae Arnold Arb. 23: 369-376.

Ze
oO

1942.
A ae Rr of Euphorbia pteroneura. Desert 15: 21-23, 1 pl. 19
—_—__— n American apie with description of eleven new species.
Jour. wick hea Sci. 33: 11-20. 1943.
—— Novelties in American Bi ces Jour. Arnold Arb. 24: 165-189. 1943.
Paradichlorobenzene is worth knowing. sia aa 147-148. 1942.
Spellings and misspellings. Desert 15: 75-78. 1943.
——— Thalictrum ametrum Greene: an interesting ood case. Madrono
7: 1-4. 1943.
—— What isacephalium? Cactus & — Jour. 14: 169-172, 7 fig. 1942.
What is the trinomial typicus?, I. Bull. Torrey Bot. Clu b 70: 310-324. 1943.
———— & Mertcatr, F. P. Novelties in Chinese Phyllanthus. Lingnan Sci. Jour.
20: 193-197, 2 pl. 1942
Jounston, I. M. New ieee is from Mexico, V. Jour. Arnold Arb. 24: 90-98.
1943

———— Noteworthy species from Mexico and adjacent United States, I. Jour. Arnold
Arb. ae 227-236

— Publication fetes for ae , ae parts of the Pacific Railroad Reports.
Jour. Arnold Arb. 24: 237-242.

KEMMERER, A. R., Fraps, G. S. " ——— P. C. The relation between the
vitamin- A- meee see in corn and the number of genes for yellow color.
Cereal Chem. 19: 525-528. 1942.

Kosuskl, C. E. Oleaceae. " Smith, A. C. Fijian plant studies, II. Sargentia 1:
96-97. 1942.

———— Studies in the Theaceae, XII. Notes on the South American species of
Ternstroemia. Jour. Arnold Arb. 23: 298-343. 1942.

———— Studies in the Theaceae, XIII. Notes on the oar and Central American
species of Ternstroemia. Jour. Arnold Arb. 23: 464-478. 1942.

———— Studies in the Theaceae, XIV. Notes on the West Indian species of Tern-
stroemia. Jour. Arnold Arb. 24: 60-76. 1943

Li, H.-L. The Araliaceae of China. Sargentia 2: 1- a 14 fig. 1942.

———— Bamboo and Chinese civilization. Jour. N. Y. Bot. Gard. 43: 213-223,
6 fig. 1942.

———— Schizomussaenda, a new genus of the Rubiaceae. Jour. Arnold Arb. 24:
99-102. 1943.

MancetsporF, P. C. The origin of maize. Proc. Eighth Am. Sci. Cong. 3: 267—274.
1942,

Plant breeding as an applied science. Chron. Bot. 7: 235-236. 1942.

Merrit, E. D. Botanical survey of the Alcan Highway. Science II. 97: 574. 1943.

———— Emergency food plants and poisonous plants of the islands of the Pacific.
Washington. i-v. 1-149, illus, 1943. (War mig ap irae rae manual 10-420.)

Figments of the imagination. Science II. 97: . 1943.
The flora of Fukien Province, China. A review. — » IT. 96: 232-233.
1942.

500 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

—____— A generally overlooked Rafinesque paper. Proc. Am. Philos. Soc. 86: 72-90.
1942.
____ Microfilm records of the Linnaean Society of London. Science II. 96:
352-353. 1942

—_———_ Microfilm records of the Linnaean collections and manuscripts. Science II.
97: 110. 1943. ;

——_—_—— aes _— reproduction of rare technical publications. Science II.
96: 180-181.

oe i; na domesticas y sus relaciones con la civilizacion. Rev. Argentina
Agron. 9: 265-283. 1942

Sargentia: pes of explanation. Sargentia I: iii-iv. 1942.
The vegetation of Malaysia. Far Eastern Quart. 2: 66-76. 1943.

—_—_—_— ry, L Pandanaceae and Myrtaceae. In Smith, A. C. Fijian
plant studies, II. Sargentia = a 5, 74-78. 1942.

———— & Per ae papuanae Archboldianae, IX, oe XI, XII. Jour.
Arnold Arb. 23: 267- 297, ee 1942; 24: 34-59, 7 fig., 207-217. 1943.

Parmer, E. J. Quercus Prinus Linnaeus. Am. Midl. Nat. 29: sea 1943.

Raup, H. M. Adventures in a botany. In Peattie, R. The friendly moun-
tains. New York. 143-185. 2.

——_———— Trends in the piers of geographic botany. Ann. Assoc. Am. Geogr.
32: 319-354. 1942.

Reeves, R. G. & Mancersporr, P. C. A proposed pone! a in the tribe
Maydeae (family Gramineae). Am. Jour. Bot. 29: 815-817.

Renpver, A. New — yeeser and combinations Ei iedient of the
Ar nold Arboretum. Jou rnold ren 23: 377-381.

Sax, K. The effect of erie upon the i a x-ray induced chromo-
somal aberrations. Proc. Nat. Acad. Sci. 29: 18-21.

Smitu, A. C. The American species of Drimys. Jour. Asal Arb. 24: 1-33, 3 fig.
1943.

—__——— Cavendishia. In Cuatrecasas, J. Notas a la flora de Colombia, V. Rev.
Acad. Colomb, Cienc. Exact. Fis. Nat. 5: 38-39. 1942.

—— Fiji plant studies, II. eee results of the 1940-41 cruise of the
“Cheng ol Sargentia 1: 1-148, 5 fig. 1

—— A nomenclatural note on the 2 ee Jour. Arnold Arb. 23:
ieee loa? .

————— Studies of Papuasian Jee V. Jour. Arnold Arb. 23: 417-443. 1942.

axonomic i vig veh he Old World species of Winteraceae. Jour. Arnold

Arb. 24: 119-164, 6 fig. 1

VerpoorN, F. At the ne of the Botanische Zeitung. Chron. Bot. 7: 294-296.

43.

In memoriam J. C. Schoute. Chron. Bot. 7: 278-279. 1942.
Wyman, D. Berries for midwinter cheer. Home Garden 1(1): 44-4 43.
————— Broad leaved evergreens in good condition ees — foliage throughout the
winter (Oct. 1942 — April 1943). Arnoldia 3: 24
————— Euonymus. Arnoldia 2: 41-49, 2 pl. 194
Flowering shrubs along the hills. Home ie 1(6) : 87-90.
Foliage po of woody plants, April to September. Arnoldia A een 1942.
Hardy vines. Home Garden 1(5) 43,
——— Hot sDirciet care of broad leaved evergreens. Home Garden 1(6) : 35-37.
1943,
———— The naming of cmaprenne: varieties. oe 3: 9-13. 1943.
Planting vegetables. Arnoldia 3: ,
Shrubs for the new place. ae aie see 62-64. 1943,
Small trees for the small place. Home Garden 1(2): 58-60.
Winter foliage color of narrow-leaved evergreens. Arnoldia My ee 20. 1943.
Winter injury in the Arnold Arboretum, 1942-43. Arnoldia 3: 25-36. 1943.

1943] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 501
Staff of the Arnold Arboretum,
1942—43

ELcmMer Drew Merrit, S.D., LL.D., Arnold Professor of Botany and
Director

Joun GEorGE JAck, Assistant Professor of Dendrology, Emeritus.

ALFRED REHDER, A.M., Associate Professor of Dendrology and Curator of
the Herbarium, Emeritus.

JosepH Horace Favutt, Ph.D., Professor of Forest Pathology, Emeritus.
IrviNG WIDMER BalILEy, S.D., Professor of Plant Anatomy.

Kart SAx, Ph.D., Professor of Cytology.

Paut CHRISTOPH MANGELSDORF, S.D., Professor of Botany.

ALBERT CHARLES SMITH, Ph.D., Curator of the Herbarium.

IvAN Murray JOHNSTON, Ph.D., Associate Professor of Botany.
Hucu Miter Rauvp, Ph.D., Assistant Professor of Plant Ecology.

CLARENCE EMMEREN Kosuski, Ph.D.,* Assistant Curator of the
Herbarium.

DonaLp Wyman, Ph.D., Horticulturist.

ERNEST JESSE PALMER, Collector and Research Assistant.
CAROLINE KATHRYN ALLEN, Ph.D., Assistant in the Herbarium.
Lity May Perry, Ph.D., Technical Assistant.

LEon Croizat, J.D., Technical Assistant.

FrANS VERDOORN, Ph.D., Bibliographer.

ARMANDO DuGAND, Ph.D., Research Associate.

CHARLOTTE GeorcIA Nast, Ph.D., Curator of Wood Collections.
Hut-Lin L1, Ph.D., Technical Assistant.

JANET RYTHER SELLARS, A.B., Librarian.

VLADIMIR CONSTANTIN Asmous, A.B., Assistant Librarian.
Susan Detano McKe vey, A.B., Research Assistant.
CoNSTANCE MANSFIELD GILMAN, Business Secretary.

Louts Victor SCHMITT, Superintendent.

WILLIAM HENry Jupp, Propagator.

* On leave of absence for service in the U.S. Army.

1943]

INDEX 503

INDEX

Synonyms are printed in italics;

Abies coahuilensis, 332

— lasiocarpa, 3

oe 332, 334
332

eee tonkinensis, 366
Actinostemon anisandrus, 177
4

Adiantum Capillus- Veneris, 323
— tricholepis, 323
Adina Metcalfii, 454
Aegopogon cenchroides 403
Agatea nets 208
— salomon
Agropyron arizonicum 384
—sp

eecenicata , 384
ert avenacea: 84

385

1 monensis, 208
erate aie ma, 211

— macrophylla, 211
— Reineckeana, 211
Alyxia floribunda, 211
— fragrans
— Lamii, 212
— pugio, 211

subalpi

American Euphorbiaceae, Novelties in, 165
American Species of Drimys, The, 1
Ammoselinum Popei, 239

Amphania integrifolia, 63.

Andropogon barbinodis, 419

—cirratus, 41

— furcatus, 419

new names in bold-face type.

eins Gerardi, 419
— Hallii,
— hirtiflorus, 418

—_— as iesiuoas 418
eomexicana, 418

— ternarius, 419

— virginicus tenuispatheus, 418

Anemia mexicana, 324

Anisophyliuin emarginatum, 185

Aplotaxis gnaphaloides, 48

Arbor laurifoliae Magellanicae, 15

Archbold Expedition to N
Ferns of the Second, 440

Archboldianae, XI, XII, XIII, Plantae
Papuanae, 34, 207, 422

Argythamnia breviramea, 165

Aristida adscensionis, 39

Guinea,

— Wrightii, 400
Arnold Arboretum, 1942-43, Staff of the,
501

Arnold Arboretum during oe aoe year
ended June 30, 1943, The,

Karelin (1801-
1872) and Kirilov (1821-1842), Explor-
ers of Siberia and Middle Asia, 107

yi ao 309

— Lealii,

— eee 309

— resiliens, 309

504 JOURNAL OF THE ARNOLD ARBORETUM

Astragalus Grahamianus, 486
Astronia atro-viridis, 435

— novae-hannoverae, 436

— papetaria novo-guineensis, 434
Astronidium anomalum, 438

— sessilifolium, 437
Astrophyllum dumosum, 239
8

Aulacolepis epileuca, 84
Avena fatua, 385

Bacopa floribunda, 248
Bartey, I. CHARLOTTE G.

Co onparative Morphology of the
Winteraceae, I. Pollen and Stamens, 340
— & Cuarvottre G. Nast. The Compara-
tive ica of the Winteraceae,
I

he —

cae etn ae
man, a Valid Species, on
Bambusa lima, 78
Banisteria elegans, 222
«

223
Banisteropsis elegans, 222

—_——— erred ae
ulcherrima, 223
— leptocar rpa, 223

— Augustae, 5

— brachybotrys, 56

— Brassii, 4

— calliantha, 47
diffusiflora, 46

—m ia, 54

— Randiana, 47

[VOL. XXIV

aise Richardsoniana, 48

Bennettiodendron cordatum, 367
Benthamia japonica sinensis, 368
Bernardia amazonica, 166

165
bibliography of the published egere? of
ff and students, July 1,

—calocarpa, 213
— minima, 214

’
Blepharidachne Bigelovii, 382
Blepharoneuron tricholepis,
sie cal oliva fructu, 15

Boi

Shai — 309

Bonnaya reptans, 253

Botanical oe ‘af the Pacific Railroad
Reports, Publication Dates for the, 237

Bouteloua aristidoides, 409

— rigidiseta, 409
— scorpioides, 406

1943]

Bouteloua simplex, 405

— trifida, 408

— uniflora, 410

Brachiaria Meziana, 412

Brachypodium mexicanum, 377

— pubifolium, 7

Brass, III, Papuan Grasses Collected by
Si ae

British Guiana, New and Noteworthy

6
Brucea acuminata, 445
Bubbia, 1
—§ pol ae 141

— Howeana, 14
— idenburgensis, 150
— isoneura, 143

— Ledermannii, 149
— longifolia, 150

— sylvestris, 150

INDEX 505

Bubbia Cee 150
—W

hiteana, 145
Buchloé dactyloides, 411
26

Byrsonima incarnata, 221
— ahaha 222

—ru yy
Rey 222
Caconapea, 248

Callisteris texana, 95
Camassia Prog 239

— Gawle

Cana cashmeriana, 486
Canella

Capraria il

humilis, 246
Caraipa simplicior, 220
Cardamine africana borbonica papuana,
207

es ye
— papuan

Carpels, The P aneeiet Morphology of
the Winteraceae, II., 472

Cavendishia Cuatrecasasii, 466
ae 8
— rose =
— speci 467
Nei hes 417, 418
17

orus, 41

Centranthera Pirin 262

— hispid
Cephalom ote vem 440
Cephalomedinil,
Cerasus cornuta,
Ceratostema a ee 462
Chaetium bromoides, 416
Chaetochloa gibbosa, 416
— leucopila, 417
oasis acini 416
Chamaesyce, 182

B riana, 186
— caecorum, 187
— catamarcensis, 187

8
— emarginata, 185

506 JOURNAL OF THE ARNOLD ARBORETUM

pennies hirta, 188
— laeticincta, 188

87
— portucasadiana, 186
— potentilloides, 184
— Selloi, 184

tevidensis, 189
Chaptalia. gossypina, 486
Cuase, AGNES. Papuan Grasses Col-
lected by L. J. an Ill, 77
Cheilanthes aemula, 3
—alabamensis, 318

Chenopodina Moquini, 229
Chloothamnus sp., 77
Chloris oo 405

irgata, 405
riba cords ta, 486
Cladothrix puna carnosa, 232
Cleidion denticulatum, 166
— oblongifolium, 166
— prealtum, 167
— , 167
Clematis Archboldiana, 36
— glycinoides, 34
— Gouriana, 34

[VoL. XXIV

Clematis Gouriana malaiana, 34

— Vitalba Setane: 34
Clethra Liangii, 449
— polyneura, 44
Cleyera albo-punctata, 76

Comparative Morphology of the Wintera-
ceae, IT. ae The, 472

CopELAND B. Ferns of the Second
ee oe to New Guinea,

oo Na ena 368

Cortex Winteranu

Corylopsis spe es 445

Cottea pappophoroides, 382

CroizaT, LEON. a velties in American
Euphorbiaceae, 16

Cucumis pseudo- Colocynthis, 486

Cyclosorus gregarius, 440
Cynodon Barberi, 84

— Dactylon, 404

Cystopteris fragilis, 308

— — tenuifoli

Cytology and Speciation in the Genus
Populus L., A Study of, 275

Dactylostemon peti 177

Danthonia mexicana, 385
Dasylirion eee 92
ew:

Sioeeier arid poh aa 368
Dendrobium alpes
Dendrocalamus sai 78

— Sp.;
Dendropanax parviflorus, 367
3

Digitaria chinensis, 8

1943]

Digitaria ane 412

— violascens, 86

midtomee Morrisiana, 369
— potingensis, 369

siderophyllus, 450
Distichlis stricta, 379
— texana, 379

Ditaxis breviramea, 165 -
Doliocarpus s Savannarum, 218

Dans 10,

Drimys § Drimys,
— § ne 10, 22, 154

_ acutifolia, 139

amblericaulis 1

I , 14
— angustifolia, 32
— arfakensis, 136
— aromatica, es 128
— — alpina,

— confertifolia, 11
— coriacea, 140
— crassifolia, 152
—cyclopum, 139
—cyclopum, 140

INDEX

Drimys nin 140

— Deplan , 142
phl 1: 3 140

_.dipetala 124

, 20
— granadensis, 21, mak 29532

— piperita, 140

507

508 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Drimys pittosporoides, 135 Elymus canadensis, 384
— polymorpha, 1 Elyonurus barbiculmis, 4
— polyneura, 149 Elytranthe ampullacea Sea 364
— punctata, 1 —— puberula, 364
— purpurascens, 128 — — tonkinensis, 363
— reducta, 134 a cochinchinensis, 364
— reticulata, 137 — — Harmandii, 364
— reticulata, 137 —— ae 364
—retorta, 31 kinensis, 363
—yrivularis, 152 RR HT en eabrescens 365
— rosea, 136 Emmenopterys Rehderi, 100
—rubiginosa, 135 Ephedra antisypidlition 339

— semecarpoides, 144 — aspera,
— sororia, —com
— stipitata, 130 — pedunculata, 339
— Traversii, 159 — Torreyana, 338
—umbellata, 150 — trifurca, 3
— uniflora, 23 Equisetum laevigatum, 325
— vaccinioides, 134 Eragrostis caudata, 378
— vascularis, 33 — curtipedicellata, 377
— Versteegii, 134 — diffusa,
—verticillata, 136 — erosa, 378

— Vickeriana, 130 — intermedia, 379
— Vieillardi, 152 —lugens, 379
— Winterana, 15 — megastachya, 378
— Winteri, 13, — mexicana, 3
— Winteri, 15, 18, 23, 29, 32 —neomexicana, 378
— Winteri andina, 17 — obtusiflora, 377
— — angustifolia, 32 — Palmeri, 378
— — chilensis, 18 —reptans, 377
— — chilensis, 18 sp.,
— — — andina, 17 Erinus verticillatus, 246
—— condeittolia. 12 Eriochloa gracilis, 412
— — fernandeziana, 12 — puncta
— — granadensis, 22 Eritrichium ‘rasssepalum, 239
—— magellanica, 15 — pusillum, 239
—— Morenonis, 15 Erpetina radioed, 424
—-— punctata, 15 Eulalia g A

quinoensis, 17 —leptost

—— revoluta, 31 Euphorbia acerensis, 181
— — semiglobosa, 30 —alboma a
— xerophylla, 128 — anomala
Drimytaceae, 120 —aureocincta, 181
Drimytacées, 120 — caecorum,
— Drimytées, 120 chica vec, 179
— Exospermées, 120 ica,
Drymi. ——ventanicola, 179
Drymys, 8 — chamaerrhodos, 184
Dryopteris augescens enromaiis 308 — dictyosperma, 180
—normalis Lindheimeri —dilatata, 239
Echinochloa colonum, ne — Eichleri, 185
— crusgalli, 8 — emarginata, 185
— crus-galli, 416 — Fendleri, 239
— Wa — Hinkleyorum, 182
Echinodorus cordifolius, 376 — hirta, 188
Eddya hispidissima, 238 —hirta destituta, 188
Ehretia hispida, 238 — hirtella, 185

Eleusine indica, 404 — Huanchahana, 178

1943]

INDEX

a aca Huanchahana penazuelensis, Festuca oreoboloides, 79
Forsythia Vahl, Nomen Genericum Con-

a Spaced ciaiey 179
a, 179

8
— -— pilcomayensis, 178
—— se ey 178
Aeielice typic
_— ieee: 180
— Lorentzii, 185
— Mevyeniana, 184
— Hehe a 188
— Milii,
_- ae ae
+ qvaltiolis montevidensis, 189

88
—— portulacodes normalis, 179
— portulacoides acutifolia, 179

— sciadophila, 180
— Selloi, 18
— serpens

servandum, 48

r
Freziera undulat 6
Fritillaria rice 486

— parvifolia

Gardenia soph, 374
Garnotia Mezii, 8

Gaultheria ai 486
Gilia aggregata texana, 95
—calot 5

Giltbertia , ya

368
Goniothalamus chartaceus, 364
~ J+ Brass, ITT;

ee os ollected by L

Papuan, 77
Gratiola ciliata, 253

— tii, 239 — pusilla, io
Eu phorbiaceae, Novelties in American, 165 — reptans abe
Euphrasia, 26 — ruelloides, 2
—callosa, 273 — veroncfi, 253
—cucullata, 269 — viscosa

Excavatia, 213

m, oe teri
Ferns of the Second Archbold Expedition oe parviflora, 368

to New Guinea,
Festuca ligulata, ee
— nubigena, 79
—nubila, 79

rel
Helicia Hen 363
— pallidifiora, 363
— Tsangii, 363

509

510 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXIV

Helicia vestita, 363

—~— mixta, 444

Her pestis floribunda, 248

Heteropogon contortus, 420
0)

—re s, 8
Hilaria Belangeri, 403
— cenchroides, 403

194
Himantandraceae, The Family, 190

Hippomane biglandulosa, 175
— glandulosa, 1

Holcus redolens, 84
Hollrungia aurantioides, 210

Holostemma Brunonianum, 486

Hornemannia bicolor, 249

Ichnanthus vicinus, 86
Illicium, 12
“Illustrations

Bot of
Himalayan Mountains,” Royle’ s, 484

— ciliata, 253
— procumbens, 251
— ruelloide

Indo-China, Notes on the Flora of, 362

lonidium enneaspermum, 2
Isachne albens, 8

ngii, 456
Jatropha flabellifolia, 168
a 168
— Hoffmanniae, 168

a 7 ema 168
Pax

Paice poe M. New Phanerogams
from Mexico, V, 9

— Noteworthy Species from Mexico and
Adjacent United States, I, 227

— Plants of Coahuila, Eastern Chihuahua,
and pears Zacatecas and Durango,
I, W,

Be ea Dates for the Botanical
Parts of the Pacific Railroad Reports,

237
Juniperus Ashei, 337
—erythrocarpa, 337

—tetragona, 337

— — oligosperma, 337

Karelin, G. S., List of plants named in
honor of, 1

Karelin and Kirilov, Bibliography of
works by,

Karelin iad Kirilov, Biographies of, 115

Karelin (1801-1872) and Kirilov (1821-—
rath Explorers of Siberia and Middle

07

Kirilov oie 1842), Explorers of Siberia
iddle Asia, Karelin (1801-1872)

and, 107
am. I. P., List of plants named in
honor of, 1

KopuskI, CLARENCE E. Studies in the
Theaceae, XIV. Notes on the West
Indian Species of Ternstroemia, 60

Koeleria cristata, ser

ier coccinea,

Province, oe Notes on the

Lagurus ae
Lasianthus kw: idee nem 458

Laurifolia Magellanica cortice acri, 15

Lepiolomn cognatum, 412

1943 ] INDEX oll

Li, Hur-Lin. Notes on the Flora of Marsilea sp., 325

Indo-China, 362 ES: a
— Notes on the Flora of Kwangsi Prov- Mazus,

ince, China, 444 = goaden toll, 249
Ba i sear A New Genus of the — japonicus,

Rubiaceae, 99 Medinilla acutifolia, 425
eranee ahs 373 —albida, 427
— suspensum, 482 —anisophylla, 430
Lilium Thomsonianum, 486 — anomala, 425
Limnophila, 246 —arfakensis, 425

romatica, 246 — Bakeria 2

— fragrans, 246 — Brassti, 425, 426, 427
—ru a) —callia

sessiliflora, 247 — be
Lindernia, 24 —cauliflora, 423, 425
— anagallis, 252 — cephalantha, 433
— angustifolia, 252 — Erpetina, 424
— antipoda, 2 — exigua, 424
— ciliata, 25 — Forbesii,
— cordifolia, 252 — Hollrungiana, 423
— crenata, 2 — Ka , 429
— crustacea, 250 — lancifolia, 432
— hirta, 250 —1 , 426
— papuana, 251 — longistylis, 425
— pusilla, 251 — Lorentziana, 42
— pusilla, 250 — luraluen 430
— trichotoma, 252 — lysipetal
— viscosa, 251 nsfeldiana, 427
Lindsaea subtripinnata, 441 — Markgrafii, 42
Linociera eae illata, — nodosa, 425
Lonicera bracteata, 486 ae rea 423

Lophotocarpus calycinus, 376

— fluitans — quadrifolia, 423
Roranthis ee: 364 —rubescens, 433
— tienyensi —rubiginosa, 428
Luculia ee 4 — sessilis, 434
Lycurus Sa aeas — tenuipedicellata, 424
— phleoides, — tulagie
—_—— a 385 — vagans, 43
Macropiper methysticum, 353 — Versteegii, 424
— puberulum, 3 56 Melica alba, 380
Macrosolen cachinckinendi. 364 — montezumae, 380
Madhuca eC aias 368 —nitens, 380
Tsangii, 36 Mella,
Magallana, il — floribunda, 248
Magellania, 10 emecylon papuanum, 438
Magnoliaceae, 190 Merri.t, E. D. & L. M. Perry. Plantae
— Drimydoideae, 120 Papuanae peas XI, XII,
— Drimyoideae, 120 XIII, 34, 207,
— Illicieae, 120 Mexico, V, New Paracas from, 90
— — § Winterineae, 120 Mexico and Adjacent ieee States: I,
— Wintereae, 120 No Seater Species from,
Manihot orinocensis, 169 ape
Manisuris altissimus, 420 —san ee
Marsilea Fournieri, 324 ap ishing Meh potdiata 214
— minuta, 324 Microcarpaea minima, 251

— mucronata, 325 Microchloa Kunthii, 403

512

Microstegium gratum, 88
Milium globosum, 87
Milla Bryani, 90
Mimulus violaceus, 255
Mokofua delicatula, 62

— arizonica, 388

—asperifolia, 386

— biloba, 386

— pore 391

— crispiseta,

= eras ai
bia

—re , 388
—rigens, 392, 393
— rigida, 389

100

wangsiensis, —

. longipetala,

— Roxburghii, an

Myristicaceae and bake canna
of South ee Plants,
worthy, 4

Naias len 375

Nama biflorum Marshii, 96

Studies
xX. Note-

JOURNAL OF THE ARNOLD ARBORETUM

[VoL. XXIV

Nama Marshii, 96
—serpylloides confertum, 96
& I. W. Barey.

teraceae, I. Pollen and Stamens, 340

—& I. W. Barey. The Comparative
Morphology of the Winteraceae, II.
Carpels, 472

—A. C. Smiru, & I. W. Battey.

Family Himantandraceae, 190

New and Noteworthy Polypetalae from
British Guiana, 218

New Guinea, A Second oe of the
Scrophulariaceae of, 2

New Guinea, Ferns of . Second Arch-
bold Expedition to, 440

New Phanerogams from Mexico, V, 90

Nolina affinis, 9

— caudata, 91
pasirrencite oo 90

rantha,

_— ace ead

Nomen Genericum eee For-
sythia Vahl, 482

Notes on the Flora of Indo-China, 362

Notes on the Flora of Kwangsi Province,
China, 444

Noteworthy Species from Mexico and Ad-

The

— aurea, 3
— bryopoda, 317
— candida, 316

— cochisensis, 314

Novelties in American Euphorbiaceae, 165
Ochrosia calocarpa, 21
— glomerata
Old World Species of a Taxo-
nomic Notes - Ma
Olea cordatula,
Oleandra a 441

1943]

Ophiorrhiza kwangsiensis, 453
— succirubra, 453

Oplismenus hirtellus, 415
Oreiostachys producta, 77

Ouratea cataractarum, 224
—cernuiflora, 225

Pacific Island ‘Plants, II. Notes on the

ie for the Bee 'Parts of the,
237

Paederia pertomentosa, 458

Panicum agrostoides, 415

—— ee 414
— filipe
— Ghiesbreghtii 415

— — oigosanthes eeieeae 413
pedicellatum, 414

— vicinum, 86

— virgatum, oS

— zizanioides, 8
Pappophorum nae 383
eee 383

, 383
Papuan pantae Collected by L. J. Brass,

Beans minutiflora, 207
Parsonsia curvisepala, 215
— flavescen

— Helicandra, se

— lata, 216

— mollissima, 216

—crinitum, 413
— distichum, 412

INDEX 513

Paspalum Hartwegianum, 413

tramineum,
Passiflora cupiformis, 447
ien

k
— Papilio, 447
Payena cambodiana, 368
Pellaea atropurpurea, 311
i 9

PENNELL, FRANCI Second Sum-
mary of the Scrophulariaceae of New

ERRILL. Plantae
penteey beers XI, XII,

57
Periclymenum foi foliis laurinis, 15
-& E.

Phacelia infundibuliformis phanerandra,

96
— integrifolia arenicola, 97
usta,

ifolia, 98
Phalaris canariensis, 411
— caroliniana,
Phanerophlebia auriculata, 308
—umbona

— latisquama, 330
— Montezumae, 331

514 JOURNAL OF THE ARNOLD ARBORETUM

Pinus osteosperma, 329

1
Piper, Studies of Pacific Island Plants, Il.
n the Pacific Species of, 347

—Jatifolium, 349, 353

— latifolium

ni hf acnlivrad 347, 356
— — fasciculare, 35

ee. Pinion Archboldianae,

XII, XII, 34, 207, 422

Plants of Coahuila, Eastern Chihuahua,
and — Zacatecas and Durango,

1,.31,3
my ber we 466

—erectifolia, 82
—involuta, 377
— lunata, 81

— multmodis, 81

357
Piptochaetium fimbriatum, 396
9

Poa papuana, 84

6
Pollen and Stamens, The

[VOL. XXIV

Comparative
0

Morphology of the Winteraceae, I., 34

an]
S
~—
=
3
~~ Ss.
Sg 8
fen)
p.. |
R
=
&
oo

Polypetalae from British Guiana, New
8

and Noteworthy,
Polypodium eeythrolepls; 323

m,
~ plesosrum Bakes 324
— polyle pis, 3

ns Michauxianum, 324

u
Speciation in n the Genus

375
Primula ae. 486
— rosea, 486
Prinsepia utilis, 486
P

Study of Ps and

463
Pseudosmodingium anomalum, 235
332

Pseudotsuga aa
Pseudowintera, 153

— axillaris, i

— axillaris

— axillaris ake 158

Pteris cordata, 309
Ptilocalyx Greggii, 239

Publication Dates for the Botanical Parts
of the Pacific Railroad Reports, 237

Randia acuminatissima,
— leucocarpa,

— salictfolia, 456
Raugium, 482

1943 ]

Rangium se i 483
— a , 482

Ranunculus So reaae eal 38

— angustipetalus,

— us,

— coacervatus, 39

— habbemensis, 38

— lappaceus multiscapus, 40
i s, 4

— multiscapus, 40

— perindutus, 39

papuanus, 39

— perpusillus, 36

— scapigerus, -

— uncostigm

Reevesia Aa 366
— tomentosa, 446

a hiinie- a Forsythia ns Nomen
m Conservandum, 4

en
Racca ieee oo

Rheum spiciforme, 486
Rinorea aan 208

— salomonensis, 208
Roscoea ain, 486
—lutea ae

— pur 486

Races S station of a ey of the
Him n Mountain

Mh i am maine
Ge f the, 99

es enaetr 252

—an

A New

ae ke
Ryania aM sie Subulilora, 219
—to

Sacciolepis sonteaiat 87
Sagittaria longiloba, 376
sae floridana Chapman, a Valid Species,

os amphibia, 104
— astatulana

— longipes, 104

INDEX

515

SANDWITH, N. New and Noteworthy
Pol ie nae British Guiana, 218

aucup 7
ear oie 175
—— ham um, 173
— Bourge

if
— Cua eter, 172
— glandulosum, 176

— wolegs seis =e
—m ele 172
— Poeppigii, 173

7
Sarcosperma laurinum, 369
Satyria arborea, 469
Schizomussaenda, A New Genus of the

Schizophragma macrosepalum, 100
Schizostachyum lima, 78

— sp., 78
Scleropogon brevifolius, 383
New Guinea, A Sec-

°
=]
co

Sebastiania anisandra, 177
— huallagensis, 177

ee Summary of the Scrophulariaceae
of w Guinea, A,

ee lepidophylla, 328
macrathera, 326

cS
o
a=
Ft)
2.
wn
DH
e
ek
oO

— Scheelei, 417

516

Setaria asieoneg te 416
— villosissima, 41
oe n Wightianum tonkinense, 369
Sitanion Hystrix, 384
MITH, A. Studies of Pacific Island
Plants, ne _ on the Pacific Species
of Pipe
‘Studies . uth American Plants,
Noteworthy Myristicaceae and Vacci-

otes on oe Old World

Cuatmers. A Study of

Ge ie aaa in the Genus
Populus L.,

Sopubia mld

Sorghastrum een 420

Sorghum halepense, 420

Spartina Spartinae, 4 4

the Genus Populus L., A

75

Spiraea Northcraftii, 234

8
— “came 396
— Wright
Staff of the a Arboretum, 1942--43,

501

Stearn, Witt1aM T. Royle’s “Tllustra-
tions of the Botany of the Himalayan
Mountains,” 48

cael re on en 239

— Ciocary

Stemodia jaan 246

—verticillata, 2

vn
_

—eminens, 397

JOURNAL OF THE ARNOLD ARBORETUM

[VoL. XXIV

sic ne 396
— 397
— aera 4

— neomexicana, 396

— curviflora, 265

— Wallichii, 2

Studies in the en XIV. Notes on
the West Indian Species of Tern-
stroemia,

Studies of Pacific Island Plants, II. Notes
an the Pacific : of Piper, 347

Studies of South American Plants, X.
Noteworthy ‘aleicrrarean and Vacci-
niaceae, 460

Study of Bla and aaa in the
Genus Populus L.,

Styrax ie uinetias,

— oligophlebius, 451

Suaeda oe 230

— suffrutescens, 230
—— deto saa
— Torreyan

Se ohne ak

Symplocos anomala ee 452
— Delavayi, 371

— ligaticads, 67