JOURNAL

OF THE
ARNOLD ARBORETUM
HARVARD UNIVERSITY
C. E. WOOD, JR.

EDITOR

FRANCES M. JARRETT LAZELLA SCHWARTEN
ASST. EDITOR CIRCULATION

VOLUME XXxIx

CAMBRIDGE, MASS.
1958

Reprinted with the permission of the
Arnold Arboretum of Harvard University

KRAUS REPRINT CORPORATION
rk

1968

DATES OF ISSUE
No.1 (pp. 1-100) issued January 10, 1958.
No. 2 (pp. 101-212) issued April 14, 1958.
No.3 (pp. 213-378) issued July 14, 1958.
No. 4 (pp. 879-524) issued October 20, 1958.

Printed in U.S.A.

TABLE OF CONTENTS

STUDIES IN THE GENUS CoccoLoBA, V. THE GENUS IN HAITI AND
THE DOMINICAN Repusuic. By Richard A. Howard ................

A MonocraPHic Stupy oF THE WEST INDIAN SPECIES OF PHYL-
LANTHUS (CONTINUED). With four plates. By Grady Nl.
Webster

New Recorps or JAMAICAN FLOWERING Puiants, I. By Richard
A. Howard and George R. Proctor

A Note on THE IDENTITY OF THE GENUS BaLANosTrREBLUS (Mo-
RACEAE). By Frances M. Jarrett

A MonocraPHic Stupy oF THE WEST INDIAN SPECIES OF PHYL-
LANTHUS (ConcLuDED). With five plates. By Grady L.
VUZE S Re yan act ears oe errata sh. AE

THE EN or Some GENERA OF THE Lauracnan. By Carroll
ood, Jr.

A Synopsis oF THE GENUS HerrantA. With seventeen plates. By
Richard Evans Schultes

Tue GENERA OF THE Woopy RANALES IN THE SOUTHEASTERN
Unirtep States. By Carroll E. Wood, Jr.
SrATISTICS OF COMPOSITAE IN RELATION TO THE FLORA OF CHINA.
By Shw-ying Hu

STATISTICS OF COMPOSITAE IN RELATION TO THE FLORA OF CHINA
(ConcLuDED). By Shiu-ying Hu

A New Species or PHALERIA (THYMELAEACEAE) FROM NEw
Guinea. By Inly M. Perry

A Taxonomic Revision or Popocarpus, XI. THE SoutH PAciFIc
SPECIES OF SECTION POoDOCARPUS, SUBSECTION B. By Netta
Gray

ONTOGENY OF THE SPORANGIA IN XIPHOPTERIS SERRULATA AND
YRROSIA NUDA. By Kenneth A. Wilson .

Tue Drrector’s REPORT

BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND
PUDENTS: JULY: 1 LOS = UN pe SOY ODS sk rice inte nntdseede ae

STAFF OF THE ARNOLD ARBORETUM, 1957-1958
INDEX TO VOLUME XX XIX

With this issue, the Journal of the Arnold Arboretum becomes the re-
sponsibility of a new editor. It is with a deep sense of gratitude that I
speak for the editorial board, the staff of the Arnold Arboretum and other
contributors to these pages to express to Dr. CLARENCE E. KopuskI our
appreciation for a job well done.

After many years of careful attention to the time-consuming problems
which beset an editor, Dr. Kobuski has asked to be relieved of this re-
sponsibility in order to devote more of his time to the herbarium and to
his own research program. Dr. Kobuski (“K” to authors who have pub-
lished in the Journal) has taken an active part in the publication of the
Journal of the Arnold Arboretum for the past twenty-six years, except dur-
ing the period of his war duty, and has served as Editor since 1949. His
own high standards have been reflected in those of the Journal and pro-
vide a challenge for future editors,

Richard A. Howard
Director

JOURNAL

OF THE
ARNOLD ARBORETUM
VoLt. XXXIX JANUARY 1958 7 NUMBER l|

STUDIES IN THE GENUS COCCOLOBA, V.
THE GENUS IN HAITI AND THE DOMINICAN REPUBLIC

RicHArRD A. HowaArp

IN various Lists of the flora of Hispaniola published by Lindau (the
most recent monographer of the genus Coccoloba), Urban, Ekman, Barker
and Dardeau (Flore d’Haiti 99-100. 1930) and Moscoso (Catalogus Florae
Domingensis 168-171. 1943), forty-seven species and three forms of Cocco-
loba have been reported. Of these, thirty-two species and the three forms
were considered to be endemic. Few of the species are either old or known
only from the original collection or description. In general, the species
are represented by several collections, the majority of which are the result
of the meticulous collecting of Erik Ekman, who was encouraged by Igna-
tius Urban. There is no doubting Ekman’s keen eye and inclusive memory
for variations and locations. Again and again, his field notes indicate that
he had seen the plant before and that it was the same as the earlier col-
lection, or that a plant was different and “definitely not the same.”
Ekman’s success in relocating old species or in collecting additional ma-
terial was aided materially by his close co-operation with Urban. In
many cases Urban directed Ekman to seek in a specific area which he desig-
nated from Berlin, or suggested that he look for a certain variation in an
effort to recollect many of the older species. We are further indebted to
them for retracing the routes of earlier botanists and for collecting suffi-
cient new material to evaluate properly the older species. Ekman’s col-
lections were studied so promptly upon their receipt, that it appeared
that the specimens were literally just received when Urban published a
new name based on them. The rush to publish was continued after
Urban’s death by O. E. Schmidt who collaborated with Ekman, or rather,
worked over his collections. Ekman’s field notes indicate that he did not
always agree with Schmidt and wished for more careful consideration of
the entities involved.

Ekman recognized the general nature of the adventitious shoot in the
genus Coccoloba and indicated in his field notes the variations from plant
to plant and on single plants. Schmidt, in contrast, did not know the
plants in the field and failed to appreciate Ekman’s comments. As was
indicated in an earlier study of the genus Coccoloba in Cuba (Jour. Arnold

2 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

Arb. 30: 388-424. 1949), Schmidt’s failure to recognize adventitious
shoots, the dioecious condition in the inflorescence and the nature of the
pubescence led him to describe the two sexes of a single species under dif-
ferent names and to assign the adventitious shoots or the parent plants to
different species, varieties or forms. Over a period of several years I have
had the opportunity of spending several months collecting in the Dominican
Republic and in Haiti.1. During these collecting trips considerable em-
phasis was given to a field study of populations of the species of Coccoloba.
I have made careful collections of young plants, mature plants, adventi-
tious shoots and normal growth, as well as sun and shade foliage, to
represent the range of leaf variation both on individual plants and in
populations. Not all of the species reported from the island, or in fact,
collected by Ekman, could be studied in the field. elieve, however,
that the material which I have examined has made the conclusions drawn
in this paper more reliable than the previously existing work.

I now recognize twenty-four species and four hybrid populations from
Hispaniola. Fifteen species are regarded as restricted to the island. Three
of the four hybrid populations are new but unnamed and, of these four
hybrid populations, three are also to be considered endemic to Hispaniola,
but may be expected on other islands. Of the fifteen endemic species, two
(Coccoloba fawcettti and C. ceibensis) are regarded as representatives of
probable hybrid populations. Two other species are questionably distinct
from Cuban counterparts.

In addition to the fifteen endemic species, three (Coccoloba costata,
C. leonardit and C. wrightit) are also known from Cuba ae not from the
Bahamas, Jamaica or Puerto Rico. Five species (C. krugii, C. micro-
stachva, C. pubescens, C. swartgiu and C. venosa) have sone extending
eastward throughout Puerto Rico to the Virgin Islands. Of these, C.
krugii is also known from the Bahamas and from Jamaica and C. swartzii
from Jamaica. Coccoloba pubescens and C. venosa have been reported
from Jamaica but their occurrence there was questioned by me in an
earlier paper (Jour. Arnold Arb. 38: 105-106. 1957). However, they are
well known in Puerto Rico and the Lesser Antilles. Only two species
(C. diversifolia and C. uvifera) are regarded as widespread in the Carib-
bean area, but further study may place C. swartzi in the same category.
Hispaniola is indeed a center of speciation in the genus, but clearly not
to the extent recognized by earlier authors.

The following key to the species is artificial and, unfortunately, for a

*T am indebted to the trustees of the American Philosophical Society for a grant
from the Penrose Fund which made one of these trips possible. I am also indebted to

Mr. George Hamor of Hull’s Cove, Maine, ain of Barahona in the Dominican
Republic, for his hospitality and assistance in the course of my field work. My appre-
ciation is also gratefully expressed for the ae and the co-operation given by

many officials of the Dominican Seamer a the various officials of Compania Gre-
nada of the United Fruit Company; by Dr. José de Js. Jiménez and Dr. M. Canela
in services of value to this study. I am nee grateful to the directors and curators of
the herbaria cited in this paper for the long-term use of oe entrusted to their
care when various aspects of this problem took ial me.

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 3

few species requires the use of complete material, including flowers and
fruits. The majority of the species recognized are variable in vegetative
characters, as an examination of the species descriptions will verify. The
key given is applicable for all of the material which I have examined in
fruiting condition and for most of the material when in flower. The ke

is not applicable in all cases to either sterile mature shoots or adventitious
shoots. Abnormal variations on mature and adventitious shoots, such as
fasciations, pathological anomalies and contortions have not been included
in the key. Such specimens are common in herbaria and have been anno-
tated, but are relatively infrequent in the field.

Many of the characteristics employed in the key should not be used in
identification without considerable experience with the group. I have
avoided the angles of departure of the primary veins which former
monographers have used, but have used the prominence, curvature, bifur-
cations and reticulations of the primary and secondary veins individually
or as patterns. I have introduced several new characters, such as the
swelling of the nodes, the position of the base of the petiole in relation
to the ocrea, the length of the pedicels in relation to the length of the
ocreolae, the nature of the apex of the mature achene and the associated
aspect of the lobes of the fruiting perianth. Morphological studies are
needed on the bead-like swollen nodes which occur in a few species of
Coccoloba. Swollen nodes are characteristic of the family Polygonaceae,
but the exaggerated development of these in Coccoloba has not been in-
vestigated. The nodes are woody and extremely hard when dry. The pith
in the swollen section is not enlarged. It is not clear whether the develop-
ment of the nodal swelling is from the shortening of cambial initials or
from a stimulated development of additional cells.

In most species of Coccoloba the petiole arises from the base of the ocrea.
In a few species, the base of the petiole or the base of the abscission layer
is a short distance above the base of the ocrea, as indicated both by vascu-
lar pattern and superficially by the color change between internodal stem
tissue and the ocreal tissue. The relative position is readily determined in
adventitious or vegetative shoots and can be seen on fertile specimens after
the leaf has fallen. This characteristic has been checked in the field
and appears to be a reliable one, since it shows no variation on individual
plants or in populations.

In all but a very few species, the length of the pedicel is constant from
the time the flower opens until the fruit is formed. The thickness of the
pedicel varies, being much stouter when a fruit is developed than when
the pistillate flowers do not form fruit. Staminate inflorescences show little
thickening of the pedicels with age unless sterile fruits are produced.

The apex of the achene can be uniformly obtuse, acute, or constricted
to form a rounded knob. When the last is the characteristic shape, the
perianth in fruit consists of a fleshy hypanthium surrounding the body
of the achene, the lobes of the perianth forming a crown around the knob.
The latter condition has been referred to as ‘“‘coronate.” If the achene
and the perianth are coronate, the perianth lobes may remain small and

4 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

free, or may become fleshy and imbricated around the knob. Since the
crown is prominent very early, this character can be used on immature
fruiting specimens, as well as on fully mature fruit. The prominence of the
vascular supply in the fruiting perianth has been mentioned in the species
descriptions but has not been used in the key to the species. The number
of vascular bundles indicated in dry fruits as ridges and grooves seems
to be a reliable indicator for a particular species. This characteristic must
be verified when the fruit is dry and is, therefore, difficult to check in the
field. The bundles are not evident when the fruit is fresh; however, species
which later will have prominent bundles in the hypanthium are generally
less fleshy and, therefore, less edible when sampled in the field. Since
most species of Coccoloba which I have eaten have had extremely astringent
fresh fruits, there is no particular pleasure in checking for this character-
istic.

The number of flowers at each locus or node of the inflorescence is re-
lated to the functional sex of the flower. In most species studied, the func-
tionally pistillate flowers are borne singly while the functionally staminate
flowers are in clusters of two to five flowers. The number of flowers has
been given unjustifiable emphasis by Lindau, Urban and Schmidt. I do
not feel this should be used. Another characteristic used by Lindau in his
keys was the condition of exserted versus included stamens. This, too,
proved to be associated with functional or pollen-producing stamens, in
contrast to the included sterile, rudimentary or abortive stamens. I have
been unable to find any characteristics of diagnostic significance in the
flowers of Coccoloba as the genus occurs in the West Indies. The flowers
are small and variations in the size and shape of the floral parts appear
to me to be too insignificant to be of real value

Specimens are cited with standard abbreviations given for the herbaria
as cited in the third edition of Index Herbariorum. The provinces cited
for locations in the Dominican Republic are those used on the Esso
Standard Oil Company “Mapa de la Republica Dominicana” prepared
by the General Drafting Company and copyrighted in 1955. Province
boundaries on this most recent map vary considerably from earlier maps
available. The provinces are listed in alphabetical order. For Haiti, five
departments are recognized and listed in geographic order from northwest
to southwest. Navassa Island, formerly a United States possession, has
been turned over to the government of Haiti and is considered as associated
with the Département du Sud.

The species are described following the key and are listed in alphabetical
order.

KEY TO THE SPECIES OF CocCOLOBA IN HISPANIOLA
A. Perianth lobes exceeding the sae aes in length, dominant in fruit;
ocreolae increasing in size from flowering to fruiting condition; leaves
eae to membranaceous, rarely Seen.
B. Leaves oblong-lanceolate to elliptic, longer than broad, apex acum-
A, Geen carne fe oe saree O48 on oh baa . venosa.

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 5

BB. zone orbicular to reniform, as broad as long, apex rounded to emar-
inate.

a Leaves of normal shoots 3 X 3 to 9 X 9 cm. long and broad; petioles
arising from the base of the ocreae; inflorescence rachis 5—7 cm. long.
Be Pa Aa at Aes Hot AV ack eA a oe Rea Bierce C. leoganensis.

CC. Leaves of normal shoots 0.2 X 0.2 to 1.1 X 1.0 cm. long and broad;
petioles arising from above the base of the ocreae; inflorescence
TACHIS AO 1 Orem ORD ota pa elena ncn C. subcordata.

AA. Perianth lobes shorter than the hypanthium in fruit, imbricate or coronate,
leaves generally coriaceous and not membranaceous.
D. Leaves tipped with a spine or a cartilaginous point.
E. Leaves cordate, the lower leaf surface with a conspicuous reticulum
ORAM CII Sit Mee Bere Sie Ateche Ale a avai mela te getiges ea ane nge . fawcetti.
EE. Leaves ovate, oblong or elliptic, not cordate; lower leaf surface not
conspicuously reticulate.
F. Leaves oblong or elliptic.
G. Primary veins 10-20 pairs, not raised on either surface and
m.

long; fruit rounded at the apex, not coronate, the base of the
aoe attenuate but not sterile or corky; terminal Bone of leaf
Ta One eS Hal: POM CCCs meg ck he < caiee ait ne C. flavescens.
GG. Primary veins slightly elevated or evident below, 3-5 pairs;
inflorescence rachis 5-15 cm. long; leaf blade terminated by
a cartilaginous point, not sharp; fruit uncertain. C. hotteana.
FF, Leaves ovate in general outline, broadest below the middle, vena-
tion evident and slightly raised on both surfaces; fruit coronate,
the base sterile and corky.

H. Primary veins generally 2-4 pairs, the lower two commonly
separate from the others, the veins reaching to the margin be-
fore bifurcating and anastomosing; leaves generally 1-2 cm.
long, uniformly acuminate from the middle to the apex.
be lok 2° cain coc aie RI MI aI eh sD Se Ce AMEN

"ably separate from the pisene the veins arcuate and bifurcat-

mar
constricted above the middle and then acuminate i the apex.
tt sa Kh ANNO Aa erty rh Niet et PN ah ote, Mast C. fuertesit.
DD. Leaves not spine-tipped.
I. Leaves with a conspicuous reticulum of raised veins and veinlets
on the lower surface; perianth lobes imbricate in fruit, not coronate.
J. Leaves of normal shoots 1-3 cm. long; inflorescence generally
shorter than the leaves.
K. Leaves of normal shoots as broad as or broader than long,
apex rounded-truncate to emarginate. .......... C. picardae.
KK. Leaves of normal shoots longer than broad, apex apiculate to
acuminate, rarely rounded but never subtruncate or emar-
TTC ae re rt ret eT a ke C. pauciflora.

JOURNAL OF THE ARNOLD ARBORETUM _ |[VoL. xxxIx

JJ. Leaves of normal shoots larger, 6 X 4 to 50 & 80 cm. long and
broad; inflorescence generally exceeding o leaves.
L. Leaves longer than broad; fruit ovoid.

M. Leaves rounded and generally asymmetrical at the base
with one lobe usually agers the petiole, the blade com-
monly bullate between the veins. .......... C. ceibensis.

MM. Leaves normally rounded and penne at the base, occa-
sionally cuneate, the blade not bullate. ...... C. wrightiz.

LL. Leaves generally as broad as or broader than long.
N. Fruit globose to ovoid; leaves generally pilose. C. pubescens.
NN. Fruit obovoid, narrowed to a stalk at the base; leaves at
most puberulent. ............. C. uvifera X C. pubescens.
II. Leaves without a conspicuous reticulum, primary veins alone con-

spicuous

O. Flowers and fruits sessile or the pedicels short and not exceeding
the ocreolae in fruiting condition; perianth lobes coronate in fruit.
P. Veins straight, not conspicuous, arcuate before bifurcating
more or less equally and anastomosing near the margin; nodes
swollen; petioles arising near the apex of the swollen nodes but

nee the base of the ocreae; leaves of normal shoots 3.5 « 5
7X 4 cm. long and broad. ............. C. microstachya.

big oa ne arcuate ascending, the terminal dichotomies unequal;
nodes not conspicuously swollen or bead-like.

Q. Petioles arising slightly above the base of the ocreae; fruits
spindle-shaped, nearly twice as long as thick; leaf bases
usually asymmetrical. ... it

QQ. sop arising from the ee of cae ocreae ; ‘fruit globular

ate, scarcely longer than broad.
a tae shiny when dry; fruit globular, 3 mm. diameter.
C. samanensis.

RR. Leaves dull when dry.

S. Leaves of normal shoots obovate to obovate-elliptic,
broadest above the middle, uniformly acute to rounded
at the apex, not abruptly constricted; blade a dull
brown when dry; fruit globular. ....... C, albicans,

5. Leaves of normal shoots ovate to elliptic, broadest
below the middle, usually abruptly narrowed above
the middle and acuminate to the tip; blade turning
black on drying; fruit ovoid. _ C. swartsit.

OO. ste and fruits borne on pedicels wen ed the ocreolae

n length.

gp)
ie)

T. Leaves of normal shoots generally orbicular and as broad as
long or broader.
U. Branchlets with conspicuous swollen nodes, these often ap-
pearing moniliform.
V. Leaves drying black; petioles arising from above the
base of the ocreae; branchlets alee pubescent.
C. nodosa.

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 7

VV. Leaves not turning black on drying; petioles arising

from the base of the ocreae; branchlets pale puberu-

buchit.
UU. Branchlets terete, the nodes not conspicuously swollen.

W. Perianth lobes and achene coronate in fruit.

X. Inflorescence rachis 1-2.5 cm. long. .... C. buchii.

XX. Inflorescence rachis 10-20 cm. long. .... C. costata.

WW. Perianth lobes imbricate in fruit, the achene not coron-

Y. Leaves 6 X 8 to 13 X 18 cm. long and broad or
larger; venation not conspicuous; fruit obpyriform,
C. uvifera.

VY. Leaves 1.1 X 1.2 to 2.5 X 2.5 cm. long and broad;
fruit 3-4 mm. long.

7. Inflorescence rachis 0.2-0.5 cm. long; branches
generally appearing to be arranged in one plane;
venation of leaves reticulate and conspicuous on
both surfaces when dry; fruit round in cross

G:

SEC LIOLIM Ee ere Re rergr ery fet arama picardae.
ZZ. Inflorescence rachis 5-8 cm. long; branches geni-
culate, not appearing to be in o ena-

fruit strongly 3-angled in cross section. C. krugi.
TT. Leaves of normal shoots not orbicular, longer than broad.

a. Inflorescence rachis short, less than 3 cm. long.

Iw SET O ROM ALC sites ane eben eeesucu ty. an Oren wee C. buchii.

bb. Fruit not coronate, perianth lobes imbricate.
c. Leaves ovate, broadest below the middle, cordate at
HEM ASE ee ee ee ene ee ot ee C. krugit.
cc. Leaves obovate to obtriangular, broadest above the
middle, narrowed or rounded at the base.
d. Leaves rounded-truncate to emarginate at the apex.
Pe es oat a ee C. picardae.
dd. Leaves normally apiculate to acuminate, rarely
rounded but never subtruncate or emarginate at the
C. pauciflora.
aa. Inflorescence rachis normally 5-20 cm. long.
e. Flowering and ee pedicels short, exceeding the ocreo-
lae but rarely twice ng.

{. Fruit sub-coronate at the apex; ocreae, petioles and
rachises puberulent when young, the hairs generally
persisting; leaves of normal shoots generally 7 X 5 cm.
long and broad or larger; blades dark green when fresh
and golden to dark brown when dry. ...... C. costata.

_ Fruit not coronate at the apex; ocreae, petioles and
rachises glabrous; leaves of normal shoots generally

long)
loan)

8 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXXIX

4+ X 3 cm. long and broad or smaller; leaf blades pale

greenish-tan in color when fresh or dry. ..... C. krugii.

ee. Flowering and fruiting pedicels conspicuous, two to sev-
eral times the length of the ocreolae

g. Perianth lobes and achene sub-coronate when mature;

primary veins of leaves conspicuous and with second-

ary venation forming an elevated and conspicuous

reticulum on both surfaces when dry... C. wrightii.

gg. Perianth lobes imbricate, the achene rounded at the

ex, not sub-coronate; primary veins evident, sec-

ondary venation inconspicuous, not elevated and reti-

culated on the lower surface.

h. All parts glabrous... C. diversifolia.

hh. Ocreae, petioles and inflorescence rachises puberu-

lent or pubescent at least when young. C. hotteana.

Coccoloba albicans Ekman in Schmidt, Fedde Rep. Spec. Nov. 27: 103.
1929.

Small, medium or large tree (fide Ekman); branches terete, striate or
canaliculate, light gray, glabrous: ocreae 6-8 mm. long, stiff, glabrous,
cleft at the apex, frequently splitting at maturity and appearing as two
ovate-lanceolate acuminate stipules; leaves of normal shoots with petioles
4—6 mm. long, glabrous, arising from the base of the ocreae: blades obo-
vate to obovate-elliptic, 4 & 2.5,5 & 4, to 5.5 & 3.5 cm. long and broad,
coriaceous, glabrous, the apex rounded, rarely bluntly apiculate, the base
rounded, the margin slightly revolute; midrib and primary veins im-
pressed above, prominent below, the primary veins 6-8 pairs, arcuate,
anastomosing conspicuously near the margins, the ultimate venation evi-
dent below but not above; leaves of adventitious or fast-growing shoots
with ocreae 1—1.5 cm. long, the petioles 1—1.2 cm. long, the blades obovate-
elliptic to elliptic-lanceolate, rarely ovate-elliptic, 6.5 & 5, 8.5 x 6, to

5 X 5 cm. long and broad, otherwise the same; inflorescences single or
aggregated as 3 or 4 racemes, terminal or terminal on axillary shoots,
to 13 cm. long, the basal ocreae to 1 cm, long, the rachis angular, glabrous;
flowers sessile, the staminate flowers 1—3 at each locus, the pistillate flowers
I at each locus, the bracts about 0.5 mm. long, ocreolae membranaceous,
I-1.5 mm. long, expanding and splitting after flowering; hypanthium to
1 mm. long, the perianth lobes ovate to suborbicular, 1.8 x 2.0 mm. long
and broad, the functional stamens 1-1.5 mm. long, the rudimentary
Stamens less than 0.5 mm. long, the fertile pistil 2.5 mm. long, the ovary
triangular, the rudimentary pistil about 0.5 mm. long; fruit sessile, ovoid
to globose, 5 mm. long, 4.5—5 mm. in diameter, the perianth lobes slightly
coronate, the achene light tan in color.

DISTRIBUTION: Endemic to Haiti.

Haiti. Depr. pu Sup: Massif de la Hotte, Les Roseaux, Nan-Patates, Ekman
H-10693 (s-lectotype, us); Les Roseaux between Nan-Patates and Alnette, Ek-

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, Vv 9

man H-10720 (B, 8); Massif de la Hotte, Morne Rochelois, Miragoane at Quaert-
Chemis, Ekman H-9206 (s), H-7936 (a, s, US); Mirago4ne on trail to Morne
Rochelois, Eyerdam 515 (F, GH, NY, US), 519 (A, F, GH, NY, US)

It is not clear to me whether Schmidt or Ekman compiled the original
description of Coccoloba albicans. Although the species was published by
Schmidt, he gives credit to Ekman for the taxon as a new entity. Various
herbarium sheets bear labels indicating as author either Ekman or Ekman
and Schmidt. In any case, leaves from fast-growing shoots and those
from shoots of normal growth have been combined in the description,
with a resulting lack of clarity. The description above distinguishes be-
tween the branches which are mature and those which are terminal shoots
or adventitious branches. Among the specimens cited are both flowering
and sterile adventitious shoots.

The original publication cites Ekman H-10693 as the type. However,
the collection 10720 in the Berlin herbarium bears the annotation ‘“‘typus”
in Urban’s handwriting, while the Stockholm specimen of the same num-
ber has a printed label indicating that is the type. A specimen of the
number published as the type is not in the Berlin herbarium and I have
chosen to select the Stockholm specimen of Ekman H-10693 as a lecto-
type. This is the correct number as published and is better material than
is the eee collection cited in the original description and labeled as
the rele

wo Sie Ekman H-7936 and H-9206, were originally labeled
ae ee albicans Ekman, forma.’’ These are the smaller-leaved mature
branches, while the collection Ekman 10720 labeled ‘“‘typus” by Schmidt
consists primarily of adventitious shoots.

The collections cited in the original publication are from staminate
plants. The collection Ekman H-7936 labeled “forma” bears fruit. Pis-
tillate flowers are seen on the Eyerdam collections. In the original de-
scription Schmidt refers to the punctations on the lower leaf surface.
These are blocked stomata and their adjacent cells which dry darker
than other areas of the lower mesophyll.

Coccoloba buchii epee Ark. Bot. 20A(15): 32. 1926; Fedde Rep.
Spec. Nov. 24: 75.192
Coccoloba revoluta Leonard, Jour. Wash. Acad. Sci. 17: 66. rae

Coccoloba tortuensis Ekman & Schmidt, Ark. Bot. 20A(15): 32. 1926.
Coccoloba eae Ekman, Bull. Estac. Bot. Moca, Ser. 5 17: 10. 1927
(nomen.

Small tree with numerous spreading branches; branches terete, puberu-
lent to pilose, the nodes slightly tumid; ocreae membranaceous, 3-7 mm.
long, oblique to nearly bilobed at the apex, puberulent to pilose; leaves
of normal shoots with petioles 3-4 mm. long, puberulent to pilose, at
least on the adaxial surface, arising from the bases of the ocreae, the
blades ovate to elliptic-orbicular or reniform to obovate-elliptic, 2 x 1.5,
3 & 25 cm.,, thin, coriaceous, puberulent above, glabrous below, the apex

10 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxix

rounded or obtuse, the base subcordate, rounded or occasionally narrowed,
the margin entire, recurved; midrib prominent on both surfaces, the
primary veins 5 or 6 pairs, straight or slightly arcuate, strongly recurved
and anastomosing near the margin, the secondary venation minutely reticu-
late above, coarsely reticulate below; leaves of adventitious shoots with
petioles 4-5 mm. long, the blades elliptic, elliptic-lanceolate to ovate-
lanceolate, 4.5 & 3, 8.5 & 3 to 11.5 & 4.5 cm. long and broad; inflores-
cences terminal on short lateral shoots 1—2.5 cm. long, the rachis puberulent
at the base or on the lower portion, becoming glabrate; staminate flowers
1—3 at each locus, pistillate flowers borne singly at each locus, the bracts
and ocreolae membranaceous, to 1 mm. long, the ocreolae tightly cylindrical
around the pedicels, the hypanthium less than 1 mm. long, the perianth
lobes ovate, 1 mm. in diameter, the fertile stamens to 1 mm. long; fruiting
pedicels 1—1.5 mm. long, glabrous, the fruit ovoid, 5 mm. long, 4 mm. in
diameter, the base rounded, the perianth lobes subcoronate, the achene
brown

Local NAME: Papelite (H).

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. INDEPENDENCIA: near Puerto Escondido, Howard
12150 (GH). Prov. LipeRTADOR: between Restauracién and Banica, Howard 12554
(GH). Prov. Monte Cristr: El Morro, Ekman 13140 (a, s, Us, number desig-
nated by Ekman and Schmidt as type of Coccoloba ciferriana), Howard 12533
(GH), Howard 12535 (GH), Jiménez 1370 (GH). Prov. PuERTO PLATA: Arroyo
Francés, Ekman 14399 (s, us). Prov. SaMANA: Los Haitises, Boca del Infierno,
Ekman 15439a (s), 15439b (s, us), 15439¢ (s), 15439d (s). Prov. SANTIAGO:
Las Lagunas at Arroyo Arrenquillo, Ekman 16078 S).

ait. Dept. pu Norp Ouest: Moustique Mts., Bassin Bleu, E.C. & G.M.
Leonard 15235 (Ny, US), Port de Paix, Ekman H-3646 (8, s, us), Presquile du

ord-Ouest, Port de Paix, Haut Moustique, Ekman H-3647 (s), Isle de la
Tortue La Vallée, Morne Barranca, Ekman H-4107 (n-type of C. tortuensis.
s, US), H-4308 (B, s). Dept. pu Norv: St. Michel de lAtalaye, E.C. Leonard
8499 (B, GH, Us-type of C. revoluta), 7244 (ny, us), Massif du Nord, Gros
Morne, Morne Chabre. Ekman H-5024 (s), Massif du Nord, Gros Morne, Morne
Bonptre, Ekman H-4951 (s-type of C. buchii).

Ekman in his publication entitled “Excursion Botanica al Nord-Oeste
de la Republica Dominicana” (Estacion Agronomica de Moca, Ser. B. 17:
9-10. 1930) refers to a new and distinct species of Coccoloba on the top of
El Morro near Monte Cristi. He comments that this is dedicated to his
amiable companion of the trip, Dr. R. Ciferri. His field label for this
specimen (Ekman 13140) reads “C. Ciferriana Ekman.” Later collec-
tions from neighboring areas, e.g., Ekman 14399, from Arroyo Francés
near Puerto Plata, and ae 16078, Arroyo Arrenquillo near Santiago,
all bear field labels reading, “‘C. CHemiant Ekman and Schmidt.” By 1932,
however, Schmidt had changed his opinion of this material and all the
sheets cited above bear his annotation labels reading “Coccoloba Buchii.”

A study of these plants in the field reveals that both Coccoloba buchii

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 11

as interpreted by Schmidt and C. ciferriana as interpreted by Ekman are
present on the top of El Morro. Careful analyses of these populations were
made in the field and reveal a continuous variation in a series in the shape
of the leaves from sterile and fertile plants and from normal leaves to
leaves of adventitious shoots. The type specimens chosen for these species
(and for C. revoluta) represent extremes of the variation. It is more satis-
factory to consider the entire range of variation as delimiting the species
and to apply the oldest name, C. buchii, to the population. Coccoloba
revoluta Leonard represents an extreme development of pubescence.

The assignment to this species of plants from the Samana Peninsula in
eastern Hispaniola extends the range of the species to the eastward, in
violation of Ekman’s concepts of species distribution in Hispaniola. Ek-
man’s field label for his collection 15439 is preserved in the Stockholm
herbarium. He believed that this plant from the Bocas del Infierno in
Samana was a new species having affinities with C. flavescens. He com-
mented, “The plants made, however, the impression of something new.
The leaves of the saplings are larger than those of mature plants, other-
wise similar.” To this Ekman added ‘“‘a,b,c,d, different stages, a — mature,
d — sapling.” All of these specimens are sterile and Ekman 15439a, de-
scribed by Ekman as a mature shrub, has smaller ovate-elliptic leaves
4.5 & 3 to 3.5 & 2 cm. long and broad. Specimens 15439 c & d are
obviously from faster-growing leader or adventitious shoots. The leaves
of these are lanceolate-ovate to 7.5 2 cm. long and broad to elliptic-
ovate 10 & 4 cm. long and broad. It seems best to recognize this as a
normal variation, although a troublesome one from the taxonomic point
of view.

Howard 12554 is tentatively referred to this species. The flowers were
either past or else the inflorescence had aborted in the dry period. Leaves
of the adventitious shoots are larger, thicker and more prominently
veined than in the remainder of the specimens seen in the field or cited
above.

Coccoloba tortuensis was based on Ekman H-4107 and H-4308, the
former being selected by Schmidt as the type. It was described at the
same time as C. buchiit and both species were based on incomplete or
sterile material. In subsequent papers Schmidt referred additional speci-
mens to C. buchii and expanded his definition of this species. It is m

feeling that C. tortuensis should be referred to the synonomy of C. buchii.
Originally Schmidt distinguished between C. buchii and C. tortuensis on
the basis of the length of the ocrea, the size of the leaf and thickness of
the nodes and the leaf shape. In all characters considered, C. tortuensts
is sufficiently similar to C. buchii to be referred to it. The leaf shape is
slightly different and the pubescence on young parts heavier, but these are
within the range of expected variation.

Coccoloba buchii is similar in general appearance to C. praecox of Cuba,
especially in sterile condition. The two species can be distinguished on the
basis of the more tumid nodes, the longer inflorescences, the larger leaves
and petioles of C. praecox.

12 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Coccoloba krugit is also similar to C. buchii, but the former can be dis-
tinguished by the more cordate-ovate leaves, the angular rachis of the in-
florescence and the angular fruits.

Coccoloba ceibensis Schmidt, Fedde Rep. Spec. Nov. 32: 81. 1933.

Shrub or small tree; branches terete, striate, puberulent, the nodes swol-
len; ocreae 1—-1.5 cm. long, uniformly membranaceous, ferruginous puberu-
lent, bilobed and acute at the apex; leaves with petioles 5-7 mm. long,
puberulent, attached at the bases of the ocreae; blades broadly elliptic, to
ovate-elliptic or obovate-elliptic, 7 « 5, 10 *& 8.5, 12 & 10 cm. long and
broad, coriaceous, slightly bullate especially between the veins, the apex
rounded to apiculate, the base rounded, the margin recurved; midrib
slightly impressed but sharply keeled above, prominent below; primary
veins 6-8 pairs, arcuate and anastomosing near the margin, impressed
above and conspicuous below, the ultimate venation also conspicuous be-
low, the lower leaf surface sparsely short pilose-pubescent, the upper sur-
face glabrous and pitted; inflorescences terminal, 8—11.5 cm. long, the
basal ocreae to 1.5 cm. long, puberulent, the rachis angular, puberulent;
flowering material not known; fruiting racemes with broadly triangular
bracts to 1 mm. long, the ocreolae 1-1.2 mm. long, the fruiting pedicels
1-1.5 mm. long; fruit ovoid, 6 mm. long and 4 mm. diameter, the perianth
lobes half the length of the fruit, imbricate, the achene tan to brown,
shining.

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. TRuyILLo: Llano Costero, Cuenca at La Ceiba,
Ekman H-13344 (p-type, 8); Llano Costero, El Manielito, Ekman H-14281 (s):
Bayaguana at Loma Managua, Ekman H-11108 (s).

Material of Coccoloba ceibensis has not been recollected in satisfactory
condition and the species is known only from the type collection. The two
additional Ekman collections cited above are referred here but with some
question. Hkman H-11108 was collected in sterile condition in January,
1929. The second collection, Ekman H-14281, was collected in February,
1930, and Ekman states on his field notes, ‘collected before sterile.”
Schmidt noted this comment with his initials and a question mark. The
second collection is fragmentary, consisting of two small branchlets, in
quality unlike the material Ekman generally prepared. It bears a stami-
nate inflorescence. Schmidt referred this specimen to C. scrobiculata, a
species which I have referred to C. wrightii. Schmidt compared this species
with C. pubescens, as well as with C. fawcettii, which I have suggested
represents the hybrid of C. pubescens and C. fuertesii.

Coccoloba ceibensis appears to me to be a hybrid of C. pubescens and
probably C. samanensis, although the latter species has not been reported
from the same area. The strong venation of the leaves and the shape of the
fruit with the imbricated lobes of the perianth are similar to C. pubescens,
The suggested parentage of C. samanensis is based on the short fruiting

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 13

pedicels surrounded by membranaceous ocreolae, as well as the texture
and aspect of the upper leaf surface. Although I spent some time collect-
ing in the area where Ekman found these plants, I was unable to locate
additional material near stands of C. pubescens. Ekman reported a the
field label of the type specimen that the plant was “not common.’

Coccoloba costata Wr. ex Sauvalle, Fl. Cub. 139. 1868; Lindau, Engl.
t. Jahrb. 13: 155. 1890, Symb. Antill. 1: 225. 1899; Schmidt,
Fedde Rep. Spec. Nov. 27: 105. 1929; Howard, Jour. Arnold Arb.
30: 145. 1949, 38: 235. 1957.
Coccoloba eggersiana Lindau, Engl. Bot. Jahrb. 13: 153. 1890.
Coccoloba verruculosa Lindau, Engl. Bot. Jahrb. 13: 154. 1890.
Coccoloba rupicola Urban, Symb. Antill. 6: 10. 1909.
Coccoloba sp. Urban, Symb. Antill. 4: 656. 1911.
Coccolobis costata Brit. & Wils. Sci. Surv. P. R. 5: 270. Ve
Coccoloba helwigit Schmidt, Fedde Rep. Spec. Nov. 27:
Coccoloba samuelssonii Ekman & Schmidt, Fedde Rep. ee ee Dine LOS:
1929.

Small tree of shrubby growth or tree to 30 feet tall; branches stout
with a ferruginous to golden pubescence, this often persisting only in pro-
tected spots or at the apex; ocreae membranaceous 4—6 mm. long, ferrugi-
nous-puberulent to subglabrous; leaves of normal shoots with petioles
8-10 mm. long, stout, lightly puberulent, arising from the bases of the
ocreae; blades generally ovate to elliptic, less commonly suborbicular to
obovate-orbicular, 2.5 & 2.2,5 &* 5,7 & 5,11 & 8, to 18 & 12 cm.
long and broad, coriaceous, usually golden, shining above, dull brown
below, sparsely pitted above and below, often showing anomalous peltate
or variously shaped resinous stomatal excretions which may be black and
are abundant when the leaf is young, scattered and few in mature leaves,
otherwise glabrous; apex obtuse or rounded, the base generally slightly
and unequally cordate to narrowly and unequally decurrent on the
petiole; midrib and veins impressed above, prominent below, the primary
veins 5—7 pairs, arcuate, anastomosing; leaves of adventitious shoots to
35 % 22 cm. long and broad on petioles to 1.5 cm. long; inflorescence ter-
minal, rachis puberulent, 15—20 cm. long, the staminate flowers in clusters
of 2-4, the pistillate flowers solitary on pedicels 0.5 mm. long, the bracts
ovate, 0.5 mm. long, the ocreolae membranaceous, 0.5 mm. long, the
hypanthium 0.5 mm. long, the perianth lobes 0.5—-1 mm. long and broad,
the fertile stamens 1 mm. long; fruiting pedicels to 1.5 mm. long, the
fruit globose to 6 mm. long, 5 mm. thick, the perianth lobes coronate.

DISTRIBUTION: Cuba, Hispaniola, Puerto Rico.

Dominican Republic. Prov. BARAHONA: Palo Mino to Montasse on trail to
Polo, Howard 12070 (GH), sea (cH); Mt. Laho, La Cueva to Placer Bonito,
Howard 12272 (A, GH). Prov. BENEFAcToR: San Juan, Loma La Vieja near
Lemba, Ekman H-13448 7 us); Rio Arriba del Norte, Howard 8843 (cH).
Prov. La Veca: Banao, Firme del Banilejo, Ekman H-16460 (s, us). Prov.

14 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

Puerto Prata: La Rosa, Eggers 1762 (B-type of C. verructulosa, K, M, NBV);
Sosta, Ekman H-14459 (s): Puerto Plata, Eggers 2731 (B-type of C. eggersiana,
cH, M), Ekman H-14401 (a, s). Prov. TruyitLo: Villa Altagracia, Ekman
H-11231 (s).

Haiti. Dept. pu Norp Ovest: Bombardopolis, EC & GM Leonard 13509 (xy,
us). Dept. pu Norp: Massif du Nord, Le Borgne, edge of Estere Savate,
Ekman H-4855 (p-type of C. samuelssonii, s, us); Massif du Nord, Gros
Morne, Morne Bonpére, Ekman H-4944 (s, us), H-8523 (B-type of C. helwigit,
s); Massif du Nord, Hinche, Bois Charles, Ekman H-6077 (s, us); Massif du
Nord, Bayeux, Morne Brigand, Ekman H-2855 (B, s, us); Massif du Nord,
St. Louis du Nord, Morne Baron, Ekman H-4684 (s) ; Massif du Nord, St. Louis
du Nord, between Baron and Rio Jean-Claire, Ekman H-3904 (s). DEPT DE
L’Ouest: Massif des Matheux, l’Archaie, Lully to Caye-Nicolas, Ekman H-9287
(a, s, US); Massif des Cahos, Las Caobas, Belladére, Ekman H-5589 (8, 8);
Ile de la Gonave, Ekman H-8721 (s)

n a treatment of the genus Coccoloba as it occurs in Puerto Rico
(Jour. Arnold Arb. 38: 235-237. 1957), I discussed the morphological
variation of Coccoloba costata and extended the recognized range of this
species from Cuba alone to Hispaniola and Puerto Rico. A number of
species from Hispaniola formerly considered distinct and endemic were
reduced to the synonymy of this taxon. These included C. samuelssonti,
considered distinct on the shape of the leaf base and the texture and shine
of the leaf blade; C. helwigti, supposedly distinct in having orbicular
leaves with minute punctations, the latter proving to be blocked stomata
and resinous excretions from the stomata; and C. verruculosa which Lin-
dau distinguished from C. costata on the matter of pubescence as well
as the number and length of flowering pedicels. Still another Hispaniolan
endemic species, C. eggersiana, must be included in this species and reduced
to the synonymy of C. costata. In his key Lindau distinguishes C. egger-
siana from C. verruculosa by the size of the leaves and the thickness of
the inflorescence rachis. Coccoloba eggersiana is also distinguished from
C. costata in the same key on the absence of pubescence on the inflores-
cence rachis. The abundance of material cited above demonstrates numer-
ous intermediates on all characters and the type collection of C. samuels-
sonii (Ekman H-4855) shows clearly the range of variation from material
matching the type of C. costata to shoots comparable to the type speci-
mens of C, eggersiana.

The following series of specimens from near Puerto Plata and from the
Province of Samana are referred to Coccoloba costata, but with some
question: Prov. Purerto PLATA: Sosta, coral reef at Forma, Ekman
H-14460 (s, us). Prov. SAMANA: San Lorenzo Bay, Abbott 2237 (8, GH,
NY, US); 2245 (us); Samana, Laguna, Ekman H-15096 (s, us); Los
Haitises, Cueva de Cal, Ekman H-15573 (Ss); Samana, Laguna, Loma
Zaramagua, Ekman H- 15249: Los Haitises, Cayo de los Cueros, Winon
H-15516 (s).

Each of these collections differs in some aspect from the expected range
of variation now recognized for Coccoloba costata. Two collections, Ek-

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 6)

man 14460 and 15096, from Puerto Plata and Samana respectively, have
normal shoots with leaves similar to the type specimen of C. eggersiana
and are clearly confluent with the range and variation of C. costata. The
adventitious shoots of these same plants which are mounted on herbarium
sheets are oblong, varying from 7.5 & 3 to 14 & 4.5 cm. long and wide.
This leaf shape has not been found on adventitious shoots of C. costata in
other sections of its range. Ekman 15249 and 15573 from the Samana
peninsula and the two Abbott collections from the same area appear to be
vigorous shoots, possibly adventitious in origin, but with small oblong
leaves 6 X 2 to9 & 2.5 cm. long and wide. All of these collections bear
staminate inflorescences. Ekman 15573 bears field notes indicating that
the plant is “very common in the Haitises.” It also has unripe fruits
which are too small to permit the determination of either the fertility of
the fruits or their possible similarity to those of C. costata. Finally,
Ekman H-15516 from Los Haitises bears a staminate inflorescence and
unripe fruits which appear to be comparable to those of the Ekman col-
lection just mentioned. This plant, Ekman notes on his field label, is
in the same area. The leaves of collection H-15516 are ovate-
lanceolate to lanceolate-oblong. While the leaf base is similar to that
expected in C. costata, the general aspect of this specimen suggests a
relationship to C. Fre eahe. It is possible that further collections may
reveal a hybrid population of C. diversifolia and C. costata in the Samana
area, particularly inland on the limestone mountains.

Coccoloba costata Wr. ex Sauv. « Coccoloba uvifera L., hybr. nov.

Shrub or small tree (acc. Ekman), branches terete, striate; persistent
bases of ocreae 4-5 mm. long, the membranaceous portions unknown;
puberulent; leaves with stout petioles, 5-6 mm. long, puberulent, attached
at the base of the ocreae; blade orbicular to elliptic, 7 « 6, 10 & 10,
10.5 %& 9 cm. long and broad, coriaceous, the apex rounded, the base
rounded to slightly cordate, one basal lobe larger than the other, one or
both slightly overlapping the petiole, the margin slightly revolute; midrib
conspicuous on both surfaces, the veins 6 pairs, slightly conspicuous on
both surfaces, the lower three close together; inflorescence terminal; im-
mature, to 8 cm. long with a shorter basal branch, the rachis puberulent;
flowers immature; fruit not known.

Dominican Republic. Prov. Puerto PLata: Puerto Plata, cliffs facing the sea
at Arroyo Francés, Ekman H-14402 (a, $s).

Schmidt referred this collection to Coccoloba verruculosa with a ques-
tion. In the general appearance of the leaves, this plant resembles C.
uvifera. The similarity is accentuated by the nature of the leaf base and
the branching habit of the inflorescence. The similarity to C. costata is
found in the texture, the venation and the color of the leaf blades. Both
C. uvifera and C, costata are found at Arroyo Francés and it is apparent
that this single collection represents a hybrid of these two species.

16 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

Coccoloba diversifolia Jacq. Enum. Pl. 19. 1760; Hist. Stirp. Amer.
114, pl. 76. 1763.
Coccoloba cubensis Meisner, DC. Prodr. 14: 162. 1857.
Coccoloba ede ak ie Engl. Bot. Jahrb. 13: 158. 1891, and all recent

authors, not Jacq
ace longifolia cern Fedde Rep. Spec. Nov. 24: 73. 1927, not Fischer
x Lind.

Shrub or tree to 23 feet tall; branches terete, often geniculate by limited
growth, glabrous, the nodes rarely slightly swollen; ocreae coriaceous in
the persistent lower portion, membranaceous and deciduous above, 3-5
mm. long; leaves of normal shoots with petioles 7-10 mm. long, glabrous,
arising from the base of the ocreae; blades ovate, oval, oblong, elliptic,
lanceolate or obovate, variable on one branch, 4 & 3.5, 7 & 5.5, 8 & 4.5,
12 & 8 cm. long and broad, coriaceous, often shining above, dull be-
neath, glabrous, the apex rounded, obtuse, acute or acuminate, the base
cuneate, rounded or subcordate, the margin entire; midrib and primary
veins slightly prominent above, the secondary venation reticulate on both
surfaces, the primary veins 3—7 pairs, arcuate, anastomosing before reach-
ing the margin; leaves of adventitious shoots on petioles 1—2.5 cm. long,
with blades of varying shapes 17 & 8, 24 & 13, to 32 12.5 cm. long
and wide; leaves of windswept specimens often much smaller than those
of normal shoots with blades 2 « 1.3 to 3 % 2 cm. long and broad; in-
florescence terminal 4.5—18 cm. long, rachis glabrous, the flowers on pedi-
cels 2-4 mm. long, the staminate flowers 2—4 at each locus, the pistillate
flowers borne singly at each locus, the bracts ovate, less than 0.5 mm.
long, the ocreolae membranaceous, less than 0.5 mm. long, the hypanthium
1 mm. long, the perianth lobes ovate to oblong, 2-3 mm. long, 1-2 mm.
broad, the functional stamens 1 mm. long, the sterile stamens rudimentary ;
fruiting pedicels 3-4.5 mm. long, the fruit globose to obpyriform, 10 7,
12 & 8,13 & 8 mm. long and in diameter, the apex rounded, the perianth
lobes inbrieate and appressed.

ComMMON NAMES: Maivisse (H), raisin marron (H), resinier (H), uva
cimarrona, uvero (DR), uvilla (DR), zamon maron (H).

DIsTRIBUTION: Florida, Bahamas, Greater and Lesser Antilles.

Dominican Republic. Prov. La ALTaGRAcIA: La Romana, Taylor 384 (Ny, US).
Prov. Azua: Azua, Rose, Fitch & Russell 3934 (Ny, Us), 4415 (NY, US). PROV.
BARAHONA: Beata Island, Howard 12361 (GH), Palo Mino to Montasse near
Polo, Howard 12096 (GH), Barahona, Fuertes 1104 (Ff, GH, MO, NY, S$, US). PROV.
BENEFACTOR: Rio Arriba, R.A. & E.S. Howard 8976 (cH). Prov. MONTE
Cristi: El Morro, Howard 12536 (GH), 12538 (GH), 12531 (GH), Moscoso 158
(sp); Santiago Rodriguez, Moncion, Valeur 254 (A, F, MICH, MO, NY, S, US).

Puerto PLata: Sosua, Jiménez 1676 (ost) Savana de Guainamoca, Eggers
2558 (B, NY, US). Prov. SAMANA: Lajana, Abbott 1297 (us). Prov. SAN PEDRO
bE Macoris: San Pedro de Macoris, Rose, Fitch & Russell 4444 (us). Prov.
SANTIAGO: Angostura, Jiménez 1824 (A). Prov. TrujiLto: Jaina, Faris 113
(us). Disrricro pE SANTO Dominco: Ciudad Trujillo, Schiffino 138 (Gc). San

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 17,

Gabriel Island, Miller 1022 (us). Puerto Escondido, Howard 12139 (cH),
12145 (GH). LOCALITY UNCERTAIN OR NOT INDICATED: Wright, Parry & Brum-
mel 477 (G, US); Lopez, Eggers 3387 (Ny, us); Rio Verde, Eggers 2325 (mM);
Sharff 4 (f), 17a (Ff); Jaeger 313 (Ny); Sessé & Mocino 5428 (F), 948 (F).

Haiti. Dept. pu Norp Ouest: Tortue Island, La Vallée, EC & GM Leonard
11288 ee GH, Us); Mole St. Nicolas, EC & GM Leonard 13367 (us). DEPT.
pu Norp: Bayeux near Port Margot, Nash 287 (F, NY). DEPT. DE L’ARTIBONITE:
Ennery, Leonard 8909 (GH, NY, US), 9527 (Us); Petit Riviére de |’Artibonite,
Picarda 1580 (B); St. Michel de l’Atalaye, Leonard 7483 (Mo, Us), 7426 (US);
La Brande to Mt. Blanche, Nash & Taylor 1649 (Ny). DEPT. pu Sup: Ile Grande
Cayemitte, Eyerdam 318 (F, GH, Us); Navassa Island, Rekder 12 (A, Ny, US),
Ekman 10800 (s), Proctor 15478 S 1J); Jeremie, Picarda 1314 (8); Massif de
V’Hotte, Trouin, Ekman 5944 (a, s); Miragoane, Prince Paul 1313 (mM). DEPT.
DE L’OuEST: Gonave Island Gee Galette): Leonard 3265 (B, NY, US), 3261
(us); Port au Paix, La Coup River, EC & GM Leonard 11138 (mo, NY, US);
between Passe Aubert & Passe Chance aulme, Ekman 3819 (s, us); Jean Rabel,
EC & GM Leonard 12704 (a, us); Fond Parisienne, Holdridge 1822 (GH, US);
Morne Hospital, Buch 1839 (B); Petionville, Picarda 1410 (8).

This species is more common in Hispaniola than the few specimens cited
above would indicate. In many areas Coccoloba diversifolia is in general
use as a fence row tree, although I could never determine whether the
plants were started as seeds or as cuttings.

The general misapplication of the names Coccoloba diversifolia and
C. laurifolia by nearly all recent authors and their proper application has
been discussed in an earlier paper (Jour. Arnold Arb. 30: 422-424. 1949).

Coccoloba fawcettii Schmidt, Fedde Rep. Spec. Nov. 24: 76. 1927.

Small tree (fide Ekman); branches tortuous, the nodes conspicuously
swollen, the youngest parts densely golden-pubescent; ocreae 3—5 mm. long,
membranaceous, pubescent, truncate or slightly bilobed at the apex; petioles
2-3 mm. long, pubescent, arising at the bases of the ocreae; blades cordate,
3.5 & 3,4 & 4, 8 & 6 cm. long and broad, rigid and thin, papery in
texture, the apex acute or obtuse, slightly apiculate, the base cordate,
the margin entire, slightly undulate; midrib slightly prominent above,
conspicuous below, primary veins 3—5 pairs, arcuate and anastomosing at
the margin, scarcely impressed above, prominent beneath, the secondary
venation prominent. and reticulate below, the upper surface full and
glabrous, the lower surface sparsely short pubescent; inflorescence ter-
minal 1.8-3 cm. long, the rachis puberulent, the bracts triangular to 1.2
mm. long, spreading, the ocreolae membranaceous, to 1.2 mm. long, the
pedicels shorter than the ocreolae; staminate flowers unknown, the pistillate
flowers 1, rarely 2, at each locus, the hypanthium less than 0.5 mm. long,
the perianth lobes broadly ovate, 1.5-2 mm. wide and 1.0-1.5 mm. long,
the stamens rudimentary, less than 0.5 mm. long, the ovary ellipsoidal,
1.5 mm. long, sharply 3-angled; fruit not known.

DIstTRIBUTION: Endemic to the Dominican Republic. Known only
from the type collection.

18 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

Dominican Republic. Prov. BARAHONA: Mare-a-chat, Ekman H-6948 (B-type,

While the Berlin specimen of this collection has old inflorescences from
which the flowers have fallen, the Stockholm specimen still retains a few
open flowers. Schmidt states in the original description that the flowers
and fruit are unknown. It is obvious that the Berlin specimen is the
holotype. The staminate flowers and fruits are still unknown, but the
pistillate flowers are described above

The Berlin specimen consists of two fragments, one obviously from a
mature shoot with inflorescences and the other from a more vigorously
growing sterile shoot. The larger leaf size given in the description refers
to leaves on the latter shoot. Truly adventitious shoots are not known.

In his original description Schmidt compared Coccoloba fawcettii with
C. pubescens and C. fuertesii. It seems probable to me that C. fawcetti
is a natural hybrid between these two species, both of which occur in the
area where C. fawcettii was collected. Schmidt refers to the leaves of C.
pubescens as being many times larger, yet some leaves of mature shoots
in that species scarcely exceed those of the larger leaves on the type speci-
men of C. fawcettii. I was unable to find any plants of this species on a
trip to the type location. Ekman stated in his field notes that the plant
was “rare” and seen “only here.’ As a hybrid, C. fawcettii would derive
the pubescence and reticulation from C. pubescens and the leaf shape
and the swollen nodes from C. fuertesit.

Coccoloba flavescens Jacq. Hist. Stirp. Amer. 115. tab. 75. 1763.
Coccoloba pungens Urban, Symb. Antill. 5: 335. 1907; 8: 195. 1920.

Shrub, often with many trunks, to small tree with single trunk reaching
15 feet in height; bark gray, the youngest branches tan, slightly puberu-
lent, geniculate, the nodes commonly swollen; ocreae membranaceous
except at the base, to 3 mm. long, the base eatlazinols. ring-like, the
petioles arising from this ring-like base, often appearing terminal due to
geniculation of the stem; leaves of normal shoots with petioles 2-5 mm.
long, puberulent to glabrous, the blades elliptic, ovate-elliptic or lanceolate-
elliptic. 1,6, 1.2, 3.0 3-18; 4.0 & 1.6, to 5.6. 22 em. Jone-and
broad, rigid, flat or slightly umbonate, the apex acute, mucronate, the
cartilaginous mucro to 1 mm. long, sharp and stiff, the base rounded to
slightly cordate, the margin entire; venation not evident in fresh condi-
tion, only slightly evident when dry, the veins 10-20 pairs, equally de-
veloped; leaves of adventitious shoots similar in shape, 6.0 « 3.2 to 7.8
< 3.1 cm. long and wide; inflorescence terminal, 2—2.5 cm. long, the rachis
glabrous, the bracts to 1 mm. long, slightly erose at the apex, the ocreolae
less than 1 mm. long, the flowering pedicels short, those of the staminate
flowers less than 1 mm. long, deciduous or rarely persisting and elongating
slightly; staminate flowers borne singly at each locus, the hypanthium
short, about 0.5 mm. long, the lobes 5 or 6, ovate, 2-3 mm. long, 1.5—2.5

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 19

mm. wide, the filaments 1-1.5 mm. long, the pistil rudiment to 2 mm. long;
female flowers not seen; fruiting inflorescence stout, 1-4 cm. long, often
strongly angled, the peduncles decurrent on the axis; fruit bright red,
4-5 mm. diameter, 6—9 mm. long, broadest at the middle, the apex rounded,
the base slightly pedicellate, the perianth segments free to the middle,
imbricate over the achene, strongly lined but the vascular bundles not
evident.

DistTRIBuTION: Endemic to Hispaniola.

Dominican Republic. Prov. BARAHONA: Barahona, Fuertes 543 (F, GH, MO,
NY, Ss, US); Las Salinas, Howard 12059 (GH), 12062 (GH). Prov. INDEPENDEN-
cra: Lake Enriquillo to Puerto Escondido, Howard 12132 (GH); Puerto Escon-
dido to Rancho Viejo, Howard 12132 (GH).

Haiti. Derr. pu Norp: Massif du Nord, Gros Morne, Morne Bonpére, Ekman
H-4939 (s). DEPT. DE L’ARTIBONITE: Gonaives to La Hotte Rochée, Nash &
Taylor 1560 (Ny). Dept. DE L’Ouest: Fond Parisien, Etang Saumatre, Leonard
10138 (Ny, us); Massif des Matheux, Croix des Bouquets, Morne a Cabrits,
Ekman H-989 (s, us). Dept. UNCERTAIN: Cape Francois, Ehrenberg s.n. (8,
NY); Poste Coudau, Buch 1024 (B-type of C. pungens).

This is a distinctive species apparently restricted to the dry areas at
low elevation. Three leaf forms were seen in the field. The normal leaves
were found on the much-branched and geniculate branches. Much smaller
leaves were seen occasionally on fasciated lateral branches which are
borne on wand-like shoots; see Ekman H-3499. Slightly larger than nor-
mal leaves are found on shoots of less compact habit, having longer inter-
nodes and these are considered to be adventitious shoots. All three habits
and leaf sizes are commonly found on one plant. The leaves are rigid
even when fresh and the short but sharp mucro is very much in evidence
when collecting this species. The leaves when fresh have the character-
istic grayish-green color dominant in arid regions of the West Indies.

The fruits of C. flavescens are distinctive in shape and in the nature of
the perianth in fruit. The fine striations on the lobes of the fruiting peri-
anth seem to be characteristic of this species. Much larger fruits are asso-
ciated with Ekman 989. These appear to be teratological and gall-infested.
The few fruits opened were devoid of seed.

Coccoloba fuertesii Urban, Symb. Antill. 7: 210. 1912.
Coccoloba taylori Urban, Fedde Rep. Spec. Nov. 13: 446. 1914; Symb.
Antill. 8: 196. 1920; Ark. Bot. 20A(15): 29. 1926.
Coccoloba barkeri Ekman & Schmidt, Fedde Rep. Spec. Nov. 27: 104. 1929.

Tree to 40 feet; diameter at breast height ten inches, bark rough, in
characteristic l-inch squares; branches terete, the nodes swollen, glabrous
or at most papillose; ocreae to 5 mm., membranaceous above, this part
lobed to the base on one side, often flaring, thick and persistent below;
leaves of normal shoots with petioles 2-6 mm. long, papillose, arising above
the base; blades ovate to ovate-triangular, less frequently ovate-lanceolate,

20 JOURNAL OF THE ARNOLD ARBORETUM _ |VoL. XxxIx

elliptic or obovate, 2.5 & 2,4 * 3,5 K 4,5 & 5.2,6.5 X 4.5 to7 X 3.7
cm. long and wide, generally broadest above the middle, coriaceous, en-
tire, glabrous, the apex short- to long-acuminate, mucronate, the ultimate
apex generally spine-tipped, the cartilaginous spine 0.5-2 mm. long, at
times rounded to emarginate or almost bilobed through failure of the mid-
rib to develop, the base cuneate to rounded, rarely acute or subtruncate,
the margin entire; veins 5-10 pairs, subequal, prominently reticulate on
both surfaces when dry, the veins arcuate, approaching but distinct from
the cartilaginous margin of the leaf; leaves of adventitious shoots borne
above the base of the ocreae 1-1.5 cm. long but occasionally to 3.4 cm.
long, on petioles 1 cm. long, the blades broadly ovate to ovate-triangular
to elliptic or ovate-lanceolate or elliptic-lanceolate, 8.2 6.2 to 16.5 K 9
cm. long and wide, often asymmetrical, the apex acute, acuminate, obtuse
or rounded, the terminal spine present or absent, the base truncate, sub-
truncate or rounded; inflorescences equalling or surpassing the leaves,
2.5-6.5 cm, long, terminal, 1—4 at each locus; flowers sessile or on pedicels
not exceeding 1 mm.; staminate flowers 1-4 at each locus, the pistillate
flowers 1 at each locus, the bracts broadly ovate, 0.5 mm. long and wide,
the ocreolae membranaceous, flaring, 0.5 mm. long; hypanthium to 1 mm.
long, the perianth lobes oblong to ovate-orbicular, 1-1.3 mm. long and
wide, the functional stamens to 1 mm. long, the sterile stamens with rudi-
mentary anthers; functional pistil to 3 mm. long, the sterile pistil rudi-
mentary; fruit a when mature, oblong or ovate, fleshy, 6-9 mm. long,
3-5 mm. in diameter, the hypanthium red in fruit, the perianth lobes
coronate, 1-2 mm. long: fruit with a corky or sods sub- arias aaa
aa a this often marked with a horizontal constriction and fre
quently attenuate below; achene brown or tan.

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Dist. DE SANTO DoMINGO: Santo pe on road to
Puerto Plata, Wright, Parry and Brummel 474 (cH, US). Prov. BARAHONA:
Barahona, Fuertes 716 (B-type), Palo Mino to Montasse on trail to Polo, Howard
12065 (cH), Cafada Maluca, Howard 12196 (GH). Prov. Monte Cristr: Res-
tauracion, Ekman H-6250 (s). Prov. Puerto PLaTA: Sosta at La Goleta, Ekman
H-14529 (s, us). Prov. SAMANA: Samana, Pan de Azucar, Ekman H-15181
(s), Punta Arena, Ekman H-14788 (s), Los Haitises near La Llanada, Ekman
H-15452 (s). Prov. Truyitxo: Villa Altagracia, Ekman H-11238 (a, 8, US).

Haiti: Dept. Du me Ovest: Massif du Nord, Port de Paix, Haut Piton
Ekman H-4633 (B, s, us); Moustique Mts., Bassin Bleu, £.C. & G.M. Leonard
14985 (US), 14986 a 15021 (A, GH, NY, us); Presqu‘ile du Nord, Port de
Paix, Haut Moustique, Ekman H-3639 (s): Jean Rabel, E.C. & G.M. Leonard
13635 (A, GH, NY, US). Dept. pu Norp: Massif du Nord, Morne Belance, Ekman
H-4911 (s); Massif du Nord, Bayeux, Morne Brigand, Ekman H-2854 (s, US).
Dept. DE L’ARTIBONITE: Massif du Nord, Gros Morne, Morne Chabre, Ekman
sae - Massif du Nord, Morne Bonpére, Ekman H-4952 (A, S, US);

a e to Morne Belance, Taylor 1674 (B-type of C. taylori, NY); Ennery,
Beman es 2468 (B). Dept. pu Sup: Tiburon, Morne Sentier, Ekman H-10397
(B, s, US); Morne Rochelois, Charlier, Ekman H-9037 (a, s); Grand Cayemite,

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 7A

ee 304 (A, F, GH, MO, NY, oe Ekman H-8927 (s, us), Miragoane, Petit-

viére des Nippes, Eyerdam 398 (A, F, GH, NY, US). DEpT. DE L’ :
Massif de la Selle, Leogane near Citronniers, Ekman H-6464 (p-type of C.
barkeri, s, us); Massif de la Hotte, Jacmel at Savanette, Ekman H-7081 (s, us).

The full variation in leaf size, shape and texture and the development of
the terminal spine in this species must be seen in the field to be appre-
ciated. One of my collections (Howard 12065) is represented by a series
of twelve sheets taken from a single tree. This tree was 40 feet tall and
10 inches in diameter at breast height. Adventitious shoots developed in
profusion from the trunk and from along the first main branch. The range
of variation on the upper branches of the tree was significant, but this was
exceeded by the variation on individual and different adventitious shoots.
In general the leaves of the adventitious shoots were 2 to 3 times the
size of the leaves on normal shoots. A second collection (Howard 12196)
was made from a plant which consisted of a cluster of 14 trunks, the small-
est less than an inch in diameter and the largest 8 inches and 20 feet tall.
The original tree had been felled, but the stump remnant was still present.
The smaller adventitious shoots produced the expected large leaves typical
of such shoot systems, but the larger trunks retained only a few large
leaves, while the majority were of the size range characteristic of the
mature shoots. Other collections which show variations, apparently on
single plants, are Ekman H-7081 and H-14788

Urban recognized the affinity of Coccoloba taylori with C. fuertesti, but
distinguished between them on the basis of petiole length and the shape
of the leaves. The material which Urban studied consisted of a fast-
growing, probably adventitious, shoot (the type of C. fuertesii), and two
gnarled branches of a mature tree (the type of C. taylori). Comparable
variation can be found very readily on one tree.

Coccoloba barkeri is based on a specimen of mature growth with in-
florescences to 9 cm. long, these being 1-4 in number. Urban’s distinction
of C. barkeri from C. taylori on the basis of inflorescence length was made
by comparing the longest inflorescence of the type specimen of C. barkeri
with the average, if not the shortest, inflorescence of C. taylori. Recent
collections demonstrate that C. barkeri is to be included with C. taylori
in the species C. fuertesii.

Much of the material cited has been determined by others as Coccoloba
retusa or C. diversifolia. Coccoloba retusa, a Cuban species, differs in
having thinner textured leaves, less tumid nodes and more conspicuously
pedicellate flowers and fruits. Leaves of C. retusa also lack the spiny
tip found in C. fuertesii, but perhaps the most significant difference is
the absence in C. fuertesii of the solid basal section of the fruits of
C. retusa. The two species are very similar and perhaps eventually C.
fuertesti should be considered a geographical variety of C. retusa. Cocco-
loba diversifolia of recent authors is now known as C. swartzii and is
clearly distinct as to the shape of the fruit, the leaf venation and margin
and the lack of any development of pedicels.

oe JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

Coccoloba fuertesii is similar to C. incrassata and the relationship of
these two species requires further study of the populations in the fie

Coccoloba hotteana Schmidt, Ark. Bot. 20A (15): 31. 1926.

A shrub or depauperate tree to 5 feet tall; branches slender, terete, the
nodes not enlarged; ocreae 5—6 mm. long, obliquely truncate at the apex,
more or less bilobed, short ferruginous-pubescent; leaves of normal shoots
with petioles 3-4 mm. long, ferruginous-short-pubescent, inserted at the
base of the ocreae; blades elliptic, rarely ovate-elliptic or narrowly obo-
vate, 3 & 1.8,4 & 2.5, 4.5 & 2 cm. long-and broad, thin coriaceous, the
apex narrowed and generally abruptly acuminate and slightly apiculate,
the base rounded or narrowed, the margin entire, slightly revolute; midrib
lighter in color, slightly conspicuous above, prominent below, primary
veins 4 or 5 pairs, evident on both surfaces, slightly arcuate to the margins
ne anastomosing parallel to the margin, the stomata depressed and ap-

ing as punctations on both surfaces; leaves of adventitious shoots
ee elliptic, rounded or narrowed at the base, acute at the apex
or apiculate, to 8 & 3 cm. long and broad; inflorescence terminal 5—15
cm. long, glabrous to sparsely puberulent, the bracts ovate or triangular,
to 0.5 mm. long, the ocreolae membranaceous, 0.5—1 mm. long, the pedicels
about 1 mm. long; staminate flowers 1—4 per locus, the pistillate flowers
borne singly at each locus, the hypanthium 1-1.5 mm. long, the perianth
lobes ovate to suborbicular to 2 mm. long and wide, the fertile stamens
1.5-2 mm. long, the sterile stamens rudimentary, the sterile pistil rudi-
mentary, the fertil pistil 1-1.5 mm. long; fruit subglobose, 6 mm. in diam-
eter, the fruiting perianth lobes imbricate, the achene dark brown,
smooth.

DISTRIBUTION: Endemic to Haiti.

Haiti. Dept. pu Sup: Anse-a-Veau, Ekman H-5399 (s-holotype, s, Us);
Ile Grande Cayemitte, Ekman H-8954 (s), Eyerdam 285 (A, F, GH, NY, US).

The type collection of this species represented by herbarium specimens
from Berlin, Stockholm and the U.S. National Museum comprises nine
plant fragments. Normally Erik Ekman collected material of good quality
and a single specimen fills an herbarium sheet. It would be interesting to
know what happened to this material. The nine specimens of plant
material represent sterile mature growth, possible adventitious growth,
staminate, flower-bearing branches, pistillate flower-bearing branches and
fruiting twigs. One fruit is attached to an inflorescence which appears to
be from a staminate plant. Additional fruits in the packet on this sheet
are hollow and sterile. The type specimen from the Berlin herbarium
consists of three fragments, one pistillate flowering axis, one fruiting axis
and one sterile branchlet. In the packet on this sheet are three fruits on
which both Schmidt and I have based our descriptions. If the fruit belongs
with the specimen, its characteristics are of value in recognizing this
species, but I question the authenticity of this fruit.

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V i)

Coccoloba hotteana appears to be a distinct species. At the present
it is known only from a coral reef west of Anse-a-Veau and from Ile
Grande Cayemitte. Additional material is much desired. The species
can be recognized by the shape and venation of the leaf, the apiculate
apex to the leaves and the pubescence on the ocreae and inflorescence
rachis.

Ekman did not approve of Schmidt’s choice of name for this species, for
La Hotte is a mountain range, while C. hotteana is known only from low
coral reefs at sea level. He comments in the field notes of a second col-
lection, ‘“‘The specific name is misleading in a way. ‘Hotteana’ ought to
be reserved for mountain plants.”

Coccoloba hotteana Schmidt * C. uvifera L., hybr. nov.

Shrub (acc. to Ekman); branches terete, striate and canaliculate,
puberulent to short pubescent; ocreae to 1 cm. long, strongly bilobed at
the apex, puberulent, persistent; leaves of es shoots with petioles
—6 mm. long, stout, puberulent, arising from the bases of the ocreae;
blades broadly elliptic to obovate-elliptic or orbicular, 6 « 4 to 7 & 6
cm. long and broad, coriaceous, the apex rounded to short apiculate, the
base rounded, usually slightly asymmetrical, the margin flat; midrib con-
spicuous on both surfaces, the primary veins 5 or 6 pairs, conspicuous
on both surfaces, the ultimate venation reticulate and conspicuous when
dry, the upper surface glabrous, the lower surface puberulent and densely
punctate-dotted with blocked stomata, the midrib and veins puberulent;
leaves of adventitious shoots similar, to 10 & 8 cm. long and broad;
inflorescence terminal, 10 to 25 cm. long, the rachis angular, puberulent,
the bracts broadly ovate to 1 mm. long, the ocreolae 1 mm. long; pedi-
cels 2—2.5 mm. long.

aiti. Dept. pU Sup: Ile Grande Cayemitte, northern coast in the Cétes de
nee Ekman H-8950 (8, s), Eyerdam 303 (A, F, GH, MO, NY, US).

This hybrid is represented by collections of Ekman and Eyerdam num-
bered separately but made when these men were travelling together, on
the same date and obviously from the same plant. This material in some
respects resembles hybrids of Coccoloba uvifera and C. costata recog-
nized from the vicinity of Puerto Plata and some specimens (e.g., Eggers
1762, the type of C. verruculosa) assigned to C. costata as a species. It
seems distinct from these and appears to be derived from the parentage of
C. uvifera and C. Aotteana, both of which occur on Ie Grande Cayemitte.
The leaf shape, particularly the nature of the leaf apex and the venation,
appears to be derived from C. hotteana. The texture of the leaf, the nature
of the inflorescence and the shape of the sterile fruits are those of C. uvi-
fera. The inflorescences are old in the specimens cited above and a few
fruits have been retained in packets. The flowering pedicels are borne,
3—5 at each node, on the inflorescence rachis, a character usually asso-
ciated with staminate plants. As has been pointed out (Jour. Arnold Arb.

24 JOURNAL OF THE ARNOLD ARBORETUM _ |VoL. XxxIx

30: 390. 1949.), staminate plants of Coccoloba uvifera often form fruits
which are of full size and normal shape but hollow or with aborted
seeds or embryos. The fruits associated with the cited specimens re-
semble those of C. uvifera, not C. hotteana. They are obovoid, to 1.5 cm.
long and 1 cm. in diameter. The fruits are strongly narrowed at the base
and rounded at the apex with imbricated perianth lobes. The achene is
dark brown and smooth but hollow or with a very small seed and is
obviously infertile.

Coccoloba incrassata Urban, Symb. Antill. 7: 208. 1912.

Coccoloba _ ke Ark. Bot. 20A(15): 1926; Fedde Rep.
Spec. Nov. 27:

Shrub 10 feet tall to densely branched tree 20 feet tall; branches com-
pact or geniculate, the lateral branches with short internodes often ap-
pearing as short shoots, the nodes conspicuously swollen, the branches
often appearing moniliform, papillose to puberulent; ocreae membranace-
ous above, 1-2 mm. long, glabrous; leaves of normal shoots with petioles
1 mm. long, glabrous, inserted above the bases of the ocreae, the blades
ovate-triangular, 1 x 0.7, to 1.7 x 1.1 cm. long and wide, thick coriace-
ous, glabrous, the stomata and subsidiary cells clear, appearing as aii
cent dots, the apex attenuate to a spinose mucro, the spine about 1 m
long, the base rounded to subtruncate; midrib cartilaginous, ee
on both surfaces, the veins anastomosing, conspicuous on both surfaces,
2 pairs of veins near the base of the blade separated from the upper
pairs of veins, the veins reaching the margin without becoming arcuate,
then fusing with the cartilaginous ring at the margin; leaves of adventi-
tious shoots with petioles 4 mm. long, the blades lanceolate to lanceolate-
ovate or elliptic-lanceolate 2.5 & 1.4 to 3 & 1 cm. long and wide, acumi-
nate and pointed or narrowed to a rounded apex, base rounded to tri-
angular, the adventitious stems densely pilose to completely glabrous;
inflorescence terminal, shorter than the leaves or reduced to an almost
sessile 1—4-flowered cluster; bracts and ocreolae minute, less than 0.5
mm. long, the pedicels approximately 1 mm. long; staminate flowers 1 or
2 per locus, occasionally 1-4 on adventitious flowering shoots, the pistil-
late flowers borne singly at each locus on the inflorescence; hypanthium
to 1 mm. long, the perianth lobes 1 mm. long and 0.7 mm. wide, the
fertile stamens to 1 mm. long, the fertile pistil 2 mm. long; fruit 5 mm.
long and 3 mm. diameter, narrowed to a corky sterile base, coronate at
the apex, on pedicels 1 mm. long.

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. Azua: Wright, Parry & lags 476 (GH, US).
Proy. SAMANA: Los Haitises, La Marachita, Ekman 15531 Us),

Haiti. Dept. pu Norp OvEsT: Pendu, i. 1269 (B-type, oa. Mole St. Nico-
las, E.C. & G.M. Leonard 13356 (mo, us); Bombardopolis, E.C. & G.M. Leonard
13428 (A, GH, NY, US); Bombardopolis ae south of Méle George, E.C. & G.M

1958] | HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 25

Leonard 13421 (micH, us); Presqu’ile du Nord Ouest, Méle St. Nicolas at
Pap-a-foux, Ekman H-4487 (s, us); Presqu’ile du Nord, Port de Paix, Haut
Moustique, Ekman H-3648 (a, B, 8, US). DEPT. DE L’ARTIBONITE: Ennery,
Leonard 8861 (us); Massif du Nord, Ennery, Ekman H-2468 (s, us); Trois
Riviéres near Gros Morne, £.C. & G.M. Leonard 9893 (cH, Mo, us); Mt. La
Cidre, near St. Michel de l’Atalaye, Leonard 7397 (Ny, US). DEPT. DE L’OUEST:
Massif de la Selle, Anses 4 Pitres, Ekman H-6688 (s); Massif de la Selle, group
Crete-a-Piquants, Port-au-Prince, Morne Aux-Fourques, Ekman H-5921 (B-type
of C. Mansfeldii, s); Massif des Matheux, l’Arcahaie, Ekman H-9330 (3, S).
The presently accepted distinctions between Coccoloba incrassata and
C. fuertesti are not entirely satisfactory. Coccoloba incrassata was based
on a Buch collection from dry hillslopes at moderate elevation near Pendu.
Urban compared the species with C. armata of Cuba which is clearly distinct
in the manner of branching as well as in the shape and venation of the
leaves. Coccoloba mansfeldii, described by Schmidt, was based on a
sterile collection made by Ekman and is clearly a fast-growing shoot of
C. incrassata. It is possible to assemble an almost complete series of
specimens to show the transitions from the very small-leaved form typical
of C. incrassata to the larger leaves of C. fuertesti. In almost every char-
acter selected, these two species tend to be closely associated. Nevertheless
I am hesitant to merge the two without further field study of this com-
plex. On a field label associated with his collection H-4911 which was
cpa by Schmidt to C. retusa, Ekman writes, “evidently the same as
and n. ———. Resembles C. retusa Griseb. but is of course
Aca being in fact related with C. incrassata through a series of in-
termediate species.” Few taxonomists split species finer than did Ekman
and Schmidt; nevertheless it is interesting that Ekman felt he had seen
and collected intermediate “species” between the two, although he never
filled in the collecting numbers on his field label. For the present I dis-
tinguish C. incrassata on the basis of smaller leaves with fewer veins and
with two pairs of these diverging near the base of the blade. The vena-
tion in specimens of C. fuertesii also differs from that of C. incrassata,
in which the primary veins run straight to the margin before curving and
fusing with the cartilaginous leaf margin. In C. fuertesm the veins divide
or arch before reaching the margin, forming a network free from the car-
tilaginous margin. A parallel set of characters has been used to distin-
guish C. retusa and C. northropiae in Cuba and the Bahamas. Additional
field study will be required to determine the value or validity of this dis-
tinction. The close association of C. fuertesit and C. incrassata is also
indicated by the fruits, each possessing the sterile corky base to the achene.
Only a few fruiting collections of C. incrassata are known, but in these
the fruit appears to be smaller and the sterile base less differentiated than
that of C. fuertesit.

ues krugii Lindau, Engl. Bot. Jahrb. 13: 145. 1890; Symb. Antill.
. 1899; Howard, Jour. Arnold Arb. 37: 337. 1956.

Coccoloba bgrgeseniit Schmidt, Fedde Rep. Spec. Nov. 24: 75. 1927.

26 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. XXxIx

Coccoloba borgesenii forma ovato-lanceolata Schmidt, Fedde Rep. Spec. Nov.

24: 76, 1927

Shrub or small tree to 19 feet tall; branches terete, glabrous, slightly
geniculate and nodose; ocreae membranaceous; persistent, 3-5 mm. long;
leaves of normal shoots with petioles 5-6 mm. long, corky at the base,
arising from the bases of the ocreae; blades ovate to suborbicular, 2
1.8, 4 & 3.5, 5 & 4 cm. long and broad, thin-coriaceous, glabrous or
rarely with a few hairs near the attachment of the petiole, the apex obtuse
or rounded, the base cordate or rounded, the margin flat or recurved;
midrib flat above, slightly prominent below, the primary veins 4—6 pairs,
straight, bifurcating and anastomosing near the margin, flat on both sur-
faces, the secondary venation minutely reticulate below, smooth above;
leaves of adventitious shoots with petioles 1 cm. long, the blades cordate
or elliptic to 7 X 6 cm. long and broad; inflorescence terminal 5—8 cm.
long, the rachis glabrous, the staminate flowers 1-3 per node, the pistillate
flowers borne singly, the bracts broadly ovate, membranaceous, 1 mm.
long; ocreolae membranaceous, flaring to 1 mm. long; pedicels wanting
or shorter than the ocreolae, the hypanthium 1 mm. long, the perianth
lobes ovate, to 2 mm. long, the filaments of fertile stamens 1.5 mm. long;
fruit ovoid or angularly fusiform, strongly triangular in cross section,
4-5 mm. long, 3—3.5 mm. in diameter, the perianth lobes appressed, about
half the length of the fruit.

DistRiBUTION: The Bahamas, Hispaniola, Jamaica, Puerto Rico, Ana-
gada, Antigua, Barbuda, St. Martin.

Dominican Republic. Prov. Monte Cristr: El Morro, Ekman 13143 Ss,
R.A. & E.S. Howard 12532 (cH), 12534 (Gu), Jiménez 1356 (A).

Haiti. Dept. pu Norp Ouest: Isle Tortue, Morne Barranca, Ekman 4314
(B-type of C. borgesenii, s, us); Port au Paix, Vallée des Trois Rivieres, Ekman
3588 (B-type of C. borgesenii forma ovato-lanceolata, Ss, us); Port au Paix,
E.C. & G.M, Leonard 15252 (A, GH, S, US)

3

Coccoloba krugii resembles C. praecox of Cuba and C. buchii of His-
paniola. It can be distinguished from these species by the angular rachis
of the inflorescence and the angularity of the fruit.

Coccoloba bérgesenii was described by Schmidt as having a puberulent
rachis to the inflorescence. This “puberulence” appears to be a mixture of
fungal hyphae, crystals of mercuric chloride and fibers of pressing ma-
terial, Coccoloba brgesenii forma ovato-lanceolata is based on a speci-
men representing the adventitious shoots of this species.

Coccoloba krugii Lind. X Coccoloba uvifera L.; Howard, Jour.
Arnold Arb. 38: 216, 217. 1957.
Coccoloba scandens Ekman, Bull. Estac. Bot. Moca, Ser. B. 17: 14. 1927,
nomen.
Shrub of 6 feet or small tree with habit of Coccoloba uvifera, young
branches terete, striate, puberulent to pubescent; ocreae membranaceous,

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 27

oblique and slightly flaring at the apex, 1-1.5 cm. long, puberulent to
pubescent; leaves of normal shoots on petioles 7-10 mm. long, the blades
ovate to ovate-elliptic, 6 « 3, 8 * 6,11 * 8 cm. long and broad, the
apex obtuse to broadly rounded, rarely acuminate with an obtuse point,
the base oblique, cordate to rounded, one basal lobe often overlapping
the petiole; midrib and veins prominent below, sub-prominent above when
dry, the ultimate venation reticulate, the primary veins 6 or 7, arcuate and
anastomosing near the margin, 2 or 3 veins close to the base of the blade;
leaves of adventitious shoots on petioles 7-10 mm. long, the blades ae.
ovate- elliptic or rarely ovate-lanceolate, 14 « 8 to 29 x 18 cm. long
and broad; inflorescence simple, terminal, to 20 cm. long, the rachis
slender, puberulent becoming glabrate; bracts broad, triangular-ovate,
puberulent, the ocreolae membranaceous, to 1 mm. long, puberulent, the
pedicels shorter than the ocreolae; staminate flowers 2—4 per locus, the
pistillate flowers solitary at the nodes, the hypanthium to 1 mm. long,
the perianth lobes orbicular, 1.5-2 mm. long and broad, the filaments of
fertile stamens 2 mm. long; pedicels in fruit equalling to twice as long
as the ocreolae, the fruit fusiform to fusiform-ovoid, 11 mm. long, 8 mm.
in diameter, the perianth lobes not coronate.

DISTRIBUTION: Puerto Rico, Virgin Islands, Hispaniola.

Dominican Republic. Prov. Monte Cristi: Los Siete Hermanos, Howard
12524 (GH), 12525 (GH), 12526 (GH), 12527 (GH), 12528 (GH), 12529 (GH);
Ekman H-13164 (s, us).

Haiti. Dept. pu Norp Quest: Tortue Island, La Vallée, E.C. & G.M.
Leonard 15352 (us), 11320 (a, xy, US); Boucan-Guepes, Ekman H-9732 (s, us):
Mare-Rocher, Ekman H-4122 (s, us); Mole St. Nicolas, £.C. & G.M. Leonard
13183 (mo, us), Port de Paix, £.C. io G.M. Leonard 11180 (ny, us), Ekman
H-3935 (s).

This hybrid was recognized and described originally in relation to
material from Puerto Rico and the Virgin Islands. Subsequent study of
these troublesome populations and of the several series of collections cited
above have shown additional and isolated occurrences of this hybrid in
Hispaniola.

Ekman first discovered one population in 1925 on [le de la Tortue and
suggested in his field notes that it was a new species related to Coccoloba
uvifera. A few days later he encountered it again along the sea coast
west of Saline Michel near Port de Paix and indicated in his field notes
that while sterile it was not the same as the Tortue plant and that it was
“not C, uvifera, by no means.’ Schmidt annotated these specimens as
C. verruculosa Lindau; Leonard, who collected in the same areas, also
found the plant and used the same species identification. In 1929 Ekman
visited Monte Grande in the Los Siete Hermanos island group off Monte
Cristi and collected the plant there. In his field notes he identified the
plant as C. verruculosa again, but comments, ‘“‘I am beginning to feel sus-
picious about this. It is altogether too widely distributed and common
to have escaped notice of the old collectors; e.g., Jacquin. How about

28 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

C. leoganensis?” In a brief published report of this last island trip
Ekman refers to the occurrence of “Coccoloba scandens” on Monte Grande.
Coccoloba scandens is published without description and, in any case, is
a later homonym of Coccoloba scandens Benth., also a nomen nudum.
The epithet “scandens” however was particularly appropriate, for several
of the plants seen on Monte Grande had scrambling branches. As was
true with populations of this hybrid in Puerto Rico and the Virgin Islands,
there is a considerable range of variation between plants, accentuated in
some cases by the normal variation between adventitious and juvenile
shoots and those of more mature plants.

Coccoloba krugii, one of the suggested parents of this hybrid, is present
on Tortue Island, Méle St. Nicolas and near Port de Paix where the
hybrid has been collected. It was not found on any of Los Siete Hermanos,
but occurs on El Morro near Monte Cristi. Coccoloba uvifera has been
found in all locations where the hybrid occurs.

Coccoloba leoganensis Jacq. Enum. 19. 1760; Hist. Stirp. 113. pl. 178,
f. 33. 1763; Pl. Amer. Pict. t. 60, f. 30, 1780.

Coccoloba rotundifolia Meisn. DC. Prodr. 14: 154. 1857..

Coccoloba uvifera var. leoganensis Willd. Sp. Pl. 2: 457. 1799.

Small tree to 20 feet tall, d.b.h. 6 inches; branches generally tortuous,
striate, minutely pubescent, the nodes slightly swollen; ocreae 6 mm. long,
membranaceous, minutely puberulent, obliquely truncate, light gray-green,;
leaves of normal shoots with petioles 7-15 mm. long, puberulent, swollen
at the base, arising from the bases of the ocreae; blades orbicular, 3 X 3,
5 & 6,7.5 & 8 to 9 & 9 cm. long and broad, subcoriaceous, glabrous
above, glabrate below or with pubescence in axils of veins or along the
midrib, the apex rounded, emarginate or obtusely short acuminate, the
base shallowly cordate, the margin entire but generally slightly undulate;
midrib and primary veins evident above and only slightly conspicuous
below, the primary veins 5—7 pairs, the secondary venation evident but
not conspicuous; fast-growing shoots not seen; adventitious shoots from
trunk of tree slow-growing, tortuous and flattened, the leaf-bearing
branches terete; leaves with petioles 2-3 mm. long, the blades similar
in shape to those of normal branches 0.4 x 0.4 to 1.2 x 1.5 cm. long and
broad; inflorescences terminal, 5-7 cm. long, the basal ocreae to 1 cm.
long, the rachises puberulent, densely flowered, commonly thin to tenuous
in flowering condition becoming thick in fruiting condition; staminate
flowers 1-3 at each locus, the pistillate flowers generally borne singly at
each locus, the bracts ovate, to 1.5 mm. long, the ocreolae 1—-1.5 mm. long,
membranaceous, puberulent; pedicels 1.5—2.0 mm. long, puberulent. articu-
lated at the base and dehiscent after flowering, the flowers articulated, the
hypanthium 1 mm. long, commonly narrowed to the pedicel, the perianth
lobes broadly orbicular, 2-3 mm. long and broad, the functional stamens
1-1.5 mm. long, the fertile pisti] 2.5 mm. long, the sterile stamens and

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 29

pistils abortive, to 0.5 mm. long; fruit ovoid, 4.5 mm. long, 3.5 mm. in
diameter, narrowed at the base, the fruiting perianth red in color, the
lobes tightly imbricate, longer than the hypanthium; achenes obtusely
3-angled, greenish brown.

Locat NAMES: Uvero (DR), uvilla (DR).

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. DistricTo DE SANTO Dominco: Ciudad Trujillo, Schif-
fino 164 (cH). Prov. BaraHona: Along the Rio Yaque, Fwertes 188 (BM
EDIN, F, GH, K, M, MO, NY, $, US). Prov. INDEPENDENCIA: Puerto Escondido,
Howard 12185 (GH). Prov. LA VEGA: Cuesta de Piedra, Eggers 2376 (BM, GH, K,
M, NY, us). Prov. Lrpertapor: Between Restauracién and Banica, Howard
12555 (cH). Prov. Monte Crist1: Copey, Howard 12522 (cH); Villa Isabela,
Howard 12521 (cH). Prov. SANTIAGO: Santiago to Onna Eggers 2429

A, B, C, US); Sabaneta, Jiménez 1737 (GH). Prov. TRUJILLO-VALDEz: Bani,
Howard 12045 (GH), 12048

Haiti. Dept. pu Norp Ouest: Port de Paix, Vallée des Trois-Rivieres, Ekman
H-3991 (s, us); Bassin Bleu, E.C. & G.M. Leonard 14603 (A, NY, US), 14782
(us); Cabaret, Baie des Moustiques, E.C. & G.M. Leonard 11820 (MIcH, US),
12051 (A, GH, K, MO, NY, US); Jean Rabel, E.C. & G.M. Leonard 12071 (cH,
Us). DEPT. DE L’ARTIBONITE: Gonaives, Buch 654 (B); Gonaives to La Hotte
Roche, Nash & Taylor 1586 (NY); Gonaives to La Branle, Nash & Taylor
1592 (NY, US). DEPT. DE L’OUEST: Between Port-au-Prince and Léogane, Jacquin
sm. (BM-holotype); Port-au-Prince, Ekman H-2110 (s); Montagnes du Trou
d’Eau, Morne a Cabrits, Ekman H-990 (aA, s, us); Ile de la Gonave. Pte. a
Raquette, Eyerdam 54 (A, F, GH, MO, NY, us); Petit Gonave Island, Leonard
5233 (Ny, US). Dept. NoT KNOWN: Cape Francois, Ehrenberg s.n. (B, G-type
hi a, Senin NY).

Jacquin’s descriptions of Coccoloba leoganensis are brief and the illus-
tration cited above is of a single leaf which lacks any distinctive charac-
teristics. All authors since Jacquin have referred C. leoganensis to the
synonymy of C. uvifera. Willdenow established C. /eoganensis as a variety
of C. uvifera and Meisner accepted this placement, but with a query.
Mr. Dandy, of the British Museum, kindly determined for me that the
Jacquin specimen was there and verified the similarity to C. rotundzfolia.
I have since had the opportunity of seeing the material myself and it
seems that at last the elusive name “‘/eoganensis” can be properly placed.
It certainly did not seem reasonable to me that Jacquin could not distin-
guish C. uvifera from other species.

The species Coccoloba rotundifolia Meisner was compared originally
by the author with C. leoganensis and distinguished on the basis of
smaller leaves. While Meisner did have a specimen from the smaller end
of the range in size, his material can easily be matched with the more
abundant specimens cited above.

Coccoloba leoganensis is a distinctive species both as living plants
_and as herbarium specimens. The habit of the plant, a low, densely

30 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

branched and rounded or flat-topped tree is striking. When in full
flower the plants seen attract thousands of bees, so that collecting speci-
mens becomes almost too competitive. In this species alone among those
known from the Greater Antilles, the pedicels absciss immediately after
flowering in the staminate inflorescences, and, if fertilization has not been
effected, in the pistillate. A staminate inflorescence may have a few fully
mature flowers at the apex and be naked and pitted below where flowers
had been. This is particularly noticeable in fresh condition.

The fruits are described here for the first time. They are bright red
in color and astringent, without any pleasant taste. The fruit is con-
stricted at the base to a short stalk. The fleshy perianth lobes surround
the achene tightly with only three lobes visible on the surface. The
bracts and ocreolae enlarge slightly in fruit. This combination of char-
acters would place the present species in the section Campderia of Lin-
dau’s classification. In total aspect the fruiting inflorescence resembles
that of Coccoloba nivea, currently the only other West Indian species in
this section.

Considerable individual variation in the amount of pubescence on the
lower leaf surface was also noticed in living specimens in the field. The
majority of mature leaves had a small amount of pubescence in the axils
of the lateral veins on the lower leaf surface and additional pubescence
along the midrib. On other plants the leaves were entirely glabrous below.

Perhaps the most unusual feature of Coccoloba leoganensis is the size
of the leaves produced by lateral and adventitious shoots. Root sprouts
and lateral stem sprouts were found on a number of trees near Bani in
the Dominican Republic. This was the largest stand of C. leoganensis
seen and here the species was the dominant tree in the dry thorn-shrub
region at kilometer 80. The sprouts produced near the base of the trunk
and on the trunk of these trees were short, compacted and contorted, in
contrast to the normal elongate, fast-growing trunk sprouts of other
species. In addition, the leaves are exceedingly small, ranging in size from
4 < 4mm. to 12 & 15 mm. on petioles 2-3 mm. long. The leaves of these
sprouts were generally orbicular and rounded at the apex, although some
were ovate and obtuse at the apex. These sprouts all resembled C. sub-
cordata. Careful attention was required to determine that these were parts
of the plant of C. leoganensis. Even now, in dealing with the herbarium
specimens, it has been necessary to recheck field notebooks to make certain
that this point had been determined in the field and that the labels were
not in error. Several of Leonard’s collections (12071, 11820 and 14603)
posed the same problem, although they had leaves grading to a larger
size. In all of these collections, the petioles arise from the base of the
ocreae, as in C. leoganensis, and not from the apex of of the ocreae, as
in C. subcordata. Coccoloba subcordata has an entirely distinct habit in
the field and has short inflorescences, but in sterile herbarium specimens
that two may well be confused. This is the only species of Coccoloba so far
encountered in which the leaves of the sprout shoots are smaller instead
of larger than the leaves of mature branches.

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 31
Coccoloba leonardii Howard, Jour. Arnold Arb. 30: 419. 1949.

Tree to 30 feet tall: branches terete, the nodes only slightly tumid,
glabrous; ocreae subcoriaceous, the persistent part 3 mm. long, glabrous;
leaves of normal shoots with petioles 8-11 mm. long, glabrous, arising
about 0.5 mm. above the bases of the ocreae, the blade ovate, 6.5 X* 4,
8 x 5.5 to 11 & 7 cm. long and broad, coriaceous, the apex rounded to
acute, the base rounded to slightly cordate unequal, usually slightly de-
current on the petiole, the margin entire, flat; midrib not conspicuous
above, the primary veins 5—7 pairs, ascending, arcuate, anastomosing near
the margin, the ultimate venation reticulate, leaf surface glabrous, drying
to a chestnut brown or ash color, the blocked stomata as seen on the lower
surface brown in color; leaves of adventitious or fast-growing shoots ellip-
tic, rounded or apiculate at the apex, often strongly cordate at the base,
not larger than leaves of normal shoots in material seen; inflorescences
terminal, 5.5-10 cm. long in fruit; staminate flowers 2—4 at each locus,
the pistillate flowers solitary at each locus, the bracts 0.5 mm. long, the
ocreolae about 1 mm. long, the pedicels short, less than 0.5 mm. long in
fruit; hypanthium 0.5—-1 mm. long, the perianth lobes 1.5 mm. long and
broad. the fertile stamens 1 mm. long, the fertile pistil 2 mm. long, the
sterile stamens and pistil rudimentary, less than 0.5 mm. long; fruit
broadly fusiform, 10-11 mm. long, 5-7 mm. broad, the perianth lobes
slightly coronate, these 1 mm. long, many-ribbed, drying black above and
reddish brown below, the achene elongate, dark brown and shiny.

DISTRIBUTION: Haiti and Cuba.

Haiti. Dept. pu Norp Ouest: Tortue Island, Basse Terre, F.C. &
Leonard 12466 (a-holotype, Mo, Ny, US), La Vallée, E.C. & GM. Leonard
11335 (MICH, US), 11381 (A, GH, K, oa 11421 (Ny, us), i478 7 (Cs US); ete:
Petite wee Ekman 4150 (s), Saline Michel near Port au Paix, Ekman 3931

Ss). vu Norv: Morne la Vigie, Cap Haitien, Ekman 2706 (s, US);
ene ‘Nash 203 (ry, NY). Depr. pu Sup: Navassa Island, Ekman 10843
(Sus)

Coccoloba microstachya Willd. Sp. Pl. 2: 459. 1800; Lindau, Engl.
Bot. Jahrb. 13: 146. 1890; Howard, Jour. Arnold Arb. 37: 332. 1956,
BBP 21721957.
Coccoloba klotzschiana Meisn. DC. Prodr. 14: 155. 1856
Coccoloba obtusifolia Lindau, Symb. Antill. 1: 22. 1899, not Jacquin.

Shrub or tree to 20 feet tall; branches terete, the nodes tumid, pubescent
or with hair primordia, the bark gray to tan in color; ocreae membranace-
ous, cylindrical, pubescent, 4 mm. long; leaves of normal shoots with
petioles 3-6 mm. long, pubescent, arising from the bases of the ocreae;
blades variable in size and shape, ovate, ovate-lanceolate, oblong or ellip-
tie 65° 15.4 x 2,515 > 3 to 4 cm. long and broad, thin-
coriaceous, usually turning black on drying, the apex acute, acuminate,
rounded or emarginate, the base narrowed, rounded or slightly cordate,

JZ JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxx1x

the margin entire; midrib and vein evident but not conspicuous above,
prominent below, primary veins 7—9 pairs, all glabrous above, pilose or
glabrate below rarely persistently and densely pubescent; leaves of adven-
titious shoots on petioles to 7 mm. long with blades ovate-lanceolate to
oblong, reaching 12 x 2.5 or 17 & 4 cm. long and broad, these generally
more pubescent than leaves of mature shoots; inflorescence terminal, 5—10
cm. long, the rachis usually pubescent, tenuous, often geniculate, commonly
recurved; staminate flowers 2 at each locus, the pistillate flowers solitary
at each locus, the bracts broadly ovate, to 0.5 mm. long, puberulent, the
ocreolae membranaceous, puberulent, to 0.5 mm. long, erect on the stami-
nate plants, generally appressed or flattened on the pistillate, the pedicels
shorter than the ocreolae or lacking, the hypanthium less than 1 mm.
long, the perianth lobes 1-1.5 mm. long and 1 mm. wide; fruit sessile,
generally ovate with distinctly coronate oorante lobes, to 6 mm. long a
4 mm. in diameter.

DIsTRIBUTION: Dominican Republic, ati Rico, St. Thomas, St. Jan,
Tortola, Virgin Gorda, Anguilla and St. Croi

Dominican Republic. Prov. ALTAGRAcIA: La Romana, Ekman H-12098 (s,
us). Prov. Monte Crist: summit of El Morro, Howard 12537 (cH). Prov.
Puerto PLata: Puerto Plata, at La Boca, Ekman H-14381 (a, S, US); Puerto
Plata at Arroyo Francés, Ekman H-14400 (s, us); between Puerto Plata and
Maimon, Eggers 2674 (8, mM). Prov. SAMANA: Cabo Samana, Ekman H-14905
(s); Santa Barbara de Samana, Ekman H-15322 (s, s).

The variation in the size and shape of leaves and the amount of pubes-
cence on plants of this species has been discussed in earlier papers in
relation to Coccoloba swartzti. The few specimens from the Dominican
Republic cited above represent nearly the extremes of variation, Most
unusual is the shape of the leaves of adventitious shoots in several speci-
mens. These are oblong — nearly linear-oblong —in shape. A similar
variation also occurs in C. costata on the adventitious shoots of plants
from the Samana peninsula and from the eastern end of the island of His-
paniola.

Coccoloba nodosa Lindau, Engl. Bot. Jahrb. 13: 147. 1890.
Uvifera nodosa Ktze. Rev. Gen. 2: 561. 1891.

Small tree or bush 10-15 feet tall; branches of mature plant conspicu-
ously nodose with short internodes, the branchlets often moniliform where
leafless, sparsely short pubescent with golden to chestnut hairs; ocreae
membranaceous, 3 mm. long, the apex acute, short pubescent: leaves of
normal shoots with petioles 2 mm. long, puberulent, arising from the
bases of the ocreae, the blades obreniform, orbicular or oblate, 0.5 « 0.9,
I x 2 to 2 & 2.5 cm. long and broad, coriaceous, generally drying black,
sparsely short pubescent on both surfaces becoming glabrate, the apex
broadly and shallowly emarginate or truncate, the base cuneate or
rounded, the margin entire and flat, the blade flat: primary veins 3—5

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 33

pairs, mostly straight to the margin, then bifid and anastomosing, the
venation evident above but slightly prominent below; adventitious shoots
with nodes slightly swollen but not bead-like, ak internodes striate, all
densely ferruginous short-pilose, the ocreae 6-8 mm. long, deeply bilobed
at the apex, membranaceous and deciduous gee chestnut-colored and
pubescent, persistent below, this portion gray-green in color, the leaves
with petioles 2-4 mm sion densely gray pilose, arising from the bases
of the ocreae, the blades cordate to ovate, to 6.5 cm. long and 5 cm. broad,
the apex acute to obtuse, the base cordate to truncate, the margin un-
dulate, the lateral branches of such shoots producing normal-sized leaves;
inflorescence terminal 4-10 mm. long, puberulent to glabrous, the bracts
triangular, membranaceous, to 1 mm. long, the ocreolae membranaceous, less
than 1 mm. long; staminate flowers 1-3 at each locus, the pedicels 1 mm.
long, the hypanthium 0.5 mm. long, the perianth lobes broadly ovate to
orbicular, 1-1.5 mm. in diameter, the stamens 1-1.2 mm. long, the pistil
rudimentary, less than 0.5 mm. long; pistillate flowers and fruit not known.

DISTRIBUTION: Endemic to the Dominican Republic.

Dominican Republic. Districro pE Santo Dominco: Llano Costero, El
Manielito, Ekman H-11289 (a, s, us); Llanura de Santo Domingo at Las Rosas,
Ekman H-5804 (s); La Yuca, imeee 3401 (A); on road from Santo Domingo
City (Ciudad Trujillo) to Puerto Plata, Wright, Parry and Brummel 475 (cH,
NY, US). Without specific location: Bertero 928 (B-type, M), R. Schomburgk
65 (B).

Until recently I had considered the sterile collection Ekman 5809 to
represent the adventitious shoot condition of this species. Several months
ago Dr. José Jiménez of the Dominican Republic sent for identification
a small specimen of Coccoloba which I assigned to this species. Since I
had not been able to locate plants of this species during several visits to
the Dominican Republic, I asked Dr. Jiménez to obtain additional material
to show all possible leaf variation in this plant. Such specimens were col-
lected for me by Professor Marcano of the Instituto Botanico ‘Rafael M.
Moscoso” and were forwarded by Dr. Jiménez. I am indebted to both
gentlemen for the material essential to a better understanding of leaf
variation in Coccoloba nodosa. Unfortunately, the plants under observa-
tion have not flowered and hence several details of the description remain
uncompleted.

Coccoloba nodosa may eventually include C. picardae. When Urban
described the latter species he disinguished between them on the basis of
the swollen nodes and obtriangular leaves of C. nodosa. He also noted
that the flowers are in multiples in C. nodosa. This aggregation of flowers
or pedicels is usually found in the staminate flowers, while the pistillate
flowers are borne singly at each locus on the rachis. On the basis of nor-
mal shoots the leaves of C. nodosa and C. picardae are similar; however,
until pistillate flowers and fruits are obtained for C. nodosa and adventi-
tious shoots for C. picardae, the species should be kept distinct. Coccoloba

34 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

nodosa has been collected at lower elevations while C. picardae has been
found only at elevations above 5000 feet

Coccoloba pauciflora Urban, Symb. Antill. 7: 209. 1912.
Coccoloba aa as Fedde Rep. Spec. Nov. 19: 1. 1923; Ark.
Bot. 20A(15): 30. 19
Coccoloba een? ee Fedde Rep. Spec. Nov. 13: 446. 1914; Ark.
Bot. 20A(15): 30. 1926.
Coccoloba fulgens Leonard, Jour. Wash. Acad. Sci. 17: 66. 1927.
Coccoloba nalgensis Schmidt, Fedde Rep. Spec. Nov. 32: 80. 1933.

Shrub or tree to 28 feet tall; branches terete, the nodes not swollen,
short ferruginous pubescent, almost pilose or becoming glabrous, the hair
bases and the often resinous stomatal excretions frequently resembling
peltate scales; ocreae membranaceous, 3-8 mm. long on normal shoots,
oblique at the apex, ferruginous pubescent; leaves of normal shoots with
petioles 2-10 mm. long, generally pilose at least on the adaxial surface,
arising at the bases of the ocreae, the blades obovate to obovate-elliptic,
1.5 «& 0.8.2 * 1.4,3 & 1.5,4 &* 2 to 5 & 3 cm. long and wide, sub-
coriaceous, glabrous or with conspicuous or inconspicuous stomatal excre-
tions, the apex rounded to shortly and abruptly acuminate, occasionally
truncate, often asymmetrical, the base cuneate; primary veins 3-5 pairs,
prominent on both surfaces when dry; adventitious shoots terete, the
ocreae to 1.5 cm. long, conspicuously ferruginous pilose or puberulent, the
leaves with petioles 1.3-1.7 cm. long, the blades ovate-elliptic to ovate-
lanceolate and 17 & 9 or 14.5 8 cm. long and broad, these broadly
cuneate to acute at the base, broadest below the middle and acuminate at
the apex, or blades obovate to obovate-lanceolate to 14 * 7 cm. long and
wide, these broadest above the middle, cuneate at the base and acute at
the apex; inflorescences terminal on axillary shoots of varying lengths and
ages, the youngest often appearing as axillary clusters of flowers, the
axis 2-3 rarely 4 cm. long, densely ferruginous pilose or puberulent at
the base or, when young, soon becoming glabrate, the bracts broadly tri-
angular, puberulent, the ocreolae 0.5—0.75, rarely 1 mm. long, tubular and
not spreading: pedicels 1 mm. long in flower, those bearing staminate
flowers increasing to 2 mm. long before or after abscission of flowers;
staminate flowers in clusters of 2—3, the pistillate flowers borne singly at
each locus; hypanthium 0.5-0.75 mm. long, the perianth lobes orbicular,
1-1.5 mm. long and broad, the functional stamens 1.2—1.4 mm. long, the
sterile stamens less than 1 mm. long, the functional pistil 2 mm. long, the
abortive pistils less than 0.5 mm. long; fruit brilliant to dull red when
fresh, reddish brown when dry, ovoid to pyriform 6—7 mm. long, 4—5 mm.
in diameter, slightly coronate at the apex, the fruiting hypanthium ex-
tending above the middle, the lobes of the perianth imbricate and cover-
ing the achene at the apex; achene chestnut-brown.

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. BARAHONA: El Aguacate to crest on road to

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V a0

Pedernales, Howard 12572 oes Beane (cH); La Tierra Fria, Howard 12210
(a); Polo, Howard 12239 (GH). Prov. La VEGA: Constanza, ene 2955 (B-
lectotype, M, NY). PRov. a Samana, Laguna, Los Bafaderos Prietos,
Ekman H-15135 (s, us). Prov. SAN RaraEL: Loma Nalga an Maco. Ekman
H-6322 (s-holotype of C. nalgensis, s, US). PRov. SANTIAGO RODRIGUEZ: Moncion,
Ekman H-12085 (B, S, US).

Haiti. DEPT. DU oe Ouest: Massif du Nord, Port de Paix, Haut Piton.,
Ekman H-3697 (B, S, US); Bassin Bleu, Morne Haut Piton, £.C. & G.M.
Leonard 15043 (A, CH, ae 15046 (Mo, NY, US), 15064 (Us). Dept. DU NorpD:
Massif du Nord, Bayeux, Morne Brigand, Ekman H-2853 (s, us). DEPT. DE
L’ARTIBONITE: Ennery, Puilboreau Pass, Leonard 9145 (B, GH, Ny, Us-type of C,
fulgens): Massif du Nord, Gros Morne, Morne Belance, Buch 811 (B-type of C.
mornicola), Ekman H-4910 (s). Dept. pu Sup: Massif de la Hotte, Nouvelle
Touraine, Ekman H-1657 (A, 8). Dept. DE L’OuEsT: Massif de la Selle. Morne
Tranchant near Godet, Ekman H-1953 (s); Massif de la Selle, Petionville,
Ekman H-1657 (us); Massif des Matheux, Grand-Bois, Cornillon, Ekman H-
5686 (s, us); Massif de la Hotte, Petit Gone road to Morne Calumette. Ekman
H-7304 (s); Fond Verettes, Leonard 5347 (B-type of C. neurophylla, Us)

To the present, Coccoloba pauciflora has been represented only by the
original collection of Tiirckheim; Coccoloba nalgensis, also, is known from
the type collection alone and the several collections cited above, assigned
indiscriminately to either C. mornicola or C. neurophylla. Both Coccoloba
mornicola and Coccoloba neurophylla were described on sterile material
and in each case Urban published a note adding to the description when
additional material became known. Unfortunately, the additional material
has been staminate in the case of C. mornicola and pistillate in the case of
C. neurophylla. Here again arises the necessity of combining names and
species based on the erroneous idea that the flowers are perfect. The
study of a considerable number of plants and populations of this species
in the Dominican Republic has allowed me to describe the range of varia-
tion found in the leaves, pubescence and inflorescence of this species.

Coccoloba pauciflora was so named for the short inflorescence axis and
the few flowers produced. The nature of the leaf margin was also a char-
acteristic which Urban used in describing the plant. It is possible to find
short inflorescences at the top of many plants where the inflorescences
terminate axillary branches. At a lower level in the tree, however, the
lateral branches, which are little more than short shoots, have elongated
slightly and the inflorescence produced is longer.

Leaf variation is considerable, both within a population and on the
same plant, as to size, texture and apex. The shape of the leaf apex is
related either to the development of the midrib or to its failure to develop
at the apex. Three conditions are found on different plants, or on the
same tree on one branch, or on different branches: (1) the midrib can be
uniformly developed to the apex, in which case the tip of the blade is
acuminate; (2) the midrib may apparently fail to develop at the apex of
the blade, in which case this area is vascularized by the upper pair of
primary veins; or (3) a remnant of the midrib may be present or absent.
In this situation the apex of the leaf is obtuse or rounded.

36 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

In the third condition observed, the midrib fails to develop to the tip
of the blade and one of the upper pair of primary veins dominates the
other. In this case the leaf apex is asymmetrical. Examples of the three
tvpes of leaf apices on one plant are found in the collections of Ekman
H-2853, H-1657 and Howard 12586, the latter collected especially for
this purpose. .

The smaller range of leaf variation in this species is represented by the
specimen Ekman H-6322 which Schmidt selected as the type of his Cocco-
loba nalgensis. The herbarium sheets of this collection bear an unpub-
lished specific name based on the province of Azua where, it is said, this
collection was made. Loma Nalgo de Maco, as now known, is in San
Rafael Province. While Schmidt describes the leaves as 1-2 cm. long and
0.8-1.2 cm. wide, the majority of the leaves on the three examples of
this number are at the smaller end of the range given. Several more recent
collections (e.g., Ekman H-12805 and Howard 12572) are intermediate
between the type collection of C. nalgensis and the majority of the speci-
mens cited above. One collection (Howard 12572) consists of several
specimens from a single tree made to show a full range of variation from
the smallest leaf and shortest inflorescence to the larger ones more typical
of the expanded species concept here employed. Unfortunately, this single
tree did not have any adventitious shoots. Plants of this series approach
C. picardae in form and it is possible that additional collections will
demonstrate that C. picardae should be included in this species, in which
case C’. picardae, the earlier name, must be used.

Several excellent examples of plants with well developed adventitious
shoots were found in the mountains around Barahona. One of these
(Howard 12239) was a 25-foot tree in full flower. Several of the lateral
branches possessed side shoots with normal leaf size and shape (obovate
with a rounded apex 3 cm. long and 1.5 cm. wide), while the apex of the
branch had longer internodes and larger leaves, some of these reaching a
length of 10 cm. and a width of 4 cm. Strict aavcnp: shoots arising
from the base of this tree were wand-like with long internodes and obovate
leaves 14 cm. long and 7 cm. wide above the middle. A second plant in the
same general area was growing on a steep hillside and at an angle. The tree
was sterile but the apex of the plant had branches with obovate leaves
averaging 3.7 cm. long and 1.7 cm. wide with acute to acuminate apices.
From the trunk of this plant were developed numerous adventitious
shoots, these all arising vertically and at an angle to the tree. The leaves
on these adventitious shoots were ovate-elliptic to ovate-lanceolate and
broadest at or slightly below the middle. The leaves ranged to 17 cm.
long and 9 cm, wide. They were broadly cuneate to acute at the base and
acuminate at the apex. The contrast between the normal foliage and that
of the adventitious shoots was startling when seen in the field and: was al-
most unbelievable when the herbarium specimens were studied in the
laboratory.

An Ekman collection from the Samana peninsula is referred here. Al-
though it is sterile and represents an adventitious shoot, it matches the

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 37

material of Howard 12239. Schmidt studied this collection and referred
to it as “C. subtruncata forma.” Neither C. pauciflora nor “C. subtrun-
cata” have been reported from the Samana peninsula.

Coccoloba picardae Urban, Symb. Antill. 5: 336. 1907.

Shrub to tree of 30 feet; branches terete, the nodes not swollen, short
ferruginous-pubescent becoming glaborous, the branchlets often arranged
in one plane; ocreae short, 1-2 mm. tone ferruginous-pubescent; leaves
of normal shoots with petioles 1.5-2 mm. long, almost villose pubescent on
the adaxial side, arising from the bases of the ocreae; blades orbicular
forobtriangular, 1. << 1.2,51.6 >< 1.2... 1.7. x 17 to 2.5 2.5 ‘cm. long and
broad, coriaceous, stiff and rigid even when fresh, stomatal excretions evi-
dent on the lower surface, the apex rounded, subtruncate to submarginate,
often asymmetrical, the base rounded, to subtruncate or narrowed and
nearly cuneate, the primary veins 3-4 pairs, the venation reticulate and
conspicuous on both surfaces when dry; adventitious shoots or leaves
not known; inflorescences terminal on lateral short branches, often appear-
ing capitate, the rachis 2-5 mm. long, the bracts broadly triangular 0.5—
1 mm. long, the ocreolae to 1 mm. long, the pedicels 0.5—1.4 mm. long;
staminate flowers 1—2 per locus, the pistillate flowers 1 at each locus,
the hypanthium to 0.5 mm. long, the perianth lobes suborbicular 1.2—1.4
mm. long and broad, the fertile stamens 1—-1.5 mm. long, the sterile stamens
0.5 mm. long, the functional pistil 1.5 mm. long, the pistillate rudiment
less than 0.5 mm. long; fruit ovoid, 3-4 mm. long, 2-3 mm. broad, the
hypanthium shorter than the lobes in fruit, the achene not coronate.

DIsTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. BARAHONA: Crest of ridge between El Aguacate
and Pedernales, Howard 12594 (GH).
Haiti. Dept. pe L’Ouest: Massif de la Selle, Petionville on top of Morne
pee Ekman H-1163 (s, us), Picarda 784 (s-holotype), Buch 1614 (B),
3 (B), Leonard 4385 (GH, NY, US); Guimbi Galata, Morne des Commissaires,
nanan 1280 (GH, NY, US)

This species is very similar to the small-leaved variation of Coccoloba
pauciflora and may eventually prove to be part of the same complex.
For the present, the two species can be distinguished on the basis of the
very short inflorescences, the leaves broader than long, the non-coronate
fruit and the single-plane branching of C. picardae.

Coccoloba pubescens L. Syst. Nat. ed. 10. 1007. 1759; Hooker, Bot.
Mag. t. 3166. 1832; Fawcett & Rendle, Jour. Bot. 51: 123. 1913;
Howard, Jour. Arnold Arb. 38: 227. 1957.

Scortea arbor Americana, amplissimis foliis, aversaparte eit extantibus
1691.

hirsutie ferruginea refertis: Plukenet, Bie oaraphia t 222 at
Coccoloba rubescens L. Sp. Pl. ed. 2. 523. 1762.

38 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Coccoloba grandifolia Jacq. Enum. 19. 1760.
Coccolobis pubescens Sandwith, Jour. Bot. 78: 98. 1940.
Coccolobis antiguensis Sandwith, Jour. Bot. 78: 98. 1940.

Mature tree to 40 feet tall, d.b.h. 12 inches, much branched above a
well defined trunk; branches terete, swollen at the nodes, the lenticels not
conspicuous, tomentose to pilose: the ocreae to 1 cm. long, generally
completely deciduous, pubescent; leaves of completely mature plants
varying considerably in size and shape, the petioles 3-6 mm. long, in-
serted below the ocreae, densely short pubescent, the pints broadly
orbicular to orbicular-ovate, 4 « 6, 7.5 * 10 cm. long and broad, grad-
ing into size of leaves of adventitious shoots, rugose or bullate, the apex
rounded, the base cordate, the basal lobes rounded and only rarely ap-
proximate, sparsely pubescent above to glabrate, densely to sparsely
pubescent below or glabrate, the margin undulate, the venation of 5
pairs of primary veins, arcuate to the margin, strongly anastomosing,
slightly impressed above, conspicuous and reticulate below; adventitious
shoots generally strict and sparsely branched, to 30 feet tall, the branches
stout, terete, slightly swollen at nodes, strongly grooved or striate, the
ocreae 2 cm. long, membranaceous and evanescent above, coriaceous and
persistent below, the petioles stout 1-2 cm. long, densely tawny pubescent,
the blades large, generally orbicular except for the terminal leaf, fre-
quently broader than long, 30 & 40, 50 x 80 cm. long and _ broad,
coriaceous, rugose or bullate when mature, thin and plane when young,
the apex rounded, the base rounded to cordate, the basal lobes commonly
encircling the stem, the terminal leaf commonly rhombic, longer than
broad when mature, densely tomentose, the veins slightly impressed above,
all venation conspicuous and reticulate below, the midrib and secondary
veins persistent-pubescent above, the others sparsely pubescent when
young, becoming glabrate above, the veins and leaf surface pubescent
or becoming glabrate below, the margin irregular, commonly undulate;
inflorescences terminal, often stout, the basal ocreae to 7 mm. long, mem-
branaceous, the peduncle to 1.5 cm. long, the rachis minutely and often
densely puberulent, 10-18 cm. long on mature shoots, to 45 cm. long on
adventitious shoots, the bracts broadly ovate, about 1 mm. long, puberu-
lent, the ocreolae membranaceous, spreading, 1 mm. long, minutely puberu-
lent or glabrate; staminate flowers 2—4 per locus, the pistillate flowers i—3
per locus, the pedicels 2-3 mm. long, the hypanthium 0.5—-1 mm. long,
the perianth lobes broadly orbicular, 1.5 mm. long, 2 mm. wide, puberulent,
the fertile stamens 2.5 mm. long, the sterile stamens rudimentary, 0.5—1
mm. long, the fertile pistil glabrous or rarely slightly puberulent on the
ovary, the sterile pistils ge: rudimentary, 0.5-1.5 mm. long; fruit
globose to ovoid, 5-6 mm. long and 4-5 mm. in diameter, the fruiting
perianth imbricate at the apex, not coronate, the fruiting hypanthium
with conspicuous vascular bundles; ee sub-globose, dark brown,
shining, slightly triradiate at the apex, the fruiting pedicels puberulent,
3—4 mm. long.

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 39
Local NAME: Gamelle (H), Hojancha (DR).

DisTRIBUTION: Hispaniola, Puerto Rico, Barbuda, Antigua, Mont-
serrat, Nevis, Guadeloupe, Dominica, Martinique, St. Lucia.

Dominican Republic. Dist. DE SANto Domrnco: Ciudad Trujillo, Schiffino 137
(cH); Cuenca, R.A & ES. Howard 9884 (A). Prov. BARAHONA: El Caiman,
between Enriquillo and El Can, Howard 12187 (cH); Beata Island, Howard
12352 (cH); Mare-a-Chat, Ekman H-6947 (s, us). Prov. Espattyat: Moca,
Eggers 2559 (B, GH, M, NY, US). Prov. INDEPENDENCIA: Between Puerto Escon-
dido and Rancho Viejo, Howard 12143 (GH). Prov. Lipertapor: Between Res-
tauracion and Banica, Howard 12569 (GH). Prov. Puerto PLata: Hoja Anchas,
Jiménez 2088 (A). Prov. SamaNA: Pilén de Azucar, Abbott 401 (us). LocaLity
UNCERTAIN: Prince Paul s.n. (mM)

Haiti. Derr. pu Norp Quest: Bassin Bleu, E.C. & G.M. Leonard 15199 (a,
us); Mole St. Nicolas, Ekman H-4489 (s). DrEpt. pu Norp: Between Pignon
and Hinche, Holdridge 1272 (cu, us); Ranquitte, Christ 2090 (8). DEPT. DE
L’ARTIBONITE: St. ui de l’Atalaye, Leonard 7296 (Ny, US); Hinche, Ekman

H-6142 (A,S, US). pu Sup: Massif de la Hotte, Morne Rochelois, Charlier,
Ekman H-9035 (8, a “Th 9086 2 Miragoane, Eyerdam 396 (GH, NY, us):
Port-a-Piment, Ekman H-336 (s); Camp Perrin, Ekman H-249 (s). DEPT. DE

L’QuEstT: Petit Gonave Island, Leonard 5242 (s); montagnes du Trou d’Eau,
Fond-des-Oranges, Ekman H-2312 (s).

The variation in leaf shape and size in specimens of Coccoloba pubescens
encountered in the herbarium and as seen in the field has been discussed in
a previous paper (Howard, Jour. Arnold Arb. 38: 229-231. 1957). These
variations contrast the adventitious and juvenile shoot systems from
those of mature trees. A greater proportion of the specimens cited above
represents mature plants than would a comparable number of specimens
from other islands. A similar observation was made in the field; i-e., the
number of individuals of this species represented by mature Hlants. was
greater in Hispaniola than on other islands where the species seemed al-
most typified by the adventitious shoot condition of very large leaves.
In most of Haiti and western Dominican Republic the plants of Coccoloba
pubescens grew relatively unmolested. In eastern Hispaniola and on the
other islands of the Antilles, mature or undisturbed specimens were rarely
encountered and second-growth habit seemed most abundant. A popula-
tion of this species near Cuenca was visited first in 1946 and at two five-
year intervals since, but the plants, though larger and with some now in
flower, retain the characteristic adventitious leaves.

One known hybrid of Coccoloba pubescens with C. uvifera is described in
this paper and I have suggested the possible hybrid condition and origin
of C. fawcettii and C. ceibensis, with C. pubescens as one of the parents
of each.

Coccoloba samanensis Schmidt, Fedde Rep. Spec. Nov. 32: 81. 1933.

Small to medium-sized tree (fide Ekman); branches terete, lightly
striate, puberulent, the nodes slightly swollen; ocreae 6-10 mm. long,

40 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

membranaceous glistening, puberulent to glabrous; leaves of normal shoots
with petioles 6-8 mm. long, puberulent with shining hairs, 6-8 mm. long,
inserted at the bases of the ocreae; blades ovate to elliptic, rarely orbicular-
ovate, 5 X 4, 7.5 & 5 to 7-5 & 8 cm. long and broad, coriaceous, apex
obtuse, short and abruptly acuminate or rarely subtruncate, the base
obtuse to slightly cordate, the margin entire, slightly revolute, the midrib
impressed above, prominent below; the primary veins 6 or 7 pairs, in-
conspicuous above, prominent below, the ultimate venation minutely
reticulate; leaves of adventitious shoots with petioles 1.4 cm. long, similar
o those of normal shoots in shape, to 12 10 cm. long and broad;
inflorescences terminal, 8-18 cm. long, the rachis glabrous, lightly pu-
berulent to short pilose, the bracts broadly ovate, to 1 mm. long, the ocreolae
membranaceous, flaring, 2- or 3-lobed, to 1 mm. long, puberulent; stami-
nate flowers not known, the pistillate flowers on pedicels shorter than the
ocreolae, the hypanthium 0.5 mm. long, the perianth lobes obovate to
elliptic, 1.2 mm. long, puberulent, the stamens abortive, about 0.5 mm.
long, the ovary to 1.5 mm. long; fruit ovate with conspicuous coronate
perianth lobes, 3 mm. long and 3 mm in diameter, the vascular bundles
conspicuously developed, the achene globular, smooth, tan in color.

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. SAMANA: Los Haitises, Boca del Infierno, Ekman
H-15392 (s); Samana, slopes of Pan de Aztcar, Ekman H-15175 (a, B-type,
Ss, US), H-15095 (s); Samana, Laguna, Los Bafiaderos Prietos, Ekman H-
5125.18),

This species is poorly known and is represented in large part by sterile
material. The collection Ekman H-15392 has a very few fruits in a packet
and two other specimens possess a few flowers. In general appearance this
species is similar to Coccoloba costata, although in the details of smaller
fruits, the coronate perianth lobes and the sessile flowers, it is distinct from
C. costata, as well as from other species of Hispaniola. This is one of the
few species of Coccoloba with the leaves shiny on the upper surfaces
when dry. The coloration of the vein pattern is conspicuous when dry,
giving the impression of a minute network or reticulum.

Coccoloba subcordata (DC.) Lindau, Engl. Bot. Jahrb. 13: 131. 1890;
Symb. Antill. 7: 209. 1912.

Erythroxylon subcordatum DC. Prodr. 1: 575. 1824.

Low shrub to 6 feet tall with numerous arching branches, these branch-
ing in one plane, the short-shoots conspicuously developed, the branch-
sate terete, ferruginous pubescent, the nodes not enlarged; ocreae 2—5

long, membranaceous, of uniform texture, obliquely truncate to
slightly bilobed at the apex, appressed, minutely puberulent; leaves of
normal shoots with petioles 1.5-2.5 mm. long, puberulent, arising from the
upper portion of the ocreae, the blades broadly suborbicular to ovate,
2x 2,4 & 4, 11 * 10 mm. long and broad, subcoriaceous, glabrous

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 41

above, glabrate below, rarely puberulent on the veins, commonly shining on
both surfaces and slightly paler in color below, the apex rounded to
emarginate, the base rounded to rounded-cordate, the margin entire,
often slightly undulate, primary veins 3-5 pairs, occasionally clustered
near the base, the veins forking and anastomosing near the margin, reticu-
late, only slightly prominent on both surfaces when dry; inflorescences
terminal on lateral branches or short-shoots, 4-10 mm. long, the rachis
puberulent or glabrate, the bracts short triangular to 0.5 mm. long, the
ocreolae membranaceous, flaring at the apex, to 1 mm. long; - pedicels
solitary or rarely 2 in the axil of each bract, to 2 mm. long at eaauinty
the hypanthium tapering from the apex of the pedicel, about 0.5 mm.
long, the lobes ovate, about 1 mm. long and broad; fertile stamens I—1.5
mm. long, the filaments slightly united at the base, the sterile stamens
aborted or rudimentary, less than 0.5 mm. long, the functional ovary
strongly 3-angled; fruit broadly ovoid, broadest below the middle, 4 mm.
long, 3-3.5 mm. in diameter, the fleshy perianth bright red, the perianth
lobes 14-24 the length of the fruit, strongly imbricated, the achene pale
tan in color.

DISTRIBUTION: Endemic to Hispaniola.

Dominican Republic. Prov. BaRAHONA: Las Salinas, Fuertes 822 (F, GH, US),
Howard 12060 (cu); Beata Island, Howard 12488 (GH). Prov. BENEFACTOR:
San Juan, Loma de Jayaco, Ekman H-13471 (s, us). PRov. INDEPENDENCIA:
Between Lake Enriquillo and Puerto Escondido, Howard 12130 (cG). PRov. NoT
KNowN: Bertero s.n. (B, c-type of Erythroxylon subcordatum, GH-photos, M).

Haiti. Dept. pu Norp Ouest: Jean Rabel, £.C. & G.M. Leonard 12761
(A, GH, NY, ue): Presqu’ile du Nord-Ouest, Baie de Henne, Ekman H-4536 (s,
US). Dept. D L’ARTIBONITE: Gonaives, Leonard 10058 (GH, NY, US). DEPT.
De LOUEST: Cul de Sac east of Gautier, Holdridge 1241 (cH, Us), 1166 (GH,
us); Fond Parisien, Etang Saumatre, Leonard 4056 (c, GH, Ny, US); Montagnes
du Trou d’Eau, Morne a Cabrits, Ekman H-1023 (a, s); Massif des Matheux,
Magasin Carriés, Ekman H-3302 (A, S, US).

The habit of this plant in the field is strikingly different from all other
species of Coccoloba which I have encountered. The plant occurs as a
low shrub in arid regions, especially in thorn-shrub zones. There is no
single trunk to the plant, but numerous branches arise in a cluster and
each branch-system arches. The lateral branches from the shoots are all
arranged on two sides of the stem so that the aspect of the plant is of
flattened leafy shoots.

There were no signs of damage to these plants seen in the field. Although
the habit of the plant suggested that it consisted of adventitious shoots,
no evidence of a central trunk was found. While all growth was slow on the
plant, a few branches showed signs of more vigorously growing twigs.
These possessed slightly longer internodes and leaves around 1 cm. in
diameter. This relatively small-sized leaf blade was in contrast to the
even smaller leaves of the rest of the plant. The secondary branches are

42 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

characteristically short shoots : ey slow growth with compacted
nodes and no internodal region

The sterile herbarium specimens of Coccoloba subcordata are difficult
to distinguish from the adventitious shoots of C. leoganensis. The flattened
branches and the short shoots, as well as the petiole arising from the apex
of the ocrea, enable one to distinguish this species from the adventitious
shoots of C. leoganensis either in the field or in the herbarium, however.

The similarity in appearance of these two species is disturbing. Cocco-
loba subcordata was originally described by De Candolle as a questionable
species of Erythroxylon and was based on a Bertero specimen. Martius
(Abhdl. Bayr. Acad. 3: 303. 1841) suggested the correct affinity for the
plant and Lindau transferred the species, publishing the new combination.
Interestingly enough, several of the specimens cited above had been in-
correctly referred to the genus Erythroxylon.

Coccoloba swartzii Meisner, DC. Prodr. 14: 159. 1856; Lindau. Engl.
Bot. Jahrb. 13: 157. 1890; Howard, Jour. Arnold Arb. 30: 420
1949, 37: 317-339. 1956.

Coccoloba swartzii var. (?) portoricensis Meisner, DC. Prodr. 14: 160. 1856.

Coccoloba barbadensis Lindau, Engl. Bot. Jahrb. 13: 148. 1890, not Jacquin.

Coccoloba diversifolia Lindau, Symb. Antill. 1: 223. 1899, and most recent
authors, not Jacquin

Trees 24 to 60 feet tall, branches terete, the youngest puberulent, be-
coming glabrate, the nodes slightly tumid; ocreae 10-12 mm. long, the
basal portion 3-5 mm. long, coriaceous, persistent, the upper portion
membranaceous and deciduous, puberulent to glabrate; leaves of normal
shoots with petioles 10-18 mm. long, puberulent or glabrate, the blades
ovate to elliptic, 2.2 & 1.3,7 X 5,11 & 9,15 & 7.5 cm. long and broad,
coriaceous, usually turning black on drying, glabrous, having pit-like de-
pressions on the upper surface and small glands on the lower surface, the
apex acute, often rounded, the base narrowed, rounded or slightly cordate
and usually oblique, the margin entire; midrib and veins inconspicuous or
flat above, prominent below, the primary veins 6 or 7 pairs, arcuate, anasto-
mosing, the secondary venation conspicuous, reticulate: leaves of ad-
ventitious shoots with petioles 1.5-2.5 cm. long, the blades generally
ovate to lanceolate 23 & 8.5, 45 & 18.5, to 70 & 25 cm. long and broad,
the apex acute to acuminate, the base rounded; inflorescences terminal,
10-15 cm. long, the rachis glabrous or with glandular exudate, rarely
papillose; staminate flowers in clusters of 3—5 flowers at each node, tightly
surrounded by membranaceous ocreolae which form a truncate cylinder
after the flowers have fallen, the pistillate flowers borne singly at each node,
the bracts ovate, 0.5 mm. long, the ocreolae 1—-1.5 mm. long, membra-
naceous, the pedicels shorter than the ocreolae; hypanthium 0.5 mm. long,
the perianth lobes 1—1.5 mm. long, the fertile stamens 1—1.5 mm. long, the
sterile stamens rudimentary, 0.5 mm. long; fruit ovoid, 8-10 mm. long, 6

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 43

mm. diameter, the perianth lobes 1-1.5 mm. long and coronate in fruit; the
achene dark brown.

DISTRIBUTION: Jamaica, Bahamas, Dominican Republic, Puerto Rico,
St. Croix, St. Jan, Virgin Gorda, St. Thomas, Saba, St. Kitts, Montserrat,
Antigua, Guadeloupe, Dom nied Martinique, St. Tcl and Barbados.

Dominican Republic. Districtro pE SANTO DomriNnco: San Isidro, Ekman
H-11014 (A, s, us). Prov. Puerto Piatra: Puerto Plata, Wright, Parry and
Brummel 472 (GH, us), 473 (GH, US). Prov. Truyjitto: Villa Altagracia,
Taylor 414 (Ny), 431 (Ny, US), 433 (B, NY, US).

A full discussion of the variation in form of this species and the cor-
rect application of the names Coccoloba barbadensis, C. diversifolia and
C. swartzii was published as the second paper of this series (Jour. Arnold
Arb. 37: 317-339. 1956). If the var. portoricensis were recognized, the
specimens seen from the Dominican Republic would be referred there.
However, gradation from Puerto Rico to Jamaica, including this outlying
population in the Dominican Republic, does not warrant the recognition
of Meisner’s variety.

Coccoloba uvifera L. Syst. Nat. ed. 10. 1007. 1759.

tol

Polygonum uvifera L. Sp. Pl. 365. 1753.
Guaiabara uvifera House, Amer. eat Nat. 8: 64. 1922.

Tree of strand areas, 6—50 feet tall, the branches terete, stout, papillose
to pilose, the nodes not tumid; ocreae rigid, coriaceous at the base, membra-
naceous at the apex, 3-8 mm. long, papillose to pilose; leaves of normal
shoots with petioles stout, 7-10 mm. long, papillose to pilose, the blades
orbicular to reniform, 6 & 8, 11 *& 13, 13 & 18 cm. long and broad,
thick and fleshy when fresh, coriaceous when dry, glabrous and minutely
punctate on both surfaces, the apex rounded, truncate or emarginate, the
base rounded to broadly cordate, one lobe often extending around the
petiole, the midrib and primary veins prominent on both surfaces, fre-
quently brightly colored when fresh, the primary veins 3-5 pairs, usually
straight, bifurcate and weakly anastomosing near the margin, commonly
barbate in the axils of the basal veins, the secondary venation minutely
reticulate or obscure; leaves of adventitious shoots usually variable in
size and shape, commonly obovate; inflorescences stout, 15—30 cm. long, the
rachis puberulent; staminate flowers in clusters of 1-7, the pistillate
flowers solitary at each locus, the bracts ovate, 1-1.5 mm. long, 2 mm.
broad, puberulent, the ocreolae membranaceous, 1 mm. long, puberulent,
the flowering pedicels 1-2 mm. long, the perianth yellow-white or greenish,
the hypanthium 2-3 mm. long, the perianth lobes 4 mm. long, 3-4 mm.
wide, the fertile stamens to 4 mm. long; fruiting pedicels 3-4 mm. long;
fruit obpyriform, 1.2-2 cm. long, 8-10 mm. in diameter, narrowed at the
base, rounded-truncate at the apex, the perianth lobes appressed against the
apex of the achene, the perianth rose-purple when mature, the achene black.

44 JOURNAL OF THE ARNOLD ARBORETUM _[vot. xxx1x

Loca NAMES: Raisin la mer (H), Uva caleta, Uva de mar, Uvero de
playa, Uva de playa (DR).

DIsTRIBUTION: Along the shores of Florida, Bermuda and through
the Caribbean Islands to Mexico, Central and South America. Spontaneous
in Africa and the Pacific Islands.

Dominican Republic. Prov. LA ALTAGRACIA: Punta Macao, Howard 9767
(A); east of Jovero, Abbott 2878 (us); Llano Costero, Jaina, Ekman 19475
(us). Prov. BARAHONA: Barahona, Fuertes 247 (us), 1143 (F, GH, US): Beata
Island, Howard 12489 (cH); Alta Vela Island, Howard 12453 (cu); El Caiman
near Enriquillo, Howard 12188 (GH), 12191 (GH). Prov. Monre Cristt:
Los Siete Hermanos, Monte Grande, Howard 12523 (cu). Prov. SAMANA:
Samana Abbott 1187 (us).

Haiti. Dept. pu Norp Ouest: Port de Paix, E.C. & G.M. Leonard 11169
(A, GH, US); Ile de la Tortue, La Vallée, Leonard 11701 (us). Derr. pu Norp:
Bayeux near Port Margot, Nash 907 (ny). Dept. bE L’OuEsT: Petit Gonave
Island, Leonard 5239 (cH, us); Morne a Bateau, Port-au-Prince. Ekman 8162
(s, US

This common strand plant is certainly more abundant in Hispaniola than
the collections cited above would indicate. It is also commonly cultivated
as an ornamental shrub or tree.

Coccoloba uvifera L. & Coccoloba pubescens L.; Howard, Jour.
Arnold Arb. 36: 225. 1955.

Shrub or tree to 18 feet tall with habit of Coccoloba uvifera, i.e., some
branches erect, some branches semi-prostrate; branchlets terete, sulcate
when dry, pubescent or puberulent, the nodes slightly enlarged; ocreae
to 1 cm. long, deeply cleft, the basal portion persistent; leaves of normal
shoots with petioles to 1 cm. long, puberulent or pubescent, attached at
the bases of the ocreae; blades usually orbicular, 8.5 « 9.5, 11 & 14,
12.5 & 13,15 & 19 cm. long and broad, thick or fleshy, plane or slightly
bullate, drying yellow-green in color, short pubescent on the midrib and
primary veins above, the rest glabrous, persistently short pubescent on the
veins below, the lower surface dark-glandular-dotted, the apex rounded,
occasionally broadly and shortly mucronate, the base cordate, the lobes
overlapping, the margin entire, slightly undulate, the primary veins usually
5 pairs, impressed above, arcuate and conspicuously anastomosing near
the margin; adventitious shoots with ocreae 1—2 cm. long, the leaves with
petioles to 1 cm. long, stout, the blades broadly orbicular or slightly
rhombic, 22 24 to 27 x 36 cm. long and broad, otherwise as in ma-
ture shoot leaves; inflorescences 12-25 cm. long, terminal and generally
paired, with one raceme shorter than the other, the rachis puberulent:
staminate flowers not seen; pistillate flowers 3-10 at each locus, these
seemingly scorpioid in development, the bracts triangular, to 1 mm., the
ocreolae membranaceous, | mm. long, the pedicels to 2.5 mm. long, puberu-
lent, the hypanthium short, to 0.5 mm. long, the perianth lobes ovate, 1.5—2
mm. long, 1-1.5 mm. wide in bud, the stamens rudimentary, the anthers

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 45

abortive, the pistil 1.5—-2 mm. long; fruit obovoid, 12-13 mm. long, 6-7 mm.
diameter, narrowed to a short stalk-like base 2-3 mm. long, the apex
rounded, the fleshy perianth lobes imbricated, red; achene obovoid to glo-
bose, light brown, smooth or slightly pitted.

Dominican Republic. PRov. BARAHONA: El Caiman between Enriquillo and
El Can, Howard 12189 (GH), 12191 (GH); Beata Island, Howard 12499 (cH).

Haiti. Dept. pu Sup: Miragoane, Eyerdam 397 (GH, NY, US); Anse a Veau,
Picarda s.n. (GH).

In 1950 Mr. George Hamor of Barahona discovered an unusual stand of
Coccoloba along a coral shelf and rocky beach area south of Barahona. He
later arranged transportation to the spot and showed me this definite hy-
brid. The plant described here occurs between a coastal stand of C.
uvifera (Howard 12188) and an inland stand of C. pubescens (Howard
12187). Fully a dozen mature plants were found in this location. Not
only is the plant intermediate in geographic location, but all of its char-
acteristics show its hybrid origin from C. uvifera and C. pubescens. In
habit the plant resembles C. uvifera, being a plant of low stature with
some branches semi-prostrate and spreading while others are erect, giv-
ing each plant the definite clump-like appearance of C. uvifera. The leaf
shape of the normal leaves on mature branches is that of C. uvifera except in
texture, in which they resemble C. pubescens, having the conspicuous re-
ticulate venation on the lower surface. The pubescence of the leaves com-
bines that of the parent species. The leaves of the adventitious shoots more
closely resemble those of C. pubescens in size, shape and aspect than com-
parable leaves of C. uvifera. In the arrangement of the flowers and the
pubescence, the characters remind one of C. pubescens, although the
presence of a smaller raceme arising laterally from the base is more com-
mon in C. uvifera. Only female flowers are known. The fruits are small-
er than typical Coccoloba uvifera but resemble them in shape and are un-
like those of C. pubescens. Fruits were abundant on the hybrid plants
and appeared in the field to be fully developed. However, none of the one
hundred fruits collected would germinate a month later, while eighty
per cent germination was obtained from a collection of C. oer made at
the same time.

A similar stand of the hybrid was found on Beata Island two weeks
later and again both parents were present.

The collection Eyerdam 397 is referred to this new hybrid. The speci-
men appears to have been taken from adventitious shoots and possesses
larger leaves more closely resembling Coccoloba pubescens. As is generally
true of flowering material collected from adventitious shoots, the in-
florescence of this specimen is larger, approaching 35 cm. in length.

While the suggested hybrid origin of Coccoloba antiguensis Sandwith
from Antigua has been rejected and that species referred to the synonymy
of C. pubescens, there is no doubt in my mind as to the valid nature and
the origin of the present hybrid. The hybrid nature of this collection is
obvious in the field and equally so in the specimens cited.

46 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

The other hybrid plants and populations involving Coccoloba uvifera
considered in this paper are C. costata * C. uvifera, C. hotteana & C.
uvifera and C. krugit * C. uvifera.

Coccoloba venosa L. Syst. Nat. ed. 10. 1007. 1759; Fawcett and Rendle,
Jour; Bot..51% 123. 1919;

Coccoloba punctata L. Sp. Pl. ed. 2. 523. 1762

Uvifera arbor americana, fructu aromatico punctatus, Pluk. Alm. 394, t.
257, - . 1696, as to leaf only.

Coccoloba nivea Jacq. Hist. Stirp. Amer. 115, pl. 78. 1763; Enum. Pl. 19
1762.

Guaiabara venosa House, Amer. Midl. Nat. 8: 64. 1922, as Guazbara.

Trees to 45 feet tall; branches terete, glabrous, the nodes not tumid;
ocreae membranaceous, deeply cleft, acuminate on one side, or truncate,
to 2 cm. long, glabrous or with flattened glands; leaves of normal shoots
with petioles 5-10 mm. long, oe: the blades ae nian to
elliptic. .6.% 410: 45. 165 365, 21 «0.27 « 105 . long and
broad, membranaceous, glabrous except for clusters of hairs in on axils of
the veins, sparsely slandular below, the apex short-acuminate, the base
narrowed and slightly cordate or cuneate or obtuse, the midrib and primary
veins slightly prominent on both surfaces, the primary veins 8-13 pairs,
straight or arcuate, bifurcate and anastomosing at the margins; leaves of
the adventitious shoots about the same size, the internodes much elongate
and the ocreae to 4 cm. long; inflorescences terminal or terminal on short
lateral branches, the rachis puberulent, angular; staminate flowers in
clusters of 2—5, the pistillate flowers solitary, the bracts lanceolate-ovate,
to 1.5 mm. long, black, puberulent to pilose or commonly with a fringe
of hairs at the apex; ocreolae to 2 mm. long, membranaceous, enlarging
with the expanding bud, each flower with an meeola: the flowering pedicels
1-2 mm. long, glabrous; hypanthium less than 0.5 mm, long, the perianth
lobes broadly ovate, 1.5-2 mm. long and broad, slightly unequal, the
fertile stamens to 1 mm, long; fruiting pedicels 1.5—-2.5 mm. long, the
perianth lobes fleshy, white or pink, enclosing the black achene, the hy-
panthium scarcely evident in the fruit, the fruit broadly ovoid, 3-4 mm.
long and broa

DiIsTRIBUTION: Cuba (introduced), Hispaniola, Puerto Rico, Jamaica
(?), Virgin Islands, Lesser Antilles and Trinidad.

Dominican Republic. District pE SANTO Dominco: between Ciudad Trujillo
and La Caleta, Ekman H-14231 (s). Prov. La ALTacRActA: Llano Costero at
La Romana, Ekman H-12089 (s,s). Prov. SAMANA: Cabo Samana near Puerto
Colorado, Ekman H-15333 (s). Prov. Serpo: Monte Redondo, east - Jovero,
Abbott 2792 (8, US). PROV. UNKNOWN: Cupey, Eggers 2682 (NY,

Haiti. Dept. pu Norp Ouest: Ile de la Tortue, La Vallée, Ekman H-9758
(s, us). Dept. pu Norp: Massif du Nord, Port Margot, Bayeux, Ekman
H-2699 (s, us). Without at location: Sessé & Mocino 952 (F), 5437 (®).

1958] HOWARD, STUDIES IN THE GENUS COCCOLOBA, V 47

Coccoloba wrightii Lindau, Engl. Bot. Jahrb. 13: 151. 1890; Howard,
Jour. Arnold Arb, 30: 418. 1949.
Coccoloba scrobiculata Lindau, Engl. Bot. Jahrb. 13: 140. 1890.
Coccoloba subtruncata Urban, Symb. Antill. 7: 211. 1912.
Coccoloba saxicola Britton, Bull. Torrey Bot. Club 50: 37. 1923.

Shrub or small tree to 21 feet tall; branches terete, the nodes not
swollen, glabrate; ocreae membranaceous, 4-6 mm. long, puberulent to
tomentose or glabrate; leaves of normal shoots with petioles 4-7 mm. long
arising from the bases of the ocreae, the blades ovate, elliptic, obovate
or rarely ovate-lanceolate, 5 « 2.5,8 & 4,10 X 7 to 11 X 10 cm. long
and wide on mature shoots, coriaceous, umbonate between the veins,
shining above when young but dull on both surfaces when mature, the
apex acute to abruptly short acuminate or truncate, the base narrowed
to obtuse, usually slightly oblique, the primary veins 4—6 pairs, arcuate,
impressed above, conspicuous below, the lower surface more or less dotted
with stomatal excretions; adventitious shoots with ocreae to 2 cm. long;
leaves with petioles 2.5 cm. long, the blades broadly ovate to elliptic,
15 x 14 to 20 & 17 cm. long and broad, the apex of these leaves rounded
to obtusely short mucronate, the bases rounded to subcordate; inflo-
rescences terminal, 3-10 cm. long, the rachis pubescent or with resinous
excretions, the bracts ovate, to 0.5 mm. long, the ocreolae membranaceous,
1 mm. long, the flowering pedicels 1 mm. long, increasing in length either
in fruit or after staminate flowers have fallen to 3 mm.; staminate flowers
2-3 per locus, the pistillate flowers borne singly at each locus; hypan-
thium to 1 mm. long, the perianth lobes 1-1.5 mm. long and broad, the
fertile stamens united at the base for 1 mm., the free portions 0.5—1
mm. long, the sterile stamens less than 1 mm. long, the functional pistil
to 2 mm. long; fruit ovoid, slightly contracted at the base, rounded but
only slightly coronate at the apex, 7-9 mm. long, 4-5 mm. in diameter.

DIsTRIBUTION: Cuba and Hispaniola.

Dominican Republic. Prov. Banoruco: between El Aguacate and Pedernales,
Howard 12585 (GH). Prov. La VecA: Constanza, Turckherm 3304 aoe of C.
subtruncata, F, GH, M, MO, NY, S$, US); Arroyo Pantuflo near Constanza, Ekman

89 (a. s, US); Bonao, Ekman 16450 (s, us). Without ree location:
Schomburgk 123 (8), Preneloup 492 (B, US), Bertero s.n. (B).

Coccoloba scrobiculata Lindau was described, collections by Schomburgk
and Preneloup being cited, in the same publication as C. wrighti Lindau.
Lindau attempted to distinguish between them in a key by indicating that
the lesser venation was flat and inconspicuous above in C. scrobiculata
while it was more prominent in C. wrightti. This is scarcely a reliable
characteristic in the genus and I have no doubt that only one species is
involved. No recent collections have been referred to C. scrobiculata and I
have chosen to accept the better known and documented C. wrightii as the
species. Although C. scrobiculata was described a few pages earlier, but
at the same time as C. wrightii, I am considering it a new synonym. The

48 JOURNAL OF THE ARNOLD ARBORETUM _ |[voL. xxxiIx

venation pattern of material called C. scrobiculata by Lindau is easily
included in the range of variation of C. wrightii and in all characteristics
visible in the scanty flowering material, the two are identical.

Coccoloba subtruncata, described some years later by Urban, was based
on a collection made near Constanza by H. von Tiirckheim. The species
has been recollected in the same area by Ekman and additional collections
are available from other areas. Urban’s original diagnosis was presumably
based on the one sheet of the Tiirckheim collection in the Berlin herbarium.
I have on loan nine sheets of this number which are obviously the same
but which would necessitate a new description to be accurate. Recent
material (e.g., Ekman H-16450 and H-14089) in fruit allows a complete
diagnosis of this species which obviously is the same as Coccoloba wrightti
of Cuba and must be referred to synonymy there. In general, the His-
paniolan specimens have less pubescence when mature than do the Cuban
plants. However, the type collection of C. subtruncata, in spite of Ur-
ban’s description, exhibits the same pubescence as C. wrightii, at least on
the young shoots and the tips of the ocreae. Coccoloba wrightii has been
considered to be endemic to Cuba, but its range is now extended to the
Dominican Republic and specimens should be found in Haiti.

A twisted tree represented by Howard 12585 was alongside a new road
from El Aguacate to Pedernales in the Dominican Republic. In the
course of road-building this tree had been pushed over at an angle and
from the lower portion erect adventitious shoots had developed with large
and extremely thick-coriaceous leaves. These adventitious stems were
8-10 mm. thick near the apex in contrast with the much smaller diameter
of the normal growth. The largest leaves on the shoot had stout petioles
2—2.5 cm. long and broadly ovate to elliptic leaves to 19 15 cm. long
and broad. The apex of the blade was rounded to short and obtusely
mucronate and the bases were rounded to subcordate. The terminal por-
tions of this plant produced shoots which were identical with those in
the type collection of Tiirckheim. Similar-sized leaves of adventitious
shoots of Coccoloba wrightii have already been reported and the previous
description is amended only to include leaves which are rounded to sub-
truncate at the base.

The collection Ekman H-15135 from Los Bafiaderos Prietos near Laguna
on the Samana Peninsula was named by Schmidt as “C. subtruncata
forma.’ Ekman’s field notes state, a ‘‘small tree, alas, sterile.” This ma-
terial seems more appropriately referred to Coccoloba pauciflora Urban. It
is obviously from adventitious shoots and the normal foliage is not repre-
sented. Neither C. pauciflora nor C. wrightii has been reported from the
Samana Peninsula.

”)

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 49

A MONOGRAPHIC STUDY OF THE WEST INDIAN
SPECIES OF PHYLLANTHUS *

GRADY L. WEBSTER
With four plates

Subgenus VII. Botryanthus Webster, Jour. Arnold Arb. 37: 345. 1956.

Trees or shrubs with non-phyllanthoid branching, the leaves on the
main axes not reduced to scales, ultimate axes not regularly deciduous.
Monoecious; flowers borne in thyrses or axillary clusters. Male flower:
calyx-lobes 5 or 6; disk-segments 5 or 6, often massive; stamens 3, filaments
connate, anthers dehiscing horizontally or obliquely; pollen grains globose,
areolate. Female flower: calyx-lobes 5 or 6; disk cupuliform; ovary of 3
carpels, smooth; styles erect or spreading, usually connate (at least
basally) into a column, the tips entire to bifid. Fruit capsular, sometimes
very large; seeds trigonous, essentially smooth.

Subgenus Botryanthus is an entirely Neotropical group which includes
possibly 35 species, most of these occurring in Central and South America.
In addition to the dubious sect. Diplocicca from Brazil (based on P.
octomerus Muell. Arg.), other South American sections may eventually
be defined; but the majority of the species in the subgenus (including all
the West Indian ones) clearly are referable to sect. Elutanthos.

From the phylogenetic point of view subg. Botryanthus is of especial
interest because it is closely related to subg. XylopAvyila. Its areolate
pollen grains are essentially identical to those of the latter subgenus and,
since its branching is non-phyllanthoid (and hence presumably unspecial-
ized), subg. Botryanthus might logically be considered the ancestral
group. The reduction series in inflorescence within sect. Elutanthos (which
can be traced from the elaborate “‘panicle” of P. grandifolius to the axillary
flowers of P. nutans ssp. grisebachianus) could furthermore be cited
in support of such a supposition. On the other hand, it must be noted
that in many ways P. grandifolius is a very specialized plant, for it has
capsules and seeds extraordinarily large in the genus, and its male flowers
with three connate stamens give no indication of primitiveness. A rather
convincing argument could be put forth to read the phylogeny in precisely
the opposite direction, and to assume that subg. Botryanthus has evolved
from some group within subg. Xylophylla (such as sect. Asterandra)
by a regression from phyllanthoid to non-phyllanthoid branching. Such
a hypothesis would involve the assumption that a great increase in leaf
and fruit size (such as occurs in P. grandifolius) would place a selective
advantage on any reversion from deciduous to permanent reproductive
axes. At present a decision for either of the alternative hypotheses can-

* Continued from volume XX XVIII, p. 373.

50 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

not be made, but with the accumulation of additional evidence (particular-
ly from cytology) it should be possible to do so

Sect. 15. Elutanthos Croiz. Jour. Wash. Acad. Sci. 33: 12. 1943.

Shrubs or trees with unspecialized ramification, leaves distichous on
persistent axes. Monoecious; flowers in cymules in the axils of foliage
leaves or of bracts (the inflorescence then thyrsoid), or sometimes solitary.
Male flower: calyx-lobes 6; disk-segments 6, often massive; stamens us-
ually 3, filaments completely united into a column; anthers dehiscing
more or less horizontally (or deflexed); pollen grains globose, areolate,
the areoles usually polybrochate. Female flower: calyx-lobes 6; disk patel-
liform; styles erect, connate at least below, entire to bifid. Capsule ob-
scurely rugulose; seeds smooth.

Type species: Phyllanthus glaucescens H.B.K. | = P. grandifolius L.].

As established by Croizat, sect. Elutanthos included seven species of
Central and South America; but there are undoubtedly a number of
others which are to be referred here, including the following six species
from the West Indies. The West Indian plants differ from the mainland
P. grandifolius and its allies in their smaller capsules and less elaborate
inflorescence, but the similarities are so striking that there is obviously a
close affinity.

The combination of an indefinite branching pattern and usually raceme-
like inflorescence distinguishes the species of sect. Elutanthos from all other
woody West Indian species. However, the expression of inflorescence is
quite variable and the typical thyrses are not developed in P. nutans ssp.
grisebachianus and some forms of P. nutans ssp. nutans. Where the raceme-
like thyrses occur, they often appear to be terminal, but close inspection
will show that each thyrse does not represent a continuation of the branch
axis but is rather inserted just below the abortive tip of the axis; often
an additional thyrse is produced at one or more additional nodes below.
In addition, cymules may also occur in the axils of foliage leaves; in
such cases, these axillary cymules tend to be male and the ones in the

“racemes” female. The homology between the more or less naked thyrses
and the leafy branch-ends is as ill-defined as that between branch-orders,
for the degree of distinctiveness of thyrse development is at least partially
dependent on the vigor of the branches. Sometimes there may be long
‘“Jeader” shoots which bear short thyrsiferous branchlets in distichous or-
der, but in other instances the thyrsiferous branchlets may themselves be of
the penultimate order and produce additional leafy axes as well as
“racemes” of flowers.

Ecologically the West Indian representatives of sect. Elutanthos are
characterized by a xerophytic tendency, most of the species appearing
to grow in open scrub formations at low altitudes. The widespread P.
nutans, however, occurs in a considerable variety of habitats, including
some of the wettest rain forest in Jamaica.

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS St

KEY TO THE SPECIES

1. Calyx-lobes of male flowers less than 1.5 mm. long; capsules trigonous;
plants of the southern Caribbean.

2. Styles subentire (truncate-emarginate), completely erect and connivent
(with no reflexed portions); disk of female flower extremely massive;
plants completely glabrous; Curagao. .............. 43. P. bo tryanthus

2. Styles with bifid sharply reflexed tips; disk of female flower tenuous;
branchlets and inflorescence axes hirsutulous; Barbados. ..... ......
Pee er ctahe ee ait cae Seca eye ee 44, P. anderssonii

1. Calyx-lobes of male flowers over 1.5 mm. long; capsules rounded; plants

of the northern Caribbean.

2. Staminal column with 3 apiculae alternating with the anthers; leaf-blades
densely hirsutulous beneath, sparsely hirsutulous above; style-tips entire
or merely emarginate; Elaitin, A ere ae oe 47, P. urbanianus.

2. Staminal column without apiculae; leaf-blades glabrous or sparsely
hirsutulous beneath, glabrous above; styles definitely bifid.

3. Leaf-blades rigidly coriaceous, the main veins distinctly ae above;
stipules indurate, dark and atin, colored as the branch; Cuba. ......
ns area ens eg het se Us (co oe en Re Pee: 46. P. Bee is

3. Leaf-blades chartaceous, the veins not sunken above; stipules not
InGucates jamaica anc: ubae set ee eee ee Ac, 45. P. nutans

43. Phyllanthus botryanthus Muell. Arg. in DC. Prodr. 15(2): 323
1866. (PLATE XXIV, figs. A-B).
Diasperus botryanthus (Muell. Arg.) O. Ktze. Rev. Gen. 2: 598. 1891.
Phyllanthus euwensi Bold. Fl. Dutch W. Ind. 2: 50-51, pl. 4. 1914.
Glochidion a ie (Muell. Arg.) Pax & Hoffm. Naturl. Pflanzenfam.
ed. 2, 19c: . 1931.

Glabrous shrub or small tree c. 1-4 m. high; branches slender, terete,
furrowed, reddish-brown, c. 1.5—-4 mm. thick; leader shoots up to 25 cm.
long, lateral branchlets (of current year) c. 3-20 cm. long, with 5-15
leaves. Leaves: stipules triangular, mostly 0.8-1.1 mm. long and 0.5-—0.6
mm. broad, acute, scarious, reddish- or blackish-brown, persistent (at
least the basal portion), more or less entire. Petioles dark, grooved above,

ovate, c. (3—) 4.5—7.5 cm. long, (2—) 2.5—5 cm. broad, acute or obtuse
(rarely emarginate) at the tip, obtuse or rounded at the base; above dark-
ened in drying, the midrib and laterals very slightly raised; beneath
more or less pruinose, the midrib salient, the laterals (c. 6-8 on a side)
arching, slightly raised, branching to form a reticulum; margins unthick-
ened, plane.

Monoecious; flowers mostly in bisexual cymules on leafless thyrses

per cymule, or the sexes more or less segregated onto different thyrses.
Male flower: pedicel capillary, c. 1 cm. long. Calyx-lobes 6, subequal,
elliptic-oblong, c. 1-1.2 mm. long and 0.5—0.8 mm. broad, obtuse, purple-

Ay JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

stained except for the narrow scarious entire margins, midrib unbranched
or nearly so. Disk-segments 6, massive, subcubical, foveolate, c. 0.3-0.4
mm. broad. Stamens 3; column c. 0.35-0.4 mm, high; anthers sessile,
discrete, triangular, acute, c. 0.25-0.35 mm. long, 0.35-0.4 mm. broad;
anther-sacs divergent, the slits not confluent, dehiscing horizontally; pol-
len grains mostly 22—26 ,» in diameter, areoles oligobrochate, c. 5 p across.

Female flower: pedicel capillary, (10--) 13-16 (-19) mm. long. Calyx-
lobes 6, triangular-oblong, c. 1.2-1.3 mm. long and 0.7—-0.8 mm. broad,
acute or subacute, purple-stained as the male, the narrow scarious margin
more or less entire, the midrib sparingly branched. Disk convex, extremely
massive, circular or angled, nearly 2 mm. across, foveolate. Styles erect,
connate or coherent into a column 0.3-0.45 mm. high, slightly dilated
and merely truncate-emarginate at the apex.

Capsule oblate, trigonous, c. 3 mm. high and up to 6 mm. broad, some-
what rugulose, the veins obscure or conspicuous. Columella 1.7—2 mm.
high. Seeds plano-convex, plump, umbonate (heliciform), 3.3-3.6 mm.
long, 2.7—-2.9 mm. broad, light brown, smooth (finely striolate).

Collected in flower and fruit May to January.

Type: Colombia, Carthagena, Triana 3664 (P, LecToTYPE; K, W,
ISOTYPES).

DIsTRIBUTION: coastal plain, northern South America (Map XVIII).

ARUBA: Kristalberg, 1885, Suringar (L). CURACAO: road near Antony-
berg, Boldingh 5141 (C); road near Wacao, Boldingh 5252 (L): calcareous soil
near San Pedro, Boldingh 5279 (NY, isotype of P. euwensii); Tafelberg,
Curran & Hamman 164 (A); Hofje Abau, Curran & Hamman 180 (A); rif-
hospitaal, Hato, Savonet, Suringar (L). BONAIRE: klip Slachtbaai, Suringar
(L). (Additional localities cited by Boldingh, loc. cit.)

The Curassavican plants represent merely a small outlying population of
this species which is widespread and common in lowland areas along the
northern coasts of Colombia and Venezuela. The specimens from the
Dutch islands show no evident differences from the mainland plants, so that
Boldingh’s proposed P, euwensii cannot be maintained even at subspecific
rank. Boldingh did not give any distinguishing characters for his intended
new species, nor did he discuss its relationships; apparently he was un-
aware that the species also occurred on the mainland or that it had already
received a name.

Because of its essentially entire styles associated in a column, P.
botryanthus was placed by Mueller in his artificial sect. Hemiphyllanthus:
it was therefore one of the species transferred to Glochidion by Pax and
Hoffmann (loc. cit.), who arbitrarily assigned all species of Phyllanthus
with entire styles to Glochidion. This betrayed mere bibliographic ac-
quaintance with the plants, for P. botryanthus and P. ovatus (the other
West Indian representative of Mueller’s sect. Hemiphyllanthus) are neither
closely related to one another nor to species of Glochidion.

The habit, floral structure, and pollen grains of P. botryanthus together

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 53

indicate an unquestionable relationship with the other species of sect.
Elutanthos. Within the section, the closest West Indian relation appears
to be P. anderssonii, which has very similar vegetative parts and male
flowers but utterly different female flowers with bifid reflexed styles and
a tenuous disk. In contrast, the female flowers of P. botryanthus are dis-
tinguished by the subentire erect styles and extraordinarily massive disk,
which even in the bud stage is massive and swollen and which expands pre-
cociously to cause the calyx-lobes to become reflexed. However, the
Mexican P. oaxacanus has a similar massive disk and, despite its very
different styles, is probably the most closely related species.

Ap ANDERSSONII
© P. BOTRYANTHUS

@ PN. GRISEBACHIANUS oy reali aaa a

» PN. NUTANS SNe ee (eee
——_* P OAXACANUS nee tet |

« P. PACHYSTYLUS | ti ie

+ P. URBANIANUS | : (EN |

Map XVIII. Distribution of some Caribbean species of sect. Elutanthos.

44. Phyllanthus anderssonii Muell. Arg. in DC. Prodr. 15(2): 395.
1866 (as P. anderssoni). (PLATE XXIV, figs. C-D).
Phyllanthus barbadensis Urb. Symb. Ant. 3: 287. 1902.

A shrub c. 1 m. high [ex Eggers], with slender leader shoots up to c.
30 cm. long bearing distichous branchlets which terminate in naked
thyrses (or leaders serving directly as thyrse-bearing axes); ultimate axes
(branchlets) reddish brown, terete, sparsely to rather densely hirsutulous,
mostly 4-10 cm. long, 0.6-1 mm. thick, with c. 6-15 nodes. Leaves:
stipules triangular-lanceolate, 0.75—1 mm. long, 0.3-0.5 mm. broad, acute
or acuminate, more or less hirsutulous, scarious, becoming reddish brown
and at least the base persistent, entire or denticulate. Petioles dark and
slender, flattened and hirsutulous adaxially, convex and glabrous abaxial-
ly, 1-1.8 mm. long. Leaf-blades membranous or chartaceous, elliptic to
mostly ovate, c. 2.5—4.5 cm. long, 1.2—-2.5 cm. broad, obtuse or subacute
at the tip, cuneate at the base; above olivaceous or drying blackish,

54 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. Xxx1x

hirsutulous proximally on the midrib but otherwise smooth and glabrous,
nerves slightly raised; beneath glabrous, pale or pruinose, the midrib dis-
tinctly raised, the laterals (c. 7-10 on a side) ascending, slightly raised,
connecting intramarginally, the tertiaries forming a rather inconspicuous
reticulum; margins scarcely thickened, plane.

Monoecious; cymules mostly unisexual; male cymules several-flowered,
borne at proximal axils of branchlet usually axillary to foliage leaves;
female cymules with 1-4 flowers, borne on the thyrsiform terminal por-
tions of branchlets (the leaves becoming more or less abruptly reduced
to scales); one or two cymules at the transitional region often bisexual.

Male flower: pedicel capillary, up to c. 8 mm. long. Calyx-lobes 6,
chartaceous, subequal, elliptic to oblong or obovate, c. 11.3 (—-1.5) mm.
long, 0.75—1 mm. broad, obtuse or subacute, entire or obscurely crenulate,
the midrib unbranched. Disk-segments 6, very massive (about as large as
the anthers), deeply furrowed and pitted, c. 0.3-0.4 mm. broad. Stamens
3: column c. 0.5—0.7 mm. high, c. 0.25 mm. thick; anthers sessile atop the
column, discrete, the common connective convex, c. 0.2—0.3 mm. long and
0.35—-0.4 mm. broad; anther-sacs divergent, dehiscing horizontally, the
slits not confluent; pollen grains c. 21-25 » in diameter, areoles poly-
brochate, c. 5—7 p across.

Female flower: pedicel slender, 7-13 mm. long. Calyx-lobes 6, sub-
equal, ovate-oblong or obovate, 1—-1.4 mm. long, 0.7-0.9 mm. broad,
rounded or subtruncate at the tip, entire or obscurely crenulate, the midrib
sparsely branching. Disk cupuliform, enclosing up to 1% of the ovary,
the tenuous rim crenulate, pitted. Styles connate or loosely coherent into
a column c. 0.4—-0.7 mm. high, the free ends sharply reflexed, bifid up to
2% their length, the tips subulate, obtuse and entire or again emarginate or

ifid.

Capsule oblate, trigonous, 6-sulcate, dark purplish brown, obscurely
venose, c. 5.5-6 mm. broad, the valves c. 4.5 mm. long. Columella 1.8—2
mm. high. Seeds plano-convex, plump, umbonate, 2.7—3.3 mm. long, 1.7—
2.3 mm, radially, 1.8-2.4 mm. tangentially, pale brown with very irregular
longitudinal bands of slightly raised transversely elongated reddish-brown
cells; hilum submedian.

Collected in flower Jan., Feb., June, Nov.; in fruit Feb., Nov.

Typr: “Caracas,” collector unspecified (Prodromus Herbarium, G;
HOLOTYPE). The typification of this species is unfortunately fraught with
difficulties which at this time cannot be wholly resolved. In Mueller’s
original description the origin of the type collection was indicated as fol-
lows: “Prope Caracas (hb. holm. sub n. 288! a cl. Dr. Andersson miss.) .””
Presumably the specimen was not collected by Andersson, since during his
trip on the “Eugenie” he did not visit anywhere within the Caribbean
area, his closest approach being the Pacific coast of Panama (cf. end-map
in Skogman, Fregatten Eugenies Resa Omkring Jorden. 1854-55). If the
plant was really collected near Caracas it seems unlikely that it could have
escaped notice subsequently, although this cannot be entirely ruled out.
It appears more likely that the specimen was actually obtained on Barbados

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS os)

but was subsequently mislabelled. A search through the literature indicates
that there is at least one possible source of the collection, for Robert
Schomburgk spent several months on Barbados in 1846-47 and his citation
(Hist. Barbados 593. 1848) of P. nutans can hardly refer to any species
other than the present one. The figure “222” for the citation in Schom-
burgk’s list does not agree with the “288” cited by Mueller, but the latter
might be a field number. Of course, none of these circumstances constitutes
proof that Andersson’s specimen is a duplicate of a Schomburgk collec-
tion, so that it is impossible to decide conclusively whether the type
collection of P. anderssonii is from Venezuela or from Barbados. Although
it differs in a few minor particulars, the type specimen of P. anderssonii
corresponds so well (e.g., in its hirtellous axes, massive male disk-segments,
and sharply reflexed styles) with the specimens of P. barbadensis that it
must be conspecific; and Mueller’s name must thus take precedence over
the much later one of Urban.

DISTRIBUTION: endemic to Barbados (Map XVIII).

BARBADOS: Lion Hill Gully, St. James, Dash 334 (NY); Forster Hall
Wood, Eggers 7130 (A, GOET, US; LecToTYPE COLLECTION of P. barbadensis) :
Cole’s Cave, St. Thomas, Freeman & Bovell (NY); Highland Gully, St. Thomas,
McIntosh (K); Jack-in-the-box Gully and under Hackleton’s Cliff, Waby 15
(K); Bathsheba, Warming 27 (C).

Beard (Nat. Veg. Leeward & Windward Isl. 166. 1949) has pointed
out that practically all of the natural vegetation on Barbados has dis-
appeared and that many plants endemic to Barbados have become ex-
tinct. However, since P. anderssonii has been collected at a number of
different localities up to 1935, it seems possible that it may still survive,
even though Beard failed to encounter it in his own reconnaissance of the
island.

Urban compared P. anderssoniu (as P. barbadensis) with P. nutans, to
which indeed it shows some similarities; but its closest relationship is un-
doubtedly with P. botryanthus. In the face of their present distributions,
the affinity between P. anderssonii and P. botryanthus has some anomalous
aspects. It seems curious, for instance, that the female flowers (which
are the best distinguishing character) of P. botryanthus are much more
highly modified than those of P. anderssonii, for one might expect that
it would be the latter, which presumably has been derived by coloniza-
tion from South America, that would show the greater specialization. It
also seems surprising that no related forms occur in a gap of about 500
miles between the state of Miranda, Venezuela (the easternmost known
station for P. botryanthus) and Barbados. One possible explanation is
that P. anderssonii may represent a relict of an extensive population which
occupied the ancient land-mass of “Paria”; according to Schuchert (Hist.
Geol. Ant.-Car. Reg. 19. 1935) this region extended from Maracaibo to
Barbados at some time during the Cenozoic, but later foundered and is
now represented only by some of the off-shore islands of northern South
America.

56 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

45. Phyllanthus nutans Sw. Prodr. 27. 1788; FI. Ind, O¢e. 1103, 1800:
Muell. Arg. in DC. Prodr. 15(2): 375. 1866; Fawc. & Rend. Fl. Jam.
4: 253-254. 1920.
Diasperus nutans (Sw.) O. Ktze. Rev. Gen. 2: 600. 1891.

A diffuse extremely variable shrub or slender tree 1—7 m. high, irregu-
larly branching, the new axes borne distichously on older branches, often
ending in nodding racemiform inflorescences, reddish brown, terete or
angled, smooth or scabridulous or hirsutulous, 2-25 cm. long, 1-2.5 mm.
broad, internodes 4-40 mm. long. Leaves: stipules scarious-chartaceous,
deciduous or persistent, ovate to narrowly lanceolate, (2—) 3-5 (-9) mm.
long, (1-) 1.5-4 (—6) mm. broad, obtuse to acute at the tip, truncate
to cordate at the base, entire, yellowish, stramineous, or brownish. Petioles
smooth or scabridulous to copiously hirsutulous, usually somewhat angled
or margined, 1.5—4.5 mm. long. Leaf-blades chartaceous, quite variable
in size and shape but most often ovate or elliptic, c. 3.5-8 (—11) cm. long,
2-5 (—8) cm. broad, obtuse or more rarely acute at the tip, cuneate to
rounded at the base; above olivaceous, dull, the veins scarcely raised;
beneath paler, sometimes glaucous or purplish-tinged, the midrib and
lateral veins (5—8 on a side) raised, brownish or stramineous. the reticu-
lum of veinlets usually conspicuous; margins unthickened, plane or revo-
lute.

Monoecious; inflorescence variable; male cymules_ several-flowered,
borne axillary or on pseudoterminal thyrses; female cymules 1—3-flowered,
usually confined to the pseudoterminal thyrses (at least in ssp. nutans).

Male flower: pedicel capillary; 8-15 (-30) mm. long. Calyx-lobes 6,
chartaceous, or somewhat fleshy, subequal, oblong to obovate, 2.3-3.2 mm.
long, 1.3-2.2 mm. broad, rounded and obscurely crenulate or denticulate
at the tip, often reddish at least below with thin creamy-yellow scarious
margins but sometimes greenish throughout, the midrib with a few often
conspicuous lateral branches. Disk-segments 6, thickened and fleshy, or-
bicular or reniform, obscurely to conspicuously foveolate-pitted, c. 0.35—
0.7 mm. broad. Stamens 3; column c. 0.71.1 mm. high, tapering slightly
to the apex; anthers sessile atop the column, discrete, usually deflexed
but sometimes horizontal, broadly triangular to elliptic in outline, c. 0.4—
0.6 mm. long and broad; anther-sacs slightly to markedly divergent, de-
hiscing horizontally or obliquely downwards, the slits confluent across the
apex; pollen grains c. 21-26 » in diameter, the areoles polybrochate,
c. 6-8 pw across.

Female flower: pedicel terete, slender, smooth and glabrous or some-
times sparsely hirsutulous, reddish or olivaceous (sometimes pruinose),
(6—) 10-27 mm. long. Calyx-lobes 6, erect at anthesis, biseriate, the outer
elliptic-oblong and obtuse at the tip, the inner obovate and broader and
more rounded at the tip; lobes 23.2 mm. long, 1-2.5 mm. broad, colored
as the male, the scarcely raised midrib simple to conspicuously branched.
Disk patelliform, fleshy, 6-angled or slightly lobed, foveolate. Styles erect,
the undivided portions connate into a column 0.5—2.3 mm. high, the free

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS Oi,

ends ordinarily sharply reflexed, bifid or parted nearly to the stylar column,
the tips narrow or dilated, flattened, acute, 0.4—-1.2 mm. long.

Capsule oblate-spheroidal, obscurely 6-ribbed, rounded in outline, c.
6 mm. high and 10 mm. broad, rugulose, the veins completely obscure.
Columella c. 3-4 mm. high. Seeds trigonous (only slightly asymmetrical
if at all), 4.2-7 mm. long, 2.8-4 mm. radially and tangentially, smooth,
mottled light brown; hilum submedian.

Flowering probably throughout the year.

The populations of this extremely variable species of Cuba, the Cayman
Islands, and Jamaica may be assigned to one or the other of the two
following subspecies.

45a. Phyllanthus nutans ssp. nutans (PLATE XXIV, figs. E-F).
Phyllanthus nutans Sw. Prodr.
Phyllanthus nutans B ee Baill. ‘Adansonia 2: 15-16. 1862.
Phyllanthus nutans var. trojanus Webster, Contr. Gray Herb. 176: 47. 1955.

Stipules thin, usually precociously camo although ae on
young growth, AE to lanceolate, (2—) 3-5 (—9) mm. long, (1—) 1.5-4
(-6) mm. broad, truncate or cordate at the base, glabrous, Serammincous
or greenish. Leaf-blades mostly ovate but often elliptic, glabrous beneath
(except sometimes at the very base), usually obtuse or subacute at the tip
and obtuse to rounded at the base; margins usually plane. Male cymules
axillary or often at the proximal nodes of thyrses; female cymules
1—3-flowered, usually borne on more or less nodding pseudoterminal naked
thyrses (1.e., ultimate axes with reduced leaves), sometimes the lowermost
cymules in the axils of partially or wholly unreduced leaves. Calyx-lobes
of male and female flowers subentire or obscurely denticulate, thickened at
the base, the midrib simple or sparingly branched. Stylar column 0.5—
2.3 mm. high; style-ends usually sharply reflexed, the lanceolate acute
tips 0.4-1.2 mm. long.

Type: southern Jamaica, Swartz.
DISTRIBUTION: Jamaica and ee Islands (Maps XVIII-XXI).

CAYMAN ISLANDS. Granp CayMAN: Grape Tree Point, dry rocky wood-
land, Proctor 11977 (GH); between Old ee and Wintersland, Proctor 15245
GH).

JAMAICA. ee locality: Alexander (A, GOET), Hooker (W), Jacquin
(W), Swartz (S, oTYPE; A, C, G, P, S, IsotypEes), Wilson 232 (NY
Wullschlaegel Sn, (GH), 900, 1053, 1114 (GOET), 1318 (W). HANOVER:

ish River Mountains, Britton & Hollick 2167, 2170 (F, NY), Harris 10260 (F,
NY. US); Dolphin Head, Britton 2314 (F, NY), Harris 10309 (F, US),
Webster & Wilson 5075 (A, JAM). WESTMORELAND: Negril, rocky wooded
hills, Britton & Hollick 2082 (NY). St. James: Chatham, alt. 300 m., Guwil-
bride & Barkley 22/174 (MICH). St. ExizaBetu: petee an woodland, Hark
9778 (F, US); New Buildings, south of Gutters, Howard & Proctor 14992 (A);
Santa Cruz Mountains, near Hampton School, alt. 2400 ft., Webster & Proctor
5293 (A, JAM). Tretawny: Ramgoat Cave, Howard 14129 (A); Tyre, alt.

58 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xXxx1x

1750 ft., Proctor 9936 (GH); road to Troy, Harris 8687 (F, NY, US);

and vicinity, Britton 929 (NY), Perkins 1331 (GH, HOLOTYPE of var. tr pee
MANCHESTER: Brown’s Town to Porus, hillside, Britton 3272 (F, NY); vicinity
of Mandeville, S. Brown 248 (A, NY); 1.5 miles north of Shooters Hill,
Howard 14107 (A). Sr. ANN: interior of St. Anns, Purdie (P, type collection
of var. purdiaena); Guys Hill, Moneague, Alexander (G); Union Hill, near
Moneague, Britton & Hollick 2747 (F); Discovery Bay, Hunnewell 18844 (GH).
CLARENDON: Croft’s Mountain, alt. 2500 ft.. Harris 11212 (F, NY). St.
CATHERINE: Old Harbour Bay, Little Goat Island, rocky woods, Britton &
Hollick 1855 (F, NY); Great Goat Island, southeastern side, Harris 9301 (A,
C, JAM, NY), 9337 (A. C, NY); Devil’s Race Course, Proctor 7213 (MICH).
St. ANDREW: valley of Yallahs iver: Alexander (NY); Rock Fort, Campodss
6412 (NY); near Hope, Harris 8601 (JAM, NY), 8950 (JAM, NY, US); Hope
River gorge, August Town, Powell 297 (A). PORTLAND: Swift River gorge at
Eden, Proctor 11868 (GH); Uncommon Hill, Proctor 8555 (GH); Port Antonio,
Cave Hill near railway station, Wight 199 (F, NY); John Crow Mountains, above
Ecclesdown, rain-forest, Howard, Proctor & Stearn 14769 (A). St. THOMAS:
Mansfield, Britton 3557 (NY); Golden Valley, Harris 5423 (F, NY); Plantain
Garden River gorge, northwest of Whitehall, Proctor 7419 (GH); Whitehall to
Big Hill, Proctor 7671 (JAM); Big Level, southeast end of John Crow Moun-
tains, Proctor 11820 (GH), Webster & Proctor 5516 (A, BM, JAM. MICH
US)

Not only is P. nutans one of the most variable of the West Indian
species, but its Jamaican representative (ssp. mutans) is certainly one of
the most widespread woody plants on that island; it occurs from sea-level
(at Rock Fort) to 2,500 ft. in the hills of the interior, and from such
arid localities as Great Goat Island to dripping rain forest in the John
Crow Mountains where the precipitation certainly exceeds 200 inches
per year. The only sizeable area on the island where it appears to be
absent is the upper slopes of the Blue Mountains (above 3,000 ft.); al-
though it has most often been collected on limestone, it has also been
found growing on serpentine (e.g., Proctor 7419).

In view of its ubiquitous distribution on Jamaica, it is not surprising
that ssp. nutans should exhibit so much variability; but any attempt to
categorize these variations can only encounter great difficulty. Even the
distinction between the two subspecies is not very well-marked and may
prove to be untenable when more collections are available from the Cay-
man Islands and Cuba. Since it has not proved very useful to express
the intraspecific variation in terms of conventional taxa, a series of maps
plotted for individual characters has been prepared. Many characters, of
course, show a purely random distribution, as Map XIX shows for the
presence or absence of pubescence. It is curious, however, that in the
related species of this section, P. barbadensis and P. botryanthus, there
is no variation in this respect, all individuals being hirsutulous and glabrous
respectively. The distribution of reddish color in the calyx, plotted on
Map XX, is a somewhat more doubtful case, for it might appear that
there is a bicentric distribution of green calyces; however, the number of
samples is small (due to the difficulty in ascertaining the color if there

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 59

MAP XIX

glabrous

@ hirsutulous

MAP XX

© greenish

i

@ reddish

© deciduous - thyrsoid
© deciduous - axillary
€: persistent - thyrsoid

© persistent- axillary

s XIX-XXI_. Distribution of certain morphological characters in popu-
nea : Phyllanthus nutans Sw. The symbols separated by heavy lines in the
upper left-hand and right-hand corners refer to the disjunct populations in the
Cayman Islands and eastern Cuba, respectively. The numbers associated
with the dots in Map XIX indicate the mean stylar length (in tenths of a milli-
meter) of the individual samples, while the numbers in Map XX refer to the
mean seed length in tenths of a millimeter. All herbarium specimens from defi-
nite localities are plotted in Map XIX, but some of these do not appear in
Maps XX and XXI because of incomplete data.

60 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

are no label data) and it seems likely that further collections may break
down the apparent distinction.

There are, however, a number of characters which show undoubted
geographically correlated variation and it might be thought that these
could serve as the basis for the recognition of subspecies or varieties. For
instance, it is quite evident that in seed size, length of stylar column, and
inflorescence-type the plants from the western part of the island show
differences from the eastern populations. But a closer inspection will show
that although there is a general east-west separation (and this holds, on a
larger scale, between Jamaica and Cuba), the characters vary independ-
ently of one another to such an extent that no satisfactory minor taxa
can be defined within the Jamaican plants as a whole. Thus var. trojanus,
which was previously defined on the basis of plants with a long stylar
column, must be relegated to synonymy, for the plants from Dolphin Head
which otherwise agree with the plants from Troy in their inflorescence and
leaves have much shorter stylar columns. The var. purdiaeanus recognized
by Baillon and Mueller is an even less significant variation (of conspicu-
ously bracteate inflorescences) which has a purely random distribution.

The two most striking character differences within P. nutans certainly
deal with the stipule and inflorescence types, and it is indeed upon this
fact that the two subspecies are recognized. It must be admitted, however,
that even here the correlation is far from perfect. The Cuban ssp. grise-
bachianus clearly differs from most forms of ssp. mutans in its solitary
axillary flowers and persistent stipules; but some plants in western Jamaica
have essentially axillary flowers, while in the Cayman Islands and in the
John Crow Mountains at the eastern end of Jamaica occur some anomalous
forms with persistent stipules but which in other respects are more or less
typical for ssp. mutans. It is clear, in the case of the John Crow plants,
that the persistent stipules have been derived independently of those in
ssp. grisebachianus; and a collection from the central part of the John
Crow range (Howard & Proctor 14769) is furthermore so divergent that
it was at first thought to represent a distinct species. In this collection the
leaves are conspicuously corrugated, quite unlike any other specimens of
P. nutans, even though the difference becomes obscured in drying. How-
ever, the inflorescence i is typical for ssp. nutans, and the persistent stipules
are shared by a collection from the Big Level area of the John Crow
Range (Webster & Proctor 5516) which seems otherwise to represent ssp.
nutans. Thus, despite the importance of the morphological divergence in
the John Crow Mountains population, it does not seem practicable to
designate it as a species or even subspecie

The plants from the Cayman Islands, Aer not showing any diver-
gence as extreme as the bullate-leaved form from the John Crows, are of
great interest in that they bridge rather nicely the gap between the two
subspecies. The two collections from Grand Cayman seem assignable to
ssp. nutans by virtue of their ovate leaves, but they are to some extent
transitional, because Proctor 15245 has the axillary flowers of ssp. grise-
bachianus combined with the deciduous stipules of ssp. mutans, while

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 61

Proctor 11977 has the thyrsoid flowers of ssp. nutans combined with appar-
ently persistent stipules as in ssp. grisebachianus. The collection from
Little Cayman (Kings LC42), on the other hand, is clearly referable to
ssp. grisebachianus, since it has elliptic, revolute leaves, persistent brownish
stipules, and flowers axillary (or mostly so). While it is perhaps arbitrary
to assign the Grand Cayman plants to ssp. mutans, there is in any event
no doubt that the Cayman Islands population forms a connecting link
between the two subspecies. Possibly the key to understanding the present
distribution of P. nutans lies in the geological history of the Cayman
Islands.

45b. Phyllanthus nutans ssp. grisebachianus (Muell. Arg.) stat. nov.
(PLATE XXIV, figs. G-H).

he ees ees Muell. Arg. Linnaea 32: 26. 1863; DC. Prodr.
15(2): 6.
Diasperus ne (Muell. Arg.) O. Ktze. Rev. Gen. Pl. 2: 599. 1891.
Stipules scarious, chartaceous, persistent (at least the basal portion),
lanceolate, 2-3.5 mm. long, 0.7—1.5 mm. broad, truncate at the base, gla-
brous or hirsutulous, dark brown. Leaf-blades elliptic, glabrous or hirsu-
tulous beneath, acute at tip and base, margins narrowly revolute. Flowers
entirely axillary, solitary or the male and female paired at each axil.
Calyx-lobes of male and female flowers denticulate, thin, the midrib con-
spicuously branched. Stylar column c. 0. 7 mm. high; style-ends re-
flexed, dilated, bifid or notched, the tips recurved. Seeds c. 4.5 mm. long.

oO

Type: eastern Cuba, Wright 582.

DISTRIBUTION: eastern Cuba (Sagua-Baracoa range) and Cayman
Islands (Mars XVIII-XXI]I).

CAYMAN ISLANDS: Little Cayman, South Town, see pati (BR, NY).
CUBA. OrrenTE: “Cuba Orientali,’ Wright 582 (G, HoLoTyPE; A, BR, G,
GH, GOET, 1sotypes); Monte Verde, 10 May 1859, Ww core WG ex. p. (BR,
G, GH, $; mixed in some collections with Margaritaria scandens).

As here circumscribed, ssp. grisebachianus is readily distinguishable
from ssp. mutans except in the Cayman Islands. The association of the
Kings collection from Little Cayman with ssp. grisebachianus rather than
with the Grand Cayman plants (which are assigned to ssp. nutans) is
based on its pointed revolute leaves hirsutulous beneath (on the midrib)
and its dark persistent stipules. However, it must be granted that the
Little Cayman specimen shows several discrepancies as compared with
the Cuban plants, for it has female pedicels 15-30 mm. long and female
flowers which are neither all solitary nor all axillary to unreduced leaves.
whereas the Cuban collections show pedicels 6-14 mm. long and have
completely solitary, axillary flowers. The population on Little Cayman
may, therefore, be classified as transitional between the two subspecies
but with the stronger leaning to ssp. grisebachianus, whereas the Grand

62 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Cayman plants fall closer to ssp. mutans. Additional collections, particu-
larly from Cayman Brac and from Cuba, might provide a decisive test
of the practicability of the present classification. It may prove to be
unfeasible to maintain two subspecies if additional break-down in the
characters is shown to exist; but, for the present, the Cuban and Jamaican
populations appear to be sufficiently distinct from one another to warrant
separation,

46. Phyllanthus pachystylus Urb. Symb. Ant. 3: 286
(PLATE L jig. 2; PLATE a en I-L).

A slender sparsely branching shrub (usually with a single main stem)
becoming 1—1.5 m. high; main stem 3-5 mm. thick, terete, bark burnished
and reddish brown becoming greyish and fissured; branches mostly steeply
ascending, reddish brown angled, furrowed, glabrous or rarely hirsutulous,
becoming mostly 8-25 cm. long, c. 1.2—2.5 mm. thick, with c. 5-15 nodes.
Leaves: stipules lanceolate to linear-lanceolate, (1.2—) 1.5-2 (—3) mm.
long, 0.4—-1 mm. broad, acute, glabrous, becoming darkened and indurate,
persistent, entire. Petioles stout, glabrous or rarely hirsutulous, 2.5—5 mm
long. Leaf-blades becoming rigidly coriaceous, mostly elliptic-oblong
(varying to narrowly elliptic, ovate, or lanceolate), mostly 3.5—7 (—9) cm.
long, (I1-) 1.5-3.5 (-4.5) cm. broad, obtuse, rounded, or retuse at the
tip, acute to obtuse at the base; above dark purplish (drying blackish)
when young, becoming more or less olivaceous-plumbeous, essentially
smooth (often somewhat wrinkled), the midrib and lateral veins noticeably
sunken; beneath much paler, alveolar-pruinose, the midrib prominently
raised, the spreading to ascending laterals (5-7 on a side) somewhat
raised, the reticulum of veinlets often visible; margins thickened, con-
spicuously revolute.

Monoecious; cymules mostly bisexual, each with 1-3 female and 2—10
male flowers, in the axils of bracts on naked pseudo-terminal thyrses (the
thyrses produced from one or several axils immediately below the tip of
the branchlet); cymules sometimes unisexual.

Male flower: pedicel capillary, smooth or rarely hirsutulous. 7-12 mm.
long. Calyx-lobes 6, rather fleshy, biseriate (but sometimes obscurely so),

PLATE XXIV. Flowers oF sect. Elutanthos.

Fics. A-B. Male and female flowers of Phyllanthus botryanthus Muell. Arg.
(Haught 6556 ae Fras. C-D. Male and female flowers of Phyllanthus
anderssonit Muell. Arg. (Warming 27 |C|). Fics. E-F. Androecium and gynoe-
cium of Pieitiand iis nutans Sw. ssp. nutans (Proctor 15245 |GH]. Webster &
Wilson 5075 [A}|). Fics. G-H. Gynoecium and female calyx-lobe of pee
nutans ssp. grisebachianus (Muell. Arg.) Webster (Wright 1436 [S]). Fras. I-L.
Androecium, gynoecium, inner and outer female calyx lobes of itl
pachystylus Urb. (androecium, Howard 6199 |GH]; others, Ekman 15037 [S
Fics. M—P. Androecium, gynoecium, and inner female calyx-lobes of Died.
thus urbanianus Mans. ie H10435 |S]).

Jour. ARNOLD Ars. VoL. XXXIX PLaTE XXIV

WEBSTER, WEST INDIAN PHYLLANTHUS

64 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxIx

(2—) 2.3-3 mm. long, outer lobes usually oblong and obtuse, c. 1.3—1.6 mm,
broad, with midrib simple or nearly so, inner lobes usually obovate, c.
1.5—2 mm. broad, with midrib usually sparsely branched (occasionally sim-
ple) ; lobes entire or sparsely denticulate. Disk-segments 6, usually rather
massive, reniform to elliptic, foveolate, c. 0.3-0.6 mm. across. Stamens 3;
column rather stout, (0.6—) 0.75-0.9 (-1) mm. high, mostly 0.4-0.6 mm.
thick; anthers sessile atop the column, basally connate, the common con-
nective plane or umbonate, triangular, c. 0.3-0.55 mm. long and 0.4—
0.6 mm. broad; anther-sacs divergent. dehiscing horizontally, the slits
not pera = grains c. 21-24 » in diameter, areoles polybrochate,
4-6 pa

Pemale en pedicel slender, terete, more or less reddish, glabrous or
rarely hirsutulous, 2.5-6 (-8) mm. long, 0.3-0.5 mm. thick. Calyx-lobes
6, distinctly biseriate, (2—) 2.5-3.5 (—4) mm. long, entire or obscurely
denticulate; outer lobes narrowly oblong with midrib almost or quite un-
branched (rarely copiously pinnately branched), c. 1-1.5 (-1.9) mm.
broad; inner lobes obovate or spathulate with midrib always pinnately
branched, c. 1.4-2.2 mm. broad; lobes entire or obscurely denticulate.
Disk shallowly cupuliform, angled, finely crenulate. Styles connate or
coherent into a massive column (0.5—) 0.7-1 (—1.5) mm. high which is
not sharply demarcated from the ovary; free ends of styles recurved,
0.4—0.6 (—0.9) mm. long, more or less dilated, parted % to 34 their length,
the tips broadly triangular to lanceolate, obtuse to acute.

Capsule rounded in outline, c. 5 mm. high and 8 mm. broad, rarely re-
maining entire, the valves pedis Columella rather massive, (3-) 4
(-4.5) mm. high. Seeds trigonous, nearly symmetric, (4—) 4.5—5.4 mm.
long, 2.7-3.1 mm. radially, 2.9-3.5 mm. tangentially, light brown, smooth
(very finely striolate); hilum submedian; micropylar end sometimes de-
veloping a conspicuous whitish caruncle

Collected in flower and fruit April through September.

Type: Cuba, Oriente, Wright 1947.

DIsTRIBUTION: endemic to the Sagua-Baracoa massif, eastern Cuba
(Map XVIII).

CUBA. ORIENTE: Sierra de Nipe, near Rio Piloto, Ekman 2274, 6028, 15037
(S); Charrascal de la Cueva, Mayari, Ledn et al. 19888 (MICH); wooded
hillside, San José, Howard 6199 (GH, NY); near Woodfred, deciduous woods
and thickets, Shafer 3617 (NY); edge of savannas near Sagua de Tanamo, 3
April 1861, Wright 1947 (GH, Lecrotype; G, GOET, isotypes); pinares near
Moa, Acuna 12503 (US), Bucher 102, 107 (NY), Clemente 3555 (MT),
Marie-Victorin et al. 21565 (A, MT), 21705 (MT); Franklyn Mine, Clemente
& Alain 3890 (MICH); Playa La Vaca, Clemente 4918 (MICH); Cerro de
Miraflores, Marie-Victorin et al. 21557 (A, MT); dense pine woods 15 kms.
southwest of Moa mill, Howard 5955 (GH); pinelands on serpentine between
Rio Moa and Rio Yagrumaje, Webster 3757, 3771 (MICH); Cayo Chiquita. 8
km. south of Moa, Webster 3848 (GH, MICH); pine scrub 10 km. south of
Moa, Webster 3895 (GH, 1 oo scrublands 16 and 18 km. south of Moa,
Webster 3906, 3907 (GH, MICH

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 65

This species has a very characteristic appearance in the field due to its
shiny, coriaceous, purplish leaves and nodding inflorescences which may
even be twining in dense undergrowth. Unlike the other West Indian
species of sect. Elutanthos, which are predominantly calciphiles, P. pachy-
stylus appears to be confined to serpentine; and like many species in other
genera growing on the limonite soils of the Moa region, it shows a greater
degree of apparent morphological adaptation to dry conditions than do its
related congeners. Ekman noted on his labels that in the Sierra de Nipe
the species is rare and becoming extinct; but in the Moa area it is certainly
still thriving and in fact is one of the commonest undershrubs in the open
pinelands there.

The closest relationship to P. pachystylus appears to be shown by
P. nutans ssp. grisebachianus, which occupies a clearly allopatric range in
the Monte Verde area, although the gap in range may be no more than
about 20 or 25 miles. The thinner stipules and leaf-blades, solitary
axillary flowers, longer female pedicels, and less massive styles of that
plant present so many distinctions that there can be no doubt as to the
specific distinctness of P. pachystylus.

47. Phyllanthus urbanianus Mansf. Repert. Sp. Nov. 32: 86. 1933.
(PLATE XXIV, figs. M-P).

A small shrub (probably with the aspect of P. pachystylus); main stem
subsimple, c. 2.5 mm. thick, terete, bark dark reddish brown, reddish hirsu-
tulous: ultimate branches reddish brown, terete, somewhat furrowed, red-
dish hirsutulous, c. 20-35 cm. long, 0.9-1 mm. thick, with c. 20-45 nodes.
Leaves: stipules narrowly lanceolate, 1.2-1.8 mm. long, 0.25-0.5 mm.
broad, acute, thin and scarious, olivaceous becoming dark brown and per-
sistent. Petioles rather slender, reddish- or hyaline-hirsutulous, c. 3-4 mm.
long. Leaf-blades chartaceous, elliptic, c. 4-6.5 cm. long and 1.2-2.7 cm.
broad on main stem, decreasing to 1.5—-3 cm. long and 0.4—1.2 cm. broad
at tips of branches, acute at the tip, acute to obtuse at the base; above
dark olivaceous, reddish hirsutulous along the scarcely raised midrib and
laterals; beneath pale, copiously hirsutulous throughout, the midrib and
laterals (4 or 5 on a side) raised, light brownish; margins unthickened,
narrowly revolute.

Monoecious; cymules mostly bisexual, each with a single central female
flower and several lateral males, in the axils of semi-reduced (still leaf-like)
bracts on distal branches (which are homologous with the “naked” thyrses
of P. nutans); individual cymules occasionally replaced by depauperate
inflorescence axes.

Male flower: pedical capillary, smooth to copiously hirsutulous, c. 8-10
mm. long. Calyx-lobes 6, c. 2.2-2.5 mm. long, more or less biseriate:
outer lobes narrowly oblong, c. 0.75 mm. broad, inner lobes elliptic-oblong,
slightly over 1 mm. broad; lobes all rounded at the tip, reddish with rather
ill-defined yellowish margins, the midrib simple or sparingly branched.
Disk-segments 6, flattened, roundish, foveolate, c. 0.2-0.3 mm. across.

66 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Stamens 3; column c. 0.8 mm. high, slightly constricted above; anthers
sessile, c. 0.25 mm. long and 0.4 mm. broad, alternating with three erect
apiculae c, 0.2 mm. long; anther-sacs rather broadly divaricate, dehiscing
horizontally; pollen grains 18-21 ,» in diameter, areoles transitional be-
tween oligobrochate and polybrochate, c. 4—

Female flower: pedicel slender, terete or nearly so, sparsely to copiously
hirsutulous, 12-15 mm. long. Calyx-lobes 6, biseriate: at anthesis, outer
lobes linear-oblong, obtuse, with simple midrib, 2—2.3 mm. long and
0.8-0:9 mm. broad, inner lobes obovate, rounded at the tip, with sparingly
branched midrib, 2.7-3 mm. long and 1.3-1.5 mm. broad; lobes chartace-
ous, reddish, essentially entire, becoming reflexed in fruit. Disk shallowly
cupuliform, 6-angled, rather fleshy, crenulate. Styles connate into a mas-
sive column c. 1 mm. high and 0.5 mm. broad; free ends of styles recurved,
dilated, oblong, obtuse or emarginate, c. 0.4—0.6 mm. long,

C apsule c. 4mm. high and 7 mm. broad, somewhat rugulose. Columella
c. 2.5 mm. high. Seeds trigonous, slightly asymmetric (somewhat um-
bonate at one corner), c. 4.2 mm. long, 2.7 mm. radially, 2.9-3 mm. tan-
gentially, smooth, mottled light brown; hilum submedian.

Type: Haiti, Dept. Sud, Massif de la Hotte, western group, Les Roseaux,
Hab. Gros-Roche, rocky forest, hard limestone, alt. 400 m., rare, 27 June
1928, Kkman H-10435 (S, HoLotypre: A, US, tsorypes).

DISTRIBUTION: known only from the type collection (Map XVIII).

This rare endemic species is of particular phytogeographic interest, be-
cause it is the only representative of the section on Hispaniola. It resem-
bles both P. pachystylus and P. nutans (especially ssp. grisebachianus),
but differs in its more hirsutulous parts and Spa staminal
column and styles, so that there would appear to be no reason to ques-
tion its specific distinctness. The Cuban and a ae. of P.
urbanianus provide another good demonstration (in addition to species
relationships in sects. Cyclanthera and Hemiphyllanthus) of the profound
floristic division between the Sellean peninsula of Haiti and the remainder
of Hispaniola.

Subgenus VIII. Xylophylla (L.) Pers. Syn. Pl. 591. 1807; emend.

X ylophylla L. Mant. 2: 147-148. 1771.

‘Trees or shrubs with phyllanthoid branching, the branchlets pinnatiform
or bipinnatiform; monoecious or very rarely dioecious. Male flower; calyx-
lobes 4-6; disk of as many segments, these free or united; stamens 2-15,
free or more commonly united; anthers dehiscing vertically to horizon-
tally; pollen grains globose, areolate. Female flower: calyx-lobes 5 or 6
(rarely 4); disk cupuliform or patelliform; ovary of 3 carpels: styles
bifid or multifid, sometimes dilated at the tips. Fruit capsular; seeds 2 in
each locule.

Included in this large, entirely American subgenus of about a dozen
sections and 60 species are the majority of the neotropical woody species

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 67

of Phyllanthus. The West Indies are definitely the center of distribution
and apparently also of the evolution of the group, but a few additional
sections (e.g., Oxalistylis and Ciccastrum) are confined to South America.
In Mueller’s treatment in the “Prodromus” the sections and species here
brought together were much scattered, the following of his sections be-
longing (at least in part) to subg. Xylophylla: 12, 16-19, 34 (in small
part), 35-36, and 44. Persoon’s original conception of subg. Xylophvlla
as including only the phylloclade-bearing species was of course much
narrower than that here adopted, and in fact corresponds to sect. XWo-
phylla alone.

Because of the dominant position of representatives of subg. X ylophylla
in the West Indies, its relationships are of particular interest. However,
although certain lines of affinity are very apparent within the group, there
are several unresolved problems which make impossible an accurate trac-
ing of phylogeny. It is certain that there is a close relationship between
subg. Xylophylla and sibg. Botryanthus, and the approach is nearest be-
tween sects. Asterandra and Elutanthos, respectively. Some of the Central
American species of sect. Elutanthos (e.g., P. grandifolius) resemble sect.
Asterandra so clearly that there can be little doubt of a significant kinship.
However, there are some obstacles, to be discussed more fully farther on,
which make it at least uncertain that subg. XylopAylla can be directly
derived from subg. Botryanthus via sect. Asterandra. The small number
of stamens in the flowers of Botryanthus and the large capsule of such
species as P. grandifolius are features which appear to be derivative; so
that it is possible to read the evolutionary series the other way around
and to postulate that subg. Botryanthus has been derived from sect.
Asterandra by reduction in the androecium accompanied by loss of phyl-
lanthoid branching and increase in fruit size.

Within subg. Xylophylla two main phyla can be discerned: a series be-
ginning with sect. Williamia and running through sects. Thamnocharis
and Orbicularia: and a series proceeding from sect. Asterandra to sects.
Epistylium, Hemiphyllanthus, and Xylophylla. Standing alone is sect.
Omphacodes, which is aberrant in many respects and which in fact resem-
bles sect. Ciccopsis of subg. Cicca more than it does any single section of
subg. Xylophylla. It is classified here because of its areolate pollen grains,
but without strong conviction, and its relationships need to be further
investigated.

The initial dichotomy between sect. Williamia and sect. Asterandra has
some puzzling aspects which cannot yet be resolved. That the higher
stamen number in Williamia may be a primitive character is attested by
the clearly documented reduction-series to a lower number in the derived
sections; furthermore, the seeds of Williamia are less highly modified than
those of Asterandra. However, the South American sect. Oxalistylis, which
is closely related to Asterandra, has more unspecialized seeds, and perhaps
can be thought of as a group more or less coérdinate with Williamia, if
indeed the two are not descended from some immediate common ancestor.

The anatomical evidence from leaves, as shown earlier in this study

68 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

(Jour. Arnold Arb. 37: 220. 1956), demonstrates for many taxa of this
subgenus an interesting correlation between floral modification and increas-
ing sclerification of foliar tissue. The rationale for this would appear to
be that the evolutionary history of many of the groups of subg. XwopAvlla
has been one of increasing adaptation to xeric conditions. However, the
adaptive radiation has been rather complex, so that no over-all generaliza-
tions on the ecology of the species can be made. In both the Williamia and
Asterandra lines, however, the end-products of evolution (viz., sects. Or-
bicularia and Xylophylla) comprise species so altered in appearance that
their ancestry could scarcely be guessed if it were not possible to trace it
back through intervening species.

KEY TO THE SECTIONS
1. Branchlets pinnatiform.

2. Styles each terminating in a dilated more or less crenate to lacerate stigma.
3. Disk-segments of male flower free or at least not completely united;

seeds thin-walled, neither fissured nor mottled; petioles without un-

dulate marginal ridges

4. Leaves obtuse or cme at the tip; styles not calyptriform or
if so then branchlet leaves opposite; stamens 3-15. .. 16. Williamia

4. Leaves acuminate; styles calyptriform or united into a massive

mn; stamens 2 or 3 (rarely 4); at least some inflorescences
cauliflorous (except in P. axillaris). .............. 21. Epistylium
3. Disk-segments of male flower connate into a massive ring.

4. Leaves acuminate, neither revolute nor golden beneath, the petiole
with conspicuous undulate- eee oe seeds es walled, mot-
tled (as in Ricinus); stamens 3-5. .............. Asterandra

4, Bee blunt at the tip, ee sélden beneath, : ene with-

d margins; seeds thin-walled, — and fissured;
ae MCN les aug agen cakes een den ibe . Glyptothamnus 19

2. Styles bifid, the branches usually slender, never ae (rarely adaxially
auriculate).

3. Capsule somewhat fleshy, tardily dehiscent; leaf-blades chartaceous,
the blade decurrent on the petiole; stamens 3. Ming ota 19. Omphacodes
3. Capsule dry, promptly dehiscent; leaf-blades chartaceous to coriaceous,

the blade not decurrent on the petiole; stamens 2—

4. Leaf-blades small (less than 2 cm. long), with conspicuous meso-
phyllar sclereids, often concave beneath and with revolute margins;
seeds less than 3 mm. long; stipules more or less persistent; flow-
ers mostly appearing after the leaves. .......... 18. Orbicularia

™ Sect. be deka tect sect. nov. Frutices monoicae, foliis coriaceis revolutis

)
integro, stylis dilatatis laceratis; <ainibus fuscis sulcatis. — Species typica Phyllanthus
chryseus Howard (sectional epithet from Gr. glyptos, penn and thamnos, shrub, in
allusion to the appearance of the massive flowers and leaves).

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 69

4. Leaf-blades larger, the sclereids aggregated within the plane, thick-
ened marginal rim (rather than in the mesophyll); seeds over

3 mm. long; stipules caducous; flowers appearing with the leaves.
17. Thamnocharis

2. Branchlets not greatly dilated, bearing typical leaves, the penultimate
AxISeITIChUState Ol SCUIMV. «nie cag ge 20 Bor ee eens 23. Hemiphyllanthus
2. Branchlets dilated and transformed into phylloclades, the leaves normally

absent except on seedlings; axes never incrustate or scurfy. 24. Xylophylla

ect. 16. Williamia (Baill.) Muell. Arg. Linnaea 52: 4. 1863; DC. Prodr.
1542.) O75 LeOO.
Williamia Baill. Etud. Gen. Euphorb. 559, pl. 27, figs. 9-10. 1858.

Shrubs with phyllanthoid branching; leaves chartaceous to coriaceous,
stipules deciduous or persistent. Monoecious, cymules bisexual, the flowers
maturing after the leaves. Male flower: calyx-lobes 5 or 6; stamens 3-15
(very rarely 2), filaments united into a column, anthers in 1-3 whorls;
pollen grains areolate. Female flower: calyx-lobes 5 or 6; disk plane,
angular, often massive; ovary sessile or definitely stipitate; styles erect or
spreading, free or connate, the distal ends dilated and lacerate or dentate.
Capsule oblate, dry, not veiny; seeds colliculose or verruculose, less than
3 mm. long.

Type species: Williamia pruinosa Baill. (= Phyllanthus discolor Poepp.
ex Spr.).

As here defined, the circumscription of sect. Williamia is enlarged from
that previously held (Contr. Gray Herb. 176: 57. 1955; Jour. Arnold
Arb. 37: 220. 1956) by the inclusion within it of sect. Williamiandra.
The emended section thus comprises seven Cuban species which are in-
dubitably related, although they are morphologically so far divergent from
one another that the group might at first sight appear to be an unnatural
one. A well-marked evolutionary series of increasing specialization may
be traced between P. discolor, with its thin, unsclerified leaves and an-
droecium of up to 15 stamens, and P. incrustatus, which has highly scler-
ified leaves and only 2-4 stamens. However, despite the prominence of
these phylogenetic trends, the relationships of the species of the three
subsections cannot be visualized as a simple and clear-cut “family tree.”

The affinity of sect. Williamia with sects. Orbicularia and Thamnocharis
is particularly clear in view of the fact that the latter appear to be its
direct offspring. In fact, the ancestry of both sections may with reason-
able confidence be traced back to subsect. Discolores, since Orbicularia
and Thamnocharis would appear to have been derived from progenitors
similar to P. microdictyus and P. discolor, respectively. In both instances
differentiation of the derived taxa has involved increased foliar sclerifica-
tion, but this has occurred in a strikingly different manner, for the sclereids
are scattered through the mesophyll in sect. Orbicularia, whereas they are

70 JOURNAL OF THE ARNOLD ARBORETUM _[vot. -xxxrx

aggregated within the marginal rim of the lamina in sect. Thamnocharis.

Because of its ancestral relationship to sects. Orbicularia and Tham-
nocharis, as well as the relatively unspecialized morphological features of
its subsection Discolores, sect. Williamia has been placed first in the linear
arrangement of sections under subg. Xvlophylla. However, this is not
intended necessarily to imply that Williamia is the most primitive of al
the taxa within the subgenus. The differences in stamen number on which
Mueller laid such stress appear far less significant in the light of present
knowledge, for the stamen number of P. discolor is mostly 10-13 instead
of the 15 reported in the literature; and in any event, even the former
number is by no means a certain indication of primitiveness. The sister
species of P. discolor, P. microdictyus, is more similar in floral characters
to the South American sect. Oxalistylis. It appears that P. discolor and
P. microdictyus may best be regarded as vicarious descendants from a
common ancestor, specialization having progressed further in the flowers
of the former and the leaves of the latter

The flowers of P. salviaefolius, the type (and only?) species of sect.
Oxalistylis, are so similar to those of P. microdictyus, particularly with
regard to the gynoecium, that the possibility of uniting sects. Oxalistylis
and Walliamia might possibly be considered. However, the South Amer-
ican species differs by having several female flowers per cymule, differently
ornamented seeds, and very different leaves much more like those of sect.
Asterandra; so, for the time being, it seems justifiable to maintain the
two sections as distinct.

KEY TO THE SUBSECTIONS AND SPECIES

1. Leaves of branchlets alternate; stigmas (style-tips) not calyptriform.
Stems smooth, lenticels (if present) sparse and inconspicuous; stamens
Lee Soe eka eb we hey ce ane exec ee ead oe Subsect. 16a. Discolores
a; Leaves chartaceous, not highly sclerified.
4. Calyx-lobes of female flower 2-3 mm. long; styles less than 2 mm.
long; stamens mostly 9-13; leaf-blades irregularly reticulate above.
48. P. discolor

bo

4. Calyelobes of female flower 7—8.5 mm. long; styles mm
long; stamens 6-10; leaf-blades ane and evenly reticulate
the areoles straight- aed. bp waveree fae oe 49. Pit eae ee
3. Leaves coriaceous, with abundant mesophyllar sclereids. .. 1...
ES th ite ne Gay nte Sak oe ee Guage A ped Eg oo P. cristalensis
2. Stems incrustate, the ie breaking up into small platelets separated by
spongy tissue; stamens (2—) 3-6. . Subsect. 16b. Incrustati

a

3. Stipules of branchlets acicular- conduplivaie, ey caducous; staminal
column prolonged into a terminal apiculum; disk of female flower with
a crenate upturned rim; leaves mostly 1.5-2.5 cm. long. ......
ee Raa 4 eee ead odes eee eee Sly. 3 williamioides
3. Stipules a branchlets nearly plane, indurate and es staminal
column not apiculate; disk of female flower plane, en
4. Leaf-blades 2.5-4.5 cm. long, prominently ae beneath;

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 71

~

pedicel of female flower 10-15 mm. lomo asta ens. o, Olu a as c5
52. P. excisus

ee ee sg fe Se a ene Subsect. 16c. Mirifici
Oulyespecicct (ier We eas ae ek OR ey age ear ae 54. P. mirificus

Subsect. 16a. Discolores, subsect. nov.
Williamia Baill. Etud. Gen. Euphorb. 559, pl. 27, figs. 9-10. 858.
Phyllanthus sect. Williamia (Baill.) Muell. Arg. in DC. Prodr. 15( 22, 328:
1866 (ex p.).

Stems smooth, lenticels obsolete or sparse, leaves of branchlets alternate,
chartaceous or (in P. cristalensis) coriaceous and sclerified; stamens 6—15
(number unknown in P. cristalensis); styles erect, the tips moderately
dilated.

Type species: Phyllanthus discolor Poepp. ex Spr.

There can be no doubt that subsect. Discolores occupies a significant
position in the evolutionary sense, for sects. Thamnocharis and Orbicularia
may with a high degree of probability be spoken of as derived from
P. discolor and P. microdictyus, respectively. If, as suggested above, the
high stamen number of P. discolor is secondarily derived from a 5- or
6-merous condition, there has at any rate been a reduction in number
in the species of the derived taxa (subsect. Incrustati, sects. Thamnocharis
and Orbicularia).

The inclusion of P. cristalensis within this subsection can only be pro-
yisional. for its floral characters are still insufficiently known. On the
basis of its vegetative characters, it would appear to be related to P. micro-
dictyus and P. excisus, since it has the smooth axes of the former and
the highly sclerified leaves of the latter. Should its present position appear
justified after the examination of adequate fertile material, P. cristalensis
would represent a nearly schematic connecting link between subsects.
Discolores and Incrustati.

48. Phyllanthus discolor Poepp. ex Spr. Syst. 3: 21. 1826.
(PLATE XXV, figs. A-C).
Phyllanthus pruinosus Poepp. ex Rich. in Sagra, Hist. Nat. Cuba 11: 2106.
1850; non sensu Muell. Arg. in DC. Prodr. 15(2): 387. 1866.
Williamia pruinosa (Poepp.) Baill. Etud. Gen. Euphorb. 560, pl. 27, figs. 9-10.
1858

Diasperus discolor (“Spr.”) O. Ktze. Rev. Gen. 2: 599. 1891.
Diasperus pruinosus (“Rich.”’) O. Ktze. op. cit.

00.
Phyllanthus decander Sessé & Moc. Fl. Mex. ed. 2, 212-213. 1894.

A shrub becoming c. 1-2 m. high, sparsely branching, the main stem(s)

iz JOURNAL OF THE ARNOLD ARBORETUM [Val ReNIK

slender (mostly 2-3 mm. thick), light brown, smooth, terete, with brown
pith. Cataphylls not indurate, early deciduous: stipules triangular to
lanceolate, acute or acuminate, 4-8 mm. long and 1-2 mm. broad (smaller
on weak shoots), truncate at the base, thin and scarious, dark reddish
brown with narrow paler denticulate or entire margins; blade linear-lan-
ceolate, acuminate, 4-6 mm. long and 0.7-1 mm. broad. Deciduous
branchlets 8-15 (—25) cm. long, 0.7-1.3 mm. thick, light brown or stra-
mineous, terete to distinctly flattened, with (5—) 7-14 (-16) leaves: first
internode 1.5—3 cm. long, median internodes mostly 1-2 cm. long. Leaves:
stipules thin and scarious, early deciduous (except at tips of branchlets) ,
linear-lanceolate, acuminate, (2.5—) 3.5-6.5 (—7) mm. long, (0.4-) 0.7-
1.2 mm. broad, brownish, with entire margins. Petioles 2-4 mm. long,
subterete, dark brownish, smooth. Leaf-blades chartaceous, mostly ovate
or elliptic, (2—) 2.5-4 (-5.5) cm. long, (1-) 1.3-3 (-3.5) cm. broad,
narrowed to an obtuse (less commonly rounded or emarginate) incon-
spicuously apiculate tip, acute to truncate or rarely subcordate at the
base; above olivaceous or bright green, the midrib and the arching,
crooked laterals slightly raised; beneath whitish- or creamy-pruinose (due
to a waxy coating), the midrib salient, the laterals (c. 5-7 on a side)
raised, minor veinlets forming a fine reticulum: margins unthickened,
plane or casually revolute.

Monoecious; branchlets usually but not always floriferous; cymules
bisexual, each most often with 1 or 2 (3) central female and c. 4—6
lateral male flowers; bracteoles of cymules triangular, dark brown, scarious,
persistent.

Male flower: pedicel (1.5—) 2-6 mm. long. Calyx-lobes 5, chartaceous,
subequal or usually unequal, rounded at the tip, entire or minutely den-
ticulate near the apex; outer lobes ovate or oblong, 1.2—2.5 mm. long and
0.9-1.3 mm. broad; inner lobes obovate or spathulate, 1.7—3.5 mm. long
and 1.3-1.9 mm. broad, 3- to 5-nerved from the base. Disk-segments 5,
roundish, smooth and entire, c. 0.3-0.5 mm. broad. Stamens (7—-) 9-13
(—15); filaments connate into a column c, 1-2 mm. high, the free portions
0.2-0.75 mm. long, erect or arching and spreading; stamens spirally
arranged or in two superposed whorls with a whorl of 3 (—5) terminating
the column; anthers c. 0.2-0.4 mm. long, 0.25—0.5 mm. broad, the upper
erect or ascending, the lower often spreading or deflexed: anther-sacs
slightly divergent, dehiscing longitudinally, the slits contiguous but usu-
ally not confluent; pollen grains 18-27 » in diameter.

Female flower: pedicel slender, becoming (3—) 5-13 (-17) mm. long.
Calyx-lobes 5, distinctly unequal, rounded at the tip, entire, with several
veins from the base (at least in the larger); outer lobes ovate or oblong,
1.2-1.8 mm. long, 0.6-1.3 mm. broad: inner lobes (2—) 2.5-3 mm. long,
1.5—2.8 mm. broad. Disk rather massive, obtusely 5-angled, foveolate, the
margins plane and entire. Ovary oblate or foveolate, with 3 low but dis-
tinct ribs, usually sessile; styles erect, somewhat unequal, free or connate
at the base, c. 0.3-0.7 mm. long, the dilated ends (stigmas) roundish or
crescent-shaped, crenulate or lacerate (but often appearing entire due

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 73

to recurved tips of lobes), sometimes adaxially auriculate, 0.5-0.8 mm.
road.

Capsule oblate, smooth or somewhat rugulose, rounded in outline, red-
dish brown, c. 2.1-2.3 mm. in diameter. Columella slender, 1.2-1.5 mm.
long. Seeds acutely trigonous, symmetric, 1.7-1.8 mm. long, 1.1-1.3 mm.
radially, 1.2-1.25 mm. tangentially, brown with somewhat irregular longi-
tudinal lines of transversely elongate dark reddish brown dots; hilum tri-
angular, c. 0.4 mm. long.

Collected in flower and fruit January through August.

Typr: “Ad rivulos Cubae in Sumidero,” Poeppig (W, LEcTOTYPE; BR,
F, isorypEs). Poeppig’s material appears to include more than one collec-
tion. The lectotype collection is the one associated with the printed label
reading ‘‘Phyllanthus discolor En pl. Cub. MSS. . .’: but several additional
collections made in Cuba in 1823 by Poeppig (all apparently in Matanzas
province) are preserved in the Vienna herbarium. A specimen in the Paris
Museum bearing the label ‘““Phyllanthus pruinosus Poepp. (ad Spreng.
ad Phyll. polygonoid. Nutt. ductus) Cuba. Sylvae lucidae” and a specimen
collected for Sagra (also in the Paris Museum) represent the type material
of P. pruinosus Poepp. ex Rich.

Although applied by Mueller and others to the plant known in this
work as P. caroliniensis ssp. saxicola, the name P. pruinosus Poepp. ex
Rich. is actually a synonym of P. discolor. Its publication was apparently
inadvertent, for Poeppig’s specific name, when it appears on the labels in
script specially printed for his collection, is usually associated with speci-
mens of P. caroliniensis ssp. saxicola. Evidently there was a mixture of
labels in the case of the Poeppig specimens which went to Paris, so that
Richard was misled into describing the present plant under a different
name than that intended by Poeppig. Mueller (Linnaea 32: 30. 1863)
independently published P. pruinosus for the other plant, as designated
by Poeppig, but this name must of course be rejected as a later homonym.

DIsTRIBUTION: serpentine barrens and hillsides, central and western
Cuba (Map XXIT).

CUBA: without locality, Sagra (A,P,W: syntypeEs of P. pruinosus). PINAR
peL Rro: Sumidero, between Hoyo Colorado and Francisco, at a brook, Ekman
18205 (S); Los Organos, Sabanilla, Wright 1941 ex p. (F, GH, GOET, MO,
NY, P, S, US; data ex GH); pinelands, Loma Cajalbana, La Palma, Alain 2362
(GH); Sagua, Bahia Honda, Wright 1949 (G, GH, GOET;° data ex GH);
Rangel, Zambumbia Hill, Ledn 12775 (MICH, NY). Hapana: La havane, 1821,
Leman (P); Guanabacoa [ex Flor. Mex. ed. 2], Sessé & Mocino 4566 (F, TYPE
COLLECTION of P. decander); eruptive rock soil, Madruga, Britton, Britton, &
Shafer 631, 690 (NY), Leén 3330 (MT, NY). Maranzas: cuabales northwest
of Pan de Matanzas, southeast of Canasi, Ekman 16469 (S); savanna, Sabanilla
de la Palma, Ledn, Roca, & Edmund 9652 (NY); Tetas de Camarioca, serpen-
tine barren; Britton, Britton, & Wilson 14063 (NY); Sumidero, Poeppig (W,
LECTOTYPE; BR, F, MO, 1sotypes)?°; San Miguel de los Banos, Killip 13867 (US),

20 This collection is also a syntype collection of P. pruinosus.

74 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXXIXx

Leon & Roca 8907 (NY). Las Vittas: Palm barren, Santa Clara, Britton &
Cowell 10202 (NY); arroyo Ciento Viejo, 12.5 km. east of Santa Clara, Howard
5081 (GH, MT), Webster et al. 250 (GH); Trinidad Mountains, El Cumbre.
thickets at a brook, Ekman 13949 (S): near Pico Potrerillo. road to Aguada
del Santo. open hillsides, quite common, Ekman 13977 (S); Sierra de San Juan,
Mina Carlota (southeast of Cumanayagua), alt. 300-400 m., Howard 5663 (GH,
MT, NY); Buenos Aires, Roig & Acuna 6152 (NY); Loma de Tibisial, Sancti-

ae
® =P CRISTALENSIS
@ P DISCOLOR
A P MICRODICTYUS
© PX PALLIDUS

Mar XXII. Distribution of sect. Williamia subsect. Discolores in Cuba.

Spiritus Mountains, Ledn & Clement 6662 (NY); Lomas de Banao, Luna 754
(NY); Lomas del Banao, Loma del Obispo, Ekman 16282 (S), Ledén 1300 (NY).

Among the woody species of Phyllanthus in western Cuba, P. discolor
is second only to P. orbicularis in abundance and frequency. As a conse-
quence of the restriction of P. discolor to areas of serpentine outcrops, it
occurs as a number of more or less isolated populations which exhibit a
noticeable variation in leaf shape and stamen number. However, the
variation within individuals is so high and the sampling of populations so
inadequate (despite the rather considerable number of specimens exam-
ined) that the extent of geographical variation is difficult. to assess. This
inexact knowledge of the normal range of variation within the species
presents a distinct hindrance in the analysis of some highly anomalous
specimens from northern Pinar del Rio.

These discordant collections, which all come from the Cajalbana region
between La Palma and Bahia Honda, represent the plant to which the
name P. pallidus has been applied. Grisebach (Goett. Nachr. 1865: 168.)
differentiated P. pallidus from P. discolor on the basis of its glaucous,
thicker leaves and greater crowding of anthers on the staminal column,
while Mueller (DC. Prodr. 15 [2]: 328. 1866) proposed to distinguish
it by its fewer flowers, the female with equal calyx-lobes and shorter
pedicels, and by the pale leaves with indistinct venation. However, even
a superficial survey of the specimens at hand is sufficient to show that
none of these characters is stable or diagnostic. Consequently, the Cajal-
bana population was recently (Contr. Gray Herb. 176: 57. 1955) reduced
to varietal rank under P. discolor: but even this disposition scarcely ap-

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 1

peared satisfactory in view of the extreme fluctuation of characters and
the lack of geographical separation between the two entities.

An intensive study of a population sample taken between La Mulata and
San Juan de Sagua, together with a reexamination of available herbarium
specimens of P. pallidus, has led to the following interpretation of the
situation.

Phyllanthus < pallidus Wr. ex Griseb. Goett. Nachr. 1865: 168. 1865, pro sp.
(= P. discolor X P. comptus.)

Phyllanthus sagraeanus Urb. Symb. Ant. 9: 182. 1924.
Phyllanthus discolor var. pallidus (Wr. ex Griseb.) Webster, Contr. Gray
57. 1955

CUBA. Prnar vex Rio: Cajalbana, Bahia Honda, Wright 1950 (GOET, HoLo-
TvPE of P. pallidus; G, GH, tsotypes); Loma de Cajalbana, cuabales. Ekman
10471 (NY, S; TYPE COLLECTION of P. sagraeanus) ; dense cuabales, eastern slope
of Loma Cajalbana, Ekman 17347 (S); orillas del arroyo, cerca de la cumbre
de la Cajalbana, Ledn & Charles 4956 ex p. (NY); San José de Sagua to San
Marcos, on serpentine rock, Shafer 11971 (F, MO, NY, US); between La Mulata
and San José de Sagua, Webster 4652, 4653, 4054, 4657 (undistributed) ; Bahia
Honda, dry rocky hillside, Wilson 9417 (NY); locality questionable. Wright
194nex Dp AUN Ys, Ps 5,.US):

Some of the specimens cited have been annotated by formula rather
than by the hybrid name, since they resemble P. discolor rather strongly
and show only a subordinate influence of P. comptus. The type collection
of P. sagraeanus is the most nearly intermediate, and Wright 1950 (the
type of P. pallidus) approaches this condition. The most striking vegeta-
tive difference between the two hybridizing species, the type of cataphyll
(much more massive in P. comptus), is obscured in most specimens by the
prevalence of discolor characteristics. Several collections, notably Ledn &
Charles 4956 and Wright 1941, contain a mixture of P. pallidus and a
form indistinguishable from typical P. discolor; in other collections there
are intermediates of varying degrees.

The decision to regard the plants assigned to P. pallidus as hybrids be-
tween P. discolor and P. comptus has been taken after an extended anal-
ysis involving comparisons of leaf anatomy, pollen and seed fertility, and
gross morphology. Anatomical studies of cleared leaves indicate that in
P. pallidus the veinlets are c. 20-40 » in diameter and are quite inter-
mediate between the tenuous veinlets of P. discolor (mostly 10-20 ». thick)
and the knobby ones of P. comptus (mostly 30-60 » thick). Of even
greater interest, although not entirely conclusive, are the results of a study
of pollen grains taken from herbarium specimens and stained in lacto-phenol
with cotton blue, which show that there is pollen sterility (though to a
highly variable degree) in specimens which are transitional between P.
discolor and P. pallidus, and essentially complete sterility in specimens
which are morphologically good P. pallidus.

Since the assignment of specimens to P. pallidus or to the “transitional”
column is somewhat arbitrary, Table IT does not make sufficiently explicit

76 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

TABLE II +
POLLEN AND SEED FERTILITY

Transitional between

P. discolor P. discolor and P. pallidus
Alain 2362 94 Ekman 17347 99
Ekman 13977 99 (+) Shafer 11971 99
Ekman 16469 10, 46 Webster 4652-1 83 (—)
Howard 5081 72 (+) 4652-2 99 (—)
Howard 5563 96 (+) 4652-3 91
Webster 250 99 (+) 4652-4 92
4652-5 20
P. pallidus 4652-6 14
Ekman 10471 — (*) 4032-7 70
Webster 4653-1 — (—, *) 4652-8 97
4654 64, 93 (-, *) 4652-9 100
4057 28, 40 (—) 4652-10 67
4652-11 98
4652-12 61 (—)
4653-2 86
4653-3 * (*)
figures following each collection number refer to the percentage of goo
pollen grains as determine examination of 200 grains from each co ection; in a
few cases a se determination has been made from a different flower on the same

in doubtful instances they have been scored as defective. The plus and minus signs in
parentheses refer to the production of seeds, any collection with fewer than 50% viable
seed having been scored minus; lack of either symbol indicates that the degree of fer-
tility could not be ascertained from the specimen. The asterisk denotes the preduction
of monstrous flowers,

the important fact that the most nearly “typical” specimens of P. pallidus
(1.€., ones such as Ekman 10471 which are exactly intermediate between
P. discolor and P. comptus) are completely sterile, only the monstrous
rudiments of female flowers being produced. Other collections, exemplified
by Webster 4654 and 4657, definitely lean toward P. pallidus but have
more or less normal flowers which are partially fertile. In the transitional
plants trending toward P. discolor the fertility is as variable as the mor-
phological features. Webster 4653—3 epitomizes the situation: it has
vegetative features approaching typical P. discolor, but its completely
monstrous flowers betray its reproductive instability.

The absence of collections intermediate between P. pallidus and the
other presumed parent species P. comptus may appear strange, but is
perhaps explained (at least in part) by the apparent rarity of the latter,
which has been collected only three times, However, if this explanation is
correct it raises the even knottier problem of why the hybrid form should
be so much commoner than. one of its parents, particularly since it ap-
pears to become fertile only when back-crossed to P. discolor, Possibly
some ecological factor plays an important role in determining the peculiar
character of the population.

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS ta

The prevalence of “pallidus” characters in the Cajdlbana population of
P. discolor might be adduced as an instance of ‘“‘introgressive hybridiza-
tion,” although it is not evident that any characters of P. comptus have
been diffused into any populations of P. discolor outside of the Cajalbana
area. Much more striking than any leakage of characters is the exag-
gerated variability of the local population as a whole; a considerable
number of specimens show anomalous characteristics. An outstanding
example of this is Ekman 10433, which vegetatively can scarcely be dis-
tinguished from typical P. discolor and which has highly fertile pollen;
but its flowers are exceptional in the long staminal column and styles and
the stipitate ovary. These peculiarities, especially the stipitate ovary, sug-
gest a resemblance to the related species P. microdictyus, which is known
only from Oriente province. However, it appears probable that these
unusual features have some connection with the local hybridity and do
not really signify any relationship to the species of Oriente.

The enhanced degree of variability is particularly marked in the stamen
number, as shown in Table III. Since P. comptus usually has only 5 or 6
stamens, one would expect that its hybrids with P. discolor might have a
number approaching the usual 9-13 of the latter. Insofar as the lower
numbers (under 10) are concerned, this is reasonably well borne out. But
the occurrence in P. pallidus of numbers much higher than those in
P. discolor is entirely unexpected. Although the number of androecia
counted (50 each of P. discolor and P. pallidus) is unfortunately small,
there can be little doubt that mere chance cannot account for 11 of the
pallidus flowers having a higher number than any of the discolor ones.

It seems quite likely that this wholly anomalous increase in stamen
number in P. pallidus is the result of the same cause which determined
the formation of a stipitate ovary in Ekman 10433, i.e., the morphogenetic
disturbance in the hybrid plant which is presumably due to imbalance
between the chromosome complements of the parents. It is conceivable
that the situation in P. pallidus presents us with an insight into one
mechanism of phylogenetic change, for if aberrant plants of P. pallidus
with androecia of 16-18 anthers should happen to be selected out as a
separate fertile population, a new species characterized by higher stamen
number might result. Possibly the prevalent numbers (9-13) in P. discolor
arose in this same way, for the primitive number in the genus is certainly
S or 6 as in subg. Kirganelia; and since P. discolor has no other particu-
larly primitive characters, it is probable that its increased stamen num-
ber, too, is derived.

The relationships of P. discolor are reasonably well-defined, for on
purely morphological grounds its affinity to the following species, P. micro-
dictvus, is apparent; and the general resemblance (though much less close)
to P. comptus supports the evidence (from the existence of P. pallidus)
that P. discolor can cross with that species. Furthermore, P. comptus in
its morphological details is transitional between P. discolor and the other
species of sect. Thamnocharis, so that it presumably has evolved from
P. discolor or some common ancestor. The hybridization occurring today

78 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. XXxIx

TABLE III *

VARIATION IN STAMEN NUMBER IN
P. discolor AND P. pallidus

Stamen number P. discolor P. pallidus
18 ¢) 1
17 6) 2
16 0 2
15 0 6
14 1 4
13 12 6
12 17 7
re 6 9
10 8 5

9 4 1
8 1 4
7 1 1
6 0 1

* The figures in the two right-hand columns refer to the number of flowers having
the specific stamen number. The 1 mean number of flowers scored per collection was
about 3, but there was some variation due to differences in the number of available

male flowers on the specimens; nee ac eee in sampling may be partially counter-
acted by the considerable spread of variation on each plant (e.g., from 9-14 in one
individual of P. discolor). It should be ere that the P. pallidus column includes col-
lections of various transitional forms, or in other words all specimens from the Cajal-
bana area which are not obviously ‘ ‘good” P. discoloi

must be due to comparatively recent changes in the distribution of one
or both species and, most probably, to a spread of P. discolor into the
restricted range of P. comptus. It is interesting that there is no evidence of
crossing between P. discolor and P. orbicularis of sect. Orbicularia, al-
though the latter is also present near almost every station of P. discolor and
shows an affinity (via P. microdictyus) evident enough so that crossing
would a priori not appear impossible. Presumably there is a much stronger
genetic barrier between P. discolor and P. orbicularis than between the
former and P. comptus; this offers an interesting subject for further anal-
ysis on cytological lines.

PLATE XXV. Fiowers or sect. Williamia, sussecrs. Discolores AND In-
crustatt,

Fics. A-C. Male flower, female flower, and female calyx-lobe of Phyllanthus
discolor Poepp. ex Spr. (Webster et al. 250 [GH]). Fics. D-F. Male flower,
gynoecium, and female calyx-lobe of Phyllanthus microdictyus Urb. (Marie-Vic-
torin & Clement 21525 [A]). Fics. G-I. Male flower, gynoecium, and female
calyx of Phyllanthus williamioides Griseb. (Webster 4014 |GH]). Fics. J-K.
Male and female flowers of here excisus Urb. (Ekman 4055 [S]). Fics.
L—-M. Male and female flowers of Phyllanthus incrustatus Urb. (Ekman 3848
[S]).

Jour. ARNOLD Arp. VoL. XXXIXK PLATE XXV

yi

‘

y,

‘ / y,
\\ \, /y

WEBSTER, WEST INDIAN PHYLLANTHUS

80 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

49. Phyllanthus microdictyus Urb. Symb. Ant. 9: 183. 1924.
(PLATE XXV, figs. D-F; PLATE XXVII, fig. A).

A shrub c. 1 m. high with the habit of a miniature tree, the stem simple
or sparsely branching near the top; bark smooth, dark brownish, becom-
ing finely fissured on old stems but not breaking up into plates. Cataphylls
deciduous: stipules triangular-lanceolate, mostly 3-4 mm. long and (1.1-)
1.5-1.7 mm. broad (smaller on weak shoots), acute, truncate at the thick-
ened base, lightly keeled, dark brown and scarious-indurate; blade nar-
rowly lanceolate, acuminate, c. 2.5-3 mm. long. Branchlets (6—) 9-17
(—19) cm. long, 0.9-1.5 mm. thick, brownish, smooth and pruinose, usually
flattened, with (6—) 8-12 (—14) leaves; first internode (15—) 20-40
(—45) mm. long, median internodes c. 10-20 mm. long. Leaves: stipules
linear-lanceolate, acuminate, (1.5—) 2.5-3 mm. long, 0.4-1 mm. broad,
scarious, deciduous. Petiole 2.5—-4.5 mm. long, subterete, brownish. Leaf-
blades flexibly chartaceous, ovate or elliptic, (1.5-) 2.5-5 (-6.5) cm.
long, (1.3—) 2—3.5 (—4) cm. broad, emarginate (or less commonly obtuse
or rounded) and minutely and inconspicuously apiculate at the tip, mostly
rounded to cordate at the base; above dull green (bright red when first
expanded), minutely foveolate, the midrib incised, the main laterals
raised, repeatedly branching and anastomosing to form a fine raised reticu-
lum with straight-sided meshes; beneath greyish-pruinose, the midrib
raised, the laterals (c. 4 or 5 on a side) and ultimate veinlets very slightly
raised, pale, forming a fine and inconspicuous reticulum; margins plane,
not thickened.

Monoecious; usually the proximal 2 or 3 nodes of a branchlet with male
cymules of 1—3 flowers, the succeeding nodes with bisexual cymules of 1
central female and 1 or 2 lateral male flowers.

Male flower: pedicel capillary, (S—) 7-15 mm. long. Calyx greenish
white, sometimes pinkish tinged: calyx-lobes 6 (rarely 5), chartaceous,
subequal, oblong to obovate, 2.5-3.5 mm. long, 1.3—2.3 mm. broad, entire
and rounded at the tip, faintly triplinerved or the midrib pinnately
branched. Disk-segments 6 (rarely 5), round or trigonous, smooth and
entire, c. 0.25—0.45 mm. broad. Stamens 6-10 (—11), the filaments 1—1.7
mm, long, connate into a rather stout column; anthers subsessile (the
free portion of the filament no longer than the anther) or the terminal
ones short-stalked, inserted on the column spirally or in 2 or 3 whorls,
ascending or spreading, c. 0.25-0.3 mm. long, 0.4—-0.5 mm. broad; anther-
sacs more or less divergent, globose before anthesis, dehiscing horizontally
or obliquely, the slits apically contiguous but not confluent: pollen grains
20-24 » in diameter, with c. 7 or 8 areoles per amb.

Female flower: pedicel capillary, (S—) 7-12 (-14) mm. long. Calyx
colored as in the male; calyx-lobes 6 (rarely 7), strongly convex, sub-
equal or unequal (the outer one or two lobes often manifestly shorter than
the innermost), oblong to obovate, the larger lobes 7—8.5 mm. long, 3.5—5
mm. broad, entire and obtuse or rounded at the apex, 3—5-nerved from
the base, the laterals ascending and anastomosing in a rather conspicuous
net. Disk 5S-angled, fairly massive in the bud but becoming fused with

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 81

and incorporated into the massive gynophore. Ovary oblate-spheroidal,
about as high as (or somewhat higher than) the gynophore, pruinose;
styles erect, (3.5—) 4-5 mm. high, connate 14 to 34 their length, grad-
ually dilated at their ends into conspicuously lacerate 3- or 4-parted tips.

Capsule oblate, emarginate at the apex, c. 5 mm. in diameter, smooth,
reddish brown, held within the erect appressed calyx-lobes. Columella
conical, somewhat constricted at the tip, 2.2-2.4 mm. long. Seeds [those
examined not quite mature] plano-convex, 2.5—2.7 mm. long, 1.5-1.6 mm.
radially and tangentially, dark shiny reddish brown, nearly smooth (min-
utely colliculose); hilum submedian, triangular, light brown.

Flowering and fruiting April through July and possibly later, but ap-
parently sterile during the winter.

Type: Cuba, Ekman 3705.

DIsTRIBUTION: lowlands near the coast, northeastern Oriente Province.
Cuba (Map XXII).

CUBA. Crienre: Moa region, Cayoguan, Acufa 12478 (SV | Herb. Roig n.
8593], US); Mina Cromita, Cayoguan, south of Punta Gorda, Clemente &
Alain 4074 (MICH); Rio Cayoguan, prés du pont sur le chemin de la mine
Delta, Marie-Victorin & Clement 21525 (A, MT), 21749 (MT); same locality,
cut-over forest of Clusia, Bactris, Arthrostylidium, et al., Webster 3809 (GH,
MICH); ad Taco Bay prope Baracoa in pinetis, 2 December 1914, Ekman 3705
(S, HOLOTYPE; NY, IsoTypEe fragment and photograph; sterile).

A very distinctive and well-marked species, P. microdictyus appears to
be quite restricted in range; the first four collections cited are all essen-
tially from the same locality, so that there are, in effect, only two known
stations. Although Ekman’s type specimen is sterile, the characteristic
venation of the leaves — identical with that of the fertile collections from
Moa — leaves no doubt as to the identity or typification of the species.
Ekman’s designation of the Taco Bay collection as being from pinelands
suggests a somewhat different habitat from the flood-plain woods along
the Rio Cayoguan, but the species would in any case appear to be a rela-
tively mesophytic one.

Because of their evolutionary significance, the several affinities of P.
microdictyus deserve special mention. Its large female flowers with long
styles and stipitate ovary are so similar to those of the South American
P. salviaefolius that there can be little doubt of a fairly close relationship,
although the latter species differs vegetatively by having hirsutulous axes
and acuminate, more coarsely veined leaves. However, P. microdictyus
is also an obviously near relation of P. discolor, which it greatly resembles
vegetatively. Furthermore, the incipient production of mesophyllar scle-
reids in P. microdictyus presages their full development in P. cristalensts :
and since it would appear that it is through the latter (or a similar species )
that the highly specialized representatives of subsect. Jncrustati and sect.
Orbicularia have originated, it would not be incorrect to regard P. micro-
dictyus as the progenitor of this entire phylogenetic line.

82 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

50. Phyllanthus cristalensis Urb. Repert. Sp. Nov. 28: 212-213. 1930.

Glabrous shrub up to 1.5 m. high; branches shiny reddish brown, be-
coming grey with age, c. 1.5-3 mm. thick. Cataphylls black, indurate,
soon deciduous: stipules lanceolate, c. 1.5-2 mm. long, acuminate, entire;
blade narrower. Branchlets mostly 4-11 cm. long, 0.5-0.8 mm. thick,
smooth, reddish brown, subterete (slightly compressed proximally), with
8-15 leaves; first internode c. 5-15 mm. long, median internodes 4—7 mm.
long. Leaves: stipules reflexed, subpersistent, triangular-lanceolate, 1.7—2
mm. long, 0.7-1.1 mm. broad, rather blunt at the tip, entire, dark brown,
somewhat indurate. Petiole 1.5-2.5 mm. long, slightly flattened, dark
brown, nearly smooth. Leaf-blades coriaceous, broadly elliptic to sub-
orbicular, c. 1-2.2 cm. long, 0.8-1.9 cm. broad, emarginate and with a
deciduous apiculum at the tip, rounded to emarginate at the base; above
dull brownish olivaceous, mottled (the cell outlines of the irregularly
anastomosing veins visible under a lens), the subprominent midrib plane
or sunken, the laterals rather obscure; beneath greyish or brownish, the
midrib and laterals (c. 4 or 5 on a side) plane, not very prominent, the
reticulum obscure; margins plane, scarcely thickened

Monoecious (presumably); flowers [and fruit not seen, description
ex Urban] few, solitary or paired, pedicels 3-5 mm. long.

Male flower: calyx-lobes 5, narrowly ovate, 2 mm. long, barely | mm.
broad; stamen number not determined; filaments connate, anthers dis-
crete. Female flower: pedicel narrowly obconic-thickened above; calyx-
lobes 5, narrowly ovate, 2 mm. long; styles not seen.

Capsule oblate, rounded-trigonous, c. 4 mm. in diameter; seeds brown-
ish, minutely puncticulate, c. 2 mm. long, 1.7 mm. broad.

Type: Cuba, Ekman 15993.

DISTRIBUTION: endemic to the Sierra Cristal, eastern Cuba (AIap
XXIT)

CUBA. OrIENTE, Sierra Cristal: at the tributary of the Rio Lebisa in char-
rascales, alt. 600-1100 m., 15 Dec. 1922, Ekman 15967 (S); in low Arthrosty-
lidium thickets which cover the top, alt. 1100-1325 m., 15 Dec. 1928, Ekman

3 (NY, isotype); ridge west of Pico Cristal, mossy elfin forest, 4 Mar.
1954, Jervis 3206 (GH).

Although still poorly known, P. cristalensis is evidently a very distinct
species on the basis of its vegetative characters. Vegetatively it is inter-
mediate between subsects. Discolores and Jncrustati, having the smooth
axes of the former combined with the highly sclerified leaves of the latter.
However, until adequate flowering material can be collected its relation-
ships cannot be satisfactorily determined.

Subsect. 16b. Incrustati, subsect. nov.

Phyllanthus sect. ai illiamiandra Griseb. Goett. Nachr. 1865: 171. 1865.
Ramsdenia Britton, Mem. Torr. Bot. Club 16(2): 72. 1920.

Stems and anes incrustate with dark bran-like flakes of bark:

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 83

leaves of branchlets alternate, coriaceous (heavily sclerified within)
stamens (2—) 3-6; styles erect, conspicuously lacerate.

Type species: Phyllanthus williamioides Griseb.

?

Grisebach proposed to distinguish sect. Williamiandra on the basis of
its androecium and styles, but his description of the androecium is errone-
ous and his discussion misleading. The type species, P. williamioides,
has only 5 or 6 stamens (not 10 as reported by Grisebach), and the con-
ical apex of the staminal column cannot serve as a sectional character since
it does not occur in the two related species. Grisebach’s stated distinction
between the styles of Williamiandra (correctly likened to those of sect.
Oxalistylis) and those of sect. Williamia is wholly incorrect, for he errone-
ously describes the styles of the latter as bifid. Except for the difference
in stamen number, there are no floral characters available for separating
Williamia from Williamiandra; but since the species of the two groups
can be distinguished vegetatively, it seems appropriate to place them into
two subsections of the inclusive sect. Williamia.

Britton’s proposed genus Ramsdenia, which included P. excisus and
P. incrustatus, corresponds almost exactly to the present subsection; he
did not discuss its relationships or clearly define its characters, and he was
apparently unaware of the existence of P. williamioides.

51. Phyllanthus williamioides Griseb. Goett. Nachr. 1865: 169-170.
1865; Muell, Arg. in DC. Prodr. 15(2): 328-329. 1866.
(PLATE XXV, figs. G-I).

Diasperus williamioides (Griseb.) O. Ktze. Rev. Gen. 2: 601. 1891.

A small shrub 0.5—1.5 (—2) m. high, with two or more stems clustered
on a gnarled caudex (stem unbranched in young plants); bark of stem-
base and root broken up into conspicuous patterns of squares. Stems and
branchlets incrustate but the bark-platelets smooth and glabrous, dark
brown, the exposed spongy tissue reddish brown. Cataphylls blackish,
indurate but mostly soon deciduous: stipules lanceolate, (1.5—) 1.8—2.7 mm.
long, 0.7-1.2 mm. broad; blade c. 1.8-2.5 mm. long, 0.5 mm. broad.
Branchlets erect or spreading, (5—) 8-15 (—19) cm. long, 1-1.2 mm. thick,
reddish-brown- or blackish-incrustate, terete or angled, with mostly 10-25
(—35) leaves; first internode (S—) 7-12 (—16) mm. long, median internodes
2-6 (-10) mm. long. Leaves: stipules acicular-lanceolate (convexly con-
duplicate), soon deciduous, 2.5—3.5 (-5) mm. long and 0.5—1 mm. broad,
rather blunt at the tip, chartaceous, dark brown. Petiole 1.5—2.5 mm.
long, brownish, smooth, flattened. Leaf-blades coriaceous, mostly orbicular
or nearly so (sometimes slightly obovate or broader than long), (1—) 1.5—
2.5 cm. long, (1—) 1.3-2.2 cm. broad, retuse or emarginate at the tip, cor-
date at the base; above olivaceous and sublucid or dully plumbeous, min-
utely foveolate-reticulate, the midrib slightly incised, the lateral veins and
veinlets forming a fine slightly raised reticulum; beneath brownish- or
olivaceous-plumbeous (minutely white-dotted with wax-covered stomata),
the midrib plane, the slightly raised laterals (c. 5 or 6 on a side) anas-

84 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

tomosing with the conspicuous tertiary veinlets to form a close reticulum;
margins plane, not thickened.

Monoecious, flowers in axillary bisexual cymules; central flower female,
lateral ones male (the male mostly developing after fruiting calyx has
fallen, so that the arrangement on the branchlet appears to be of proximal
racemes of male flowers and distal solitary female flowers); bracteoles
blackish and indurate.

Male flower: pedicel c. 1.8-2 mm. long. Calyx greenish white, whitish,
or reddish tinged; calyx-lobes 5, thin but coriaceous, subequal, oblong to
spathulate, (1.3- ) 1.5-2.3 mm. long, 0.7-1.5 mm. broad, denticulate near
the apex, more or less triplinerved but nerves rather obscure. Disk-seg-
ments 5, flattened or concave, roundish or triangular, smooth and entire,
0.25-0.35 mm. broad. Stamens 5 or 6, the filaments connate into a rela-
tively slender column 1-1.1 mm. high and 0.25-0.4 mm. thick; anthers
in two series, the upper of 3, the lower of 2 or 3, the apex of the column
terminated by an apiculum; anthers slightly apiculate, sessile on the
column or very nearly so, the upper erect, the lower ascending or spreading,
c, 0.3-0.4 mm. long and broad; anther-sacs sub-parallel, dehiscing longi-
tudinally, the slits not confluent; pollen grains 24-27 » in diameter, the
areoles polybrochate, c. 5—6 » across, c. 10 per amb,

Female flower: pedicel short and stout, only (0.7—) 1.2-1.8 (—2.4) mm.
long. Calyx colored as in the male; calyx-lobes 5-7, quite unequal, den-
ticulate at the apex, incrassate at the base, nerves several but obscure,
outer lobes oblong, obtuse, 2.5—3.5 mm. long, 1.5—2 mm. broad; inner lobes
obovate, rounded at the tip, 5-5.5 mm. long, 2.2-2.8 mm. broad. Disk
a thick brownish crenate-rimmed cup, enclosing the base of the ovary.
Ovary sessile, smooth, carinate between the septae; styles erect, free, 1.8-
2.7 mm. high, the dilated tips more or less 3-fid, the three branches with a
total of c. 8-10 slender lacerae.

Capsule subglobose, c. 4 mm. in diameter, the valves olivaceous, smooth,
not veiny, 3.5 mm. long, often retained within the appressed calyx-lobes.
Columella 2.2 mm. long. Seeds trigonous, somewhat asymmetric (excen-
trically carinate on the back), 2.2-2.3 mm. long, 1.2-1.5 mm., radially
and tangentially, blackish brown, colliculose on back and sides; hilum
subterminal, elliptic or ovate, c. 0.5 mm. long.

Type: Cuba, Wright 1944.

DISTRIBUTION: endemic to the Baracoa region, Oriente province, Cuba
(Map XXII).

CUBA. OrtenTtE: Cuchillas de eee Wright 1944 (GOET, HOLOTYPE;
GH, 1sotypeE; locality ex isotype); Altos de Farola, c. 20 miles south of Baracoa,
pineland at 1750 ft., 3 Aug. 1951, cee 4014 (GH, MICH, NY); among pines
and tree ferns, moist arroyo c. 12 miles east of Baracoa, 18 Mar. 1954, Jervis
3332 (GH).

Wright’s type collection differs from the other specimens in its distinctly
obovate leaves. However, in the field, leaves of this shape were found on
shoots from the base of plants which higher up bore orbicular leaves; it

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 85

thus appears that Wright made his collection from a stump sprout or very
depauperate plant.

The present species is very closely related to P. incrustatus, but is usually
distinguishable at first sight by virtue of its smooth bark platelets (these
usually scabridulous in P. incrustatus) and larger more elongated leaves.
Actually, the concave deciduous stipules and androecium of 5—6 stamens
are the best diagnostic features of P. williamioides, and will always permit
its ready separation from P. incrustatus.

As in many species of subg. Xyvlophylla, the leaves of P. williamioides
are distinctly reddish-tinged when young, though less strikingly so than in
P. pachystylus. The undersurface is pale yellow at first and only later
becomes green; this behavior, again, is characteristic of many other species.

/ © PEXCISUS
S @ P INCRUSTATUS
EG Ps * P MIRIFICUS

/ Te @ P. WILLIAMIOIDES
[goer aaah V peer

i WI

/

i, i

ASS h / .

(ha ea Sa Ee TS 3
76 73

7 7

Map XXIII. Distribution of sect. Williamia subsects. Incrustati and Mirifici
in eastern Cuba.

52. Phyllanthus excisus Urb. Repert. Sp. Nov. 13: 449-450. 1914.
(PLATE XXV\V, figs. J-K).

Ramsdenia excisa (Urb.) Britton, Mem. Torr. Bot. Club 16(2): 72. 1920.

Shrub c. 2-3 m. high; pith brownish; stems and branchlets incrustate
with smooth dark reddish brown platelets, the interspaces of spongy tissue
light reddish brown. Cataphylls blackened, indurate, more or less per-
sistent: stipules triangular-lanceolate, 2—-2.5 mm. long, 1.5—1.6 mm. broad,
acuminate; blade narrower. Branchlets stout, 9-27 cm. long, 1.7-3 mm.
thick, reddish brown and incrustate, subterete (more or less angled dis-
tally), with (10—) 15-35 nodes; first internode (5—) 10-15 (—20) mm.
long, median internodes 5-18 mm. long. Leaves: stipules lanceolate,
2.5-3 mm. long, 1.1-1.5 mm. broad, acuminate, quite oblique at the base,
blackish, indurate, appressed or spreading, persistent. Petioles 1.5-3 mm.
long, brownish, transversely furrowed, with a deep median adaxial groove.

86 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

Leaf blades coriaceous, mostly ovate or elliptic, (2.5—) 3-4.5 cm. long,
2.5-3.5 cm. broad, emarginate at the tip, obtuse to subcordate at the
base; above olivaceous-plumbeous, minutely foveolate-reticulate, the
midrib plane or slightly impressed, the lateral veins obscure; beneath
brownish grey when dried (not white-dotted), the midrib saliently raised,
the laterals (c. 6-8 on a side) and tertiary veinlets slightly raised, anas-
tomosing to form a rather prominent reticulum; margins unthickened,
plane.

Monoecious; flowers in axillary bisexual cymules, each of a solitary
central female flower and a few (c. 3 or 4) male flowers on bracteolate
lateral axes; bracteoles indurate, persistent in old axils.

Male flower: pedicel capillary, c. 4-5.5 mm. long. Calyx whitish (ex
Shafer); calyx-lobes 5, subequal, oblong to obovate, (3—) 3.3-4 mm.
long, 1.5—-2 mm. broad, entire, the venation obscure. Disk-segments 5S,
roundish, entire, c. 0.4—0.5 mm. across. Stamens 5—6, filaments united into
a rather stout non-apiculate column 1.5—-1.8 mm. high and 0.5—0.6 mm.
broad; anthers biseriate (upper series of 3, lower of 2 or 3) but both
whorls crowded at the top, not evidently apiculate; anther-sacs subparallel,
dehiscing longitudinally, the slits not confluent; pollen grains 19-22 yu in
diameter, the areoles polybrochate, c. 7-11 » long, 4—5 per amb.

Female flower: pedicel nearly straight, slender, (10-) 12-15 mm. long.
Calyx greenish-white (ex Shafer); calyx-lobes 5, coriaceous, thickened at
the base, subequal (the outer slightly narrower), oblong to obovate, the
larger lobes 5—7 mm. long [up to 8 mm., ex Urb.], 2.5-4.5 mm. broad,
entire and rounded at the tip, the venation obscure. Disk massive, ob-
tusely 5-angled, entire (or minutely crenulate), obscurely foveolate, the
margin plane, somewhat thinner. Ovary sessile, carinate dorsally; styles
erect, free or shortly connate at the base, 1.8-2 mm. high, the dilated ends
deeply parted into 6-7 (—10) narrow divaricate lacerae; outline of stylar
apex roughly triangular, two arms extending adaxially, the third abaxial
and with more lobes.

Capsule subglobose, emarginate, c. 5 mm. in diameter, smooth or slightly
rugulose, brownish, not veiny. Columella c. 3 mm. long. Seeds |seen only
in immature condition] c. 3 mm. long, brownish, probably with the orna-
mentation of P. williamiotdes when mature.

Typr: Cuba, Shafer 4447.
DiIstTRIBUTION: endemic to the Baracoa region, Oriente province, Cuba

(Map XXIII).

CUBA. OrtENTE: Navas to Camp Buena Vista, alt. 650 m., 23 March 1910,
Shafer 4447 (NY, Lecrotype; F, MO, US, Isotypes); prope Baracoa, ad
Maravi in pinetis, 26 Dec. 1914, Ekman 4055 (S).

Apparently a rare species of restricted range, P. excisus is well distin-
guished from the two related species of subsect. /ncrustati by its larger
ovate leaves, larger flowers on longer pedicels, and different styles. Its

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 87

strong resemblance to P. microdictyus of subsect. Discolores can hardly
be mere coincidence, and it can be rather confidently regarded as a
xerophytic derivative of that plant. The flowers of both species are similar,
but P. excisus has fewer stamens and, of course, its edie stems and
coriaceous leaves easily separate it om ee microdictyu

The apparent rarity of P. excisus, as compared ae ‘the other Oriente
species of sect. Williamia, may prove to be illusory, for the region between
Navas and Baracoa has been much less visited by collectors than the Moa
district, which up until recently was much more accessible.

53. Phyllanthus incrustatus Urb. Repert. Sp. Nov. 13: 449. 1914.
(PLATE XXV, figs. L-M).

Ramsdenia incrustata (Urb.) Britton, Mem. Torr. Bot. Club 16(2): 73. 1920,

Shrub 0.5—2 m. high; stems and branchlets usually scabridulous-hirtel-
lous at the tips, below incrustate with scabridulous platelets of bark (rarely
the platelets smooth). Cataphylls blackish, indurate, subpersistent: stip-
ules triangular to narrowly lanceolate, (1- ) 1.5-3.7 mm. long, 0.5-1 mm.
broad, acute to acuminate (sometimes with very attenuate tips), sometimes
Seite keeled, becoming reflexed or spreading; blade narrowly
lanceolate, (0.7—) fess (-—5) mm. long, also more or less reflexed.
Branchlets (5—) 7-17 (—26) cm. long, 1-1.7 mm. thick, incrustate with
dark brown usually scabridulous platelets separated by lighter reddish
brown furrows, with (10—) 15-40 (-—60) leaves; first internode 3-10 mm.
long, median internodes 2-6 mm. long. Leaves: stipules persistent, ap-
pressed or reflexed, triangular to linear-lanceolate, (0.8—) 1.2—2.5 (-3.2)
mm. long, 0.3-0.9 mm. broad, acute or acuminate, dark brown, more or
less indurate. Petiole 1-1.8 mm. long. Leaf blades coriaceous, mostly or-
bicular or suborbicular (sometimes broader than long), c. 7-12 (-15) mm.
long, 6-12 (—14) mm. broad, usually retuse or emarginate and incon-
spicuously apiculate, truncate or more commonly cordate at the base;
above dull olivaceous or plumbeous, minutely foveolate-reticulate, the
midrib impressed, the laterals usually rather obscure; beneath greyish
brown or purplish brown (minutely white-dotted with wax-covered sto-
mata), the slightly raised laterals (c. 5 or 6 on a side) anastomosing with
the veinlets to form an inconspicuous reticulum; margins plane, not
thickened

Monoecious; flowers in axillary bisexual cymules; first (central) flower
of cymule female, rapidly maturing and deciduous, succeeded by several
male flowers on one or both lateral abbreviated axes; branchlets thus
appearing (as in P. williamioides) to have short racemes of male flowers
at proximal nodes and solitary female flowers at distal nodes.

Male flower: pedicel (0.5—) 1-1.5 mm. long. Calyx whitish (in living
condition [ex Shafer], reddish when dried); calyx-lobes 5, subcoriaceous
but thin, subequal, oblong to obovate or spathulate, (1.3—) 1.5—2.4 mm.
long, (0.7) 1-1.5 mm. broad, rounded at the tip, entire or occasionally
denticulate, nervation obscure, only the midrib at all prominent. Disk-

88 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XXxIx

segments 5, flattened or concave, roundish or trigonous, c. 0.25-0.35 mm.
across. Stamens 3 or 4 (rarely 2), filaments connate into a slender non-
apiculate column 0.6-1.2 mm. high and 0.15-0.25 mm. thick; anthers
sessile, aggregated into a mass at the top of the column, 0.25-0.35 mm.
long; anthers-sacs parallel, dehiscing longitudinally and vertically, the
slits not confluent; pollen grains 12-19 » in diameter, the areoles poly-
brochate, c. 6—7.5 » long, c. 3-5 per amb.

Female flower: pedicel slender or incrassate above the middle, 1-3 mm.
long. Calyx colored more or less as in the male; calyx-lobes 5, rather
unequal, rounded at the tip, entire, the outer oblong, 1.6-2.7 mm. long
and 0.8—-2.3 mm. broad, the inner mostly spathulate or obovate, 1.8-3.3
mm. long and 1.2-3.3 mm. broad; nervation obscure. Disk massive, 5-
angled, entire, plane, foveolate. Ovary sessile; styles free or shortly con-
nate below, spreading or ascending, flattened, 0.5-0.6 mm. long, the dilated
apices 5—6-lacerate.

Capsule subglobose, emarginate, c. 3-4.5 mm. in diameter, the valves
smooth, dark brown, not veiny. Columella 1.5—2 (—2.5) mm. long. Seeds
trigonous, asymmetrically carinate on the back, 1.8-2.7 mm. long, 1.1—
1.8 mm. radially, 1.35-1.45 mm. tangentially, reddish brown, nearly
smooth.

Type: Cuba, Shafer 4020 (NY, LectoryPeE; the original specimen of
this collection in Herb. Krug & Urban destroyed during World War IT).

DISTRIBUTION: wooded areas, northeastern Oriente province, Cuba
(Map XXIII).

CUBA. OrtiENTE: Cumbre Cayo [Cayoguan], Acuna 12477 (US); Mina
Delta, c. 500 m. alt., échine de serpentine sur le chemin de la mine Delta, Warie-
Victorin & Clément 21754 (A, MT); Camp La Gloria, south of Sierra Moa,
Shafer 8113 (F, MO, NY, US); damp thickets between Rio Yamaniguey and
Camp Toa, alt. 400 m., 22-26 Feb. 1910, Shafer 4020 (NY, LECTOTYPE); moist
woods, Navas to Camp Buena Vista, alt. 650 m., Shafer 4453 (NY, US); Minas
de Iberia ad Taco Bay, alt. c. 800 m., Ekman 3833, 3848 (S).

Evidently the commonest and most widespread of the three species of
the subsection, P. incrustatus is morphologically the most specialized, at
least with respect to its smaller leaves and androecium of fewer stamens.
Although occasional large-leaved forms, such as Ekman 3848, simulate
P. williamioides, the ensemble of characters of P. incrustatus is distinctive.
In addition to its smaller leaves and fewer stamens, it also differs from
P. williamioides in its persistent stipules and somewhat larger number of
leaves per branchlet.

Although its small, roundish, coriaceous leaves suggest an affinity of
P. incrustatus with the species of sect. Orbicularia, it does not appear likely
that P. incrustatus is the progenitor of that group. Rather, it seems that
the species of subsect. /ncrustati represent a more or less parallel line of
development with the representatives of Ordicularia.

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 89

Subsect. 16c. Mirifici, subsect. nov.?!

Stems and branchlets smooth but with conspicuous lenticels; leaves of
branchlets opposite, coriaceous (heavily sclerified within); stamens 5;
styles obsolete, the greatly dilated stigmas (style-tips) forming a cap
which covers most of the ovary.

TYPE SPECIES: Phyllanthus mirificus Webster.

The single very distinctive species of this monotypic taxon, only re-
cently discovered in eastern Cuba, is unique among the New World
species of Phyllanthus in its opposite leaves (on the branchlets) and
hypertrophied stigmas. Although its leaf arrangement and androecium
suggest certain Old World taxa, its areolate pollen grains, leaf structure,
and aspect provide an incontestable basis for placing it in sect. Williamia.
Its affinity with sect. Jncrustati is evident from even superficial compari-
son, but its many distinctive characters amply justify the erection of a
cde subsect. Mirifici.

54. Phyllanthus mirificus Webster, Contr. Gray Herb. 176: 58. 1955.
(PLATE XXVI, figs. A-C).

Definitely woody, presumably a shrub; branches slender, terete, not
over 3 or 4 mm. thick (toward the ends), smooth, greyish- or reddish-
brown, with conspicuous elliptic to linear lenticels. Cataphylls blackish
and indurate but soon deciduous: stipules triangular-lanceolate, 2.3-3 mm.
long, 1-1.5 mm. broad, acuminate, entire, carinate on the back; blade
linear-lanceolate, c. 2 mm. long. Deciduous branchlets 8—13.5 cm. long,
1.5—2 mm. thick (those on lateral axes only 1 mm. thick), terete, smooth,
brownish, becoming lenticellate proximally or throughout, with c. 6-8
(—11) pairs of leaves; first internode 6-20 mm. long, median internodes
3—20 mm. long. Leaves of branchlets all opposite (one node very slightly
higher than its neighbor): stipules blackish and indurate but mostly
soon deciduous, triangular-lanceolate, 2.2-3 mm. long, 0.6—-0.9 mm. broad,
acuminate, entire. Petioles 1.8-3 mm. long, brownish, plane and grooved
down the middle adaxially, convex and corrugate-rugulose abaxially. Leaf-
blades rigidly coriaceous, elliptic or ovate to suborbicular, 1.5—2.8 cm.
long, 1.3-2.1 cm. broad, emarginate at the apex (the small short-conical
dark brown apiculum of the juvenile leaf represented by a scar, or appar-
ently absent), obtuse or rounded to subcordate at the base; above oliva-
ceous or brownish, the midrib conspicuous, pale, plane or slightly sunken,
the laterals obscured by the close-meshed slightly raised reticulum of
tertiary veinlets; beneath more or less greyish-pruinose, the midrib nearly
plane, the laterals (c. 5 or 6 on a side) and veinlets anastomosing in a sub-
prominent reticulum; margins sharply differentiated, thin and acute, plane
or reflexed (but not revolute).

*! Subsect. Mirifici, subsect. nov. Ramis ramulisve lenticellatis; foliis ramulorum
oppositis, coriaceis ; staminibus 5; stigmatibus dilatatis, calyptratis. — Species typica
Phyllanthus ean Webste

90 JOURNAL OF THE ARNOLD ARBORETUM _[VvoL. xxxIx

Monoecious; flowers in axillary bisexual cymules, each with a solitary
female and mostly 2 or 3 male flowers; bracteoles ovate, blackish and
indurate, persistent.

Male flower: pedicel slender, c. 1—-1.2 mm. long. bck ee _
dried; calyx-lobes 5, chartaceous, obovate, c. 1.5 mm. long, 0.9-1.3
broad, rounded at the tip, entire, the mide eee ae
Disk-segments 5, ellipsoid, somewhat thickened, entire, c. 0.25—0.3 mm.
across. Stamens 5; filaments connate into a column c. 0.5—0.8 mm. high,
unequal in length (two anthers inserted lower than the other three) ;
anthers subsessile, ascending or vertical, blunt, 0.3-0.4 mm. long, 0.4—
0.45 mm. broad; anther-sacs divergent, dehiscing vertically, the slits not
confluent; pollen grains 16-18 » in diameter, areoles polybrochate, c. 6 p
across, c. 5—7 per amb.

Female flower: pedicel (at anthesis) 1.5-2.5 mm. long, subterete, en-
larged and massive above. Calyx reddish when dried; calyx-lobes 6, sub-
equal, biseriate, the outer broadly ovate, the inner broadly obovate,
c. 1.5-1.8 mm. long, 1.1-1.4 mm. broad, entire, the veins obscure. Disk
rather massive, bluntly angled, plane, the margins entire. Ovary sub-
globose, sessile, almost completely covered by the lower margins of the
three dilated stigmas (style-ends) which are reflexed and appressed to
form a close-fitting calyptra; stigmas with auricles connivent into a
blunt beak at the top, the lateral margins entire, the distal margins crenu-
late-notched, 0.9-1.1 mm. long, 0.9-1 mm. broad across the distal edge.

Fruit and seeds unknown.

Type; Cuba, Leon e¢ at, 22613.

DISTRIBUTION: scrublands and pinelands, northeastern Oriente province,
Cuba (Map XXIII).

CUBA. ORIENTE: Charrascal del Coco, south of Moa, July 1945, Ledn,
Clemente, & Alain 22613 (MICH, HoLotypE; LS, IsotypE); charrascos y
pinares, Sierra de Moa, alt. c. 750 m., 25 July 1953, Alain 3380 (GH); pineland
barrens, Charrascos de Pena Prieta, Toa, alt. 600 m., 30 Dec. 1953, Alain 3616
(GH; sterile).

This extraordinary species, which remained undetected up until recent
times, now appears to be as widely distributed in the Sagua-Baracoa range
as some of the other specialized members of sect. Williamia. Its opposite
branchlet-leaves and calyptriform stigmas at once set it apart from all
of its West Indian congeners, and its androecium of 5 stamens in two
unequal sets appears very similar to that in certain Old World species of
subg. Kirganelia. Nevertheless, its phyllanthoid branching and areolate
pollen grains show that it belongs in subg. XylopAylla, and both foliar
and floral characters attest its affinity with the species of sect. Williamia.
However, no one species of the section appears particularly close to
P. mirificus, although P. excisus has a similar aspect (the type collection
was, in fact, determined as that species), while P. williamioides has similar
androecia and short-pedicellate flowers. In some ways, P. miurificus is

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 91

intermediate between subsects. Discolores and Incrustati; this is especially
true of its development of copious lenticels, which presents a condition
transitional between the smooth stems of subsect. Discolores and the scurfy
axes of subsect. /ncrustati. Because of this intermediary position and of
its very divergent features, P. mirificus seems best assigned to a special
subsection of its own.

Sect. 17. Thamnocharis Webster, Contr. Gray Herb. 176: 59. 1955.

Shrubs with phyllanthoid branching; cataphylls large, indurate; leaves
coriaceous, stipules caducous. Monoecious; cymules bisexual, the flowers
appearing with the expanding leaves. Male flower: calyx-lobes 4—6; disk-
segments 4-6; stamens 2-6 (—8), filaments united (apparently free in P.
comptus), anthers dehiscing vertically; pollen grains areolate. Female
flower: calyx-lobes 4—6, coriaceous; disk entire, angled; styles free or
connate, bifid, the style-branches narrowed to acute tips. Capsule sub-
globose, dry, not veiny; seeds smooth or rugulose.

Type spEcIES: Phyllanthus cinctus Urb.

The relationships and nomenclature of this Cuban section and its three
constituent species were discussed at some length in the original place of
publication (op. cit. pp. 59-62). However, subsequent investigation has
shown (on the basis of the presumed hybridization between P. discolor
and P. comptus) that there must be a fairly close affinity between sects.
Williamia and Thamnocharis; consequently, the position of the latter in
the linear arrangement of sections has been altered. Furthermore, it is
now clear that sect. Tkamnocharis need not be compared with the Asiatic
sect. Eriococcodes, for the latter definitely belongs in subg. Eriococcus,
and its floral similarities are simply due to parallel evolution. Despite the
dissimilarity (in gross appearance and in floral details) between P. cinctus
and P. discolor, a significant affinity between subsect. Discolores (of sect.
Williamia) and sect. Thamnocharis appears highly probable.

Before male flowers of P. comptus and before its hybridization with
P. discolor were known, sect. Thamnocharis was thought to be related to
sect. Epistylium. The Jamaican species of that section resemble the Cuban
species of sect. TAamnocharis in their palm-like habit, reduced androecium,
and similarly veined leaves. It may yet be possible to show that sect.
Epistylium is closely related to the Cuban plants, but since its representa-
tives (P. cauliflorus and P. cladanthus) differ in having persistent stipules,
horizontally dehiscing anthers, and dilated lacerate styles, they are placed
at some distance in the present linear arrangement. Even more similar to
sect. Thamnocharis, at least superficially, is the monotypic sect. Glypto-
thamnus which closely mimics P. cinctus; but that group appears to be
much more closely allied to sect. Epistylium in its technical characters.
These rather puzzling suggestions of reticulate relationships indicate that
the phylogeny of the groups in question can by no means be considered
finally settled.

Jour. ARNOLD Ars. VoL. XX XIX PLATE XXVI

WEBSTER, WEST INDIAN PHYLLANTHUS

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 93

KEY TO THE SPECIES

1. Calyx-lobes and stamens 5 or 6; styles free; pedicel of female flower 10-
14 mm. long; seeds vermiculately marked, 4.5 mm. long or more; cataphyllary
stipules ribless, less than 1 cm. long. .................... 5S. P. comptus
Calyx-lobes 4; stamens 2; styles united into a column; pedicel of female
flower not over 6 mm. long; seeds smooth, not over 4 mm. long; cataphyllary
stipules longitudinally ribbed, 1 cm. long or more.

—

2. Male calyx-lobes c. 2 mm. long, with mostly 1-3 nerves from the base;
anthers triangular, 0.5-0.8 mm. long; female calyx-lobes reflexed at the
tips, the inner (longer) c. 3-5 (-6) mm. long; female disk tenuous, rim-
laleeteceee, Till ACTOSS cee oes oa okt eee ah ne eee eed 56. P. cinctus

2. Male calyx-lobes c. 3-3.5 mm. long, with mostly 5-7 nerves from the

base; anthers lanceolate, 0.9-1.2 mm. long; female calyx-lobes not reflexed
at the tips, the inner (longer) becoming 6-8 mm. long; female disk very
massive, forming a mound 3-4 mm. broad and 0.5—-1 mm. high, its fove-
OlaLeEMM eG eDLesSeGw cds, i54 Bitrate dnc hate ee tele 57. P. ekmanii

55. Phyllanthus comptus Webster, Contr. Gray Herb. 176: 61. 1955.

Glabrous shrub; branches terete, smooth, pale brown becoming greyish,
c. 3-4 mm. thick. Cataphylls massive and coriaceous but deciduous: stip-
ules triangular-lanceolate, acuminate, c. 6—-7.5 mm. long, 2-3 mm. broad,
obliquely truncate at the base, brownish with blackened indurate tips;
blade narrower. Deciduous branchlets (5.5—) 7-14 cm. long, 1-1.5 mm.
thick, stramineous or pale reddish brown, smooth, subterete (somewhat
flattened proximally), with only 5—7 nodes; first internode (15—) 20-30
mm. long, median internodes c. 15-30 mm. long. Leaves: stipules cadu-
cous, ovate-lanceolate, 5—6 mm. long, 2.7-3 mm. broad, acuminate, more
or less denticulate along the margin, brownish, scarious-chartaceous with
more or less indurate darkish tips. Petioles plane adaxially, convex
abaxially, smooth, stramineous or reddish brown, 2.5—4 mm. long. Leaf-
blade coriaceous, ovate or elliptic, c. (2.5—) 3-5.5 cm. long, (1.2-) 1.5-3
cm. broad, obtuse and minutely apiculate at the tip, obtuse or rounded at
the base: above sublucid, olivaceous, the midrib and chief laterals slightly
raised; beneath paler, the midrib and main laterals (4 or 5 on a side)
slightly raised, forming a subprominent reticulum, veinlets obscure; mar-
gins cartilaginous-thickened, light brownish, more or less reflexed.

Monoecious: flowers appearing on new branchlets with the expanding
leaves, the male early deciduous; cymules usually bisexual, of 1 female
and several (c. 5-8) male flowers.

PLATE XXVI. Fiowers oF sect. Williamia, sussect. Mirifici, AND OF SECT.

Thamnocharis.

Fics. A-C. Male flower, female calyx, and gynoecium of Phyllanthus mirificus
Webster (Alain 3380 [GH]). Fics. D-F. Male flower, female flower, and
gynoecium of Phyllanthus cinctus Urb. (Jervis 3355 [GH]). Fics. G-J. Male
flower, androecium, female calyx-lobe, and gynoecium of Phyllanthus ekmaniu
Webster (Jervis 3037 [GH]).

94 JOURNAL OF THE ARNOLD ARBORETUM _[vo.. xxx1x

Male flower (described from immature buds): pedicel up to c. 2.5 or
3 mm. long. Calyx-lobes 6 (rarely 5), greenish, nervation obscure: disk-
segments 6 (rarely 5), flattened, foveolate; stamens (4—) 5 or 6 (-8),
filaments shorter than anthers and apparently not united (but staminal
column perhaps not yet developed); anthers reniform, c. 0.4—-0.5 mm.
broad; anther-sacs divergent, dehiscing more or less vertically, the slits
apically confluent.

Female flower: pedicel slender, subterete below, gradually dilated and
obtusely angled above the middle, 10-14 mm. long. Calyx-lobes 6 (rarely
5), unequal, spathulate, rounded and entire at the tip, in flower 2.5-4 mm.
long and 1.2-1.7 mm. broad, in fruit 3.3-4.2 mm. long, c. 1.7 mm. broad,
the nerves obscure. Disk flat, 6-angled, becoming dark brown, not con-
spicuous. Ovary reddish brown, sulcate, emarginate at the top; styles free
except at the very base, ascending, c. 1.5 mm. high, bifid; style-branches
divergent, recurving, terete, narrowed to the tips.

Capsule-valves c. 7—-7.5 mm. long, reddish brown, the veins obscure.
Columella 3.8—4 mm. long. Seeds trigonous, asymmetric, carinate or some-
what irregular on the back, (4.5—) 4.8-5.1 mm. long, (2.5—) 3 mm. tan-
gentially (across the back), dark brown (to the naked eve), with a
vermiculate pattern of brownish-black raised lines on a light brown back-
ground; hilum submedian, c. 0.4—-0.5 mm. across, the raphe conspicuous.

Type: Cuba, Acuna 18222.

DISTRIBUTION: endemic to the Cajalbana region, Pinar del Rio province,
Cuba (Map XXIV).

—

@® PCINCTUS
© PCOMPTUS
* PEKMANII

|e

Map NNIV. Distribution of sect. Thamnocharis in Cuba.

CUBA. Pinar bet Rio: en arroyos, La Cajalbana, La Mulata, 28 Sept. 1952,
Acuna 18222 (SV, HOLOTYPE; fruiting); banks of a rivulet, cuabales, east of
Loma Cajalbana, La Palma, Oct. 1952, Acuna (LS, MICH); La Cajalbana,
junto a cahada, Camino de Sagua, 12 Mar. 1954, Acuna & Schubert 19120
(GH, PARATYPE; male and female flowers).

This recently described species, which would still remain unknown were
it not for the efforts of Ing. Julian Acuna, needs further study. Good
material of the male flowers is sparse or lacking in most specimens of this
and the other two species of sect. T/amnocharis, because the flowers appear

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 95

with the developing leaves and the male ones are soon deciduous. The
androecium of P. comptus is of especial interest because of the apparently
free filaments and the variable number of stamens. A count of 25 flowers
yielded the following distribution (the first figure being the stamen num-
ber): 4— 3, 5— 9, 6— 11, 7—1, 8—1. This range of variation nicely
bridges the gap between the androecium of P. discolor, with 9-14 stamens,
and the other two species of sect. Thamnocharis, with only 2. The pre-
sumption that P. comptus represents a phylogenetically, important con-
necting link is thus strengthened.

Although there appears to be no serious doubt regarding the kinship
of P. comptus with P. cinctus and P. ekmanu, the differences between the
Pinar del Rio and Oriente plants are so profound that their separation
must be ancient; they could be placed in different subsections if there were
any point in subdividing so small a section as Thamnocharis. The recogni-
tion of the hybridization between P. comptus and P. discolor (discussed
in detail under the latter) has further complicated the picture. Although
these two species obviously retain some degree of genetic compatibility,
the cataphylls, leaves, styles, and seeds of P. comptus are so different from
those of P. discolor that each species must represent the end-product of
a long-separate evolutionary line. In common with several other species
endemic to the Cajalbana area, P. comptus bears the earmarks of a relict
species the nearest relatives of which have long since disappeared.

56. Phyllanthus cinctus Urb. Symb. Ant. 9: 191-192. 1924; emend.
Webster, Contr. Gray Herb. 176: 60. 1955.
(PLATE XXVI, figs. D-F).

Conami (?) ovalifolia Britton, Mem. Torr. Bot. Club 16: 73-74. 1920; non
Phyllanthus ovalifolius Forsk., 1775.
Phyllanthus brittonit Alain, Contr. Mus. La Salle 11: 1. 1952.

Subshrub or shrub 0.15—-1 m. high, with the habit of a miniature tree,
the primary stem unbranched, greyish, furrowed, c. 4-8 mm. thick. Lower
leaves (missing on many specimens) with petioles c. 5-7 mm. long, leaf-
blades obovate or spathulate, 4-6.5 cm. long; upper leaves reduced to
cataphylls: stipules ek Sa -lanceolate (somewhat falcate), acuminate,
10-17 mm. long, 4-5 mm. broad (on vigorous shoots; sometimes as small
as 7 mm. long and 3.5 mm. ee on weak axes), obliquely truncate at the
base, conspicuously longitudinally corrugate-ribbed, dull reddish brown,
somewhat blackish and glandular at the base, sometimes densely hirtello-
scabridulous at the base or throughout, scarious-indurate; blade acicular,
7.5-9 mm. long. Deciduous branchlets erect to spreading, (10—) 15-23 cm.
long, 2.5-3 mm. broad, olivaceous, smooth or pustulate-scabridulous, dis-
tinctly flattened, with c. (4-) 8-15 leaves; first internode (15—) 20-50
(—65) mm. long, median internodes (10-) 15-2 (-40) mm. long. Leaves:
stipules caducous (represented in most specimens only by small scars),
broadly lanceolate, c. 3.5-4 mm. long, acuminate, entire or obscurely den-
ticulate, brownish, scarious. Petioles subterete (somewhat flattened adax-

96 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

ially), rugulose or sometimes scabridulous, 2.5—-5 mm. long. Leaf-blades
coriaceous, oblong-elliptic, tending to be slightly obovate, (3—-) 4-8 cm.
long, (1.5—) 2—3.5 (—4.5) cm. broad, obtuse or rounded and with a minute
blackish apiculum (sometimes obsolete) at the tip, acute to obtuse at the
base; above sublucid or dully olivaceous, minutely foveolate, the midrib
plane or slightly raised, the lateral and tertiary veins anastomosing in a
somewhat prominent reticulum; beneath paler, the midrib prominently
raised, the laterals (mostly 6-10 on a side) and veinlets forming a prom-
inent reticulum; margins with a reflexed thickened marginal rim (which
is colored as the rest of the undersurface).

Monoecious; flowers appearing on new branchlets with the expanding
leaves, the male early deciduous; cymules bisexual, of 1 or 2 (rarely 3)
female and c. 10-12 male flowers.

Male flower: pedicel slender, 3-5 mm. long. Calyx purplish or dark
red; calyx-lobes 4, chartaceous, biseriate, the outer broadly ovate or ob-
long, the inner obovate or suborbicular, 1.9—2.1 mm. long, 1.4-2 mm.
broad, rounded and entire at the tip, mostly 3-nerved, the irregular veins
dark and conspicuous. Disk-segments 4, irregularly cubical or subglobose,
foveolate-pitted, 0.4-0.6 mm. across. Stamens 2; filaments completely
connate into a rather slender terete column usually 0.6-0.9 mm. high and
0.2-0.25 mm. thick; anthers erect, sessile atop the column, discrete or
fused back-to-back below, triangular-ovate, (0.5—-) 0.6-0.8 mm. long,
0.4-0.5 mm. broad; anther-sacs subparallel, dehiscing vertically, the slits
not confluent; pollen grains c. 20-24 ,» in diameter, areoles oligobrochate
or transitional to polybrochate, c. 12-15 per amb, 4-6 w across.

Female flower: pedicel terete and slender at the very base but abruptly
dilated above, not angled, 4-5 mm. long. Calyx purplish or dark red;
calyx-lobes 4, coriaceous, biseriate, subequal or unequal, at anthesis more
or less spreading with the tips reflexed, elliptic to broadly ovate-oblong, the
larger lobes c. 3-5 (—6) mm. long and 2-4 (—5) mm. broad, rounded or
obtuse at the tip, triplinerved but the laterals much less conspicuous than
in the male calyx, mesophyll densely crystalliferous. Disk squarish, the
4 coalesced segments forming a shallow undulate-crenulate, foveolate-pitted
cup c. 2 mm. across. Ovary reddish brown, sulcate, the tips of the carpels
slightly projecting above the insertion of the styles; styles erect, 2.3-3.5
mm. high, connate into a comparatively slender column 1.7—2.4 mm. high
and 0.4—0.55 mm. thick; stylar branches divergent, recurving, terete and
narrowed to subacute tips.

Capsule obtusely angled, c. 5.5-6 mm. in diameter, smooth, reddish-
brown, not veiny. Columella 2.5-3 mm. long. Seeds trigonous, symmetric,
3.7-4 mm. long, 2.4-2.8 mm. radially and tangentially, reddish brown,
smooth aaa colliculose) ; hilum subterminal, elliptic, c. 0.5 mm. long.

Collected flowering Feb.—Apr., Aug.; fruiting Feb— Mar., July.

Type: Cuba, Shafer 8446.

DISTRIBUTION: serpentine areas, usually in pinelands, eastern Oriente
province, Cuba (Map XXIV)

1958] WEBSTER, WEST INDIAN SPECIES OF PHYLLANTHUS 97

A. OrrenTE: Cananova, sur le charrascal serpentineux du Cerro de
Miraflores, 16-23 Apr. 1943, Marie-Victorin, Clement, & Alain 21634 (MT);
vicinity of Moa, Arroyo Moa, 10 Apr. 1945, Acuna 12479 (SV, US); Moa,
pinares, summer 1939, Mrs. Bucher 61 (NY, SV): Moa, rio de la scierie, 16-23
Apr. 1943, Marie-Victorin, Clement, & Alain 21635 (MT); Moa, 29 Aug. 1917,
Roig 1545 (NY, SV); wet pinelands near Punta Gorda, 14 July 1947, Leén &
Clemente 23057 (MICH); rich woods, alluvial valley of Rio Yamaniguey,
27 Feb. to 1 Mar. 1910, Shafer 4228 (NY, US), 4274 (F, NY); between Yamuri
Arriba and Bermejal, Feb. 1911, Shafer 8446 (NY, Lectotype); Cuchillas de
Baracoa, c. 1 mi. north of Rio Yumuri, pinelands, alt. c. 2000 ft., 18 Mar. 1954,
Jervis 3355 (GH).

Of the three species of sect. Thamnocharis, P. cinctus is the commonest
and most widespread, occupying relatively mesophytic riparian habitats
as well as drier pinelands or scrub. It correspondingly is the most variable
of the three species, and Jervis 3355 is particularly divergent due to its
large female calyx (the longer lobes being 6 by 5 mm. instead of 3-5 by
2-4 mm.) and smaller leaves with only 5—7 lateral veins. In these respects
it partially closes the morphological gap between P. cinctus and P. ekmanti
and provides additional support for the possible alternative course of
grouping these plants as two subspecies of a single species. Nevertheless,
the gap — although narrowed — remains, and until further knowledge of
the range of variation can be obtained, the present arrangement appears
to be the best.

Also occurring in the Moa region within the range of P. cinctus is the
narrowly endemic P. chryseus (sect. Glyptothamnus), which is strikingly
similar in its habit, tetramerous calyx and androecium of two stamens.
However, it differs in so many important respects (e.g., persistent stipules,
unisexual inflorescence, annular male disk, lacerate styles) that the simi-
larity to P. cinctus would seem to be ascribable to convergent development
rather than to a close affinity.

57. Phyllanthus ekmanii Webster, Contr. Gray Herb. 176: 60. 1955.
(PLATE XXVI, figs. G-J; PLATE XXVII, fig. B).

Phyllanthus cinctus Urb. Symb. Ant. 9: 191-192. 1924 (as to description, not
as tot

Subshrub or treelet, with the habit of P. cinctus, 0.3-1 m. high. Lower
leaves of main axis (missing on many specimens) with petioles 5-7 mm.
long, leaf-blades obovate or broadly elliptic, 3.5-6.5 cm. long and 2—4 cm.
broad; upper leaves reduced to cataphylls: stipules triangular-lanceolate,
acuminate, 10-17 mm. long, 4-5 mm. broad (on vigorous shoots; some-
times as small as 7 mm. long and 3.5 mm. broad), truncate at the base,
conspicuously longitudinally corrugate-ribbed, dull reddish brown, scari-
ous-indurate; blade acicular. Deciduous branchlets erect to spreading,
(8-) 10-23 cm. long, 2.5-3 mm. broad, olivaceous, smooth or pustulate-
scabridulous, distinctly flattened, with c. 7-15 leaves; first internode 15—
45 mm. long, median internodes 10-25 mm. long. Leaves: stipules cadu-

98 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxix

cous (represented in most specimens only by small scars), lanceolate, c.
1.5—2 mm. long, acuminate, obscurely denticulate, brownish, scarious.
Petioles somewhat flattened adaxially, rugulose, sometimes scabridulous,
(2—) 2.5-4 mm. long. Leaf-blades stiffly coriaceous, elliptic to slightly
ovate, (2.5—) 3-5 (—6) cm. long, 1.3-3 (—3.5) cm. broad, obtusely rounded
or emarginate at the tip (the minute blackish apiculum nearly or quite
obsolete), cuneate to obtuse or rounded at the base; above sublucid.
minutely foveolate, the midrib plane or slightly raised, the lateral and ter-
tiary veins anastomosing in a prominent somewhat raised reticulum; be-
neath paler, the midrib prominently raised, the lateral (c. 4-6 on a side)
and tertiary veins forming a reticulum more prominent than that above:
margins with a reflexed thickened brownish or somewhat orange marginal
rim.

Monoecious; flowers appearing on new branchlets with the expanding
leaves, the male early deciduous; cymules bisexual, of 1 or 2 female and
c. 2-5 male flowers.

Male flower: pedicel slender, becoming 4-9 mm. long, abruptly dilated
and fleshy above. Calyx dark reddish (rarely creamy-white?); calyx-
lobes 4, chartaceous, biseriate, suborbicular, 3—3.7 mm. long, 2.3—2.7 (—3)
mm. broad, rounded and entire at the tip, ordinarily with 5—7 nerves from
the base but these not very conspicuous. Disk-segments 4, subrectangular,
more or less flattened, rugulose, c. 0.7-1.1 mm. across. Stamens 2; fila-
ments completely connate into a stout column c. 0.5-0.8 mm. high and
0.6 mm. thick; anthers erect, sessile atop the column, fused back-to-back
from 1% to all their lengths, narrowly triangular-lanceolate, c. 1-1.2 mm.
long, 0.5-0.7 mm. broad; anther-sacs slightly divergent, dehiscing ver-
tically, the slits not confluent; pollen grains c. 24-28 ,» in diameter, areoles
oligobrochate, mostly 15—18 per amb, c. 4—6 p across.

Female flower: pedicel terete and slender at the base but abruptly
dilated above, not angled, 2.5-5 (-—6) mm. long. Calyx dark reddish
(rarely creamy-white?); calyx-lobes 4, coriaceous, biseriate, subequal,
broadly ovate or elliptic (the outer strongly convex, the inner condupli-
cate), erect (the tips not reflexed), the larger lobes 6-8 mm. long and
3-5 mm. broad, rounded at the tip, with c. 6 or 7 subparallel conspicu-
ously reticulate-ramifying nerves, mesophyll densely crystalliferous. Disk
squarish, very massive, forming below the ovary a mound-like pedestal
c. 0.5—-! mm. high and 3—4 mm. broad, the foveolate rim depressed.

Ovary reddish brown, ea sulcate; styles erect, 3-5 mm. high, con-
nate into a column (2. I ) 2.5-4 mm. high and c. 0.4 mm. thick (some-
what dilated upwards); eee branches divergent, slender, recurving,
terete and narrowed to subacute tips.

Capsule valves c. 5 mm. long, reddish brown, smooth, not veiny. Colu-
mella nearly 3 mm. long. Seeds trigonous, nearly symmetric, 3.3—3.4 mm.
long, 2—2.2 mm. radially, 2.2-2.5 mm. tangentially, reddish brown, smooth
tas colliculose) ; hilum subterminal, ovoid or elliptic, c. 0.5 mm.

"Collected flowering Feb., Apr., Aug.; fruiting May, June, July.

N2eq. 297

KEIN OF uss

ro ee oe

«g. £ L Ebmoa

Fic. A. Phyllanthus microdictyus Urb. (Ek-
25

man 3705 iS, HOLOTYPE]), Fic. B. Bislliaiis panies Webster (Ekman 2
the unreduced leaves on the lower part of the main axis

3 [S, HOLoTYPE]); note

TIXXX “IOA ‘day ATONay “wolf

IIAXX SLV1d

100 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

Type: Cuba, Ekman 2523.

DISTRIBUTION: restricted to serpentine areas of the Sierra de Nipe,
Oriente province, Cuba (Map XXIV).

CUBA. ORIENTE, SIERRA DE Nipe: Cayo del Rey, Pinar Colorado. 16 Apr.
1940, Carabia 3587 (MICH, NY); Rio Piloto, locis rupestribus, alt. 750 m.,
18 Aug. 1914, 15 May 1915, Ekman 2523 (S, HOLOTYPE; NY, ISOTYPE), 5704
(S); charrascales, Rio Piloto, 10 June 1915, 27 Apr. 1919, Ekman 6026, 19166
(S); charrascales, ad viam Bio, 27 Apr. 1919, Ekman 9583 (S); exposed ridge,
Pico Estrella, 18 Feb. 1954, Jervis 3037, 3065 (GH).

This species endemic to the Sierra de Nipe is so closely related to
P. cinctus that Urban’s confusion (in associating specimens of the former
with the name of the latter) is understandable. Although some of the dis-
tinctions between the two taxa are not quite absolute and may even further
break down upon study of additional collections, there are so many points
of difference that it seems preferable to rank them as closely related
allopatric species rather than as two subspecies of one variable species.
The larger size of the male flower, at least, always distinguishes P. ekmanii;
and its hypertrophied female disk, which somewhat recalls the gynophore
of P. microdictyus, is very different from the unmodified disk of P. cinctus.
In addition, the female calyx and styles of P. ekmanii appear to be def-
initely larger than those of P. cinctus, but the dimensions of these organs
are subject to such variation after anthesis that on the basis of specimens
at hand it is difficult to make comparable measurements. There are, fur-
thermore, additional differences in the mean values of such characters as
leaf-size and vein-number which lend support to the suppositon that P.
ekmanti and P. cinctus are distinct species.

In contrast to P. cinctus, which ranges over much of the Sagua-Baracoa
massif, P. ekmani has a much more restricted range; it has thus far been
collected only in the southern part of the Sierra de Nipe but not in the
vicinity of Loma Mensura or Bandera. The only divergent specimen no-
ticed was Jervis 3037 which, according to the collector, had “creamy-
white” rather than reddish flowers, Unfortunately no flowers were present
on the single available specimen of this collection number; but since Jervis
3065 from the same locality has typically reddish flowers, it would appear
that the whitish-flowered form is only a trivial local variant.

(To be concluded )

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXXIX APRIL 1958 NUMBER 2

NEW RECORDS OF JAMAICAN
FLOWERING PLANTS, I

RicHArp A. HOWARD AND GEORGE R. PROCTOR

FoR THE PAST SEVERAL YEARS the authors have been conducting surveys
of the vegetation growing on bauxitic and related soils in Jamaica, B.W.LI.
(Jour. Arnold Arb. 38: 1-41, 151-169. 1957). In the course of this work
it was possible to visit interesting geographical areas, sometimes en route
between areas of mining operations, sometimes in areas off the beaten path
where bauxite is only suspected to occur. In all of these forays our work
was encouraged by Mr. C. Bernard Lewis, director of the Institute of
Jamaica, and was financed in part by contributions from the Kaiser Bauxite
Company and the Reynolds Jamaica Mines, Ltd. To Mr. Lewis and to
officers and representatives of the contributing aluminum mining com-
panies, we express our appreciation. We are also grateful to Dr. Lily M.
Perry for her assistance with the Latin descriptions. The specimens cited
are to be found in the herbaria of the Arnold Arboretum (A) or the Gray
Herbarium (cH) and the Institute of Jamaica (13). A few specimens
have been seen in the Hope Botanic Garden Herbarium now at the Depart-
ment of Botany of the University College of the West Indies (ucw1).
Duplicate specimens of our collections will be distributed to other herbaria.

Acrosynanthus jamaicensis, sp. nov. RUBIACEAE

Frutex 3 m. altus, ramis teretibus, ramulis compressis hispidulis; stipulis
late triangularibus ‘apiculatis ciliatis usque 2 mm. longis, intus petiolo
coalitis; foliis 3-4.5 cm. longis, 4-8 mm. latis lineari- lanceolatis, apice
acutis basi cuneatis decurrentibusque, margine integris et valde recurvis
utrinque resinaceis, supra atroviridibus hispidulis, subtus sparsim hispidulis,
albescentibus (vivis vel siccis); petiolo 1-2 mm. longo, hispidulo, resi-
inflorescentia terminali cymosa, cymis 3-floris, pedunculo 1 cm.
longo, bracteis minutis, minus quam 0.5 mm. longis; pedicello 1-2 mm.
- hypanthio obovoideo, sepalis 4-5, ovato-lanceolatis, minus quam

mm. longis, crassis, apice patenti- srecurvatis hispidulis; corollae lobis 4,
uno Siemmmaue majore, 2.5-3 mm. longis, oblongis, apice rotundatis
utrinque dense pubescentibus (pilis clavatis), tubo similiter pubescente,
2.5-3 mm. longo; staminibus 4 tubi corollae basi affixis, filamentis minus
quam 0.5 mm. longis, antheris elongatis acutis vel acuminatis 1 mm. longis;

102 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. XxXxIx

disco piloso; stylo glabro, stigmate bilobato, ovario 2-loculari, loculis 8—12-
ovulatis; fructu maturo pallido-brunneo, 3 mm. diametro, seminibus pal-
lido-brunneis tenuiter scrobiculatis.

Jamaica. PARISH OF TRELAWNY: Shrub on limestone sides of steep ravine,
Ramgoat Cave district of the “cockpit” country, Howard 14133 (A-type), Howard
& Proctor 14391 (a, 13); Mrs. Bernard Benenan as (J)

This is the first record of the genus Acrosynanthus from outside Cuba.
The genus, as currently known, was established by Urban Se Antill.

7: 544. 1913) and monographed by Standley (N. Amer. FI. 32: 43. 1918).
ae vege three additional 7 ee were described, bringing . total
eight species, all from the Oriente province of Cuba. The genus 1s

more or less distinct within the Rae where it has been assigned;
however, it is badly in need of restudy and probably of redefinition. The
present species with five sepals, 4 petals, one larger than the others, having
a pubescent corolla inside and out and four stamens attached at the base
of a very short corolla tube does not fit readily into the limitations of
the genus as recognized by Urban and Standley. Acrosynanthus jamaicen-
sis is easily recognized by the resinous covering of the leaves and young
stems, as is often found in species of Phialanthus and Antirhea. The white
color of the under surface of the revolute leaves draws attention to this
plant in the field

Alvaradoa lewisil, sp. nov. SIMARUBACEAE
Frutex 4 m. altus, ramis virgatis peer foliolis 22-25, ellipticis
vel oblongis supra medium latissimis, 1.5—2. . longis, 0.7—1 cm. latis,

apice rotundatis vel retusis, basi angustatis a seu margine valde
recurvis, supra roameibie subtus pallida flavoviridibus, glabris; petio-
lulo 1 mm. longo puberulo vel breviter adpresso-pubescente; inflorescentia
terminali racemosa arcuatim pendula, rhachi aurea vel fulva puberula,
pedicellis 9-11 mm. longis puberulis; fructu glabro, immaturo rubro, ma-
turo flavo, lanceolato-ovoideo, 12-18 mm. longo, 6-7 mm. diametro, medio
latissimo, utrinque angustato, carpellis (sterili fertilibusque) in magnitu-
dine aequalibus.

Jamaica. PARISH OF TRELAWNy: A shrub on the steep face of a lmestone
ravine, Ramgoat Cave district in the “cockpit” country, Howard 14128 (A-type).

Alvaradoa is a small genus of five species recently monographed by
Cronquist (Brittonia 5: 133-137. 1944). Two of the species previously
known have glabrous fruits, as does this. Alvaradoa lewisii differs from
A. jamaicensis, the only species previously reported from Jamaica, in hav-
ing leaflets oblong, broadest above the middle, with the margin strongly
recurved, and the fruits longer than broad, but broadest at the middle and
cuneate or narrowed at the base. Alvaradoa jamaicensis is based on a Prior
specimen from Union Hill and most recent collections are from the central
part of the island. Alvaradoa lewisii is from the “cockpit” country in the
western third of the island of Jamaica, a region sf noteworthy local en-
demism. The other glabrous-fruited species is A/varadoa arborescens, from

1958 | HOWARD & PROCTOR, JAMAICAN PLANTS, I 103

Cuba, distinguishable from A. lewisii in having all three carpels equally
developed with the styles at the same level, and in the shape of the fruit
which tapers at the apex and the base.

Alvaradoa lewisii is named in honor of Mr. C. Bernard Lewis, director
of the Institute of Jamaica, known for his work on the Pedro Cays, the
Cayman Islands and many parts of Jamaica.

Clusia portlandiana, sp. nov. GUTTIFERAE

Frutex vel arbor epiphytica 4-7 m. alta; foliis obovatis planis, 13—
18 cm. longis, 7.5-12 cm. latis, superiore 1% latissimis, apice rotundatis
vel truncatis, versus basin angustatis, sessilibus, costa supra leviter canali-
culata subtus prominente, venis numerosis parallelis; inflorescentia ter-
minali corymbosa, 12-15 cm. longa, 12—15—flora; bracteis primarlis an-
guste triangularibus, 9-13 mm. longis, basi 6 mm. latis, apice acutis,
infrequenter foliiformibus obovatis 3 cm. longis 2 cm. latis; floribus fem-
ineis pedunculatis; pedunculo 1 cm. longo; bracteolis 4 oppositis et decus-
satis; sepalis 4 orbiculari-ovatis carnosis, siccis margine scariosis, 4-5 mm.
longis latisque; petalis 5, ovatis carnosis 5-6 mm. longis latisque, carinatis,
margine scariosis; fructu oblongo, 1-1.5 cm. longo, | cm. diametro, 5—
loculari, apice rotundatis, stigmatibus 5, sessilibus; floribus masculinis
pedunculatis; pedunculo 6-8 mm. longo; bracteis 4; sepalis 4, petalis 5
ut in flore feminea, staminibus numerosis, filamentis 4 mm. longis, antheris
1 mm. longis.

amaica. PARISH oF PorTLAND: In mist forest on limestone, John Crow Moun-
tains. alt. 1500-2500 feet, 1.5-2.5 mi. southwest of Ecclesdown, Howard & Proc-
tor 14765 (a, 3J, fruit), Proctor 9797 (aj, fruit), Proctor 993 (a-type; IJ, fruit),
Webster & Wilson 5136 (A, male); summit of John Crow Mountains, east 0
Millbank, Swabey 13018 (ucw1, male).

Clusia portlandiana is similar to C. venosa and C. krugiana of the Lesser
Antilles and Puerto Rico, differing from both in the larger inflorescence,
longer peduncles and oblong fruit. In comparison with the other species
reported from Jamaica, C. portlandiana is similar only to C. clarendonensts,
differing in the larger leaves, larger inflorescence and the more numerous
female flowers.

Dipholis bullata, sp. nov. SAPOTACEAE

Arbor parva usque 8 m. alta; foliis obovatis vel fere orbicularibus,
plerumque 4-13 cm. longis 3.5-6.5 cm. latis, apice obtusis vel rotundatis,
glabris coriaceis saepissime plus minusve bullatis propter margines re-
volutos, supra nitide atroviridibus subtus pallidioribus, costa subtus prom-
inente, venis cetera subobscuris, petiolo 4-10 mm. longo; floribus in axillis
2-4, pedicello crasso minute sparsimque adpresso-strigoso, 5-10 mm., in
fructu usque 15 mm. longo deinde suberoso-lenticellato; sepalis similibus
D. montana, corolla ca. 5 mm. longa, tubo 2.5—3 mm. longo, lobis cucul-
latis, appendicibus lateralibus planis acuminatis sparsim erosis; filamentis
ca. 1.2 mm. longis, antheris 0.6-0.8 mm. longis; staminodiis late ovatis,

104 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxx1x

margine erosis, apice acutis vel subacuminatis filamentis subaequilongis
vel paullo longioribus; stylo ca. 1 mm. longo, ovario glabro 5-loculari;
fructu immaturo anguste ovoideo, apice truncato-apiculato, nitide olivaceo,
ca. 15 mm. longo 5 mm. diametro, 1-spermo, semine immature, hilo ut
videtur laterali.

Jamaica. ParisH or PorTLAND: In elfin-woodland over limestone (elevation
about 2500 feet), John Crow Mountains, 1.5-2.5 miles southwest of Ecclesdown,
January 24, 1956, Howard, Proctor & Stearn 14759 (A-type; IJ, fruit); Howard,
Proctor & Stearn 14755 (with fewer, more immature fruit); September 14,
1956, Howard & Proctor 14842 (flowers) (A, IJ).

Though obviously related to Dipholis montana of the Blue Mountains
area, the present species differs rather strikingly in its somewhat larger,
shining, bullate leaves, fewer flowers per cluster, longer pedicels, slightly
smaller flowers and differently-shaped fruits. It is entirely unlike D. octo-
sepala, another Jamaican congener with which D, montana has been com-
pared

Myrica jamaicensis, sp. nov. MYRICACEAE

Arbor parva 3-4 m. alta, ramulis sparsim piloso-pubescentibus, pilis
brevibus crispis albisque; foliis ellipticis vel obovato-ellipticis, 3.5 x 2,
4.5 & 2.5, 6 & 3 cm., apice rotundatis vel truncatis, retusis, basi cuneatis
decurrentibusque, margine undulatis supra medium grosse sinuato-denticu-
latis, utrinque aequaliter minute foveolato-puncticulatis et glandulosis,
bullatis, costa subtus sparsim, supra breviter pilosa, venis primariis 5—6-
paribus subtus sparsim breviter pilosis deinde glabrescentibus; petiolo
1-2 mm. longo, crispe et breviter piloso; inflorescentia feminea 8-9 mm.
longa, bracteis 0.6-0.7 mm. longis apice obtusis; fructu subgloboso 3 mm.
diametro, ceraceo-papillato, dense glanduloso.

Jamaica. PARISH OF PoRTLAND: John Crow Mountains at 2500 feet, 2.5 miles
southwest of Ecclesdown, Howard & Proctor 14832 (A-type; IJ).

This species differs from the two common species of Jamaica and the
West Indies, Myrica cerifera and M. microcarpa, in having broader leaves,
rounded and retuse at the apex, the margin undulate and sinuate and not
evidently toothed, the blades bullate between the prominent primary veins.
The plant is less pubescent than the other species and, while the staminate
inflorescence is not known to us, the fruits of this new species are more
resinous than others we have seen. The similarity of M. jamaicensis is
with.M. microcarpa rather than with M. cerifera.

Phialanthus myrtilloides Griseb. RUBIACEAE
Jamaica. PARISH OF TRELAWNY: a shrub 2.5 m. tall, with arching branches,
on limestone and steep slopes of ravine, Ramgoat Cave district of “cockpit”
country, Howard 14134 (a).
Two species of Phialanthus have been described and reported from
Jamaica, with an additional seven species known from Cuba and one from
Cuba and the Bahamas. The specimen cited above does not fit either of

1958] HOWARD & PROCTOR, JAMAICAN PLANTS, I 105

the species known from Jamaica and is assigned to P. myrtilloides, a
species described from Cuba but later recognized from the Bahamas, thus
extending its range. Phialanthus is not well represented in herbaria and
has not received comparative studies in the field. Certainly the taxonomic
characters used by Standley in his treatment of the genus (N. Amer. FI.
32: 281. 1934) are very weak. Further study will probably show that
there are fewer species and greater variation than are currently recognized
and nothing is to be gained by adding still another species at this time.
The specimen cited is more heavily resinous than the other species of the
genus. The leaves are lanceolate, broadest at the middle and only slightly
recurved at the margin. The inflorescences are sessile and four flowers are
produced. The corolla equals the calyx lobes in flower, but the calyx lobes
soon expand until in fruit they exceed the tube in length. The stamens
are shorter than the corolla lobes. The specimens we have seen from the
Bahamas indicate a considerable variation in the size and shape of the
leaves and the amount of resin produced. The Jamaican specimen can be
included in this range of variation.

Schefflera stearnii, sp. nov. ARALIACEAE

Frutex 2-3 m. altus, glaber, inflorescentia excepta; ramis crassis tereti-
bus prominenter lenticellatis; petiolo striato, 9-16 cm. longo prope basin
lenticellato, ligulo simili S. sciodaphyllum,; foliolis 5-8, majoribus minori-
busque intermixtis, firme coriaceis, supra nitido-atroviridibus, subtus oliva-
ceis, obovato-oblongis usque 18 cm. longis, 8 cm. latis, basi truncatis vel
inaequalibus, apice abrupte acuminatis, margine eartilaginico integris ali-
quantum recurvis, costa praecipus subtu us prominente, venis lateralibus
utrinque prominulis, petiolulo 1-5 cm. longo; inflorescentia terminali
ramosa, ramis ca. 6, 13-22 cm. longis, minute denseque ferrugineo-fur-
furaceo-pubescentibus, capitulis pedicellatis (pedicello 3-11 mm. longo),
racemosis, 3—6-floris inter flores plus minusve setulosis; calyce coriaceo,
inferiore x, dense ferrugineo-puberulo, 1.5-2 mm. longo, ca. 2 mm. dia-
metro, minute dentato; petalis Bebernlict 1.5-2 mm. longis; antheris
0.5-1 mm. longis; stylis viridibus, ca. 1 mm. longis, apice valde incurvis;
fructu non viso.

Jamaica. ParisH oF PorTLAND: in mossy elfin-woodland over limestone, ele-
vation about 2500 feet, John Crow Mountains, 1.5—-2.5 miles southwest of Eccles-
down, March 9, 1957, Proctor 16255 (1J- type) ; Howard, Proctor & Stearn 14761
(A, IJ, in bud)

In his treatment of Sche flera in North American Flora (28B(1): 25-29.
1944), A. C. Smith ascribes two species to the Jamaican flora. One of these,
S. troyana, is characterized by its dense white tomentum; the other, S.
sciodaphyllum, by being nearly glabrous in contrast. Other, less striking
differences can be observed by closely comparing the descriptions of the
two species. Another name, Sciadophyllum praetermissum, is reduced to
synonymy under Schlefflera sciodaphyllum, probably in the belief that it
represents but a juvenile and more pubescent flowering stage of the latter

106 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. xxxIx

species. This is a point which needs to be settled by further collecting.

The present new taxon entirely lacks the white tomentum of Schefflera
troyana and, in fact, differs from that species in many details. From
S. sciodaphyllum it differs by its shorter petioles and fewer, shorter, bi-
colorous leaflets of a different texture; by much shorter inflorescence
branches covered with a rusty-scurfy pubescence different both quantita-
tively and qualitatively from the sparse, minute, whitish hairs of S. scio-
daphyllum; by the fewer and strictly sessile (instead of pedicellate) flow-

rs; by the slightly smaller calyx and shorter, puberulent petals; by the
much shorter anthers of a different shape; and by the green (instead of
carnose) styles.

Both Schefflera stearnit and S. sciodaphyllum occur in the John Crow
Mountains more or less adjacent to each other. In our opinion, therefore,
the differences between them cannot be attributed to contrasting environ-
ments

Weinmannia portlandiana, sp. nov. CUNONIACEAE

Arbor 2 m. alta, ramulis ultimis compressis atro-fuscis sparsim et breviter
pubescentibus; foliis oppositis trifoliolatis glabris, foliolo terminali lanceo-
lato-elliptico, 44.5 cm. longo, 1.5-2 cm. lato, apice acuto, basi cuneato,
margine versus apicem crenato versus basin integro, venis supra leviter
impressis subtus subobscuris, foliolis lateralibus oblongis vel ellipticis apice
rotundatis basi obliquis uno latere cuneatis altero rotundatis, margine
saltem supra medium crenatis, 2.5 & 1.7 cm2.7 1.7 cm., petiolo
1.2-1.4 cm. longo alato, alis obovatis; inflorescentia terminali 8 cm.
longa pseudoracemosa, rhaci sparsim et breviter pubescente; floribus her-
maphroditis fasciculatis (2-5), bracteolis late ovatis usque 1 mm. longis,
pedicello 4.5 mm. longo sparsim pubescente; calyce 4—partito, lobis ovatis
usque 1 mm. longis; petalis oblongis 2-3 mm. longis; filamentis 4 mm.
longis; pistillis 2, usque 5 mm. longis.

Jamaica. PARISH OF PORTLAND: John Crow Mountains at 2500 feet, about
2.5 miles southwest of Ecclesdown, Howard & Proctor 14839 (A-type; IJ), Proc-
tor 11351 (13), Howard, Proctor & Stearn 14770 (13).

The genus Weinmannia has been known previously in Jamaica by the
variable and widespread W. pinnata L. and the pubescent form of the
latter, W. hirta Sw. The present species differs from W. pinnata by havy-
ing glabrous and strictly ternate leaves, the leaflets of which are larger
than even the extreme forms of W. pinnata. Weinmannia pinnata as it
occurs in the islands from Cuba to Grenada needs further taxonomic study.
It is probable that a number of good subspecies or varieties will eventu-
ally be established in this complex. A comprehensive field study of vari-
ous populations is required, however, to understand the morphological
variations which occur and are represented in herbaria.

ARNOLD ARBORETUM

and
THE INSTITUTE OF JAMAICA

1958 | JARRETT, NOTE ON BALANOSTREBLUS 107

A NOTE ON THE IDENTITY OF THE GENUS
BALANOSTREBLUS (MORACEAE) *

FRANCES M. JARRETT

THE OBJECT OF THIS NOTE is to provide an identification for Balanostre-
blus Kurz, a monotypic genus of the Moraceae which has been credited
to the Asiatic flora. It was assigned by Bentham and Hooker (Gen. PI.
3: 377. 1880) to the subfamily Artocarpoideae and tribe Artocarpeae
(using the modern names for these groups) and placed among the Amer-
ican genera of the tribe, next to Sorocea St.-Hil., a small genus of the trop-
ical forest.

Balanostreblus was described by Kurz in 1873 (New Burmese Plants.
Part III. Jour. Asiat. Soc. Bengal 42: 247. ¢. 19) and his account is given
in full below.

BALANOSTREBLUS, nov. gen. Flores monoici; masculi ignoti (ex inflores-
centiis valde juvenilibus probabiliter amentace i?). Feminei racemosi: perian-

m cum ovario connatum, sursum liberum et ovarium omnino includens, apice
anaes Ovarium semisuperum, 1-ovulatum, ovulo pendulo; stylus perbrevis,
e perianthii orificio protrudens; stigmata 2, brevia, crassa, villosula. Drupa
perianthio carnoso inclusa, monosperma. ne: lactescens, subglabra, foliis alter-
nis grosse spinescenti-dentatis. Genus imperfecte cognitum sed distinctissimum
Antiari affine.

BALANOSTREBLUS ILICIFOLIUS, nov. sp. Arbor ramulis scabriuscule
puberulis; folia elliptica ad lato-ovalia, petiolo terete 1—2 lin. longo glabro suffulta,
basi saepius subinaequali acuta v. obtusa, rigide coriacea, spinoso-acuta, grosse
spinoso-dentata, 1-3 poll. longa, glabra, supra nitida costa supra immersa subtus
unacum nervis lateralibus arcuato anastomosantibus valde prominente; flores
parvi, viridiusculi, pedicello brevi crasso suffulti, in racemum axillarem brevem
collecti; periant thium obturbinatum, rugulose-tuberculatum, c. 2 lin. longum;
drupae pisi minoris magnitudine, rubrae, rugulosae, carnosae, glabrae. — Chitta-
gong (Hf. et Th. sub Sapii sp. No. 4); Ava (J. Anderson).

The generic characters were thus based by Kurz on the female inflores-
cences, The plate, which is reproduced here, shows these attached to a
leafy shoot and also includes dissections of the flowers. The inflorescences
are raceme-like, though presumably having the cymose origin typical of
the Moraceae, and their appearance is in contrast to that of the other Old
World genera of the Artocarpeae, in which the inflorescences are capitate.
The pedicellate flowers are likewise distinctive in structure and may be
regarded as having the ovary sunken in and fused to a fleshy receptacle
which is surmounted by a short, tubular, perforate perianth from which

his paper is based on part of a thesis presented to the Univ ersity of Cambridge,
ar for the degree of Ph.D. Thanks are due to the Directors of the Royal Botanic
Gardens, ae hae the Muséum National d’Histoire Naturelle, Laboratoire de Phanéro-
r their hospitality and to the latter for the loan of the collection of

ear raat “licifolius.

108 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

ny

SPE

fe 98
fate
cae

As en
Pate
ce

eos
=

—pr

Original illustration of Balanostreblus ilicifolius Kurz, Jour, Asiatic Soc.
Bengal 42: ¢. 19. 1873

1958] JARRETT, NOTE ON BALANOSTREBLUS 109

the bifid style projects. In this, as in the general aspect of the inflores-
cences, the genus resembles Sorocea.

Kurz did not state from which plant the drawing was made, but the
identity of this was established by Hutchinson in 1918 (Kew Bull. 1918:
147-153) when, in a paper which was primarily a revision of Taxotrophis,
a genus belonging to the Moroideae, he also discussed the typification of
Balanostreblus ilicifolius. He examined Kurz’s material from the her-
barium of the Botanic Garden, Calcutta, and found that it consisted of
the specimen from Chittagong, Hooker and Thomson 4 (also at Kew),
and a specimen described as “cultivated at the Botanic Garden.” Appar-
ently no specimen from Ava was extant, since Hutchinson did not receive
one, and he concluded that this might have been a living plant, though
he mentioned that there was a collection under this name at Kew made
by Anderson at Bhamo, about 180 miles to the north-east of Ava. Hutchin-
son stated that there was no doubt that the illustration had been prepared
from the cultivated plant, which was female. The plant from Chittagong
was male and this he identified as a collection of Taxotrophis ilicifolia
Vidal, a variable species, which may have spiny-toothed leaves rather
similar to those of the cultivated plant with which it had been matched
by Kurz. Having thus shown that Balanostreblus ilicifolius was a mixtum
compositum, Hutchinson redefined and redescribed the genus (l.c., 152),
basing it entirely on the cultivated specimen and providing a new illustra-
tion. He still assumed that this plant was of Asiatic origin and suggested
that it might be Anderson’s collection from Ava. He thought that the genus
should probably be removed to the Broussonetieae in the subfamily Moro-
ideae. However, its characters are not in accord with that group.

In the course of a review of the characters of the genera of the Moraceae,
Balanostreblus attracted my attention. The female inflorescence did not
resemble that of any Old World genus and Bentham and Hooker’s placing
of the genus still seemed from the descriptions of Kurz and Hutchinson
(in the absence of the lectotype) to be the most satisfactory, although
the general classification of the Moraceae may be in need of revision. The
genus thus appeared to be anomalous in the Asiatic flora.

Another possibility as to the origin of the plant was suggested by the
finding under Balanostreblus, during a visit to the herbarium of the Mu-
séum d’Histoire Naturelle, Paris, of a specimen which had been gathered
by L. Pierre in the Botanic Garden at Calcutta in 1863, when he was assist-
ant there. This was ten years earlier than Kurz’s publication and the
specimen had later been identified as Balanostreblus ilicifolius on the basis
of Hutchinson’s paper. It matched the description and plate exactly and
it seems justifiable to assume that Pierre and Kurz made their collections
from the same tree. The label gave the origin as “Brasilia ?? India ?”
In view of this and the similarity already remarked between the female
inflorescences and those of the South and Central American genus Sorocea,
the sheet was compared with collections of that genus in the same herba-
rium. It was found to match almost perfectly Guillemin 131, Corcorado,
Brasil méridional, 1838, which is the type of Sorocea guilleminiana Gaudi-

110 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxIx

aud, Bot. Voy. Bonite, ¢. 74, 1844, except that in the latter most of the
indiviclual receptacles are enlarged and contain mature seeds. This species
is characterized by the verruculose surface of the receptacles.

It is therefore concluded that Balanostreblus Kurz emend. Hutchinson
is based on an introduced species of Sorocea, which is identified as S. guil-
leminiana, pending a revision of the genus. This was presumably brought
in during the introduction of Cinchona to India, which took place at about
this time and in the course of which several shipments of plants were sent
from the New World under the auspices of the Royal Botanic Gardens at
Kew.

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS ital

A MONOGRAPHIC STUDY OF THE WEST INDIAN
SPECIES OF PHYLLANTHUS *

GRADY L. WEBSTER

With five plates

Sect. 18. Orbicularia (Baill.) Griseb. Fl. Br. W. Ind. 34, 1859.

Orbicularia Baill. Etud. Gen. Euphorb. 616. 1858.
Roigia Britton, Mem. Torr. Bot. Club 16: 73. 1920.
Dimorphocladium Britton: ibid. 74.

Phyllanthus sect. Dimorphocladium (Britton) Pax & Hoffm. Naturl. Pflanzen-

fam. 19c: 63. 1931.

Shrubs with phyllanthoid branching, axes smooth and glabrous; leaves
with mesophyllar sclereids, mostly coriaceous, stipules (at least proximal
ones) mostly persistent. Monoecious; cymules male and bisexual, female
flowers usually only one per cymule. Male flower: calyx-lobes 6 (rarely
5); disk-segments free or coalescent; stamens 3—7, filaments united en-
tirely or below into a column; pollen grains areolate. Female flower:
calyx-lobes 6 (rarely 5); disk tenuous to rather massive; ovary sessile
or slightly stipitate; styles free or connate below into a column, the free
ends bifid, style-branches often revolute at the tips. Capsule oblate, veins
conspicuous or obscure; seeds trigonous, verruculose.

Type species: Orbicularia phyllanthoides Baill. |= Phyllanthus orbi-
cularis HBK. |

Although a well-marked and apparently monophyletic group, which here
is given essentially the same circumscription as that of Carabia (Ecol. Mon.
15: 335. 1945), sect. Orbicularia is rather difficult to characterize as dis-
tinct from neighboring sections in subg. Xylophylla. The most closely
related section, from which sect. Orbicularia has probably been derived,
is sect. Williamia. The species of the latter with sclerified leaves (subsect.
Incrustati) can be distinguished from the present group only by their
incrustate axes and lacerate style-tips.

After a prolonged study which has been pursued intermittently over
a period of several years, the following treatment of sect. Orbicularia is
presented with some diffidence. Despite personal field observations in
Cuba in 1951 and 1953 and the analysis of a considerable number of
herbarium specimens, a definitive resolution of the taxonomic problems in
the group has not been achieved. Field studies have been of value in dem-
onstrating (to the author’s satisfaction, at least) that some of the confu-
sion in the literature is traceable to the striking differences in leaf form
brought about by different ontogenetic changes and/or responses to dif-
ferent ecological situations. Thus it appears that the very glossy and

* Continued from volume XXXIX, p. 100.

12 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxrx

foveolate leaf on which P. foveolatus was described is only the older, more
coriaceous one of P. myrtilloides ssp. erythrinus; and the convex, very
glossy leaf which supposedly characterizes P. coelophyllus is merely the
kind of leaf developed by P. baracoensis in drier situations. However, a
much more serious difficulty to a satisfactory classification in the section
is the extreme range of variability of some species, particularly P. myr-
tilloides and P. orbicularis. It is also possible that in certain instances
hybridization has further complicated the picture by giving rise to anom-
alous specimens which appear to transgress the rather indistinct specific
lines. It would be extremely difficult to account for the several peculiar
forms from the Sierra de Nipe (described by Urban as species) except on

°o oO

74 72°)| 707} 16

Map XXV. Distribution of sect. Orbicularia: black dot, P. ella eee
Muell. ee ae P. orbicularis HBK.; vertical lines, remainder of species
in the secti

the supposition that they represent the results of crosses between P.
chamaecristoides and P. phiebocarpus. Finally, it must be remarked that
Urban’s usually sound taxonomic judgment failed him when treating sect.
Orbicularia, for he described numbers of “paper species” (such as P.
breviramis, P. cardiophyllus, and P. melanodiscus) which cannot be re-
tained at any taxonomic rank.

In attempting to formulate a usable and objective classification of sect.
Orbicularia, it has been necessary to abandon, in part, a purely morpho-
logical standard for delimitation of taxa and to rely heavily on the prin-
ciple of geographical replacement as formulated by Huxley (New System-
atics 36. 1940) and elucidated by Van Steenis (FI. Mal. ser. I. 5(3):
cxcv—cxcvi. 1957). The result is that, whereas Alain in the “Flora de
Cuba” (1953) recognized 19 Cuban species referable to sect. Orbicularia,
only 7 are accepted in the present treatment. It is realized that this may

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 113

appear excessively radical especially since it leads to some apparent incon-
sistencies, but nevertheless it seems the closest approximation to the facts
of nature that can be devised at this time. It is doubtful that the accumu-
lation of additional material will bring any improvement as long as it can
be analyzed only by orthodox herbarium methods; a completely satis-
factory classification of the section must surely await an extensive popula-
tion analysis with emphasis on cytological methods.

KEY TQ THE SPECIES

—e

. Leaves spathulate, acute at the base, rigidly coriaceous, the margin neither
reflexed nor revolu

ae ee eats leaf-blades of main stem 2-4 cm. ae branchlets

(otard UG evn 15 71a 0:2: Meare eRe em On te ae ae EOE UreT OM oe Mae 58. P. formosus

De shpat ae leaf-blades of main stem, if unreduced, not over 1.5 cm.
long; branchlets at most 5 cm. long.

3. Pedicel of female flower mostly over 10 mm. long; anthers sessile or

nearly so atop the column; calyx bright pink or purplish. ...«...

SRO RGEC Ct eS ere ROU Re ae Oe OT et ane ane 59. P. comosus

3. Pedicel of female flower less than 5 mm. long; filaments united about

ne way, the anthers long-stipitate; calyx whitish or See pinkish-

{10 Ae eee eet freee on cy es ei 2 60. P. orbicularis

—y
.

Leaves, if spathulate, not rigidly coriaceous nor with plane margins.
2. Stamens mostly 5-7 (rarely 4, very rarely 3); leaves not all expanding
with the flowers, not prominently reticulate on both sides with raised veins.
3. Styles usually more than 1 mm. long, united into a column mostly

.5 mm. high or more.
4. Pedicel of female flower mostly more than 3 mm. long (if exactly
3 then anthers stipitate and cataphylls deciduous).

rephrase toh) ae ns eae R SEN Seem an 8a g Ts 61. P. ‘myrtilloides
4. Pedicel of female flower mostly less than 3 mm. long (rarely to

5. Filaments of stamens mostly 1.5-2.5 mm. high, united in the
lower half; branchlets mostly with less than 10 elliptic to
orbicular leaves: cataphylls persistent. .... 60. P. orbicularis

. Filaments of aes not over 1 mm. high, completely united
or nearly so; branchlets mostly with 15-45 oblong to obovate,
often falcate, leaves; cataphylls deciduous. .................
ee er enn \ teen eee 62. P. chamaecristoides

3. Styles free or barely united at base, mostly less than 1 mm. long.

4. Cataphylls See eee branchlets 3-6 cm. long; leaves Le
petiole less than 1 mm. long, blade obovate or obcuneate, 3-7 m
long, prominently apiculate: outer calyx-lobes (of both sexes) nar-
rowly oblong, mostly less than 1 mm. broad. .. 63. P. scopulorum

4. Cataphylls ore branchlets mostly 6-12 cm. long; leaves
with petiole 1-2.5 mm. long, blade broadly elliptic to orbicular,
mostly 10-20 mm. long, at most obscurely apiculate; calyx- lobes
subequal, the outer 1 mm. broad or more. ....................
EEA Win tat Sk a OR ee 4. P. nummularioides

2. Stamens mostly 3 (rarely 2 or 4); leaves and flowers expanding together,

Nn

114 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

leaf-blades with a prominent raised reticulum on both sides; styles free
or nearly so, spreading, less than 1 mm. long. ..... 65. P. phlebocarpus

58. Phyllanthus formosus Urb. Repert. Sp. Nov. 13: 450-451, 1914.
(PLATE XXVIII).

Dimorphocladium formosum (Urb.) Britton, Mem. Torr. Bot. Club 16: 74.
1920.

Low shrub (4-6 cm. high ex Shafer), sparsely branching; main stems
terete, stout (c. 5 mm. thick), more or less incrusted by the persistent
stipules, foliage clustered at top. Leaves of main stems (i.e., those sub-
tending branchlets) apparently not reduced to cataphylls: stipules black-
ish, indurate, acicular-lanceolate, 8-11 mm. long, attenuate-acuminate, the
thin scarious lacerate margins more or less deciduous. Leaf-blades nar-
rowed to an ill-defined petiolar base c. 4-7 mm. long, rigidly coriaceous,
spathulate, 20-40 mm. long, 6-9 mm. broad, rounded and emarginate at
the tip (apiculum of young leaf early deciduous), acute at the base; above
olivaceous, sublucid, smooth (minutely foveolate), the impressed midrib
distinct, the lateral veins (c. 10-12 on a side) less prominent; beneath
paler, midrib strongly raised proximally, lateral veins obscure; margins
differentiated beneath, thin and acute, plane. Deciduous branchlets spread-
ing to erect, 8-10 cm. long, 0.8-1 mm. thick, olivaceous, densely papillate-
scabridulous, terete below, somewhat compressed above, with up to 23
leaves; first internode 8-13 mm. long, median internodes 3-5 mm. long.
Leaves: stipules persistent, mostly 3.5-7 mm. long, midrib portions be-
coming blackish and indurate, strikingly dimorphic: one of each pair ob-
long-lanceolate, abruptly acuminate, with broad scarious conspicuously
lacerate margins, the other narrowly linear-lanceolate, attenuate-acumi-
nate, with very narrow entire scarious margins. Petiole 1 mm. long or less.
Leaf-blades as on main stems but much smaller: asymmetrically obovate
or broadly oblong, 6-11 mm. long, 3.5—-5 mm. broad.

Male flower [not seen, description ex Urban]: Calyx-lobes 5, ovate, red-
dish. Stamens 4; filaments (in bud) apparently connate; anthers free,
connective broad, anther-sacs discrete, dehiscing longitudinally.

Female flowers and fruits unknown.

Type: Cuba, Oriente, trail from Camp La Barga to Camp San Benito,
damp thickets, alt. 450-900 m., 22-26 Feb. 1910, Shafer 4102 (NY,
LECTOTYPE). The original holotype in Herb. Krug & Urban (B) presum-
ably has been destroyed.

DISTRIBUTION: known only from the type collection (Map XXVIII).

Although still imperfectly known, P. formosus is obviously a distinctive
and isolated species on the basis of its vegetative characteristics, particu-
larly the large unreduced leaves of the primary axes and the dimorphic
stipules. Technically, the branching of the species could scarcely be called
phyllanthoid in view of the apparent failure of the leaves on the main
axes to become reduced to cataphylls. However, it is possible that such

Jour. ARNOLD Ars. VoL. XX XIX PLaTE XXVIII

Hasir or Phyllanthus formosus Urb. (Shafer 4102 [NY]).

116 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

reduction may be found to occur when additional specimens are collected.
Furthermore, since the branchlets are deciduous and similar to those of
P. comosus and since the reduction of leaves to cataphylls is often very
tardy in that species, it appears that in P. formosus true phyllanthoid
branching occurs but is masked by the reversion of cataphylls to ex-
panded leaves.

The rigid spathulate leaves of P. formosus are so similar to those of
P. comosus that the two species must be rather closely related, although
when better material of P. formosus is available there may prove to be
important distinctions in the reproductive parts. The proposal by Britton
to erect a separate genus Dimorphocladium for P. formosus is quite with-
out merit. The description of this genus was occasioned mainly by the
fact that Britton failed to realize that the distinction between permanent
axes and deciduous branchlets, which he noticed in P. formosus, also oc-
curred in many other species of Phyllanthus.

59. Phyllanthus comosus Urb. Repert. Sp. Nov. 13: 451. 1914.
(PLATE XXIX, figs. A-B).

Roigia comosa (Urb.) Britton, Mem. Torr. Bot. Club 16: 73. 1920.

Glabrous diffusely branching shrub c. 0.5—2 m. high, with short leafy
branchlets often clustered at branch-tips; branches at first smooth, angu-
lar, reddish-brown, waxy, becoming terete, greyish, and with fissured bark
in age, c. 1.5-3 mm. thick. Leaves of branches often unreduced on lower
portions (similar to branchlet leaves), but distally becoming abruptly
reduced to cataphylls. Cataphylls blackish, indurate, more or less per-
sistent: stipules triangular to lanceolate, 0.6-1 mm. long, 0.5—0.6 mm.
broad, acuminate; blade acicular, c. 0.5-0.8 mm. long. Deciduous branch-
lets mostly 0.5-2 (-3) cm. long, 0.3-0.5 mm. thick, stramineous, with
flaky deciduous waxy coating, furrowed, terete, with (3—-) 4-8 (-15)
leaves; first internode 1.5-3 mm. long, median internodes 1.5—-3 mm. long.
Leaves: stipules persistent, triangular-lanceolate, 0.5—0.8 mm. long, 0.25—
0.4 mm. broad, brownish, the scarious tips reflexed. Petiole 0.7-1.5 mm.
long. Leaf-blades rigidly coriaceous, obovate to narrowly spathulate, (7—)
9-16 mm. long, 2.5-5 (—6) mm. broad, obtuse to rounded or slightly emar-
ginate at the apex (rare individual blades with a short inconspicuous
apiculum), narrowly acute at the base; young leaves more or less pinkish-
or purplish-tinged; mature leaf-blade above bright green, subfoveolate
(with a minute subhexagonal reticulum of raised cell walls), midrib slightly
raised; beneath somewhat paler, midrib plane, lateral veins obscure; mar-
gins subdifferentiated, not especially thickened, plane.

Monoecious; flowers mostly solitary, usually only one female flower per
branchlet (at the third or fourth node), other flowers male (occasionally
a male and female flower at the same axil).

Male flower: pedicel capillary, 10-15 mm. long. Calyx pinkish (in
life) ; calyx-lobes 5 or 6, membranous, crystalliferous, rather unequal, ob-
long to spathulate, (3.2—) 3.5-4 mm. long, 1.5-2.5 mm. broad, obtuse or

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 117

rounded at the tip, entire, midrib branching, veins more or less anastomos-
ing. Disk a fleshy more or less 3-angled ring (the 5 or 6 disk-segments
completely coalesced), plane, entire, crimson-colored when fresh. Stamens
5 or 6; filaments connate into a slender sometimes apiculate column
1-1.8 mm. high, less than 0.3 mm. thick; anthers subsessile (free parts of
filaments about as long as or shorter than the anthers), crowded into a
single whorl atop the column (i.e., attached at more or less the same level,
although 2 or 3 may be inner to the others), broadly ovate and definitely
emarginate, c. 0.2-0.3 mm. long, 0.4-0.5 mm. broad; anther-sacs divari-
cate, well-separated on the connective, dehiscing obliquely or horizontally,
pollen grains c. 22—29 » in diameter.

Female flower: pedicel slender, curving, (10—) 12-17 (—21) mm. long,
terete below, angled above, purplish. Calyx pink, becoming darker in fruit;
calyx-lobes 6, chartaceous, subequal (outer ones narrower), elliptic-oblong
to spathulate, 3.5-5 mm. long, 1.8-3 mm. broad, obtuse or rounded at the
tip, entire, midrib pinnately branched, veins more or less anastomosing
(obscure in fruiting calyx). Disk obtusely 5- or 6-angled, plane, crim-
son when fresh, with a thickened narrow entire rim. Ovary nearly sessile,
prominently ribbed along the septa; styles erect, connate below into a
column c. 0.5-0.7 mm. high (or sometimes nearly free), divergent above,
c. 0.8-1.2 mm. high (not counting rolled-up portion), parted c. 4% to %
their lengths; style-branches divergent, entire, circinately revolute, nar-
rowed to the tips.

Capsule oblate, c. 4-4.5 mm. in diameter, smooth, stramineous, not
veiny; valves more or less retained within the marcescent calyx. Seeds
trigonous, symmetric or sometimes slightly asymmetric, 2-2.6 mm. long,
1.3-1.6 mm. radially, 1.1-1.7 mm. tangentially, when mature dark brown
with rows of slightly raised reddish-brown dots; hilum subterminal, elliptic,
c. 0.3 mm. long.

Type: Cuba, Shafer 4242.

DISTRIBUTION: serpentine lowlands and hills, northeastern Cuba (Map
XXVIII).

UBA. OrieNnTE: Cerro de Miraflores, Cananova, Ledn, Clemente, & Howard
20302 (MICH); Ledn 21161 (GH); Webster 3883 (GH, MICH, NY, US);
savanna southeast of Playa de Vaca, Jervis 1650 (GH, MICH); Moa, Mrs.
Bucher 96 ex p. (SV); serpentine hills near mouth of Rio Yamaniguey, Shafer
4242 (NY, LEcToTYPE; F, US, ISOTYPES).

This species, because of the delicate purplish-pink tinging of leaf-mar-
gins and flowers, is one of the most attractive of the West Indian species
of Phyllanthus. Unfortunately, it does not appear to lend itself to cultiva-
tion, for attempts to germinate seeds were entirely unsuccessful. In the
Cerro Miraflores, the species was observed to grow in the low scrub (char-
rascal) on rocky serpentine soil, associated with such characteristic Moa
plants as Dracaena cubensis, Euphorbia helenae, and Scaevola wright.

Although P. comosus is probably most closely related to P. formosus, as
has been discussed under the latter, it in some respects resembles local

118 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

populations of P. orbicularis. Both species were observed growing in the
same localities at Playa de Vaca and Cerro de Miraflores, but never exactly
at the same site (i.e., plants of the two species were never observed to
grow side by side). ie these two areas P. orbicularis appears to occupy
somewhat lower altitudes, but their detailed ecological relationships re-
main to be elucidated. The presence of comosus-like features such as
spathulate leaf-blades on the local specimens of P. orbicularis suggests
transfer of characters between the two species. However, it is also possible
that part, or all, of this similarity may be due to the fact that the eastern
race of P. orbicularis is closely related to P. comosus; this is suggested
by the fact that spathulate leaf-blades occur in P. orbicularis in areas con-
siderably outside of the known range of P. comosus (e.g., the Sierra de
Nipe and Maravi River near Baracoa). The verification of any hypothesis
of the modification of P. orbicularis by “introgression” of P. comosus char-
acters must therefore be tested with the alternative possibility in mind.

60. Phyllanthus orbicularis HBK. Nov. Gen. & Sp. 2: 111-112. pl.
106. 1817; Muell. Arg. in DC. Prodr. 15(2): 331-332. 1866.
(PLATE XXIX, figs. C-D).
Orbicularia phyllanthoides Baill. Etude. Gen. Euphorb. 617.
si al orbicularis a genuinus, 6 ellipticus, et y obovatus Muell. Arg. loc.

be orbicularis (HBK.) O. Ktze. Rev. Gen. 2: 600. 1891.
Phyllanthus rotundifolius Sessé & Moc. Flor. Mex. ed. 2. 212. 1894.
Orbicularia orbicularis (HBK.) Moldenke, Rev. Sudamer. Bot. 6: 178. 1940.

Glabrous shrub c. 0.5-2 m. high, usually diffusely branching, lateral
branches sometimes reduced to short-shoots; branches c. 1.5—3 mm. thick,
reddish brown becoming greyish, bark sometimes breaking into thin plates.
Cataphylls blackish, indurate, not reflexed, more or less persistent: stipules
lanceolate, (1.5—) 2-3 mm. long, acuminate, entire or denticulate; blade
linear-lanceolate, c. 1.5-2 mm. long. Deciduous branchlets sometimes
clustered on spur-shoots, (0.5—) 1-3 (—5) cm. long, 0.3-0.8 mm. thick,
reddish brown, subterete, smooth, with (2—) 3-8 (—11) leaves; first inter-
node (2—) 4-8 (—10) mm. long, median internodes mostly 2—6 mm. long.
Leaves: stipules reflexed, persistent, lanceolate, (1—) 1.5-2 (—2.5) mm.
long, acuminate, scarious to indurate, margins entire. Petioles 0.3—-1 mm.
long. Leaf-blades coriaceous, sometimes very thick and rigid, plane or
convex, broadly elliptic or obovate to suborbicular or obcuneate (some-
times broader than long), c. 5-10 (—12) mm. long, 5-13 mm. broad,
rounded or emarginate at the tip (apiculum obsolete or minute), acute
to rounded or often emarginate at the base; above olivaceous, in age
plumbeous or dark brown and mottled, veins obscure to conspicuously
raised and reticulate, foveolate; beneath yellowish or brownish, midrib
and laterals (c. 4-6 on a side) slightly raised, branching crookedly, the
reticulum often prominent; margins differentiated (with narrow epidermal

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 119

cells) but no thicker than blade, plane or occasionally reflexed but never
revolute.

Monoecious; cymules with mostly 2—5 flowers, male or bisexual, female
flowers one per cymule.

Male flower: pedicel 2—5 (—7) mm. long. Calyx whitish or pink-tinged;
calyx-lobes 6 (rarely 5), membranous or chartaceous, subequal, elliptic
or oblong to obovate or spathulate, (2—) 2.5-4 (—5) mm. long, 1.3-2 mm
broad, obtuse and entire at the tip, midrib sparingly to conspicuously
branched. Disk-segments usually 6, flat or concave, rather fleshy, cuneate
or quadrate, c. 0.3-0.5 mm. broad, closely contiguous or sometimes united
in pairs, dark red (drying brownish). Stamens mostly 6 (less commonly
5 or 7, rarely 3 or 4); filaments (1—) 1.5-2.5 mm. long, unequal (3 an-
thers usually distinctly higher) erect, united usually about halfway into
a column c. 0.8—1.3 mm, long; anthers stipitate in two whorls (free por-
tion of filament usually much longer than anthers), ovate, dehiscing more
or less vertically or obliquely (occasionally lower anthers horizontal or
nearly so), c. 0.2-0.3 mm. long, 0.3-0.5 mm. broad; pollen grains (21-)
23-27 (—29) uw in diameter.

Female flower: pedicel 0.5-2.5 (-—3.5) mm. long. Calyx whitish or
pinkish-tinged; calyx-lobes 6, subequal, scarious or chartaceous, elliptic
or oblong to spathulate, 3-4 (—6) mm. long, 1.5—2.5 (-3.5) mm. broad,
obtuse at the tip, entire, midrib usually conspicuously branched but veins
often obscure. Disk hexagonal, flat, rim slightly thickened, entire, red
(drying brownish). Ovary sessile or slightly stipitate; styles erect and
connate into a column (0.3—) 0.5-0.8 (—1.2) mm. high, free ends erect to
recurved, parted usually c. 4 their length (rarely parted to the column),
tips of style-branches often revolute.

Capsule 3—3.8 mm. in diameter, reddish brown, smooth or slightly rugu-
lose, veins obscure. Columella 1.2-1.8 mm. jone. Seeds trigonous, some-
times asymmetric, (1.6—) 1.8-2.2 (—2.4) mm. long, 1.1-1.4 (—1.6) mm.
radially and tangentially, reddish brown with rows of dark slightly raised
points; hilum elliptic, 0.3 mm. long, micropylar end rarely obscurely carun-

Collected flowering and fruiting throughout the year.

Type: “Insula Cubae prope Regla et Havana,” Humboldt (P, TYPE
COLLECTION).

DISTRIBUTION: thickets, serpentine barrens, Cuba (Maps XXV_ and
XXXVI).

CUBA. Without specific locality, Sessé e¢ al. 4565 (F, probably TYPE COL-
LECTION of P. rotundifolius). PINAR DEL Rfo: San José de Sagua to San Marcos,
Shafer 11969 (F, NY, US); La Cajalbana area, Acuna & Alain 15669 (SV),
Ekman 10474 (S), Leén & Charles 4959 (NY), Webster 4650, 4655 (GH);
Bahia Honda, Ledn (MT), Wilson 9412 (F, NY, US), Wright 1942 ex p. (GOET,
S); Cuabal de Lechuza near San Claudio, Ekman 12988 (S). HABANA: Regla
and Havana, Humboldt (P, TyPE COLLECTION); Madruga, Britton et al. 613
(NY), Ledn 3328 (NY), Van Herman 15328 (SV); Guanabacoa, Loma de la

120 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xxxrx

Jata, Ekman 16535 (S), Shafer & Ledn 12055 (NY). Matanzas: cuabales NW
of Pan de Matanzas, Ekman 16476 (MICH, S); Cuabal del Espinal, east of
Canasi, Leén & Roig 12949 (NY); Tetas de Camarioca, Britton et al. 14078
(F, GH, NY, US). Las Vittas: Motembo, sabana, Ledn 9368 (NY); Santa
Clara, Britton et al. 6187 (NY), Britton & Cowell 13304 (NY, US). CAMAGUEY:
La Ciega, Caobilla, Acufa 13518 (US); hills near Camagiiey, Britton et al.
13232 (F, GH, NY, US); Camagiiey to Santayana, Britton 2416 (F, NY, ee
Sabana de la Matanzas, Roig, Luaces, & Arango (SV). ORIENTE: Holguin

base of Cerro de Fraile, Ekman 3223 (S); Holguin, Aguasclaras, Ekman 71660

Mar XXVI. Distribution of P. orbicularis HBK.

(MICH, S); base of Loma Pilon, near Holguin, Shafer 1213 (F, NY, US);
Holguin, Holguin to Mayari, Wright 1942 ex p. (GH, GOET); Sierra de Nipe,
along creek 5 km. south of Woodfred, Howard 6132 (GH, MICH, NY, US);
Mayari, Rio Miguel, Alain et al. 5845 (GH); Cananova, Cerro de Miraflores,
Marie-Victorin et al. 21493 (A, MT), Webster 3884 (GH, MICH, NY, US);
Playa de Ja Vaca, Acuna 12505, 13159 (SV), Clement 3655 (GH, MT, US),
Marie-Victorin et al. 21475, 21747 (A, MT), Webster 3867 (GH, MICH, NY,
US); Moa, Mrs. Bucher 95, 96 ex p. (NY, SV); plain between Moa and Ya-
guaneque, Ledn et al. 20289 (NY); hills near mouth of Rio Yamaniguey, Shafer
4246 (F, NY, US); pinelands near sea-shore, Rio Maravi, near Baracoa, Ekman
4032 (S)

PLATE XXIX. FLoweErs oF sect. Orbicularia.

Fics. A-B. Phyllanthus comosus Urb. A, androecium and male calyx lobe;
B, gynoecium and female calyx lobe (Jervis 1650 [GH]). Fics. C-D. Male and
female flowers of P. orbicularis HBK. (Webster 4650 [GH]). Fics. E-F.
Male and female flowers of P. myrtilloides Griseb. ssp. myrtilloides (Alain 3073
[GH]). Fics. G-H. Gynoecium and androecium of P. myrtilloides ssp.
alainit Webster (Alain 5563 USE). Fics. I-J. Phyllanthus myrtilloides ssp.
spathulifolius (Griseb.) Webster. I, androecium and male calyx-lobe; J, gy-
noecium and female calyx-lobe (Webster 4897 [GH]). Fics. K-L. Male and
female flowers of P. chamaecristoides ssp. baracoensis (Urb.) Webster (Ekman
4326 [Sj). Fics. M-N. Male and female flowers of P. scopulorum (Britton)
Urb. (Webster 3800 [GH]). Fics. O-P. Androecium and gynoecium of P.
nummularioides Muell. Arg. (Allard 16066 [US]). Fires. Q-R. Male and female
flowers of P. phlebocarpus Urb. (Carabia 3573 [GH]).

Jour. ARNOLD Ars. VoL. XXXIX PLATE XXIX

WEBSTER, WEST INDIAN PHYLLANTHUS

es JOURNAL OF THE ARNOLD ARBORETUM _ [VoL, XxxIx

Phyllanthus orbicularis is the commonest and most widespread woody
species of the genus in Cuba; it occurs always in relatively dry serpentine
areas in scrub thickets, savannas, or palm barrens. In its original publica-
tion the species was als said to occur in “opacatis Orinoci prope Cari-
chana,” but this record remains unconfirmed and is surely an error; Hum-
boldt must have confused the Cuban plant with a South American species
of sect. Microglochidion. It is apparent, even on superficial inspection,
that the species contains a number of strikingly divergent local popula-
tions. The most outstanding of these is represented by Kkman 7660 from
the vicinity of Holguin, which has much larger flowers than plants from
all other localities. However, there is also a notable difference between
plants of the eastern and western parts of the islands, so that it is possible
to recognize two major subspecific races: a western race characterized
by smaller male flowers, branchlets with more leaves, and thinner leaves
with the nerves conspicuous above; and an eastern race with larger male
flowers, shorter few-leaved bancblets: and thick rigid leaves with the
nerves obscure above. The boundary between the two races would be
drawn west of the Sierra de Nipe, and if the differences were really sharp
two subspecies could be defined. It does not seem worth-while to recog-
nize any formal subspecific categories, however, in view of the impossibil-
ity of defining two natural subspecies which could be separated by a key
In the case of the number of leaves per branchlet, there appears to be a
cline running-from Pinar del Rio, where the mode is commonly 6 to 8, to
the Moa region of Oriente, where it is usually 4 to 5. Unfortunately the
number of samples with male flowers is too small to demonstrate whether
there is a cline in this character too, but it seems not unlikely.

There appears to be a considerable amount of fluctuating or random
variability within P. orbicularis, particularly with regard to stamen num-
ber and stylar configuration. In the original description of the species the
stamen number was given as 4-10, and Mueller noted it as 6-9, rarely to 12.
These reports have not been confirmed, for although a few flowers with
only 4 (or very rarely 3) stamens have been observed, none have been
seen with more than 7. In view of the large range of variation in the
species, it is not impossible that Nowers with 10, or even 12, stamens occur
but it seems more likely that the observations in the literature are erro-
neous.

There is yet another factor which must be considered in attempting
to account for the variation patterns shown by P. orbicularis. It is possible
that the distinctive characters of the eartern race of the species may be
partially the result of hybridization with P. comosus. The obovate leaves
with acute bases and the spathulate calyx-lobes of the Moa plants of
P. orbicularis strongly suggest the influence of P. comosus. In the Cerro de
Miraflores both species were seen growing together in the scrubland, where
there certainly does not appear to be any ecological barrier to their cross-
ing. However, no plants with intermediate characters were observed and
it could not be determined whether or not hybridization is occuring at the
present time.

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 125

61. Phyllanthus myrtilloides Griseb. Mem. Amer. Acad. Sci. 8: 158.
1860.

Glabrous bushy shrub or small tree c. 0.5—4 m. high; branches of cur-
rent year straight, slender (c. 1-3 mm. thick), smooth, terete, sometimes
furrowed or cracking open, dark brown or greyish. Cataphylls blackish,
indurate, often reflexed, more or less deciduous: stipules triangular to
lanceolate, mostly 1-3 mm. long, acute to acuminate, entire or denticulate;
blade linear-lanceolate, c. 0.5-1.5 mm. long. Deciduous branchlets spread-
ing or ascending, (2—) 3-8 (—11) cm. long, 0.3—-0.8 mm. thick, stramineous
or olivaceous, subterete, more or less smooth, with mostly 8-18 (-25)
leaves; first internode mostly 2-12 mm. long, median internodes mostly
2-10 mm. long. Leaves: stipules reflexed, subpersistent or often mostly
deciduous, subentire, lanceolate, acute or acuminate, the proximal ones
partly or entirely blackish and indurate, c. 1-3 mm. long and 0.4—0.8 mm.
broad, the distal ones brownish and scarious, c. 0.5—1 mm. long. Petioles
mostly 1-2.5 mm. long. Leaf-blades chartaceous to rigidly coriaceous,
spathulate to obovate or suborbicular, mostly 8-25 mm. long and 4-15 mm.
broad, obtuse to rounded or sometimes emarginate at the tip, the scarious
apiculum obsolete or up to c. 1 mm. long, acute to rounded or rarely sub-
cordate at the base; above olivaceous or plumbeous, dull to sublucid (or
sometimes polished in age), veins plane or slightly raised, obscure to sub-
prominent (veinlets obscure); beneath yellowish or whitish, midrib, main
lateral veins (c. 3-5 on a side) and sometimes the veinlets slightly raised,
reticulum subprominent to obscure; margins plane to conspicuously revo-
ute.

Monoecious; proximal cymules usually with 2 or 3 male flowers, some
distal cymules unisexual with 1 female and 1-3 male flowers, or flowers
subsolitary (rarely 2 female flowers per axil).

Male flower: pedicel mostly 7-15 (—25) mm. long. Calyx whitish or
pinkish-tinged; calyx-lobes 6 (rarely 5), membranous or somewhat fleshy,
subequal or sometimes quite unequal, the outer lobes oblong to obovate,
the inner ovate to spathulate, c. 1.7-3 mm. long, obtuse or rounded and
entire to denticulate at the tip, midrib branched or unbranched. Disk-
segments usually 6, flat, thin or somewhat fleshy, roundish or squarish,
free or occasionally connate, mostly 0.4-0.5 mm. across. Stamens 6 (rarely
5, very rarely 4); filaments connate into a column 0.7-2 mm. long; anthers
sessile to stipitate (the free distal portions of filaments up to 0.8 mm.
long), broadly ovate, mostly 0.25—0.4 mm. long and 0.3-0.5 mm. broad;
anther-sacs divergent, the upper (inner) dehiscing more or less vertically,
the lower (outer) dehiscing more or less horizontally, slits apically con-
tiguous but not confluent; pollen grains c. 18-26 m in diameter.

Female flower: pedicel slender, c. 3-15 (—20) mm. long. Calyx whitish,
ereenish, or sometimes pinkish-tinged; calyx-lobes 6 (rarely 5), scarious
to subcoriaceous, subequal, elliptic or oblong to obovate (inner lobes often
broader) ; larger lobes c. 3-4 (—4.5) mm. long, mostly 1.5-2.5 mm. broad,
obtuse or rounded at the tip, entire or denticulate, midrib usually more or

124 JOURNAL OF THE ARNOLD ARBORETUM _ |[voL, xxxrIx

less branched but veins often obscure. Disk plane, tenuous to rather mas-
sive, more or less angled, sometimes inconspicuously pitted. Ovary sessile,
smooth, 3-sulcate, ribless or inconspicuously carinate; styles united c. /%
their length into a column c. 0.5—-1.5 mm. high, free ends bifid or parted
to the column, style-branches tapering to slender more or less revolute
tips.
Capsule oblate, c. 2-2.7 mm. high, 3-4 mm. in diameter, dark reddish
brown, veins subprominent or completely obscure. Columella c. 1.5—2 mm
long. Seeds trigonous, sometimes slightly asymmetric (one face carinate),
1.5-2.2 mm. long, 1-1.5 mm. radially and tangentially, reddish brown
becoming fuscous, with regular or irregular rows of dark slightly raised
points; hilum subterminal, elliptic to triangular, c. 0.3-0.4 mm. long,
micropylar end sometimes (in ssp. spathulifolius) carunculate.

In the greatly enlarged circumscription here adopted, P. myrtilloides
is a variable polytypic species common and widespread in the serpentine
lands of northern Oriente province, Cuba. It may appear excessively con-
servative to combine such different plants as P. spathulifolius, with
chartaceous narrowly obovate revolute leaves and whitish or greenish
flowers, and P. erythrinus with plane nearly orbicular coriaceous leaves
and at least the floral disk dark reddish or purplish. It must be admitted
that — due to insufficient sampling —clear zones of intergradation be-
tween the five taxa have not been demonstrated, and indeed in most cases
ae aa specimens have not been observed. However, the five taxa

n the P. myrtilloides complex have in toto neatly allopatric and adjoining
ranges, and display varying combinations of essentially a single overall
pattern of characteristics. Although these taxa could still be maintained
as five closely related species, their degree of relationship is so close that it
seems more realistic to regard them as five unusually distinctive subspecies.

KEY TO THE SUBSPECIES

1. Anthers all stipitate, dehiscing vertically or obliquely; fruiting pedicel c. 3-7
(rarely to 9) mm. long; leaves prominently apiculate (at least when young),
HOt OF SCATCEly TEVOlUGs 2 v4 246 oe Aa Kea ssp. myrtilloides (61b)

. Anthers not all stipitate, at least the lower whorl sessile or subsessile and
dehiscing more or less Seep fruiting pedicel mostly 8-20 (rarely as
low as 5) mm. long; leave

—

2. Style- branches ee ae recurving from the top of the stylar
column; leaves plane, not revolute. ................ ssp. alainii (61c)
2. Style-branches not auriculate; leaves various.

3. Young leaves and floral disk dark reddish or purplish; ioe neither

strongly apiculate nor revolute; staminal column mostl

long. ssp. po ee (61a)

Young leaves and floral disk greenish or at least not purplish; leaves

strongly revolute.

4. Cataphylls reflexed; staminal column 0.5—1 mm. long; larger calyx-
lobes of male flower c. 2 mm. long, midrib unbranched; seeds
Pe i. WONG s- bic5 ss oo hae Sede wc eed ssp. shaferi (61d)

ww

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS WAR)

4. Cataphylls not reflexed; staminal column 1-1.7 mm. long; large
calyx-lobes of male flower c. 2.5 mm. long or more, midrib
branched; seeds 1.6-1.8 mm. long. ...... ney nian (61e)

6la. Phyllanthus myrtilloides ssp. erythrinus (Muell. Arg.), stat.
nov.
Phyllanthus purpureus Wright ex Griseb. Goett. Nachr. 1865: 168. 1865; non
P. purpureus Muell. Arg., 1864.

Phyllanthus erythrinus Muell. Arg. in DC. Prodr. 15(2): 332. 1866.
Diasperus erythrinus (Muell. Arg.) O. Ktze. Rev. Gen. 2: 599, 1891.
Orbicularia foveolata Britton, Mem. Torr. Bot. Club 16: 73. 1920.
Phyllanthus cardiophyllus Urb. a Ant. 9: 190. 1924.
Phyllanthus melanodiscus Urb. :
Phyllanthus foveolatus (Britton) Alain, Contr. Occ. Mus. La Salle 11: 2. 1952.

Shrub 0.5-3 m. high; cataphylls not reflexed, stipules 1-2 mm. long,
the tip deciduous, the blade c. I1-1.5 mm. long. Branchlets (2.5—) 4-8
(-11) cm. long, often bright yellow proximally, with (6—) 8-17 (-25)
leaves; first internode (4~-) 6-12 (-15) mm. long, median internodes
4-12 mm. long. Leaves: proximal stipules c. 1-1.5 mm. long, distal stip-
ules c. 0.5—0.8 mm. long; petioles c. 1.5-2.5 (-3.5) mm. long; leaf-blades
chartaceous to coriaceous, conspicuously purplish when young, mostly
broadly elliptic to suborbicular, mostly 13-25 mm. long and 10-18 mm
broad, the apiculum very short (0.3 mm. long or less) or obsolete; mar-
gins plane or reflexed but never definitely revolute.

Male flower: pedicel 10-20 mm. long; calyx more or less purplish-
tinged; calyx-lobes mostly 2.5-3 (-3.5) mm. long, 1.8-2 mm. broad,
midrib usually conspicuously branched; stamens 6, staminal column (1—)
1.5-2 mm. long; anthers substipitate (free portions of filaments about as
long as to shorter than the anthers), dehiscing vertically or obliquely.
Female flower: pedicel (5—) 10-20 mm. long; calyx-lobes (3—) 3.5-4 mm.
long, 1.5—2.5 mm. broad, midrib more or less branched; disk flat, colored
dark reddish or See styles erect, coherent or usually connate into
a column c. 0.5—0.8 mm. high, free ends parted c. % their length. Seeds
1.6-1.8 mm. long, oe ie mm, radially and tangentially.

Collected in flower Mar., June, July, Nov.—Dec.; in fruit, July, Dec.

Type: Cuba, Oriente, Wright 1943.

DISTRIBUTION: scrub or pineland, mountainous serpentine areas, eastern
Cuba (Map XXVIII).

CUBA. ORIENTE: Sierra de Moa, alt. 800 m., Alain 3414 (GH); Baracoa, El
Yunque, north side, Ekman 3546 (S, TYPE COLLECTION of P. cardiophyllus) ; be-
tween Taco and Nibujon, charrascales, pinales, Ekman 3724 (S, TYPE COLLECTION
of P. melanodiscus); pinelands, road between Baracoa and Florida, Ekman 3989
(S); Lomas de Cuaba [Duaba], Ekman 4265 (S); upper valley of Rio Navas,
thickets by river, Shafer 4412 (GH, NY); Loma Santa Teresa, near El] Yunque,

126 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

Shafer 7734 (NY), 7736 (F, GH, NY); Camp La Gloria, south of Sierra Moa,
Shafer 8271 (NY, HoLotyPE of Orbicularia foveolata); pine woods, Baracoa,
Underwood & Earle 1349 (NY); Sierra de Moa, c. 20 km. south of Moa lumber
mill, Webster 3899 (GH, MICH, NY, US); pinal near Baracoa, Wright 1943
(GOET, HOLOTYPE; G, GH, ISOTYPES). A sterile specimen from Charrascos de
Pena Prieta, Toa, Alain 3597 (GH), is probably referable to this subspecies.

Of the five subspecies of P. myrtilloides, the present one appears to be
the most widespread and also the most xerophilous. Although it has been
collected in river thickets (Shafer 4412), it seems to be mainly a plant of
open pinelands, so that ecologically it corresponds more to P. baracoensis
than to any of the other subspecies of P. myrtilloides. The purplish color
of the unfolding leaves and floral disk is most conspicuous in the living
plant and at once distinguishes it so well that in the field it might be taken
for a distinct species. However, in almost all respects except pigmentation
ssp. erythrinus is very similar to ssp. myrtilloides, although it can be dis-
tinguished by its shorter stipules, non-apiculate leaves, and usually longer
fruiting pedicel.

Despite its rather formidable synonymy, ssp. erythrinus does not appear
to be an unusually polymorphic taxon. The two species proposed by Urban,
P. cardiophyllus and P. melanodiscus, are trivial variants. Alain (Flora
de Cuba 3: 51, 1953) has already reduced the latter, and the supposed
difference in leaf-shape of P. cardiophyllus is quite inconsequential. In
contrast, Britton’s Orbicularia foveolata (which Alain has transferred to
Phyllanthus) cannot be disposed of so easily. The holotype (Shafer 8271)
is a nearly sterile specimen with strikingly shiny coriaceous suborbicular
leaves which are conspicuously foveolate due to the prominent anticlinal
walls of the hexagonal epidermal cells. Shafer noted on the label that the
flowers were white, which would be unusual for the usually pinkish-flowered
plants of ssp. erythrinus; unfortunately, the floral material is too inade-
quate to provide the necessary additional information as to floral charac-
ters. However, the glossy leaves which are supposedly the main distin-
guishing feature of P. foveolatus can be matched closely by those on older
branches of a plant encountered in the Sierra Moa (Webster 3899) prob-
ably not far from Shafer’s original locality. It thus appears that the pro-
posed species P. foveolatus was founded on a specimen which differs from
typical ssp. erythrinus only in its ontogenetic stage (and possibly also in
its development in a drier environment). Foveolate-leaved specimens may
be expected throughout the range of the subspecies.

61b. Phylanthus myrtilloides ssp. myrtilloides
(PLATE XXIX, figs. E-F).
Phyllanthus myrtilloides Griseb. Mem. Amer. Acad. Sci. 8: 158. 1860.
Diasperus myrtilloides (Griseb.) O. Ktze. Rev. Gen. 2: 600. 1891.
Bush or small tree c. 2-4 m. high; cataphylls more or less reflexed,
stipules mostly 1-1.5 mm. long, blade c. 1 mm. long. Branchlets (3-)

—

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 127

5-11 cm. long, with 10-15 (—20) leaves; first internode 2-7 mm. long,
median internodes 3-7 mm. long. Leaves: proximal stipules 1.5—3 mm.
long, distal stipules c. 1 mm. long or less; petioles c. 1-2.5 mm, long; leaf-
blades chartaceous or subcoriaceous, elliptic to suborbicular, mostly 10-20
(—27) mm. long and 7-14 (-17) mm. broad, obtuse to rounded or some-
times emarginate and, when young, conspicuously apiculate at the tip
(apiculum c. 0.5-1 mm. long, more or less deciduous in age), acute to
obtuse at the base, margins plane or reflexed (or at most slightly and
casually revolute).

Male flower: pedicel 5-12 mm. long; calyx whitish (drying reddish),
calyx-lobes c. 2-3.2 mm. long, the outer elliptic-oblong and c. 0.8— ays cahaate
broad, the inner obovate and c. 1.2—1.8 mm. broad, midrib unbranched or
nearly so; disk-segments free; stamens 6 (rarely 5), staminal column
mostly 1-1.5 (-1.8) mm. long, anthers usually stipitate (free portion of
filaments definitely longer than at least the inner anthers), dehiscing
vertically or obliquely. Female flower: pedicel 3-7 (-9) mm. long; calyx
greenish white (drying reddish), calyx-lobes mostly 3-4 (-4.5) mm. long
and 1.5-2.5 mm. broad, midrib unbranched or nearly so; styles erect and
connate into a column c. 0.5-1 mm. high, the free ends bifid c. }2 their
length. Seeds c. 1.5-2 mm. long, 1-1.3 mm. radially and tangentially.

Tyre: Cuba, Oriente, Wright 1435 ex p.

DISTRIBUTION: rain-forest, mountains of Oriente province, Cuba (Map
XXVII).

CUBA. ORIENTE: wet woods near Palenquito, Yateras region, alt. 500 m.,
20 July 1953, Alain 3073 (GH); border of Laguna del Galano, Toa, alt. 900 m..,
2 Jan. 1954, Alain 3837 (GH); woods, Monteverde, alt. 800 m., Apr. 1889,
Eggers 5110 (F, GOET, US); La Prenda, north of Guantanamo, 22 July 1921,
Hioram 4775 (MICH); Monte Verde, 23 Dec. [1858], Wright 1438 ex p.
(GOET, HoLotyPE; BR, G, GH, NY, S, W, ISOTYPES ).

In its usual phase, as in the Monteverde region, ssp. myrtilloides is dis-
tinguishable from the remainder of the species by the definitely stipitate
anthers, the free portions of the filaments approaching the elongated condi-
tion as in P. orbicularis. The collection from the lake at the top of Monte
Galano (Alain 3837) is quite atypical in having small male flowers with
the anthers nearly sessile on the staminal column. However, since it agrees
with ssp. myrtilloides in its reflexed cataphylls and short fruiting pedicel,
it is classified here provisionally.

The present subspecies is certainly closely related to ssp. erythrinus,
but clearly differs in its apiculate leaves, usually reflexed cataphylls, and
lack of purplish coloration in leaves or flowers; furthermore, ssp. myrtil-
loides appears to be a more mesophytic plant. However, it is not certain
that these variable morphological characters will still provide valid dis-
tinctions when additional collections become available. Another closely
related plant is ssp. alainii, which is vegetatively similar but differs in its
auriculate styles and shorter filaments.

128 JOURNAL OF THE ARNOLD ARBORETUM _[vot. xxx1x

MAP XXVII P MYRTILLOIDES
x alainii

Es @ erythrinus
- ? * myrtilloides
fe Z Lf MEE™D o shaferi
| ; 7 aL
/ LG

Sy)

© spathulifolius

: Ney i
o
- Oy ee * f sey
¥ .'
S)

20

P- CHAMAECRISTOIDES

/ AN : a es
/ ) a

7 |

J Vics } ) | é

2 em EROS Bis —
b6 ;

7

MAP XXVIII a @ baracoensis
2 © chamaecristoides
tf 2 x P COMOSUS
f © P FORMOSUS
| be Me pf FLED
5 WKY i ss
1 Sy AN 5 6 " [,

btg

MAP XXIX

oO P PHLEBOCARPUS
@ P. SCOPULORUM

ti

Mars XXVII-XXIX. Distribution of some species of sect. Orbicularia.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 129

6lc. Phyllanthus myrtilloides ssp. alainii, ssp. nov.”*
(PLATE XXIX, figs. G-H).

Shrub; cataphylls more or less spreading (not aang stipules c. 1.5
mm. long, with deciduous tips, the blade c. 1.5 mm. long. Branchlets 4-8
cm. long, with c. 9-15 leaves; first internode ae mm. long, median inter-
nodes 3-8 mm. long. Leaves: proximal stipules 2—3.5 mm. long, distal
stipules c. 1 mm. long; petioles c. 1.5—2.5 mm. long; leaf-blades char-
taceous, broadly elliptic or obovate, c. 10-25 mm. long and 8-15 mm.
broad, the apiculum obsolete or nearly so, acute or obtuse at the base, mar-
gins plane or recurved.

Male flower: pedicel c. 10-15 mm. long; inner (larger) calyx-lobes c.
2.5-3 mm. long and 1.8—2.7 mm. broad, midrib sparsely branched; stamens
6 (rarely 4 or 5), staminal column 0.7-1.1 mm. high, anthers sessile or
subsessile atop the column, the upper dehiscing vertically, the lower more
or less horizontally. Female flower: pedicel 7-15 mm. long; calyx-lobes
mostly 3.5-4 mm. long and 2-3 mm. broad, venation obscure; disk flat,
tenuous; styles erect and connate up to the bifurcation point into a column
c. 1 mm. long, style-branches reflexed from top of column, adaxially
auriculate, less than 1 mm. long, revolute at the tips. Seeds c. 2.1 mm.
long, 1.4 mm. radially and tangentially.

Type: Cuba, Oriente, Alain et al. 5475.
DISTRIBUTION: Sierra Cristal, Oriente province, Cuba (Map XXVII).

CUBA. OrieEnTE. Sierra Cristal: southern ee Ee ee 1956, Alain, Acuna,
& Lépez Figuieras 5475 (GH, HOLOTYPE), 5478, 9 (GH); shore of Arroyo
Cristal, 2-7 Apr. 1956, Alain, Acuna, & Lopez Panera 5633 (GH); manacales
at the headwaters of the Ric Teepicn alt. 600-700 m., 14 Dec. 1922, Ekman
15939 (S; sterile).

This apparently not uncommon plant of the Sierra Cristal vegetatively
resembles ssp. shaferi and ssp. spathulifolius and is, in fact, quite closely
allied to the former subspecies. However, its auriculate style-branches,
non-revolute leaves and large male flowers are in sum so distinctive that
it merits recognition as a distinct entity. There is also a definite vegetative
resemblance between ssp. alainii and ssp. myrtilloides but the leaves of
ssp. alainii are not apiculate, the female pedicel is usually longer, and the
androecium of completely connate filaments is quite different from that of
ssp. myrtilloides.

It seems appropriate to name this plant for Brother Alain, the assiduous
botanist of the Colegio de la Salle, Vedado, Habana. His recent collections

* Phyllanthus myrtilloides ssp. alainii, ssp. no

Frutex stipulis cataphyllorum patentibus 1.5 mm. ais ngis; ramulis 9—15-foliatis;
foliis chartaceis ellipticis or aaa 10-25 mm. ane nec apiculatis nec revolutis; flore
masculo laciniis calycis 2.5-3 . longis, columna staminum 0.7-1.1 mm., am thieeis
sessilibus; flore seers pedicel is 15 mm. longo, ‘Taci iniis calycis plerumque 3.5-4 mm.
longis, columna stylor .1mm., stigmatibus auriculatis, reflexibus, revolutis; semina
c. 2.1 mm. longa.

130 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxix

of several species of sect. Orbicularia have greatly assisted in analyzing
this taxonomically difficult group.

old. Phyllanthus myrtilloides ssp. shaferi (Urb.), stat. nov.
Phyllanthus shaferi Urb. Repert. Sp. Nov. 13: 448. 1914.

Shrub 1-3 m. high; cataphylls usually reflexed, stipules c. 1-1.5 mm.
long, blade up to 1.5 mm. long but often greatly reduced. Branchlets 2-6
cm, long, with c. 10-16 (—18) leaves; first internode 1.5—-4 mm. long,
median internodes 2-6 mm. long. Leaves: proximal stipules c. 1-2.2 mm
long, distal stipules c. 0.5-1.3 mm. long; petioles c. 0.8-1.5 (—2.5) mm.
long; leaf-blades chartaceous often becoming convex and _ coriaceous,
broadly elliptic to obovate or sometimes suborbicular, c. 6-10 (—15) mm.
long, (3—) 4-8 (-10) mm. broad, rounded at the tip, the conspicuous
apiculum (of juvenile leaves) deciduous, margins usually conspicuously
revolute.

Male flower: pedicel 6-15 mm. long; calyx-lobes mostly 1.7-2.1 mm.
long and about as broad, midrib branched or unbranched; disk-segments
rather massive, free; stamens 6 (rarely 5), staminal column 0.5—-1 mm.
high, anthers sessile on the column, upper dehiscing vertically, lower hori-
zontally. Female flower: pedicel 7-15 mm. long; calyx-lobes c. 3-4 mm.
long and 1.7—2.5 mm. broad, midrib branched but veins obscure; disk
rather massive; styles united very nearly to the bifurcation point into a
column (0.5—) 1-1.5 mm. high, style-branches sharply reflexed from the
top of the column, the tips revolute. Seeds 1.9-2.2 mm. long, 1.2-1.5 mm.
radially and tangentially.

Type; Cuba, Oriente, Shafer 1715.
DISTRIBUTION: serpentine regions, Sierra de Nipe, Cuba (Map XXVIII).

CUBA. OrtENnTE. Sierra de Nipe: Cayo del Rey, Arroyo Canapu, Apr., May
1940, Carabia 3603, 4086 (GH, MICH, NY); pinar de Mayari, Carabia 3626
(GH, MICH, NY); Salto del Sojo, Carabia 3723 (GH, NY); along Rio Piedra
in pinelands, alt. c. 500 m., 3 July 1914, Ekman 1786 (MICH, S); charrascales,
at stream, 25 July 1914, Bae 2196 (S): Bayate prope Rio Piedra, 18 F 1915,
Ekman 4671 (S, sterile): ad marginem Rio Piloto, 20 Apr. 1919, Ekman 9501
(S); La Planch, July 1941, Howard 6181 (GH, MT, NY, US); pinelands,
La Casimba, 27 July 1940, Leon & Alain 19235 (MICH): pindlands ae of
Sierra Nipe, alt. 600-700 m., 16-18 Oct. 1941, Morton & Acuna 3110 (US);
near streamlet, pinales southeast of Paso Estancia, 1-2 May 1909, ae i715
(NY, LeEcToTYPE; F, isorype); Arroyo del Medio, above the falls, alt. 450-
550 m., 22 Dec. 1909, Shafer 3266 (F, GH, NY, US). Sierra Cristal: border of
Rio Miguel, Mayari, 2-7 Apr. 1956, Alain, Acuna & Lopes Figuetras 5940 (GH).

In the Sierra de Nipe ssp. shaferi is a very common plant along water-
courses. As illustrated by Marie-Victorin (Contr. Inst. Bot. Univ. Montr.
68: 62. 1956), it has a distinctive aspect with small convex revolute leaves
and pendent long-pedicellate flowers. However, it is very closely related
to ssp. alainii, and the auriculate style-branches and non-revolute leaves

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 131

of the latter are the only really distinctive characters. Also similar to
ssp. shaferi is ssp. spathulifolius of the Moa region, but that plant has
narrower leaves, longer staminal column, and different styles.

6le. Phyllanthus myrtilloides ssp. spathulifolius (Griseb.), stat.
nov. (PLATE XXIX, figs. J-J).
Phyllanthus spathulifolius Griseb. Goett. Nachr. 1865: 169. 1865; Muell.
in DC. Prodr. 15(2): 332. 1866; Alain, Flora de Cuba 3: fig. 9. 1953.
Diasperus spathulifolius (Griseb.) O. Ktze. Rev. Gen. 2: 601. 1891.
Phyllanthus myrtilloides 6 spathulifolus (Griseb.) Gomez de la Maza, Anal.
Soc. Hist. Nat. Madrid 23: 53. 1894

Shrub 0.5—3 m. high; cataphylls deciduous, not reflexed, stipules (0.6—)
1-1.5 mm. long, blade c. 0.5-0.8 mm. long. Branchlets (2—) 3-6 (-8)
cm. long, with mostly 7-12 (—15) leaves; first internode c. 2-5. mm. long,
median internodes 4-8 mm. long. Leaves: proximal stipules c. 0.7-1.5
mm. long, distal stipules c. 0.4-0.7 mm. long; petioles c. 1-1.5 mm. long;
leaf-blades chartaceous, obovate or spathulate, 7-13 mm. long and 5-9
mm. broad, obtuse or rounded at the tip (apiculum nearly obsolete even
in young leaves), acute at the base, margins usually distinctly revolute.

Male flower: pedicel 8-16 mm. long; calyx whitish; calyx-lobes mostly
2-3 mm. long and 0.7-1.8 mm. broad, midrib distinctly branched; disk-
segments free; stamens 6 (rarely 5), staminal column (1—) 1.2—1.5 (-1.7)
mm. high, anthers substipitate to sessile (free portion of filament at most
about as long as or slightly longer than anther). Female flower: pedicel
(6-) 8-13 mm. long; calyx greenish white; calyx-lobes mostly 3—3.5 mm.
long, 1.2-2 mm. broad, midrib branched but veins inconspicuous; styles
erect, coherent or connate into a column c. 0.5 mm. high, free ends parted
c. % their length, the style-branches divergent, tips revolute. Seeds 1.6—
1.8 mm. long, 1.1-1.3 mm. radially and tangentially, often carunculate.

Type: Cuba, Oriente, Wright 1438).

DISTRIBUTION: riparian woods, northern Oriente province, Cuba. (Map
XXVITI).

CUBA. OrrentE: Monte Grande de Centeno, south of Moa, 4 Aug. 1945,
Leon, Alain, & Clemente 22673 (MICH); bank of Rio Moa, c. 20 km. south of
Moa, 21 July 1951, Webster 3896 (GH, MICH, NY, US), 3897 (GH, MICH);
margin of Rio Castro near Sagua de Tanamo, 3 Apr. 1861, Wright 1438b (GOET,
HOLOTYPE; G, S, W, ISOTYPES).

In many respects, including its general aspect, ssp. spathulifolius is a
very distinctive group and might be thought to merit specific distinction,
although Gomez de la Maza long ago reduced it to a variety of P. myrtil-
loides. Vegetatively ssp. spathulifolius is not unlike ssp. alainii of the
Sierra Cristal but its styles are much more similar to those of ssp. myrtil-
loides. In the Sierra Moa, ssp. spathulifolius was observed growing with
Exostema in the bed of the Rio Moa within a few miles of a small popula-
tion of plants of ssp. erythrinus. Here the two plants appear so different

132 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxx1x

from one another, both morphologically and ecologically, that they gave
the impression of being distinct species. However, when the other related
populations in Oriente are taken into consideration, they are seen to be
merely at the opposite ends of a nearly continuous spectrum.

The locality cited for Wright’s collection has been taken from a label
on an isotype sheet of ssp. myrtilloides (GH). There is no material of
the type collection of ssp. spathulifolius in the Gray Herbarium (due,
presumably, to the vagaries of distribution), but circumstantial evidence
suggests that the label cited above belongs with specimens of ssp. spathuli-
olius.

62. Phyllanthus chamaecristoides Urb. Symb. Ant. 9: 185-186, 1924,

Low shrub c. 1-2 m. high; branches straight, c. 1-4 mm. thick, terete,
dark brown, smooth or furrowed. Cataphylls blackish, indurate, some-
times reflexed, glabrous or puberulent, deciduous: stipules lanceolate,
c. 1.5-3 mm. long, acuminate, entire or denticulate; blade narrowly lanceo-
late, c. 1-1.5 mm. long, acuminate, entire. Deciduous branchlets ascend-
ing or spreading, c. 3-7 (—13) cm. long, 0.25-0.5 mm. thick, stramineous
to reddish brown, subterete, with mostly 15-45 leaves; first internode
1-5 mm. long, median internodes 1-4 mm. long. Leaves: stipules reflexed,
persistent or deciduous, triangular-lanceolate, acuminate, the proximal
ones partly or entirely blackish and indurate, 1-2.5 mm. long, the distal
ones brownish and scarious, 0.5-1 mm. long. Petioles 0.3—1.2 mm. long.
Leaf-blades brittle-chartaceous to subcoriaceous, sometimes convex, ellip-
tic or oblong (and sometimes falcate) to obovate or spathulate, 3-10 mm.
long, 2-5 (—7) mm. broad, subacute to obtuse or rounded and obscurely
to prominently apiculate at the tip, acute to obtuse (sometimes inequi-
lateral) at the base; above olivaceous or plumbeous, dull or sublucid
(sometimes glossy in age), midrib and lateral veins slightly raised, sub-
prominent to obscure; beneath pale, sometimes whitish or yellowish, mid-
rib and steeply ascending lateral veins (c. 5—7 on a side) slightly raised,
prominent to obscure, lateral veins nearly or quite unbranched; margins
plane, reflexed, or revolute.

Monoecious; flowers mostly solitary, the proximal male, occasional
distal ones female; male flowers sometimes paired.

Male flower: pedicel 2.5—-7 mm. long. Calyx greenish white (more or
less reddish when dried) ; calyx-lobes normally 6, more or less membranous,
biseriate, c. 1.5—-2 mm. long, the outer lobes elliptic or oblong, 0.8—1.3 mm
broad, the inner lobes broadly elliptic to obovate, 1.1—1.7 mm. broad; lobes
obtuse or rounded and entire or minutely denticulate at the tip, midrib
simple or with steeply ascending lateral veins. Disk-segments normally
6, roundish or squarish, somewhat fleshy, entire, c. 0.3-0.5 mm. across.
Stamens usually 6 (rarely 5 or 7), filaments united into a rather stout
sometimes apiculate column 0.5-0.9 mm. high; anthers subsessile (the
free portion of the filament usually shorter than the anther) in 2 approxi-
mate whorls atop the column, broadly ovate, rounded across the top, c.
0.2-0.3 mm. long, 0.3—-0.4 mm. broad; anther-sacs divergent or divaricate,

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 133

dehiscing vertically (upper whorl) or horizontally (lower whorl), the slits
not confluent; pollen grains 18-23 (—26) m in diameter.

Female flower: pedicel 0.3-2 (—2.5) mm. long, smooth, slender, angled.
Calyx greenish white (drying reddish) ; calyx-lobes normally 6, chartaceous-
scarious, subequal, elliptic-oblong to spathulate, c. 2.5-3 mm. long, 1-1.7
mm. broad, rounded or obtuse at the tip, entire or obscurely denticulate,
midrib obscure and unbranched or nearly so. Disk thin and flat, 6-lobed
or angled, crenulate or entire. Ovary sessile, inconspicuously sulcate or
ribbed; styles erect, mostly 1.3-2.5 mm. high, united below into a column
c. 0.8-1.5 mm. high, parted % to 34 their length; style-branches usually
sharply reflexed from the point of bifurcation, c. 0.7-1.2 mm. long, diver-
gent, narrowed to slender sometimes revolute tips.

Capsule oblate, c. 3.5 mm. in diameter (or slightly less), dark reddish
brown, essentially smooth, veins obscure. Columella 1.2-1.5 mm. long.
Seeds trigonous, symmetric or slightly asymmetric, 1.6-1.7 mm. long,
1-1.2 mm. radially, 1-1.3 mm. tangentially, dark greyish- or reddish-
brown, with evenly spaced rows of slightly raised dots; hilum subterminal,
elliptic or ovate, c. 0.2-0.3 mm. long.

The two subspecies of this rather widespread plant appear different in
so many respects that some workers might be reluctant to combine them.
In almost all instances, they can be distinguishd vegetatively because of
the hirtellous cataphylls of ssp. chamaecristoides. However, the male and
female flowers are nearly identical (within the usual range of variation )
in both groups and the allopatric range adds support to the probability
that we are here dealing with subspecies which replace one another geo-
graphically rather than with distinct “Linnaean” species. The practical
task of characterizing P. chamaecristoides taxonomically is made more
difficult by the fact that ssp. chamaecristoides and ssp. baracoensis appear
to hybridize with two different species, P. phlebocarpus and P. scopulorum,
respectively.

The species most closely related to P. chamaecristoides are probably
P. myrtilloides and P. scopulorum. Subspecies shaferi of P. myrtilloides
is vegetatively somewhat similar and also has connate filaments and a
rather similar stylar column, although it differs in its much longer pedicels.
A much greater vegetative similarity is shown by P. scopulorum, but that
species has much shorter, free styles and also usually longer pedicels.

KEY TO THE SUBSPECIES
Cataphylls and branchlet stipules hirtellous; leaf scarcely or not apiculate, mar-
gins plane or reflexe ssp. chamaecristoides
Cataphylls and branchlet stipules glabrous; leaf with a usually conspicuous re-
flexed apiculum, margins usually revolute. ................-. ssp. baracoensis

62a. Phyllanthus chamaecristoides ssp. chamaecristoides

Phyllanthus chamaecristoides Urb. Symb. Ant. 9: 185-186. 1924.
Phyllanthus apiculatus Urb. op. cit. 184-185 (non P. apiculatus Merr., 1920).

134 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXXIX

Terminal cones of cataphylls conspicuously scurfy-hirtellous; cataphylls
caducous. Branchlets (2.5—) 4-7 (-13) cm. long, with mostly 25-45
(-55) leaves. Leaves: stipules at first copiously hirtellous on both faces
and margins, later becoming more or less glabrate; leaf-blades mostly
asymmetrically elliptic or oblong, often more or less falcate, with an ob-
scure or obsolete apiculum at the tip, more or less inequilateral at the base
(one side acute, the other obtuse), c. 6-10 mm, long, 2-4 mm. broad;
margins plane or reflexed, not revolute.

Female flower: pedicel 0.9-1.4 mm. long; stylar column c. 0,9-1.5 mm.

high.
Collected in flower Feb., June, July, Nov.; in fruit Nov.
Type: Cuba, Ekman 2127.

DIstRIBUTION: thickets and savannas, sepentine soil, Sierra de Nipe
(Map XXVIII).

CUBA, OrIENTE: Sierre de Nipe: Loma de Estrella, Ekman 1732 (S), 2127
(S, HoLotype; A, NY, S, ISOTYPES), 9838 (S); Bayate, Arroyo Piedra, Ekman
4655 (S); Bayate, Rio Piloto, Ekman 5908 (A, NY, S; isotypes of P. apicu-
latus); Bayate, Pinalito, Ekman 10020 (S).

This population, which can be distinguished from all others in sect.
Orbicularia by its hirtellous cataphylls, is known (in the typical form, at
least) only from the southern part of the Sierra de Nipe. The specimen
of ssp. baracoensis collected in the Sierra de Nipe by Carabia probably
comes from the northern part of the range, but it is not yet clear how
close the distributional limits of the two subspecies approach one another.

The collection from the Arroyo Piedra (Ekman 4655), on the basis of
which Urban proposed the species P. apiculatus, does not differ from other
collections in any important respect. The leaves are somewhat less falcate
and the stipules even more densely hirtellous, but the floral structure is
essentially the same,

62b. Phyllanthus chamaecristoides ssp. baracoensis (Urb.), stat.
nov. (PLATE XXIX, figs. K-L).
Phyllanthus baracoensis Urb. Symb. Ant. 9: 186-187. 1924.
Phyllanthus coelophyllus Urb. loc. cit.

Terminal cones of cataphylls completely glabrous; cataphylls deciduous.
Branchlets 2-5 (—8) cm. long, with (12—) 15-25 (-35) leaves. Leaves:
stipules glabrous; leaf-blades symmetric, becoming convex, broadly rhom-
bic-obovate to spathulate, with a usually conspicuous reflexed more or
less deciduous apiculum, 3—7 (—9) mm. long, 2-5 (—7) mm. broad; margins
recurved or usually strongly revolute.

Female flower: pedicel 0.3-2 (—2.5) mm. long; stylar column (0.6—)
0.8—1.1 (—1.4) mm. high.

Collected in flower and fruit Jan.-Aug.

Type: Cuba, Ekman 4326.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 13

DISTRIBUTION: pinelands, thickets, and stream-banks, serpentine areas,
eastern Cuba (Map XXVIII)

CUBA. Oriente. Sierra de Nipe: Loma del Winch, Carabia 3821 (MICH,
Nae Moa region: pinares de Moa, Acufa (SV 13154); Moa, Mrs. Bucher 31,

6 (SV); Rio de la Sabana, Yaguaneque, near Cananova, Leon, Marie- eee
' Clement 20706 (MICH); Cerro de Miraflores, Ledn 21147 (MICH);
Brefia woods, Ledn, Clement, & Alain 22496 (MICH); Playa de Vaca, es
Victorin, Clement, & Alain 21462 (MT); Centeno, bords du rio, Marie- Victori in,
Clement, & Alain 21471 (MT); charrascal du Cerro de Miraflores, Marie-Vic-
torin, Clement, & Alain 21491 (MT); Moa, pinéde, Marie-Victorin & Clement
21748 (MT); ravines, Cerro de Miraflores, Webster 3888 (GH, MICH); pine-
land between Cerro de Miraflores and Moa, Webster 3893 (GH, MICH).
Baracoa region: Minas de Iberia ad Taco Bay, Ekman 3839 (S, ISOTYPE of
P. coelophyllus); charrascales near Rio Macaguanigua, Ekman 4326 (S, HOLO-
type; NY, S, isotypes); Mesa de Prada, Jauco, Leén 11771 (NY, SV).

In contrast to ssp. chamaecristoides, ssp. baracoensis is a much more
widely distributed plant, extending from the Sierra de Nipe on the west
to near the eastern tip of the island at Jauco. The Carabia collection from
the Sierra de Nipe is rather surprising, since it is far disjunct from the
rest of the known population, and one would rather have suspected that
ssp. chamaecristoides would occur at the Loma del Winch. However, the
collection is quite typical for ssp. baracoensis, although the leaves are un-
usually small, and there is no doubt as to its determination.

The group most closely related to ssp. baracoensis (besides ssp. chamae-
cristoides) is P. scopulorum, which has a very similar habit and which
occurs together with ssp. baracoensis in the Moa region. In this area ssp.
baracoensis shows an increase in female pedicel length; in Wedster 3893,
for example, the pedicels are up to 3.5 mm. long, or at the border of the
size-range of P. scopulorum, but the long united styles of the specimen
place it definitely in ssp. baracoensis. More nearly intermediate is Marie-
Victorin et al. 21746 (MICH, MT; not cited above), which has female
pedicels 6-8 mm. long as in P. scopulorum but united styles as in ssp.
baracoensis. The origin of these aberrant specimens is still not clear, but
if P. scopulorum is really a distinct species there is certainly a strong
presumption that it hybridizes with ssp. baracoensis.

63. Phyllanthus scopulorum (Britton) ve Symb. Ant. 9: 187. 1924.
LATE XXIX, figs. M-N).

Orbicularia scopulorum Britton, Mem. Torr. Bot. Club 16: 73. 1920.

Glabrous bushy shrub up to 1-2 m. high; branches straight, brittle,
c. 1-2.5 mm. thick, terete, bark dark brown or greyish, smooth or fur-
rowed, sometimes wax- -incrusted. Cataphylls blackish, indurate, reflexed,
largely persistent: stipules triangular-lanceolate, (1.5-) 2-3 mm. lone.
0.5-1 mm. broad, long-attenuate at the tip, entire or obscurely denticulate;
blade lanceolate, c. 1-2 mm. long, 0.5 mm. broad. Deciduous branchlets
ascending or. spreading, mostly 3-5 (—6) cm. long, 0.2-0.4 mm. thick,

136 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

stramineous becoming greyish, subterete, with c. 15-25 (-30) leaves;
first internode 1—2.5 mm. long, median internodes 1.5—3 mm. long. Leaves:
stipules reflexed, lanceolate, acuminate-attenuate, entire or denticulate
at the base, the proximal ones blackish, indurate, persistent, mostly 1.5—
3.5 mm. long, the distal ones scarious, brownish, subpersistent, 0.5-1.5
mm. long. Petioles olivaceous or stramineous, 0.4-0.8 mm, long. Leaf-
blades flexibly coriaceous, mostly obovate or obcuneate, sometimes falcate,
mostly 3-6 (—7) mm. long, 1.5-4 (—5) mm. broad, broadly obtuse to
rounded or subtruncate at the tip, the conspicuous attenuate apiculum of
the juvenile blade reflexed and deciduous in age, acute at the base: above
olivaceous, sublucid (or glossy in age), very minutely foveolate, veins
somewhat raised and subprominent to quite obscure; beneath paler, green-
ish or whitish, often with waxy atoms or minutely scabridulous, midrib
and lateral veins (c. 3-5 on a side) much more prominent than above,
usually raised, veinlets obscure; margins usually conspicuously revolute
(at least in older leaves).

Monoecious; cymules mostly with 1 or 2 flowers; proximal cymules
male, occasional distal cymules bisexual or with a single female flower.

Male flower: pedicel 3-6 mm. long. Calyx usually pinkish-tinged;
calyx-lobes 6, biseriate, membranous, 1.5—2.5 mm. long, the outer linear-
to oblong-obovate, c. 0.5-0.8 mm. broad, the inner obovate or spathulate,
c. 0.8-1.2 mm. broad; lobes obtuse or rounded at the tip, entire or ob-
scurely denticulate, midrib usually unbranched or nearly so. Disk-segments
6, squarish, concave, rather fleshy, minutely pitted, c. 0.2-0.3 mm. across.
Stamens 6 (rarely 7), filaments united into a slender column 0.5—0.9 mm.
high; anthers sessile or subsessile in 2 approximate whorls atop the column,
broadly ovate, rounded or truncate across the tip, c. 0.2-0.25 mm. long,
0.4-0.5 mm. broad; anther-sacs divergent or subparallel, dehiscing verti-
cally (upper anthers) or horizontally (lower anthers), the slits not con-
fluent; pollen grains c. 18-23 , in diameter.

Female flower: pedicel becoming (3—) 4-8 (-12) mm. long. Calyx
usually pinkish-tinged; calyx-lobes 6, thin, biseriate as in the male, c.
1.7—2.3 mm. long, the outer c. 0.5-0.8 mm. broad, the inner c. 0.8—1.3 mm.
broad, obtuse or rounded at the tip, entire or obscurely denticulate, midrib
sparsely branching but veins usually obscure. Disk undulate-lobed or
divided into segments, somewhat thickened, dark. Ovary sessile, with
inconspicuous sutural ribs or bands; styles free or sometimes connivent
or coherent, erect, ascending, or spreading, c. 0.3—-0.7 mm. high (i.e., to
point of bifurcation), bifid; style-branches usually reflexed, about as long
as base of style, the tips slender, inturned or revolute.

Capsule oblate, c. 3-3.5 mm. broad, dark reddish brown, smooth, with
yellowish sutural stripes and rather conspicuous veins. Columella 1.2—1.5
mm. long. Seeds trigonous, slightly asymmetric, carinate on one of the
lateral faces, 1.4-1.7 mm. long, 1.1-1.3 mm. radially and tangentially,
reddish-brown with even rows of raised dots; hilum subterminal, roundish,
c. 0.25—0.3 mm. across.

Collected in flower and fruit April through July.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS UBS

Type: Cuba, Shafer 4006.
DISTRIBUTION: pinelands, northeastern Oriente province, Cuba (Map
XXIX).

UBA. Oriente: Moa region: Camino Delta no. 1, Acuma 12497 (SV, US),
12498 (US); Yagrumaje del médio, Clément 3620 (MT); chemin du Cayo
Chiquito, Clément 3650 (GH, MT, US); Sierra de Moa, edge of creek in pine
woods 15 km. southwest of Compania de Moa mill, Howard 5854 (GH. MT, NY,
US); charrascal del Coco, south of Moa, Ledn, Clemente, & Alain 22637
(MICH); Jicotea rivulet, pinelands, east of Moa, Ledn, Clemente, & Howard
20165 (MICH); pineland between Rio Moa and Rio Yagrumaje, Webster 3756
(MICH); banks of Rio Cayoguan c. 3-4 mi. upstream from delta, Webster
3800 (GH, MICH, NY), 3803 (MICH); ecotone between hardwoods and pine-
land, Cayo Chiquita, 8 km. south of Moa, Webster 3849 (GH, MICH); thickets
near Camp Toa, alt. 400 m., Shafer 4006 (NY, HOLOTYPE).

This species endemic to the Moa region grows both in pinelands and
streambeds, although in the latter habitat it is sometimes replaced by
P. myrtilloides ssp. spathulifolius. From both that plant and from P.
chamaecristoides ssp. baracoensis, it is distinguished by its much shorter
free styles, while in pedicel length it is intermediate between them. As
previously mentioned, there is a distinct possibility that it hybridizes with
ssp. baracoensis and it is conceivable that it may cross with ssp. spathuli-
folius as well.

In many respects P. scopulorum is intermediate between P. phlebocarpus
and P. chamaecristoides ssp. baracoensis; it has the free styles and long
female pedicel of the former combined with the stamen number and vegeta-
tive features of the latter. The possibility that it is in fact a hybrid popu-
lation must therefore be considered. However, weighing against such an
assumption is the fertility of P. scopulorum and the fact that in the Moa
region it is much commoner and more widespread than P. phlebocarpus.
Consequently, P. scopulorum is here provisionally accepted as a distinct
species.

The typification of P. scopulorum presents additional difficulties, because
the type specimen is entirely sterile and originated somewhat outside of
the known range of the species as determined by the remainder of the
collections. Until Shafer’s type locality can be revisited, therefore, it is
not certain that the name P. scopulorum is really applicable to the plant
of the Moa region described above. However, until more evidence is forth-
coming the Shafer specimen is best considered conspecific with the others.

64. Phyllanthus nummularioides Muell. Arg. Linnaea 32: 5. 1863,
DC. Prodr. 15(2): 333. 1866. (PLATE XXIX, figs. O-P).
Diasperus nummularioides (Muell. Arg.) O. Ktze. Rev. Gen. 2: 600. 1891.
Low shrub (becoming 1 m. high, Ekman); branches erect, straight,
slender (c. 1.5—-3 mm. thick), smooth, subterete, reddish brown. Cata-
phylls blackish, indurate, reflexed, more or less deciduous: stipules lanceo-
late, 2-3 mm. long, 0.5—-1 mm. broad, attenuate-acuminate, entire, glab-

138 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxix

rous; blade very similar, c. 2-3 mm. long. Deciduous branchlets (4—)
6-12 (-15) cm. long, 0.4-0.7 mm. thick, stramineous to reddish brown,
subterete, smooth, with mostly 10-20 (—25) leaves; first internode (2-)
4—8 mm. long, median internodes (2—) 4-11 mm. long. Leaves: stipules
reflexed, lanceolate, entire, the proximal ones blackish, indurate, mostly
persistent, 1.5-2.5 mm, long, the distal ones scarious, brownish, sub-
persistent, c. 0.7-1.1 mm. long. Petiole 1—2.5 mm. long. Leaf-blades
firmly chartaceous, broadly elliptic or obovate to suborbicular, mostly
10-20 (—23) mm. long and 7-17 mm. broad, rounded or subtruncate and
minutely apiculate at the tip (apiculum of distal leaves not over c. 0.2
mm. long), obtuse to rounded at the base; above olivaceous, dull or sub-
lucid, very minutely foveolate, midrib and lateral veins plane or somewhat
sunken, often prominent; beneath yellowish- or greyish-green, midrib and
main lateral veins (c. 4 or 5 on a side) more or less equally raised, veinlets
also raised and reticulum subprominent: margins thin, scarious, plane
or reflexed but not revolute.

Monoecious [but female flowers sparse and specimens thus occasionally
appearing to be entirely male]; cymules mostly with 2-4 male flowers,
occasional distal cymules with 1 female and 1-3 male flowers.

‘Male flower: pedicel capillary, 5-12 mm. long. Calyx whitish; calyx-
lobes 6, membranous, subequal, elliptic to obovate, mostly 1.5—2.5 mm.
long, 1-1.5 mm. broad, obtuse at the tip, subentire, midrib simple or
pinnately branched. Disk-segments 6, roundish, entire, not evidently
pitted, c. 0.2-0.3 mm. broad, free or united by pairs. Stamens 6 (rarely
5); filaments united into a column c. 0.5—0.9 mm. high; anthers briefly
stipitate (free portion of filament c. 0.2—0.5 mm. long) or the outer sub-
sessile, ovate, emarginate, c. 0.25—0.3 mm. long and 0.3-0.4 mm. broad;
anther-sacs divergent, the slits confluent at the apex, the inner (upper) de-
discing more or less vertically, the outer dehiscing more or less horizontally;
pollen grains 18-26 p in diameter.

Female flower: pedicel slender, 8-17 mm. long, terete below, angled
and slightly thickened above. Calyx whitish; calyx-lobes 6, thin, sub-
chartaceous, subequal, oblong to broadly elliptic or obovate, c. 2-3.2 mm.
long, 1.2—2 mm. broad, obtuse or subacute at the tip, entire or obscurely
denticulate, midrib pinnately branched. Disk entire and 6-angled or cut
into distinct segments 0.4—-0.5 mm. across, flat or undulate, not massive.
Ovary sessile; styles free or slightly coherent at the very base, erect or
ascending, bifid c. 1% their length, the undivided portion c. 0.3-0.5 mm.
Jong, the branches slightly spreading to sharply recurving.

Capsule not seen entire (probably c. 4 mm. in diameter) ; valves reddish
brown, smooth, not veiny, c. 2.5 mm. long. Seeds [not seen fully mature]
c. 1.5 mm. long, reddish-brown, with rows of slightly raised dots.

Collected in flower and fruit Oct.-Feb.

Type: St. Domingo, Bertero (G, HOLOTYPE; MO, W, IsotTypEs),

DIsTRIBUTION: pinelands, Central Hispaniola (Map XXV).
DOMINICAN REPUBLIC. La Veca: pine forests on rocky slopes north of

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 139

Piedra Blanca, alt. 200-500 m., 14 Oct. & 23 Dec. 1947, 9 Jan. 1948, Allard
16066 (GH, NY, US), 18085 (US), 18861 (S, US); Bonao, Loma Peguera, alt.
c. 250 m., 8 Feb. 1929, Ekman H11487 (A, S, US).

As the sole representative of sect. Orbicularia outside of Cuba, P, num-
mularioides is of particular phytogeographic interest. All of the Allard
and Ekman collections are from the same mountain area north of Piedra
Blanca and Ekman (in a note on a label) has suggested that Bertero may
have collected the plant at the same place, since the Loma Peguera is near
the road between Sto. Domingo [Ciudad Trujillo] and Santiago. The
complete isolation of this small population strongly suggests its origin by
long-distance dispersal from Cuba

The relationships of P. nummularioides are rather difficult to determine.
In a number of respects, including reflexed cataphylls and essentially
free styles, it closely resembles P. scopulorum. However, since that plant
has such a different aspect and also differs by the several characters shown
in the key, it may not be the most closely related species. Much more
similar vegetatively is P. myrtilloides, especially ssp. erythrinus, which
differs, however, in its larger flowers, purplish coloration, and connate
styles. It is possible that P. nummularioides should be considered only a
subspecies of P. myrtilloides but it is retained at specific rank for the time
being because its free styles, at any rate, distinguish it from all forms
of that polytypic species.

65. Phyllanthus phlebocarpus Urb. Symb. Ant. 9: 189. 1924.
(PLATE XXIX, figs. Q-R).
Phyllanthus breviramis Urb. op. cit. 192.
Phyllanthus estrellensis Urb. op. cit. 188.

Glabrous shrub; branches straight, slender (c. 1.5—3 mm. thick), smooth,
terete, becoming dark greyish or brown. Cataphylls blackish, indurate,
soon deciduous: stipules broadly triangular, c. 1-1.5 mm. long, acute,
entire; blade linear-oblong, 1-1.2 mm. long. Deciduous branchlets erect
or ascending, 2-6 cm. long, 0.4-0.5 mm. thick, stramineous, terete or
obtusely angled, smooth, with mostly 7-12 (—20) leaves; first internode
2-8 (-10) mm. long, median internodes 2-8 mm. long. Leaves: stipules
reflexed, deciduous, triangular-lanceolate, entire, the proximal ones black-
ish and indurate, the distal ones brown and scarious, 0.6-1 mm. long, up
to 0.5-0.6 mm. broad. Petioles c. 1-2 mm. long, stramineous. Leaf-blades
thinly chartaceous, broadly elliptic or obovate to suborbicular, (8-) 10-18
(—20) mm. long, (6- ) 8-13 (-19) mm. broad, mostly retuse or emarginate
at the tip (apiculum completely obsolete), obtuse to truncate or sub-
cordate at the base; above olivaceous, dull to lucid, midrib, veins, and
veinlets, yellowish, plane or raised, anastomosing in a fine but prominent
reticulum; beneath yellowish, dull to lucid, midrib, lateral veins (c. 4-6
on a side), and veinlets prominently raised in a conspicuous reticulum;
margins unthickened, plane

Monoecious; proximal cymules with 1 or 2 male flowers, occasional

140 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

distal cymules with a solitary female flower (often only one female flower
on a branchlet).

Male flower: pedicel 2.5—4 (—6) mm. long. Calyx whitish; calyx-lobes
normally 6 (rarely 4 or 5), membranous, biseriate, subequal, oblong or
obovate (inner lobes somewhat broader than outer), 1-1.5 (-1.8) mm.
long, 0.6-0.9 (—1.2) mm. broad, midrib unbranched. Disk-segments isom-
erous with calyx-lobes, roundish or squarish, obscurely stipitate, entire,
minutely foveolate, 0.25—-0.4 mm. across. Stamens 3 (rarely 2 or 4),
filaments completely connate into a slender column c. 0.4—0.6 mm. high;
anthers sessile atop the column, united by the connectives, ovate, c. 0.25—
0.4 mm. broad; anther-sacs divergent, dehiscing obliquely or horizontally,
the slits not confluent; pollen grains (14-) 16-20 (—22) , in diameter.

Female flower: pedicel very slender, 6-9 mm. long, slightly thickened
above. Calyx whitish; calyx-lobes 6, thin, biseriate, subequal, oblong to
narrowly obovate, 1.5—2 mm. long, 0.8-1.2 mm. broad, rounded and entire
or obscurely and minutely denticulate at the tip, midrib unbranched or
nearly so. Disk 6-lobed or divided into 6 discrete, squarish, somewhat
thickened, dark segments. Ovary sessile, minutely verruculose, 3-sulcate;
styles very shortly basally connate into a column c. 0.2 mm. high, thence
spreading, the free ends parted c. 34 their length, the branches c. 0.2—0.4
mm. long, divergent, the narrowed tips recurved.

Capsule not seen entire; valves c. 2.5 mm. long, greenish or reddish-
tinged, with a rather conspicuous reticulum of sunken veins. Columella
1.2-1.3 mm. long. Seeds asymmetrically trigonous (carinate on one face),
1.3-1.6 mm. long, 1-1.2 mm. radially and tangentially, reddish brown,
with fine more or less evenly spaced reddish-brown slightly raised dots;
hilum subterminal, rounded-triangular, c. 0.3—0.4 mm. across.

Collected in flower Apr.-May; in fruit May, July, Oct.

Type: Cuba, Ekman 2271a.

DisTRIBUTION: scrubland and savannas, on serpentine, northern Oriente
province, Cuba (Map XXIX),

CUBA. OriENTE: Sierra de Nipe: Cayo del Rey, Pinar Colorado, Carabia
3573 (GH, NY); Loma La Mensura, Carabia 3751 (GH, MICH, NY); Cayo
del Rey, Loma de Bio, Carabia 4068 (GH, MICH, NY); Bayate, in collibus,
Ekman 2019 (S); Loma de Estrella, savanna, Ekman 2271a (S, HOLOTYPE; NY,
ISOTYPE); charrascales ad viam Bio, Ekman 9574 (S, HOLOTYPE of P. brevira-
mis); northern slope of Sierra de Nipe, alt. 400 m., Morton & Acuna 2984
(US). Moa region: brefiales de Playa Vaca, Acufa (SV 13157); Mina Franklin,
Acuna 12504 (US); Moa, Ferras 15004 (SV); charrascal, Playa de La Vaca,
Marie-Victorin, Clément, & Alain 56700 (MT

The following more or less aberrant collections are probably also referable
here:

CUBA. ORIENTE: Bayate, Ekman 2018 (S; det. by Urban as P. aff. norlin-
dit ?); Loma de Estrella, savanna, Ekman 2271b (S, HOLOTYPE of P. estrellensis),
2272 (S, NY; TYPE COLLECTION of P. norlindii).

As here interpreted, P. phlebocarpus is a rather widespread and variable

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 141

plant of open vegetation on serpentine lands, extending from Sierra de
Nipe to Moa. Usually the species is easily recognizable by its broad, thin
leaves which are conspicuously reticulate on both sides, unlike those of
any other species of sect. Orbicularia. The androecium of 3 stamens and
the short spreading nearly free styles also sharply distinguish it. Further-
more, P. phlebocarpus appears to differ from other species of the section
in being at least partially deciduous, the flowers and new leaves develop-
ing together on the expanding branchlets during the renewed growing
season in April or May. In other species the flowers appear in sequence
as each branchlet successively matures. Urban’s proposed species P.
breviramis was based on a specimen of P. phlebocarpus in the spring
“flush” of growth. The “pistillode” described by him is merely the
apiculate tip of the staminal column, and in all essential respects the
specimen does not deviate from typical P. phlebocarpus.

Heretofore, the status of P. phlebocarpus has been obscured (in the
literature and in herbaria) by the existence in the Sierra de Nipe of
anomalous plants with narrower leaves, shorter petioles, and a variable
stamen number of 3-5; on the basis of such specimens Urban described
two new species, P. estrellensis and P. norlindii. However, the circum-
stances of the collection of these specimens casts strong suspicion on the
validity of these proposed concepts as representing true species in nature.
It is notable that Ekman collected the type specimens of all three “species”
at the same locality (ie., savanna at Loma Estrella) and, in fact, con-
founded P. estrellensis and P. phlebocarpus under the same number; this
appears significant, since Ekman rarely made mixed collections of different
species. The only other collection which has been referred to P. norlindi
(Ekman 2018) was also taken at the same locality with P. phlebocarpus
(Ekman 2019) and — judging from the consecutive collection numbers —
again from adjacent plants.

The known occurrence of P. estrellensis and P. norlindi only in con-
junction with P. phlebocarpus suggests that they may represent only some
sort of modification of that species. It is possible that they represent
hybrid forms between P. phlebocarpus and P. chamaecristoides ssp.
chamaecristoides, for the latter was also collected by Ekman at both
localities; and the leaf-shape, petiole and pedicel-length, and stamen-
number of P. estrellensis and P. norlindii are more or less intermediate
between the two putative parental species, P. estrellensis suggesting a back-
cross with P. phlebocarpus, and P. norlindii being more truly intermediate.
The type specimen of P. norlindii is copiously flowering and fruiting, and
the pollen fertility is about 80%, but the seeds do not appear to be viable.
However, it should be stated clearly that evidence of hybridity is thus far
purely circumstantial and is advanced only as a working hypothesis. The
situation is complicated by the intrinsic variability of P. phlebocarpus.
For example, such specimens as Carabia 4068, although typical for P.
phlebocarpus in floral characters, have leaf-shapes which match those on
sheets of P. estrellensis and P. norlindii. Whatever their genetic constitu-
tion the plants to which those two names have been applied would appear

142 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

to be classifiable under P. phlebocarpus. It does not seem necessary at
the present time to designate them by formula, even if their hybrid origin
could be demonstrated.

The relationships of P. phlebocarpus are not easy to determine, owing
to its isolated position within sect. Orbicularia. Vegetatively it resembles
some forms of P. myrtilloides, particularly ssp. erythrinus, and this is
perhaps its closest definable affinity

Sect. 19. Omphacodes Webster, Contr. Gray Herb. 176: 59. 1955.

Shrubs with phyllanthoid branching; branchlets often borne on spur-
shoots; leaves chartaceous, stipules subpersistent. Monoecious, cymules
bisexual. Male flower: calyx-lobes 5; stamens 3 or less commonly 4,
filaments united; pollen grains with large areoles. Female flower: calyx-
lobes 5, often deciduous; disk shallowly cupuliform; styles free, bifid,
branches rather thick. Capsule massive, the exocarp somewhat fleshy,
cocci indehiscent; seeds asymmetric, smooth.

Type species: Phyllanthus subcarnosus Wright ex Muell. Arg.

The single variable species of this section occupies an isolated position
in subg. Xylophylla and it is possible that it should be referred to subg.
Cicca instead. The female flower resembles that of the commonly culti-
vated P. acidus (sect. Cicca), the subindehiscent fruit that of P. elsiae
(sect. Aporosella), and the leaf venation and branch-spur formation that
of P. pseudocicca (sect. Ciccopsis). Probably sect. Ciccopsis should be
regarded as the most closely related group in subg. Cicca. Among the
sections of subg. Xylophylla, perhaps the most similar is Asterandra,
which has leaves with somewhat similar petiolar ridges, male flowers with
three (or more) monadelphous stamens, areolate pollen grains, and a
globose fruit which is fleshy until rather late in ontogeny.

The areolate pollen grains of P. subcarnosus (PL. IX, fig. 41) are of
particular interest since they can be regarded as derived from a microspore
with three confluent bordered colpi such as that of P. acidus (Jour. Arnold
Arb, 37: 240. 1956). If the areolate grains of the other taxa in subg.
Xylophylla can be shown to have evolved from the microspore of P.
subcarnosus, then sect. Omphacodes would have to be regarded as a phylo-
genetically important group and possibly the most primitive group within
subg. Xylophylla. However, the specialized fruit, reduced androecium,
and peculiar styles of sect. Omphacodes offer eather eeu obstacles
to such an evolutionary hypothesis. Furthermore, it is by no means im-
possible that the areolate pollen grains of sect. ieee ee io been
derived independently of those in the other sections of the subgenus. It
is true that the areolate grains of sect. Macraea can be distinguished from
those of subg. Xylophylla by the different placement of the germ-pores,
while in sect. Omphacodes there is no such difference. Thus there is a
slight presumption in favor of the idea that the pollen of sect. Omphacodes
may be primitive within subg. Xylophylla but the case is by no means

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 143

proved. Until some of the South American species which appear also to
lie in the “neutral zone” between subgenera Cicca and Xylophylla can
be critically studied, the phylogenetic position of sect. Omphacodes must
remain somewhat uncertain.

66. Phyllanthus subcarnosus Wright ex Muell. Arg. in DC. Prodr.
15(2): 379. 1866. (PLATE XXX, figs. A-B).

Diasperus subcarnosus (Wright) O. Ktze. Rev. Gen. 2: 601. 1891.
Phyllanthus leonis Alain, Contr. Ocas. Mus. Colegio La Salle 12: 1. 1953.

Glabrous shrub or tree c. 2-4 m, high; branches straight, distally c.
3-6 mm. thick, terete or sometimes distinctly angled, smooth, dark brown
or grey, lenticels inconspicuous. Cataphylls indurate, brownish, not re-
flexed, persistent: stipules triangular, blunt or acute at the tip, 1.5-2
(—2.5) mm. long, 1-2 mm. broad; blade linear-lanceolate, c. 1.5-2.5 mm.
long. Deciduous branchlets borne scattered on main branches or clustered
on lateral spur-shoots, mostly 7-15 (—25) cm. long and 1.5—2.5 mm. thick,
compressed, sharply wing-angled, stramineous, smooth, with 5-12 (—15)
leaves; first internode mostly 10-30 mm. long, median internodes c, 10-30
mm. long. Leaves: stipules appressed, persistent or (distally) sometimes
completely deciduous, triangular-lanceolate, acuminate, 1.2—2.5 mm. long,
0.8-1.5 mm. broad, brownish, entire, the base becoming more or less in-
durate, the scarious tip often breaking off. Petioles 1.5-4 (—5) mm. long,
brownish, with two straight or slightly undulate adaxial ridges decurrent
from the blade. Leaf-blades chartaceous, elliptic or slightly ovate (less
commonly obovate or suborbicular), 3-11 cm. long and 2—7 cm. broad
(reduced blades at proximal nodes smaller), emarginate or obtuse to sub-
acute and with a brownish more or less deciduous apiculum up to 0.5 mm.
long at the tip, obtuse to rounded or occasionally subcordate at the base;
above olivaceous to silvery, obscurely to conspicuously foveolate, the de-
pressed or plane midrib rather conspicuous, the lateral veins obscure;
beneath paler or silvery, more or less obscurely foveolate, the midrib
salient (and excurrent at the apex), the lateral veins (c. 5-7 on a side)
slightly raised, anastomosing short of the margin, enclosing subprominent
reticula of veinlets; margin unthickened or scarcely so, plane or recurved.

Monoecious; cymules bisexual (proximal ones sometimes entirely male),
of one female and several male flowers.

Male flower: pedicel capillary (slightly thickened above), c. (3—) 4-6
mm. long. Calyx whitish (in life); calyx-lobes 5, subequal, more or less
spreading, suborbicular to broadly elliptic or obovate, 1.3-2.7 mm. long
and 1-1.8 mm. broad (at full anthesis, somewhat smaller in mature bud
stage), subentire, thickened at the junction with the fleshy receptacle but
distally thin, the midrib unbranched. Disk-segments 5, free or casually
united, somewhat fleshy, squarish to triangular or round, obscurely pitted,
c. 0.25-0.5 mm. across. Stamens 3 or less commonly 4, filaments connate
into a column (0.3—) 0.5—0.8 mm. high; anthers subsessile atop the column,
more or less ascending, c. 0.25-0.3 mm. long and 0.35-0.5 mm. broad;

144 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxrx

anther-sacs divergent, the slits apically contiguous but not confluent,
dehiscing obliquely; pollen grains spheroidal, c. 20-22 » in diameter, with
large polybrochate mostly pentagonal areoles c. ™% the diameter of the
grain.

Female flower: pedicel becoming (8—) 10-20 mm. long, terete and
slender below, dilated and sharply angled above. Calyx-lobes 5, charta-
ceous, caducous (falling at anthesis and represented only by prominent
scars) or sometimes persistent and becoming reflexed in fruit, elliptic-
oblong to obovate, c. 1.8-2.5 mm. long, 1.2—1.8 mm. broad, subentire,
centrally thickened with very thin scarious margins, the midrib unbranched.
Disk shallowly cupuliform, tenuous, the undulate margin not pitted, c.
0.25—0.4 mm. high but always much shorter than the ovary. Ovary sub-
globose; styles free, spreading, bifid nearly to the base, the divergent
thickened and rather fleshy branches c. 0.5—0.7 mm. long, blunt at the tips.

Capsule subglobose, when dried 6-13 mm. in diameter, the exocarp
fleshy but apparently thin, drying reddish brown; cocci more or less
indehiscent, the outer and inner walls firmly united. Columella not seen.
Seeds asymmetrically trigonous, those of each pair often unequal, (3—)
3.5-5 mm. long, (2—) 2.7-3.3 mm. broad, shiny reddish brown, smooth
(sometimes very obscurely striate); hilum subterminal, c. 0.7 mm. long;
micropylar end often with a small yellowish caruncle.

Collected in flower June, July, Aug.; in fruit Feb., Mar., July, Aug.

Type: Cuba, Wright 1946,

DISTRIBUTION: infra-mangrove coastal scrub and inland mountainous
areas, Cuba and Hispaniola (Map XXX).

CUBA. Pinar pet Rio: Mendoza, in forests at Boquerén, Ekman 18746
(S, SV); behind manglares, halfway between Malas Aguas and San Cayetano,
Webster 4683 (GH, MICH); Sierra de los Organos, grupo del Rosario, San
Diego de Tapia, at the edge of the Rio Mani-Mani (= Rio San Miguel), Ekman
12665 (S); Zambumbia Hill, Rangel, Leén 12715 (MICH, NY); Rio San
Miguel from Volador to Mal Paso, Wilson 9374 (F, GH, NY, US); oe
in woods pee manglares, Wright 1946 (G, HoLoTYPE; F, GH, GOET, MO,
P, S, isotypes); Toscano, a s Calaveras, woods bordering on manglares,
Ekman 17427 (S); Morillo, a bordering on manglares, Ekman 17401 (S).
ORIENTE: Puerto Padre, Curbelo X91 (NY, SV); Sierra Maestra, Rio Yara,
Nagua, in thickets, Ekman 14180 (S), Leén 10980 (NY); Sierra de Nipe, at
base of Loma Mensura, Ekman 3157 (S); Sierra de Nipe, El Taller, prope
Rio Piloto in dumetis, Ekman 9676; Cerro de Cananova, Sagua de Tanamo,
Clemente & Crisdgono 6234 (GH, tsotype of P. leonis); Rio de la Sabana, Ya-
guaneque, Cananova, Ledn et al. 20724 (MICH); El Yunque, Baracoa, Bucher
(SV 14556); Farallon La Perla, Shafer 8747 (NY); woods at the foot of Sierra
de Imias, Imias, Ledn 12137 (NY); rocky banks of Jauco River, Jauco, Ledn
11843 (NY, SV).

HISPANIOLA. “Saint-Domingue,” Poiteau (A, P).

As here circumscribed, P. subcarnosus is a highly variable species oc-
curring in scattered localities in Cuba and Hispaniola, Although wide-
spread, it has been collected only rarely in the fertile condition, and its

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 145

variation pattern presents many puzzling aspects. Originally it appeared
that the population from Oriente province, described by Alain as P. leonis,
could be recognized as a subspecies distinct from the Pinar del Rio plants
on the basis of the following features: longer branchlets (up to 30 cm.)
and leaves (mostly 5-10 instead of 3-5 cm. long), larger male calyx-lobes
(2 mm. long or more) and female pedicels, usually persistent rather than
deciduous female calyx-lobes, and larger capsules (9-13 mm. as compared
to 6-7 mm. in the very few capsules measured from Pinar del Rio) and
seeds (4.5-5 mm. long as opposed to 3-3.8 mm.). This would appear to
be an impressive number of distinctions, but it must be remembered that
the size measurements are mostly based on inadequate sampling and that,
in most instances, there is either some overlap in the range of variation
or else a relatively narrow gap. When the Hispaniolan plant is taken into
account, the single collection of Poiteau is sufficient to erase some of the
supposed distinctions, for it has the large capsules of P. leonis combined
with the small male flowers of typical P. subcarnosus and is intermediate
in leaf-size.

™~..
“ ”
5 é y ey
F<3 i. x ° so 100 tsa. | an
eng Pan) ‘Ty =
er 1 So) tite
“aon ee = = NN
Poe
t eae
a es ays
‘ ‘4 ae RR or
ne Q
Pon
x NS A a
Sars z
iba ~ SSS
, <a ee
as a 8 n z 7 Lise err rors

Map XXX. Distribution of P. subcarnosus Wr. ex Muell. Arg.

It is still possible that when better samples are available from mosv
areas of the range of P. subcarnosus it may prove feasible to distinguish
or more subspecific taxa; the leaf-shape of the Hispaniolan plants
is distinctive, and the specimens from the Sierra de Nipe have unusually
large leaves. However, it is outside the province of this study to base the
systematic disposition on any such hypothetical prediction. Consequently,
for the time being no subspecific entities are recognized within P. sub-
carnosus, with the proviso that a much more intensive study of the varia-
tion is indicated.

One of the major difficulties in the analysis of the present species is
the lack of adequate floral material. It is not possible, for instance, to
decide conclusively whether the variation in size of the male flowers is
purely random or follows a geographical pattern. When sect. Omphacodes
was first described, the stamen number was given as 3. This proves tc be
the prevalent but not the exclusive number, for a count of 50 flowers from
Webster 4683 yielded a score of 38 flowers with 3 stamens and 12 flowers

146 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

with 4. This tendency toward a higher stamen number is of interest in
view of the tetramerous condition of the androecium in some of the prob-
ably reiated sections of subg. Cicca. In contrast to the stamen number,
no deviations were observed in the number of calyx-lobes and carpels, which
were constantly 5 and 3, respectively.

The ecological behavior of P. subcarnosus is very interesting and merits
field study. On the northern coast of Pinar del Rio it occurs in the
shrubby woods directly behind the mangrove zone, associated with such
plants as Eugenia axillaris and Comocladia dentata; Curbelo’s collection
from Puerto Padre probably came from a similar habitat. However, it
occurs inland both in the serpentine areas of the Sierra de Nipe and the
calcareous regions of the Sierra Maestra. If it has really been derived
from subg. Cicca, then it is probable that the original habitat of the species
was in the sub-mangrove zone as in P, elsiae and P. acidus. The distribu-
tion of the present-day populations of P. swbcarnosus may illustrate there-
fore, an important West Indian evolutionary phenomenon, viz., the colon-
ization of inland areas by plants which have migrated along. halophytic
shoreline areas.

Sect. 20. Asterandra (Kl.) Muell. Arg. Linnaea 32: 5. 1863.
Asterandra Kl. Arch. Naturgesch. 7: 200. 1841.

Slender shrubs or trees with palm-like habit, the trunk with branchlets
clustered at the apex; cataphylls inconspicuous; branchlets pinnatiform;
leaves chartaceous, stipules reflexed. Monoecious; cymules mostly bisex-
ual. Male flower: calyx-lobes 5, disk-segments coalescent into a massive
ring; stamens 3-7, filaments united; anthers dehiscing horizontally or
downwards; pollen grains areolate. Female flower: calyx-lobes 5 (6);
disk as in male; styles erect, connate about halfway, the tips dilated, hori-
zontally spreading. Capsule dry at maturity, cocci rather massive; seeds
rounded, mottled, smooth, thick-walled.

Type species: Asterandra cornifolia (HBK.) Kl. [= PAyllanthus jug-
landifolius Willd. ssp. cornifolius (HBK.) Webster. |

Section Asterandra, although comprising only a single polytypic species,
has a rather broad distribution which includes most of the West Indies
and tropical South America. It is certainly very closely related to sect.
Oxalistylis, which has very similar leaves and flowers but differs in its
usually hirsutulous leaves and axes, free male disk-segments, longer styles
with differently shaped stigmas, and thin-walled ridged or striate seeds.
Possibly further study will show that the two sections should be combined,
in which case Oxalistylis would have priority as a sectional epithet.

67. Phyllanthus juglandifolius Willd. Enum. Hort. Berol. Suppl.
64-65.
Phyllanthus grandifolius sensu Muell. Arg. in DC. Prodr. 15(2): 329. 1866;
et auct. seq., non

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 147

Shrub or tree 1.5-10 m. high, the slender simple or sparsely branching
trunk often with branchlets clustered at the apex in the manner of a palm;
axes nearly smooth (minutely scabridulous on the youngest parts) or
minutely hirtellous. Cataphylls inconspicuous, reflexed: stipules chartace-
ous, lanceolate, c. 1.7—2.5 mm. long. Deciduous branchlets 25-120 cm.
long, 2-5 mm. thick, olivaceous, smooth or minutely hirtellous, more or
less angled, with c. 15-45 leaves; first internode 15-70 mm. long, median
internodes 10-50 mm. long. Leaves: stipules triangular-lanceolate, soon
reflexed, c. 1-2.5 mm. long, 0.8-1.5 mm. broad, chartaceous, olivaceous.
Petiole glabrous or scabridulous, 2-5 mm. long, the leaf-blade margins
decurrent on the adaxial side as two strongly undulate and conspicuous
flanges. Leaf-blades chartaceous, elliptic- or oblong-lanceolate (proximal
ones broadly elliptic to suborbicular), c. 5-20 cm. long, 2.5-6 cm. broad,
abruptly short-acuminate, obtuse to cordate at the base; above olivaceous,
smooth, the midrib, veins, and veinlets all evident but scarcely raised; be-
neath paler (sometimes glaucous), smooth or minutely scabridulous, some-
times minutely scurfy or hirtellous on the midrib and major veins, mid-
rib salient, lateral veins (c. 6-10 on a side) and veinlets rather prominently
raised, forming a conspicuous reticulum; margins unthickened, plane or
recurved.

Monoecious; branchlets most often floriferous at every node; cymules
simple to thrice compound, the peduncle up to 2 mm. long, more or less
adnate (and flowers thus slightly supra-axillary); female flowers up to
3 or 4 in proximal axils, usually 1 or 2 in middle region of branchlet, often
lacking distally (distal cymules thus entirely male); male flowers 2—10
per cymule, rarely solitary or absent

Male flower: pedicel capillary, c. 10-20 mm. long. Calyx yellowish-
green or greenish-white; calyx-lobes 5, chartaceous (rather fleshy at the
base), somewhat unequal, the outer lobes elliptic or oblong and narrower,
the inner lobes broadly elliptic to obovate or suborbicular, c. 1.8-3 mm.
long and 1.5—2.5 mm. broad, blunt and often sparsely denticulate at the
tip, otherwise entire, the scarious margin narrow and indefinite, the (1—)
3—5 nerves usually obscure. Disk annular, 5-angled, very massive, usually
deeply pitted, 1.5-3 mm. across. Sava 3-7, filaments connate into a
rather massive column 1—1.5 mm. high and 0.5-0.75 mm. thick; anthers
sessile atop the column, horizontal or deflexed below the umbonate top of
the column, mostly c. 0.6-0.7 mm. long and 0.7—0.8 mm. broad (reduced
ones smaller) ; anther-sacs subparallel, not confluent, dehiscing horizontally
or obliquely downward; pollen grains 23-30 » in diameter.

Female flower: pedicel slender, 5-30 mm. long, smooth or rarely scabri-
dulous, obscurely angled. Calyx yellowish green; calyx-lobes 5, herbaceous,
rather thick, subequal or distinctly unequal, elliptic or oblong to obovate,
in fruit becoming 2—4 mm. long and 1.2—2.5 mm. broad, obtuse or rounded
and entire or obscurely denticulate at the tip, the midrib slightly raised
on both sides, the laterals obscure. Disk massive, 5-angled, closely re-
sembling the male. Ovary smooth, carinate; styles erect, connate about

148 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

halfway into a thick column, the flattened dilated tips bifid or emarginate,

c, 0.5—0.7 mm. long, more or less horizontally spreading.

Capsule oblate, trigonous, deeply sulcate between the cocci, c. 7-12 mm.
high and 9-17 mm. in diameter, smooth and glabrous, not veiny. Columella
4—6 mm. long, massive. Seeds plano-convex, plump, when mature 4.2—6
mm. long, 3.2—-4 mm. broad, irregularly brown-and-gray mottled, smooth;
hilum central, raphe linear-triangular.

It was shown previously (Jour. Arnold Arb. 37: 10. 1956) that Mueller’s
application of the name P. grandifolius to this West Indian and South
American species was an erroneous adoption of horticultural practice, the
true P. grandifolius L. being a Mexican species of subg. Botryanthus. As
here interpreted, P. juglandifolius is a very widespread and polymorphic
species which ranges over the West Indies and much of tropical South
America.

KEY TO THE SUBSPECIES

Leaves mostly elliptic-lanceolate, obtuse to rounded at the base, c. 5-11 (rarely
to 15) cm. long, glabrous beneath; stamens 3-5; fruiting pedicel 6-11 (-14

long; mature capsule mostly 10-12 mm. in diameter. ...... ssp. juglandifolius
Leaves mostly oblong-lanceolate, rounded to cordate at the base, c. 12-20 (rarely
only 9 or 10) cm. long, glabrous or hirtellous beneath; stamens mostly 6, less
commonly 5 or 7; fruiting pedicel 15-30 mm. long; mature capsule mostly 12-15
(ody) Wan I AOUEE. os ivy eas daso ber as eee nes erry ee es ssp. cornifolius

67a. Phyllanthus juglandifolius ssp. juglandifolius
(PLATE I, fig. 4; PLATE XXX, figs. C-D).

Phyllanthus juglandifolius Willd. emer Hort. Berol. Suppl. 64-65. 1813.

Agyneia berterit Spreng. Syst. Veg. 3: 1826.

Phyllanthus grandifolius y genuinus a Arg. in DC. Prodr. 15(2): 329.
(ex p.); non P. grandifolius

Phyllanthus quinquefidus Sesse & Moc. Flora Mex. ed. 2. 212. 1894.

PLATE XXX. Fiowers or sects. Omphacodes, Asterandra, Epistylium, AND
Glyptothamnus.

Fics. A-B. Male and female flowers of P. subcarnosus Wr. ex Muell. Arg.
(Webster 4683 [GH]). Fics. C-F. Flowers of P. juglandifolius Willd. C, male
flower of ssp. juglandifolius and androecium as seen from above (Webster 4028
[GH]). D, female flower of ssp. juglandifolius and styles as seen from above
(Howard 6468 [GH]). E-F, androecium (as seen from above) and gynoecium
of ssp. cornifolius (HBK.) Webster (Haught 3067 [GH]). Fics. G-I. Male

, ovary in long section, and female flower of P. cladanthus Muell. Arg.
(Proctor 11800 [GH]). Fics. J-K. Female flower and gynoecium of P. cauli-
florus (Sw.) Griseb. (Britton & Hollick 2027 |NY]). Fics. L-O. Male flower,
female flower, gynoecium in long section, and female calyx-lobe of P. axillaris
(Sw.) Muell. Arg. (Howard 14131 [A]). Fics. P-Q. Phyllanthus chryseus
Howard (Webster 3853 [GH]). P, part of male flower, showing androecium,
disk, and one calyx-lobe. Q, disk and gynoecium of female flower.

Jour. ARNOLD ARB, VoL. XXXIX PLATE XXX

WEBSTER, WEST INDIAN PHYLLANTHUS

150 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

Shrub or treelet 1.5—3 m. high; branchlets 25-60 (—75) cm. long, smooth
or minutely scabridulous, with (12—) 15-35 leaves; leaf-blades elliptic-
lanceolate (less commonly oblong-lanceolate), c. 5-11 (-15) cm. long,
2.55 cm. broad. Male flower: calyx-lobes 1.8-2.5 mm. long, 1.2—2.5 mm.
broad; disk 1.5—2 mm. across; stamens mostly 3 or 4, less commonly 5.
Female flower: pedicel becoming 6-11 (-15) mm. jong; fruiting calyx-
lobes 2—2.8 mm. long. Capsule c. (S—) 7-8 mm. high and (9—) 10- 12.5
mm. in diameter; mature seeds 4.5—5.5 mm. long, 3.2—-4 mm. broad.

Collected flowering February through October; fruiting April through
December.

Type: Herb. Willdenow (B, HoLoTYPE). The type sheet has no exact
data, but from Willdenow’s description it appears to have been taken from
a plant cultivated in the Berlin Botanical Garden.

DistTRIBUTION: mainly Greater Antilles, a disjunct population occurring
in Brazil (Mar XXXI).

CUBA. Prvar DEL Rio: Sierra de Rangel, Taco-Taco, Acuna 5950 (SV),
Wright 586 ex p. (S, US); Taco-Taco River, Aspiro, Santa Cruz de los Pinos,
Alain & Clemente 1429 (MICH). Las VILLAs: — Mountains: San Blas
to Buenos Aires, Gavinas, Howard 6468 (GH, MO, MT, US); Buenos Aires, alt.
2500-3500 ft., Tack 7435 (A, US), 8030 (S, US), W ebster 4771 (GH, MICH).
CAMAGUEY: potrero de La Ciega, Caobilla, Acufta 13516 (SV). OrteNTE: Sierra
Maestra, Rio Yara, gravel-beds, Ekman 14667, 14830 (S); Moa, Bucher 11646,
14085 (SV); Los Llanos, Baracoa, Bucher 10115 (SV); Jagiiey, Eggers 5304
(A, F, US); Mt. Liban, Linden 1796 (BR, G, P, W); Alto Yateras, Guantanamo,
es 633 (SV); 20 miles south of Baracoa, alt. 2100 ft., W abster 4028 (GH,

MICH); Monte Verde, Wright 586 ex p. (BR, C, G, GH, MO, P, S, W).

HAITI. Norv-Ovest: Bassin Bleu, road to Gros Morne, Leonard & Leonard
14701 (GH, US); Méle River 2 miles from La MGle-St.-Nicolas, Leonard &
Leonard 13086 (US); Méle-St.-Nicolas, road through Méle gorge, Leonard &
Leonard 13125 (A, MO, US); Presqiv’ile du Nord-Ouest, Port-de-Paix, Haut-
Moustique, Ekman H3670 (S). Sup: Morne de la Hotte, northeast slopes, alt.
800 m., Ekman H212 (S); Aux Cayes, Ekman H18 (A, S

DOMINICAN REPUBLIC. “St. Domingue,” Bertero (P); “Santo Do-
mingo,” Prenleloup 524 (G, US). Puerto Prata: Pto. Plata, Rio Mameyes,
Eggers 1652 (C, G, US). Santiaco: Rio Amina, el Corozo, alt. 500 m., Jimenez
H15 (US); Jicomé, Mera 2055 (US). La Veca: Rio Yaque, near Jarabacoa, alt.
550 m., Fuertes 1603 (A, G, L, W). SAN PEpRo pE Macoris: San Pedro de
Macoris, Rose, Fitch, & Russell 4208 (US); 20 km. west of San Pedro de
Macoris, Howard & Howard 9505 (GH). Serpo: Llano Costero, Higiiey, Arr.
Caguero, Ekman H12146 (S); La Romana, river basin, Taylor 365 (US).

PUERTO RICO. Bertero (P; TYPE COLLECTION of Agyneia berterii), Plée
848 (P), Riedlé (P), Wydler 307 (F, G, P). AcuapitLa: Aguada, Sintenis
5565 (L). Arecrpo: Manati, Mango, Sintenis 6611 (MO, US). San Juan:
Catano, limestone hill, Briton. Britton, & Brown 6988 (F, US); Bayamon,
mountains, Sintenis 998b (US), Stahl 1073 (US). Ponce: El Tendal, Coamo
River, Britton, Britton, & Brown 6015 (F); arroyo near Ponce, Britton &
Britton 9562 (US). GuayAMma: Cayey, Rio Morillos, Sintenis 2291 (G, GH, S);
Cayey, La Cruz, Sintenis 2387 (F, G, P).

1958] WEBSTER, WEST INDIAN PHYLLANTHUS Loe

VIRGIN ISLANDS. Sr. THomas: 1798, Riedlé (P); Signalhill, alt. 500 m.,

Eggers 374 (BR, G, GOET, W)
67b. Phyllanthus juglandifolius ssp. cornifolius (HBK.), stat. nov.
(PLATE XXX, figs. E-F).

Phyllanthus cornifolius HBK. Nov. Gen. & Sp. 2: 115. 1817.

Asterandra cornifolia Kl. Arch. Naturgesch. 7: 200. 1841.

Phyllanthus grandifolius a cornifolius (HBK.) Muell. Arg. in DC. Prodr.
15(2): 329. 1866.

Phyllanthus grandifolius 8 salzmanni Muell. Arg. ibid.

Phyllanthus grandifolius y genuinus Muell. Arg. ibid. (ex p.).

Shrub or palm-like tree 2-10 m. high; branchlets (40—) 50-100 (—120)
m. long, smooth or hirtellous, with (25—) 30-45 leaves; leaf-blades oblong-
lanceolate, (10—-) 13-17 (—20) cm. long, 3-5 (—7) cm. broad, glabrous
beneath or sometimes hirtellous along the midrib and main veins. Male
flower: calyx-lobes 2.1-3 mm. long, 1.5—2.8 mm. broad; disk 1.8-3 mm.
across; stamens mostly 6, less commonly 5 or 7 (mostly 5 in Trinidad
plants). Female flower: pedicel becoming (10—) 15-25 (—30) mm. long;
fruiting calyx-lobes c. 3-4 mm. long. Capsule c. 8-12 mm. high, (10-)
12-15 (-17) mm. in diameter; mature seeds 4.5—5.5 mm. long [6 mm.
ex Urban], 3.5—4 mm. broad.

Collected flowering (in Trinidad) in June; fruiting in June and August.

Type: Ecuador, Guayaquil, Herb. Humboldt 3850 (P, SYNTYPE).

DisTRIBUTION: widespread in tropical South America, reaching its
northern limit in Trinidad (Map XXX1).

TRINIDAD: without specific locality, Finlay (TRIN 2462), Von Rohr 91
(C); Moruga, Britton & Broadway 2463 (GH); Cedros, L’Enviense, Broadway
(TRIN 8538); San Fernando Hill, Lewis (TRIN 9164); 12 mile post, Penal
Rock ate Williams (TRIN 12172); Southern Watershed Reserve, Walliams
(TRIN 12179).

The following specimens are representative of the South America ma-
terial studied of this subspecies:

BRAZIL. Banta: Bahia, pea (G, GH; TYPE COLLECTION of P. grandi-
folius var. salzmanni). Maran Ao: Rio Pindare, Rapoza-Moncao, Froes 11661
(F.) Rio DE JANEIRO: eee near Campos, Glaziou 13491 (C Ne

ECUADOR. Guayas: Balao, Eggers 14123 (A); south of Milagro, Hitchcock
20567 (GH). MAnasi: near Guale, Haught 3067 (GH).

PERU. Loreto: Lower Rio Huallaga, Santa Rosa, Williams 4880 (F).

VENEZUELA. Zutta: Maracaybo, Moritz 260 (GH); Perija, Maracaibo,
Karsten (W).

Urban (Repert. Sp. Nov. 15: 404. 1919) accepted P. cornifolius as a
species distinct from P. juglandifolius on the basis of its leaf-shape, longer
female pedicels, supposedly differently shaped style-tips, and larger fruit.
He cited from Trinidad Broadway 2727 (unfortunately not examined dur-

152 JOURNAL OF THE ARNOLD ARBORETUM _ [vov. xxxiIx

ing the present study) and described the capsule as 14 mm. high, 18 mm.
in diameter, and with seeds 6 mm. long. These dimensions, while presum-
ably correctly stated, are larger than those observed in any other specimen;
and the distinctive features of P. cornifolius as presented by Urban thus
appear impressive. However, it is significant that his discussion is cau-
tiously restricted to a comparison of the Trinidad collection with the West
Indian specimens only of P. juglandifolius. When the variation throughout
the entire range of both taxa is analyzed, the fluctuation and overlap of
characters together with the complete geographical replacement of the two
is strongly indicative that the populations in question are best ranked as
subspecies of a single variable species.

Map XXXI. Caribbean distribution of P. juglandifolius Willd.: small dots,
ssp. juglandifolius; large dots, ssp. cornifolius.

Our knowledge of the variation patterns within ssp. cornifolius is very
fragmentary in comparison with that of ssp. juglandifolius, which is at
once homogeneous and better represented by herbarium specimens. How-
ever, although the scarcity of comparable herbarium material makes it im-
possible at this time to prove whether or not there is a significant overlap
in measurements, it is clear that the mean values of many dimensions
(e.g., height, length of branchlets and fruiting pedicels, size of leaves, and
diameter of fruits) in ssp. cornifolius are definitely larger than those in
ssp. juglandifolius.

The Trinidad population of P. juglandifolius is no better known than

1958] WEBSTER, WEST INDIAN PHYLLANTHUS LoS

the South American ones with respect to the number of fertile collections
available; but in any event it appears to be somewhat transitional between
the two subspecies. For example, the leaves are always glabrous as in
ssp. juglandifolius, but in size and shape they are wholly characteristic
of ssp. cornifolius. Again, the stamen number in the Trinidad plants ap-
pears to be mainly 5, which is the number of overlap between the sub-
species; and a single flower of Von Rohr 91 had only 3 stamens. This
is the only instance within ssp. cornifolius of a number lower than 5, so
that it may simply be a completely exceptional occurrence; but the exam-
ination of a good series of male flowers from Trinidad is needed to clarify
the matter

The detailed distribution of ssp. cornifolius on Trinidad is interesting
in that all collections have been made in the southwestern corner of the
island, San Fernando Hill being the point farthest north. Judging from
the vegetation map of Beard (Nat. Veg. Trinidad, frontisp. 1946), the
plant is generally found in the semi-evergreen seasonal forest, where it
presumably plays the part of a pioneer plant as ssp. juglandifolius was
observed to do in Cuba.

Sect. 21. Epistylium (Sw.) Griseb. Fl. Br. W. Ind. 33. 1859.
Epistylium Sw. FI. Ind. Occ. 1100.
Epistylium sect. Euepistylium Baill. Sil Gen. Euphorb. 647. 1858.
Phyllanthus sect..Catastylium Griseb. ibid.

Shrubs or trees with phyllanthoid branching; branchlets pinnatiform,
clustered at the apex of the more or less unbranched stem; leaves char-
taceous or coriaceous, stipules persistent. Monoecious; cymules bisexual,
borne on racemiform axillary or cauline thyrses. Male flower: calyx-lobes
4 or 5; disk-segments 4 or 5; stamens 2 or 3, filaments united; anthers
more or less deflexed; pollen grains areolate. Female flower: receptacle
and base of calyx massive; calyx-lobes 5, erect; disk tenuous, lobed or
parted into 5 segments; stigmas massive, sessile atop ovary or terminating
an elongated gynoecium. Capsule dry at maturity, angled; seeds 2 or pos-
sibly sometimes 1 per locule.

Type species: Omphalea axillare Sw. (= Phyllanthus axillaris (Sw.)
Griseb. )

As here construed, sect. Epistylium comprises three species endemic to
Jamaica which have the palm-like habit of sect. Asterandra and show a
similar calciphilous habitat preference. Although the latter section differs
from sect. Epistylium in its woodier seeds, much more massive floral disk,
and very different inflorescence, its resemblance (at least to P. cladanthus)
in habit, stipules, anthers, and styles suggests a rather close relationship.
However, some of the representatives of subg. Cicca, especially P. acidus
(sect. Cicca), also are suggestively similar to sect. Epistylium in the pro-
duction of cauliflorous thyrses; and the tetramerous flowers, reduced fe-

154 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. XXxIx

male disk, and tendency to reduction in seed number in sect. Cicca repre-
sent a much closer approach to the condition in sect. Epistylium than do
the corresponding features in sect. Asterandra. On the other hand, the
tricolporate pollen grains, free stamens, and very different styles of sect.
Cicca would appear to preclude an intimate relationship with sect. Epis-
tylium. Another possibly related group is sect. Omphacodes, which does
have areolate pollen grains and which shows some vegetative resemblance
to the Jamaican plants. However, in that section cauliflory does not occur,
and the stipules and styles are very different. It is most difficult to decide
whether the resemblance between Epistylium and Cicca is genetically sig-
nificant, or whether the various floral similarities may not rather be sim-
ply correlated with the cauliflorous condition. The multiple and seem-
ingly contradictory affinities of sect. Epistylium pose a most interesting
problem for further and more intensive investigation.

If the nature of the ties between Epistylium and its possible ancestors
is obscure, the affinity in a different direction is much more clear. Un-
doubtedly related to sect. Epistylium and probably descended from it is
sect. Hemiphyllanthus, which scarcely differs in anything more than pubes-
cent axes and bipinnatiform branchlets. The leaf venation of the Haitian
P. maleolens and P. myriophyllus is similar to that of P.. cladanthus, and
the massive receptacle and stylar column of P. cauliflorus is not unlike
that of P. ovatus or P. megapodus. The more highly modified species of
sect. Xylophylla are also related to sect. Epistylium, either directly or via
sect. Hemiphyllanthus.

The three species of sect. Epistylium, although clearly differing both in
floral and vegetative characters, are so obviously related that no purpose
would be served in separating P. cladanthus into a separate sect. Catastyl-
ium; Mueller recognized the weakness of the distinctions but out of inertia
recognized two sections. The least specialized of the three species is clearly
P. cladanthus, which has 3 or 4 stamens and a subglobose ovary. Its re-
flexed stipules are much more like those of sect. Asterandra than are the
stipules of either P. cauliflorus or P. axillaris.

KEY TO THE SPECIES

1. Stipules lanceolate, reflexed, not fused with branchlet; branchlets subterete,;
leaves oblong-lanceolate, chartaceous; male flower with 5 calyx-lobes and
ae ete id a ee Oe ae ee eee ee ee 68. P. cladanthus
Stipules triangular, massive, not reflexed, more or less fused to branchlet;
male flower with 4 calyx-lobes and 2 stamens.

2. Branchlets angled, not flattened; leaves chartaceous, oblong-lanceolate;
inflorescences at least partially cauline; pedicel of female flower 2-3 mm.
long or more; ovary extended into a long stylar column. ..............
bette a yon ena hace aes ee ee ee od a ee aie 69. P. cauliflorus
Branchlets flattened; leaves coriaceous, mostly elliptic; inflorescences
strictly axillary to leaves on branchlets; female flower subsessile; ovary
ellipsoid (stylar column confounded with upper part of ovary). ........
70. P. axillaris

pony

bo

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 1s)

68. Phyllanthus cladanthus Muell. Arg. Linnaea 32: 46. 1863; DC.
Prod. 15(2): 413. 1866; Fawc. & Rend. Fl. Jam. 4: 258. 1920.
(PLATE XXX, figs. G—l).

heres cauliflorus sensu Griseb. Fl. Br. W. Ind. 33. 1859; non Omphelea
auliflora Sw

Ren
eee Cadan thus (Muell. Arg.) O. Ktze. Rev. Gen. 2: 598. 1891.

Slender tree c. 5-10 m. high; trunk c. 1 dm. thick or less, usually un-
branched, smooth and reddish brown when young, becoming greyish.
Cataphylls coriaceous, reflexed: stipules triangular or broadly lanceolate,
c. 2.5-4 mm. long, 1.7-4 mm. broad, blunt, greyish; blade narrower.
Deciduous branchlets steeply ascending, (15—) 20-60 cm. long, c. 2-4
mm. thick, reddish brown, smooth, terete or somewhat angled, with c.
10-20 leaves; first internode (3—) 5-8 (—12) cm. long, median internodes
c. 1.5-3 cm. long. Leaves: stipules persistent, reflexed, indurate, lance-
olate, 2.5-5 mm. long, 1.5—2.5 mm. broad (becoming broader with age),
acuminate or blunt-tipped, dark reddish brown and polished. Petioles
4—7 mm. long, the laminar flanges adaxially decurrent. Leaf-blades char-
taceous, ovate- to more commonly oblong-lanceolate, abruptly acuminate,
9-15 cm. long, 3-5.5 (—7) cm. broad, obtuse to rounded or rarely sub-
cordate at the base; above olivaceous, the incised midrib prominent, the
delicate lateral veins plane or very slightly raised; beneath paler, sub-
lucid, the salient midrib proximally keeled, the main lateral veins (c. 8-12
on a side) and veinlets tenuous but raised, forming a conspicuous retic-
ulum; margins plane or narrowly reflexed.

Monoecious; cymules usually bisexual, borne on naked thyrses, the
latter usually fascicled at old nodes on main axis; additional thyrses
sometimes produced on branchlets. Thyrses (2—) 5-20 cm. long, with
c. 7-15 nodes; larger thyrses often compound, with up to 15 lateral axes
and with conspicuous indurate cataphylls proximally. Cymules each with
1 femaie and up to c. 10 male flowers.

Male flower: pedicel slender, 3-8 mm. long. Calyx pinkish (ex Proctor) ;
calyx-lobes 5 (rarely 6), thin and scarious, unequal: outer lobes oblong
to elliptic, c. 1-1.5 mm. long and 0.8—1.2 mm. broad; inner lobes broadly
obovate to suborbicular, c. 1.3-1.8 mm. long and 1.2—1.7 mm. broad; lobes
rounded at the tip, entire, the midrib simple or sparingly branched avove
the middle. Disk-segments 5 (rarely 6), dark, roundish, thickened, pitted,
crenulate, concave, c. 0.3-0.6 mm. across. Stamens 3 (rarely 4); filaments
completely connate into a column c. 1 mm. high; anthers sessile atop the
column, steeply deflexed (almost upside-down), (0.6—) 0.7—1 mm. long,
0.5-0.8 mm. broad; anther-sacs subparallel, dehiscing pseudo-vertically
(actualiy obliquely downwards), the slits not confluent; pollen grains
mostly 25-30 » in diameter, with c. 15 oligobrochate areoles per amb,
areoles c. 4—6 pw across.

Female flower: pedicel slender and terete (dilated only just beneath
the calyx), c. 7-15 mm. long. Calyx pinkish; calyx-lobes 5, erect, strongly
imbricate, somewhat unequal, fleshy with thin scarious margins, mostly

156 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL, xxxIx

broadly elliptic to suborbicular, c. 1.3-1.7 mm. long, 1-1.5 mm. broad, entire
and rounded at the tip, the midrib sparsely branching. Disk a very incon-
spicuous membranous crenate ring (with small teeth alternating with the
calyx-lobes), hidden under the ovary. Ovary subglobose, c. 1.5-2 mm.
high and 2.5-3 mm. broad at (or shortly after) anthesis, shallowly sulcate;
styles undeveloped, the dilated petaloid stigmas (style-tips) sessile, mas-
sive, more or less triangular, bluntly auriculate apically, crenulate along
the distal margin, c. 0.5-1.5 mm. high and 1.3—2 mm. broad.

Capsule (not seen fully mature) oblate, bluntly trigonous, c. 8 mm. in
diameter, smooth, not veiny. Seeds (immature) 2 per locule, c. 2.5 mm.
long, brownish.

Collected flowering Mar., June, Aug.; fruiting Mar.

Type: Jamaica, Wilson (GOET, fragment of TYPE).

DIstRIBUTION: limestone hills, west central to eastern Jamaica (Map
XXXII)

JAMAICA. TreLtawny: Cockpit country, rocky wooded hills, Tyre, ee
Sept. 1906, Britton 570 (NV); Crown lands, near Troy, alt. 2000-2500 ft.,
June, Aug., 1904, Harris 8722 (A, F, JAM, NY, US), 8761 (F, JAM, Rd
PorTLAND: John Crow Mountains, mist forest on dogtooth limestone 1.5 miles
uae of Ecclesdown, alt. c. 1000 ft., 6 Aug. 1954, Webster & Wilson 5161
(A). THOMAS: woodlands, eastern slopes of south end of John Crow Moun-

S, _ Mar. 1909, Harris & Britton (F, JAM, NY, US); Big Level, wooded
jawestone hill, alt. 1500-2000 ft., 16 Mar. 1956, Provigy 11800 (GH).

Although it was the basis for the separate section Catastylium estab-
lished by Grisebach, P. cladanthus is too closely related to P. cauliflorus
to be maintained in a distinct group. Vegetatively it so closely resembles
that species that Grisebach confounded them, but Mueller was able to
straighten out the confusion. However, even in the sterile condition P.
cladanthus is ordinarily readily distinguishable from P. cauliflorus by its
very different reflexed stipules.

Harris, in his detailed notes made on this species as he encountered it
near Troy, remarked of the flowers that they were “produced along the
main stem and branches, below the clusters of leaves, and extending down-
wards for a considerable distance. Occasionally the side branches are
covered with flower fascicles.” This description would appear to indicate
that P. cladanthus, like P. cauliflorus, produces axillary inflorescences on
the branchlets in addition to the usual cauliflorous ones.

The distribution of P. cladanthus parallels that of numerous other
Jamaican species in the great disjunction between the western localities
in the “cockpit” country at Troy and the eastern stations in the John Crow
Mountains. There are no apparent differences between the specimens of
the two areas, although the variation has certainly been inadequately
sampled heretofore. There are no obvious habitat differences in the inter-
vening region where the species does not occur, and such areas as Mt.
Diablo, at least, are so well known that it is unlikely the species has
escaped detection there.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS LEE

69. Phyllanthus cauliflorus (Sw.) Griseb. Fl. Br. W. Ind. 33. 1859;
emend. Muell. Arg. Linnaea 32: 46. 1863; and in DC. Prodr. 15(2):
412. 1886; Fawc. & Rend. Fl. Jam. 4: 258. 1920.

(PLATE XXX, figs. J-K).

Omphalea cauliflora Sw. Prodr. 95. 1788.
Epistylium cauliflorum (Sw.) Sw. FI. Ind. Occ. 1099, pl. 22, figs. e, f, h. 1800.
Diasperus cauliflorus (Sw.) O. Ktze. Rev. Gen. 2: 598. 1891

Slender tree with usually unbranched trunk, becoming c. 5—6 m. high.
Cataphylls not observed. Deciduous branchlets 25-55 cm. long, mostly
2.5-3 mm. thick, olivaceous or reddish-brown, distinctly angled, smooth,
with 8—20 leaves; first internode 6-12 cm. long, median internodes c.
1.5-3.5 cm. long. Leaves: stipules persistent, not reflexed (tip appressed
or spreading), indurate and rather massive, becoming more or less fused
with the branchlet, triangular, c. 3-4 mm. long and 2.5-3 mm. broad on
lower part of branchlet but decreasing to only 1.5 mm. long distally, blunt-
tipped, greyish. Petioles 3-7 mm. long, the laminar flanges adaxially
decurrent. Leaf-blades chartaceous or subcoriaceous, flexuous, ovate- to
more commoniy oblong-lanceolate, 7-14 cm. long, 3-6 cm. broad, abruptly
short-acuminate, obtuse to rounded or subcordate at the base; above more
or less jucid when dried, the midrib incised, the lateral veins and veinlets
plane or slightly raised, subprominent; beneath sublucid, the midrib
carinate and very prominent, the lateral veins (c. 8-10 on a side) and
veinlets raised, forming a prominent reticulum; margins plane (or slightly
and narrowly reflexed).

Monoecious; cymules bisexual, borne on naked thyrses, the latter
usually fascicled at nodes on spur-shoots from main trunk but sometimes
also produced one or two together in axils of leaves on branchlets (axis
of thyrse sometimes reduced so that flowers may appear to be in axillary
clusters). Thyrses c. 3-18 cm. long, with 4-20 nodes; cymules each with
1 female and several male flowers.

Male flower: pedicel slender, 3-5 mm. long. Calyx yellowish green;
calyx-lobes 4, rounded at the tip, biseriate, unequal: outer lobes elliptic-
oblong, c. 0.9-1.2 mm. long, inner lobes suborbicular or broader than long,
1.4-1.8 mm. long and 1.3-2.1 mm. broad; midrib of outer lobes simple,
or inner lobes often sparingly branched distally. Disk-segments 4, dark,
rather massive, entire, c. 0.25—-0.6 mm. across. Stamens 2; filaments com-
pletely connate into a column c. 0.6-1 mm. high; anthers sessile atop the
column, triangular-ovate, blunt-tipped, c. 0.5—0.7 mm. long, 0.4—-0.6 mm.
broad; anther-sacs subparallel, dehiscing obliquely downward or almost
horizontally, the slits not apically confluent; pollen grains 15-18 pw in
diameter, areoles mostly polybrochate, c. 7 or 8 per amb, c. 5—6 wu across.

Female flower: pedicel becoming 2—3.5 mm. long, slender and subterete
below, rather abruptly incrassate and angled above. Calyx blood-red (ex
Swartz); calyx-lobes 5, erect, strongly imbricate, massive and fleshy at
the base, elliptic to suborbicular, at anthesis c. 1-1.2 mm. long and 0.9-1.2
mm. broad, later increasing up to 1.8 mm. long, rounded at the tip, entire,

158 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. xxxIx

iad eu heres branched. Disk divided into 5 thin, erect, oblong
segm ; m. long or less. Gynoecium cylindrical- ellipsoidal (i.e.,
the massive a column about as long as and nearly as broad as the
ovary proper); stylar column increasing to 1.5—2.5 mm. long soon after
anthesis; stigmas massive, more or less triangular, bluntly auriculate and
obcordate (apically channelled), subentire along the distal margin (often
with a downwardly projected lateral lobe), c. 0.4-0.7 mm. high and as
broad or broader.

Capsule (not seen mature) ovoid, pointed, obscurely ribbed, dark, not
veiny. Seeds (not seen) 2 per locule (ex Swartz

Type: montane forests, western Jamaica, Swartz (C, G, S; SYNTYPES).
DistTrRIBUTION: wooded limestone hills, western Jamaica (Map XXXII).

JAMAICA. ey ahs woods, summit of Dolphin Head, 17 Mar. 1908, Britton
& Hollick 2853 (F, NY). eee: es coastal thicket, Negril, 9-

Mar. 1908, Ballon 6 Hollick 2027 (F, N

On the basis of available collections, it ena that Mueller and Fawcett
and Rendle were correct in accepting Swartz’s distinction between P.
cauliflorus and P. axillaris. The two arene are much more closely related
to one another than to P. cladanthus, having in common the peculiar in-
crassate stipules, nearly indentical male flowers, and elongated gynoecia.

OQ
%
© P. AXILLARIS A :
@ P CAULIFLORUS
@ P CLADANTHUS nr

Map XXXII. Distribution of the species of sect. Epistyliam.

Swartz distinguished P. cauliflorus from P. axillaris by its greater height,
longer oblong leaves, cauliflorous inflorescence, more elongate gynoecium,
and beaked fruit with two seeds in each locule (instead of solitary as in
P. axiliaris). In general these distinctions are still largely valid, although
a few qualifications are necessary. Recent collections of P. axillaris show
that it is not necessarily only 2—4 ft. high as stated by Swartz but may
attain over 3.5 m. in height. Furthermore, P. cauliflorus does not produce
exclusively cauline inflorescences, for they are partially axillary in Britton
& Hollick 2027. However, this does not efface the inflorescence distinction
since, as far as is known, P. axillaris is never cauliflorous, and P. cauliflorus
always produces some cauline flowers.

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 159

The validity of the fruit characters suggested by Swartz must still rest
on his personal observations, since herbarium material is inadequate. His
descriptions are so accurate in other respects that there is no reason to
doubt his ascription of paired seeds in the locules of P. cauliflorus and
solitary ones in the locules of P. axillaris. However, it remains to be proved
whether this distinction will hold when a large.number of capsules of both
species can be examined.

70. Phyllanthus axillaris (Sw.) Muell. Arg. in DC. Prodr. 15(2): 412.
1866; Fawc. & Rend. Fl. Jam. 4: 258-259. 1920.
(PLATE XXX, figs. L-O).

Omphalea axillaris Sw. Prodr. 95.

Epistylium axillare (Sw.) Sw. FI. Ind. ae 1097. pl. 22, figs. a-d, g,i-k. 1800.
Omphalea epistylium Poir. Encycl. Meth. Suppl. 4: 140. 1816

Phyllanthus epistylium (Poir.) Griseb. Fl. Br. W. Ind. 33. 1859.
Diasperus axillaris (Sw.) O. Ktze. Rev. Gen. 2: 598. 1891.

Glabrous shrub c. 0.5—3.5 m. high, with usually unbranched trunk topped
by crown of branchlets. Cataphylls not seen. Deciduous branchlets 20—40
cm. long, c. 2-4 mm. thick, greyish and subterete proximally, reddish-
brown and flattened (and more or less marginally angled) distally, with
c. 15-25 leaves; first internode 6-8 cm. long, median internodes quite
variable in length (successive pairs of nodes sometimes approximate).
Leaves: stipules persistent, triangular, mostly 2-4 mm. long and 2.5—3.5
mm. broad (smaller at tip of branchlet), not reflexed (tip more or less
spreading), massive and indurate, becoming more or less fused with the
branchlet, blunt-tipped, greyish. Petioles 2.5-4 mm. long, with decurrent
adaxial laminar flanges, sometimes corrugate-fissured. Leaf-blades coria-
ceous, sometimes quite rigid, ovate- to elliptic-lanceolate, (6—) 8-11 cm.
long, (2.5—) 3-5.5 cm. broad, abruptly short-acuminate, mostly obtuse
to rounded at the base; above drying dull plumbeous or brownish grey,
the midrib incised but veins and veinlets distinctly raised and prominent;
beneath sublucid, often coppery when dried, the midrib carinate, the
lateral veins (c. 5—8 on a side) and veinlets raised, forming a very promin-
ent reticulum; margins slightly and narrowly reflexed.

Monoecious; cymules bisexual, each usually with one female and several
male flowers, borne on short naked thyrses c. 1-3 cm. long which are
apparently always axillary to leaves on branchlets (i.e., flowers never
cauline).

Male flower: pedicel slender, mostly 3-5 mm. long. Calyx yellowish
green; calyx-lobes 4, rounded at the tip, entire, biseriate, unequal: outer
lobes elliptic, c. 1-1.3 mm. long and 0.7—1.2 mm. broad; inner lobes sub-
orbicular or broader than long, c. 1.3-1.5 mm. long and 1.7—2 mm. broad;
midrib simple or sparingly branched distally. Disk-segments 4, dark,
massive, rugose-crenulate, concave, c. 0.3-0.5 mm. across. Stamens 2;
filaments completely connate into a column c. 0.9-1.2 mm. high; anthers
sessile atop the column, nearly horizontal, triangular-ovate, blunt-tipped,

160 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxtx

c. 0.7-0.8 mm. long and 0.5—-0.6 mm. broad; anther-sacs subparallel, de-
hiscing horizontally or slightly obliquely downwards, the slits not apically
confluent; pollen grains c. 21-25 » in diameter.

Female flower: pedicel sub-obsolete, less than 1 mm. long. Calyx cream-
colored or greenish; calyx-lobes 5, erect, epee imbricate, massive and
fleshy at base, elliptic to suborbicular, c. 1-1.3 mm. long and 0.8 (outer)
to 1.3 (inner) mm. broad, rounded at the ioe ae midrib simple or
sparsely branched distally. Disk divided into 5 thin, erect, oblong seg-
ments c. 0.1—-0.2 mm. long. Ovary ellipsoid, at or shortly after anthesis
c. 1.5—-1.8 mm. high and 1-1.3 mm. broad; styles undeveloped (confounded
with ovary), the dilated stigmas sessile atop the ovary, massive, more or
less triangular, bluntly auriculate and obcordate (apically channelled),
subentire along the distal margin, c. 0.5-0.7 mm. high and 0.7-1 mm.
broad.

Capsule trigonous, green (ex Howard); valves dark reddish brown, c.
5 mm. long. Columella c. 2.5 mm. long. Seeds solitary in each locule (ex
Swartz), c. 3.5 mm. long, slightly over 2 mm. across the back, pale brown
with dark brown slightly raised more or less elongated flecks; hilum sub-
median, triangular, c. 0.3 mm. long.

Collected in flower Apr. (Swartz), July, Sept.; in fruit July.

TYPE: mountains of western Jamaica, Swartz (G, SYNTYPE).
DISTRIBUTION: limestone areas, western Jamaica (Map XXXII).

JAMAICA, Tretawny: Ramgoat Cave area, limestone hilltops, hillsides, and
cliff-faces, 26 Sept. 1954, 4 July 1955, 19 Jan. 1956, Howard & Proctor 14131,
14371, 14639 (A).

From the other two species of sect. Epistylium, P. axillaris is readily
distinguishable on account of its shorter, thicker leaves, flattened branch-
lets, and axillary inflorescences. Apparently it also differs in its solitary
rather than paired seeds, as pointed out by Swartz, although this requires
confirmation,, The gynoecium of P. axillaris appears to be intermediate
between that of P. cladanthus and that of P. cauliflorus but this appear-
ance may be deceptive. The complete loss of any distinction between style
and ovary in P. axillaris might represent a further modification of the
long-styled gynoecium of P. cauliflorus. This would appear the more
probable, since in other respects— such as its more rigid leaves and
flattened branchlets — P. axillaris appears to be the most specialized
representative of sect. Epistylium.

Sect. 22. Glyptothamnus Webster, Jour. Arnold Arb. 39: 68. 1958.

Dendriform small shrub with pinnatiform branchlets; cataphylls indu-
rate; leaves coriaceous, margins revolute, stipules indurate and persistent.
Monoecious; cymules mostly unisexual., Male flower: calyx-lobes 4, disk-
segments coalescent into a massive ring; stamens 2, filaments united;
anthers dehiscing horizontally; pollen grains areolate. Female flower:

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 161

calyx-lobes 5; disk massive; styles dilated, lacerate. Capsule globose, not
sulcate; seeds dark, fissure

Type species: Phyllanthus chryseus Howard.

The single species of this monotypic section is so distinctive in many
features, such as its small indurate cataphylls and stipules, revolute leaves
golden beneath, very massive disk, and fissured seeds, that it cannot be
regarded as closely related to any other group. As noted by Howard, the
leaves and cataphylls show some resemblance to those of P. subcarnosus
(sect. Omphacodes), but of course the flowers are completely different.
The closest floral similarity is perhaps to be found in sect. Asterandra,
where there is a very apparent resemblance in the floral disk and styles;
however, the vegetative parts in sect. Asterandra are quite dissimilar. The
Jamaican species of sect. Epistylium are vaguely similar but differ in many
important respects, such as inflorescence, floral disk, and styles. Possibly
P. chryseus could be regarded as a xerophytic derivative (adapted to serpen-
tine) of the mesophytic and calciphilous P. juglandifolius, but if so the
affinity must be indirect. Phyllanthus chryseus is an excellent example of
a highly specialized relict species of ancient origin.

71. Phyllanthus chryseus Howard, Jour. Arnold Arb. 28: 121. 1947.
(PLATE XXX, figs. P-Q).

Glabrous shrub resembling a miniature tree, with bluish-green foliage,
the erect woody unbranched stem 2—8 dm. high, 5—7 mm. thick, reddish
brown and pruinose above, greyish below. Lower leaves of stem with
petioles 5-10 mm. long, leaf-blades elliptic or oblong-obovate, rounded or
obtuse at the tip, c. 6-8 cm. long and 2.3-4 cm. broad; upper leaves
reduced to cataphylls: stipules triangular-lanceolate, acuminate, 3.5—5.5
mm. long, 2-3 mm. broad, oblique at the base, indurate, smooth, reddish
brown and polished, with a single excentric keel; blade linear-lanceolate,
2.5-5 mm. long. Deciduous branchlets mostly 10-20 cm. long, 1.5—2.5
mm. thick, olivaceous, smooth, distinctly flattened but not sharply angled,
with mostly 8-20 (—25) leaves; first internode (of well-developed branch-
lets) 25-55 mm. long, median internodes 10-35 mm. long. Leaves: stipules
persistent, triangular, 1-2 mm. long, 0.9-1.8 mm. broad, bluntly pointed,
shining and indurate, reddish brown. Leaf-blades rigidly coriaceous,
broadly elliptic to orbicular (or sometimes broader than long), c. 2—4.5
cm. long, 1.5—3.5 cm. broad, rounded or emarginate at the tip (the minute
scarious apiculum reflexed), obtuse to rounded (or sometimes truncate
or subcordate) at the base; blades distinctly reddish when young, when
mature bluish green above (turning olivaceous or plumbeous on drying),
minutely foveolate, the midrib usually incised, other veins obscure; be-
neath golden-yellow (or greenish yellow) when living, turning coppery
when dried, the midrib salient, laterals (c. 5 on a side) slightly raised but
inconspicuous; margins thickened, conspicuously revolute.

Monoecious, cymules usually unisexual; two proximal nodes of branch-

162 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

let most often with racemiform cymules of 3 (less commonly up to 7)
male flowers; subsequent nodes with solitary female flowers alternating
with male cymules at somewhat irregular intervals.

Male flower: pedicel slender, 4.5—7.5 mm. long. Calyx greenish or
reddish-tinged, the receptacle massive; calyx-lobes 4, coriaceous, biseriate
(strongly imbricate in the bud), subequal, suborbicular (mostly broader
than long), 2.8-4.5 mm. long, 4.5—6 mm. broad, rounded at the tip, entire,
midrib with well-developed but rather inconspicuous branches. Disk very
massive, squarish in outline, surrounding the androecium (inclosing it in
a hollow), deeply pitted, greenish in life but turning reddish when dried.
Stamens 2; filaments completely connate into a slender column less than
1 mm. high; anthers sessile atop the column, discrete (separated by a notch
on each side, the connective between the anther-sacs emarginate so that
there may appear to be 4 anthers), semicircular-notched in outline, c.
0.4—0.5 mm. long, 0.7-0.8 mm. broad; anther-sacs divergent, curved, de-
hiscing horizontally, the slits not confluent; pollen grains 18-21 ,» in
diameter; areoles polybrochate, 5 or 6 per amb, 5- or 6-sided, c. 6-9 p
across.

Female flower: pedicel slender, 11-17 mm. long, terete or obscurely
angled, slightly and gradually broadened upwards. Calyx greenish, the
receptacle massive; calyx-lobes 5, coriaceous, spreading at anthesis, broadly
ovate, c. 4-7 mm. long and broad, rounded at the tip, entire, midrib con-
spicuously branched. Disk massive, 5-angled, smooth, entire, yellowish
at anthesis (drying reddish brown). Ovary sunken in the disk, capped
by the petaloid style-tips (stigmas), these free, spreading and horizontally
appressed, sessile, obcuneate, 1.5—2 mm. long, 2.3-2.7 mm. broad, yellow-
ish, conspicuously lacerate.

Capsule spheroidal, c. 7.5 mm. in diameter, reddish brown, rugulose,
not veiny. Columella slender, 4.5-6 mm. long. Seeds trigonous, symmetric,

. 4.7-5 mm. long, 2.6—3 mm. radially and tangentially, very dark reddish
ae (sometimes nearly black), deeply and irregularly transversely or
somewhat obliquely fissured on back and sides; hilum submedian, nar-
rowly elliptic, c. 0.4-0.5 mm. long.

Collected in flower and fruit May, July.

Type: Cuba, Howard 5829.

DISTRIBUTION: endemic to a small area of serpentine hillsides in the
Moa region, eastern Cuba.

CUBA. OrtentE: Moa, summer 1939, Mrs. Bucher 75 (NY); common in
woods along ravine 15 km. southwest of Moa, 26 July 1941, Howard 5829 (GH,
HOLOTYPE; US, ISOTYPE); pinelands, Arroyo Jicotea, Moa, July 1941, Leon,
Clemente, & Howard 20169 (MICH); Moa, plateau de 400 m. entre le rio
Cabanas et le rio Yagrumaje, 27-31 May 1943, Marie-Victorin & Clément 21755
(A, MT); pinales c. 15 km. south of Moa, 19 July 1951, Webster 3853 (GH,
MICH, NY, US).

One of the most distinctive endemic species of the serpentine flora of
Moa, P. chryseus appears to be confined to a very narrow range south of

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 163

Moa between the Rio Cabafias and the Rio Yagrumaje. In the field it
presents a striking appearance due to its miniature-tree habit and stiff
round leaves bluish above and yellowish beneath (cf. photograph taken
by Marie-Victorin, Contr. Inst. Bot. Univ. Montreal 68: 164. 1956).
Although so restricted in range, the plant was locally quite common,
associated with such characteristic Moa species as Scaevola wrightii and
Anastraphia recurva.

Sect. 23. Hemiphyllanthus (Muell. Arg.) Muell. Arg. Flora 1865: 370.
1865: DC. Prodr. 15(2): 323. 1866; emend. Webster, Contr. Gray
Herb. 176: 62. 1955.

Glochidion sect. Hemiphyllanthus Muell. ee Linnaea 32: 59. 1863; Pax &
offm. Natiirl. Pflanzenfam. 19c: 58. 193

Shrubs or small trees with phyllanthoid branching; axes incrustate or
tomentulose; branchlets bipinnatiform, leaves well-developed at least on
ultimate axes, leaf-blades membranous to coriaceous. Monoecious; cymules
unisexual or bisexual, produced only on ultimate axes of branchlet. Male
flower: calyx-lobes 5; disk-segments 5; stamens 2-6, filaments free or
united; anthers dehiscing obliquely or horizontally; pollen g grains areolate.

emale flower: calyx-lobes 5; disk cupuliform, dissected, or obscure;
styles free or connate, branches erect to reflexed. Capsule trigonous, cocci
fragile; seeds trigonous, verruculose or (in P. maleolens) smooth.

Type species: Phyllanthus ovatus Poir.

Geographically, ecologically, and morphologically, sect. Hemiphyllanthus
is one of the most distinctive groups of West Indian Phyllanthus. The
bicentric distribution of the section, the representatives of which occur in
two widely disjunct areas (southwestern Haiti and the Lesser Antilles),
raises most interesting problems regarding the past migrations of the group.
All of the six species of the section agree in being rain-forest calciphiles,
and they probably all have the palm-like (‘‘schopfbatimchen”) habit char-
acteristic of pioneer plants of rain-forest areas. The bipinnatiform branch-
lets have a distinctive tomentum of reddish multicellular scales except in
P. maleolens, the branchlets of which are merely incrustate.

The most specialized fern-like Lesser Antillean species, such as P.
mimosoides, are so distinctive in appearance that they do not appear to
belong with any other taxa. However, they can be related (via P. ovatus)
to the two representatives in Hispaniola, and the latter show a definite
resemblance to the Jamaican section Epistylium. Phyllanthus matcolens
of the present section has leaves and stipules similar to those of P. cauli-
florus in sect. Epistylium, and the flowers of P. maleolens are rather similar
to those of P. LPT i The most important morphological gap between
sects. Epistylium and Hemiphyllanthus is of course the difference between
their pinnatiform and bipinnatiform branchlets. Assuming that there is
a close relationship between P. cauliflorus and P. maleolens, it is easy to
suggest a hypothesis of the origin of the bipinnatiform branchlets of sect.

164 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxx1x

Hemiphyllanthus. Vf the axillary thyrses which sometimes appear on the
branchlets of P. cauliflorus were to become thickened, rigid, and leafy, the
result would be a bipinnatiform branchlet like that of P. maleolens, with
leaves on both orders of axes but with flowers confined to the ultimate
axes. Whatever the exact course of evolution, it is apparent that the phylo-
genetic origin of sect. Hemiphyllanthus is to be sought within sect. Epi-
stylium.

The disjunct distribution of the representatives of sect. Hemiphyllanthus
is reflected in the distinct morphological differences between the Hispani-
olan and Lesser Antillean species-groups, which could be defined as sub-
sections if any formal subdivision were worthwhile. The origin of the
section must have been to the west of the Lesser Antilles, possibly in
Hispaniola, and the radiation in which P. mimosoides has taken the leading
part must represent a later burst of evolution. It is notable that in sect.
Hemiphyllanthus, in contrast to such groups as sect. Orbicularia, the
species are much more sharply defined, only the most elementary taxonomic
judgment being required to distinguish them

Another aspect of the phylogeny of sect. Hemiphyllanthus, certainly of
no small interest, is its relationship to sect. Xylophylla. It is clear that the
compound phylloclade of most representatives of that group is homologous
with the bipinnatiform branchlet of the present section; and it is further-
more evident that the two sections are related and that Hemiphyllanthus
is the less specialized of the two (in vegetative characters, at least). How-
ever, this is not the same thing as saying that sect. X ylophylla was derived
from sect. Hemiphyllanthus. As discussed under the former, the relation-
ship of the two sections can perhaps best be stated as one of more or less
coordinate origin.

KEY TO THE SPECIES

1. Stipules of branchlets thickened, indurate, and blackish; leaf-blades well-
developed on both primary and ultimate oa axes; pedicel of female
flower slender and terete, 3 mm. long or lon

2. Branchlets merely incrustate; leaves cor en shiny above, 10-35 mm.
long; stamens 3, filaments free, anthers dehiscing obliquely upwards.

Se he wie Pete Nese SGN ner he See ee he Age aedeaen Be hae tie © maleolens
2. Branchlets reddish-tomentulose; leaves chartaceous, dull above, 4-7 mm.

long; stamens 2, filaments connate, anthers dehiscing horizontally.

Stem ateeye kanes ae ee ee ge ude aan heehee Bek 73. P. myriophyllus
Stipules of branchlets scarious, neither indurate nor blackened; leaves of
primary branchlet axes reduced to cataphylls; pedicel of female flower thick-
ened and fleshy or else less than 3 mm. long.

2. Leaves ovate or elliptic, symmetric at base, mostly 30-50 mm. long;
styles undivided, united into a column 2.5—3.5 mm. high which is exserted
DEYOUNG ING WAIVE. oo. ccts anh eb ny baa a Wee eo bw eee e eos 74. P. ovatus

2. Leaves cc aero at base, — mm. long; styles definitely bifid, column
shorter and not exserted from
3. Pedicel of female flower at ek dilated into an incrassate recepta-

—

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 165

cle usually broader than the calyx; styles fused into a massive column
higher than the ovary; cataphylls of primary branchlet axis densely
tomentulose when young; ultimate axes of branchlet with mostly
15-25 leaves, blades mostly 18-30 mm. long. ..... 75. P. megapodus

_ Pedicel of female flower more slender; styles free or shortly connate,
spreading or reflexed; cataphylls of primary branchlet axis glabrous;
ultimate axes of branchlet with mostly 30-60 leaves (or more), blades
5-13 mm. long.

Pedicel of female flower slender, calyx-lobes 1.5-3 mm. long,

Ww

4.
spreading; styles spreading; stamens 3; leaves smooth beneath.
ie ee Ree OR pe ee oh ee eee 76. P. mimosoides
4. Pedicel of female flower incrassate above, calyx-lobes not over

mm. long, erect; styles reflexed and appressed to outside of
calyx; stamens usually 5; leaves minutely scabridulous beneath.
Be Aten Seite ar abd Mache hate cme a eee 77. P. acacioides
72. Phyllanthus maleolens Urb. & Ekm. Ark. Bot. 22A (8): 60. 1928.

(PLATE XXXI, figs. A-B).

Slender tree 2-8 m. high, with unbranched trunk, evil-smelling (fide
Ekman); main axis reddish-incrustate. Deciduous branchlets bipinnati-
form; primary axis 25-50 cm. long, (2—) 3-4 mm. thick, leafy (but leaves
often very soon deciduous), ferruginous-incrustate on younger parts
(smoother and greyish in age), terete, with c. 12-30 nodes; first internode
c. 325 cm. long, median internodes c. 1-3 (—4) cm. long. Leaf-blades of
primary axis similar to those on ultimate axes but rather smaller; stipules
persistent, more or less spreading or reflexed at the tip, indurate, broadly
ovate-triangular (often broader than long), c. 1.5-2.5 mm. long, 1.5-3.5
mm. broad, blunt-tipped, greyish. Ultimate axes ascending, mostly 6-13
cm. long (occasional axes only 3-5 cm. long), c. 1.3-2 mm. broad, reddish
brown and incrustate, more or less flattened, sharply wing-angled between
the stipules, with 10-25 (—30) leaves; first internode 2.5-5 mm. long,
median internodes 3.5-6 mm. long (internodes up to 10-12 mm. long on
sterile axes). Leaves: stipules persistent, more or less reflexed, becoming
blackish and somewhat indurate, c. 1.2-2.2 mm. long, bluntly pointed,
more or less decurrent. Petioles reddish brown, plano-convex, 1-2 (—2.5
mm. long. Leaf-blades subcoriaceous, elliptic or slightly obovate, (10—)
15-35 mm. long, (5—) 10-20 (-25) mm. broad, retuse at the tip (the
minute apiculum of young leaves deciduous), symmetrically acute or
obtuse at the base; above olivaceous, lucid, the nerves (except the midrib)
inconspicuous; beneath paler, yellowish green, the midrib conspicuous and
raised, the lateral veins (c. 4 or 5 on a side) raised, somewhat crooked,
giving off a few subanastomosing veinlets; margin usually narrowly revo-
lute.

Monoecious; cymules axillary, on ultimate axes of branchlet, bisexual,
of 1 female and up to 8 or 10 male flowers; bracteoles thickened and indu-
rate.

166 JOURNAL OF THE ARNOLD ARBORETUM _[vor. xxxrx

Tale flower: pedicel slender, 4-7 mm. long. Calyx yellowish, some-
times reddish-tinged; calyx-lobes 5, chartaceous, broadly elliptic or obo-
vate, convex and sometimes cucullate, somewhat unequal, 1.2-1.5 mm.
long outer lobes c. 0.75-1 mm. broad, inner lobes c. 1-1.4 mm. broad,
rounded at the tip, entire, midrib sparsely branched or unbranched in
smallest lobes. Disk-segments 5, suborbicular, entire, slightly thickened,

0.25—0.4 mm. broad. Stamens 3; filaments free, up to 0.4—0.5 mm. long
or anthers appearing subsessile; anthers ovate, rounded and emarginate
at the tip, c. 0.5—-0.7 mm. long, 0.6-0.9 mm. broad; anther-sacs subparallel
or divergent, dehiscing longitudinally (more or less obliquely), slits apically
contiguous but not confluent; pollen grains c. 22.5 p in oe with c.
9-11 polybrochate areoles per amb, germ-pores conspicu

Female flower: pedicel 6-10 mm. long, slender, mens ee angled
above, straight or curved. Calyx-lobes 5, chartaceous, spreading at anthesis
(more or less reflexed in fruit), inequel. oblong to obovate, 1-1.2 mm.
long, 0.6-1 mm. broad, obtuse at the tip, entire, midrib inconspicuously
branched. Disk somewhat fleshy, 5-angled, crenulate. Ovary oblate, sul-
cate; styles free or slightly connate at the base, erect, 0.7 mm. high, bifid,
sharply bent at the crotch, style-branches somewhat dilated and flattened,
one or both toothed or again bifid

Capsule oblate, trigonous, c. 5 mm. in diameter, dark reddish brown,
obscurely rugulose. Columella stout, c. 1.3-1.5 mm. high and nearly as
broad. Seeds asymmetrically trigonous (plano-umbonate), c. 2.5-2.7 mm.
long, 1.7-2 mm. radially and tangentially, light brown, smooth; hilum
submedian, ovoid, partly extending onto the lateral face, c. 0.6-0.7 mm.
long and 0.4-0.5 mm. broad.

Collected in flower June-Sept.; in fruit July, Sept.

Type: Haiti, Ekman H6849.

DIsTRIBUTION: mountains, southwesterr Hispaniola (Map XXXI"").

HAITI. Ouest: Morne des Commissaires: Grand-Gosier, Morne Sincilio, alt
c. 1300 m., 3 Sept. 1926, Ekman H6849 (S, HoLoTYpPE); Grand-Gosier, Ravine-
Fanchon, alt. c. 1675 m., 4 Sept. 1926, Ekman H6880 (A, S); Boucan Chat, alt.
1600 m., 10 June 1942, Holdridge 1266 (GH. a Mare Sal, alt. 1600 m.,
28 July 1942, Holdridge 1381 (GH, MICH, MO, US

DOMINICAN REPUBLIC. BARAHONA: eae de Bahoruco, Sierra de
los Comisarios, above Gros-Figuier, alt. 1500 m., 29 Sept. 1926, Ekman H6782
(S)

The offensive odor from which this species takes its name was noted
by Ekman only for the plant encountered in Barahona, and was not men-
tioned by Holdridge. Since the production of a distinct odor by the vege-
tative part of the plant (Ekman’s specimen bears no flowers) is a distinctly
unusual, if not unique, character in Phyllanthus, Ekman’s observation
needs to be confirmed. The formation of an unpleasant odor at night has
been reported for P. epiphyllanthus, but this appears to be associated with
flowering.

The citation of specimens from both Hispaniolan republics is somewhat

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 167

misleading, for the total range of P. maleolens is apparently confined to
the Morne des Commissaires, and it barely crosses the international line
into the Dominican Republic. According to the collectors’ notes, the
species grows in “Laubwald” (broadleaf forest) on calcareous soil at higher
altitudes than those known for any other West Indian species of subg.
Xylophylla.

The closest relative of P. maleolens is certainly its counterpart in the
Massif de la Hotte, P. myriophyllus. That plant, however, not only differs
in its smaller flowers with shorter pedicels, dimerous androecium, and re-
flexed styles, but vegetatively — by virtue of its smaller leaves and tomen-
tulose axes — resembles the Lesser Antillean species. Distinctly similar

P ACACIOIDES
P MALEOLENS
P MEGAPODUS
P MIMOSOIDES

P MYRIOPHYLLUS
P OVATUS

©

@®oeo0od

Map XXXIII. Distribution of the species of sect. Hemiphyllanthus.

to P. maleolens in several respects is P. ovatus of Martinique, which has
however, completely different female flowers with highly modified styles.

Since in many respects — notably its lack of tomentum, large leaves,
free stamens, and unmodified female flower — P. maleolens is the least
specialized of the members of section Hemiphyllanthus, its extra-sectional
affinities are of particular interest. There is no doubt that these affinities
are all with the Jamaican sect. Epistylium, despite the fact that all three
species of that group have simply pinnatiform branchlets. Each one of the
three species shows some similarity to P. maleolens, P. cauliflorus in its
stipules, P. cladanthus in its comparatively long pedicels, and P. axillaris
in the flattened branchlets. It would appear, therefore, that P. maleolens
evolved from the common ancestor of the three living species of sect.
Epistylium.

168 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxix

73. Phylanthus myriophyllus Urb. Ark. Bot. 17(7): 36. 1921.
(PLATE XXXI, figs. C-D).

Slender shrub or tree c. 2-3 m. high, with unbranched trunk. Deciduous
branchlets bipinnatiform; primary axis 30-80 cm. long, 3-4 mm. thick,
leafy (leaves subpersistent), densely scurfy with ferruginous tomentum
(becoming glabrate on older parts), terete, with 40-110 nodes; first inter-
node 10-25 mm. long, median internodes 5-10 mm. long. Leaf-blades of
primary axis precisely as on ultimate axes; stipules at first broadly tri-
angular and reddish brown, c. 1.5-2.5 mm. long, the tip deciduous, the
blackish, massive, indurate base persistent, c. 1.5—2.5 mm. broad. UIti-
mate axes ascending or spreading, (5—) 7-14 cm. long, 0.7—1 mm. broad,
chestnut-brown, conspicuously fluted with obtuse ribs decurrent from the
stipule-bases, tomentulose between the ribs with simple pale to dark
ferruginous hairs, with 25-70 leaves; first internode c. 1 mm. long, median
internodes 1.5—2.5 mm. long. Leaves: stipules persistent, decurrent at the
base, at first ovate-lanceolate and c. 1 mm. long (with denticulate margin),
later the tip inflexed and base incrassate, blackish, indurate, c. 0.7-1 mm.
broad. Petioles reddish-brown, 0.3—0.4 mm. long. Leaf-blades firmly char-
taceous, smooth on both sides, broadly elliptic-oblong or ovate-oblong,
(4—) 5-7 mm. long, (2.5—) 3—-4.5 mm. broad, obtuse or bluntly apiculate at
the tip, symmetrical and obtuse to truncate or subcordate at the base;
above dark olivaceous, dull, the nerves completely obscure or the plane
midrib visible; beneath pale green, lucid, the midrib raised, the lateral
veins (c. 4 or 5 on a side) ascending, inconspicuous, veinlets not visible:
margins slightly thickened and subrevolute.

Apparently monoecious (possibly sometimes dioecious); cymules axil-
lary, on ultimate axes of branchlet, the proximal with 3-5 male flowers,
the distal with some male and some bisexual cymules (each of the latter
with a single female flower).

Male flower: pedicel capillary, 2-4 mm. long. Calyx yellowish white:
calyx-lobes 5, thin and scarious, obovate or suborbicular, unequal, c. 1-1.2
mm. long, 0.7—1.1 mm. broad (the outer narrower), obtuse or acute, entire
or obscurely crenulate, midrib simple or sparsely branched. Disk-segments
5, oval, thin and flattened, entire, c. 0.25 mm. across. Stamens 2; fila-
ments connate into a column c. 0.3 mm. high; anthers subsessile, broadly
ovate, rounded and minutely emarginate at the tip, c. 0.3 mm. long and
0.5 mm. broad; anther-sacs divergent, dehiscing more or less horizontally,
the slits not apically confluent; pollen grains c. 16-18 p» in diameter, with
c. 5—7 polybrochate areoles per amb.

Femaie flower: pedicel 3-3.5 mm. long at anthesis, slender, only very
slightly and gradually broadened upwards, terete, smooth, reddish brown.
Calyx yellowish white; calyx-lobes 5, membranous, more or less spreading,
unequal: outer lobes obovate-oblong c. 1 mm, long and 0.6—-0.7 mm.
broad; inner lobes suborbicular, c. 1.2 mm. long and broad; lobes entire,
midrib unbranched. Disk conspicuous, fleshy, 5-angled, crenulate. Ovary
oblate, shallowly sulcate; styles slightly connate at the base, flattened,

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 169

dilated, c. 0.7 mm. long, parted c. 34 to 4% their length, the branches re-
flexed over the ovary (but remaining inside the outspread calyx), tapering
to acute tips. Fruit unknown.

Collected in flower May, Nov.

Type: Haiti, Ekman H140.

DISTRIBUTION: mountains, western end of southern peninsula of Haiti
(Map XXXIII)

HAITI. Sup, Massif de la Hotte: along stream, northwest slopes, montane
forest, alt. c. 800 m., 10 May 1917, Ekman H140 (S. HoLotyPE); Camp-Perrin,
northern slope of Morne Vandervelde, in “Jardins Coutard”, laterites on erup-
tive, alt. c. 900 m., 29 Nov. 1925, Ekman H5185 (S, US); Pestel, rocky ridge of
M. Delcour, alt. 1000 m., 27 Aug. 1927, Ekman "H9007 GS

Not only is P. myriophyllus interesting as yet another example of the
remarkable endemism of the Massif de la Hotte, but it is phylogenetically
significant in being transitional between the large-leaved species P. maleo-
lens and the “mimosoid” species of the Lesser Antilles. It seems unlikely,
however, that P. myriophyllus can be regarded as directly ancestral to
P. mimosoides or its relatives; it rather illustrates a parallel reduction in
leaf-size to that which can be traced from P. ovatus to the small-leaved
Antillean species.

74. Phyllanthus ovatus Poir. in Lam. Encycl. Method 5: 297. 1804
(as P. ovata); Muell. Arg. in DC. Prodr. 15(2): 323. 1866.
(PLATE XXXI, figs. E-F).
Phyllanthus grandifolius sensu Spreng. Syst. Veg. 3: 22. 1826; non L.
idion ovatum (Poir.) Muell. Arg. Linnaea 32: 71. 1863; Pax & Hoffm.
Natiirl. Pflanzenfam. 19c: 58. 1931.
Diasperus ovatus (Poir.) O. Ktze. Rev. Gen. 2: 600. 1891.

Shrub 1-4 m. high, with usually unbranched trunk; stem ferruginous-
tomentulose at the apex, becoming more or less glabrate below. Cata-
phylls inconspicuous, indurate, very similar to those of main axis of
branchlet. Deciduous branchlets bipinnatiform; primary axis (8—) 10-25
cm. long, c. 2.5-3 mm. thick, distally ferruginous-tomentulose, proximally
sparsely hairy or glabrate, subterete, with 10-20 (—35) leaves; first inter-
node 20-60 mm. long, median internodes mostly 10-20 mm. long. Leaves
of primary axis reduced to cataphylls: stipules persistent, spreading or
somewhat reflexed, indurate, ovate-triangular, 1-2 mm. long, 1.3-2.3 mm.
broad, obtusely pointed, brownish or greyish; blade narrower, 1—1.2 mm.
long. Ultimate axes ascending, when well developed c. 5-15 cm. long,
1-1.5 mm. broad, olivaceous, angled or somewhat flattened, smooth and
glabrous, with mostly 5-20 (—30) leaves; first internode mostly 5-15 mm.
long: median internodes mostly 5-20 mm. long. Leaves: stipules persistent,
sometimes reflexed, indurate, triangular-lanceolate, mostly 1-1.5 mm.
long, acuminate, entire. Petioles not sharply set off from blade, c. 0.5—1.5
mm. long. Leaf-blades chartaceous, smooth on both sides, symmetrically

170 JOURNAL OF THE ARNOLD ARBORETUM | [vor. xxxrx

or asymmetrically elliptic or ovate, mostly 30-50 mm. long and 20-35
mm. broad, sometimes (especially on distal part of branchlet) only 15-30
mm. long and 10-20 mm. broad, acute or subacute at the tip, acute or
obtuse at the base; above olivaceous or plumbeous, minutely foveolate,
midrib incised and prominent but other veins obscure; beneath paler, more
or less smooth, the midrib conspicuous and raised, the lateral veins (c. 5—7

a side) somewhat raised, ascending, straight, veinlets obscure; margins
unthickened, plane or narrowly recurved.

Tonoecious; cymules axillary, on ultimate axes of branchlet; proximal
cymules male and distal ones bisexual (of 1 female and c. 6-8 male
flowers), or all bisexual.

Male flower: pedicel thickened upwards, 1.5-3.5 mm. long. Calyx dry-
ing reddish brown; calyx-lobes 5 (rarely 6), chartaceous, oblong-obovate
to suborbicular, subequal, c. 1.5—1.8 mm. long, 1.1—1.6 mm. broad, obtuse
or rounded at the tip, entire with narrow pale scarious margins, midrib
unbranched. Disk-segments 5, more or less reniform, entire, somewhat
thickened, inconspicuously pitted, c. 0.4-0.6 mm. across. Stamens 3
(rarely 4); filaments completely connate into a column c. 0.4-0.6 mm.
high; anthers sessile atop the column, fused by their connectives (androe-
cium plane on top, 3-lobed), triangular-ovate, obtuse, c. 0.4-0.6 mm.
broad; anther-sacs discrete, divergent, dehiscing horizontally; pollen grains
c. 22-25 w in diameter, with c. 15 oligobrochate areoles per amb.

Female flower: subsessile at anthesis, the thick fleshy obpyramidal
pedicel lengthening in fruit up to 1.52.2 mm. long. Calyx-lobes 5,
coriaceous, erect at anthesis (spreading or reflexed in fruit) strongly im-
bricate, more or less unequal, ovate or elliptic, 2-2.7 mm. long, 1.5-2.5
mm. broad, obtuse at the tip, midrib apparently unbranched. Disk some-
what fleshy, crenulate or notched, dark reddish brown. Ovary oblate; styles
completely united into a thick fleshy urn-shaped column c. 2.5—3.3 mm.
high (far exceeding the ovary); stigmas (undivided style-tips) triangular,
obtuse, more or less spreading, subentire, c. 0.4—-0.7 mm. long.

Capsule not seen entire; valves c. 4 mm. long, dark reddish brown, not
veiny. Columella 2.7 mm. long. Seeds (not seen fully mature) somewhat
asymmetrically trigonous, c. 4 mm. long and 2.5 mm. radially and tan-
gentially, light brown with irregular rows of slightly raised dark reddish-
brown dots; hilum ovate-triangular

Collected in flower Feb., Mar., July, Aug.; in fruit, Feb., July

Typr: Martinique, Herb. Poiret.

DistTRIBUTION: forested regions, Martinique (Map XXXIII).

MARTINIQUE: without specific locality, Herb. Lamarck (G, P); Herb.
Poiret (P, TYPE COLLECTION); Sieber Fl. Martin. 224 (F, GOET, MO, P, W);
Hauteurs de la Grand-Riviére, de Case-Pilote, Morne-Rouge, 1880, 1899, Duss
53 (F, GH, MO, NY, near 4057 (NY, US); Hauteur de la Case-Pilote, July
1870, Hahn 323 (K, P); Case-Naoire, collines peu buisées, Feb. 1868, Hahn
406 (A, G, P); same a ts Feb. 1869, Hahn 643 (P); taillis a lisiers forestiers
et alturales, Morne Vert, Bernadette, alt. 560 m., 17 July 1942, Stehlé 5065 (F).

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS byl

In its occurrence in moist forested regions of northern Martinique,
P. ovatus conforms with the ecological preference of other species of the
section. According to Stehlé it is found at the margin of the forest, and
is thus heliophilous as well as mesophytic. The species is certainly very
distinct but is perhaps most closely related to P. maleolens, from which
it differs in its tomentulose axes, thinner more pointed leaves, thickened
female pedicel, and connate nearly entire styles. Also related to P. ovatus
is P. megapodus, which has somewhat similar female flowers but which
vegetatively is much closer to P. mimosoides and P. acacioides.

As was pointed out previously (Contr. Gray Herb. 176: 62. 1955),
the referral of P. ovatus to Glochidion by Pax and Hoffman on the basis
of its undivided styles is unjustified, since it contradicts all other indica-
tions of affinity. Mueller was correct in reversing himself as to the generic
disposition of this species, even though he did not place it in the proper
circle of relationship. There is indeed a certain similarity between the
stylar column of P. ovatus and the much shorter one of P. botryanthus (the
species associated with it by Mueller), but the two species are only dis-
tantly related, and the gynoecial resemblance must be ascribed to parallel
evolution.

75. Phyllanthus megapodus Webster, Contr. Gray Herb. 176: 62-63.
1955: (PLATE XXXI, figs. G-H).

Phyllanthus mimosoides [lusus] macrophyllus Muell. Arg. in DC. Prodr.
15(2): 381. 1866; non Phyllanthus macrophyllus Muell. Arg. Flora 1865:
3/0; 1869.

Shrub or small tree up to 3-5 m. high, presumably with the habit of
P. mimosoides. Cataphylls indurate, densely tomentulose: stipules and
blade triangular-lanceolate, recurving, c. 2.5-3 mm. long. Deciduous
branchlets bipinnatiform; primary axis 20-40 cm. long, 2-3 mm. thick,
sparsely tomentulose and soon glabrate, smooth, terete, with (6—) 10-18
nodes; first internode 50-130 mm. long, median internodes mostly 15-40
mm. long. Leaves of primary axis reduced to cataphylls: stipules per-
sistent, appressed or reflexed, subindurate, triangular-lanceolate, c. 2.5—4
mm. long, 1.5-2.5 mm. broad, acuminate (but tip often broken off),
auriculate at the base, densely reddish-tomentulose (when young; more or
less glabrate in age); blade lanceolate, usually spreading or reflexed,
reddish-tomentulose, c. 2-3 mm. long. Ultimate axes ascending, (5—) 7-15
(-18) cm. long, 0.8-1.3 mm. broad, olivaceous or stramineous, flattened
ventrally (above) and angular-convex dorsally (beneath), with ribs de-
current from stipules on both sides, smooth and glabrous, with (13-)
15-25 (-27) leaves; first internode (1—) 3-7 mm. long, median inter-
nodes 4—9 mm. long. Leaves: stipules more or less deciduous, not reflexed,
scarious, lanceolate, c. 1-1.8 mm. long, reddish, more or less ciliate at apex
and on margins (tomentulose when young), decurrent at the base. Leaf-
blades subsessile (petioles ill-defined, c. 0.5 mm. long or less), chartaceous,
smooth on both sides, asymmetrically broadly oblong or obovate, sub-

ees JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxtx

falcate, (12—) 18-30 mm. long, (5—) 8-15 mm. broad, obtuse or rounded
and apiculate at the tip, at base oblique on adaxial side; above olivaceous,
minutely foveolate, veins inconspicuous; beneath midrib conspicuously
raised, lateral veins (c. 7 or 8 on a side) straight and slightly raised, vein-
lets completely obscure; margins not especially thickened, plane or re-
curved

Monoecious; cymules axillary, on ultimate axes of branchlet; proximal
cymules with 1-5 male flowers; distal cymules either male or of a solitary
female flower, or sometimes bisexual (with 1 female and 1 or 2 male
flowers).

Male flower: pedicel slender, 3-5 mm. long. Calyx whitish; calyx-lobes
5, membranous, somewhat spreading, broadly elliptic to obovate, sub-
equal, c. 1.8-2.5 mm. long, 1.3-2.1 mm. broad, obtuse or rounded at the
tip, entire, midrib unbranched. Disk-segments 5, roundish, rather thin,
entire, c, 0.3-0.5 mm. across. Stamens 3; filaments connate into a column
c. 0.5—0.8 mm. high; anthers subsessile atop the column, discrete, emar-
ginate, c. 0.3 mm. long and 0.6-0.7 mm. broad; anther-sacs spheroidal,
divaricate, dehiscing horizontally, slits not apically confluent; pollen grains
c. 21-24 p» in diameter, with c. 15 areoles per amb.

Female flower: pedicel at anthesis c. 1.5-2 mm. long, terete and slender
at the very base but above greatly dilated, obpyriform, fleshy, as broad
as or broader than the calyx; fruiting pedicel c. 3-4 mm. long and 1.3-2.2
mm. broad (when dried). Calyx-lobes 5, thick and fleshy, at anthesis
stiffly erect and imbricate-connivent around the ovary, elliptic to ovate
(becoming obovate-oblong in fruit), c. 1.8-2.5 mm. long and 1.2—2 mm.
broad, obtuse at the tip, entire, midrib unbranched. Disk nearly obsolete
(scarcely 0.1 mm. high, or less, with 5 small points alternating with the
calyx-lobes). Ovary oblate-spheroidal, smooth; styles erect, united into a
massive column c, 2 mm. high and nearly as broad; style-branches reduced
to blunt apical projections or becoming up to 1 mm. long.

Capsule not seen entire; valves c. 5.5 mm. long. Columella massive,

PLATE XXXI. MALE AND FEMALE FLOWERS OF SECT. Hemiphyllanthus AND
secT. Xylophylla.

Fics. A-B. Phyllanthus maleolens Urb. & Ekm., (Ekman H6849 [S]). Fics.
C-D. Phyllanthus myriophyllus Urb. (Ekman H5185 [S]). Fics. E-F. PAyl-
lanthus ovatus Poir. (Duss 53 [GH]). Fics. G-H. Phyllanthus megapodus

[MO]). Fics. I-J. Phyllanthus mimosoides Sw. (Hodge & Hodge 3439 [GH]).
Fics. K-L. Phyllanthus acacioides Urb. (Broadway 4189 [MO]). Fics. M-N.
Phyllanthus montanus (Sw.) Sw. (Proctor 11594 [GH]). Fics. O-P. Phyllan-
thus latifolius Sw. (Webster 4875 [A]). Fries. Q-S. Phyllanthus arbuscula
(Sw.) Gmel. Q-R, male and female flowers of Race A (Harris 9191 Us|).
S, female flower of Race C (Stearn 387 [JAM]). Fics. T-U. Phyllanthus
angustifolius (Sw.) Sw. (Webster 5558 [A]). Fics. V-W. Phyllanthus proctoris
Webster (Macfadyen [K]). Fics. X-Z. Phyllanthus epiphyllanthus L. X-Y,
male and female flowers of ssp. epiphyllanthus (Nash & Taylor 875 [NY]).
Z, female flower of ssp. dilatatus (Muell. Arg.) Webster (Ekman 10271 [S]).

PLATE XXXI

Jour. ARNOLD Ars. VoL. XXXIX

WEBSTER, WEST INDIAN PHYLLANTHUS

174 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxix

2—2.8 mm. long. Seeds (not seen mature) slightly over 3 mm. long, with
evenly spaced slightly raised reddish-brown dots

Type: Martinique, Sieber Fl. martin. 396.

DistTRIBUTION: Dominica and Martinique, probably in upper montane
rain-forest (Map XXNIII).

DOMINICA: forest clearings or ag river, Fon he a lesser peak of the
western ridge of Morne Diablotin, alt. c. 1000 m., 14 SS 940, W. & B. Hodge
2841 (GH); forest, Hampstead, 1903, ties 640 (F, N

MARTINIQUE: Sieber Fl. martin. 396 (W, HOLOTYPE; BR, G, L, MO, P,
ISOTYPES),

In Dominica P. megapodus is apparently a rare plant of the upland rain
forests, where it appears to replace the much more common lowland species,
P. mimosoides. It is not yet clear, however, whether the replacement is so
complete that the two species are allopatric. The record from Martinique
requires confirmation, since it is not certain that all the plants distributed
by Sieber under the “Flora Martinicensis” or ‘Flora Trinitatis” labels
were actually collected on those two islands. Of course, if P. megapodus
is as uncommon on Martinique as on Dominica it may have escaped detec-
tion since the visit of Sieber’s collector, Kohaut.

Although in aspect it resembles P. mimosoides so markedly that Mueller
interpreted it as merely a large-leaved form, P. megapodus has very differ-
ent female flowers which are much more like those of P. ovatus. Vegeta-
tively P. megapodus is well distinguished from P. mimosoides having fewer
than 20 ultimate axes per branchlet instead of 30-60 or more, tomentulose
branchlet cataphylls, and larger leaves. From P. ovatus, P. megapodus
differs in its non-exserted stylar column, smaller asymmetrical leaves, and
especially in its hypertrophied female pedicel. In the last feature P. mega-
podus appears to represent the generic extreme, although it is somwhat
approached by the thickened pedicel of P. acacioides.

76. Phyllanthus mimosoides Sw. Prodr. 27. 1788; Fl. Ind. Occ. 1101—
1102. 1800; Muell. Arg. in DC. Prodr. 15(2): 380-381. 1866.
(PLATE XXXI, figs. I-/).

Diasperus mimosoides (Sw.) O. Ktze. Rev. Gen. 2: 600. 1891.

Shrub c. 1-5 m. high, with a slender usually unbranched trunk and a
terminal crown of leafy branchlets; stem apex ferruginous-tomentulose,
older parts glabrate. Cataphylls indurate: stipules triangular, recurving,
c. 2—2.5 mm. long, becoming acropetally displaced to flank the base of the
branchlet; blade c. 1.5 mm. long. Deciduous branchlets bipinnatiform;
primary axis (20—) 30-70 (—100) cm. long, 2-4 mm. thick, more or less
ferruginous-tomentulose (proximally becoming glabrate), terete, with (20—)
30-60 (—100) nodes; first internode (25—) 40-100 (—130) mm. long,
median internodes (5—) 7-15 (—20) mm. long. Leaves of primary axis
reduced to cataphylls: stipules persistent, not reflexed, more or less indu-
rate, lanceolate, (2—) 2.5-4 mm. long, 0.7-1.3 mm. broad, acuminate,

1958] WEBSTER, WEST INDIAN PHYLLANTHUS LeRS

stramineous or reddish, with a dark sharply demarcated subtriangular
fleshy basal area; blade thinner, scarious, linear-lanceolate, attenuate-
acuminate, sometimes obscurely marginally ciliate below, mostly 2-3.5
mm. long and 0.2-0.5 mm. broad. Ultimate axes ascending, (3—) 5-10
(-12) cm. long, 0.5—0.9 mm. broad, olivaceous, flattened and angled with
sharp ridges decurrent from stipules, smooth and glabrous (except for one
or few tufts of hairs often present on internodal part of dorsal ridge), with
(20—) 30-60 (—80) leaves; first internode c. 1-1.5 (—2) mm. long, median
internodes c. 1-3 mm. long. Leaves: stipules persistent, not reflexed,
scarious, lanceolate, mostly 1-1.5 (—1.8) mm. long, acuminate, decurrent
at the base (adaxial margin running down center of the axis as a ridge).
Leaf-blades subsessile (petioles only 0.2-0.4 mm. long), thinly chart-
aceous, smooth on both sides, asymmetrically oblong or oblong-obovate
and often falcate, c. 5-11 (-13) mm. long, 2-4 (-6) mm. broad, mostly
rounded or subtruncate and apiculate at the tip, at base abruptly oblique
on adaxial side and straight on abaxial side; above olivaceous, minutely
foveolate, veins obscure; beneath pallid, midrib slightly raised and running
out into apiculum, lateral veins (4-6 on a side) ascending, straight, in-
conspicuous or quite obscure; margins unthickened, plane or slightly
revolute.

Monoecious; cymules axillary, on ultimate axes of branchlet; proximal
cymules with usually 1 or 2 male flowers; distal cymules male or a few
with a single female flower, this solitary or accompanied by 1 or 2 males;
bracteoles fimbrillate.

Male flower: pedicel capillary, 1-3.5 mm. long. Calyx whitish; calyx-
lobes 5, thin and scarious-membranous, spreading, oblong-elliptic to broadly
elliptic (outer lobes narrower), (0.8—) 1-1.5 mm. long, c. 0.5-1.3 mm.
broad, obtuse at the tip, entire, midrib unbranched (or branches very
obscure), usually somewhat raised on ventral surface. Disk-segments 5,
roundish, flattened, smooth, entire, c. 0.2-0.3 mm. across. Stamens 3;
filaments connate into a stout or slender column 0.2—0.8 mm. high; anthers
subsessile atop the column, discrete, very broadly ovate, obtuse, emar-
ginate, 0.2-0.3 mm. long, 0.3-0.6 mm. broad; anther-sacs divaricate, de-
hiscing horizontally, the slits apically contiguous, discrete or confluent;
pollen grains 16-18 » in diameter, with c. 7-10 polybrochate areoles
per amb.

Female flower: pedicel scarcely evident at anthesis, in fruit becoming
0.5-1.5 (—2) mm. long, smooth, terete, not markedly dilated above. Calyx
whitish; calyx-lobes 5, thin and membranous or scarious, spreading from
anthesis onwards, subequal, elliptic-oblong to obovate or spathulate, c.
1.5-2.5 mm. long (or up to 3 mm. long in fruit), 0.4-1.5 mm. broad,
obtuse and entire at the tip, midrib unbranched. Disk divided into 5
reddish-brown linear to filiform segments 0.2-0.5 mm. long (or the smaller
one or two reduced to short points). Ovary spheroidal, deeply sulcate;
styles united below into a column 0.3-0.8 mm. high (or sometimes nearly
free), ascending or spreading, the free ends c. 1.5—-2 mm. long, divided %

176 JOURNAL OF THE ARNOLD ARBORETUM _ [vor. xxxrx

to 4, their length into two slender divergent, often twisted terete branches
with bluntly pointed tips.

Capsule oblate-spheroidal, c. 4-5 mm. in diameter, olivaceous or dark
reddish-brown, not veiny. Columella slender, 1.2-2.5 mm. long. Seeds
trigonous, sometimes rather asymmetric, 1.52.5 mm. long, 1.2-2 mm.
radially and tangentially, light brown with evenly spaced slightly raised
reddish-brown dots; hilum rounded-triangular, 0.7-1 mm. across.

Collected in flower Feb.—Oct.; in fruit Feb.—Aug.

Type: Antigua, de Ponthieu (ex Swartz, 1788). The type collection
is probably represented by the Ponthieu collection in the herbarium of
the Conservatoire Botanique (G), and by a fragment in Swartz’s herba-
rium (S); but neither of these has any definite indication of locality. A
sheet collected by Ponthieu on Dominica and cited by Swartz in the “Flora
Indiae Occidentalis” was examined at the British Museum; but this can
hardly be considered the type. Since the sheet at Geneva has been an-
notated as “peut-etre l’echantillon dulequel Swartz a fait la description”
and since it is a more ample specimen than that at Stockholm, it is here
chosen as the lectotype.

DISTRIBUTION: rain-forest areas, Lesser Antilles (Map XXXIII).

ANTIGUA: de Ponthieu (G, LEcToTYPE; S, IsoTYPE; both presumably from
Antigua).

_ MONTSERRAT: Ryan (BM, C); Fergus Mountain, Shafer 340 (F, NY);
West (W).

GUADELOUPE: without specific locality, Duchassaing (C, GOET, P),
L’Herminier (P), Perrottet (G), Quentin 389 (P), Richard (P); Bassin Bleu,
Duss 208 (P); Bois des Bains-Jaunes, du Matouba, de Bouillante, etc., alt. 450—-
990 m., Duss 2445 (F, GH, NY, US); Matouba, Forsstrom (S); Sofaya, above
Ste. Rose, alt. 700 m., Holdridge 449 (NY); Ste. Rose, alt. 20 m., Questel 852
(US); Plateau du Palmiste, aoiedegied 4669 (A, P); mornes basaltiques, Honel-
mont, Stehlé 205 (US); hauteurs de Vernon, Petit-Bourg, alt. 450 m., Stehlé
266 (A, S, US); Comperon at Font arabe, alt. 200 m., Stehlé 2630 (U S),

DOMINICA: without specific locality, Bryant (NY), Fishlock 35 (NY), Im-
ray 315 (GOET), de Ponthieu (BM), Ramage (BM); St. Hilaire Trace, alt.
300 m., J. S. Beard 645 (A, US), P. Beard 1460 (S); Sugarloaf prope Prince
Ruperts, Eggers 770 (G, GOET, P, W), 1064 (US); bank of St. Mary’s River, at
mouth of Pegoua River, Hatton Garden Estate, near sea level, W & B. Hodge
3075 (GH); along stream 1 mile north of Calibishie, W. & B. Hodge 3172
(GH); Carib trail from Salybia to Hatton Garden, W. Hodge 3219 (GH);
Pegoua River, Deux Branches, Concorde Valley, W. & B. Hodge 3439 (GH);
Hampstead River, c. 2 miles from mouth, La Chaudiére, alt. 100 m., W. & B.
Hodge 3558, 3662 (GH).

MARTINIQUE: dans le haut de la riviére de la Grande-Riviére, Duss 2045

TRINIDAD: Sieber FI. Trinitatis 153 (MO, P, W).

By far the commonest and most widespread species of the section, P.
mimosoides is apparently the only one which has been taken into cultiva-
tion, where it has attracted the interest of morphologists; illustrations

1958] WEBSTER, WEST INDIAN PHYLLANTHUS Le)

showing its distinctive habit have been published by Goebel (Organogr.
Pflanz. 84. 1898) and Troll. The species is abundant on Guadeloupe,
where it is called “‘batard de fougere”, and Dominica, where it is referred
to as “tamarind grand bois”. According to Beard (Nat. Veg. Windward
& Leeward Isl. 106. 1949) it forms an extensive ground stratum in the
secondary rain-forest on Montserrat. On Martinique, in contrast, Duss
(Fl. Phanerogam. Ant. Fr. 23. 1897) found it very rare; and the species
has presumably become extinct on Antigua. The record of Wickstrom
(Kgl. Vet. Acad. Handl. [Stockholm] 1825: 423. 1826) from St. Bar-
thélemy, based on a Forsstrom collection, is surely erroneous and perhaps
was based on misdetermined specimens of P. amarus which are preserved
in the Riksmuseet, Stockholm. One would in any event not expect a meso-
phytic species stich as P. mimosoides to occur on a low barren island such
as St. Barthélemy. The collection from Trinidad, Sieber 153, offers greater
difficulties. It is possible that this record is correct, but unless the collec-
tion was made from a cultivated plant it represents a remarkable range
extension. The failure of subsequent collectors to encounter the plant on
Trinidad suggests that Sieber’s locality may be erroneous, and in any event
the record needs confirmation.

In view of its wide distribution, it is not surprising that P. mzmosoides
is a rather variable species; the populations on each island tend to be some-
what different from those on neighboring islands. The plants of Dominica
are perhaps the most divergent, differing from those of Guadeloupe and
Montserrat in their somewhat larger leaves, flowers, and seeds. Some char-
acters, such as the size of male flowers and degree of stylar union, appear
to be intrinsically variable in all populations. However, even those which
show some geographical correlation are not sufficiently well-marked to
justify the recognition of subspecific taxa.

Although P. acacioides is vegetatively quite similar to P. mimosozdes, its
female flowers (as noted in the key) are so different that there can be
no doubt the two species are perfectly distinct. There is also an obvious
resemblance between P. mimosoides and P. megapodus, but the latter is
not only vegetatively different but also has female flowers which are even
more dissimilar than are those of P. acacioides. It is possible that the
larger size of vegetative and floral parts in the Dominican population of
P. mimosoides may be partly attributable to hybridization with P. mega-
podus, although there is no direct evidence for this, but in any event this
would not prove that P. megapodus is the most closely related species.
The sum of the evidence would appear to suggest that P. mimosoides is
closest to P, acacioides in morphological characters, but is distinctly iso-
lated even from that species.

77. Phyllanthus acacioides Urb. Symb. Ant. 3: 287-288. 1902.
(PLATE XXXI, figs. K-L.)
Slender shrub with habit of P. mimosoides, up to 4 m. high; stem densely
reddish-tomentulose near apex, glabrate below. Cataphylls indurate: stip-

178 JOURNAL OF THE ARNOLD ARBORETUM _ [vor. xxxrx

ules triangular, appressed, acuminate, c. 2 mm. long; blade lanceolate,
convex, nearly 2 mm. long. Deciduous branchlets bipinnatiform; primary
axis c. 30-50 cm. long, 1.5—2 mm. thick, reddish- or purplish-tomentulose,
becoming more or less glabrate, scabridulous, terete, with c. 40-70 nodes;
first internode 15-25 mm. long, median internodes 5-9 mm. long. Leaves
of primary axis reduced to cataphylls: stipules persistent, not reflexed,
more or less indurate, triangular or triangular-lanceolate, c. 1.2—1.8 mm.
long, 1-1.7 mm. broad, acuminate, stramineous or reddish, entire and
glabrous, with dark fleshy basal area as in P. mimosoides; blade about as
long as stipules, lanceolate, long-acuminate, strongly convex on back. Ulti-
mate axes ascending, 5-11 cm. long, 0.5—0.6 mm. broad, purplish-brown
to olivaceous, more or less flattened and bluntly angled with ridges de-
current from the stipules, glabrous (occasionally with a few deciduous
tufts of hair proximally), usually scabridulous, with 35-60 leaves; first
internode 1.5—2.5 mm. long, median internodes 1.5—-3 mm. long. Leaves:
stipules persistent, not reflexed, scarious, lanceolate, 1-1.3 mm. long,
0.2-0.3 mm. broad, acuminate, marginally ciliate when young (otherwise
glabrous), adaxially decurrent at the base. Petioles 0.3-0.5 mm. long.
Leaf-blades subchartaceous, asymmetrically oblong- or obovate-falcate,
c. 6-11 mm. long, 2-4 mm. broad, obtuse or abruptly subacute at the tip
(apiculum more or less deciduous), at base oblique on adaxial side and
straight on abaxial side; above olivaceous, smooth, veins obscure; be-
neath pallid (albescent), finely and densely scabridulous, midrib raised
and running to tip, lateral veins obscure; margins somewhat thickened,
plane or reflexed.

Monoecious; cymules axillary, on ultimate axes of branchlet; cymules
mostly bisexual, of 1 female and 1 or 2 male flowers, but perhaps some-
times unisexual; leaves subtending flowers at branchlet tip sometimes
reduced to scales.

Male flower: pedicel capillary, c. 2.5-5 mm. long. Calyx-lobes 5, char-
taceous, elliptic or obovate, c. 1-1.3 mm. long, 0.7-1 mm. broad, obtuse
or rounded at the tip, entire, midrib unbranched. Disk-segments 5,
roundish to cuneate, somewhat fleshy, smooth, entire, c. 0.25-0.4 mm.
across. Stamens usually 5 (rarely 4 or 6); filaments completely connate
into a stout column 0.3-0.5 mm. high; anthers subsessile, one or two
slightly displaced in rare androecia of 6 stamens, anthers in two super-
posed whorls, discrete, ovate, obtuse, c. 0.2 mm _ long, 0.2-0.3 mm. broad;
anther-sacs aca flattened, dehiscing horizontally, slits apically con-
tiguous but discrete; pollen grains c. 13-14 » in diameter, with c. 5 or 6
polybrochate areoles per amb.

Female flower: pedicel c. 0.7—1.3 mm. long at anthesis, broadly dilated
above. Calyx-lobes 5, thick and fleshy, at anthesis stiffly erect and im-
bricate-connivent around the ovary, broadly elliptic or ovate to sub-
orbicular, c. 0.5—0.8 mm. long and broad, rounded at the tip, entire, mid-
rib unbranched, Disk nearly obsolete (reduced to a tenuous ring with
minute projections at angles). Ovary smooth, subspheroidal; styles united
at the base into a very short and stout column, bifid c. 44 their length, in

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 179

bud erect and slightly protruding above the calyx, at maturity greatly
elongating; stylar branches flattened, becoming 2-2.5 mm. long. tapering
from base to the acute tips, sharply reflexed on outside of calyx.

Capsule not seen entire; valves c. 3 mm. long, reddish-brown. Seeds
trigonous, c. 1.5 mm. long, light brown.

Type: Tobago, Eggers 5840.
DISTRIBUTION: endemic to Tobago (Map XXXIII).

TOBAGO: near Lot 42, shaded cool places, 21 Apr. 1913, Broadway 4189
(F, G, GH, MO, W); in sylvis Montis Morne d’or, alt. 500 m., Eggers 5840
(K, LEcToTYPE); St. George-Castara Road, 27 May 1930, Marshall (TRIN
12384),

Broadway described this plant as a gregarious single-stemmed shrub
with the appearance of Jacaranda caerulea and growing in shaded cool
places. The female flowers are remarkable for the great change in stylar
configuration associated with anthesis; in the bud the styles are erect
and protrude slightly above the calyx, but at anthesis they elongate greatly
and recurve abruptly so that they are appressed to the outside of the
rather fleshy calyx.

Although not all the diagnostic characters cited by Urban are depend-
able, P. acacioides clearly differs from P. mimosoides not only in its dis-
tinctive styles but also in its reduced female disk, pentamerous androecium,
and whitened scabridulous undersurface of leaf-blade. Because of its dis-
tribution, P. acacioides might be taken for an outlying species derived
from a disjunct population of P. mimosoides. However, the fact that the
female flower of P. acacioides shows a greater resemblance to that of P.
ovatus than it does to that of its “sister” species shows that great caution
must be observed in proposing phylogenetic speculations. Thus although
P. mimosoides is overall the species most similar to P. acacioides, it should
not be regarded as necessarily its direct ancestor.

Sect. 24. Xylophylla (L.) Baill. Etud. Gen. Euphorb. 623. 1858.

Xylophylla L. Mant. 2: 147-148. 1771.

Genesiphylla L’Heérit. Sert. Angl. 29. 1778.

Hexadena Raf. Sylva Tellur. 92. 1838.

Lomanthes Raf. ibid.

Phyllanthus b. Xylophylla Endl. Gen. Pl. 1120. 1840 (without indication of
k)

Phyllanthus sect. Typophyllanthus subsect. Genesiphylla (L’Hérit.) O. Ktze.
Lex. Gen. Phaner. 434. 1904

Shrubs or small trees with phyllanthoid branching; axes smooth and
glabrous; cataphylls of main stem indurate, clustered at the apex in a
scaly cone; branchlets bipinnatiform with ultimate axes transformed into
usually leafless phylloclades, or (in P. epiphyllanthus) entire branchlet
converted into phylloclade; nodes represented by marginal notches, leaves
usually all reduced to scales, occasionally well-developed. Monoecious;

180 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

cymules unisexual or bisexual, produced at notches of phylloclades. Male
flower: calyx-lobes 6 (rarely 5); disk-segments usually 6 and discrete;
stamens normally 3 or rarely 4 (very rarely 5), filaments united at least
at the base; anthers dehiscing more or less horizontally; pollen grains
globose, areolate. Female flower: calyx-lobes 6 (rarely 5); disk of discrete
segments or cupuliform to urceolate; styles free or connate below, style
branches often again lobed or bifid. Capsule oblate, smooth to tuberculate;
seeds trigonous or sometimes (when only 1 per locule) ovate and flattened,
verruculose.

Type species: PAyllanthus epiphyllanthus L. (Xylophylla latifolia L.,
ex p.).

This West Indian section of about 10 specific and subspecific taxa
includes some of the most familiar representatives of the genus; the type
species, P. epiphyllanthus, was the first American member of the genus to
become well-known in Europe. Because of their unusual morphological
features, plants of sect. Xylophylla have received considerable attention
from both horticulturists and morphologists, but except in the little-quoted
work of Dingler (Flachsprosse der Phanerogamen. 1885) few attempts
have been made to understand the evolution and relationships of the
phylloclade-bearing species of Phyllanthus.

As here construed, sect. X ylophylla has a circumscription narrower than
that of Mueller (DC. Prodr. 15[2]: 427-432. 1866) or Pax and Hoffman
(Naturl. Pflanzenfam. 19c: 64-65. 1931). These authors included, in
addition to the West Indian species, four or five phyllocladiferous Brazilian
species such as P, klotzschianus and P. flagelliformis. Convergent evolu-
tion has progressed so far that on superficial inspection certain forms
of P. klotzschianus and P. montanus might almost appear to be conspecific.
However, an attentive examination of the flowers belies this impression, for
the flowers of the Brazilian species differ in many details, including their
deeply emarginate anthers, and their tricolporate coarsely reticulate pollen
grains are completely different from the areolate grains of the West Indian
species. It therefore appears certain, as suggested earlier in this study
(Jour. Arnold Arb, 37: 111. 1956), that the Brazilian species have evolved
phylloclades quite independently of the West Indian ones; they are to
be transferred to sect. Choretropsis.

As emended, sect. Xylophylla comprises only West Indian species, all
except two of which are confined to Jamaica. At the present time sect.
Xylophylla gives the appearance of a successful group in which evolution
is probably actively in progress. Several of the species, particularly P.
epiphyllanthus and P. angustifolius, occur in large and conspicuous popula-
tions, which show signs of incipient speciation. The representatives of the
section afford a most interesting epitome of the evolutionary process as it
occurs in many groups of higher plants on the Caribbean islands, and merit
more intensive study than has been accomplished here.

The relationships of sect. Xylophylla are plainly closest to sects. Epis-
tylium and Hemiphyllanthus, the former appearing to be the group from

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 181

which both sect. Xylophylla and sect. Hemiphyllanthus were derived. An
hypothesis of the probable origin of bipinnatiform branchlets has already
been mentioned under the latter group. The peculiar branchlet structure
of sect. Xyvophylla is thus a specialization involving the replacement of
the original leaves of a bipinnatiform branchlet by the expanded axes them-
selves. However, the relationship of P. montanus, the most primitive
species of sect. Xylophyilla, to P. cauliflorus and P. axillaris of sect. Epis-
tylium is so striking that it seems possible that sect. Xylophylla may have
arisen from sect. Epistylium coordinate with, rather than derived from,
the less specialized sect. Hemiphyllanthus.

e P. ANGUSTIFOLIUS
P EPIPHYLLANTHUS

© dilatatus

O domingensis

@ epiphyllanthus

Cage er Ce

Map XXXIV. Distribution of sect. Xylophylla; heavy black line indicates
limit of range of species other than P. angustifolius and P. epiphyllanthus.

In order to make clear the homology between the branchlets in the
present section and those in sect. Hemiphyllanthus, the terminology used
for vegetative structures in the species descriptions has been slightly modi-
fied from that presented in the introductory section of this study. The com-
pound phylloclade, as illustrated in text-fig. 5, is here described as a
branchlet with the ultimate axes transformed to phylloclades, the branch-
lets of all species of sect. Xyvlophylla being regarded as modifications of
bipinnatiform branchlets as in sect. Hemiphyllanthus. In P. montanus,
which is the most primitive species of sect. X ylophylla, the primary branch-
let axis is unmodified and the term phylloclade is applicable only to the
lateral ultimate axes. In P. angustifolius the primary axis is flattened and
as green as the ultimate axes, so that the branchlet of this species can be
regarded as a truly compound phylloclade. Finally, in the most special-
ized species, P. epiphyllanthus, the lateral axes are suppressed and the
dilated primary axis of the branchlet serves directly as a simple phylloclade.

182 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. Xxxix

Although it cannot be used as a key character, due to its relative in-

frequency of occurrence, the production of leaves on the usually leafless
phylloclades of sect. Xylophylla is a phenomenon of some taxonomic
interest; the tendency toward suppression of leaves parallels, as might be
anticipated, the trend toward an increasingly leaf-like character of the
branchlet axes. In P. montanus leaves commonly occur on seedling phyl-
loclades and sprout-shoots and, in a diminutive form, may sometimes be
produced on normal ance. In P. latifolius and P. angustifolius leaves
normally occur on seedlings and also appear on wound-shoots, and this 1s
probably true of P. proctoris as well. In P. arbuscula and P. epiphyllanthus,
on the other hand, leaves have never been observed at any stage in
ontogeny.
The identification and taxonomic analysis of the Jamaican species of
sect. Xylophylla is beset with a number of practical difficulties due to the
high percentage of sterile specimens; flowering in some species appears to
be erratic and infrequent. The key has been constructed with this in mind
and should be workable even for sterile specimens as long as they are not
sprout-shoots or unusual modifications.

KEY TO THE SPECIES

1. Phylloclades borne distichously on main axis of deciduous branchlet

2. Branchlet slow-growing, rather long-persistent, main axis greyish or
brownish, not colored as the phylloclades; pemer! : female flower
OS—15 Mitty LONG. 65600 e sea cacanendex es . P. montanus
Branchlets expanding rapidly, not long- meeeictent: main axis greenish
and of consistency of lateral axes (ohyllociades),
3. Pedicel of female flower only 0.5-1.5 mm. long; disk of female flower
dissected into lobes or segments; phylloclades more or less rhombic-
lanceolate, with mostly 20-50 nodes; apical cone 10-15 mm. broad
cataphylls scarcely or not ciliate. .......... 79. P. latifolius
Pedicel of female flower mostly 2 mm. long or more; disk of female
flower entire; phylloclades elliptic- - obovate- lanceolate ;
4. Cataphylls of main axis remaining brown or grey, ciliate (if at

all) only toward the base; phylloclades mostly 1-2 cm. broad with

20-40 nodes; ovary smooth. ................ 80. P. arbuscula
Cataphylls of main axis becoming dark or blackish-brown, copiously
ciliate on margins at least when youn hylloclades mostly 0.5-1
cm. broad with 10-25 nodes, often ean at end of branch-

bo

Ge

=

let axis.

5. Phylloclades aa he lanceolate, mostly well over 3
mm. broad, with (8-) 10-20 (—24) nodes; styles not dilated,
the ends 3—5-fid into fea tips. 81. P. eae

5. Phylloclades (floriferous gee sacral near only 1-3 mm.

broad, with 7-10 (-13) nodes; styles dilated and ene
ends ‘merely crenulate, aa a sort of pre over the
OVAL. Giidiesargan $A eed alee Oho e bee . proctoris
1 Phylloclades each representing an entire branchlet, borne ee directly
OU Ia SteNls 2.2 nn lecan wena ee na eer 9 eh ee 83. P. epiphyllanthus

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 183

78. Phyllanthus montanus (Sw.) Sw. FI. Ind. Occ. 1117. 1800; Muell.
Arg. in DC, Prodr. 15(2): 429. 1866; Fawc. & Rend. Fl. Jam. 4:
261-262. 1920. (PLATE XXXI, figs. M-N).

Xylophylla montana Sw. Prodr. 28. 1788.
Diasperus montanus (Sw.) O. Ktze. Rev. Gen. 2: 600. 1891.

Shrub or small tree, usually c. 2-5 m. high; branches 2-8 mm, thick,
terete, smooth, greyish-brown. Apical cone inconspicuous, c. 2 mm. long
and 2—2.5 mm. broad; cataphylls deciduous, pale, thickened, deltoid, not
over c. 1.5 mm. long, blunt, entire; blade about as long, massive. Branch-
lets ascending, bipinnatiform, ultimate axes transformed into phylloclades;
primary (penultimate) axis 5-30 (—40) cm. long, 1.5-3 mm. broad, prox-
imally terete, distally often flattened, greyish or brownish, with 7—20
(-30) lateral axes; first internode (0.5—-) 1.5-4 cm. long, median inter-
nodes mostly 0.5—2 cm. long. Cataphylls of primary axis subpersistent
or deciduous; stipules ovate, thickened, mostly 1-1.5 mm. long, blunt,
entire or sparsely ciliate; blade shorter or about as long. Phylloclades
flexible or rather stiff, elliptic to lanceolate (rarely linear-lanceolate), ob-
tuse to long-attenuate at the tip, 3-20 cm. long, 0.5-3 cm. broad, with
10-30 (—50) slightly notched nodes; midrib conspicuous, more or less
plane or salient beneath, veins tenuous and not very prominent; margins
slightly (if at all) differentiated. Euphylls rarely produced except on
sprout-shoots, blade elliptic, up to 11 mm. long and 6 mm. broad, obtuse
or subacute at the tip, acute at the base, distinctly paler beneath, midrib
prominent but veins and veinlets rather obscure. Cataphylls of ultimate
axes reddish brown, scarious with often narrow, paler margins; stipules
suborbicular-ovate, auriculate, 0.5-1.3 mm. long, 0.4-0.8 mm. broad,
subentire or denticulate; blade subulate, less than 1 mm. long.

Monoecious; cymules usually bisexual, pulviniform, each with a single
(rarely 2) female flower and c. 3-12 male flowers.

Male flower: pedicel slightly thickened, stiff, 1-3 mm. long. Calyx
usually reddish-tinged; calyx lobes 5, chartaceous or hard-scarious, erect
to somewhat spreading, unequal, one or two outer lobes usually elliptic
to ovate, less than 1 mm. long, inner lobes suborbicular, c. 1.5 mm. long
and broad, blunt-tipped, midrib simple or very sparsely branched above.
Disk-segments 5, rather massive, roundish concave, c. 0.4—0.5 mm. across.
Stamens 3; filaments completely connate into a column c. 0.5 mm. high;
anthers sessile atop the column, broadly deltoid, blunt or emarginate,
c. 0.2-0.25 mm. long and 0.3-0.4 mm. broad; anther-sacs dehiscing hori-
zontally; pollen grains 17-21 » in diameter.

Female flower: pedicel stout, c. 0.5-1.5 mm. long. Calyx greenish or
(ex Swartz) dark purplish; calyx-lobes 5 or 6, at anthesis somewhat fleshy
and erect, becoming scarious and spreading in fruit, subequal (outer lobes
narrower), c. 1-1.5 mm. long, 0.8-1.5 mm. broad, more or less denticulate,
midrib unbranched. Disk divided into distinct thin, petaloid segments
c, 0.2-0.3 mm. long. Ovary ellipsoid or oblate, smooth, shallowly sulcate;
styles free, spreading or ascending, sometimes slightly recessed into the

184 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxiIx

top of the ovary, c. 0.3-0.5 mm. long, bifid, the tips blunt, sometimes
dilated.

Capsule oblate, rounded-trigonous, c. 4 mm. in diameter, reddish brown,
sometimes somewhat glaucous, rugulose, not veiny. Columella c. 1.8 mm.
long. Seeds asymmetrically trigonous (plano-umbonate), 1.7-1.9 mm.

MAP XXXV

@ P MONTANUS

0-20 ms

MAP XXXVI "
e

@ P. LATIFOLIUS

——— 5 20) ‘ins

MAP XXXVI O
JPEN

OF PROCTOR :
@ P. ANGUSTIFOLIUS .

=———— 2) ms

ApS XXXV-XXXVII. Distribution of some Jamaican species of sect.
Pe a

long, 1.2-1.4 mm. broad, reddish, verruculose with rather densely spaced
slightly raised dots; micropylar end sometimes carunculate.
Collected in flower Dec., Feb—May; in fruit Feb., July, Oct.

Type: western Jamaica, Swartz (S, HOLOTYPE),

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 185

DISTRIBUTION: wooded hills, western and central Jamaica (Map
XXXV).

JAMAICA: Beer specific locality, Purdie (A), Swartz (S, HOLoTYPE, G, K,
ISOTYPES). HANOVER: Dolphin Head, Britton 2211 (NY), Harris 9251 (JAM,
NY); hills near ere Britton 2422 (NY); Shepherds Hall, 1 mi. east of
Great Valley, Proctor 7255 (JAM). TRELAWNY: near Troy, Hae 8736 (NY);
Sherwood Content, hilltop, Proctor 11064 (GH); Ramgoat Cave area, Howard &
Proctor 14412 (A). MANCHESTER: Somerset, northwest of Mandeville, limestone
cliff, Proctor 11594, 11595 (GH). St. ANN: Hollymount, Mt. Diablo, Harris
8986 (A, JAM, NY, S), Maxon 1906 (US); Pedro district, Purdie (K);
Ramble Estate, 2.5 mi. SW of Claremont, Webster & Proctor 5637 (A). CLAREN-
poN: Peckham Woods, Harris 10991, 11013 (NY), 12797 (JAM, NY), Proctor
8430 (GH), Stearn 15 (A), Webster & Proctor 5431 (A). St. CATHERINE:
Lluidas Vale, Hunnewell 19766 (GH). St. ANDREW: cascade of Falls River,
Kingston, Prior (K).

Phyllanthus montanus is a species with most interesting features, since
it clearly is the most primitive living representative of sect. Xylophylla
and furthermore is obviously related to the species of sect. Epistylium.
The primary axis of the branchlet of P. montanus is so similar to the
branchlet of P. axillaris, even to the stipules, that a reasonably close
affinity between the two species seems undeniable. One difficulty in assess-
ing the relationships of P. montanus is that it is an extraordinarily variable
species in its vegetative features. Specimens from Peckham Woods, which
have short, often twisted branchlets with short lateral axes, scarcely sug-
gest any relationship to a species of sect. Epistylium, but two unusual
collections made near Somerset by Proctor have long primary axes with an
unmistakable resemblance to those of P. axillaris. Despite these differences,
intraspecific variability of P. montanus does not appear to show any definite
geographic correlation, and there is no necessity to describe subspecific
taxa.

Although it is the most isolated species in the section, P. montanus
shows certain points of resemblance to P. latifolius — specifically, the short
pedicels and dissected female disk — which suggest that the latter is its
closest relation within the section. However, P. latifolius is vegetatively
much more like the other species of the section in having a definitely
modified primary branchlet axis, and its relation is certainly not very
close.

79. Phyllanthus latifolius Sw. Fl. Ind. Occ. 1109. 1800.
(PLATE XXX], figs. O-P; PLATE XXXII).
Phyllanthus 1. Folis latioribus, &c. Browne, Hist. Jam. 188. 1756.
Xylophylla latifolia L. Mant. 2: 221. 1771 (ex p.; excl. typ.).
Lomanthes latifolia Raf. Sylva Tellur. 92. 1838.
Diasperus latifolius (Sw.) O. Ktze. Rev. Gen. 2: 599. 1891.
Phyllanthus isolepis Urb. Symb. Ant. 3: 290. 1902

Shrub or small tree c. 1-4 m. high; branches of current year’s growth
3-5 mm. thick, terete, brownish or greyish. Apical cone conspicuous, c.

186 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

10-15 mm. long and broad, irregular in outline. Cataphylls of main axis
deciduous, scarious-indurate; stipules lanceolate, (4—-) 6-10 mm. long,
acuminate, squarrose; blade 3.5-11 mm. long. Branchlets ascending, bi-
pinnatiform, ultimate axes transformed into phylloclades; primary axis
(S—) 8-20 cm. long, (1—) 1.5-2 mm. broad, flattened, edges obtuse, oliva-
ceous, smooth, with (4-) 7-12 (-15) lateral axes (phylloclades) ; first
internode (1—) 2—4 (—8) cm. long, median internodes mostly 1-2 cm. long.
Cataphylls of primary axis deciduous; stipules and blades linear-lanceolate,
attenuate-acuminate, c. 1-2 mm. long, reddish brown, not ciliate. Phyllo-
clades usually rigid, most often rhombic- or obovate-lanceolate but some-
times (especially on sprout-shoots) narrowly lanceolate, (3.5—) 5-8 (—10)
cm. long, (0.7—) 1.5—3 cm. broad, narrowed (often abruptly so) to a blunt
or sometimes caudate tip, with (15—) 20-50 (-—60) nodes; midrib more
or less prominent, veins steeply ascending, tenuous or subprominent; mar-
gins not differentiated. Euphylls occasional on sprout-shoots: blade obo-
vate, c. 3-5 mm. long, acute at the tip and base, midrib rather prominent
beneath but veins otherwise obscure. Cataphylls of ultimate axes subor-
bicular, convex, trifid, not over c. 0.5 mm. long, reddish brown, ciliate.

Monoecious; cymules male or (more often) bisexual, with 1 female and
several (up to c. 10) male flowers.

Male flower: pedicel c. 1-3.5 mm. long. Calyx reddish; calyx-lobes 6,
scarious-membranous, spreading (at least at the tips), subequal, obovate,
c. 0.7-1.3 mm. long and 0.4-0.9 mm. broad (occasional outermost lobes
greatly reduced), entire, midrib unbranched. Disk-segments 6, not pitted,
c. 0.2-0.3 mm, broad. Stamens 3; filaments (0.5—) 0.7-0.9 mm. long,
united below into a stout column c. 0.3—-0.7 mm. high; anthers broadly
ovate, emarginate, c. 0.2—-0.3 mm. long, 0.4—0.5 mm. broad; anther sacs
dehiscing more or less horizontally, pollen grains 15—19 p» in diameter.

Female flower: pedicel 0.5-1.5 mm. long. Calyx reddish; calyx-lobes 6,
scarious or chartaceous, spreading, obovate or suborbicular, 0.5-0.8 mm.
long, 0.4—0.7 mm. broad, entire, midrib unbranched. Disk separated into
6 lobes or segments c. 0.2—0.4 mm. across. Ovary oblate or turbinate, sul-
cate, smooth; styles basally connate into a short column 0.2—0.4 mm.
high; style-tips spreading, c. 0.6-0.8 mm long, with mostly 3-5 lobes
(occasional style-tips merely bifid).

Capsule oblate, c. 4 mm. in diameter, reddish brown, rugulose, not veiny.
Columella c. 1 mm. long. Seeds asymmetrically trigonous, c. 2 mm. long,
1—1.2 mm. broad, reddish brown, verruculose with slightly raised dots.

Collected in flower June, July, Sept.; in fruit July.

Type: Jamaica, Browne (Herb. Linn. 1105-1 LINN; HoLotyPe).

DIstRIBUTION: dry rocky areas, southern Jamaica (Map XXXVI).

JAMAICA: without specific locality, Browne (LINN), Jacquin (BM), March
(GH), Masson, Shakespeare (BM). St. CATHERINE: near Salt Island, Health-
shire Hills, Britton 3058 (NY), Harris & Britton 10532 (JAM, US); 2 mi.
NNW of Guanaboa Vale P.O., 900 ft., Proctor 7151 (MICH); Horse Cave, Hell-
shire Hills, Proctor 7602 (GH); near Bartons, c. 5 mi. north of Old Harbor, alt.

Jour. ARNOLD Ars. VoL. XXXIX

PLATE XXXII

3
Ss)
“4
=
an
fos}
Lar
ca
wn
“a
se
Cu
o .
AH
Bros
o
ae.
a
UO Ss
55
KH 0
Ton}
Lig
aq
os
ma
48
FO et
ror ag
oO
MD en
Go
rn
ao
apm
he
a ae
a
ow
a 3D
Oo 0
Bea
av
om
plan a 4 |
ean
om
4
tot 6)
jo)
ciclo
SN
ae
aD O96
oon
ined
- >
we
=
& i=

‘ebster & W

if

abit of Phyllanthus latifoliu
(J

H

188 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

c. 1000 ft., Webster & Wilson 4862 (A). St. ANDREW: Ferry River, Britton 2828
(NY); Ferry Pen, Campbell 6280 (NY; type collection of P. isolepis Urb.) ;
Fresh River north of Ferry, Proctor 8276 (GH), Webster & Wilson 5129 (A);
Red Hills, Britton 3469 ve Grant & Barkley 22J080 (GH, MICH); Long
Mountain, Barry (JAM, MICH), Harris 8843 (NY), 8845 (NY), Maxon 10531
(NY, US), Webster & fs 4875 (A, MICH); Mona, Barry (JAM); Cane
River Valley, alt. 250-400 ft., Harris 9631 (US, NY), 10065 (JAM, NY, US).

Phyllanthus latifolius has a relatively restricted distribution, being
known thus far only from dry hills in St. Catherine and St. Andrew par-
ishes, although it may eventually be found to overlap into Clarendon and
St. Thomas. It is perhaps the most xerophytic species in the section, al-
though some forms of P. angustifolius and P. epiphyllanthus also occur
in similarly dry habitats. The detailed distribution of P. latifolius offers
some interesting problems with regard to the interaction between species.
On Long Mountain, it was found to occupy the dry, lower southwestern
slopes, while P. angustifolius occupied the northeastern slopes of the rela-
tively mesophytic woods on top. This suggests an ecological separation
between the species, but in some areas (e.g., Portland Ridge) P. angusti-
folius grows in habitats which are the sort occupied by P. datzfolius. It is
striking that the two species have never been observed growing side by
side, even where (as at Long Mountain) their populations are in close
proximity; this mutual exclusiveness, which suggests the effects of com-
petition, deserves to be investigated critically from the ecological point
of view.

As has been mentioned above, P. latifolius approaches closer to P. mon-
tanus than any other species of the section. The characters which suggest
such a relationship — the dissected female disk and short pedicel — are
also the ones which serve to separate P. latifolius from its close relative,
P. arbuscula. These two species have the same kind of cataphylls and
rather similar phylloclades, but those of P. latifolius are often thicker,
more rhombic, and with denser clusters of flowers. The plants of P.
angustifolius which occur within the range of P. latifolius may usually be
easily distinguished from “narrow-leaved” forms of the latter by their black-
ish, ciliate cataphylls and phylloclades with fewer nodes which are often
paired at the branchlet-tip.

80. Phyllanthus arbuscula (Sw.) Gmel. Syst. 2: 204. 1791.
(PLATE XXXII, figs. Q-S).

Xylophylla arbuscula Sw. Prodr. 28. 1788.

Xylophylla angustifolia b. linearis Sw. ibid.

Phyllanthus speciosa Jacq. Collect. 2: 360. 1788; Ic. . a pl. 616. 1792.

Phyllanthus linearis (Sw.) Sw. Fl. Ind. Occ. 1113.

Xylophylla speciosa (Jacq.) Sweet, Hort. Brit. ed. . ee 18277,

Genesiphyla speciosa (Jacq.) Raf. Sylva Tellur. 92. 1838.

Phyllanthus linearis q genuinus Muell. Arg. in DC. Prodr,. 15(2): 430. 1866
(excl. descr.).

Diasperus speciosus (Jacq.) O. Ktze. Rev. Gen. 2: 601. 1891.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 189

Phyllanthus inaequaliflorus Fawe. & Rend. Jour. Bot. 57: 66. 1919.
c. & i

Phyllanthus dingleri Webster, Jour. Arnold rae 37-4 O56:

Shrub or small tree up to 7 m. high, usually with a single main trunk
and a few erect branches, these c. 3-5 mm. thick, smooth, light brown or
greyish. Apical cone as broad as long, c. 5-10 mm. across, outline irregular
due to exsertion of phyllocladar cataphylls; cataphylls deciduous, pale,
thickened: stipules deltoid to lanceolate, 2.5—7 (—9) mm. long, 1-3 mm.
broad, obtuse to acuminate, entire (rarely ciliate toward the base); blade
deltoid to lanceolate, acuminate, about as long as the stipules. Branchlets
ascending, bipinnatiform, ultimate axes transformed into phylloclades; pri-
mary axis (2.5—) 5-25 (—30) cm. long, 1-2.5 mm. broad, compressed and
obtusely angled, adaxially sulcate, greyish green, with 4-16 lateral axes
(phylloclades); first internode 1.5—5 (—6) cm. long, median internodes
0.5-2.5 mm. long. Cataphylls of primary axis mostly deciduous (occa-
sional ones more or less persistent): stipules linear-lanceolate, (1.5—) 3-6
(-8) mm. long, acuminate, entire (or rarely sparsely ciliate toward the
base), pale brown, basally auriculate; blade linear-lanceolate, about as
long as the stipules. Phylloclades thin and flexible to somewhat rigid,
elliptic to lanceolate, (2.5—) 4-11 cm. long, (0.5—) 1-2 (—2.7) cm. broad,
tapering (sometimes abruptly) to an acute or acuminate tip, often paler
(yellowish) beneath, with (10-) 20-40 (—50) conspicuously notched
nodes; midrib and veins very often conspicuous and raised on both sides;
margins usually with a distinctly differentiated rim running between
notches. Euphylls never observed, even on seedlings. Cataphylls of phyl-
loclades reddish brown, fragile and readily breaking off at the base, trifid
(stipules united to the blade), tips acuminate, c. 0.5—1 mm. long; base
auriculate, auricles ciliate-dentate.

Monoecious; cymules bisexual, of 1 female and c. 3 or 4 males, or
proximal cymules male.

Male flower: pedicel capillary, (1.5—) 2-5 (-8) mm. long. Calyx cream-
colored or greenish to scarlet; calyx-lobes 6 (rarely 5), membranous to
coriaceous, more or less unequal, outer lobes obovate to elliptic and inner
lobes ovate, or all lobes ovate, 1-2 mm. long, 0.8—1.8 mm. broad. Disk-
segments 6, roundish, entire, thin, flat, c. 0.3-0.6 mm. across. Stamens 3;
filaments 0.3-0.6 (—0.8) mm. high, united into a massive column 0.2—
0.5 mm. high; anthers emarginate, 0.25—0.4 mm. long, 0.4—0.6 mm. broad;
anther-sacs dehiscing horizontally; pollen grains c. 18-21 » in diameter.

Female flower: pedicel thicker than in male, stiff, (2.5—) 3.5-10 (—14)
mm. long at anthesis, increasing to 6-15 mm. long in fruit. Calyx-lobes
cream-colored or greenish,* membranous to coriaceous, erect to spreading,
c. 1-1.5 (—3) mm. long and broad (reduced outer lobes often smaller),
obtuse (outer lobes sometimes with a brownish scarious tip), entire, mid-
rib unbranched. Disk annular or shallowly cupuliform to urceolate and

* or scarlet in western populations.

190 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

enclosing the ovary, up to c. 1 mm. high. Ovary oblate, smooth; style
free and spreading or basally united and then more or less erect, stylar
column up to 1 mm. long; style-ends often twice bifid, tips with 3 or 4
lobes.

Capsule oblate, trigonous, c. 4-5 mm. in diameter, dark reddish brown,
rugulose, not veiny. Columella 2—2.2 mm. am Seeds asymmetrically
trigonous (plano-umbonate), 2.7—3.3 . long, 2-2.5 mm. radially and
tangentially, with irregular era ee of eave raised reddish-
brown dots

Collected 4 in flower Feb., Apr., June—Nov.; in fruit Apr.—July.

Type: Jamaica, Catherine Hill, Swartz (BM, HOLOTYPE; S, ISOTYPES).
DISTRIBUTION: upland areas, Jamaica (Map XXXVIII).

This species consists of some distinctive populations but no clear-cut
subspecies. The specimens are cited according to these races, which are
lettered but not named; the proposed species to which each race roughly
corresponds is given in parentheses.

ace A. Calyx scarlet, fleshy or coriaceous, lobes erect; disk of female
flower urceolate, enclosing the ovary; styles united into a column as high
as or higher than the ovary; (P. dingleri).

JAMAICA: sea-coast, Swarts (BM, S; TYPE COLLECTION of P. swartzii).
HANOVER: Dolphin Head, woods on summit, Britton & Hollick 2850 (NY), T.
Farr (GH), Proctor 7276 (GH, MICH); Kempshot, alt. 550 m., Britton 2428
(NY). Sr. James: Lapland, alt. c. 2000 ft., Harris 9191 (A, JAM, US). Man-
CHESTER: Somerset, see of Mandeville, alt. 2300 ft., Proctor 11605 (GH).

Race B. Calyx scarlet, more or less coriaceous; disk of female flower
saa enclosing the ovary; styles united into a definite column; seeds
often 1 per locule; phylloclades often small and thick; (P. coxianus).

JAMAICA. Tretawny: Kimloss, Howard 14136 (A); Stonehenge, Howard
14156 (A); Ramgoat Cave, Howard & Proctor 14394, 14401, 14407 (A);
Troy, Britton 596 (NY), Harris 9368 (A, JAM, NY), 8565 (A, NY), Proctor
7998 (JAM). MANcuHEsTER: 1 mile west of Christiana, alt. 3000 ft., Howard &
Proctor 14336 (A). St. ANN: St. Anns, 1850, Prior | Alexander | (GOET, ie
VY; TYPE COLLECTION of P. coxianus); 2 mi. west of Albion, alt. 2500 ft.,
Howard et al. 14617 (A); Ramble, Claremont, alt. 1700 ft., Rauiccts & Hors
oe (BM, US); Union Hill, near Moneague, Britton & Hollick 2743 (NY,

CLARENDON: Peckham Woods, alt. 2500 ft., Proctor 8432 (GH), Webster &
ae 5404 (A, JAM, MICH).
Race C. Calyx whitish, or partly pinkish-tinged, thin-textured; disk
of female flower covering c. 1% ovary or less; styles slightly united or
nearly free; phylloclades nena and thin; (P. jumequmlsfoss’.

JAMAICA. TreLawny: near Troy, alt. 2000-2500 ft., Harris 8714 (NY,
S, US), 8771 (JAM, NY), Maxon soe (US), Webster et al. 5380 (A, MICH).
MANCHESTER: Marshall’s Pen; alt. 2300 ft., Stearn 387 (A, JAM); Mandeville
Brown 135 (NY, US) - Fairfield, W Sere 1012 (GOET, M, W;; data ex M).
St. ANN: Lydford (Golden Grove), Howard & Proctor 13982 (A): Mt. Diablo,

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 191

Tee 15306, 15307 (GH); Holly Mount, Mt. Diablo, Harris 8988 (BM,
AM, NY; TYPE COLLECTION of P. inaequoliforus), Webster & Wilson 5012
a JAM, MICH): Grier Mount, Webster & Proctor 5627 (A, JAM, MICH).

Race D. Calyx whitish or greenish, membranous; disk of female flower
covering less than 1% the ovary; styles free or very nearly so, spreading;
phylloclades usually broad and thin; (P. speciosus).

JAMAICA: without ees (probably a Andrew), March (GOET), Swartz
(S, probably isoTyPE), Wright (W). St. ANDREW: John Crow Peak, Bancroft
(J.P. 1263) (JAM); Silver an Hylton ae Catherine Hill, Swartz (BM,
HOLOTYPE). PoRrTLAND: Hayc ock Mt., above Balcarres, alt. 2750-3500 ieee
Proctor 8074 (JAM); Spring Bank, 2.5 mi. WSW of Port Antonio, Proctor
6722 (MICH). St. THomas: House Hill to Cuna Cuna Gap, alt. 550-725 m
Maxon 8966 (NY, US); Corn Puss Gap, Webster & Wilson 4893 (A, JAM,
MICH); Mansfield, near Bath, alt. 300-500 m., Maxon 2404 (NY, US); Big
Level, alt. 2200-2500 ft., Webster & Proctor 5517 (A, JAM, MICH).

The long synonymy of P. arbuscula is a good indication of the extraor-
dinary variability of this common Jamaican species. For a long time an at-
tempt was made by the writer to distinguish the species proposed by
Fawcett and Rendle, first on the specific level and then on the subspecific.
After prolonged consideration, however, any attempt to distinguish sub-

la@i RACE A:
@ RACE B.
® RACE C.

O RACE D. | 0. 20 ni.

Map XXXVIII. Distribution of P. arbuscula (Sw.) Gmel., showing variation
in floral characters according to races described in text. Small circles indicate
sterile specimens.

specific taxa within P. arbuscula has been abandoned, for although “sub-
species” could be recognized and grouped in a key, this would give a mis-
leading impression of permanence and solidity to what are more probably
clines, rather than geographically delimited subspecies. If one compares
a flowering specimen of the Dolphin Head race (A), with one from the
Blue Mountains race (D), it appears at first glance that two different
species are at hand; but if the specimens are taken instead from the area
in the center of the island, between Troy and Mt. Diablo, serious difficul-
ties immediately arise. All combinations of the characters mainly con-
cerned (flower color and texture, female disk size, and stylar configura-

192 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

tion) can be found within a relatively small area. Thus, in Manchester
parish, Proctor 11605 from Somerset represents “good” P. dingleri, with
red flowers, urceolate female disk, and united styles; but the Brown and
Wullschlaegel collections from Mandeville and Fairfield are, despite their
peculiar phylloclade shape, scarcely different from the typical form of the
species in the Blue Mountains; and Stearn 387, with whitish flowers but
united styles, is transitional to the Dolphin Head type. Similar examples
could be added if it seemed necessary to elaborate the point, that the char-
acters which appear so distinct at the two extremities of the island undergo
“dissolution” in the center. The situation might possibly be interpreted
as representing that of two extensively hybridizing subspecies, but the
area of intermediates would probably be so large that such a classification
would serve no useful purpose. The variation is so evidently clinal in
character that it would be extremely interesting to map it in detail but,
unfortunately, the large number of sterile specimens makes a thorough
analysis impracticable at this time.

A character of especial interest is the common occurrence in plants of
the Cockpit Country race (B) of capsules in which all or some of the
locules produce only a single seed. As is usual in the Phyllantheae, the
single seeds are so differently shaped that they look quite unlike the paired
ones, Unfortunately, it is not yet possible to estimate the taxonomic im-
portance of this feature because seeds are not known from the other three
races. Thus far it appears that the production of solitary seeds is confined
(in sect. Xylophylla) to P. arbuscula.

The name X wophylla arbuscula, the basis for that here accepted for this
species, was reduced by Swartz himself to the synonymy of P. speciosus
Jacq., a course subsequent writers have followed. It has been assumed
that Swartz withdrew his own name because of Jacquin’s priority in pub-
lication, but Mr. William Stearn (personal communication) has discovered
evidence that Swartz’s name was actually the earlier. Swartz may have
withdrawn his proposed name out of modesty or perhaps because Jacquin’s
was published under the correct genus; but in any event, the plant col-
lected by Swartz in the Blue Mountains must take the name of P. arbuscula.

Phyllanthus arbuscula is rather closely related to P. latifolius but, as
has been discussed under that species, it clearly differs both in floral
and vegetative characters. More closely related is the species from the

cause of lack of floral material. In contrast, P. angustifolius is much less
similar morphologically since it has dark, ciliate cataphylls, usually nar-
rower indistinctly veined phylloclades with fewer nodes, and a rugulose
ovary. However, at least one specimen (Webster & Proctor 5517) has
characteristics intermediate between P. arbuscula and P. angustifolius,
suggesting that it may represent a hybrid. The peculiar plant named
P. linearis by Swartz may possibly represent a cross between P. proctoris
and P. arbuscula; but narrow-leaved forms of the latter sometimes occur
normally (e.g. March, GOET), so that in this case the situation is not
entirely clear.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 193

81. Phyllanthus angustifolius (Sw.) Sw. Fl. Ind. Occ, 1111. 1800
(as P. angustifolia); Fawc. & Rend. Fl. Jam. 4: 262. 1920.
(PLATE XXXI, figs. T-U).

Phyllanthus 2. Foliis angustis longioribus, &c. Browne, Hist. Jam. 188. 1756.

Xylophylla angustifolia Sw. Prodr. 28. 1788.

Phyllanthus cognatus Spr. Syst. 3: 23. 1826.

Hexadena angustifolia (Sw.) Raf. Sylva Tellur. 92. 1838.

Phyllanthus angustifolius a genuinus Muell. Arg. in DC. Prodr. 15(2): 430-
431. 1866.

Phyllanthus linearis 8 cognatus (Spr.) Muell. Arg. op. cit. 430.
Diasperus angustifolius (Sw.),O. Ktze. Rev. Gen. 2: 598. 1891.
Xylophylla contorta Britton, Bull. Torr. Bot. Club 37: 353-354. 1910.

Shrub up to c. 3 m. high; branches 2.5-5 mm. thick, reddish brown or
greyish. Apical cone usually rounded in outline (cataphylls with tips
parallel or inflexed) and c. 3-4 mm. in diameter, but sometimes up to
5 or 6 mm. in diameter and irregular, the cataphyll tips somewhat diver-
gent (but not squarrose); cataphylls deciduous, light brown to blackish
brown, scarious-indurate: stipules usually broadly ovate, c. 15-2 mm.
long and about as broad, less commonly lanceolate and up-to 5 mm. long,
margins always copiously ciliate when young (becoming more or less
glabrate in age); blade ovate to lanceolate, about as long. Branchlets
bipinnatiform, ultimate axes transformed into phylloclades; primary axis
3-12 (-14) cm. long, 1.2-2.5 mm. broad, flattened with acute edges, pale
green as phylloclades, with 4-8 (—12) lateral axes; first internode (1-)
2-5 (-6) cm. long, median internodes mostly 0.5—2 cm. long. Cataphylls
of primary axis usually with deciduous tips: stipules and blade lanceolate,
not fused, c. 0.7-1.5 mm. long, copiously ciliate when young, becoming
dark reddish brown and glabrate with a narrow whitish margin. Phyl-
loclades usually flexible, elliptic- or oblong- to obovate-lanceolate, c. 3-10
(-13) cm. long, (0.2—) 0.5-1 (-1.2) cm. broad, obtuse to bluntly acu-
minate at the tip, with (8—) 10-25 nodes; veins occasionally distinct but
more often obscure and phylloclades merely longitudinally furrowed or
channelled; margins not differentiated.

Monoecious; cymules male or bisexual, of mostly 1 or 2 (rarely 3)
female and c. 2—5 male flowers.

Male flower: pedicel (1—-) 2-6 mm. long. Calyx reddish or cream-
colored; calyx-lobes 6 (rarely 5), membranous, more or less spreading, sub-
equal or sometimes unequal (outer lobes then narrower), elliptic to broadly
obovate, (0.8—-) 1-1.5 (-2.3) mm. long, entire, midrib unbranched. Disk-
segments usually 6, thin, flat, entire, roundish, c. 0.2—-0.5 mm. across, some-
times united in pairs or obsolete. Stamens 3 or less commonly 4 (rarely ie)
filaments 0.4-0.8 (—1) mm. long, united for usually 12 their length or
more into a column 0.3-0.7 mm. high, free ends spreading; anthers emar-
ginate, c. 0.2-0.3 mm. long, 0.4-0.5 mm. broad; anther-sacs rounded, de-
hiscing more or less horizontally; pollen grains c. 18-25 » in diameter.

Female flower: pedicel (1—-) 2-4 (-7) mm. long. Calyx yellowish-green
to pinkish; calyx-lobes 6, scarious, subequal or distinctly unequal, elliptic

194 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrix

to obovate, 1-1.5 (—2.2) mm. long, 0.7—1.5 (—2) mm. broad, entire, mid-
rib unbranched. Disk plane or shallowly cupuliform, c. 0.25 (—0.5) mm.
high, rim undulate or crenulate. Ovary sulcate, rugulose; styles c. 0.6-1
mm. long, free or basally united into a very short column not over c.
0.3 mm. high, somewhat flattened, shallowly to deeply 3—4-lobed, lobes
slender.

Capsule oblate, 3-4 mm. in diameter, reddish brown, rugulose, not veiny.
Columella 1-1.8 mm, long. Seeds trigonous, 1.4—2.6 mm. long, 1-2 mm.
radially and tangentially, reddish brown to fuscous, with finely raised dots.

Collected in flower and fruit throughout the year, but many collections
sterile.

Type: Jamaica, Browne (LINN, HOLOTYPE).

DIsTRIBUTION: seacoasts and lower altitudes inland, usually on lime-
stone, Swan Islands, Cayman Islands, and Jamaica (Mars XXXIV and
AAV IL),

SWAN ISLANDS: Eastern Swan Island, Moyne 6, 12 (K); Larger Island,
Nelson 103, 104, 105 (GH); “Crane Island’, 1897, Sharples (GH).

CAYMAN ISLANDS. Granp CayMAN: Georgetown, 1888, pa (ke)
1891, Hitchcock (MO); 1 mi. SE of Georgetown, Kings G.C-202 (BM, NY);
east end, Kings G.C.-116 (BM, NY); Grape Tree Point, Proctor eee 11976
(GH). Cayman Brac: 1888, Fawcett (K); 1924, Matley (BM): east end,
Kings C.B-70 (BM, NY).

JAMAICA: without specific locality, Bertero (MO; TYPE COLLECTION of P.
cognatus), MacFadyen (K), Shakespeare (BM), Swartz (BM. 5S), Wolle (GH).
ANOVER: Orange Bay Point, Harris 10262 (JAM); Lucea, 1891, Hitchcock
(MO). WerEsTMoRELAND: Negril, near lighthouse, Britton & Hollick 2075, 2075a
(NY), M. Farr (G Hl), Harris 10233 (JAM, P, US), Webster & W on 5062
(A, JAM, MICH); Negril Hills, 0.5 mile east of Little Bay, Proctor 11150
(GH). Sr. ELIZABETE: Yardley Chase, nee 1157 (NY), Harris 9666 (JAM,
S); Merriman’s Point, Proctor 15335 GH). TRELAWNY: Jackson Town,
Hunnewell 19767; 1 mi. NW of Stewart ae Pierce 24 (MICH) « Florida
each, Falmouth, West & Arnold 808 (GH). St. ANN: east of Rio mee
Patrick 306 (JAM); Discovery Bay (Dry Harbour), Dignum 86 (JAM),
Farr (GH), Hunnewell 15304, 18845 (GH); Runaway Bay, Orcutt 6119 in
US); St. Ann’s Bay, Gully Road, Britton 2515 (NY, HoLotyPE of Xylophylla
contorta); Ocho Rios, Hanewel) 18843 (GH), Proctor 9566, 15533 (A).
CLARENDON: Round Hill, Proctor 9481 (GH); Portland ane Howard 12007
(A), von der Porten (JAM), Webster 5111 (A. JAM, MICH). St. CATHERINE:
Bog Walk, Crawford 817 (NY): Great Goat Island, paves éy Hollick 1891
(NY, US), Harris 9334 (JAM, NY), 9339 (A, NY, US); Lazaretto, Proctor
9968 (GH): Port Henderson Hill, Webster & Wilson 4925 (A, JAM, MICH).
st. Mary: Cabarita Island, off Port Maria, Proctor 7549 (GH), 7550 (JAM).
St. ANDREW: Cinchona, Harris 8592 (MO, NY); between Pleasant Hill and
Green Valley, Maxon & Killip 1058 (A. GH, NY, US); Long Mountain, Davis
(MICH), Proctor 7338 (JAM), 7387 (GH), Webster 4857 (A, JAM, MICH).
PorTLAND: Uncommon Hill, Proctor 8558 (JAM); Port Antonio, Britton 879
(NY), 910 (NY, US), Fredholm 3187 (US), Hitchcock (MO). St. THomas:
road to Hagley Gap, Harris 5830 (NY, US); Sheldon Road, Harris 12889 (GH

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 195

JAM, NY, US); Yallahs River, Purdie (K); Plantain Garden River, near White-
hall, Proctor 7421, 11778 (GH); Morant Point, Bengry (GH), Britton 4105
(NY), Webster 5558 (A, JAM, MICH).

The following collections include plants which presumably were cultivated or
naturalized:

FLORIDA. Dave Co.: Deering Hammock, Cutler, 1921, Small et al. (NY).
Monroe Co.: Key West, Blodgett (NY), Small et al. 10201 (NY). CUBA:
Ramon de la Sagra (A, P). HAITI: Ovesr: garden, Petionville, collected by
Barker, Ekman H9965 (S). DOMINICAN REPUBLIC: garden, Santiago,
Jiménez 1624 (US). VIRGIN ISLANDS: St. Thomas: Dec. 1880, Eggers 358
(GH); July 1882, Eggers 769 (G); Oct. 1882, Eggers (US).

Few species of Phyllanthus have a more interesting distribution than
does P. angustifolius; the occurrence of the plant outside of Jamaica only
on the Swan and Cayman Islands is an indication that the migrational his-
tory of this species may provide the key to some of the Pleistocene bio-
geographical problems of the Caribbean. Typically, P. angustifolius is a
xerophytic species of dry littoral areas, often on cliffs or platforms of
dogtooth limestone near the sea; however, it occurs inland at a number
of localities (notably in the Blue Mountains) and thus shows some adapta-
bility to colonizing different habitats. In the southeastern end of the John
Crow Mountains it may at the present time be interbreeding with P. arbus-
cula.

Morphologically, P. angustifolius and its sister species P. proctoris are
distinguished by their dark ciliate cataphylls, often in a distinctly small
apical cone, and by the narrow phylloclades which are often paired at
the end of the primary branchlet axes. Although closely related, P. proc-
toris appears to be sufficiently distinct by virtue of the floral characters
mentioned below

Phyllanthus ea is a rather variable species and individual
specimens often look distinctive. A form with narrow, elongated and often
twisted phylloclades sometimes occurs and was the pects for the proposed
Xylophylla contorta of Britton. In the vicinity of Port Antonio occurs a
form with unusually large flowers, but it does not diverge sufficiently to
warrant taxonomic recognition. The specimens from the Cayman Islands
and Swan Islands are quite typical for the species and show no unusual
features; this may indicate that the migration of the species there took
place relatively recently.

82. Phyllanthus proctoris, nom. nov. (PLATE XXXI, figs. V-W).
Phyllanthus linearis sensu Griseb. Fl. Br. W. Ind. 35. 1859; Muell. Arg.

in DC. Prodr. 15(2): 430. 1866 (excl. 8 cognatus) et Fawc. & Rend. FI.
Jam. 4: 266. 1920; non P. linearis (Sw.) Sw

Shrub up to 3 m. high; branches of current year’s growth c. 1.5-2.5 mm.
thick, terete, greyish-brown. Apical cone roundish, c. 2.5-3 mm. long and
2—2.5 mm. broad, smooth in outline; cataphylls‘of main axes deciduous,
blackish and scarious: stipules triangular-lanceolate, c. 0.8-1.7 mm. long,

196 JOURNAL OF THE ARNOLD ARBORETUM _ [vor. xxxIx

blade lanceolate, about as long. Branchlets pinnatiform, ultimate axes
transformed to phylloclades; primary axis (3—) 5-13 cm. long, 1-2 mm.
broad, flattened, edges obtuse, pale green as ultimate axes, with c. (4—)
6 or 7 lateral axes; first internode (1—) 2-6 cm. long, median internodes
mostly 0.7-1.5 cm. long. Cataphylls of primary axis deciduous; stipules
lanceolate, c. 1 mm. long, basally auriculate, densely ciliate marginally
(when young; soon glabrate), reddish brown; blade about as long as the
stipules. Phylloclades thin and flexuous, linear-lanceolate, c. (3—) 4-11 cm.
long, 0.1-0.35 cm. broad, attenuate to the tip, finely striate or sulcate longi-
tudinally (veins obscure) with 7-10 (—13) nodes; margins not differen-
tiated. Euphylls not observed. Cataphylls of ultimate axes basally auricu-
late, c. 0.5—-0.8 mm. long, prominently ciliate on margins (when young).

Monoecious: cymules male or bisexual, of 1-3 male flowers and/or 1
female flower.

Male flower: pedicel 2—3.5 mm. long. Calyx presumably greenish; calyx-
lobes 6, membranous, more or less spreading, subequal, elliptic to obovate,
0.8-1 mm. long, 0.6-1 mm. broad, entire, midrib unbranched. Disk seg-
ments 6, roundish, c. 0.2 mm. across. Stamens 3; filaments united into a
column 0.2—0.25 mm. high; anthers subsessile atop the column, c. 0.15 mm
long and 0.25 mm. broad; anther-sacs dehiscing horizontally; pollen grains
14-17 pn.

Female flower: pedicel mostly 2-3 mm. long. Calyx presumably green-
ish; calyx-lobes 6, unequal, the outer one or two usually brownish and
scarious, only 0.7-0.9 mm. long; inner lobes chartaceous, broadly obovate,
c. I-1.2 mm. long, 0.8-1.2 mm. broad, entire, midrib unbranched. Disk
shallowly cupuliform, irregularly notched, c. 1.2 mm. across. Ovary sub-
globose, smooth; styles divergent, dilated, obcuneate with crenulate distal
margin, c. 0.5-0.9 mm. long, 0.5-0.8 mm. broad, horizontally spreading
or reflexed over the ovary and forming a calyptra. Fruit and_ seeds
unknown.

Collected in flower Jan.

Type: Jamaica, Purdie.

DISTRIBUTION: seacoast and inland hills, western Jamaica (Map
XXXVIT).

JAMAICA. HANover: a een (MO); Eton, alt. 100 ft., Harris
12871 (JAM, MO, NY, US). TMORELAND: seacoast, in. 1844, Purdie os
HOLOTYPE). ST. James: ne Hill, in garden, 1897, Poweate (NY).
ELIZABETH: Mulgrave, alt. 1300 ft., Bares 12382 (MO, NY, US). Tee
TER: Fairfield, Waullschlaegel 1011 (G, GOET, M). Parish not designated:
Macfadyen (GOET, , beds

This species has long been misunderstood, owing to the fact that Swartz’s
type represents a peculiar plant which has not been recollected. It is clear
from details of his description (such as pedicels 4 lines long and styles
bifid) that Swartz’s P. linearis (F1. Ind. Occ. 113-114. 1800) cannot be
the same as the plant described above and accepted as P. linearis by
Grisebach, Mueller, and Fawcett and Rendle. The long pedicels suggest an

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 197

affinity of Swartz’s plant with P. arbuscula which has been confirmed by
examination of the type specimen of P. linearis, The latter is not typical
for P. arbuscula and may possibly be a hybrid, but in any event Swartz’s
name cannot be applied to the present plant.

Since a new name is required, it seems appropriate to name the plant
for Mr. George Proctor of the Institute of Jamaica. Although he has not
collected this species, his extensive gatherings of representatives of sect.
Xylophylla and of other Jamaican groups of Phyllanthus have supplied
valuable data during this study.

The closest relative of P. proctoris is certainly P. angustifolius, with which
it agrees in having darkened ciliate cataphylls and often paired phylloclades
at the end of the main branchlet axis. However, the phylloclades of P.
proctoris are consistently narrower than all except a few aberrant forms of
P. angustifolius, and no specimen of the latter has such highly modified
styles. Thus although the two species are certainly very closely related,
they nevertheless appear to be quite distinct.

83. Phyllanthus epiphyllanthus L. Sp. Pl. 981. see

Shrub or small tree generally 1-3 m. high, trunk slender, erect, sparsely
to considerably branched; branches of current year’s growth 2.5-10 mm
thick, reddish brown becoming greyish, smooth, lenticels very inconspicu-
ous. Apical cone usually small, c. 2-4 mm. in diameter. Cataphylls per-
sistent, reddish brown becoming greyish, indurate, massive, the stipules
fused with the blade in the lower half or entirely coalescent into a deltoid
scale, c. 2.5-5 mm. long, more or less marginally ciliate when young but
later often appearing entire. Phylloclades flexible to rigid, borne c. 0.5-1
cm. apart on terminal branches, or densely clustered on lateral spur-
shoots, simple, often extremely variable in size and shape: linear to ob-
lanceolate, straight to markedly falcate, (3—) 5-25 (-32) cm. long, (0.3—)
0.5—2.3 cm. broad, long-attenuate to truncate at the tip, with 7-40 nodes,
these mostly in the distal half; midrib raised and rather conspicuous on
both sides, lateral veins departing at an acute angle, tenuous, neither
raised nor conspicuous; nodes mostly conspicuous along the margin, form-
ing distinct notches, each with a dense cushion of persistent bracteoles.
Euphylls never observed; cataphyll blade partially to completely united
with its stipules, c. 0.8-1 mm. long, copiously ciliate along the margin,
reddish brown, delicate and evanescent (scarcely evident except on young
expanding phylloclades).

Monoecious; cymules usually bisexual, each with 1-3 female and sev-
eral to many male flowers.

Male flower: pedicel capillary to thickish, 1-2.5 mm. long. Calyx
usually reddish- or pinkish-tinged; calyx-lobes usually 6, subequal to very
unequal, erect or spreading, sometimes connate below, oblong to ovate
or obovate (sometimes broader than long), c. 0.8-1.3 mm. long, outer lobes
often with dark brownish scarious fimbriate tips. Disk-segments usually
6, discrete or sometimes united in pairs, tenuous to massive, sessile to

198 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

pedicellate, more or less foveolate-pitted, 0.25—0.5 mm. broad. Stamens 3
(rarely 4 or with the rudiment of a fourth); filaments partially to wholly
united into a slender to massive column; anthers stipitate to sessile atop
the column, broadly deltoid, more or less emarginate, 0.15—-0.25 mm. long
and 0.25-0.5 mm. broad; anther-sacs rounded, dehiscing horizontally or
slightly deflexed; pollen grains c. 16-20 p in diameter.

Female flower: pedicel rather stout, curved, often greatly thickened
above, 0.5—2.5 mm. long. Calyx usually reddish- or pinkish-tinged; calyx-
lobes 6, chartaceous, erect or spreading, equal or unequal, oblong to obo-
vate (or broader than long), 0.7-1.5 mm. long, 0.5-1.5 mm. broad, entire
or outermost lobes with dark brownish scarious ciliate tips, midrib un-
branched. Disk variable, patelliform to urceolate and enclosing the ovary.
Ovary smooth to conspicuously tuberculate, sulcate; styles free or shortly
united into a column; style-ends spreading or deflexed, mostly dilated and
irregularly 5—8-lobed

Capsule oblate, trigonous, c. 3.5—-4 mm. in diameter, dark reddish brown,
not evidently veiny, smooth to incrustate-tuberculate; valves 2.5—3.2 mm
long. Columella c. 1.5 mm. long. Seeds asymmetrically trigonous, 1.4—
2.1 mm. long, 1—1.6 mm. broad, smooth, dark reddish brown.

Collected in flower and fruit throughout the year.

Phyllanthus epiphyllanthus, vegetatively the most specialized species of
sect. Xylophylla, is also the most widespread. It clearly appears to be
closely related to and presumably has been derived from P. angustifolius,
its simple phylloclades corresponding to the main axis of the branchlet of
that species. In floral characters the two plants are very similar but
P. epiphyllanthus usually has shorter pedicels and a greater number of
stylar branches. The extensive populations of P. epiphyllanthus may be
grouped into three distinctive and rather sharply defined subspecies which
can be separated by the following key.

KEY TO THE SUBSPECIES

1. Stipules of cataphylls completely fused with the blade into an entire scale
(rarely the extreme tips free); female disk c. 1/4 the height of the ovary
or less (rarely to 1/3); calyx-lobes free to base, spreading at anthesis.

ssp. epiphyllanthus

8 Stipules of eueey ee Sualgey! fused with the Bidde at base, the tips
free; disk covering 1/3 or more of the ovary; calyx-lobes more or less erect.
2. Disk covering c. on - 1/2 of the ovary; calyx-lobes definitely con-

nate below, forming a cup in which the disk is concealed... ssp. domingensis

Disk urceolate, entirely enclosing the ovary; calyx-lobes only barely, if

at all, united at base... 0... ee ssp. dilatatus

bo

83a. Phyllanthus epiphyllanthus ssp. epiphyllanthus
(PLATE XXXT, figs. X-Y).
Phyllanthos americana Sue flores e singulis foliorum crenis proferens Herm.
rodr. 385. 1689 [nom.]; Commelin, Hort. Med. Amstel. Rar.
199-200, ne 102. 1697,

1958 | WEBSTER, WEST INDIAN PHYLLANTHUS 199

Phyllanthus foliis lanceolatis serratis: crenis floriferis L. Hort. Cliff. 439. 1738
cl. ref. Sloan.].

Phyllanthus epiphyllanthus L. Sp. Pl. 981. 1753.

Xylophylla falcata Sw. Prodr. 28. 1788.

Phyllanthus falcatus (Sw.) Gmel. Syst. Nat. ed. 13. 2: 204. 1791.

Phyllanthus epiphyllanthus 8 genuinus Muell. Arg. in DC. Prodr. 15(2): 428.
1866; in part [excl. ref. Linden et Schomb. |.

Diasperus epiphyllanthus (L.) O. Ktze. Rev. Gen. 2: 599. 1891.

Xylophylla epiphyllanthus (L.) Britton in Small, Fl. Florida Keys 76. 1913.

Exocarpus epiphyllanthus (L.) Merr. Interpr. Rumph. Herb. Amb. 208. 1917
(as to type only).

Shrub 0.5—-2.5 m. high; cataphylls of branches with stipules and blade
completely fused into a deltoid scale (rarely the extreme tips becoming
free) (1.5-) 2.5-4 (-5) mm. long. Phylloclades flexible to rigid, linear
to oblanceolate, often falcate, 2.5-20 cm. long, 0.3-2 cm. broad, truncate
to attenuate at the tip, with (7-) 9-30 (—35) nodes.

Male flower: pedicel slender, 1-3 mm. long. Calyx-lobes spreading,
usually in two dissimilar whorls, the outer oblong to obovate, c. 0.5-1
(-1.5) mm. long and 0.5-0.9 (-1.3) mm. broad, the inner ovate to orbicu-
lar or broader than long, 0.7-1.5 mm. long and broad. Filaments mostly
united 14 to 44 their length into a column 0.2-0.6 (—0.9) mm. long.

Female flower: pedicel becoming 0.5—2 mm. long. Calyx-lobes spread-
ing, similar to the male, the outer c. 0.5-1 mm. long and broad, inner
c. 0.8-1.3 mm. long and broad. Disk patelliform or shallowly cupuliform,
at most not over 0.2 mm. high, usually enclosing the basal 14 of the
ovary (rarely to 43). Ovary rugulose to conspicuously tuberculate, at
least above; styles nearly free or less commonly united into a column
up to 0.5 (rarely to 1.3) mm. high, style-tips spreading, distally lacerate,
0.5-1 mm. long. Capsule apically rugulose to conspicuously tuberculate,
34.2 mm. in diameter, valves 2.2-3 mm. long. Seeds 1.4-2.1 mm. long,
1-1.6 mm. broad.

Type: Herb. Hort. Cliffort. (BM). The specimen was probably col-
lected in the Bahamas by Catesby and given to Linnaeus by Miller.

DiIstRIBUTION: widespread in the West Indies, mostly on limestone
rock, often near the sea (Map XXXIV).

BAHAMAS. Bruinr: North Bimini, Howard & Howard 10086 (GH, NY,
US), Millspaugh 2375 (NY); South Bimini, Millspaugh 2411 (NY). ANDROS:
Mangrove Cay, Brace 4864 (NY, US); Mastic Point, Brace 7096 (NY);
Nicols Town, Northrop & Northrop 325 (A, G, GH), Small & Carter 8954
(NY, US). New Provipence: Blue Hills Road, Britton 14, 57 (NY); Far-
ringdon Road, Britton & Brace 223 (NY, US); Adelaide, Britton & Brace 520
(NY); Nassau, Britton & Millspaugh 5357 (NY), Curtiss 4 (A, G, GH, MO, P,
US), Earle 57 (NY), Northrop & Northrop 146 (NY); Mt. Vernon Estate,
Coker 31 (NY); Cooper 48 (NY); Hog Island, Eggers 4059 (US), Wilson 8255
(NY); von Reis 227 (MICH, US); Grantstown, Wight 18 (GH, NY), Wilson
8184 (NY). Exuma: Little Galiot Cay, Britton & Millspaugh 2837 (NY);
Great Guana Cay, Britton & Millspaugh 2866, 2919 (NY); Great Exuma,

200 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

Haynes Road, Britton & Millspaugh 3034, 3038 (NY). Wat inc’s IsLanp:
Columbus Monument, Britton & Muiullspaugh 6179 (NY); Hitchcock (MO);
southeast end, Wilson 7266 (GH, NY). Lonc IsLanp: Britton & Millspaugh
6294 (NY); Water Cay, Coker 524 (NY). Crooxep IsLanp: Brace 4538 (NY).
Atwoop Cay: Wilson 7381 (GH, NY). Great RAccEep IsLtanp: Wilson 7816
(GH, NY). Lone Cay (Fortune I.): Brace 4026, 4216 (NY), Hitchcock (MO).
MayaGuaNna (Mariguana): 10 mi. west of Abraham Bay, Wilson 7431 (GH,
NY). Great Inacua: Matthewtown, Harshberger 13 (US), Nash & Taylor 875
(NY). Carcos Istanps: South Caicos, Wilson 7662 (GH, NY). Turxs IsLanps:
Hjalmarsson (GOET); Grand Turk, Lewis (GH), Nash & Taylor 3766 (NY),
Proctor 8764 (GH)

CUBA. Maranzas: Pan de Matanzas, Ekman 16454 (S). Las VILtas:
Calicita, Combs 502 (GH, MO, NY, P, US); Cienfuegos Bay, Britton & Wilson
5724 (NY), Fernando 734 (NY), Howard 5462 (NY), Jack 4190 (A), 5110
(A, NY, US), 7559 (A, US); Trinidad Mts., Ekman 13904 (S), Webster 4784
(GH); La Vigia Hill, Britton & Wilson 5532 (NY); Rio Toyaba, Britton et al.
557 (NY). CAMAGUEY: Jatovieja, Cayo Sabinal, Shafer 1059 (NY, US); Cayo
Paredon Grande, Shafer 2747 (NY, P, US); Cayo Guajaba, Shafer 728 (NY),
2813 (NY, US); Pastelillo, near Nuevitas, Ekman 15401 (S). Oriente: Las
Salinas, Puerto Padre, Curbelo X60 (NY); Santiago, mouth of Rio Aguadores,
Ekman 9232 (S); Jauco, Jervis 3381 (GH); Leon 11851, 11767 (NY).

HAITI. Ite pe 1a Tortue: Embouchure de la Rochelle, Ekman H4190 (US).

DOMINICAN REPUBLIC. Setgo: Isla Saona, Taylor 506 (NY).

PUERTO RICO. Mona Istanp: Britton et al. 1748 (NY), Noble 2827 (NY,
US), Stevens 6447 (NY). AcuaDILLA: Quebradillas, Britton et al. 1955 (NY,
US), Sargent 720 (US), Stevens & Hess 5146 (NY); Camuy, Hess 2583 (NY).
Areciso: Vega Baja, Britton et al. 6925 (NY). Mayacuez: Guayanilla, Brit-
ton & Shafer 1833 (NY, US); Yauco, Garber 120 (GH). Ponce: west of Ponce,
Britton & Britton 7354 (NY), Peller 6317 (A, G, GH, MO, NY, P, US).
SAN JuAN: Bayamon, Britton et al. 1518 (NY, US), Heller & Heller 398 (NY,
US), Johnston (NY, US), Sintenis 1028 (GH, US), ne 1074 (US), Steven-
son 382 (US). Vieques IsLtanp: 1876, Eggers (GH); Campo Asilo, Eggers
(MO); Eggers 769 ex p. (P); Cayo Puerto Real, Shafer 2757 (NY, US).

LESSER ANTILLES. Vircin Gorpa: North Sound, Fishlock 11 (GH, NY,
US). ANGUILLA: Goodwin & Goodwin 9 (NY). Sr. Marri: Boldingh 2785B
(NY), Rijgersmaa (S). Barsupa: Codrington, Beard 373 (A). AnticUA: Peli-
can Bay, Box 552 (US); Wullschlaegel 498 (GOET). GuapEeLoure: Beaupertuis
(P), Duchassaing (GOET, P), L’Herminier (G), Perrottet (G), Segretain oy
Vieux-Fort, Gosier, Duss 2444 (NY, US); Gosier, Quentin 555 (P). Ite
SIRADE: Duss 210 (P), Stehle 215 (NY). Marte-Gatante: Guadeloupe 2
Marie-Galante, Richard (P). MARTINIQUE: piace (P), Hahn 323 (G, P);
Jardin botanique, Duss 2044 (NY, US), Plee (P). Barsapos: Curran 59 (GH),
Waby 34 (US); Chancery Lane, Dash 554 (NY, US); Forster Hall Wood,
Eggers 7234 (GOET, P, US); Bathcheba, Miller 04 (US). Trrnipap: Sieber
Flor. Trinitat, 337 (G, MO; cultivated?).

Subspecies epiphyllanthus is one of the most familiar of the West Indian
representatives of Phyllanthus due to its abundance on limestone shores
throughout much of the Greater and Lesser Antilles. It has received many
common names, including “seaside laurel” and “sword bush” (Fawcett and

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 201

Rendle), “mutton parish” (in Barbuda, ex Beard), “duppy bush” (in
Cayman Islands), ‘‘farine 4 zombi,” “farine chaude,” “langue a chat” (in
Guadeloupe), and “panetela” (in Cuba). Roig (Dict. Bot. Nombres Vulg.
Cub. ed. 2. 2: 743. 1953) considers that the Cuban name alludes to
the pleasant nocturnal aroma of the flowers, which resembles a certain
kind of pastry; the epithet “farine chaude” from Guadeloupe would ap-
pear to be derived in the same way. This distinctive nocturnal fragrance
is interesting, since it has not been observed in any other species of the
genus.

The report of Stehlé and Quentin (Fl. Guad. 2(1): 51. 1937) from
St. Barthélemy is doubtless correct and refers to this subspecies, but most
other reports of P. epiphyllanthus from localities other than cited herein
are to be regarded with suspicion. The record of P. epiphyllanthus from
Jamaica on the basis of a Purdie specimen at Kew (Fawc. & Rend. Fl.
Jam. 4: 261. 1920) can presumably be discounted, since numerous sub-
sequent collectors have encountered only P. angustifolius along the shores
of that island. Since Purdie is known to have collected on Puerto Rico
and Guadeloupe (Urban, Symb. Ant. 3: 107. 1902), his specimen may
have been taken on either of those islands. The report of Xylophylla
epiphyllanthus from the Florida Keys by Small (Flora Florida Keys 76.
1913) is based on a misidentification of a doubtfully indigenous specimen
of P. angustifolius. Lemée (Fl. Guyane Fr. 2: 263. 1952) has recorded
P. epiphyllanthus from French Guiana on the basis of an ambiguous cita-
tion by Aublet (Hist. Pl. Guy. Fr. 2: 853. 1775), but if Aublet encoun-
tered the plant there it must have been cultivated. The only disjunct
record which cannot be positively rejected is that from Trinidad by Sieber,
but the failure of recent collectors to confirm it strongly suggests that
Sieber’s plant too may have been taken from cultivation.

Subspecies epiphyllanthus consists of many isolated populations which
may be grouped into three races. The plants of the Bahamas and eastern
Cuba have relatively short phylloclades with fewer nodes and a distinctly
tuberculate fruit. The plants from Matanzas and Las Villas are vegeta-
tively similar to the Bahaman ones but have peculiarly dilated more or less
calyptriform styles. The populations of Puerto Rico and the Lesser Antilles
tend to have longer phylloclades (up to 20 cm. long) with more nodes,
and nearly smooth capsules. However, the Barbados population is vegeta-
tively so like the Bahaman that it breaks down the distinction between
that race and the Lesser Antillean one; and, since it is not certain that
the distinctive styles of the central Cuban race are a consistent character,
it does not seem necessary to define formal varieties within the subspecies.

The peculiar distributional relationships of ssp. epiphyllanthus demand
further investigation. Typically the subspecies is (like P. angustifolius )
a littoral plant of coral benches, but in central Cuba and Puerto Rico it
occurs in upland rain-forest. In both eastern Cuba and Hispaniola, on the
other hand, it does not appear to go inland, but is instead replaced by
ssp. dilatatus and ssp. domingensis respectively. In areas where moun-
tainous regions approach near the coastline, it might be expected that ssp.

202 JOURNAL OF THE ARNOLD ARBORETUM _[VvoL. XxxIx

epiphyllanthus would come in contact with the other two subspecies; but
thus far they have not been observed to grow sympatrically at any point
and no specimens intermediate between any of the subspecies have been
collected.

83b. Phyllanthus epiphyllanthus ssp. domingensis, ssp. nov.~"

a oo 6 genuinus Muell. Arg. in DC. Prodr. 15(2): 428.
866 (ex p., excl.

Shrub or small tree, 1-2 m. high or more (up to 10 m., ex Jiménez) :
cataphylls of branches massive, 3-6 mm. long, blade and stipules fused
at base (c. lower 1 mm.) but distally free. Phylloclades more or less rigid,
narrowly linear to broadly oblanceolate, straight to distinctly falcate, 10—
31 cm. long, (0.4—) 0.7-1.3 (-—2.5) cm. broad, usually long-attenuate at
the tip, with (15—) 20-40 nodes.

Male flower: pedicel distally thickened, 1-2.5 mm. long. Calyx-lobes
subequal, basally connate above the receptacle into a cup c. 0.2-0.3 mm.
high; free ends of lobes erect, obovate, 0.8-1.3 mm. long and broad. Disk-
segments mostly 0.2—0.3 mm. across. Filaments almost completely united
into a column 0.5—0.8 mm. high. Female flower: pedicel becoming 1.3—2
mm. long. Calyx-lobes erect, basally connate into a cup c. 0.4—0.5 mm.
high; free ends of lobes broadly obovate, 0.8-1.2 mm. long and ann
becoming up to 1.5 mm. long in fruit. Disk cupuliform, c. 0.3-0.5 n
high, usually enclosing the ovary for about 4 to ™% its height, pene are
rimmed. Ovary smooth; styles free or united below the middle, dilated
and channelled distally, c. (0.3—) 0.5-0.7 mm. high, the deflexed limb
lobed but lobes often inturned and thus margin appearing entire. Entire
aes not seen; valves 2.8-3.2 mm. long; seeds 1.7—1.9 mm. long,

4mm. broad

Type: Dominican eon Ekman H12642.

DISTRIBUTION: wooded hillsides, sandy or calcareous soil, northern His-
paniola (Map XXXIV).

HISPANIOLA: “St. Domingue,” Poiteau (G), Richard (P).

HAITI. Norp: es Haitien, slopes of Morne Haut du Cap, alt. 500 m.,
Ekman H2727 (S, US).

DOMINICAN REPUBLIC. Monte Cristi: Moncién, banks of Rio Mao,
alt. 200 m., Valeur 164 ex p. (A, US). Santraco: Monciodn, at Rio Magua, alt.
200 m., Ekman H12642 (S, HOLOTYPE; US, tsotypPe); Arroyo Vallecito, Jicomé,
Valeur 164 ex p. (G, MO, NY). SAMANA: Samana Peninsula, Laguna, Pilon de
Azucar, alt. 100-500 m., Abbott 2360 (US).

Vegetatively, ssp. domingensis closely agrees with ssp. dilatatus in its
unfused cataphylls and usually elongated or broadened phylloclades. It

“ Phyllanthus epiphyllanthus ssp. domingensis, ssp.

Cataphyllis trifidis; laciniis calycis ad basin connatis; paeeniid omnino coalitis;
disco femineo cupuliforme 0.3-0.5 mm, alte; ovario laeve; stylis liberis vel iifeve
connatis

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 203

is transitional between ssp. dilatatus and ssp. epiphyllanthus in its floral
characters, having the smooth ovary of the former but a shallow female
disk tending toward that of the latter; however, its basally cupulate calyx
is different from either of the other two subspecies. On the whole, ssp.
domingensis is much more closely related to ssp. dilatatus than to ssp.
epiphyllanthus.

83c. Phyllanthus epiphyllanthus ssp. dilatatus (Muell. Arg.), stat.
nov. (PLATE XXXI, fig. Z).

Phyllanthus epiphyllanthus a dilatatus Muell. Arg. in DC. Prodr. 15(2): 428.

Phyllanthus epiphyllanthus 8 genuinus Muell. Arg., ibid., ex p.

Shrub or small tree up to c. 3 m. high; cataphylls of branches massive,
2-4 mm. long, the linear-lanceolate stipules and blade fused in the lower
1 mm. but distally free. Phylloclades more or less rigid, linear to oblanceo-
late, straight or somewhat curved (5—) 10-20 cm. long, (0.3-) 0.7-2.5
(—2.8) cm. broad, with c. 20-30 nodes.

Male flower: pedicel 1-3 mm. long. Calyx-lobes subequal, erect (tips
sometimes flaring), barely if at all connate below, obovate, 0.8-1.2 mm.
long, 0.6-1.5 mm. broad. Filaments completely united into a rather slen-
der column 0.5-0.8 mm. high. Female flower: pedicel 1-2 mm. long;
calyx-lobes subequal or somewhat unequal, oblong to broadly obovate,
0.9-1.3 mm. long, 0.5-1.3 mm. broad; disk urceolate, longitudinally ribbed,
entirely enclosing the ovary, 0.8-1.3 mm. high; ovary smooth; styles
0.5-0.8 mm. high, free or more commonly united into a column 0.3-0.6 mm.
high, style-tips spreading or reflexed, adaxially channelled, with 3 or 4
lobes. Capsule 3.5-4 mm. broad, smooth; valves 2.8-3.3 mm. long; seeds
1.8-2.1 mm. long, 1.4-1.5 mm. broad.

Type: Cuba, Wright 1951.

DIsTRIBUTION: calcareous uplands, eastern Cuba (Map XXXIV).

CUBA. OrtENTE: valley of Rio Cauto, Linden 2140 (G, P); Sierra de Nipe,
Rio Piloto, Ekman 3391 (S); Monte Picote, Ekman 7408, 9130 (S); Caridad
de los Indios, Jervis 1369 (GH, MICH); Monte Libano, San Fernandez, Ek-
man 10271 (S); El Palenquito, alt. 600 m., Eggers 4844 (GOET, P); cliffs
(paredones), Monteverde, Wright 1951 (G, HoLotyPE; GOET, MO, NY, P,
US, isotypes; data ex NY); Baracoa, pinelands on road to Florida, Ekman
3989a (S); top of El Yunque, 1955 Harvard Trop. Bot. Course 469 (GH).

The epithet dilatatus assigned to this taxon by Mueller is not wholly
appropriate, for some forms such as that represented by Linden 2140 have
long narrow phylloclades as in ssp. epiphyllanthus. On the basis of the
much more extensive material which has become available since Mueller’s
treatment, it is now apparent that ssp. dilatatus cannot be distinguished
vegetatively from ssp. epiphyllanthus except by its cataphylls. Even the
“typical” broad-bladed form of ssp. dilatatus (as Wright 1951) can be
matched by phylloclades from Bahaman populations of ssp. epiphyllanthus

204 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

(e.g., Coker 524 and Britton & Millspaugh 6294 from Long Island). The
diagnostic characters of ssp. dilatatus reside not in the vegetative parts
but rather in the distinctive female flowers. No form of ssp. epiphyllanthus
has such a highly developed female disk and, if it were not for the transi-
tional character of the plants placed in ssp. domingensis, the population in
the inland parts of Oriente province would have all the attributes of a
distinct species.

ACKNOWLEDGMENTS

The basic framework of this study was constructed during the writing
of a doctoral dissertation which was submitted at the University of Mich-
igan in 1953. The inspiration for the investigation was given by Dr. Rogers
McVaugh, who suggested the need for detailed study of Phyllanthus and
who directed the research program with sympathy and understanding.
Professor H. H. Bartlett generously supported field work in the southern
United States, Cuba, and Jamaica, criticised the manuscript with thorough-
ness, and revised (and in many instances entirely recast) the author’s
attempts at scientific Latin. Dr. W. H. Wagner, Jr. gave valuable advice
and encouragement on morphological techniques and suggested some im-
portant improvements in terminology. Financial assistance from the Re-
search Club of the University of Michigan and from a Rackham Fellow-
ship greatly expedited completion of the dissertation.

Investigations on the West Indian species of PAvllanthus were continued
at Harvard University under a National Science Foundation fellowship
kindly sponsored by Dr. Reed C. Rollins and Dr. I. W. Bailey, who made
available various needed facilities. Dr. Richard A. Howard assisted in
many ways, including supporting field work in Jamaica, furnishing many
interesting specimens, and imparting freely of his extensive knowledge of
Caribbean plant life.

Roy N. Jervis, of East Texas State College (Commerce), made
possible field work in Cuba in 1951, collected extensive series of specimens,
and generously supplied a valuable series of outline maps drafted by him-
self. Mr. Bernard Lewis and other staff members of the Institute of
Jamaica materially assisted in the success of field work on that island in
1954. Mr. George Proctor, of the Institute, aided in many ways in these
explorations, and has since forwarded both living and dried collections
of unusual interest. Dr. Kenneth A. Wilson collaborated helpfully on
the Jamaican field work throughout the summer of 1954. Special collec-
tions or materials of West Indian species of Phyllanthus were also fur-
nished by Hermano Alain (Colegio de la Salle, Havana), Ing. Julian
Acuna (Santiago de las Vegas), Dr. E. P. Killip, and Dr. Fred A. Barkley.

Several persons have lent assistance in the preparation of illustrations.
Dr. Ding Hou kindly furnished the drawings for PLates XIII and XIV;
Mrs. Martha Szerlip Wittels applied her talented pen to the execution of
Text-fig. 4 and Pirates V, VIII, and IX; and Dr. Richard A. Scott took
most of the photographs used for PLATES VI, VII, XI, and XII. Photo-

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 205

graphs of type specimens were taken at the University of Michigan through
the courtesy of Dr. McVaugh. The other drawings were skillfully recon-
structed from the author’s sketches by his wife, Dr. Barbara D. Webster.
In addition to the outline maps of Cuba and Jamaica furnished by Dr.
Jervis, outline maps of Caribbean America and the Western Hemisphere
from Goode’s series of base maps (University of Chicago) were utilized
in preparing distribution maps. For the publication of all of these illus-
trations in their present form and for the execution of those not specifically
mentioned, the author is alone responsible.

Particular gratitude is expressed to Dr. C. E. Kobuski and to Dr.
Carroll E. Wood for their forbearance and patience in editing what must
have appeared to be a never-ending manuscript.

Finally, the author wishes to record his particular indebtedness to his
wife, Barbara, for her untiring and indispensable assistance. Not only has
she contributed by typing manuscript, inking in most of the line drawings,
and preparing mounts for microscopic study, but she has with great under-
standing provided encouragement throughout the extended period in which
this manuscript has taken its final form.

BIOLOGICAL LABORATORIES
HARVARD UNIVERSITY

CORRECTIONS

37: 92. Changes in concept during the course of this work have resulted in a
total of 83 West Indian species (instead of ‘“90-odd”) of Phyllanthus, of which
about 75 are indigenous.

37: 99. Read 1,000 pages for 1,000 species.

37: 120. Read 1450 pz instead of 4150 p» for P. pachystylus.

37: 219. Phyllanthus elegans has not been introduced into the West Indies;
it has been confused here with P. pulcher.

37: 256. Read Phyllanthus hyssopifolioides HBK. for Phyllanthus hyssopi-
folius HBK.

37: 340. Read Isla de Providencia instead of Providenciales.

37: 344. Delete footnote 3.

37: 345. Sect. Apolepis (Subg. Conami) will not run down properly in the
key to subgenera unless the expression “entirely woody” under lead 2 is ignored.

38: 177. Read Cyclanthera for Cyllanthera.

38: 298. Plate XXII, A is more applicable here than XXII, B.

38: 323. Read fig. B for fig. A under P. fadyeniz.

39: 58. Read P. anderssonii for P. barbadensis.

206 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxIx

APPENDIX. I*
ARTIFICIAL KEY TO THE WEST INDIAN TAXA OF PHYLLANTHUS

1. Branching not phyllanthoid (i.e., ultimate axes persistent, their subtending
leaves not reduced to scales).
2. Herbs; stamens free; pedicel of female flower ~ seeds less than 2 mm.
OM: wae aoe eee bao eae Lake aah Sect. Loxopoptum (37: 346)
2. Shrubs or trees; stamens connate; pedicel female flower and seeds
Pore (han 2mm. WONG: is eases ee okra Sect. ELUTANTHOS (39: 51)
1,

Branching phyllanthoid (i.e., ultimate axes Peer their subtending leaves
usually reduced to scales).
2. Branchlets with normal leaves, not greatly dilated.
3. Branchlets pinnatiform (with a single lateral axis in spp. of Sect.
pgesiiae
Specimens with flow
5. Stamens inl anthers wholly confluent into a circumcis-
sile synandriu
6. noe ecole blackened and indurate; column of
synandrium at OP MOTE AIG. eee easy ES ie oe 44 Oh ds
“ieee P. dimorphus (Sect. PHYLLANTHUS) (38: 341)
6. Monoecious: Sipuiles paler and thinner; column of synan-
drium less than 1 mm. high. . Sect. CYCLANTHERA (38: 177)
5. Stamens not wholly confluent.
6. Dioecious.
7. Herbs; flowers in axillary cymules
Sect. PHYLLANTH ius, Subsect. PENTAPHYLLI (38: 298)
7. Trees; flowers in cauliflonous paca ee Le eee eee
Sec APOROSELLA (38: 72)
6. Monoecious (male flowers smetine soon deciduous and
specimens then appearing fem
7. Annual herb; stamens 5

ati ee. weees
be eat hy dew en oases Sect. FLoRIBUNDI (38: 51)
7. Habit various; stamens, if 5, i least partially connate.
8. Flowers, at least in part, in naked thyrses.
9. Stamens free; styles bifid, slender .........
Perec saree oon te oe eee bg. Crceca (38: 60)
9. Stamens united; styles thickened or 3-4 lobed.
10. Anthers dehiscing vertically; ea of female
flower urceolate, ne the ovary. ......
ended tere Hed te tien. (38: 75)
Anthers dehiscing Reena disk of fe-
male pone much shorter, not enclosing the
OVATY.: » dosed ee Sect. Epistytium (39: 153)
8. Flowers in ae eee to normal leaves.
9. Ovary with c. 9 or 10 carpels, styles minute and
aggregated into an irregular mass; androecium of

oS

* Numbers in parentheses refer to volume and page number of the taxon under
which the unknown plant may be traced to species.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 207

4. Specimens with fruits on
Fruits woody or fleshy, indehiscent or cue! dehisc
Sec

6. Fruits baccate

5 stamens united in two sets. ..............

Sect. ANISONEMA (38: 56)

9. Ovary with only 3 carpels, styles otherwise.

o

Annual or perennial herbs; stamens 2 or 3;
styles bifid, free except at the base, never
greatly dilated or united into an erect column.
11. Leaves as broad as or broader than long
(c. 5-10 mm.); stamens 3, free; female
disk dissected a 6 segments. .......
Sect. ApoLeris (38: 371)
11. Not with this ae care of characters.
Subg. PHyLLANTHUS (38: 170)
Shrubs or trees; stamens 2— ee styles vari-
ous, often united into a colum
11. Calyx-lobes dark red are lacerate mar-
gins; stamens 2, filaments united; axes
reddish- mieenode Deb tees ey:
Sect. Ertococcus (38: 360)
1k Calyz- lobes at most minutely denticu-
late, never lacerate. ..... Subg. XyLo-
PHYLLA (39: 68

lye

a eam Tea 56)

6. Fruits drupaceous or woody.
7. Cocci firmly united, not separating. Subg. Crcca (38: 60)
7. Cocci at length separating.

8, Leaves linear-oblong, 2-5 mm. i i ee

Lica (38: 75)

Sec
8. Leaves aes broader. Sect. ene (39: 142)
5. Fruits dry, dehisc
6. Herbs;

es Weare vatheende Su

seeds ne than 2 mm. long.
7. Eee orbicular or broader than long. ..............

ect. APOLEPIS (38: 371)
be. Parvitastas (38: 170)

6. Shrubs or trees; ae often 2 mm. long or mor

eer (39: 68)

3. Branchlets bipinnatiform (primary axis ee several to many lateral
xes)

4. Calyx-lobes 6; pubescence not reddish; leaves (at least the petioles)
and female pedicels scabridulous or hirst ulous.
Sec

P Nowioctrma (38: 363)

4. Calyx-lobes 5; pubescence more or less a. leaves and female
pedicels smooth and glabrous. Sect. HEMIPHYLLANTHUS (39: 163)

2. Branchlets with at least the lateral axes modified to phylloclades, leaf-
blades usually greatly reduced or absent. .. Sect. XYLOPHYLLA (39: 179)

208 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

APPENDIX II
WEST INDIAN SPECIES EXCLUDED FROM PHYLLANTHUS

PHYLLANTHUS ANTILLANUS (A. Juss.) Muell. Arg. — 32; 51, 1865,
(Cicca antillana A. Juss. Tent. Euphorb. pl. 4, fig. 13B. 1824.) = Margaritaria
nobilis L. f. Suppl. Pl. 428. 1781.

PHYLLANTHUS BAHAMENSIS Urb. Symb. Ant. 3: 289. 1902. (Margaritaria
bahamensis (Urb.) Br. & Millsp. Bahama Fl. 220. 1920.) = Margaritaria tetra-
cocca (Baill.) Webster, Jour. Arnold Arb. 38: 66. 1957.

PHYLLANTHUS CUNEIFOLIUS (Britton) Croizat, ae Wash. Acad. Sci. 33:
12. 1943. (Andrachne ? cuneifolia Britton, Mem. Torr. Bot. Club 16: 72. 1920.)
Not Phyllanthus; the disposition of this species is still uncertain.

PHYLLANTHUS GLABELLUS (L.) Fawc. & Rend. Jour. Bot. 57: 68. 1919.
(Croton glabellum L. Syst. ed. 10, 1275. 1759.) = Croton lucidus L. (as to
type, not as to Fawcett and Rendle’s application). Fawcett and Rendle cited
as basionym Croton glabellum L. Amoen. 5: 409. 1760 (non Syst., 1759), which
supposedly is the plant accepted below as an Astrocasia; but this procedure is,
of course, inadmissible under present rules of nomenclature.

PHYLLANTHUS HOTTEANUS Urb. & Ekm. Ark. Bot. 22A(8): 60. 1928.
= Margaritaria hotteana (Urb. & Ekm.) Webster, Jour. Arnold. Arb. 38: 66. 1957.

PHYLLANTHUS LAURIFOLIUS A. Rich. in Sagra, Hist. Nat. Cuba 11: 216. 1850.
= Savia sessiliflora (Sw.) Willd. Sp. Pl. 4: 771. 1806

PHYLLANTHUS NEOPELTANDRUS Griseb. Goett. Nachr. 1865: 167. 1865.
ae neopeltandra (Griseb.) Urb. Symb. Ant. 3: 285. 1902; Securinega
neopeltandra (Griseb.) Urb. ex Pax & Hoffm. Natiirl. Pflanzenf, 19¢: 60. 1931.)

= Chascotheca neopeltandra (Griseb.) Urb. Symb. Ant. 5: 14. 1904.

PHYLLANTHUS NopiLis (L. f.) Muell. Arg. in DC. Prodr. 15(2): 414. 1866.
= Be calles nobilis L. f. Suppl. Pl. 428. 1781.

PHYLLA US PORTORICENSIS (QO. Ktze.) Urb. Symb. Ant. 4; 338. 1905.
ene portorcensi O. Ktze. Rev. Gen. 2: 602. 1891; Conami portoricensis

. Ktze.) Britton, Sci. Surv. Puerto Rico 5(4): 475. 1924.) = Flueggea virosa
(Willd.) Baill. a Gen. Euphorb. 593. 1858

PHYLLANTHUS PUBIGERUS A. Rich. in sagia: Hist. Nat. Cuba 11: 216. 1850.
= Savia sessiliflora (Sw.) Willd. Sp. Pl. 4: 771. 6.

PHYLLANTHUS SCANDENS (Griseb.) Muell. Arg. in DC. Predr. 15(2): 415.
1866. (Cicca scandens Wr. ex Griseb. Goett. Nachr. 1865: 165. 1865.) = Mar-
garitaria scandens (Wr. ex Griseb.) Webster, Jour. Arnold Arb. 38: 66. 1957.

PHYLLANTHUS TREMULUS Griseb. Fl. Br. W. Ind. 34. 1859. SS
glabellus sensu Fawc. & Rend., non Croton glabellum L.; Astrocasia phyl
thoides Robins. & Millsp. Bot. Jahrb. Beibl. 80: 20. 1905.) = = Astrocasia ets
(Griseb.) Webster, comb. nov. The details of typification will be taken up in
a forthcoming study on the genus Astrocasia.

LANTHUS VIRENS (Griseb.) Muell. Arg. in DC. Prodr. 15(2): 415. 1866.

garitaria tetracocca (Baill.) Webster, Jour. Arnold Arb. 38: 66.

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 209

INDEX

Synonyms are printed in italics; new names in bold-face type.

Agyneia berterii, 39: 148
Andrachne brittonii, 38: 171
Andrachne cuneifolia, 39: 208

— pumila, 37:
A porosella, 38: a
Asterandra, 39: 146
Astrocasia Phsllanphoides, 39: 208
Astrocasia tremula, 39

Sei pean 39: 208
Cheram
Choris ona Pinnata, 38: 52

66

— racemosa, 3
ae oF G 39: 208
— viren
Conami Braensi, 38: 368
— conami, 65

— ovalifolia, 39: 95
— portoricensis, 39: 208

shea SHE
Dimo orhocadion,. are 111
— formosum, 39: 114

Emblica, 38: 76

t. Eriococcus, 38: 360
Bee Euepistylium, 39: 153
159

Euphorbia ludoviciana, 37:
Execarpus epiphyllanthus, ae ae

Flueggea virosa, 39: 208

Geminaria, 37: 346
— obovata, 37: 348

Genesiphylla, 39: 179
— speciosa, 39: 187

Glochidion, 38:
— sect. ieniphlants 39: 163
— botryanthum
— ovatum, 39: es

ED 39: 179
GME 39: 193

Ki ees 38:

ape Anisonemops ie 38: 56
— sect. ee 38: 56

Lomanthes, 39: 179
= pray 39: 185

Margaritaria bahamensis, A 208
— ; 39: 208

Myrobalanus emblica, 38: 77
Nirurt, 37: 343

Omphalea eit 39: 159
— cauliflora 157
istylinm, 88: 159
Orbicularia, 39:

— Phyllanthoides, oe 118
— scopulorum, 39: 135

Phyllantheae, 37: 340
Phyllanthus, sect. Anisonema, 38: 56
71

— sect. Catast ; 1
— sect ae 38: 52
eee Cicca, 3§

—— eed Eucicca, 38: 65

— sect. Dimorphocladium, 39: 111

210 JOURNAL OF THE ARNOLD ARBORETUM _ [voL, XxxIx

oe sect. Elutanthos, 39: 50 Phyllanthus, bahamensis, 39: 208
. Emblica, 3 5 — baracoensis, 34
— ne Epistylium, 39: 153 — barbadensis, 39: 53
— sect. Eriococcodes, 38: 359 — berteroanus, 38: 190
— sect. Eriococcus, 38: 360 — botryanthus, 39: 51
— sect. Euphyllanthus, a 295 —brachyphyllus, 38: 331
— sect. Floribundi, 38: — brasiliensis, 38: 365
— sect. Flueggeopsis, 38: e oblongifolius, 38: 368
— sect. Glyptothamnus, 39: 68, 160 — breviramis, 39: 139
— sect. Hemiphyllanthus, 39: 163 — brevistipulus, 38: 62
— sect. Kirganelia, 38: 51 — brittonii, 39: 95
— sect. Loxopodium, 38: 171 — buchii, 38: 333
— sect. Menarda, 38: 52 — cantoniensis, 38: 194
— sect. Nornocleaw, 38: 363 — cardio phyllus, 39: 125
— sect. Omphacodes, 39: 142 — caribaeus,
— sect. Orbicularia, 39: 111 — carnosulus, 38: 172
—sect. Paraphyllanthus, 38: 76 ——caroliniensis, 37: 347
ct. Phyllanthus, 38: 295 t us, 37: 34
— — subsect. Niruri, 38: 299 —— caroliniensis, 37: 348
— — subsect. Pentaphylli, 38: 324 —— guian :
—— su Swartziani, 38: 3 —— saxicola, 3 50
— sect. Scepasma, 38: 359 — — stenopterus, 37: 348
ect. Thamnocharis, 39: 91 — cauliflorus, 39: 157
—sect. Typophyllanthus —ceramanthus, 37: 233
subs Cheramela, 38: — chacoensis, 38: 72
— —subsect. Genesiphylla, 39: 179 — chamaecristoides, 39: 132
—— subsect. Kirganelia, 5 ——ch maccristoides, 39: 133
—sect. Urinaria, 38: 1 — baracoensis, 2
—sect. Williamia, 39: 69 — chamae peuce, 38:
— — subsect. Tisenlores, 39: 71 — chryseus, 39: 161
— — subsect. Incrustati, 39: 82 — cicca, 38: 66
— — subsect. Mirifici, 39: 89 — cinct
— sect. Williamiandra, 39: 82 — cladanthus, 39: 1
—sect. Xylophylla, 39: 179 — coelophyllus, 39: 134
— subg. Botryanthus, 37: 345; 39: 49 — cognatus, 39: 1
— subg. Cicca, 37 4; 38: 60 — comosus, 16
— subg. Conami, 37: 345; 38: 363 —compt 93
— subg. Eriococcus, 38: 359 — conami, 38: 368, 368
— subg. Kirganelia, 37: 344; 38: 51 — corcovadensis, 38: 52
— subg. Phyllanthus, 38: 170 — coxianus,
— subg. Xylophylla, 38: 61 —cristalensis, 39: 82
— abditus, 3 — cuneifolius, 39: 208
—acacioides, 39: 177 —cyclanthera, 38: 180
— acidissimus, 38: 66 — — gracillimus, 38: 183
— acidus — — lindenianus, 38: 183
—acuminatus, 38: 364 — — scabrellus, 38: 183
— alatus, — debilis, 3
—amarus, 38: 313 — decander, 39: 71
—amunicola, 38: 336 — diffsus, 38: 315
— anderssonii, 39: 63 — genuinus,
—angustifolius, 39: 193 0 gifolius, 38: 318
— — genuinus, 39: 193 — dimorphus, 38: 341
— anisophyllus, 38: 190 — dingleri, 39: 189
— antillanus, 39: 208 — discolor, 39
— apiculatus, 39: 133 — pallidus, 39: 75
— aquaticus, 38: 315 — distichus, 38
— arbuscula, 39: 187 — — nodiflorus, 38: 66

— axillaris, 39: 159 — echinospermus, 38: 353

1958] WEBSTER, WEST INDIAN PHYLLANTHUS 211

Phyllanthus, ekmanii, 39: 97 Phyllanthus, maleolens, 39: 165
— elsiae, ga : 171
— emblica — melanodiscus, 39: 125
— Sante 39: 197 — micranthus, 3 2
dilatatus, 3 03 — — fuertesii, 38: 328
— — domingensis, 39: 202 icrodictyus, 39: 80
—w— epiphyllanthus, 39: 198 — mimicus, 38:
—— genuinus, 39: 199 Mae 174
— epistylium, 39: 159 —— Se 39: 171
— erythrinus, 39: 125 ee 38
— estrellensis, 39: 139 — minor, 3 3
— euwensii, 39: $1 — mirificus, 39: 89
— excisus, 5 —montanus, 39: 183
— fadyenii, 38: 323 — muelleranus, 38: 5
— falcatus, 39: 199 —myriophyllus, 39: 168
— formosus, 39: 114 ge NS 39: 12
— foveolatus, 39: 125 — — alainii,
— fraternus, 38: 309 — — erythrinus, 3
fuertesii, 38: 328 — — myrtilloides, 39: 126
— glabellus, 39: 208 — — shaferi, 39:
— gracilissimus, 38: 183 ae ee 39: 131
— grandifolius, 39: 146 — nanus, 38:
cornifolius, 39: 151 = neopeltandrus, 39: 208
— — genuinus, 39: 1 — niruri,
— — salzmanni, 39: 151 — — debilis, 38: 307
— haplocladus, 38: 172 — — genuinus, 38: 313, 315
— ee 38: 171 — — javanicus, 3 7
hoffmannseggii, 38: 315 ——tathyroides, 3 38: 302
— eon eanie 38: 66; 39: 208 — — niruri, 3
— hyssopifolioides, 38: 171 — — radicans, 38 a
—imbricatus, 8 — — scabrellus, 38: 309
— inaequaliflorus, 39: 189 — nobilis, 38: 66; 39: 208
— inaequifolius, 38: 187 — norlindii,
— incrustatus, 39: 87 = Summillariaefolius, 38: 53, 371
— isolepsis, 39: 185 —nummularioides, 39: 137
— jamaicensis, 38: 57 — nutans, 39: 56
— juglandifolius, 39: 146 ~~ erisebachianus, 39: 61
— — juglandifolius, 39: 148 —— nutan Sait)
— —cornifolius, 39: 151 een 39: 118
— junceus, 38: 342 — orbiculatus, 38: 371
—lathryroides commutatus, 38: 300 — ovatus, 39: 169
— latifolius, 39: 185 — pachystylus, 39: 62
— laurifolius, 39: 208 — ™ pallidus, 39: 75
eonardorum, 38: 186 — pentaphyllus, 38: 344
len nis, 39: 143 pentaphyllus, 38: 348
— lepidocar pus, 38: 194 — — polycladus, 38: 350
—leprocarpus, 38: 194 — phlebocarpus, 39: 139
— leptoneurus, 38: 339 — pinnatus, 38
— lindenianu : 18 — pinosius, 38: 35
— — inaequifolius, 38: 187 — poiretianus, 38: 371
— —jimenezii, 38 5 — polycladus, 38: 350
— — leonardorum, 38: 186 — — curassavicus,
— — lindenianus, 38: 183 — eloupensis, 38: 350
— linearis, 39: 187, 195 — portoricensis,
— cognatus, 39: 193 — procerus, 38
$, 39: 187 — proctoris, 39: 195
— longifolius, 38: 66 eee 3

— maestrensis, 38: 337 — subnudus, 38: 343

212 JOURNAL OF THE ARNOLD ARBORETUM

Phyllanthus, pseudocicca, 38: 62
— pubigerus,

= rotundifolns, ae 118

— sagraeanus, 3

— scandens, 38: 66: 39: 208
35

— shaferi,
— hatha tobias 39: 131
— speciosa, 39: 18

— squamatus, 38: 343
— ean ee ee

|

i
=
9
fs
(2)

— ee 38: 180
: 208

8: 194
— virens, 38: 66; 39: 208
— wightianus, 38: 52

Phyllanthus, williamioides, 39: 83

Ramsdenia, 39: 82
— excisa, 39: 85

Reidia, 38:

rae 39: 1

omosa, a 116

Savia sessiliflora, 3
Securinega reopliondr 39: 208
Staurothylax, 3
Synexemia, 37:
= ie ei . 348
— cuneifolia, 37: 349
— obovata, 37: 349
— pumila, 37: 349

Tricarium, 38: 65

Urinaria, 37: 343
a, 38: 300
— indica, 38: 194
Williamia, 39: 69
— pruinosa, 39: 7
Wurtzia tetracocca, 38: 66

X ylophylla, 39: 66, 179

— montana, 39: 183
— speciosa, 39: 187

[VOL, XXXIX

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXXIX Jury 1958 NUMBER 3

THE CITATION OF SOME GENERA
OF THE LAURACEAE !

CARROLL E. Woop, Jr.

THE COURSE of a review of the genera of Lauraceae occurring in the
southeastern United States it has been necessary to verify the citations
of the native and naturalized genera and their typification. Hardly any
uniformity in these matters exists in the literature of this difficult family,
especially in connection with groups published prior to C. G. Nees’ basic
monographic treatment in 1836. Cinnamomum, for example, is still being
cited variously as of Linnaeus, Blume, Trew or Nees & Ebermaier. Still
another author has been proposed for this genus and for Persea and Sassa-
fras in recent papers by Kostermans (1952, 1957).

It is hoped that the notes which follow will help to clarify the biblio-
graphic citation of these and other genera grouped by Linnaeus under his
all-inclusive Laurus (Gen. Pl. ed. 5. 173. 1754). Supposing that the glands
of the third series of stamens were a constant generic characteristic (rather
than one of very nearly the entire family), Linnaeus lumped under his
“A452. LAURUS.* Tournef. 367.” plants formerly treated as separate genera
(and later restored to generic rank by one or more authors in spite of the
Linnaean influence which lasted almost fifty years). Included by Lin-
naeus were “Cinnamomum Herm. H. L. B. 656. Burm. zeyl. 28: 1.
Camphora Gronov. Diss. Persea Plum. 20. Borbonia Plum. 2. Benzoé
Boerh. Sassafras Off.” In the Linnaean interpretation, each of these
generic names was used for that of a species under Laurus.

Recently, in an historical review of the Lauraceae (1952) and in his
synopsis of the family (1957), both works representing an enormous
amount of time and effort and of great value toward an understanding of
the group, Kostermans has accepted G. R. Boehmer’s 1760 edition of
C. G. Ludwig’s Definitiones Generum Plantarum as the place of valid publi-
cation of Cinnamomum, Camphora, Borbonia, Persea, Benzoin, and Sassa-
fras. While it is most certainly desirable to establish the earliest possible
date for these names, in this instance the publication appears to be invalid

‘Continuing a series of miscellaneous notes and papers on the flora of the south-
eastern United States made possible through the hire and support of George R.
Cooley and a grant from the National Science Foundatio

214 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

under the International Code of Botanical Nomenclature. On page 63 of
this edition, under “188. LAURUS. Linn, Ed. V. n. 452,” following a
generic description identical with that in the edition of 1747, come the
pertinent lines: “Sexus et numerus partium variat, staminum tubercula
pro constanti generico charactere assumit Linnaeus; et ab eodem huc
refertur.’ There follow the names and good short descriptions (here
omitted) of:

LAURUS Tourn. 597.

CINNAMOMUM Burm. Zeyl. 28. carva H. Mal. T. 1. p. 107.
CAMPHORA Boerh. I]. 261.

BORBONIA Plum. G. 3.

PERSEA Plum. G. 44.

BENZOIN Boerh, IT, 259.

SASSAFRAS C. B. Pin. 431.

LAURUS INDICA Aldini licet monente Hallero Goett. p. 15.

These are unnumbered, placed beneath Laurus in the fashion of similar
notes and clearly intended synonyms throughout the book, and are in
italics in the index (the usual practice for synonyms). The descriptions
are also in italics, in contrast to the roman type of accepted genera in this
work.

The earlier edition of Ludwig (1747) includes on p. 35 under “133,
LAURUS. Linn.” the same generic description followed by the note “Huc
igitur a Linnaeo referuntur:” and the same generic names of the later
edition, with the exception of Laurus indica which was added by Boehmer
along with more complete citations.

Clearly the names appended to Laurus were not accepted by either
Ludwig or Boehmer, both of whom followed Linnaeus in construing that
genus in its very broadest sense. They cannot be attributed properly to
Boehmer for, being placed by him in the synonomy of Laurus, they must
be regarded as invalid. (Cf. Art. 33, Internat. Code Bot. Nomencl. 1956.)

Although Boehmer’s names are not acceptable, it would seem necessary
to follow Rehder (1949) and Little (1953) in adopting those of Lauraceae
from Trew in Blackwell’s Herbal (1757, 1760). This work is a nomen-
clatural hodge-podge following no particular system, but Sassafras, Cinna-
momum, and Camphora are adequately described and illustrated with clear
and unmistakable references linking them to the corresponding Linnaean
species. If the generic names now generally cited as dating from Miller,
Gard. Dict. Abr. ed. 4. 1754, are legitimate under the Code, those of Trew
must be equally acceptable. Cinnamomum, Camphora, and Sassafras of
Trew antedate by a number of years the other possible authors of those
names.

The citations which follow are not complete but include the earliest
place of publication, the second earliest in most instances, and other data.

BENZOIN Fabr. Enum. Meth. Pl. Horti Medici Helmstad. ed. 2. 401. 1761, nom.
rejic. (Type: Laurus Benzoin L. [= Lindera Benzoin (L.) Blume]). =

1958] WOOD, CITATION OF GENERA OF LAURACEAE 25

Lindera Thunb. Nova Gen. Pl. 2: 44. 1783 (Type: L. umbellata Thunb.) ;
Blume, Mus. ie Lugd.-Bat. 1: 323. 1851, nom. cons. Non Lindera Adanson,
Fam. Pl. 2: 425. 1763.

BorgsontA Miller, Gard. Dict. Abr. ed. 4. 1754 (based on a mixture including at
least Persea Miller and Nectandra Rol. ex Rottb.; cf. Kostermans, 1952);
Adanson, Fam. Pl. 2: 341. 1763 (a mixture, including Persea, Nectandra,
Ocotea Aubl.). Non Borbonia L. 1753 (Leguminosae).

CampuHora Trew, Herb. Blackwell. Cent. 4, signature L. ¢. 347. 1760 (Type:
Laurus Camphora L. | = Cinnamomum Camphora (L.) T. F. L. Nees &
Ebermaier]); C. G. Nees in iaeres P]. As. Rar. 2: 61, 72. 1831 (Type:
Camphora officinarum C. G s [| = Cinnamomum Camphoral). = Cin-
namomum Trew, sect. ee (Trew) Meissn. [attributed by Meissner
to C. G. Nees]

Cinnamomum Trew, Herb. Blackwell. Cent. 4, signature mM. ¢. 354. 1760 (Type:
Laurus Cananonun Ca ee Oe zeylanicum Blume]); Blume, Bijdr. Fl.
Nederl. Indié 568. 1826 (Type: C. zeylanicum [Garc.] Blume).

Persea Miller, Gard. Dict. Abr. ed. 4. 1754, nom. cons.” (Type: Laurus Persea
L. [ = P. americana Miller, 1768]); Miller, Gard. Dict. ed. 8. 1768 (Type:
P, americana Miller); Gaertn. f. Fruct. 3: 222. ¢. 221. 1805 (Type: P.
gratissima Gaertn. f. [ = P. americana Miller]).

Sassafras Trew, Herb. Blackwell. Cent. 3, signature p. t. 267. 1757 (Type: Laurus
Sassafras L. [ = S. albidum (Nutt.) C. G. Nees]); T. F. L. Nees & Eber-
maier, Handb. Med.-Pharm. Bot. 2: 418. 1831 (Type: S. officinale T. F. L.
Nees & Ebermaier, “officinalis.” | = S. albidum]).

REFERENCES

ene A. J. G. H. A historical survey of the eee Jour. Sci. &
Indus, Res Indonesia 1: 83-95, 113-127, 141-159.
. Lauraceae. Reinwardtia i 193-256. 1957. ae publ. as Comm.

Forest Res. Inst. Indonesia 57. )

LitTLe, E. L., Jr. Check list of oe eur naturalized trees of the United States
Gannie Alaska). U.S. Forest Serv. Agr. Handbook 41. 1953

ReupbeR, A. Bibliography of cultivated trees and shrubs. Arnold Arb., Jamaica
Plain, Mass. 1949.

* Conservation unnecessary.

= % wee Tf €q 2 ee
~ e | ae e |

Nhe
_Jlerrania

2 5 f a aor" A \
ee ( wees
=e eee ‘@ /
3 aaa IH rane 10H ‘aa ff
i : \ 2Hb 11H Maria . vA
ae \ j i. ‘ A
A on — 5 : previo 12 ance A
¥ 4 argoa 13 H nycterodendron if
U { PP) 5 ‘ ence 14H pulcherrima J
( al 6 H Dugandii 15H." © vor pacifica a
Ss { — 7 H kanukuensis i6 H. purpurea |
~Y 4 e 8 fonor 17 H. tomentella
\ oy 3 H laciniifolia 18 H.umbratica p Db

9T2

WOALAYOUUY GIONUV AHL AO TWNUNOL

XIXXX “I0A]

1958] SCHULTES, SYNOPSIS OF HERRANIA Za

A SYNOPSIS OF THE GENUS HERRANIA
RICHARD EvANS SCHULTES
With seventeen plates

La flor se podra mirar como la mayor maravilla del reino vegetal, y
apenas se pudiera creer que la economica y sencillisima naturaleza
hubiese gastado tantas cintas y atavios para engalanarla casi con la
ostentacion de las modas.

— Eloy Valenzuela (1784)

MorE THAN AN ACADEMIC INTEREST moves me to present this study of
the sterculiaceous genus Herrania, a close relative of Theobroma, the genus
of the cultivated cacaos. In recent years, the chocolate industry of the
world has become increasingly alarmed over the ravages of certain fungal
and virus diseases of the cultivated forms of Theobroma. These are so
serious that the industry has waned almost to extinction in several parts
of the world, and our sources of certain grades of chocolate are seriously
threatened. Extensive investigations toward a control of the disease have
been carried out in a number of scientific institutions; much has been done
in plantations toward selection of resistant strains, and several attempts
to do this from wild material have been made; furthermore, attention has
been devoted toward the possibility of overcoming the difficulties through
hybridization and other techniques.

It is my belief, however, that sooner or later an extensive program for
selection of new germ plasm from the jungle must be initiated, if the
chocolate industry is to be saved. Naturally, such a project would include
all near allies of Theobroma (Schultes, R. E. ‘El género Herrania, pariente
silvestre del cacao cultivado.” Agric. Trop. 7: 43-48. 1951; “Le genre
Herrania, parent sylvestre du cacaoyer cultivé.” L’Agron. Trop. 6: 661—
663. 1951). In 1952-53, the Imperial College of Tropical Agriculture
in Trinidad and the Colombian Ministerio de Agricultura carried out
jointly a very thorough study of Theobroma and Herrania in the Ama-
zonian sector of Colombia. This study, which was headed by the late
Prof. Richard E. D. Baker and Dr. Francis W. Cope, of the Imperial
College, and which included half a dozen additional British and Colom-
bian scientists, explored many of the major rivers of the region and made
selections of interesting or outstanding material for the cacao-improvement
scheme in Trinidad.

Herrania first attracted my attention in 1941, when I collected a then
undescribed species in southern Colombia. The process of determining
and describing this collection brought out the great need for a revisionary
treatment of the genus. Nevertheless, in view of the sparse and frag-
mentary collections then in our herbaria, as well as the unavailability
during the war of material from European institutions, it seemed advis-
able to await a more favorable time before undertaking such a revision.

218 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. XxxIx

From 1941 to 1953, I was engaged in almost uninterrupted field studies
(chiefly of the genus Hevea) in the Amazon Valley. In connection with
this work, it was often possible for me to give special attention to Herrania
in its native habitat and to make collections over a wide area. This intimate
association with the living plants enabled me to study a number of
morphological and ecological characters which are lost in the preparation
of herbarium material. It has also been possible, in some cases, to study
and collect fruits which, although often wanting in our herbaria, can be
highly significant from the taxonomic point of view.

In addition to field studies, I have consulted all available specimens
from the principal herbaria in the United States and South America in
the preparation of the following synopsis. Furthermore, in 1947 and again
in 1950, I was able to study the collections preserved in England and on
the continent of Europe.

Although it would seem that sufficient material upon which to base an
admittedly preliminary treatment of Herrania has now been consulted, we
must realize that certain concepts are still rather poorly understood. Fur-
ther field studies will almost certainly lead to alterations in our present
treatment and will undoubtedly enhance our knowledge of the composition
and geographic distribution of the genus.

ACKNOWLEDGEMENTS

So many friends have aided me in so many ways in both the herbarium
and the field that I find it difficult to thank them all individually. The late
Professor Oakes Ames, of Harvard University, followed my field and
herbarium investigations of Herrania from their beginning and, by letter
and conference, stimulated me with his many suggestions and with his
deep interest in economic botany. It is with warm appreciation that I
thank Professor Paul C. Mangelsdorf, director of the Botanical Museum of
Harvard University, for his constant encouragement in the work of pre-
paring this study of Herrania. My colleagues at the Botanical Museum,
especially Dr. Albert F. Hill and Mr. Charles Schweinfurth, have likewise
given generously of their time, advice and encouragement throughout the
years of preparation of this study. My deep gratitude must go to Dr. E. W.
Brandes and Dr. Robert D. Rands, who directed my work on Hevea rubber
for the United States Department of Agriculture, for their encourage-
ment and understanding of my botanical activities which, like the Herrania
investigation, were not directly concerned with my regular studies in latex-
bearing plants. Dr. N. Y. Sandwith of the Royal Botanic Gardens, Kew,
has been most generous with his help during the last few years of this
study. To Dr. Armando Dugand, former director of the Instituto de
Ciéncias Naturales in Bogota, and Dr. Felisberto Camargo, former director
of the Instituto Agrondémico do Norte in Belém do Para, Brazil, I am espe-
cially indebted for their interest and for many kindnesses during my years
of field work. Mr. Gordon W. Dillon, the late Senorita Inés de Zulueta
and Miss Helen Schieffer deserve my warmest thanks for their drawings
which were published in some of my earlier taxonomic papers on Herrania.

1958] SCHULTES, SYNOPSIS OF HERRANIA 219

It is difficult for me to express adequately my gratitude to Mr. Elmer W.
Smith for the highly accurate and artistic drawings which illustrate this
synopsis. Sincere thanks go also to the American Cocoa Research Institute
and ihe National Science Foundation for generous grants of money towards
the illustrating of my taxonomic studies in this genus of close relatives
of the chocolate plant.

My deepest appreciation is due to the numerous members of the Anglo-
Colombian Cacao Collecting Expedition, especially to the late Prof.
Richard E. D. Baker and to Dr. Francis W. Cope, together with which
Expedition I was able to spend many months in the field and to test the
workability and completeness of this synopsis of Herrania,

To the directors and staffs of the following American botanical institu-
tions I send my thanks for the loan of specimens or for their kind per-
mission to consult material in their care; the Gray Herbarium, the Arnold
Arboretum, and the Economic Herbarium of Oakes Ames (Harvard Uni-
versity); the Bailey Hortorium (Cornell University); the New York
Botanical Garden; the Missouri Botanical Garden (St. Louis, Mo.); the
Chicago Natural History Museum; the United States National Herbarium
(Washington, D. C.); the United States National Arboretum [Herbarium |
(Beltsville, Md.); the Philadelphia Academy of Natural Sciences; and
the Yale School of Forestry (New Haven, Conn.). I am deeply grateful
also to the officers of the following European and South American botanical
centers for their many courtesies during my periods of study in their
herbaria; the Royal Botanic Gardens (Kew) and the British Museum
(Natural History) (London); the Botanisch Museum en Herbarium
(Utrecht); the Rijksherbarium (Leiden); the Universitats Botaniska
Museum (Copenhagen); the Museum National d’Histoire Naturelle
(Paris); the Jardin Botanico (Madrid); the Conservatoire Botanique
(Geneva); the Herbier Boissier (Geneva); Jardin de )’Etat (Brussels) ;
the Botanische Staatssammlung (Munich); the Naturhistoriska Riks-
museum (Stockholm); the Instituto de Ciéncias Naturales (Bogota); the
Facultad Nacional de Agronomia (Medellin); the Instituto Agrondmico
do Norte (Belém) ; the Museu Goeldi (Belém); the Jardim Botanico (Rio
de Janeiro); the Instituto Miguel Lillo (Tucuman); and the Instituto
Interamericano de Ciéncias Agricolas (Turrialba, Costa Rica).

HISTORICAL NOTES

Herrania was described by Justin Goudot in 1844 and was named in
honor of General Pedro Alcantara Herran (1800-1872), president of
New Granada (Colombia) from 1841 to 1845. In spite of civil unrest
in the country, Herran showed much interest in the botanical investigations
of Colombia during this period. It is therefore quite appropriate that this
genus, which has its center of diversification in Colombia, should bear his
name.

‘It is equally appropriate for me to express, in this paper, summarizing our present
knowledge of Herrania, my gratitude to Dr. Alvaro Herran Medina, his great grand-

220 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

The specimens upon which this genus and its first two species, 7 errania
albiflora and H. pulcherrima, were based had been brought to the atten-
tion of the scientific world long before the actual publication of the genus
in 1844. At the meeting of the Linnean Society of London on January 15,
1828, Goudot referred to “a plant of the genus nearly akin to Theobroma
from which it differs chiefly in habit, in the form of the calyx, and the
structure of the stamens . . . ‘Arbuscula foliis digitatis, quinatis’”’ (Phi-
losoph. Mag. 3: 132. 1828). Before presenting his diagnosis of Herrania,
Goudot devoted much careful study to the new genus, for he wrote “ [since
1828] . . . I have had frequent occasions to study these plants and to
assure myself that they ought to constitute a distinct genus” (Ann. Sci.
Nat. Paris III. 2: 229. 1844). Goudot’s original diagnosis of Herrania,
reproduced below, is so detailed and so carefully prepared, with such a
clearly outlined concept as to relationship, that now there is no need to
alter either the description or the concept, and this in spite of the fact
that the number of species known has been measurably increased.

Long before Herrania was actually described, however, several species
were recognized as distinct from Theobroma. The Botanical Expedition
of Mutis in New Granada (Exped. Bot. Mutisii Novae-Granat.), which
was carried out in Colombia by a capable and devoted group of scientists
under the leadership of Padre José Celestino Mutis from 1783 to 1808,
collected some 4055 specimens, including several species of Herrania.
The greatest heritage of the expedition is the collection of water-color
paintings of Colombian plants, numbering 6900 plates and representing
2800 species. The plates, in perfect condition, are preserved in the Jardin
Botanico in Madrid and their publication has only recently been initiated.
I was fortunate in 1950 to have an opportunity to study many of these
plates, but permission to publish any of them could not be granted.

Amongst the 6900 plates, there are several extraordinarily artistic and
accurate illustrations representing three species of Herrania. These plates,
labelled simply “Theobroma” by the artists, are all included in Volume
No. 28 under number 5333. They were first determined as representing
several concepts of Herrania by Triana, who annotated each plate in
pencil and signed his annotations. These plates were made about half a
century before the genus Herrania and the three species so beautifully
represented by the Mutis plates were described by Goudot

Recently, a most significant discovery related to the work of the Mutis
Expedition was made. The Colombian historian, Dr. Guillermo Hernandez
de Alba uncovered, in 1950, a beautifully preserved manuscript prepared
by Padre Eloy Valenzuela, one of the most active members of the Expe-
dition. The manuscript is without title, but Dr. Enrique Pérez-Arbelaez
entitled it “Diario de la Expedicién Botanica del Nuevo Reino de Granada

nephew, and my good friend. Dr. Herran Medina, who served for a number of years
as Colombian Consul General in Manaos, Brazil, followed the progress of my work
in the Amazonian area of both Colombia and Brazil with enthusiastic interest and
performed innumerable services and favors without which that work would have been
much more difficult.

1958 | SCHULTES, SYNOPSIS OF HERRANIA 221

dirigida por Don José Celestino Mutis y llevado por Eloy Valenzuela
desde el dia 29 de abril de 1783 al dia 8 de mayo de 1784” (Pérez-
Arbelaez, E., Un hallazgo historico y cientifico: el diario de D. Eloy
Valenzuela. El Tiempo (Bogota), February 11, 1951). Thanks to Dr.
Pérez-Arbelaez, I may quote in full this extraordinarily detailed descrip-
tion of a species of Herrania which Valenzuela wrote in his diary and
publish a photograph of one of the pages in Valenzuela’s own hand.

This description, referring without doubt to Herrania pulcherrima, is
the earliest reference to Herrania of which we have any knowledge. In it,
Valenzuela clearly sets forth his recognition of the plant as different from
the cacao comun (Theobroma Cacao). Goudot’s validly published diag-
nosis and description of Herrania is, for all practical taxonomic purposes,
the earliest; but, for historical reasons, I am publishing herewith a tran-
script of Valenzuela’s diary-notes on his cacao esquinado. There is every
reason to believe that the plant to which these notes refer grew in the
vicinity of Mariquita, where the Expedition carried out a great part of
its labor.

Dia 13 de marzo de 1784. Sin concluir el Guacimo real, se trabaja hoy en el
cogollo de la theobroma de cintillas 0 Cacao esquinado. El otro dia repeti mi
visita a la mata y el duefio me dijo que los cuatro tallos que tiene aproximados
corresponden a una raiz sola. Los tallos seran hacia el pie tan gruesos como el
dedo pulgar sin ramo alguno tupidisimos juntamente con las hojas de pelo
chico, apelmazado y amonado; por dentro blanco, fibroso, ligeros; la parte
lefiosa es radiada con lineas que tiran para el centro; el meollo gruesesito,
blanquisimo, esponjoso, regado de puntos lucientes y cercado con algunas fibras
huecas. Una de estas partes despide algo de un licor diafano y viscido. Hojas
son alternas bastante inmediatas, patentes, decusadas. — Estipulas dos lateri-
folias algo retiradas opuestas adpressas, subuladas pollicares y de la misma
pubescencia que el tallo y hojas.— Flor: agregadas axilares y supraaxilares,
peculiares de la parte inferior; pedicelos sencillos, recurvos o levantados para
arriba.

Dia 14. Domingo.

Dia 15. Se trabaja en dibujar el tallo del cacao esquinado o theobroma de
cintillas en la parte que carga las flores y como vinieron muchas de éstas, he
tenido lugar de hacer el apunte casi por en

Cacao Esquinado. Raiz. Tallo: De la ets humana, derecho, rollizo, sen-
cillisimo, tupidisimo de pelo chico, grueso al tanto del dedo pulgar: y parece

es algo desnudo y color de café y aqui mismo tiene la epidermis menudamente
hendido-reticulada y algo levantada. — La corteza es gruesesita, verdosa debajo
de la epidermis y flexibilisima, el interior blanco, fibroso, ligero y en el corte
transversal se ve radiado con rayas que tiran al meollo desde la misma corteza:
este es blanquisimo bofo y cercado de orificios capilares por donde mana un
jugo muy cristalino y gelatinoso.

Hojas: Grandes, alternas, tupidas, de pelo chico, 1-partidas hasta el mismo
pezon comun. — Pezén: comun: patentes, pedales, multisulcados a lo largo y
geniculados en los extremos.— Hojuelas: cuneiformes en la mitad inferior,
escabradas por encima, membranaceas, y de bordes enteros, regados de puntas
rigidiusculas; la mayor excede al pezén comun: las anteriores chicas obovato-

222 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

oblongas, obtusas con punta: las exteriores se ensanchan desde la mitad y abren
en lacinias pinnadas, anchas, acuminadas, de las que son mas grandes las dos
i. : :

chica y ro antes de las primeras.— Venas: paralelas, rectas, distantes.
ee : laterifolias, algo ce og de la axila, subuladas adpressas, ae
y conformes a la extension del tall
Flor: amontonadas, tupidas, a res y supraaxilares en la parte inferior del
tallo y casi siempre en las ax xilas desnu das, pollicares ligeramente olorosas y de
i ingunos. — Pedicelo

sencillos, levantados, delgadisimos por el pie y con dos o tres subulag cortas. —

aliz: periantio mondofilo, coriaceo, colorido, paludo exteriormente; antes de
abrir es cerrado, ovado, obtus no se divide hasta cerca de la base en 3 partes
concavas, enteramente Fen aaa la una aguda; las - anchas y
agudamente semibifidas; parece que por su naturaleza es semiquinquéfido agudo
y que al reflectar se hiende algo mas, y no se apartan algunas veces los tres

s

Nectarios: de 5 piezas, subrotundas, chicas, convexas exteriormente, embros-
cadas hacia de plegadas y venulosas longitudinalmente y puestas alrededor
del receptaculo: sus apéndices: son otras tantas cintillas angostas lineares, con
6 are corridas a ‘fs largo de dos o mas venitas de color mas subido. Antes
de abrir estan envueltas en piezas paradas sobre los pétalos y el color tira al
amarillo.

: de 5 piezas alternas con las del nectario y mas interiores, lanceoladas,
Poe patentes de una longitud con el periantio, surcadas por encima a
largo y cuadunadas ligeramente en la base.

Estambres: de 10 filamentos nacidos del lado de los pétalos, cortados, gruesos,
recurvados para afuera, apareados y aproximados por los apices; anteras lineares,
univalves, rayadas a lo largo, convexas, adnatas, a otras tantas lacinulas de los
a blancas y escondidas en el seno de los nectarios; 4 estan en un fila-

mento y 2 en el lateral. ae forman como un angulo y las primeras forman
ee ae el uno en el o

rmen: superior, pentagono, pyramidal, hirsuto, chico, de color herbaceo. —
Estilo unico linear, subduplo de la altura de los pétalos. — Estigma un poco
obtuso y parece sencillo.

Fruto: bayas sentadas sobre la flor marchita, 0 bien mazorca ovada acuminada,
tripolicar, 5 angular verde lustrosa y regada de tomento caedizo; el acumen es
entero algo oblicuo; los angulos levantadisimos, anchos, enteros, sdlidos, u-
losos lateralmente y alternan con 5 venas gruesas, algo anguladas, ecu
longitudinalmente por la depresién de ellos; la corteza blanda quebradiza, sub-
Bluniosa:© en la partidura, blanca y destituida de angulos por dentro. Meollo
ovado, 5 surcado, libre de la corteza y prendido por la base. —Granos: 70
convexo- eats rugoso-reticulados, coloridos de rosado blanquisimo, arilados
con un saco fibroso empapado en jugo agridulce, y aproximados como en 5 hileras
con alguna pulpa intermedia. — Receptaculo comun, lehoso y central; a
filiformes.

El interior ;

El cérculus asoma la cabecilla en la superficie interior . . . Cotiledones sin
las priser irae es del cacao comun

Hojas seminales:

Tengo puestos algunos granos de tierra himeda para ver si nacen y con esto
determinar el modo con que se explica su interior y el tiempo que se necesita.

1958] SCHULTES, SYNOPSIS OF HERRANIA 223

Nace con mucha frecuencia en el monte y rozas de esta inmediacién: el grano
amargo y sumamente mantecoso, dicen ser apetecido y consumido por los micos.

La flor se podra mirar como la mayor maravilla del reino vegetal y apenas
se pudiera creer que la econdmica y sencillisima naturaleza hubiese gastado tantas
cintas y atavios para engalanarla casi con la ostentacién de las modas.

El periantio por fuera es de color acanelado, sucio salpicado de carmin, casi
borrado. Por dentro es lustroso y de carmin hermoso.

Nectario: en el fondo es blanquecino o descolorido; las venitas y pliegues
son de purpura oscura. Las cintillas son carmesfes con venas oscuras.

Pétalos de purpura afinadisima sub-oscura

Estambres: en los filamentos son carmesies y en las anteras blancos.

Geéermen: palido verdoso. Estilo Baas estigma blanco.

Mazorca verde: por dentro sub-rés

Dia 16. Se trabaja hoy en la hoja del cacao esquinado.

ECOLOGICAL PREFERENCES OF SPECIES OF HERRANIA

Herrania is distributed from sea level to about four thousand feet, but
the altitudinal range of each species is often strictly limited. Herrania
purpurea has been collected at sea level, and most of the specimens of this
species are from localities below two hundred feet. Herrania Mariae
occurs on the great flat Amazon planada from sea level at the mouth of
the Amazon River to approximately three hundred and fifty feet in the
western parts of the Amazon Valley. Herrania nitida and H. tomentella
can be found in the western Amazon and Orinoco drainage-areas, from
about two hundred and fifty feet above sea level up to the base of the
Andes at about nineteen hundred feet. Herrania albiflora inhabits the
slopes of the Andes between about three hundred and three thousand feet,
thus exhibiting a more surprising altitudinal tolerance. The only known
collection of H. umbratica was taken at about twenty-one hundred feet.
Although the type collection of H. pulcherrima was made on the eastern
slope of the Colombian Andes at about fifteen hundred feet altitude, most
of the other collections are from the central Andean regions at about three
thousand feet, with a distinct variety near sea level in western Colombia.
Apparently rather intolerant of altitudinal variation is Herrania laciniifolia,
which is known only from regions between three and four thousand feet.
One of the species which seems to be confined to a narrow altitudinal range
is H. breviligulata known from two collections from the eastern slopes of
the Andes between eighteen and twenty-five hundred feet. The occurrence
of H. Camargoana merits special mention, for it is unknown except in the
upper Rio Negro basin of Brazil and Colombia, where it is found near or on
the summits of barren granitic mountains rising to an altitude of about five
hundred feet above the level of the river (or approximately twelve to
seventeen hundred feet above sea level) and, rarely, in sandy riverside
savannahs. Herrania Cuatrecasana, H. Dugandii, H. kofanorum and H.
nycterodendron appear to be confined to the dense forests of the western-
most parts of the Amazon Valley at an altitude of from three hundred to
one thousand feet. Herrania lemniscata is found between four hundred

224 JOURNAL OF THE ARNOLD ARBORETUM _ [voL, xxx1x

and three thousand feet and H. kanukuensis between four hundred and
one thousand two hundred feet.

The species of Herrania do not seem to be exacting in their soil require-
ments, with the single exception of H. Camargoana which prefers and,
actually, is found exclusively on sterile white sand and on the most in-
hospitable of rocky slopes strewn with granite blocks and often nearly
devoid of soil. Most species thrive on well-drained slopes, but I have seen
Herrania purpurea in the northwestern part of Colombia along the banks
of rivers where, in the rainy season, the plants stand in from four to six
feet of water. Similarly, Herrania Mariae often grows in Amazon forest
which is subject to deep flooding half the year. Herrania lemniscata is also
reported to grow in ‘‘mixed forest in swamp and on rather swampy soil”
or “en selva anegada en la estacion lluviosa.” The widely distributed
Herrania nitida, in the Amazon Valley, is found usually just above the
reach of the deep annual inundation, although occasionally it occurs where
a few feet of water may stand for several weeks; it cannot, however, be
classified as a flood-tolerant species.

There is one point which stands out in relief concerning the ecology of
the various species of Herrania: both in altitudinal tolerance and in topo-
graphical preference there is a tremendously wide range within the genus.
Some species seem to be extremely exacting, whereas others show an un-
believable tolerance in both altitudinal and topographical distribution.

Although all species of Herrania are primarily forest trees, a number
act like weeds and propagate rapidly in agricultural clearings. The helio-
philic tendency is especially marked in Herrania nitida. In the Amazon
area, especially in Indian settlements, trees of Herrania are never sacrificed
in the clearing of a field for planting or in the weeding of a cultivated patch.
Since most cultivation is attempted on land above the reach of the annual
flood, the species most commonly met with under conditions of partial
human care is Herrania nitida, The only reason for not cutting down
Herrania trees is, apparently, the esteem in which the natives hold the
acidulous pulp surrounding the seed. Under conditions of ample room and
sunlight, many more fruits mature than in the forest, and man is thereby
repaid for his little consideration in sparing the tree. This toleration of
the tree in planted fields is undoubtedly the basis of occasional reports
that Herrania is sometimes cultivated. I have never seen Herrania under
conditions of cultivation, nor have I been able to find reliable reports in
the literature or on herbarium specimens to this effect.

USES OF SPECIES OF HERRANIA

As has been pointed out above, the principal importance of Herrania
lies in its potentialities as a source of new germ-plasm in a possible hybrid-
ization program with the cultivated species of Theobroma.

In addition to this potential use, a few species find employment among
native peoples. Everywhere Herrania trees are left standing in cultivated
ground, because the white, acidulous pulp in which the seeds are embedded

1958] SCHULTES, SYNOPSIS OF HERRANIA 225

is eaten. Apparently this pulp is most delicious just before complete ripen-
ing of the fruit, and for this reason, it is often difficult to find a mature
capsule. It is said that monkeys also search for the fruit as a food. Among
the Ingano Indians of Mocoa, in the Putumayo of Colombia, the ashes
of the bark of Herrania breviligulata are employed to dry up and ‘‘cure”’
infected wounds and ulcers (Schultes & Smith 2050). In British Guiana,
according to notes accompanying Archer 2514, Herrania lemniscata is
utilized in the preparation of a “‘beverage like chocolate.” Similarly,
Pittier reports (Plantas Usuales de Costa Rica. 72. 1908) that the Bribri
Indians of Costa Rica employ the seeds of Herrania purpurea to make a
bitter drink. It has been reported (Van Hall, Cacao. ed. 2. 74. 1932) that
formerly, in northern Colombia, the seeds of Herrania albiflora were pur-
posely mixed with those of commercial cacao to improve the flavor of the
chocolate and that these same seeds were often used to prepare a bitter
febrifuge. Herrania Mariae seeds were formerly found as an adulterant
in “Para cacao” (Van Hall, loc. cit.).

RELATIONSHIPS AND TAXONOMY

Herrania appears taxonomically to be intermediate between Guazuma
and Theobroma. It approaches Guazuma in the placement of its anthers
but has an entirely different habit and fruit. It can be separated from
Theobroma immediately by its habit (having compound-digitate leaves
and, in general, comprising small, delicate and slender treelets). It further
differs from Theobroma in the placement of its stamens, in the number
of divisions of the calyx, in the length and form of its ligules, in its wood
anatomy and in the structure of the pollen grains. The fruit of Herrania
resembles, in outward form, that of Theobroma, but the seeds of Herrania
have thick cotyledons which are almost entire and which are apparently
not folded. The field botanist has perhaps a better opportunity of noting
the differences between Theobroma and Herrania, for, in addition to the
technical characters listed above, he is able to appreciate the great differ-
ence in habit and to observe that the jorquetting habit of branching,
so characteristic of Theobroma, is not found in Herranza.

Although it would appear, on the basis of a taxonomic and morphologic
examination of recent collections and field studies, as well as on experi-
mental evidence, that Herrania may very justifiably be considered as a
genus distinct from Theobroma, it has, in the past, usually been relegated
to the position of section or subgenus under Theobroma. Schumann (in
Martius, FI. Brasil. 12(3) : 70-72. 1886), for example, considered Goudot’s
Herrania to constitute merely a section of Theobroma, which genus he
divided into the two sections: Herrania and Eutheobroma. Bernoulli
(Uebersicht der bis jetzt bekannten Arten von Theobroma. 4. 1871), on
the contrary, had divided Theobroma into four sections but kept Herrania
as a distinct genus, stating, “. . . Theobroma ab Herrania Goudot genere
proxime affini differt praecipue habitu, foliisque integris nec digitatim
quinque- sexfoliolatis.”

226 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

The general tendency in recent years has been to keep Herrania and
Theobroma as distinct genera. Cook (Contr. U.S. Nat. Herb. 17: 616.
1916), for example, followed this course. Cheesman (The economic
botany of cacao. Suppl. Trop. Agric. 1. 1932) pointed to Herrania and
Guazuma as the nearest relatives of Theobroma. Chevalier (Révision du
genre Theobroma d’apres | Herbier du Museum National d’Histoire Natu-
relle de Paris. Rev. Bot. Appl. Agric. Trop. 26: 265. 1946) excluded
Herrania from his treatment of Theobroma. A recent morphological and
taxonomic study of Guazuma (Freytag, G. F., A revision of the genus
Guazuma. Ceiba 1: 193 ff. 1951) offers data which very convincingly
support the separation of Herrania from Theobroma.

A comparative study of the pollen of Herrania, Guazuma, Theobroma
and other allied genera, outlined in detail below, has indicated that there
are appreciable differences between the pollen of Herrania and Theobroma.

There is still not complete agreement, however, in the generic interpre-
tation of Herrania. In 1940, Ducke (Rodriguesia 4: 273. 1940) treated
Herrania as a “subgenus (or section)” of Theobroma. A chromosome
study of several species of Theobroma and Herrania carried out by Mufioz
(Estudios cromosémicos en el género Theobroma L. Unpubl. Thesis. Fac.
Inst. Interam. Ciénc. Agric. Turrialba 30-35. 1948), indicates that the
number in both concepts is 2” = 20. The chromosomes are very similar,
but they are generally more slender and more definitely contrasted in
Herrania albiflora, H. purpurea and H. pulcherrima (the only three species
studied) than in those species of Theobroma which were examined, Mufoz
feels, but is not entirely certain, that Herrania should be treated as a
section of Theobroma

Most recently, Ducke (As espécies brasileiras do género Theobroma
L. Bol. Técn. Instit. Agron. Norte 28: 3-20. 1953) has reiterated his
belief that Herrania and Theobroma should be treated as a single genus.

It is pertinent to this discussion that recently Theobroma and Herrania
have been shown experimentally to be very closely allied. Ing. Agron.
Addison, geneticist at the Instituto Agronémico do Norte in Belém do
Para, Brazil, successfully pollinated Theobroma Cacao with Herrania
Mariae, but the embryos failed to develop (Addison, G. O’N., & R. M.
Tavares, Observagoes sobre as espécies do género Theobroma que ocorrem
na AmazOnia. Bol. Técn. Instit. Agron. Norte 25: 1951),

Van Hall (Cacao. ed. 1. 1914; and ed. 2. 1932) treated Herrania as a
section of Theobroma. On the other hand, Pound (Cacao and witchbroom
disease (Marasmius perniciosus) of South America. 47 ff. 1938), while
not definitely committing himself, outlined in very great detail the char-
acteristics of Herrania and pointed out the important differences between
Herrania and Theobroma,

The following key (adapted primarily from Schumann’s key to his sec-
tions Herrania and Eutheobroma of the genus Theobroma) sets forth
the gross characters by which Herrania may be separated from Theobroma.

A. Arbores parvae et graciles. Folia quinque- ad novem-digitata. Petalorum

ligulae lineares vel filiformes, comparate longissimae, cucullum multo

1958] SCHULTES, SYNOPSIS OF HERRANIA 22

superantes, in alabastro circinnato-involutae. Calyx usualiter 3-fidus.
ee saepissimae conspicue costatus, pericarpio vulgo sear crassu-
NSN E Lt cS haus Ne TPnus bales leh hoes Nee RN ARO Ae nia.

AA. ee Folia integra. Petalorum ligulae, duplo vel triplo ean super-
antes, vel breviores, in alabrastro reflexa vel erecta. Calyx usualiter 5-fidus.
Fructus saepe obscure costatus vel sublaevis, pericarpio ut videtur saepius

EI ECOTELEC II ROMi se taint tt oe Blot suse named ee een tain eae Theobroma

Herrania Goudot, Ann. Sci. Nat. III. 2: 230. ¢. 5. 1844; Walpers,
Rep. 5: 111. 1845-46; Endl. Gen. PI. Suppl. 4: 62. 1850; Karst.
Linnaea 28: 446. 1856; Tr. et Planch. Prodr. Fl. Novo-Granat. 209.
1862; Benth. et Hook. Gen. Pl. 1: 225. 1862; Walpers, Ann. Bot.
Syst. 1: 225. 1862, 7: 430. 1868; Baill. Hist. Pl. 4: 131. 1873, Dict.
Bot. 3: 49. 1891.

Lightia Schomb. Rep. Brit. Assoc. Adv. Sci. 2: 71. 1844. Non Lightia
mb. Linnaea 20: 757. 1847.
Brotobroma Karst. et Tr. ex Tr. Nuev. jén. y esp. plant. fl. Neo-Granad. 11.

Tyrer species: Herrania albiflora Goudot.

GENERAL DISTRIBUTION: From Costa Rica down the Andes to Peru,
along the Pacific coast of Colombia and Ecuador, across Venezuela and
the Guianas and in the Amazonian basin of Brazil, Colombia, Peru and,
probably, Bolivia.

ORIGINAL DESCRIPTION: Calyx 3-partitus, coloratus, deciduus, laciniis aequalibus
concavis ; aestivatio valvata. Corolla 5-petala, hypogyna, cucullato-concava,

didymis. Stylus cylindraceus. Stigm Ss i ereticecula. obtusa. Ovarium 5-
gonum, 5-loculare sessile, disco hy aca destitutum. Ovula anatropa in singulo
loculo anguloque centrali 1-seriata, horizontalia. Fructus ovato- cea
costatus, basi et apice subacuminatus, coriaceo-lignosus, indehiscens. Semin
pulpa nidulantia, ovata, angulata, testa pergamacea venosa. Embryo Senne
crassis, hinc convexis, inde planis, radicula brevissima.

Small trees. Leaves unusually large, digitate, 4—-9-foliolate. Flowers
cauline, fasciculate. Calyx 3-5-fid; sepals more or less divided, valvate
in the bud. Corolla 5-parted; petals hypogynous, cucullate-concave, api-
cally inflexed, with a linear or filiform ligule which, before anthesis, is
circinnate-involute. Androphore 5-fid, fleshy, glabrous: sterile segments
alternate with the petals, enlarged above into erect or reflexed, ovate-
lanceolate or linear, petaloid staminodes; fertile segments 1—4 (mostly
3)- antheriferous, anthers short-stipitate, with diverging, didymous locules.
Ovary sessile, 5-locular, the lower with many anatropous ovules. Style
filiform, See Stigmas 5, rather terete, obtuse. Fruit ovate-oblong,
costate, apically usually subacumminate or acuminate, toughly coriaceous,

228 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

indehiscent. Seeds enclosed in a pulp, flattened-ovate, angulate, with a
papery testa, exalbuminous; cotyledons thick, sometimes flat, with a very
short dadicle. Chromosomes: on = 20°. On three species studied).

In his “Woods of northeastern Peru”? (Field Mus. Nat. Hist. Bot. Ser.
15: 323-324. 1936), Llewelyn Williams has published what appears to
be the only detailed description of the wood of Herrania. Williams cited
five collections which were used as a basis for this description and referred
all of the collections to “Theobroma Mariae.” J have found that they
represent Herrania nitida, Until wood samples are collected in exact
correlation with herbarium specimens from many areas and for numerous
species, an understanding of specific differences in the wood of Herrania
cannot be appreciated. For this reason, Williams’ description of Herrania
nitida must serve as a guide to the wood structure of the whole genus.

Sapwood pale pink; heartwood pinkish brown. Wood has no distinctive odor
or taste; straight- or wavy-grained, coarse-textured; light in weight and soft;
requires a sharp knife to cut smoothly across grain; perishable. Growth rings
absent or present. Parenchyma indistinct. Pores fairly small or very small;

t numerous and well-scattered; solitary or in radial multiples of 2-3. Vessel
lines fairly long, not prominent, but discernible to unaided eye. Rays coarse,
lighter-colored than background, sometimes wavy, and conspicuous on cross
section; darker than background and fairly distinct on tangential; of darker
color than adjacent elements and conspicuous on radial surface. Pith light or
dark greyish brown.

The examination of the pollen grains of Herrania, to the best of my
knowledge the first which has been made, was carried out by Dr. Theodor
van der Hammen of the Colombian Servicio Geolégico Nacional. Pollen
of Herrania tomentella was analyzed. All terms are used in accordance
with Iverson and Troels-Smith’s nomenclature proposed in 1950.

Herrania pee is R. E. Schult. Pollen collection Serv. Geol. Nac. #IV 86.
Collection Col. 3

Pollen grains: oe reticulate, subsphaeroidal; granulae of the muri
visible but not separated. Lumina of reticulum irregular of size, rather large,
polygonal, smaller near the colpae. In hoes lumina, rather faint granulae are
visible. Colpae clear, edges separated; pores clear, without ectexine elements;
sometimes indications of small ae furrows. Magnitudo pollinis: media
— Magnitudo luminum: Hse macro (2-4.7 «) and smaller. Index
pollinis: subsphaeroidea (1-1.12); (‘prolate sphaeroidal’ of Erdtman). Index
areae poli: middle (0.30-0.35). Index exinae: middle (0.05—0.08 ).

o~
bo
oo
WwW
5 =

Van der Hammen reports further that Theobroma, Guazuma, Sterculia
and Herrania have tricolporate and reticulate pollen grains but that the
grains of Waltheria and Helicteres are of different types. A comparison
of the four genera with tricolporate and reticulate grains leads to the really
unexpected conclusion that, insofar as pollen morphology is concerned,
there is no evidence that Herrania and Theobroma are very closely allied.
On the contrary, the pollen grains of Theobroma resemble those of Guazuma

1958 | SCHULTES, SYNOPSIS OF HERRANIA 229

even more than they do those of Herrania. Van der Hammen compares the
grains of Theobroma Cacao and of Herrania tomentella as follows: “The
grains of Theobroma Cacao are subsphaeroidal (‘oblate sphaeroidal’ of
Erdtman); index pollinis + 0.8-0.9; magnitudo pollinis + 22 p. The
polar area is relatively much larger than that of Herrania, the colpae are
very narrow, unclear and short. Pores are small and not very clear; lumina
of ret much smaller than those of Herrania (greatest size measured 1-1.75
»), and more regular. Exine (including sculpture) relatively thicker than
those of Herrania.”

A most careful and detailed chemical analysis of the seeds of Herrania
has been made by MacLean (MacLean, J. A. R., Oil-bearing seeds of possi-
ble economic importance to West Africa. Nature 169: 589. April 5, 1952).
The material studied was reported to be referable to Herrania balaénsts
and H. Mariae, but unfortunately no voucher specimens were cited for
checking, and the specific identifications are open to doubt. The plants
were introduced from Trinidad to West Africa in 1944. The photographs
published on pages 589 and 590 would suggest that one of the species
analyzed (probably the one called Herrania Mariae) was H. nitida.

MacLean reported a very high oil content (up to 66.1%) of the seeds,
and it is suggested that optimum cultural conditions might improve this
figure. He found that the “percentage of total alkaloids is approximately
one-third that of Amelonado beans, and the theobromine-caffeine ratio is
reversed.” The caffeine values were found to be comparable with the
lower limit of range for coffee beans and kola nuts, with values greater
than 1% being obtained on two occasions. Freshly extracted fat from
Herrania seeds is liquid at 25-29°C with an odor resembling that of lin-
seed oil; its specific gravity is 0.93-0.94; and its unsaponifiable matter
is less than 1%. The figure for reducing sugars (0.4%) is four times that
of Amelonado beans. The iodine values varied from 39 to 47, and the
saponification value ran from 203 to 206. The percentage of free fatty
acids in the oil ranged from 2.3 to 2.8 The composition of the oil was
found to be 18-26% linoleic, 2-7% oleic and 74-76% of saturated acids.

Herrania comprises two groups of species which are sufficiently well
marked one from the other to permit their segregation into sections. These
two sections are not only distinct in the morphology of the flower; they
also fall naturally into two distinct, though overlapping, geographical
areas.

Sect. Herrania

Calyx patelliformis, sepalis media pro parte connatis. Subgeneris typus:
Herrania albiflora Goudot.
Sect. Subcymbicalyx R. E. Schultes, sect. nov.

Calyx subcymbiformis, sepalis plerumque fere usque ad basim liberis.
Subgeneris typus: Herrania nitida (Poepp.) R. E. Schultes.
The characters upon which these concepts are founded serve for the

230 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

first dichotomy in the following key to the species. Herrania albiflora,
H. purpurea and H. umbratica are the only species with the curious patelli-
form calyx. Their flowers, because of this condition, have a completely
different appearance from those of all other species: the patelliform calyx
lends a “closed” or compact appearance to the flower, whereas the sub-
cymbiform condition allows for a greater expansion of the petals and
their ligules and the staminodes. Furthermore, the ligules of Herrania
albiflora, H. purpurea and H. umbratica are extremely short, whilst those
of all other species are longer — with the single exception of H. breviligu-
lata, very much longer.

The species of the section Herrania are confined to Middle America
and the northernmost regions of Colombia in South America. Those of
Subcymbicalyx extend over the whole northern half of South America, in
northern Colombia overlapping the area occupied by the three species of
Sect. Herrania.

Although, at the present state of our understanding of the genus, it is
rather difficult to form any well founded phylogenetic picture, the patelli-
form calyx would appear to me to represent a derived condition and the
subcymbiform calyx, conversely, a primitive condition in Herrania.

CLAVIS SPECIERUM HERRANIAE

A. Calyx patelliformis. Sepala medio pro parte connata. Ligulae usque ad
20 mm. longae. Se NIA.
B. Petala, ligulae et staminodia alba. Sepala usque ad 7 ee long
EA dp kratus er te eke Ba Hf. albiflora.
BB. Petala, ligulae et staminodia purpurea vel sanguinea. ane 12 mm.
longa vel longiora.
C. Ligulae usque ad 15 mm. longae. Capsulae costae valde inaequales,
aliquid tenues, distantes. Fructus usque ad 9 cm. |
nla Ree ioe eae eae areca ahs ada d 1d: a purpurea.
CC. Ligulae 19 mm. longae vel longiores. Capsulae costae quasi ina
crassissimae, non distantes. Fructus 11 cm. longus vel long
SRE dh Pee Ga. ERG A Gd eee ke dade ec anesa Lis i ae are
AA. Calyx subcymbiformis Sepala a fere usque ad basim libera.
Ligulae 25 mm. longae vel sonerore Sect. SUBCYMBICALYX
D. Foliola ye incisa vel lobat
E. Foliola grosse et regulariter pee lobulis anguste triangulari-
lanceolatis, acutissimis; lamina subtus densissime et molliter stellato-
tomentosa, usque ad 60 cm. longa. Ligulae 180 mm. longae. C apsula
maturitate fusco-rubescens: costis longitudinalibus bene conspicul
et sine costis transversalibus prominentibus. .... 9. H. laciniifolia.
EE. Foliola maxime grossissime sed irregulariter pinnatilobata; lobis late
triangularibus, acutis lamina subtus aspera, stellatis cum pilis, usque
ad 80 cm. longa. Ligulae 85 mm. longae. Capsula maturitate
ochracea (?); costis transversalibus bene akan
Be Si andre Sets ough Gey ena n aoa hacdon ee Go yee na Ok 0. H. lemniscata.
DD. Foliola integra vel leviter dentato-sinuata vel eee
F. Ligulae 25-35 mm. longae. Foliola subtus cum indumento sub-
Fe Me oun e HO anaes 4a aye vd 3. H. breviligulata.

1958 | S)
FF. Ligulae

CHULTES, SYNOPSIS OF HERRANIA 231

60 mm. longae vel longiores. Foliola Peis glabra vel cum

indumento aspero vel molle, sed numquam velu
G. Foliola glabra vel subtus sparsissime aspero- arenes

BO cle oe Tel ds hig eae gree Nema Nen ere ee mea a 2 ae 12. H. nitida.

GG. Foliola variabiliter sed molliter ee vel tomentosa.

H.

Foliola valde et regulariter sinuata.
I. Capsula costis quinque mecaeanire rectis armata et

laevis vel transverse striatofibrosa; maturitate flava
Staminodia quam 17 mm. longiora, apice acuta vel con-
spicue trifida
J. Foliola maxima, 45-60 cm. longa vel saepissime longior
19-35 cm. lata, latissime lanceolato-ovata, 12 lata
duplo longior quam latior, apicem versus grossiuscule
sinuata, dentibus 5-9 cm. distantibus. Petala et ligulae
rufo-purpurea. Ligulae 2-3 mm. latae. Staminodia
aApteer Urity asc tien was erat ete Gs 14. H. pulcherrima.
_ Foliola minor, usque ad 60 cm. longa sed saepissime
brevior < 10-15 cm. lata, anguste lanceolato-elliptica,
quadruplo longior quam latior, dentibus 2-4 (rarenter
5 ein: pice ans et ligulae albicantes. ligulae
minores quam 2 . latae. Staminodia apice acuta.
kK. Foliola usque a 30 cm. longa * 11 cm. ee
Sepala inaequalia; majora 15-16 mm. 15
minora 10 mm. X 6 mm. Petala 9 mm. < a mm.
Staminodia lanceolato-elliptica, 22 mm.
errs We ind, a eae 8. H. kofanorum.
KK. Foliola usque ad. 60 cm. longa sed vulgo brevior
x lata. Sepala subaequalia, 14 mm. Petala
8 mm. X 5 mm. Staminodia nee 18 mm.
Rea a eee eo. ee
sed

—
—

formibus longis, mollibus et crassis productis; oe
sanguinea vel rarenter flava. Staminodia usque ad 14
longa, apice valde obtusa. .......... 4. H, Cen

HH. Foliola normaliter integra, subintegra vel apicem versus plus
minusve sin

L.

ma

=

ma
7

uosa.

Foliola lanceolato- daa oe ad 60 cm. longa * 22

cm. lata, basi longe sim attenuato-decurrentia.

Petioli 45-60 cm. longi.

M. Capsulae ee tenues, late cultriformes, principaliter

costa versus pilis urticantibus armatae, tactu asperae;

capsula alibi glabra vel glabrescens.

RR re nes bere Cuatrecasana.

apsulae costae crassae, He betato: Bae sine
pilis urticantibus; capsula omnino molliter indumento
stellato-velutino armata.

N. Capsula ovoidea, apice breviter cuspidata vel sub-
rotundata, costis minoribus, primariis usque ad 2
mm. altis, secundariis haud Se
Sed Cee ek oth ge ae Sa Lane ionas:

232 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxx1x

NN. Capsula ellipsoidea, apice obtusa vel attenuata,
costis majoribus, primariis usque ad 8 mm. altis,
secundariis usque ad 3-5 mm. altis.

. Capsula 10-12 cm. longa, 4-5 cm. in diametro,
apice longe et sensim attenuata, cum costis
hebetatis crassissimis, non bene distantibus,
inter costas prominenter fibroso-rugosa. Ligu-
lae usque ad 90-100 mm. longae. Staminodia
apice leviter trifida. .... 13. H. nycterodendron.

. Capsula 9 cm. longa, 4 cm. in diametro, apice

oO

oO
©

distantibus, inter costas paullo et leviter fibroso-
rugosa. Ligulae he ad 70 mm. longae. Stami-
nodia apice acuta. ........ 16. H. tomentella.
LL. Foliola pee ieee — ‘subrhomboidea, usque ad 31
cm. longa X 12 cm. lata, basi aliquid attenuata vel quasi
cuneata. Petioli plus minusve 30 cm. longi
P. Foliola obovato-oblonga, vulgo 35 cm. longa vel minora.
Sepala aequalia vel subaequalia, 21 mm. longa * 12
mm. lata, intus subdense ferrugineo-pilosa. Petala 11—
12 im. mm. Ligulae atrosanguineae vel atro-
purpureae, 100 mm. longae, 2 mm. latae. Staminodia
ee apice acutissima, integra 12-15 mm. 5
sah fede sein eee puna ach h esd Ce aes 6. H. Dugandii.
Pr, Foliola rhomboideo-obovata, vulgo 30-35 cm. longa.
Sepala valde inaequalia; majora 12-15 mm. longa
6-10 mm. lata; minora 11-13 mm. x 11-12 mm., intus
glabra vel minutissime pilosa. Petala 7-9 mm. x 6-8
mm. Ligulae albicantes, 75-100 mm. longae, usque
ad 1 mm. latae. Staminodia lanceolato-elliptica, apice
obtusa et saepe indentato-mucronata ie es ali-
quid sinuato-undulata, 20 mm. * 6-
bi iia bees ern Gaya ae eae ees oarears o4 a 7 Mariae.

1. Herrania albiflora Goudot, Ann. Sci. Nat. Paris III. 2: 230. ¢. a;
fig. 1-10. 1844; Tr. et Planch. oo Fl. Novo-Granat. 1: 209. 1862;
Schultes, Caldasia 2: 325.

Theobroma albiflorum (Goudot) De Wildeman, PI. Trop. Grande Cult. 90.
1902.

DistRIBUTION: Northern sector of the Magdalena basin in Colombia
and in westernmost Venezuela

Small tree up to 16 feet tall consisting of several round, simple or
(rarely) branching trunks, 11-14 . In diameter, with a erevich bark.
Leaves grouped at the apex of the ers digitate, stipulate, 5—6-foliolate.
Branches densely and minutely ferruginous- tomentulous, probably becom-
ing almost glabrous. Petioles terete, densely and minutely ferruginous-
villose, somewhat dilated at the base, as long as the leaves, conspicuously
striate-sulcate, up to 45 cm. long, 4-6 mm. in diameter. Stipules conspicu-

1958] SCHULTES, SYNOPSIS OF HERRANIA 233

1c. 1. Fruits of Herrania nycterodendron, H. Mariae, and H. kanukuensis
(left to right) with schematic cross sections.

ous, linear, entire, acute, caducous, ferruginous, 50 mm. long, 3-4 mm.
wide. Leaflets lanceolate-obovate, acuminate, basally long-attenuate-
decurrent, very shortly petiolulate (petiolule strong, 3-4 mm. long), mar-
ginally entire, thin-chartaceous, 20-60 cm. (usually 50 cm.) long, 9-15
cm. wide, dark green and glabrous above, pale green and almost glabrous,
or with extremely remote and microscopic stellate hairs, beneath; the
veins of both surfaces prominent, clothed with ferruginous, minute and
weak puberulence. Inflorescence fasciculate, few-flowered (5 or 6 flowers).
Flowers subglobose, 15 mm. in diameter, borne in contracted racemes on
the lower and middle part of the trunk, white, pedicellate. Pedicels very
short, cylindric, fulvous-tomentulose, subtended at the base by a minute,
very densely fulvous-tomentulose, linear bract. Flower buds small, globose,
tomentose. Calyx patelliform. Sepals 3, subequal, connate half their length,
rotund-ovate, rounded, marginally entire, yellowish white, externally
densely stellate-villose, internally glabrous, about 7 mm. long, 6 mm. wide.
Petals 5, very broadly rotund-ovate, cucullate, glabrous near the apex,
thick-membranaceous, somewhat muricate-granulose on both surfaces, 5
mm. long, 4 mm. wide, longitudinally striate-veined, internally with con-
spicuous veins, ligulate, white. Ligules linear, reflexed, white, glabrous,

234 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

membranaceous, with 6 or 7 brownish nerves, up to 2 cm. (usually shorter)
long, less than 1 mm. wide at the base. Stamen tube 5-fid, alternately
2- and 3-antheriferous with short, simple, free filaments and 2-locular,
divergent, longitudinally dehiscent anthers. Staminodes petaloid, very
broadly ovate, acute, reflexed, 6 mm. long, 5 mm. wide, densely muricate-
granulose on both sides, areal entire. Pistil short, more or less 1.5
mm. long. Style linear, erect, glabrous with a 5- parted stigma. Ovary
sessile, subglobose, pilose, 1.3 * 1 mm. Fruits capsular, oblong, apically
short-acuminate, the tip somewhat rounded, hispid, 10-costate, yellow
when ripe, 11-14 cm. long. Seeds 30—40, irregular, compressed, enveloped
in a white, mucilaginous pulp, slightly acid, covered with an internal
membranous tegument and an internal pellicle, coriaceous and exteriorly
rugose. Embryo brown, with 2 thick, unequal, rarely folded cotyledons
and a very short radicle.

Colombia. |No definite are probably near Mariquita], Hxped. Bot.
Mutisitti Novae-Granat. 3759. ANTIoQUIA: Vuelta de Acufia, Rio Magdalena
tae only oo to H. albiflora |, Pennell 3799; opposite Boca Carare,
alt. 125 m., Pennell 3832. Bottvar: Bojorque [ Bohérquez], Rio Magdalena,
eae 1580; NorosivTiquisio Trail, Lands of Loba, alt. 150-600 m., Curran
135; Boca Verde, Rio Sinu, alt. 100-300 m., Pennell 4208. CUNDINAMARCA:
Muzo, Goudot s.n. (Type) ; Rio Guaco, near Muzo, Purdie s.n. SUR DE SANTANDER:
vicinity of Puerto Berrio, between Carare and Magdalena Rivers, alt. 100-700
m., Haught 1598. Venezuela. ZULIA: vicinity of Perija, Tejera 268. Trinidad.
Royal Botanic Gardens, Port-of-Spain, No Collector Cited; Bailey s.n.; Imperial
College of Tropical Agriculture, Diego Martin Estate, Schultes 18639.

Herrania albiflora is very closely allied to H. purpurea from which it
can be distinguished by its white or cream-colored flowers. There are also
other differences: the sepals of Herrania albiflora are usually much smaller
than those of 7. purpurea; the stipules of the former are longer than those
of the latter species; the leaflets of H. albiflora are lanceolate-obovate,
whereas those of H. purpurea tend to be obovate-oblong; and the petals
of the former species are very broadly rotund-ovate, 5 mm. & 4 mm.,
whilst those of the latter are obovate, 8 mm. * 8 mm.

The habit and general floral structure of Herrania albiflora, H. purpurea
and H. umbratica are strikingly similar. These species alone in the genus
have a patelliform calyx, which gives the flower a completely different
appearance from the usual cymbiform calyx. The ligules in these three
species are likewise similar in structure and are under 20 mm. in length.

Purdie s.n., one of the earliest collections of the species, was identified
as Herrania albiflora at Kew by Planchon, who annotated the sheet, a
topotype. The report of the collector of this specimen relative to his having
seen “several hundreds of fruits” on a cultivated tree of Herrania albiflora
must be recognized as an exaggeration. Although fruits are often very
numerous in most of the species of Herrania, I have never met any condi-
tion which would indicate ‘‘hundreds of fruits”.

It has always been presumed that the type of Herrania albiflora, the

1958] SCHULTES, SYNOPSIS OF HERRANIA 235

type-species of the genus, was preserved in the herbarium at Paris, but
in June, 1950, I found Goudot material at Geneva which may very well
be the type of Herrania albiflora. In Geneva, there are three sheets repre-
senting the Goudot collection of this species, all labelled, in his hand-
writing: “C. N. 1 Herrania albiftora mihi. Annales Sc. Nat. 1844. Muzo.”
One sheet has several very young and membranaceous leaves and an en-
velope in which there are fragments of a fruit belonging possibly to an
annonaceous plant and which, by some error, have been associated with
the Herrania collection. Another sheet has a complete and mature leaf.
The third sheet has three envelopes: one contains several seeds of Herrania
albiflora; another has a few flowers and a very young capsule; the third
has a flower completely dissected, with the parts glued flat to the envelope.
An examination of these floral parts and of Goudot’s description and draw-
ing of Herrania albiflora leads me to the conclusion that, at least for the
flowers and fruits, the Geneva material is the type of the species and genus.
As to how has it been possible for Goudot type material to find its way to
Geneva one cannot be certain. There are, of course, many Goudot collec-
tions in the Delessert Herbarium (cf. A. Laségue, Musée Botanique de M.
Benjamin Delessert. 471. 1845).

It may be of interest to note that a comparison of the Goudot floral
dissection with the description of Herrania albiflora has uncovered several
minor discrepancies or omissions. The sepals, described as glabrous within,
have a very minute and sparse puberulence on the lower portion of the
inner surface; and the petals are extremely muricate-granulose externally,
as are also the very short ligules in the basal portion near their junction
with the petal.

One of the Mutis plates in Madrid, executed by the Colombian botanical
artist Francisco Javier Matiz, represents a fruiting and flowering branch
of Herrania albiflora. Not only are ripe and unripe fruits shown in excel-
lent detail, but a large number of flowers are depicted so painstakingly that
it is clear that Mutis and his colleagues could, so long ago, differentiate
between the patelliform calyx of Herrania aibiflora and the subcymbiform
calyx of the other species illustrated. No foliage is drawn on the plate of
Herrania albiflora. Triana correctly annotated this plate as ‘“Herrania
albiflora Goudot.”

In the Mutis collection of plants in Madrid, there is a sterile collection
of leaves, misidentified as “Theobroma M ariae,” which represent Herrania
albiflora, They undoubtedly belong to the plant the flowers and fruits
of which are portrayed on the Mutis plate of Herrania albiflora.

Van Hall (Cacao. ed. 2. 74. 1932) claims that “according to Goudot,
who collected this species near Muzo (Colombia), the seeds are mixed
with those of the commercial cacao for home use; they are said to improve
the taste of the chocolate.” He credits Goudot with the report that “the
seeds are also used, unmixed, for the preparation of a very bitter product
which is used by the population as a febrifuge’.

Herrania albiflora is generally recognized as a cacao-relative. Hart
(Cacao. 13. 1911) reports that “Herrania albiflora and Pachira insignis

236 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

have both been sent to the author as ‘wild’ cacao, but neither of these
trees has anything in common with Theobroma Cacao and neither of them
produces saleable samples”.

la. Herrania albiflora Goudot f. titanica R. E. Schultes, Caldasia
3(15): 442. t. pag. 443. 1945
DIsTRIBUTION: Western or Magdalena slope of the eastern Cordillera
in the Departamento del Sur de Santander, Colombia.

Usually a robust tree up to 30 feet in height, differing from Herrania
albiflora principally in having much larger leaves and flowers.

Colombia. SUR DE SANTANDER: Carare, Landazuri alt. 1000 m., Richter s.n.
(Type); vicinity of Barranca Bermeja, Magdalena Valley, patween Sogamosa
and Colorado Rivers, alt. 100-150 m., Haught 1490.

The collection L. Richter s, n. seems to present clear evidence that there
exists a very large variant of Herrania albiflora, which may probably best
be treated as a distinct forma. The leaves are not only very large for the
species, but, in Richter s. n., the entire plant is exceptionally robust for
the genus, measuring up to 10 meters in height. The height of the plant
from which Haught 1490 came is not given, but the collector notes that
it was “a small tree;” this would seem to indicate, in the case of such a
meticulous collector as Haught, that it was not, as in nearly all the species
known, a “treelet.” The leaves of Haught 1490 are much larger than in
the Richter collection.

I have found that within a species of Herrania the size of the leaves is
more or less standard, regardless of whether it grows in shaded forest or
open pasture. This causes me to feel that the unusual size of the leaves
of Herrania albiflora {. titanica does not represent a mere ecological variant.

Vegatatively, the new form is almost indistinguishable from Herrania
umbratica of the same general region, but there are important differences
in the fruits (Caldasia 2: 261-264. tt. 2, b, c, d. 1943). It is unfortunate
that, except for the drawing which accompanies the original description,
we know very little of the fruit of Herrania albiflora. The capsule of
Herrania albiflora {. titanica is similar to that of H. purpurea, being some-
what intermediate between H. purpurea and H. umbratica. It measures
12.5-14 cm, in length and 5—5.5 cm. in diameter and has low, rounded
unequal ribs; the apex is acuminate.

Richter informed me that he collected a few dried flowers from the base
of the type of Herrania albiflora {. titanica and that they showed evidence
of having been white or yellowish in life. Unfortunately, these were lost in
shipment. The color of the flower in Haught 1490 is uncertain, for the
label bears no note in this respect. But with tees id 1598 (a collection
representing Herrania albiflora) is a note stating: “a small cauliflorous
tree, cf. 1490, from which this differs in having white flowers.”

2. Herrania balaénsis Preuss, Exped. Centr.- und Siid-Amerika (1899—

1958] SCHULTES, SYNOPSIS OF HERRANIA 237

1901). 253. #. 7. 1901; Bull. Soc. Etudes Col. Ann, 9(4): 220. ¢. pag.
221. 256, fig. 1-8. 1902 (non accurate titulata).

Theobroma balaénsis (Preuss) De Wildeman,’ Pl. Trop. Grande Cult. 89. 1902.

Small tree up to 25 feet in height, with a slender, simple, very straight
cylindrical trunk about 15 cm. in diameter at the base. Leaves arranged
at the apex of the trunk, digitate, usually 7-foliolate, petiolate. Petioles
terete, sulcate, very densely ferruginous-tomentose, up to 30 cm. long,
about 5 mm. in diameter. Leaflets unequal, lanceolate-elliptic, marginally
conspicuously sinuate, apically strongly acuminate, basally long attenuate-
decurrent, sessile, firmly chartaceous, up to 60 cm. long, 15 cm. wide; leaf
blade dark green and glabrous or subglabrous (microscopically scabridulous
with minute stellate puberulence) above, brownish green and very densely
and softly ferruginous-stellate-pilose beneath; the veins prominent and
brown-tomentose on both sides, especially beneath. Inflorescence fascicu-
late, many-flowered (12—15 flowers), on the lower and middle portions of
the trunk. Pedicels robust, densely greyish stellate-pilose, 8-9 mm. long,
1-1.5 mm. in diameter, subtended at the base by a minute, very densely
stellate-pilose, linear bract. Calyx subcymbiform. Sepals 3, subequal,
externally reddish brown, internally deep purple, lanceolate-elliptic, api-
cally subacute, marginally entire, very densely muricate-papillose or
granulose, 14 mm. long, 6 mm. wide. Petals 5, very broadly ovate, strongly
cucullate, 8 mm. long, 5 mm. wide, whitish with 8-10 purplish veins,
densely muricate-papillose, sparsely stellate-pilosiusculous externally,
papillose, glabrous internally, ligulate. Ligules of the petals white and
rose-colored, filiform, glabrous, very minutely granulose, membranaceous,
100 mm. long, less than 2 mm. wide. Staminal tube 5-parted; stamens
alternately 2- and 3-antheriferous with relatively long, simple, glabrous
filaments and 2-locular, divergent, longitudinally dehiscent anthers.
Staminodes 5, petaloid, reflexed, membranaceous, elliptic, apically acute,
marginally entire, glabrous, minutely and densely muricate-granulose on
both sides, probably purplish red, 17-18 mm. long, 7-8 mm. wide. Pistil
about 5 mm. long. Style linear, erect, glabrous with a 5-parted stigma.
Ovary sessile, subglobose, 5-ribbed, densely white-pilose, 1.8-2 mm. in
diameter, 3 mm. long. Fruit elongate-ovoid, the apex long-acuminate, 10-
costate with 5 primary and 5 secondary ribs, olive-green when young,
yellow-green and hispid when ripe, about 14 cm. long.

Ecuador. Rio de Peripa, Andre K26; El Recreo, Eggers 14362; Balao, Eggers
14362. Trinidad. CuLtivatep: Imperial College of Tropical Agriculture, Diego
Martin Estate, Schultes 18638.

Herrania balaénsis appears to be most closely allied to H. kofanorum.
This relationship will be discussed under Herrania kofanorum. This
species also resembles H. Dugandii, from which it differs in being very

? This combination, as ‘Theobroma balonsis,” has recently been made by Llano
Gomez (p. 18, Cultivo del cacao. Publ. Min. Econ. Nac. Bogota. 1947), who appar-
ently was unaware of the earlier publication.

238 JOURNAL OF THE ARNOLD ARBORETUM _[VOL. XxxIx

much larger in its vegetative parts; in having conspicuously sinuate (in-
stead of nearly entire) leaflets; in having much smaller sepals which are
externally granulose (instead of stellate-pilose) and smaller petals which
are whitish (instead of red-purple) and sparsely covered with a stellate
puberulence; in having much larger staminodes; and in several other minor
respects.

Herrania balaénsis has a type of fruit which immediately sets it apart
from most other species. The structure of the fruit, according to the draw-
ing of the capsule which was published with the es ee is
similar to that of an unrelated species, Herrania kanukue

The specific epithet balaénsis refers to the locality a eh the type
was taken

Llano Gémez (loc. cit. 18) reported that this species is cultivated to
some extent in Ecuador, but whether this refers to the toleration of the
plant as a wild intruder in plantings or to actual cultivation is not clear.
It seems probable that Herrania balaénsis is actively cultivated.

3. Herrania breviligulata R. E. Schultes, Caldasia 1: 19-24. t. pag.
21, figs. 1-4, t. pag. 24, fig. 5. 1942.
DistRIBUTION: Upper reaches of the Putumayo River, Colombia, and
adjacent Ecuador.

Small, slender, graceful tree up to about 15 feet in height. Trunk erect,
branching near the apex, terete, 6-7 cm. in diameter, with ashy-brown
rimose-scrobiculate bark. Branches apparently tomentose, becoming gla-
brous. Branchlets densely villose, with rust-colored hairs, becoming almost
glabrous. Leaves large, 4- or 5-digitate, very long-petiolate. Petioles terete,
basally slightly swollen, subferruginous, tomentulose, up to 46 cm. long,
about 0.5 cm. in diameter. Leaflets sessile, unequal, lanceolate-oblong,
the margin entire, firmly chartaceous or papyraceous, mostly 20-40 cm.
long, 6-30 cm, wide, very cuspidate-acute, up to 3 cm. long-decurrent-
attenuate, puberulent on both surfaces but especially so beneath; above
light green, very sparsely and minutely stellate-pubescent with white hairs,
beneath pale green, densely and velvety stellate-villose-sericeous; the veins
prominently elevated and brown-tomentose on both surfaces but espe-
cially so beneath. Inflorescence fasciculate, 5—8-flowered. Flowers cauline,
in contracted racemes from the upper portion of the trunk, dark crimson-
purple, pedicellate. Pedicels about 5-7 cm. long, very densely brown-
tomentose, articulate, basally subtended by a short, linear, very densely
brown-tomentose bract 1-1.5 mm. long. Bud globose, 4-9 mm. in diameter,
very densely rusty-villose. Calyx 3-parted, divided almost to the base,
subcymbiform. Sepals 3, broadly elliptic-oblong, obtuse, marginally slightly
revolute, crimson-purple, 13-14 mm. long, 7-9 mm. wide, internally very
sparsely puberulent or glabrescent, externally beset rather densely with
long, white, stellate hairs, valvate in the bud. Petals 4, basally sessile,
obovate, 6—7 mm. long, 5-6 mm. wide, concave, strongly cucullate, muri-
cate-papillose or granulose on both surfaces but especially so externally,

1958 | SCHULTES, SYNOPSIS OF HERRANIA 239

5-nerved, longitudinally striate-veined, internally with prominent veins,
dark crimson-purple, ligulate. Ligules linear, 25-35 mm. long, 1.5—2 mm.
wide, hanging, basally slightly and abruptly contracted, apically slightly
inrolled in a spiral position, 3-nerved, yellowish-red, with purple nerves,
very minutely muricate-papillose or granulose. Staminal tube 5-parted,
with stamens alternately 2- and 3-antheriferous and simple, short, free
filaments. Staminodes petaloid, dark scarlet, lanceolate-elliptic, acute,
margin entire, muricate-granulose on both surfaces, 15 mm. long, 4 mm.
wide, crimson-purple. Pistil 2.8 mm. long. Style terete, simple, purplish,
with a deeply 5-parted stigma. Ovary sessile, very densely pilose, pale
yellow, subglobose, 3 mm. long, 2 mm. wide. Fruit unknown, but said to
be pale yellow when ripe.

Colombia. Putumayo: Mocoa, alt. 850 m., Schultes & Smith 2050 (Type).
Schultes & Cabrera 19082, Anglo-Colombian Cacao Collecting Expedition (Cope
& Holliday) 78; Rio Caqueté, Puerto Limon, Schultes & Cabrera 18720.
Ecuador. NApo-pastaza: near Archidona, alt. 650 m., Mexia 7320.

Herrania breviligulata seems to be closely allied to no other known
species of the genus. In some respects it resembles H. Cuatrecasana but can
be separated from this species readily by having seven (instead of four
or five) smaller leaflets with an entire (instead of crenate-denticulate)
margin, a more cuneate base and a much softer and denser indumentum
on the lower surface; an inflorescence with only six or eight (instead of
eighty or ninety) flowers; much smaller buds; pedicels which are 5-7 mm.
(instead of 20-30 mm.) long; somewhat smaller sepals and petals; ligules
25-30 mm. (instead of 130 mm.) long; and much smaller staminodes
(14 mm. X 4 mm. instead of 25 mm. & 8 mm.) which are apically acute.
Both species are known only from the upper reaches of the Putumayo
River and adjacent regions, an area which would appear to be one focus
of speciation in the genus.

Herrania breviligulata has the shortest ligules known for any of the
species with a subcymbiform calyx, a characteristic which is suggested by
the specific epithet. It is likewise distinguished from all other known
species by the very soft and dense, greyish indumentum which is velvety
on the under surface of the leaflets.

The Ecuadorean collection Mexia 7328 bears the data: “Fruit green,
deeply ribbed.” I have been able to find only one specimen of this col-
lection (that in the Riksmuseum in Stockholm) and this is without a
capsule. The fruit of Herrania breviligulata, therefore, remains unknown,
Mexia’s field notes notwithstanding.

4. Herrania Camargoana R. E. Schultes, Bot. Mus. Leafl. Harvard
Wniveed4® 120. tho290, 32 O50:

Theobroma Camargoana (R. E. Schultes) Ducke, Bol. Técn. Inst. Agron.
Norte 28: ayo)

DISTRIBUTION: Upper Rio Negro basin in Brazil and Colombia.

240 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxx1x

Small, slender and graceful tree, usually up to 10 (but sometimes up to
27) feet tall. Trunk erect, sparsely branched or else unbranched near
the top, about 4-5 inches in diameter, covered with a black bark. Branches
tomentose, but soon glabrous. Branchlets densely villose, ferruginous,
subglabrescent. Leaves very large, digitate, 7- to 9-foliolate, very long-
petiolate. Petioles terete, strongly constricted at the base, softly golden-
ferruginous-tomentellous, up to 60 cm. long, 10 mm. in diameter. Stipules
persistent, subulate, very densely tomentellous, up to 3 cm. long. Leaf-
lets sessile, oblanceolate or broadly lanceolate-ovate, slightly erect, un-
equal, membranaceous-papyraceous, acuminate, basally attenuate, mar-
ginally regularly and conspicuously sinuate in the upper half and every-
where armed with cilia-like stellate hairs (up to 1.5 mm. long), 60-75
cm. long, 16-26 cm. wide, asperous above, sparsely pilose with long and
single hairs, beneath rather softly tomentellose with long stellate hairs.
Inflorescence fasciculate, many-flowered, growing from all parts of the
trunk but principally from the lower portions. Pedicels articulate, up
to 28 mm. long, 0.8 mm. in diameter. Buds globose, up to 10 mm. in
diameter, stellate-pilose. Calyx 3-parted, divided nearly to the base,
subcymbiform. Sepals widely elliptic-oblong, subacute, marginally entire,
externally dark purple, internally blood-red, mostly 12 mm. long, 8—9 mm.
wide, glabrous within, stellate-pilose with rust-colored hairs (up to 1 mm.
long) and very minute white hairs without. Petals 5, sessile, obovate-
rotund, very strongly concave-cucullate, about 8 mm. long, 6 mm. wide,
with 5 dark purple longitudinal nerves as well as finely reticulate nerves,
elsewhere yellow, externally conspicuously muricate-verrucose, apically ex-
tended into a ligule. Ligules linear, mostly 90 mm. long, at the base 1.7
mm. wide, apically filiform, at the very base blood-red but for the greater
part of their length whitish yellow. Staminal tube 5-parted with 2-anther-
iferous stamens and short free and simple filaments. Ovary ellipsoid, 3.5
mm. long, 2—2.5 mm. in diameter, very densely and coarsely white-pilose.
Style terete, simple, yellow, with the stigma apically inconspicuously 5-
parted, 3 mm. long. Staminodes conspicuous, rhomboid-elliptic, obtuse,
marginally entire, verrucose on both surfaces, ashy purple, 14 mm. long,
5 mm. wide. Fruits numerous, almost globose, or slightly ellipsoid, apically
very abruptly long-apiculate (apicule 2—2.5 cm. long), mostly 8—-8.5 cm.
long, 3.5—4.5 cm. in diameter, basally attenuate, with persistent sepals,
longitudinally 10-costate with the primary and secondary ribs almost equal,
thin, knife-shaped, irregular in height but for the most part (in life) 5—6
mm. high, transversely irregularly but conspicuously costate with knife-
shaped ribs, which are slightly lower or often somewhat higher than the
longitudinal ribs, the junction of the longitudinal and transverse ribs
prolonged into a softly carnose, mammose or spine-like projection (which
is somewhat blunt at the tip) with stinging, stellate hairs in all parts but
especially along the ribs; the pericarp thick, usually dark red or blood-red
(but sometimes yellow) when ripe; seeds 25, obtusely round-pyramidal,
about 9 mm. long & 11 mm. wide * 7 mm. thick, imbedded in a white
pulp.

1958] SCHULTES, SYNOPSIS OF HERRANIA 241

Brazil. AMAZONAS: Rio Negro, Serra de Sao Gabriel, Schultes & Lopez 9722
(Type); Rio Negro, Frdes 21468; Rio Negro, Uaupés (Sao Gabriel), Frdes
21540; summit of Serra de Sao Gabriel, alt. 100 m., Schultes & Lépez 8758,
8759, 8762, 8763, 9162, 9619; Tapurucuara (Santa Isabel), Schultes & Lépez
8956; Serra de Uanari, Murca Pires 775; Schultes & Murca Pires 8978; mouth
of Rio Xié, Schultes & Lopez 9205; Nazaré, Schultes & Lopez 9240; Serra
Jacamin, Schultes 9747; between Sao Felipe and Karapana, Schultes & Lopez
9869; Rio Padauiary, Frdes 22673; Rio Uaupés, between Ipanoré and con-
fluence with Rio Negro, Serra Wabeesee, on left bank below Bela Vista, Schultes
& Lopez 9144 (Type of flowers); Schultes & Murca Pires 9130; Murca Pires
ELI O:

Colombia. VaupEs: Cafio near north end of Inambu, Anglo-Colomb. Cacao
Coll. Exped. (Bartley & Holliday) 56; Rio Negro, San Felipe and vicinity
(below confluence of Rio Guainia with Casiquiare), alt. about 600 ft., Cafo
Marijabo, Schultes, Baker & Cabrera 18050; Anglo-Colomb. Cacao Coll. Exped.
(Bartley & H niliday) 45,

Restricted apparently to the uppermost Rio Negro basin of Brazil and
Colombia, Herrania Camargoana seems to have as its closest ally the
Guianan and Venezuelan H. lemniscata. This relationship is strikingly
evident when one compares the fruits of the two concepts. Both species
have relatively small capsules in which there are transverse ribs nearly
as large as the cultriform longitudinal ribs, and soft, pointed mammoid
projections at each junction of the longitudinal and transverse ribs. The
former species, however, has much longer and more upturned projections
than the latter, and would seem, in this as in some other characters, to
represent an extreme in the evolution of the genus. There would appear
to be a rather easily traceable trend from Herrania Mariae through H.
lemniscata to H, Camargoana, on the one hand, and to H. laciniifolia on
the other.

The coloration of the flowers of Herrania Camargoana and H. lemnis-
cata is similarly complex and may also indicate a relationship. No other
known species of Herrania can match these two for complexity of floral
coloration. Herrania Camargoana has sepals which are dark blood-red
externally but scarlet internally; petals which are ashy red or purple
with yellowish stripes; staminodes which are ashy purple-maroon with
white-vellow patches internally but entirely dark red externally; and
ligules, red without and white-yellow within, which are folded or in-
rolled, so that the red is enclosed and is not seen directly. Herrania lem-
niscata has, according to field notes (Steyermark 60558), sepals which
are white in the uppermost two-thirds and rose-salmon below, with rose
stripes; and staminodes (called “‘petals” on the label) which are dull
yellow with dull rose specks in the lower half.

The shape and size of the leaflets, however, differ strikingly in the two
concepts. Herrania Camargoana has oblanceolate or broadly lanceolate-

with the upper half regularly and conspicuously sinuate. Herrania lem-
niscata has leaflets which are at least 80 cm. long and 40 cm. wide with
the margin very deeply pinnatilobate with usually four irregular and, for

242 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

Fic. 2. Crown of Herrania Camargoana.

the most part, widely triangular or widely lanceolate-acuminate segments,
each up to 18 cm. long and 9-10 cm. wide.

Herramia Camargoana is unusually abundant near the summits of the
isolated granitic mountains of the upper Rio Negro valley. It has also
been found in sandy patches along the banks of the rivers themselves. A
search through the Spruce edilecions and notes from this area has failed
to turn up any evidence of Herrania Camargoana in the extensive material
which this early explorer gathered in the long period (1851-1856) which
he spent in diligent study of the area to which the species is confined.

There has been no collection of Herrania Camargoana made as yet from
Venezuela, but it is undoubtedly represented in the upper Rio Negro drain-
age area of that country

fHerrania Comarcoana was named in honor of Dr. Felisberto Camargo,
founder and first director of the Instituto Agronémico do Norte in Belém
do Para, Brazil.

1958 | SCHULTES, SYNOPSIS OF HERRANIA 243

5. Herrania Cuatrecasana Garcia-Barriga, Caldasia 2: 57. ¢#. 2, 1941.
DISTRIBUTION: Upper reaches of the Putumayo River in Colombia.

Small tree 9 feet tall with whitish, maculate bark. Petioles terete,
densely brown-stellate-tomentose. Leaves digitate, large, long-petiolate,
7-foliolate. Leaflets sessile, spreading, oblanceolate-oblong, the margins
very remotely and obscurely crenate-denticulate, acuminate, basally long
and gradually attenuate-decurrent, the lower leaflets about 33 cm. long,
9 cm. wide, the middle 50 cm. long, 17 cm. wide; the leaf surface papy-
raceous, above sparsely and minutely stellate-pubescent, the veins brown-
tomentose, beneath softly stellate-pilose, the veins more prominent and
hirsute above than beneath. Inflorescence fasciculate, 80—90-flowered.
Flowers cauline, growing from the upper portions of the trunk, pedicellate.
Pedicels slender, articulate, stellate-tomentulose, 2—3 cm. long, basally sub-
tended by a linear or filiform, pilose, caducous bract. Buds ellipsoid or
globose, 8-9 mm. in diameter. Sepals 3, equal, oblong, obtuse, dark pur-
lish red, outwardly stellate-pilose with long hairs, inwardly with very
short, reddish hairs, 15-18 mm. long, 7-11 mm. wide. Petals concave,
rotund-ovate, strongly cucullate, glabrous, muricate-papillose, longitudi-
nally striate-nerved, 10-11 mm. long, 7-8 mm. wide, with a long, filiform
ligule, 130 mm. long, 1.5 mm. wide. Staminal tube 5-parted, with the stamens
all bearing 4 anthers. Staminodes petaloid, purple, rhomboid-lanceolate,
attenuate towards the apex, rather obtuse or acute, 25 mm. long, 8 mm.
wide. Ovary shortly depressed, ovoid, densely pilose with whitish yellow
hairs, 2 mm. long. Stigmas 3. Fruit ellipsoid, about 11-12 cm. long and
7-8 cm. in diameter, apically attenuate-acuminate, 10-costate, with 5
primary and 5 secondary ribs, cultriform, covered, especially along the ribs,
with very minute stinging, stellate hairs, rind very thin and brittle, yellow
when ripe. Seeds about 60, regular, compressed, triangular in outline,
14 mm. long, 12 mm. wide, oma thick, enveloped in a white, mucilaginous
pulp, covered with a coriaceous tegument, exteriorly rugose.

Colombia. Putumayo: Rio Guamués, San Antonio de Guamués, alt. about
310 m., Cuatrecasas 11168 (Type); Rio Uchupayaco, alt. 300 m., Schultes
3342: Mocoa and vicinity, alt. 1800-2400 ft., Schultes & Cabrera 19100; Anglo-
Colomb. Cacao Coll. Exped. (Cope & Holliday) 80; Rio Caqueta, Puerto
Lim6n, Schultes & Cabrera 18712, 18715; Rio Putumayo, Puerto Ospina and
vicinity, Schultes & Cabrera 18976; Montclar, Anglo-Colomb. Cacao Coll. Exped.
(Cope & Holliday) 86.

cuador. Rio Sucumbios, 17 hrs. by motor upstream from Puerto Ospina.
Anglo-Colomb. Cacao Coll. Exped. (Cope & Holliday) 84.

Apparently the closest ally of Herrania Cuatrecasana is H. tomentella.
This relationship is discussed under Herrania tomentella

Herrania Cuatrecasana has been considered to be related to H. Mariae
through the concept H. Mariae var. putumayonis. It differs from H. Mariae
chiefly in having leaflets which are very much more long-attenuate-decur-
rent at the base, in the indument and size of the leaflets, in the length
of the pedicels, in having obtuse instead of acute sepals, and in having

244 JOURNAL OF THE ARNOLD ARBORETUM _[VoL. xxxIx

very characteristic and conspicuous rhomboid-lanceolate staminodes, and
in the length of the petals. It is much more softly and densely pubescent
than Herrania Mariae. It is also apparently a very much smaller and
weaker plant than H. Mariae.

Herrania Cuatrescasana likewise resembles H. Dugandi in some re-
spects. The former differs from the latter principally in having dentic-
ulate, instead of almost entire, leaf margins and in being puberulent on
both surfaces of the leaflets instead of being glabrous above. Furthermore,
Herrania Cuatrecasana has many-flowered (80-90) inflorescences, whereas
H. Dugandii has an inflorescence composed of but twenty flowers or fewer.
The pedicels of the flowers of the former species are twice as long as those
of the latter, and there are important differences in the shapes and sizes
of the floral parts and leaflets.

Vegetatively, Herrania Cuatrecasana bears some resemblance to H,
nycterodendron; this will be discussed under that species.

The description of the fruit of Herrania Cuatrecasana is based upon
Schultes 3342 from the Putumayo. The leaves and dried remains of floral
parts adhering to the ripening fruits as well as floral parts which were col-
lected on the ground at the base of the tree have enabled me to determine
the specimens as representing the species in question.

The specific epithet Cuatrecasana honors the botanist Dr. cs vee
casas, formerly of the Instituto de Ciéncias Naturales in Bo and
the Chicago Natural History Museum, now at the Smithsonian ana

6. Herrania Dugandii Garcia-Barriga, Caldasia 2: 59, 61. ¢. 3. 1941.

DISTRIBUTION: Westernmost Amazonia, especially in the Putumayo
basin of Colombia.

Small tree, 9 feet tall, sparsely branched at the apex of the trunk.
Petioles terete, densely hirsute with reddish stellate hairs, 30 cm. long.
Leaves digitate, long-petiolate, 7-foliolate. Leaflets sessile and spreading,
the margins almost entire or near the apex slightly sinuate, obovate-oblong,
acute or obtusely subacuminate, basally gradually attenuate, the lower
18 cm. long, 6 cm. wide, the middle 31 cm. long, 12 cm. wide; leaf surface
firmly papyraceous or thinly coriaceous, above glabrous or very remotely
strigillose, the nerves subimpressed and pubescent, beneath stellate-hirsute,
roughish, the nerves reddish hirsute. Inflorescence axillary, 15—20-flowered.
Flowers cauline in contracted, subumbelliform racemes, on the upper por-
tions of the trunk, dark purple, pedicellate. Pedicels very densely brown-
tomentose, articulate, 1 cm. long or shorter, basally subtended by a short
linear, villous, caducous bract. Bud ovoid-oblong, 11 mm. in diameter,
18 mm. long. Calyx subcymbiform. Sepals 3, oblong, obtuse, dark purple,
externally subdensely armed with stellate hairs, internally with very
short rust-red hairs, up to 21 mm. long and 12 mm. wide. Petals con-
cave, obovate, basally slightly attenuate, striate-nerved, glabrous, 11-12
mm. long, 5-6 mm. wide, apically strongly cucullate, with a long ligule.
Ligule filiform, dark purple, 100 mm. long, 2 mm. wide. Staminal tube

1958] SCHULTES, SYNOPSIS OF HERRANIA 245

5-parted. Staminodes narrowly lanceolate, basally and apically attenuate,
apically very acute, squamulate-rugulose on both surfaces, 12-15 mm.
long, 5 mm. wide. Ovary 4 mm. long, 3 mm. in diameter, densely covered
with yellow hairs. Stigmas 5. Fruit unknown.

Colombia. Putumayo: Rio Putumayo, Puerto Porvenir, above Puerto Ospina
near La Loma, alt. 230-250 m., Cuatrecasas 10742 (Type). AMAzonas: Trapecio
Amazonico, Rio Loretoyacu, alt. about 100 m., Schultes 6038.

Herrania Dugandii seems to be most closely allied to H. kofanorum, a
species from the same general region. Both are rather similar in their
vegetative parts, although the leaflets and certain floral parts of the latter
are very much larger than those of the former. Herrania kofanorum fur-
ther differs from H. Dugandii in having much more coriaceous leaflets
which are more deeply and regularly undulate-sinuate; in having very
strongly unequal sepals, the outer one of which is apically slit (the sepals
of Herrania Dugandii appear at once to be much longer than H. kofanorum
because they are much narrower in relation to their length; but, in real-
ity, they are smaller); in having round-ovate petals which measure 9 mm.
s 7 mm. (instead of obovate, basally attenuate petals measuring 11-12
mm. < 5-6 mm.); in having a more filiform ligule which measures 80—
100 mm. in length X 1 mm. (instead of 100 mm. X 2 mm.); and in
having much larger staminodes, which measure 22 mm. X 6 mm. (in-
stead of about 12-15 mm. & 5 mm.). These size differences combine
with slight differences in shape of floral parts to give the flowers of the
two species rather dissimilar aspects. It is obvious, however, that the
two are closely related. Further collections and studies, especially when
the fruit of both are known, may indicate that Herrania kofanorum should
be treated as a variety of H. Dugandit.

The specific epithet Dugandii honors Dr. Armando Dugand, outstand-
ing botanist, and director from 1940 to 1952 of the Instituto de Ciéncias
Naturales of the Universidad Nacional, Bogota, Colombia.

7. Herrania kanukuensis R. E. Schultes, Caldasia 2: 11. 1943; Bot.
Mus. Leafl. Harvard Univ. 13: 277. 1949; ibid. 14: ¢. 33. 1950.
Theobroma Mariae (Mart.) Decaisne ex Goudot var. lobata Pulle, Rec. Trav.
Bot. Néerl. 9: 151. 1912.
DisTRIBUTION: Southern Venezuela, adjacent Brazil and British and
Dutch Guiana.

Small tree, slender and graceful, up to 16 feet in height. Trunk with
a somewhat striate, brownish-black bark. Branchlets very densely brown-
tomentose. Leaves large, 5- or 6- digitate, stipulate, very long-petiolate.
Stipules caducous, linear, acute, somewhat rigid and dry, very densely
brown-tomentose, 15 mm. long, basally about 1.5 mm. wide. Petiole
strong, up to about 53 cm. long, 6-7 mm. in diameter, very densely and
softly brown-tomentose, basally rather swollen, then abruptly constricted.
Leaflets sessile, unequal, strongly obovate, abruptly acuminate, basally

246 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

abruptly cuneate, almost entire, often near the apex subundulate-sinuate
and often very conspicuously, though minutely, mucronate (with the pro-
longations of the lateral veins up to 1 mm. long), firmly papyraceous, the
central leaflets 30-44 cm. long, 13-16 cm. wide, above dark green or
exceedingly sparsely hirsute, aspero-strigose with long white hairs notice-
able along the veins, the veins very densely and softly stellate-pilose and
rust-colored, the nerves on both surfaces, but especially beneath, promi-
nently raised. Inflorescence fasciculate. Flowers not completely known
(remains of the persistent calyx showing 3 densely brown-stellate-tomen-
tose, elliptic-lanceolate, acute sepals, 14 mm. long, 6 mm. wide; the 5
petals, concave, strongly cucullate, elongate-obovate, about 8 mm. long,
4—5 mm. wide, ligulate). Fruit long-pedunculate (peduncle rather robust,
eae stellate-puberulent, up to 3 cm. long, about 2-3 mm. in diameter),

—8 cm. long, about 4 cm. in diameter, apically very shortly cuspi-
a or subrotundate, basally rounded, extremely velutinous with dense
and minute stellate- miberulence: beans stinging hairs, 10-costate, with
5 narrow and low (1 mm. wide and rarely up to 2 mm. high) blunt primary
ribs and 5 secondary ribs which are similar but smaller and barely notice-
able; pericarp extremely thin and apparently somewhat fragile, yellow
when ripe. Seeds probably more than 60, triangular or angulate-ovate
in outline, complanate, 12 * 10 & 10 mm., 4 mm, thick, in a white
pulp.

Brazil. Rio Branco: lower Rio Branco, Tapanaruca, Frées 23003. British
Guiana. Northwestern slopes of Kanuku Mountains, drainage-area of Mokumoku
Creek, tributary of Takutu River, alt. 150-400 m., A. C. Smith 3541 (Type).
Dutch Guiana. Upper River Corantijne, Hulk 26; River Corantijne, Kaurikreek,
Gongegrijp 2111; Stahel & Gonggrijp 3015; River Coppename, Gonggrijp 2565:
Placer L’Aiva, Gonggrijp 4126; River Marowijne, Gonggrijp 4101; near Ameri-
kan Kondre, Lanjouw & Lindeman 2304; River Tapanahonie, Jaikreek, Gong-
grigp 4117

Herrania kanukuensis stands rather apart from the other known species
of the genus. Its fruit differs strikingly from that of all other species,
with the single exception of the unrelated H. balaénsis of Ecuador. The
capsule is relatively small (measuring 9 cm. & 5 cm.) with a rounded
or very shortly tipped apex (not elongate ovoid or ellipsoid with an acumi-
nate tip, as in most species); the rind is very thin and brittle when dry
(in contrast to the usual thick, leathery and fibrous condition), and the
ribs are not prominent.

The few fragmentary floral parts which remained adhering to the fruit
of Smith 3541 indicate that there are also floral differences between
Herrania kanukuensis and other species.

A duplicate type of Herrania kanukuensis at Kew has several leaflets
which tend to be slightly irregularly dissected, suggesting a condition
which approaches that of some specimens of H. lemniscata with abnor-
mally developed leaflets. Herrania kanukuensis would appear vegeta-
tively to be somewhat intermediate between H. Mariae and H. lemniscata,

1958 | SCHULTES, SYNOPSIS OF HERRANIA 247

but the indumentum on the under surface is more softly tomentose with
brownish stellate hairs than in either H. Mariae or H. lemniscata.

In 1932, Uittien (in Pulle, Fl. Surin. 3: 44. 1932) reduced Theobroma
Mariae var. lobata to synonymy under T. Mariae, identifying all of the
then available material from Dutch Guiana as representing this Ama-
zonian concept. In 1943, it appeared to me that the Surinam concept
described as Theobroma Mariae var. lobata represented the plant which
Schomburgk described from nearby British Guiana as Lightia lemniscata,
and I placed it in synonymy under Herrania lemniscata (Schultes, Cal-
dasia 2: 13. 1943), a species with remarkably lobate leaflets. During the
war, the Utrecht material was unavailable for study. Recently, I have
had an opportunity of consulting all of the Surinam specimens and am
convinced that Theobroma Mariae var. lobata and Herrania kanukuensis
represent the same concept.

From the numerous collections, for the most part from Surinam, it is
now obvious that Herrania kanukuensis is both a widespread and, at least
locally, an abundant element of the flora of Surinam, eastern British Guiana
and the adjacent rim of northern Brazil.

The specific epithet Ranukuensis is derived from the name of the Kanuku
Mountains in British Guiana, where the type specimen was collected.

8. Herrania kofanorum R. E. Schultes, Bot. Mus. Leafl. Harvard Univ.
14: 126. t. 28, upper fig. t. 34. 1950

DISTRIBUTION: Upper Putumayo valley of Ecuador and Colombia.

Small tree, slender and graceful, up to 15 feet tall, usually with one
trunk from each root, columnar, apically branched or unbranched, covered
with an ashy black scrobiculate and scabrid bark, up to 7 cm. in diameter.
Branches ferrugineous-tomentose, becoming almost glabrous, subterete and
sulcate. Branchlets similar but more densely tomentose. Leaves at apex
of trunk, large, digitate, very long-petiolate, 7-foliolate, stipulate. Stipules
caducous, linear, acute, 2.5-3 cm. long, about 3 mm. wide, dry, outside
hispidulous or strigillose, inside usually subglabrous. Petioles robust,
terete but very obscurely sulcate, basally slightly dilated, subferruginous,
very densely and softly tomentose, up to 30 cm. long, 6 mm. in diameter.
Leaflets sessile, unequal, lanceolate-oblong, apex with a cusp about 2 cm.
long, basally attenuate-decurrent, margin conspicuously and regularly un-
dulate-sinuate; blades firmly coriaceous, mostly 17-30 cm. long, 1
cm. wide, above dark green, glabrous or exceedingly sparsely and very
minutely strigillose-pilose with caducous white hairs, brownish-hirsute
along the principal veins, beneath brownish green, very densely and softly
stellate-pilose, ferruginous-tomentose along the principal veins; veins on
both surfaces but especially beneath rather prominently raised. Inflores-
cences fasciculate, up to 20-flowered. Flowers cauline, long-pedicellate, on
the lower parts of the trunk, in contracted racemes. Buds large, elongate-
globose, 1.8 cm. in diameter, stellate-pilose, rather brownish red. Pedicels

248 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

strong, terete, articulate, very densely and minutely stellate-pilose, with
ashy-colored are ‘mostly 9-10 mm. long, 1-1.5 mm. in diameter, at
the base with small linear, acute and densely tomentose bracts, 2-4 mm.
long. Calyx subcymbiform, divided almost to the base. Sepals 3, very
strongly unequal, subchartaceous, entire, within minutely papillose, sub-
glabrous and probably eventually glabrous, without sparsely stellate-strigil-
lose, with hairs up to 1 mm. long and also with extremely minute stellate
hairs, valvate in bud; the inner 2 sepals elliptic, acute, 10 mm. long, 6
mm. wide; the outer round-ovate, 15-18 mm. long, 15 mm. wide, apically
rounded and often deeply (up to 2 mm.). cut, the slit extended interiorly
as a furrow nearly to the base. Petals 4 or 5, round-obovate, sessile, con-
cave, strongly cucullate, glabrous, on both surfaces (but especially on the
outside) muricate-papillose or granulose, dark red, 5-nerved, striate-nerved
with longitudinal purplish veins, ligulate, 9 mm. long, 7 mm. wide. Ligules
filiform, pendulous, membranaceous, entirely glabrous but basally minutely
granulose, about 1 mm. wide, 80-100 mm. long, at the base strongly dilated.
Staminal tube 5-parted, the stamens with alternately 2 and 4 anthers and
strongly flattened, short, free filaments. Staminodes conspicuously petal-
oid, lanceolate-elliptic, acute, basally attenuate, coarsely muricate-papil-
lose on both surfaces, entire, 22 mm. long, 6 mm. wide. Ovary sessile,
elongate-ovoid, distinctly 10-ribbed and 5-locular, yellow, very densely
stellate-pilose, 2.5-3 mm. in diameter. Pistil flattened, 3 mm. long, gla-
brous, simple. Fruit unknown, but said to ripen yellow.

Colombia. Putumayo: path between Puerto Ospina and Concepcién, alt. 250
m., Schultes 3670. CAQuETA: upper Putumayo River, Caucaya, Anglo-Colomb.
Cacao Coll. Exped. (Cope & Holliday) 89. Ecuador. Rio San Miguel or Sucum-
bios, between Rio Putumayo and Quebrada Teteyé, alt. 260 m., Schultes 3478
(Type).

Herrania kofanorum differs from its close ally H. balaénsis in being
smaller, in having leaflets only half as large, and in having the sepals
very conspicuously unequal instead of nearly alike. Herrania kofanorum
has two inner sepals which are elliptic, 20 % 6 mm. and an outer one
which is rotund-ovate, 15-16 * 16 mm., whereas H. balaénsis has three
lanceolate-elliptic sepals which measure 14 x 6 mm. Furthermore, the
outer sepal of Herrania kofanorum is so constructed that it is often con-
spicuously slit to a depth of 2 mm., and this slit is prolonged as a furrow
to the base of the interior of the sepal. (When the split is not present,
there is a markedly thin furrow.) Nothing similar is seen in H. balaénsis.
Both of these species are closely related to Herrania Dugandii.

The specific epithet refers to the Kofan Indians who inhabit the area
where the species is known to occur.

9. Herrania laciniifolia Goudot ex Tr. et Planch. Prodr. Fl. Novo-
Granat. 209. 1862, nomen subnudum; Garcia-Barriga, Caldasia 1:
55. t#. 1,4. 1941.

1958] SCHULTES, SYNOPSIS OF HERRANIA 249

Theobroma laciniufolium (Goudot ex Tr. et Planch.) De Wildeman, Pl. Trop.
Grande Cult. 90. 1903.

DISTRIBUTION: Middle Magdalena Valley, Colombia.

Small tree 12 to 18 feet tall with a simple, erect trunk, about 5 cm. in
diameter. Leaves very large, 7-digitate, round in outline, long-petiolate,
grouped towards the top of the trunk. Leaflets sessile, strongly patent, the
margins deeply pinnatisect with irregular segments (similar to leaves of
Carica Papaya), for the most triangular or lanceolate-acuminate, up to
22 cm. long and 3 cm. wide (but usually smaller), papyraceous or mem-
branaceous, basally long-decurrent-attenuate, laciniate towards the apex,
the apex itself acuminate, 38-58 cm. long, above sparsely stellate-pilose
with suberect hairs along the nerves, the central and lateral veins densely
red-tomentulose, beneath rather pale and rather densely and softly stellate-
pubescent, with the veins as above. Petiole 25-31 cm. long, terete, striate,
densely red-tomentulose, basally swollen. Flowers cauline, fasciculate,
borne on upper portion of trunk, pedicellate, with the pedicel about 2 cm.
long and stellate-pubescent. Sepals 4, lanceolate, rather concave, apically
very acute, 17 mm. long, about 7 mm. wide, pubescent on both surfaces,
externally with fewer and larger reddish stellate hairs, internally near
the apex incanous, with smaller and denser hairs. Petals 5, glabrous,
broadly elliptic, very concave, purplish (with yellowish veins) or “‘green-
ish white’ (Kalbreyer 2047), apically strongly cucullate, ligulate, mar-
ginally revolute, 12 mm. long, 8 mm. wide. Ligule filiform, glabrous, red or
whitish, 180 mm. long, 1 mm. wide. Staminal tube 5-parted, the stamens
all 4-antheriferous. Staminodes petaloid, widely lanceolate, purple, 17
mm. long, 5 mm. wide, the margin slightly sinuous, acute or 3-dentate.
Ovary globose or ovoid, 3 mm. long, densely covered with pale yellow
hairs. Fruit coriaceous, oblong-ovoid, basally rounded, apically long atten-
uate-acuminate, the tip itself rather obtuse, 10-costate, the 5 primary ribs
conspicuously raised, the 5 secondary ribs much less raised, transversely
rugose, brownish red when ripe, 10.5 cm. & 5 cm. in diameter; peduncle 4
cm

Colombia. [No precise locality], Exped. Bot. Mutisii Novae-Granat. 937;
“New Granada, 4,000 ft.”, Kalbreyer 2047. CUNDINAMARCA: Pefio de Conejo
basin of Rio Magdalena, Goudot s. n. ToLIMA: Falan, region of Calamon nte,
alt. 1120 m., Garcia-Barriga 8375; Mariquita, alt. 547 m., Pérez-Arbelaez 10303.

Herrania lacintifolia is apparently not an abundant element of the
flora of Colombia, for it has been collected but thrice in a century, although
it grows in one of the most populous parts of the nation. It is most closely
allied to Herrania lemniscata from which it may be distinguished chiefly
by having leaflets which are deeply and regularly pinnatisect (instead of
very deeply and irregularly pinnatilobed). The under surface of the

* This combination has recently been made independently as “Theobroma ee
Min

folia” (Goud. ex Tr. et Planch.) Llano Gomez in “Cultivo del cacao” (Publ.
Econ. Nac., Bogota) 19. 1947.

250 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

leaflets of Herrania laciniifolia is very much more densely and _ softly
stellate-pilose than in Hf. lemniscata; the leaflets are, in general, smaller;
and the ligules measure 180 mm, (instead of 85 aig) in length. The
fruit of Herrania laciniifolia ripens red, whereas that of H. lemniscata,
in common with all other species but one, ripens bright yellow. This other
exception is fHerrania Camargoana of the upper Rio Negro basin. The
fruit of Herrania Camargoana ripens scarlet, but this species does not
appear to be closely allied to H. lacinifolia, and the similarity of fruit
color is probably coincidental.

In the collection of Mutis plates preserved at Madrid, there are sev-
eral illustrations representing Herrania lacinitfolia. Drawn by Matiz, these
are all in black and white, not in color. One plate has a leaf with a com-
plete leaflet and a length of stem with several flowers and buds; another
plate has analytical drawings of the flowers and fruits; a third has analyses
only of the floral parts

The Mutis specimen in Madrid (cited above) is sterile, but it was un-
doubtedly taken from the tree from which the plate was made. In view
of the scarcity of collections of Herrania laciniifolia, it is unfortunate that
a definite locality for the Mutis collection is not available. Both the speci-
men and the plates agree perfectly with the type and the later material of
this remarkably distinct species.

10. Herrania lemniscata (Schomb.) R. E. Schultes, Caldasia 2: 13.
1943; Pittier et al. Cat. Fl. Venez. 2: 134, 139. 1947: R. E. Schultes,
Bot. Mus. Leafl. Harvard Univ. 13: 281. ¢. 30. 1949.

Lightia lemniscata Schomb. Rep. Assoc. Advancem. Sci. 13: 71. 1844. Nomen
subnudum

DisTRIBUTION: British and Dutch Guiana and the Orinoco basin of
eastern Venezuela, extending into northeastern Colombia.

Small tree, with a simple, round, slender trunk up to 30 feet tall, with
leaves grouped at the tip of the trunk. Branches apparently tomentose,
becoming subglabrous. Leaves very large, 7-digitate, very long-petiolate,
stipulate. Stipules caducous, minutely subulate, extremely densely and
softly ferruginous-tomentose, about 2 cm. long, basally 2 mm. wide. Peti-
oles robust, terete, sulcate, apically widely complanate-flabelliform, very
densely and softly rusty tomentose, slightly swollen then immediately
strongly constricted, up to 45 cm. long, 8-10 mm. in diameter. Leaflets
sessile, unequal, the central one up to 80 cm. long, 40 cm. wide, acuminate,
basally long-decurrent-attenuate, the margin very profoundly pinnati-
rer with (usually) 4 irregular, usually triangular or broadly lanceolate-
acuminate segments, in the longest part of leaflet up to 18 cm. long, 9-10
cm. wide, papyraceous or membranaceous, above dark green, subglabrous
or beset with very remote stellate eo densely ferruginous-strigillose
along the nerves, beneath pale green, densely and softly stellate-pilose, the
nerves densely ferruginous-tomentose, the veins prominent on both sur-
faces. Inflorescence fasciculate, bearing probably up to 30 or 40 (but

1958 | SCHULTES, SYNOPSIS OF HERRANIA 2a

usually about 20) flowers. Flowers borne in contracted racemes on the
lower portion of the trunk, dark crimson-purple, pedicellate. Pedicels
strong, terete, minutely and densely stellate-tomentellose, articulate, 7-8
mm. long, basally subtended by a short, linear, tomentose bract, 1 mm.
long. Buds globose, 9-10 mm. in diameter, very densely and minutely
stellate-tomentellose and stellate-pilose. Calyx divided almost to the base,
subcymbiform. Sepals 3, reflexed, almost equal, thick, internally very
densely puberulent, externally very densely stellate-tomentose and stellate-
pilose (2 distinct types of hairs), with an entire and strongly inflexed
margin, brownish red; the two larger sepals ovate, subobtuse, 12-13 mm.
long, 7 mm. wide; the smaller one lanceolate-elliptic, apically subobtuse,
13-15 mm. long, 5-8 mm. wide. Petals 5, sessile, elongate-obovate, con-
cave, strongly cucullate, entire, dark blood-red (sometimes possibly yellow-
ish) with 5 prominent black or purplish nerves, 6-7 mm. long, 4 mm. wide,
glabrous, on both surfaces but densely muricate-papillose externally, thick,
lizulate. Ligules hanging, membranaceous, filiform, glabrous, without
nerves, basally 1-1.5 mm. wide, about 85 mm. long, blood-red, sometimes
becoming yellowish. Staminal tube 5-parted with stamens alternately 2-
and 4-antheriferous, filaments short, free and simple. Staminodes con-
spicuous, petaloid, dark blood-red, sometimes apically yellowish, elliptic,
slightly and obscurely undulate, acuminate, 10-12 mm. long, 4-5 mm.
wide, glabrous, densely muricate-granulose on both surfaces. Ovary sessile,
ovoid-globose, 4.5—-5 mm. long, 2.5 mm. in diameter, distantly 10-costate
with 5 primary and 5 secondary ribs, yellow, very densely hispid-pilose.
Style terete, glabrous, simple, with an obscurely 5-parted stigma. Fruit
long pedunculate (with a strong, glabrous or glabrescent peduncle up to
4.5 cm. long and 3 mm. in diimereo) perfectly ovoid, 7 cm. long, 4 cm.
in diameter, apically very shortly and abruptly cuspidate, basally almost
rounded, densely and minutely stellate-puberulent, apparently lacking
stinging hairs, 10-costate with 5 narrow, comparatively low, subcultriform
primary ribs which are 2 to 4 mm. high and 1-2 mm. thick (in the dried
specimen) and 5 similar but smaller ribs, the very thin pericarp strongly
reticulate-costate between the longitudinal ribs, probably yellow when
ripe. Seeds apparently more than 60, triangular or angular-ovate in out-
line, flattened, 12 * 11 11 mm., above 4 mm. thick, buried in a white
pulp.

British Guiana. Banks of Barima River. Schomburgk s. n. (Type); [no
precise locality], im Thurn s. n. (Type of fruit); Essequibo River, White Greek.
Groete Ck., Forest at . Brit. Guian. Field No. F1763, Research No. 4500.
Colombia. SANTANDER: 15 km. east of Puerto Berrio, alt. about 250 m., Scolnick,
Araque & Barkley ain Venezuela. Near mouth of Orinoco River Rusby &
Squires 252: Pakaraima Mountains, Myers 3371. Bovivar: Salto de Para, Medio
Caura. alt. 120 m., L. Williams 11339; lower part of Quebrada O-paru-mo, tribu-
tary of Rio Pacairao, below Santa Teresita de Kavanayen, alt. 915-1065 m.,
Stevermark 60558.

Herrania lemniscata is undoubtedly one of the most strikingly distinct
species of the genus. It can be distinguished at once by its very large and

Zoe JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

broad leaves with pinnatilobed leaflets. It is, apparently, most closely
related to H. lacintifolia of central Colombia, but its leaflets are much more
coarsely incised, with fewer and wider lobes. Florally, H. lemniscata does
not seem to be very distantly allied to H. Mariae and H. kanukuensis.

It would appear that the Guianan Herrania lemniscata occupies a some-
what intermediate position between the Colombian #. lacintifolia on the
one hand and the Amazonian H. Camargoana on the other. This relation-
ship has been discussed under the latter species.

The name Lightia lemniscata was published by Schomburgk in 1844
without an adequate description and without the citation of specimens.
It must be considered a nomen subnudum. Several years later, Schom-
burgk (Linnaea 20: 756. 1847) reduced the name to synonymy under Her-
rania Mariae, and, wishing to perpetuate a generic name honoring Governor
Light of British Guiana, he transferred the name Lightia to a new genus
in another family.

In 1848, he (Schomburgk, Fauna und Flora von British-Guiana. 993.
1848) listed under Herrania Mariae specimen(s) which he had collected in
British Guiana along the River Barima and its affluents. It is probable,
then, that this represents the area from which the type material of Lightia
lemniscata came.

Fortunately, I have had an opportunity to study a beautiful collection
(Archer 2514) from this same region. I have considered it as a topotype,
although it must be remembered that Schomburgk’s mention of the “River
Barima and its affluents” circumscribed a rather extensive area.

All of the material of Herrania from British Guiana and northeastern
Venezuela which I have seen is (with exception of collections from near the
Brazilian border) referable to one species. This species is distinct from
others of Middle and South America. I am inclined to believe that these
specimens are referable to the concept which Schomburgk called Lightia
lemniscata

In the New York Botanical Garden there is a specimen collected in the
easternmost part of Venezuela near the mouth of the Rio Orinoco in 1896
(Rusby & Squires — On the label of this specimen, a handwritten
notation states: “ = coll. by Schomburgk in Brit. Gui.” On the basis
of this annotation a the near homogeneity of the Guianan collections,
I validated Schomburgk’s nomen subnudum and transferred it to the genus
Herrania in 1943,

When I validated the Schomburgk concept, making the new combina-
tion Herrania lemniscata, European collections were unavailable for study.
Recently, additional and extensive material which I have seen in England
indicates that Herrania lemniscata is indeed a very distinct species. Schom-
burgk’s original water-colors, made in the field, are preserved at the British
Museum (Natural History). Included in the collection is an excellent
painting of this concept (as Lightia lemniscata) which depicts with un-
usual accuracy the habit of the small tree. There is, likewise, what appears
to be a Schomburgk water-color attached to one of the herbarium sheets
at Kew.

1958] SCHULTES, SYNOPSIS OF HERRANIA 205

The collection Myers 3371 consists onlv of flowers. It is referred with
some reservation to Herrania lemniscata. The collection L. Williams 11339
has been reported in the literature (L. Williams, ee nr botanicas
en la Guayana venezolana. 309. 1942) as Herranta Mar.

By using the specific epithet lemmniscata, See reel to call
attention to the very long ribbon-like ligules which adorn the flower.

11. Herrania Mariae (Mart.) Decaisne ex Goudot, Ann. Sci. Nat. IIT.
2: 233. 1844
Abroma Mariae Mart. Denkschr. Regensb. Bot. Gesell. 3: 297. t¢. 6, 9. 1841.
Theobroma Mariae (Mart.) Schum. ex Mart. Fl. Brasil. 12(3): 71. ¢. 15.
1886; Ducke, Rodriguesia 4: 273. ¢. 5, fig. 2. tt. 6, 7. 1940.
DisTRIBUTION: General in the Amazon basin, with the exception of the
northern and northwestern sectors.

Small or large tree, up to 30 (doubtfully up to 60) feet tall; trunks
(often 5 or 6 from a root) erect, columnar, 7-30 cm. in diameter, with
brownish-black or black, rimose or scrobiculate bark. Branches some-
times closely crowded, forming a subglobose crown, in which the lower
branches are subhorizontally spreading, rather flexuous, sometimes few and
wide-patent; ferruginous-tomentose but becoming subglabrous, subterete,
sulcate. Branchlets similar, spreading. Leaves large, digitate, long-petio-
late, 6-9 (usually 7-) -foliolate, stipulate. Stipules caducous, linear, acute,
2.5—4 cm. long, about 3 mm. wide, dry, tomentose. Petioles strong, terete,
subsulcate, basally somewhat swollen, 30-50 cm. long, 7 mm. in diameter,
sibienmeinous. very densely and softly tomentose with erect, brown,
stellate hairs. Leaflets thin-membranaceous to thin-papyraceous, sessile,
paar usually strongly rhomboid-ovate, apically acuminate (acumen
about 2 cm. long), basally attenuate- deeurene entire or very slightly
Bee towards the apex and conspicuously armed with small, spine-
like, hirsute, mucronulate prolongations of the veins up to 1 mm. long; the
leaf surfaces usually 27-54 cm. long; 7-19 cm. wide, above dark green,
very sparsely and minutely hirsute (or, in rare specimens, subglabrous)
with white hairs (very rarely with brownish stellate hairs), more densely
hirsute along the principal veins and along the margin, beneath pale green,
subasperous or soft with minute stellate-sericeous hairs, ferruginous-tomen-
tose along the principal veins; the veins prominent on both surfaces, but
conspicuously elevated beneath. Inflorescences fasciculate, usually few-
flowered (with about 10-15 flowers) but frequently many-flowered (80-
90). Flower buds subglobose, large, 17 mm. in diameter, densely stellate-
hispid. Flowers cauline, long-pedicelled, growing from the middle and
lower portions of the trunk in contracted racemes. Pedicels densely fulvo-
tomentose, hispidulous with minute, appressed strigillose hairs and also
sparsely setose, articulate, 1.5-5 cm (mostly 3.5-5) long, basally sub-
tended with a short linear, acute, densely tomentose bract 3 mm. long.
Calyx subcymbiform, divided nearly to the base. Sepals 3-5, strongly
unequal, entire, externally densely stellate-strigose, internally glabrous,

254 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

or very minutely pilose, thick-membranaceous, brown, valvate in bud later
reflexed, longitudinally striate; the interior broadly elliptic-oblong, api-
cally subacute, with coarse hairs frequently up to 1 mm. long, 12-15 mm.
long, 6-10 mm. wide; the exterior very broadly rotund or broadly ovate,
apically abruptly and obscurely acuminate, with coarse hairs 1 mm. long,
11-13 mm. long, 11-12 mm. wide. Petals 4 or 5, basally sessile, concave,
rotund-obovate, apically very strongly cucullate, 7-9 mm. long, mm.
wide, glabrous, muricate-papillose on both sides, 5-nerved, pale purplish
red with black or purple veins, apically ligulate. Ligules filiform, pendulous,
membranaceous, pale yellow or white, with purple venation, minutely
granulose in all parts but especially so at the base, basally up to 4 mm.
and apically hardly 1 mm. wide, 75-100 mm. long (very rarely less),
apically slightly coiled. Staminal tube 5-parted, stamens alternately 2-
and 4-antheriferous, filaments short and free, anthers yellow. Staminodia
conspicuous, petaloid, reflexed, red, lanceolate or lanceolate-elliptic, api-
cally obtuse and usually indentate mucronate or serrate, somewhat sinuate-
undulate, glabrous, muricate-granulose on both sides, up to 20 mm. long,
6-7 mm. wide. Ovary sessile, elongate-subglobose, pentagonal (distinctly
10-costate and 5-locular), densely stellate-pilose, rose-colored or yellow-
white. Style short, subcylindric-pentagonal, reddish. Stigmas 5, filiform,
rose-colored. Fruit baccate, elliptic-ovoid, apically acuminate, long-
peduncled (peduncle up to 2.5 cm. long), conspicuously 10-costate; the
5 primary ribs large, protruding, acute-scutelliform, about 8 mm. tall, 5
secondary ribs similar but smaller, about 4 mm. tall, conspicuously fibrous-
striate transversely between the ribs, very densely stellate-hispid with
stinging hairs along the ribs, up to 10-12 cm. long, 6—7 cm. in diameter,
pericarp thick, subsucculent (not conspicuously fibrous), yellow when ripe.
Seeds 30-40 (possibly —60), obtusely rhomboid, flattened, with a sub-
coriaceous testa, about 10 cm. long, 9 mm. wide and 4 mm. thick, in a
white, acidulous pulp.

Brazil. AMAzonaAsS: Rio Solimées, von Martius s. n. (Type); Paleta, Teffé,
Krukoff’s 4th Exped. Brazilian Amazon 4523; Fonte Béa, Frdes 20630; Rio
Jurua, Marary, Ule 5031; Rio Amazonas, Taperinha, near Santarém, Capes
04; Rio Madeira, Humayta, near Tres Casas, Krukoff’s 5th Exped. Brazilian
Amazon 6085; Riosinho, Juruema, Frdes 21041. ParA: Belém, Ducke 595;
Museu Goeldi. oe Pires & Black 740; a Schultes 8072; Be ibaura.
Black 47-1916. Colombia. Amazonas: Trap Amazonico, interior regions
of trapecio between Amazon and Pu ano “watersheds, alt. about és
Spa cai 6759; mouth of Rio Atacuari, Black & Schultes 46-223. Peru. MADRE
E Dros: near Iberia, Schultes 6461.

Due primarily to the detailed drawing of Herrania Mariae which was
published in Flora Brasiliensis and to the general availability and reliability
of this work, botanists have had a tendency to consider as representing
this species collections from a wide area but which actually belong to very
diverse concepts. Almost all collections in our herbaria have been referred
to Herrania Mariae. Indeed, very recently, Ducke has stated that Theo-

1958 | SCHULTES, SYNOPSIS OF HERRANIA 259

broma Mariae is “the only species in the Brazilian Amazonia representing
the subgenus (or section) Herrania ... ,” that this “species is found
throughout the hylea (including the Guianas) ... ,” and that “it is
probable that, in addition to aérorubra, still other species of the subgenus
Herrania, described from northern and northwestern South America, will
in the future be reduced to synonymy under Theobroma Mariae” (Rodri-
guesia 4: 273. 1940).

Herrania Mariae is probably the tallest species in the genus and, unlike
almost all other species (except H. pulcherrima), it often has many trunks
growing from one root. Martius described the plant as being 20-30 feet
in height, and the drawing which was published with the original descrip-
tion shows an extremely robust and corpulent tree with a basal diameter
of some 15—30 cm. and with a very heavy, round and full crown. My field
work with Herrania leads me to believe that this drawing, herein re-
produced for historical reasons, is erroneous and greatly exaggerates the
size of the trunk and crown. Ducke commented similarly on this illustra-
tion when he wrote (loc. cit.), “In Martius’ drawing, reproduced in ‘Flora
Brasiliensis,’ 7. Mariae appears as a much-branched tree which does not
in any way correspond with the real habit of our plant ... .” The few
collections of this widespread species which are at hand at the present time
exhibit some variation, but an abundance of comparative material is lack-
ing for determining with precision the character and extent of the variation.
I am certain that when more abundant collections and field observations
have been made, it will be necessary to recognize several definite geograph-
ical variants. An example of our inability to treat with complete certainty
some collections is seen in Krukoff 6085 which may possibly be an hybrid
between Herrania Mariae and H. nitida. Schultes 6238 and 6461, un-
fortunately sterile, represent perhaps one of the extremes exhibited by the
available material of Herrania Mariae. They are the southernmost collec-
tions of the genus, occurring on high land which never floods (in contrast
to the flood-land habitat of the banks of the Amazon), and appear to
have on the under surface of the leaflets a much denser and softer indumen-
tum which tends to be rather greyish in life.

In the herbarium at Munich, there are seven very ample specimens of
Martius’ type collection. A study of this material has enabled me to
evaluate the concept more critically. The type material has extremely
membranaceous to very thinly papyraceous leaflets which have a rather
softly asperous indumentum on the lower surface and a sparse asperous
pilosity on the upper. The leaflets are strongly rhomboid-obovate, a shape
which is peculiar to Herrania Mariae. The petiole, shorter and less robust
than in most species, is densely tomentose, and the inflorescence is many-
flowered.

Herrania Mariae is related to H. Cuatrecasana, H. nitida and H. pulcher-
rima. All four species have a similar type of fruit with very large and
protruding cultriform ribs which are unequal and with strong transverse
wrinkles between the ribs and more or less at right angles to them.

Herrania Mariae resembles H. Cuatrecasana very closely in some char-

256 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xxx1x

acters, and I once suspected that the latter might well be considered a
variety of the former. Further field work and additional collections of
Herrania Cuatrecasana, however, and the discovery of H. Mariae var.
putumayonis give us reason to maintain the two concepts as completely
distinct. The most obvious difference between the two species is the shape
of the leaflets: very long oblanceolate-oblong and very long and grad-
ually attenuate towards the base in Herrania Cuatrecasana and rhomboid-
obovate and more abruptly attenuate towards the base in H. Mariac; the
leaflets of the former are also usually very much longer than those of the
latter species, and the petiole is much longer, stouter and more softly
ferruginous-tomentose. Herrania Mariae has lanceolate or lanceolate-
elliptic, somewhat sinuate-undulate and apically often serrate or indentate-
mucronulate staminodes which are, for the most part, 20 mm. long and
6—7 mm. wide, whereas H. Cuatrecasana has definitely rhomboid-lanceolate,
entire and apically non-serrate staminodes which measure 25 mm. long
and 8 mm. wide. The ligule of the former species usually measures from
75 to 100 mm. in length, of the latter, 130 mm

The differences which set Herrania Mariae apart from H. nitida are
very evident. The latter is separated at once from the former by its com-
plete lack of indumentum on the leaflets (or, when slightly tomentulose
underneath, by the sparsity and asperous nature of the hairs); by its
leaflets which are usually lanceolate-elliptic, more firmly chartaceous and
most often entire; and by a number of floral characters.

Herrania pulcherrima is most easily separated from H. Mariae by the
enormous size of its leaflets (up to 60 cm. long) which are oblong, for the
most part about half as wide or wider, with the secondary and tertiary
veins extraordinarily conspicuous and med beneath, giving the upper
surface, in most specimens, a verrucose appearance; by the unusually long
and strong petioles; and by the larger flowers which are dark red in all
parts, excepting the ligules which in some specimens are cross-banded
scarlet and whitish. Both of these species have a characteristically trifid
staminode.

With the specific epithet of Herrania Mariae Martius honored Maria,
Queen of Saxony, whose father, King Maximilian Joseph of Bacar,
patronized Martius’ extensive botanical explorations in South America.

Van Hall (Cacao. ed. 2. 74. 1932) states that seeds of this species are
occasionally found as an adulterant in “Para cacao”, but it is not certain
that his identification can be taken as reliable: his confusion of Herrania
Mariae with another species is obvious from the statement that “it is very
common in the forests of Surinam near Paramaribo and also deep into the
interior’ (loc. cit.).

lla. Herrania Mariae (Mart.) Decaisne ex Goudot var. putumayonis
Rad: ae Bot. Mus. Leafl. Harvard Univ. 14: 129. ¢. 30, upper
fig. 19
oe Western part of the Amazon Valley, especially in the
Putumayo River basin.

1958] SCHULTES, SYNOPSIS OF HERRANIA 257

Small tree up to 12 or 14 feet tall, differing from Herrania Mariae chiefly
in having much larger (up to 52 cm. long, 18 cm. wide), lanceolate-elliptic
(not conspicuously rhomboid) leaflets; stronger and longer petioles; larger
flowers up to 17 mm. in diameter, with the buds globose, and shorter
ligules up to 70 mm. long, but usually somewhat less.

Brazil. AMAZONAS: Rio Jurua, Lago Cerrado, Traill 65. Peru. Loreto: Rio
Putumayo, between Rio Igaraparana and Rio Yaguas, alt. 100-150 m., Schultes
4010 (Type).

Additional material may indicate that this concept is deserving of specific
rank. At the present time, however, it would seem advisable to treat it
as representing a variety of Herrania Mariae. The fruit of Schultes 4010
is hardly distinguishable from that of typical Herrania Mariae. The
flowers have several differentiating characters, the most conspicuous of
which is the shorter ligule. Vegetatively, the collection is extremely similar
to Herrania nycterodendron (with the type plant of which it was growing)
and differs markedly from H. Mariae chiefly in the departure from the
typical rhomboid form of the leaflets and in their unusually large size.
The type plant of Herrania Mariae var. putumayonis consisted of four or
five trunks in a clump, whereas H. Mariae is a treelet with a single trunk,
although the condition of several trunks from one root is not uncommon
in H. Mariae.

In the Paris herbarium there is a specimen, the collector and date of
which we are ignorant, referable with reservation to Herrania Mariae var.
putumayonis. It was collected at the upper Amazon town which is now
called Teffé: ‘fluv. Amaz. Ega. 2660. Abroma n. sp. Arbor debilis.” The
specimen is sterile, but it was thus early recognized as a distinct concept.

Herrania Mariae var. putumayonis may represent a western variant of
the species which is most abundant in the eastern half of the Amazon basin.
The varietal epithet refers to the Putumayo River.

12. Herrania nitida (Poepp.) R. E. Schultes, Caldasia 2: 16. ¢. pag. 17.
1943

Abroma nitida Poepp. in Poepp. & Endl. Nov. Gen. ac Sp. Pl. 3: 73. 1845.
Brotobroma aspera Karst. & Tr. ex Tr. Nuev. jén. y esp. pl. fl. Neo-Granat.
1854,

Herrania aspera (Karst. & Tr. ex Tr.) Karst. Linnaea 28: 447.

Theobroma nitidum (Poepp.) Schum. ex Mart. Fl. Brasil. mou 72. 1886.
Non T. nitida Bernoulli, Neue mes all. Schweiz. Gesell. gesam.
Naturw. 24(3): 15. ¢. 17, fig. 3

Herrania atrorubens Hub. Bull. So. fe II. 6:

Theobroma aspera (Karst. & Tr. ex Tr.) Van Hall,* oe eS 2. 49. 1932.

*This combination was made, but imperfectly so, by van Hall who, in a key,
refers to “7. aspera (Karsten) Schumann”, He refers to Schumann’s account in
Martius’ Flora Brasiliensis where aspera is not mentioned under Theobroma, Her-
rania or eee tobroma; and there is no evidence that Schumann ever made the com-
binat

258 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

Herrania nitida (Poepp.) R. E. Schultes var. aspera (Karst. & Tr. ex Tr.)

Schultes, Bot. Mus. Leafl. Harvard Univ. 14: 130. 1950, pro parte.

DISTRIBUTION: Widespread in the western half of the Amazon Valley
and the Orinoco basin in Colombia.

Small tree, rather weak, graceful, probably up to 12 feet tall, with dark
brown, roughened bark; the trunk usually less than 7 cm. in diameter at
the base. Branches sparsely and minutely tomentose, becoming glabrous.
Petioles subcomplanate, conspicuously striate-sulcate, ferruginous, very
minutely closely appressed stellate-tomentulose, basally somewhat swollen,
up to 40 cm. long, 5 mm. in diameter. Leaves grouped at the apex of the
trunk, 7—9-digitate, long-petiolate, stipulate. Petiole usually slender,
asperous, very minutely ferruginous-tomentulose, sometimes becoming
subglabrous, terete but very deeply sulcate, up to about 40 cm. long, basally
4—5 mm. in diameter. Stipules linear, up to 2.5 cm. long, 1 mm. wide,
usually brown stellate-setulose. Leaflets sessile or nearly so, unequal,
lanceolate-oblong, acuminate, basally very long attenuate-decurrent, mar-
ginally entire, rigid-chartaceous to subcoriaceous, light green, shiny, gla-
brous above, glabrous or rarely with very sparse and deciduous minute stellate
hairs remotely placed along the nerves beneath; the central leaflets 25—45

m. long, 7-14 cm. wide; the lateral leaflets much smaller and often
asymmetrical. Inflorescence fasciculate; often very numerous on the basal
portion of the trunk, up to 30—40-flowered. Flowers cauline, pedicellate.
Pedicels articulate, densely appressed, tomentulose with occasional strigose
setae, 5 mm. long, less than 1 mm. wide. Calyx subcymbiform. Sepals
3, yellow-red, externally very coarsely stellate-setose, internally glabrous:
outer sepal rounded-ovate, or triangular-ovate, up to 14 mm. long 10 mm.
wide (but usually smaller); inner sepals elliptic, acute, up to 17 mm. long
(but usually shorter) 6 mm. wide. Petals 5, broadly obovate, rotund,
4—5 mm. long, 4-5 mm. wide at the top, strongly cucullate, fleshy mem-
branaceous, very densely muricate-papillose on both sides but especially
without, glabrous, longitudinally marked with 5 or 6 dark red veins, yellow-
ish-rose or rose-white, ligulate. Ligules linear, hanging, basally slightly
contracted, apically slightly coiled, pink or scarlet, up to 80 mm. long,
membranaceous, glabrous, longitudinally marked with five purple nerves.
Staminal tube 5-parted, alternately 2- and 4-antheriferous with simple.
flattened, free filaments. Staminodes conspicuous, petaloid, glabrous,
muricate-granulose, elliptic, acute, marginally slightly undulating. dark
blood-red without, vellowish-red within, 9 mm. long, 5

5 mm. wide. Ovary
5-locular, subcylindric, densely and very minutely stellate-pilose, yellow,
about 1.8 mm. long, 1 mm. in diameter. Style strongly flattened, 3 mm.
long, with a simple stigma. Fruit baccate, ovoid, acuminate, 11 cm. long,
5 cm. in diameter, dull, rich green, 10-costate, the 5 primary ribs thin,
cultriform, basally up to 2.5-3 mm. and apically 1.5 mm. thick, 8-9 mm.
high, the 5 secondaries similar but smaller, minutely and very densely
beset with stinging hairs along the ribs; pericarp between the ribs usually
smooth and not fibrous-rugose, very thin. sometimes with a few stinging
hairs but becoming glabrescent. Peduncle short and strong, usually under

pon

1958 | SCHULTES, SYNOPSIS OF HERRANIA 259

14 mm. long, 2.5 mm. in diameter. Seeds 30-40 or more, flattened, roughly
triangular in outline, 9 mm. * 8 mm. X 5 mm., in a white pulp

Brazil. Amazonas(?): Glaziou 9635. AMAzoNAS: Rio Purus, varzea at Sobral,
Traill 64; Rio Solimées, Santo Antonio do Ica, Ducke 7618; Benjamin Constant,

Colombia. Amazonas: Rio Ammaconas, La Victoria, L. Wiliams 2816, 2843;
Leticia, alt. about 100 m., Schultes 6000, 6016, 6141, 6142, 6143, 6144, 6145,
6146, 6147, 6149, 6192a; Rio Loretoyacu, Schultes 6304, 6045, 6058, 6118, 6124,
6640, 6878, 8129, Schultes & Black 8286, 8377; Rio Apaporis, between Rio
Pacoa and Rio Kananari, region around Soratama, alt. 250 m., Schultes &
Cabrera 13628, 13630, 13632, 14880, 14882; Schultes & Garcia-Barriga 14016,
Rio Caqueta, La Pedrera, Schultes 5876; Anglo-Colombian Cacao Collecting
Expedition (Baker & Cope) 17, 19, 22, 24. CagueTA: Rio Orteguaza, Tres
Esquinas, Schultes 3698; upper Rio Putumayo, Montclar, Anglo-Colomb. Cacao
Coll. Exped. (Cope & Holliday) 85; Caucaya, Anglo-Colomb. Cacao Coll.
Exped. (Cope & Holliday) 87, 88; Rio Caucaya, Laguna Primavera, Anglo-
Colomb. Cacao Coll. Exped. (Cope & Holliday) 92; Rio Caguan, Arbolitos,
Anglo-Colomb. Cacao Coll. Exped. (Cope & Holliday) 100; Camp Three,
Anglo-Colomb. Cacao Coll. Exped. (Cope & Holliday) 112; Cartagena, Anglo-
Colomb. Cacao Coll. Expedition (Cope & Holliday) 106; Rio Caqueta, Camp
Five, Anglo-Colomb. Cacao Coll. Exped. (Cope & Holliday) 120. Putumayo:
Rio Putumayo, Puerto Ospina, alt. about 280 m., Schultes 3405, Schultes &
Cabrera 18933; Rio Caucaya, Schultes 3730. VauprEs: upper Apaporis basin,
path between Rio Itilla and Rio Macaya, alt. 300 m., Schultes 5351; Rio
Macaya, Cachivera del Diablo, alt. 350 m., Schaltes 5491; near confluence of
Rio Ajaju and Rio Macaya, Puerto Hevea, alt. 350 m., Schultes 5529; middle
Apaporis basin, Rio Kananari, Buenos Aires, alt. about 250 m., Schultes 5685;

10; Rio Vaupés, Puerto ‘Naré, Schultes 5359; near Miraflores, Schultes 5715; La
Jirisa, Schultes 5755; near mouth of Rio Kubiyu, Schultes & Cabrera 14537;
Circasia, alt. 800 ft., Schultes & Cabrera 19665; between Mitt and Javareté,
Igarapé, Murutinga, near Tipiaca, Schultes & @ihrer 19284: Rio Kuduyari.,
Cerro Yapoboda, Schultes & Cabrera 14343; near Irabasu, Schultes, Baker &
Cabrera 18439; lower course of Rio Kaduyari, alt. 700-800 ft., near Yararaca,
Schultes, Baker & Cabrera 18553, Rio Paca, Wacaricuara, Anglo-Colomb. Cacao
Coll. Exped. (Bartley & Holliday) 55c; Rio Inirida, Cano Caribe, Anglo-Colomb.
Cacao Coll. Exped. (Bartley & Holliday) 67; Rio Papunawa, near junction
with Rio Inirida, Azglo- Bes Cacao Coll. Exped. (Bartley & Holliday) 76.
Meta: Villavicencio, alt. 400 m., Triana s. n., Killip 34247: Llanos de San
Martin, near Villavicencio, alt. 400 m., Triana 5333; Cordillera La Macarena.
path between Rio Guejar and Cano Guapayita, Cano Yerly, Jdrobo & Schultes
768, Schultes 11627; savannahs near San Juan de Arama, Rio Guejar, near
aoe strip “Los Micos”, alt. about 500 m., Jdrobo & Schultes 1325.
uador. Rio Pastaza, near Andrés, Spruce 4969.
[No precise locality], “Herb. Pavon”’. SAN MARTIN: Rio Tocache,
De 1979 (Type). Loreto: Rio Maranon, near Pongo de Manseriche,
Tessmann 4024; Iquitos, alt. about 100 m., Killip & Smith 27431, Murca Pires
ick 873; lower course of Rio Huallaga, between Yurimaguas and Balso-
puerto, alt. 135-150 m., Killip & Smith 28234; Rio Amazonas, Caballo Cocha,
BS: 7 UWOMmS 25925 upper course of Rio Itaya, San Antonio, alt. 145 m., L.
Williams 3345; Paraiso, alt. 145 m., L. Williams 3364.

260 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxrx

Herrania nitida is the most widespread species in the genus, extending
from the eastern slopes of the Andes in Colombia, Ecuador and Peru
throughout the entire Amazon basin, being especially concentrated in the
western half. As should be expected, therefore, there is evident considerable
variation within the concept. It is undoubtedly the most variable species
of Herrania, and further collections and studies will probably indicate
that some of the variations are deserving of taxonomic recognition. To
date, however, I have been able to separate out only one variant which
is recognized now as a forma.

One can think of Herrania nitida as a kind of evolutionary center. It is
in itself most distinct from all other species, but in certain respects some
of its variants approach the variants of other species. Its foliage resembles
that of Herrania albiflora and H. purpurea to a remarkable degree, but
there are no floral or fruiting characters which point to even a remote
relationship. Herrania nitida has often been confused with H. Mariae,
usually because, in spite of its specific epithet, it has variants with leaflets
softly tomentose (even though always sparsely so), especially in the central
and eastern parts of the Amazon Valley. Indeed, one collection (Krukoff
6085) is so intermediate between these species that I have indicated (under
Herrania Mariae) that it may possibly represent a hybrid of the two. Very
significant floral characters (coloration, size and shape of the staminodes,
form of the style, and placement of the anthers) serve easily to set the two
apart; and the peculiar rhomboid-obovate leaflets of Herrania Mariae
contrast strikingly with the lanceolate-oblong leaflets of H. nitida. We
can see from the similarity in the fruit, however, that these two species
are more closely allied than one would suspect from a study of the floral
and vegetative parts.

The flower of Herrania nitida is one of the smallest and most delicate
in the genus. Usually there is a general tendency in the staminodes and
even in the petals for a yellowish tinge, with nerves which are the more
conspicuous in these parts because of their dark purplish color against the
yellow. The ligules are almost always yellowish or white in the upper half,
but there are variants which have very dark scarlet ligules. The leaflets
of H. nitida are generally drooping, a characteristic which I have not seen
commonly in other species.

Herrania nitida prefers well-drained sloping soil, usually of a semi-
lateritic consistency, and is rarely found where the annual flood of the
rivers remains long enough and becomes deep enough to produce a drown-
ing effect. It is most often rather abundant in the areas where it occurs.

In 1950, when I reduced Herrania aspera to a varietal status under H.
nitida, I pointed out that there has been confusion in the presentation of
this concept since 1857, when Karsten stated that it was found “‘in vallis
Orenocensis marginibus ad pedem Andium bogotensium meridensiumque

. et littora fluminis Magdalenae’’. It was difficult to accept the occur-
rence of the same species on both sides of the great Andean cordillera, and
Karsten’s assertion may have been based on the study of a sterile specimen
of Herrania albiflora Goudot. Later, Triana and Planchon (Prodr. Novo-

1958] SCHULTES, SYNOPSIS OF HERRANIA 261

Granat. 209. 1862) erroneously reduced Herrania aspera to synonymy under
H. pulcherrima Goudot. A further confusion resulted with my treatment
of Herrania aspera as a variety of H. nitida.

A study of the fruit and flowers of the Macarena material indicates that
this plant, which represents a hitherto undescribed concept and which has,
in part, been referred in the past to H. aspera, bears little relationship to
Herrania pulcherrima and none to H. nitida. The confusion which has re-
sulted in the past appears to be directly a result of two factors: incomplete
material and Karsten’s failure to cite a definite specimen which we could
take as a type. The presumed type of Herrania aspera has been taken as
Triana 5333 from the llanos of Villavicencio. In view of the material now
available and of field studies in the Macarena not far from the type locality
of Herrania aspera, I am now reducing H. aspera to synonymy under H.
nitida and am describing H. tomentella to accommodate the soft-pilose,
large-leaved plant which is common in the western part of the Llanos and
in the Macarena.

Herrania atrorubens is also herewith reduced to synonymy under 4.
nitida. When Huber described Herrania atrorubens, he cited his collection
7935 from the Alto Amazonas of Brazil as the type and only material of
the concept. He stated that it differed from Herrania Mariae in being
smaller and in having dark red flowers. In 1944, in treating this binomial
(Caldasia 2: 329. 1944), I wrote, “I have been unable to examine the type
of this concept. Without typical material, I have found it impossible to
estimate its validity as a species, but it would seem that the colour charac-
ter alone would hardly suffice for the creation of a new specific concept.”

Now, having completed an extensive study of the classical material of
the genus, I have been unable to locate the type of Herrania atrorubens.
One would expect it to be preserved in the Museu Goeldi in Belém do Para
or in the Herbier Boissier in Geneva, but a search in these two institutions,
as well as in other principal Brazilian and European herbaria has not
uncovered Huber’s material.

In the light of the experience gained during the study of a wide range
of material, it would seem from an evaluation of the meager characters
given by Huber and from the geographical data given for the type collection
that Herrania atrorubens may be reduced safely to synonymy under H.
nitida.

12a. Herrania nitida (Poepp.) R. E. Schultes f. sphenophylla R. E.
Schultes, Bot. Mus. Leafl. Harvard Univ. 14: 131. 1950.
Herrania ee Cae R. E. Schultes var. sphenophylla R. E. Schultes,
alda a2. 1
ae ee ees K. Schum. f. minor Diels, Notizbl. 15: 48. 1940.

DISTRIBUTION: Western part of the Amazon Valley.

A small tree which differs from Herrania nitida principally in having
smaller and lanceolate-elliptic or very narrowly obovate leaflets, in having
a fewer-flowered inflorescence, in having flowers which are usually larger

262 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxrx

and redder (the ligules usually entirely red), in having a smaller fruit
(7.5-8 cm. long, 4 cm. in diameter), and in being humbler in stature.

Brazil. AMAzonas: Rio Jutahy, See Jununenes, poles 21040. Colombia.
Putumayo: Umbria, alt. 325 m., Klug 1853. AMAzoN : Trapeécio Amazonico,
Rio Loretoyacu, Black & Schultes ee 331. Peru. ene ‘Rio Ucayali. Tess-
mann 3287; Guamitanacocha, Rio Mazan, alt. 100-125 m., Schunke 45 (Type).

The very striking difference in size between the leaves and nearly all
other parts of Herrania nitida and H. nitida {. sphenophylla cannot be
laid to ecological variation. When this difference was first noted (on the
basis of Schunke 45) it was recognized as varietal. Later studies in the
held, however, as well as information obtained through an examination of
additional collections, indicate that the concept is probably better treated
as a form, since little other than color and size differences, constant though
they be, are evident.

In 1940, Diels described Theobroma Mariae f. minor from the Rio
Pastaza in eastern Ecuador. The type specimen seems to have been
destroyed during the recent war. If we may judge from the few characters
given in Diels’ original description and from geographical distribution,
it might be safe to assume that the concept represented either Herrania
nitida or, more probably, H. nitida f. sphenophylla. In 1943, I pointed
out this possibility (Caldasia 2: 332. 1944), stating: “I have been unable
to examine herbarium material or photographs of this form, and, until an
opportunity to do so presents itself, I shall be unable to treat it critically.
In most of the characters enumerated in the original description, it would
seem that . . . it approaches Herrania nitida var. sphenophylla, although,
of course, no mention is made of the fundamentally important character
of leaf-pilosity.”

13. Herrania nycterodendron R. E. Schultes, Caldasia 2: 21. tt. pag.
22, 26. 1943; Bot. Mus. Leafl. Harvard Univ. 14: ¢. 35. 1950.

DIsTRIBUTION: Westernmost parts of the Amazon Valley.

Small, slender, graceful tree up to 25 feet tall, the trunk usually simple,
erect, terete, apically leafy, up to 8-9 cm. in diameter at the base, covered
with an ashy-yellowish, scrobiculate bark. Leaves large, 7-digitate, stipu-
late, very long petiolate. Stipules caducous, linear, acute, 2.5-4 cm. long,
more or less 3 mm. wide, dry, tomentose. Petiole strong, terete but very
obscurely sulcate, basally slightly swollen, subferruginous, extremely
densely and softly tomentose, up to about 60 cm. in length, 8-9 mm. in
diameter. Leaflets sessile, anequal lanceolate-oblong, with a rather acute
cusp up to 2 cm. long, basally long- and gently decurrent-attenuate, entire
(or minutely and obscurely subundulate) but often very conspicuously
armed with the hirsute spinule (up to 1 mm. long) formed by the pro-
longation of the veins; firmly chartaceous or papyraceous, the central
leaflets up to 60 cm. long, 22 cm. wide, above dark green and subnitid,
subglabrous or very sparsely and minutely hirsute, minutely tomentulose

1958] SCHULTES, SYNOPSIS OF HERRANIA 263

along the principal veins, beneath pale green, softly stellate-pilose, very
densely and softly ferruginous-tomentose along the main nerves; the veins
prominently raised on both surfaces but especially so beneath. Inflores-
cence fasciculate, up to 40-flowered. Flowers cauline, long-pedicellate, in
contracted racemes on the lower portions of the trunk. Pedicels very
slender, appressed-tomentose, articulate, 2-2.5 cm. long, basally with a
short, linear, apically acute bract which is densely tomentose and 3 mm.
long. Bud subglobose, large, about 7-8 mm. in diameter, densely and
minutely stellate-puberulent, brown. Calyx subcymbiform, divided almost
to the base. Sepals 3, strongly unequal, thick, brownish-purple, valvate
in the bud, externally minutely stellate-pilose, internally very minutely
puberulent; the outer sepal broadly rotund-obovate, apically rotund-obtuse,
entire, 19 mm. long, 15 mm. wide; the inner 2 elliptic, entire, apically
subacute, about 15 mm. long, 7-8 mm. wide. Petals 5, basally sessile,
widely rotund, concave, apically strongly cucullate, 10 mm. long, 8-9 mm.
wide, glabrous, muricate-papillose on both surfaces but especially so on the
outer, pale yellow with 7 purple nerves, ligulate. Ligules filiform, hanging,
membranaceous, glabrous on both surfaces but basally minutely granulose,
2 mm. wide at base, up to 90-100 mm. long, dark purplish with prominent
black nerves. Staminal tube 5-parted; stamens alternately 2- and 4-
antheriferous, the filaments glabrous, slender and free. Staminodes con-
spicuous, petaloid, reflexed, lanceolate-elliptic, acute, entire, 19-21 m
long, 7-9 mm. wide, muricate-granulose on both surfaces. Ovary sessile,
ellipsoid, 10-costate and 5-locular, densely pilose, 3 mm. long, 2 mm. in
diameter, yellow. Style filiform, glabrous, apically profoundly divided into
5 parts. Stigmas 5, thick. Fruit ellipsoid, 10-12 cm. long, 4-5 cm. in
diameter, apically long and gradually attenuate, often slightly constricted
near the tip, apically acute or often rotund-obtuse, basally indented and
pedunculate (peduncle woody, up to 4-5 cm. long, 3-4 mm. in diameter),
with an extremely dense and minute velvety-stellate indumentum on all
parts, without stinging hairs, very profoundly 10-costate, with 5 thick
and strongly blunt-rounded primary ribs and 5 similar but smaller second-
ary ribs; pericarp thick-crassulent, sublignose, ashy-yellow when ripe.
Seeds up to about 100, triangular or triangrlar-ovate in outline, flattened,
9mm. X 8mm. X 2 mm. thick, in a white pulp.

Brazil. Amazonas: Rio Solimées, Fonte Boa, Frodes 20578.

Colombia. Amazonas: Trapécio Amazonico, Rio Amazonas, Leticia, Schultes
6017; interior regions of Trapécio Amazénico, between Amazon and Putumayo
watersheds, alt. about 400 m., Schultes 6777; path near Quebrada Agua Negra
(headwaters of Rio Hamacayacu), Black & Schultes 46-389; Rio Loretoyacu,
Schultes 6335; Rio Caqueta, La Pedrera, Anglo-Colomb. Cacao Coll. Exped.
(Baker & Cope) 14, 20. CaquretA: Rio Caucaya, Laguna Primavera, Anglo-
Colomb. Cacao Coll. Exped. (Cope & Holliday) 93; Rio Caguan, Cartagena,
Anglo-Colomb. Cacao Coll. Exped. (Cope & Holliday) 108; Camp Two, Anglo-
Colomb, Cacao Coll. Exped. (Cope & Holliday) 110; Camp Three, Axglo-
Colomb. Cacao Coll. Exped. (Cope & Holliday) 113; Camp Four, Anglo-Colomb.
Cacao Coll. Exped. (Cope & Holliday) 109; Rio Caqueta, at confluence with

264 JOURNAL OF THE ARNOLD ARBORETUM _[vot. xxx1x

Rio Caguan, Anglo-Colomb. Cacao Coll. Exped. (Cope & Holliday) 121; Camp
Six, Anglo-Colomb. Cacao Coll. Exped. (Cope & ue 128; Piedra Blanca,
ee Ae oe Coll. Exped. (Cope & Holliday) 13

. Loreto: Rio Putumayo, Remanso, alt. 180 m., a 4011 (Type);
Ee es . Rio Zubineta, alt. 200 m., Klub 2069: “Corbata”, opposite
Isla Salamanca, alt. 180 m., Schultes 4012; near Caucaya, ‘Anglo-Colomb. Cacao
Coll. Exped. (Cope & Holliday) 97; Rio Amazonas, Mishayacu, near Iquitos,
alt. 100 m., Klug 1588

Herrania nycterodendron most closely resembles H. Mariae and H.
Cuatrecasana. It is immediately set apart from these species, however, by
its curious type of fruit. The fruit of Herrania nycterodendron has a dry,
somewhat coarse and fibrous rind which is covered completely with a soft
indument of velvety hairs; it lacks the stinging hairs which are usually
present in this genus. The ribs of the fruit of Herrania nycterodendron are
broad and rounded with deep furrows. The fruit is apically much more
bluntly rounded, in most cases, than is that of related species.

Froes 20578 is a sterile collection which has been referred to Herrania
nycterodendron with some reserve

The natives of the Peruvian bank of the Rio Putumayo near Remanso,
Isla Salamanca, refer to Herrania nycterodendron as “bat-tree” or “tree
of the bats”. This curious common name may be due, as several natives
explained to me, to the fact that the soft, velvety indument of the fruit
feels like the fut of small bats which are common in the vicinity. It may
also be due to the fact that the fruits cluster on the basal portions of the
stem in such a manner as to suggest bats which are accustomed to pass the
day hanging from the lower parts of the trunks of small trees in the dark
forests. The Witoto name, mu-sé’-na, is also applied to the marraca (Theo-
broma glauca Karst.), and mu-sé-ge-ke, the diminutive, is very commonly
used to refer to Herrania nycterodendron and possibly also to H. Mariae
var. putumayonis.

14. Herrania oe Goudot,..Ann.. Set. Nat: JIL. 22 232, ¢. 3.
figs. 11, 12. 1844

Theobroma pulcherrima (Goudot) De Wildeman,’ Pl. Trop. Grande Cult. 89,
1902.

DISTRIBUTION: The mountainous regions of Central Colombia.

Small tree up to 15-24 feet tall, with a simple, columnar trunk 15-25
cm. in diameter. Bark thin, ashy- brown, scrobiculate, glabrous except near
the apex where it is covered with a ferruginous indument. Branchlets
villose, obscurely sulcate, Leaves 10-15, large, digitate, very long-petiolate,
5—7- foliate. Petioles robust, terete or ‘suleate, very densely ferruginous-
villose, as long as the leaflets, 7-10 mm. in diameter. Leaflets sessile, un-
equal, lanceolate-ovate, marginally coarsely and regularly sinuate towards

" s combination has been made independently by later workers, apparently
unaware of the earlier publication. Cf. Pittier, Man. Pl. Usuales Venez. 147. 1926.

1958] SCHULTES, SYNOPSIS OF HERRANIA 265

the apex, acuminate, basally attenuate, very coriaceous, mostly 45-60 cm.
long (occasionally much longer), 19-35 cm. wide; above dark green,
shining, coarsely muricate or minutely subtuberculate, glabrous (or ex-
tremely remotely armed with caducous hairs), the nerves minutely but
densely tomentulose; beneath brownish green, densely and softly ferrugi-
nous stellate-pilose, softly and densely villose-sericeous along the veins.
Stipules caducous, linear, tomentose, 3 cm. long. Inflorescence fasciculate,
many-flowered (20-30). Pedicels 11 mm. long, 1.2 mm. in diameter,
scabrid-hirtellous and minutely stellate-pilosiusculous. Buds ovoid, densely
fulvo-tomentose, 10 X 6 mm. Flowers large, crimson-red. Calyx sub-
cymbiform. Sepals 3, subchartaceous, broadly ovate or elliptic-ovate, en-
tire, rounded or subacute, externally brownish red, stellate-tomentose and
stellate-puberulent, internally glabrous and crimson-red, up to 19-20 mm.
long, 12-13 mm. wide, valvate in the bud. Petals 5 or (usually) 6, sessile,
obovate, concave, strongly cucullate muricate-papillose or granulose on
both sides, but especially externally, longitudinally striate-veined, the 5
veins prominently purple or black internally, crimson or dark red, ligulate.
Ligules pendulous, 80-110 mm. long, 2-3 mm. wide, glabrous, strongly
marked with 3 dark red veins, basally strongly and abruptly contracted,
dark purplish red. Stamen tube 5-parted with alternately 2- and 3-anther-
iferous stamens and simple, short, free filaments. Staminodes very con-
spicuous, petaloid, lanceolate-elliptic, apically 3-fid, marginally entire,
muricate-granulose on both sides, dark purple-red, 23 mm. long, 7 mm.
wide. Pistil up to 7.5 mm. long. Style slender, simple yellow, 3.5 mm. long,
the stigmatic tip deeply 5-fid. Ovary sessile, 5-locular, ovoid, very densely
pilose, pale yellow, 4 mm. long, 3 mm. in diameter. Fruit ellipsoid, atten-
uate-acuminate, 10-costate, with 5 large primary and 5 smaller secondary
cultriform ribs, covered, especially along the ribs, with minute stinging
stellate hairs, the rind very thin when dried but crassulent in life, yellow
when ripe. Seeds probably about 50, compressed, triangular in outline, in
a sweet white pulp.

Colombia. META: Iraca, San Juan [de Arama], Llanos Orientales. Goudot
s. n. (Type). Boyacd: region of Mount Chapon, northwest of Bogota, El] Umbo
region, alt. 3,000 ft., Lawrence 437. CUNDINAMARCA: Municipio de El Penon,
Hacienda “Curiche,” alt. 1000 m., Jaramillo 202.

The type specimen of Herrania pulcherrima has always been thought to
be in Paris, but, as in the case of H. albiflora, there is material in Geneva
which may well be the true type from which Goudot’s description and
illustrations (at least of the flowers) were made. Goudot spoke of the
type plants as inhabiting the great forests situated between the Rios
Ariari and Guayabero, affluents of the upper Orinoco, in the Colombian
Llanos.

The Geneva material consists of two sheets. It is labelled in Goudot’s
hand, “Herrania pulcherrima mihi. An. Sc. Nat. 1844, Llanos del Orinoco,
pueblo d’Iraca, San Juan, Flos: Dec.” One sheet consists of a piece of
golden-tomentose stem about one foot long, a very young leaf, and young

266 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

capsules. One of the envelopes has the native Coreguaje Indian name (re-
ported by Goudot in the original description): “cacao cahouai — Llanos”’.
Another envelope, on the outside of which Goudot has written “C N. 2
theobroma affinis Herrania pulcherrima,” has a completely and beautifully
dissected flower, the separate parts glued to the inside of the envelope.
There can be no doubt but that Goudot made his drawing of the flower of
Herrania pulcherrima (loc. cit. t. 5, figs. 11, 12) from this same dissection.

The leaf which is preserved at Geneva could not have served as a basis
for Goudot’s excellent description, but a study of the material and the
description would seem to indicate that the Paris material represents that
from which the original description of the leaf was drawn.

It may be of value to publish a few notes on Goudot’s dissection of the
flower. The three sepals are laid flat, the very slightly puberulent inner
surface exposed. Two are rather broadly ovate, about 15 mm. long and
5 mm. wide (all measurements taken dry), apically rounded; the third,
somewhat elliptic, 18 mm. long and 4 mm. wide, apically bluntly pointed.
The five petals are all about equal, strongly cucullate, very densely muri-
cate-papillose or granulose externally, papillose internally in six longi-
tudinal lines, the ligules up to 90 mm. long, 2.5 mm. wide immediately
above the constriction at its junction with the petal. The staminodes are
lanceolate-elliptic, 15 mm. long, 4~4.5 mm. wide, muricate-granulose, and
apically so strongly trifid that the tip appears to be mucronate. This was
noted by Goudot when he described the staminodes as apically ‘“‘mucronés
et échancrés;” but, in his drawing, he indicated the tip as extremely acute.
The ovary is very densely yellowish tomentose.

It is unusual to find a species of Herrania which occurs both east and
west of the Andes, as well as in the valley between the several Andean chains
in Colombia. Yet that appears to be the distribution of Herrania pulcher-
rima, Goudot said that he had found it in the deep valleys of the eastern
Andean chain, near Savana-Grande and Payme where, however, it seemed
to be rare and isolated. I have seen no Goudot specimen from this locality,
but it is very significant, I think, to note that all earlier and a number of
the later collections were made not in the eastern Ianos but within the
Andean cordillera.

Vegetatively, Herrania pulcherrima can easily be confused (and has
been confused) with H. tomentella, a species growing in the eastern llanos
at the foothills of the Andes where the type of H. pulcherrima was collected.
The differences between these two species are discussed under Herrania
tomentella.

The earliest reference to Herrania pulcherrima is Eloy Valenzuela’s
minute description of the plant written in Mariquita in the Departamento
del Tolima, Colombia, in 1784, while he was engaged in the work of the
Mutis Botanical Expedition in New Granada. For historical reasons, this
description has been reproduced in full under the generic description at
the beginning of the synopsis.

In the collection of Mutis plates, there is a most strikingly beautiful and
accurate water-color of a section of Herrania pulcherrima in full flower.

1958] SCHULTES, SYNOPSIS OF HERRANIA 267

A number of diagnostic characters of this species are most clearly shown:
the congested, many-flowered inflorescences, the very abbreviated pedicels,
and the long and membranaceous ligules with alternate scarlet and whitish
bands. Of this colored plate, there are two copies in black and white. No
foliage seems to have been drawn. A search in the Mutis collection of
plants in Madrid has failed to produce a specimen of Herrania pulcherrima.

The specific name pulcherrima, meaning “very beautiful,” could not be
more appropriate. It recalls Valenzuela’s picturesque remark, which I
have used as a theme for this synopsis; that the flower of Herrania pulcher-
vima or cacao esquinado “could be considered as the greatest marvel of
the plant kingdom, and one can hardly believe that nature, as frugal and
simple as she is, would have used so many ribbons and so much ornamenta-
tion to adorn herself almost as ostentatiously as in the fashions”.

14a. Herrania pulcherrima Goudot var. pacifica R. E. Schultes, Bot.
Mus. Leafl. Harvard Univ. 14: 131. ¢. 28, lower fig. 1950.

noe ania pacifica ee Rev. Acad. Col. Ciénc. Exact. Fisic. Nat. 7. 27: 307.
. Nomen n
DISTRIBUTION: Pacific coastal a of Colombia and northern Ecuador
and the Gulf of Uraba in Colombia

A small tree up to 25 feet tall, differing from Herrania pulcherrima
chiefly in having strongly membranaceous leaflets which are minutely
stellate-pilose and not muricate or subtuberculate above; lateral leaflets
usually strongly oblique; flowers which are commonly much smaller, with
the petals and ligules yellow or white; and smaller fruit (11.5 cm. long,
7 cm. in diameter).

Colombia. ANtTrioguIA: north of Dabeiba, road to Turbo, Univ. Calif. 3rd
Bot. Exped. Andes 1942 (Metcalf & Cuatrecasas) 30173; near Guapa, 53 km.
south of Turbo, alt. about 50 m., Haught 4607; Uraba, Municipio de Mutata,
Villa Arteaga, ale about 150 fo Schultes & Cabrera 18707a. VALLE: Pacific
Coast, Rio Yurumangui, Caimanero, Cuatrecasas 16010; a Calima, Quebrada
La Brea, alt. 30-40 m., Schultes 7324. CuHocd: Rio San Juan, vicinity of
Palestina, alt. 0-30 m., Cuatrecasas 21337; Rio ee ies. Anglo-Colomb.
Cacao Coll, Exped. (Bartley & Holliday) 174. Ecuador. PICHINCHA: Santo
Domingo de los Colorados, alt. m., Acosta-Solis 10923; “foot of western
cordillera”, alt. 100 m., Rimbach a

This concept, as indicated by the varietal epithet, would seem to repre-
sent a western or Pacific coastal variant of Herrania pulcherrima, a species
which, in its typical form, is endemic to the Cordilleras of Colombia

15. Herrania purpurea (Pitt.) R. E. Schultes, Caldasia 2: 333. 1944;
Caldasia 3: 23. t. pag. 24, fig. 1, 2. 1944; Bot. Mus. Leafl. Harvard
Univ. 13: 282. 1949.

Theobroma purpureum Pitt., Fedde Rep. Sp. Nov. 13: 319. 1914; Stand.
Contr. U. S. Nat. Herb. 27: 262. ¢. 51. 1928; Standl., Contr. Arnold Arb.

268 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

5: 104. 1933; Standl., Field Mus. Nat. Hist. Bot. Ser. Publ. 392: 688. 1937;
Leon, Inst. Interam. ‘Ciuc. Agric. Bol. Técn. 2: 6. 1949

DISTRIBUTION: Costa Rica, Nicaragua, Panama and northwesternmost
Colombia

Small tree up to 10 feet tall, without branches along the trunk, except
near the apex, with grey bark becoming glabrous but densely yellow-villose
when young. Leaves usually 5-digitate, stipulate. Stipules linear, acute,
dark purple, more or less stellate-villose, caducous, up to 5 cm. long, 3 mm.
wide. Petioles terete, obscurely sulcate, somewhat ferruginous-villose,
strong, basally rather swollen, 30-45 cm. long. Petiolules strong, very
short, up to 5 mm. long. Leaflets unequal, obovate-oblong, basally cuneate,
apically broadly and obtusely acuminate, entire or very obscurely sinuate,
thin-chartaceous, glabrous above, sparsely and minutely stellate-villose
beneath (the nerves densely stellate-villose on both surfaces), almost as
long as the petioles, 22-35 cm. long, 6-13 cm. wide. Inflorescences fasci-
culate, 5—8-flowered. Flowers cauline, in contracted racemes on the lower
and middle portions of the trunk, dark purple, pedicellate. Pedicels terete,
articulate, brown-tomentose, about 4 mm. long, basally subtended by a
short, linear, naviculiform bract which is densely brown-tomentose exter-
nally, glabrous internally. Buds globose, 7-9 mm. in diameter, villose.
Calyx patelliform. Sepals 3 (rarely 4), united for half their length, broadly
ovate or (rarely) elliptic-ovate, obtuse, entire, 12 mm. long, 9 mm. wide,
brownish purple, glabrous within, ferruginous, densely stellate-tomentose
without. Petals 5, sessile, obovate, strongly cucullate, up to 8 mm. (fre-
quently less) long, 4 mm. wide, muricate-papillose on both sides but denser
externally (especially along the nerves), pale purple without, veins deep
purple within, longitudinally striate-nerved, 5-veined, ligulate. Ligules
linear, very narrowly lanceolate, basally emarginate, acute, hanging, dark
purple, about 15 mm. long, 1.8-2 mm. wide. Staminal tube 5-parted, short,

mm. long; stamens alternately 1- and 2-antheriferous; filaments simple,
short, free; anthers about 1.5 mm. long, longitudinally dehiscent, yellow.
Staminodes petaloid but not very conspicuous, ovate, apically acute, re-
flexed, densely muricate-granulose, red-purple, 9 mm. long, 8 mm. wide.
Ovary ovoid, 10-sulcate, villose, 2.5 mm. long, 1.3 mm. in diameter. Style
glabrous, 5-parted. Fruit not numerous, elliptic- ovoid, often irregularly
twisted, up to 9 cm. long, 5 cm. in diameter, apically rotund- obtuse, slightly
constricted near the apex, basally hardly indented, with a strong, com-
paratively long, peduncle, 10-costate, the 5 primary and the 5 secondary
ribs almost equal, blunt-rounded, 5 mm. and 4 mm. high, respectively, and
very densely armed with stinging stellate hairs, between the ribs striate-
fibrous and armed with stinging hairs; pericarp crassulent-coriaceous or
subligneous, yellow when ripe. Seeds 25, obtusely rhomboid, flattened,
about 1 cm. long, 1.3 cm. wide and up to 0. 6 cm. thick.

Colombia. ANTIoQuIA: Golfo de Uraba, region around Turbo, road between
Turbo and Rio Grande, Schultes 5754; Rio Micuri, Schultes 5755: Municipio
de Pavorandocito, outskirts of Pavorandocito, alt. 80 His Guitiérres. 2000; Mu-

1958] SCHULTES, SYNOPSIS OF HERRANIA 269

nicipio de Mutata, Villa Arteaga, alt. about 150 ft., Schultes & Cabrera 18653,
18593: Schultes & Lépez 10464; Anglo-Colomb. Cacao Coll. Exped. (Bartley
& Holliday) 164, 169, 170.

Costa Rica. Palmar, Rio Grande de Terraba, Pittier & Durand 3926, 6721;
Pacific Coast, Boca Culebra, alt. 50 m. Pittier 12158. Limon: La Colombiana
Farm of United Fruit Company, alt. about 70 m., Standley 36832; Finca Monte-
cristo, Rio Reventazén, below Cairo, alt. about 25 m., Standley & Valerio
48421, 48545, 48584; Hamburg Finca, Rio Reventazon, below Cairo, Standley
& Valerio 48792; hills above tramline, Los Negritos Farm near Rio Reventazon,
Dodge & Neverman 7178. Gotro Dutce: Playa Blanca, Valerio 461.

Nicaragua. Seemann s. n.

Panama. Monte Lirio, Hayes 398. CANAL ZonE: Near El Paraiso, alt. 30-100
m., Pittier 2574; forests along Rio Indio de Gatun, Maxon 4835; Valley of
Masambi, road to Las Cascadas Plantation, alt. 20-100 m., Pittier 2675; Barro
Colorado Island, Gattn Lake, Maxon, Harvey & Valentine 6804, Kenoyer 443,
Gatun Lake, Standley 31319; alt. 120 m., or less, Standley 40911; Armour
House to second bay north, Bangham 549, Bailey & Bailey 31, Shattuck 198;
near end of Fairchild Trail, Wetmore & Abbe 73; hills north of Frijoles, Stand-
ley 27434; Gamboa, Standley 28416; near Fort Randolph, Standley 28647;
Obispo, Standley 31722; near Madden Dam, Alston 8861. CHtrIgui: Progreso,
Cooper & Slater 283. Bocas DEL Toro: Laguna de Chiriqui and vicinity, Hart
96, von Wedel 1112; Changuinola Valley, Dunlap 448, von Wedel 976, 1721,
Lucas 2. Darien: trail between Pinogana and Yavisa, alt. 15 m., Allen 282.

ANAMA: vicinity of Cafia, alt. 900 m., Goldman 1974; Changuinola, Cooper
& Slater 12a; Changuinola and Sixaola, Rowlee & Stock 1029; Marraganti and
vicinity, alt. 10-200 ft., R. S. Williams 662.

Herrania purpurea is obviously most closely allied to H. albiflora, a
relationship which is discussed under H. albiflora. Both species are alone
in the genus in having a curious patelliform calyx and extremely short
ligules which give the flowers an entirely different appearance from those
of all other species which have a subcymbiform calyx, usually with very
long, filiform ligules. Herrania albiflora and H. purpurea, therefore, are
considered to form a distinct section of the genus.

Further study may indicate that Her .nia purpurea might better be
treated as a variety of H. albiflora; but, at the present state of our under-
standing, the two would seem to represent well established specific concepts
which geographically are sharply delineated.

Herrania purpurea is the only species of the genus known to occur
outside of South America. It has its main center of distribution in lower
Middle America but it is represented in the adjacent part of Antioquia
(and probably in the northern Chocé) — the northwesternmost corner of
Colombia.

The binomial Herrania purpurea was published as a nomen nudum in
the first edition of Thomas Belt’s “The Naturalist in Nicaragua” (1874,
p. 116). Belt wrote: “About here grows a cacao (Herrania purpurea)
differing from the cultivated species (Theobroma Cacao).” I have been
unable to discover a description of the plant or a publication of the bi-
nomial prior to 1874. In the preface of his book, Belt stated that “Prof.

270 JOURNAL OF THE ARNOLD ARBORETUM _[VvoL. xxxrIx

D. Oliver of Kew has kindly named for me some of the plants.” In the
collection of Herrania at Kew, I did not find any specimen from Nicaragua
collected prior to 1874 and annotated with this binomial. Dr. N. Y. Sand-
with of Kew has kindly searched through the archives and reports that he
can find nothing which might suggest that Oliver had published the bi-
nomial.

Pittier’s description of Theobroma purpureum was based upon a Pana-
manian collection, and he made no mention of a prior publication of this
specific epithet. There is no doubt that the binomial which Belt published
refers to the same concept which Pittier later and independently described
and for which he used the identical specific epithet. In accordance with
the International Code of Botanical Nomenclature, therefore, we must
consider Pittier’s Theobroma purpureum as the first valid use of the
specific epithet.

According to Pittier (Fedde Rep. Sp. Nov. 13: 319. 1914; Standley,
Field Mus. Nat. Hist. Bot. Ser. 18 (Publ. 392): 688, 1937), the Bribri
Indians of Costa Rica employ the roasted seeds for preparing a bitter drink.

16. Herrania tomentella R. E. Schultes, Bot. Mus. Leafl. Harvard Univ.
16% :205, 213; ¢.32, 1954.
Herrania nitida (Poepp.) R. E. Schultes var. slaty (Karst. & Tr. Tir)
R. E. Schultes, Bot. Mus. Leafl. Harvard Univ. 14: 130. 1950, a es
DIsTRIBUTION: Eastern foothills of the Andes in the Orinoco drainage
area of Colombia.

A small tree, slender and graceful, commonly up to 12 feet in height.
Trunk erect, about 3 inches in diameter, covered with blackish bark,
sparsely branched near the top or unbranched. Branches tomentose.
Branchlets densely villose, with golden-rust-colored and persistent hairs.
Leaves very large, digitate, 7-foliate, very long-petiolate. Petioles round,
somewhat constricted at the base, very densely and softly golden or fer-
ruginous, tomentellous, up to 60 cm. long, 9-10 mm. in diameter. Stipules
persistent, linear, densely rough-tomentellous, up to 3 cm. long, 2 mm.
wide. Leaflets sessile, oblanceolate or broadly lanceolate-ovate, erect,
strongly unequal, membranaceous to papyraceous, acuminate, basally
attenuate, the margin both regularly and lightly sinuate-dentate in the
upper half but especially towards the apex and everywhere armed with
cilia-like stellate hairs, 30-50 cm, long, 13—20 cm. wide, above rough to
the touch with sparse, single, brown hairs, beneath rather softly and densely
tomentellous with long golden-rust-colored stellate hairs. Inflorescence
fasciculate, relatively few-flowered, growing from the lower portion of the
trunk. Pedicels articulate, 7 mm. long, 1.5 mm. in diameter, densely stellate-
pilose. Buds globose, 15 mm. in diameter, densely stellate-pilose. Calyx
3-parted, divided almost to the base, subcymbiform. Sepals commonly
unequal, rather carnose in life, dark purplish, strongly valvate in the bud,
externally rather coarsely stellate-pilose, internally very minutely granu-
lose-pulverulent; the 2 interior sepals round-ovate, the margins entire,

1958] SCHULTES, SYNOPSIS OF HERRANIA 20h

apically perfectly rounded, about 14 mm. long, 10 mm. wide; the exterior
sepal usually triangular-elliptic, the margin entire, apically subacute, 13-14
mm, long, basally 6—7 mm. wide. Petals 5, basally sessile, obovate or ovate,
apically very strongly concave-cucullate, about 8 mm. long, 7 mm. (often
up to 8 mm.) wide, dark blood-red with purple nerves, externally minutely
muricate-verrucose, ligulate. Ligules linear, about 70 mm. long, basally
3 mm. wide, filiform near the apex, dark blood-red but near the tip pinkish.
Staminal tube 5-parted with stamens bearing 1 and 2 anthers alternately
and with short, free filaments. Staminodes petaloid, dark blood-red,
membranaceous, elliptic, marginally entire, acute, 14-15 mm. long, 6—7
mm. wide, somewhat verrucose on both surfaces. Fruits not numerous,
ellipsoid, up to 9 cm. long, 4 cm. in diameter, long-attenuate but near the
tip slightly constricted, the tip itself obtuse and frequently twisted, basally
not indented, pedunculate, with remnants of the persistent sepals; peduncle
articulate, 3 cm. long, 4 mm. in diameter, everywhere densely and very
minutely velvety-pilose, soft to the touch and without stinging hairs, very
deeply 10-costate, the 5 primary ribs thick and bluntly rounded, 8 mm
high, the 5 secondary ribs similar but smaller, 4-5 mm. high, transversely
rather fibrous-rugose, the pericarp thick, almost woody, reported to ripen
yellow. Seeds about 60, embedded in a white pulp.

A description of the pollen grain of Herrania tomentella is given under
the generic description at the beginning of this monograph.

ee Oe precise locality], Rocha s. n. Meta: Villavicencio, alt. 300
> Sprague 135; Sierra de la Macarena, Playa Bonita, alt. 4 400 m.,
Pee larobo & Fernindez 1420; Cafio Entrada, alt. 550 m., Philipson,
robo & Jaramillo 2199; Sabanas de San Juan de Avama, Rio @uejar. near
landing-feld Los Micos, alt. about 500 m., /drobo & Schultes 612, 721; path
between Rio Guéjar and Cafio Guapayita, alt. about 500-600 m., /drobo &
Schultes 787, 1192 (Type); Cafio Yerly, Schultes 11629; Sabanas de San Juan
de Arama, Rio Guéjar, Schultes 11821.

Herrania tomentella resembles, in its foliage, H. pulcherrima and H.
Cuatrecasana. It differs from the former in having a much more finely
sinuate margin, in having a smooth (instead of a rather muricate-sub-
tuberculate) upper surface, in being more finely tomentose beneath, and
in being membranaceous (rather than coriaceous) and generally smaller.
From the latter, it can be distinguished by differences in the shape and
margin of the leaflets: those of Herrania Cuatrecasana are conspicuously
long attenuate-decurrent towards the base and have very remotely and
obscurely crenate-denticulate margins.

In form of the fruit, Herrania tomentella approaches H. Cuatrecasana
more closely than H. pulcherrima. The capsule of Herrania pulcherrima
has strongly cultriform ribs with stinging hairs, whereas that of H. tomen-
tella has broadly rounded ribs without stinging hairs. Furthermore, floral
differences between H. pulcherrima and H. tomentella are marked, espe-
cially in the staminodes which are apically trifid in the former but acute
in the latter.

212 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

Although there are a number of resemblances between the capsule of
Herrania tomentella and that of H. Cuatrecasana, the soft indumentum
and lack of stinging hairs in the former are in sharp contrast to the condi-
tion in the latter where, except for stinging hairs along the ribs, the surface
is glabrous or glabrescent. There are likewise several floral differences.

The leaflets of Herrania tomentella are borne in a partly erect position.
This is also true of Herrania Cuatrecasana and H. pulcherrima and possi-
bly of all species which have a noticeably swollen callus at the base of the
leaflets. In this erect position of the leaflets, Herrania tomentella differs
strikingly in habit from the only other species known in the Macarena,
H. nitida, which has leaflets which tend to be rather reclinate.

A study of the fruit of the material from the Macarena has clarified a
confusion of long standing. Although in the past specimens of Herrania
tomentella have been referred to H. pulcherrima or to H. nitida (as H.
aspera or H. nitida var. aspera), a study of the capsule, until recently un-
known, shows conclusively that H. tomentella has its relationships in other
directions. The history of the confusion between Herrania tomentella and
H. aspera has been discussed in detail under the heading of H. nitida.

17. Herrania umbratica R. E. Schultes, Caldasia 2: 261. t. pag. 263,
figs. a-d. 1943; R. E. Schultes, Bot. Mus. Leafl. Harvard Univ. 17:
86. t. 24. 1955.

DISTRIBUTION: Department of Santander, Colombia.

Small, slender, graceful tree up to about 16 feet in height, with the
branches grouped at the top of the trunk. Trunk erect, up to 15-18 cm.
in diameter, the bark probably brownish black; the root long, fusiform.
The branchlets densely ferruginous-tomentose. Leaves very large, 7-digi-
tate, very long-petiolate, stipulate. Stipules membranaceous, 3—6 cm. long
(according to the collector). Petioles strong, sub-terete but obscurely
sulcate, slightly swollen near the base, rusty, densely but softly tomentose,
up to about 60-65 cm. long, basally 10 mm. and apically 4-5 mm. in
diameter. Leaflets unequal, sessile, papyraceous, lanceolate-oblong, rather
acutely cuspidate with a tip about 2 cm. long, basally subattenuate-
cuneate, entire; above dark green and almost glossy glabrous, minutely

nerves prominent and rather more densely stellate-pilose; the central
leaflet 55-60 cm. (according to the collector, 40-70 cm.) long, 20-22 cm.
wide; the lateral leaflets smaller. Inflorescence fasciculate, many-flowered.
Flowers cauline, arising from the lower part of the trunk in abbreviated
racemes, short-pedicellate. Buds globose in anthesis, 18-22 mm. in di-
ameter. Pedicels up to 5 mm. long, densely fulvo-tomentose, articulated
basally and subtended by a minute, linear bract. Buds globose, mostly
10-12 mm. in diameter. Calyx patelliform, obscurely 2-parted. Sepals
very fleshy, 2, connate most of their length, subequal, rotund-ovate, entire,
apically rounded, more or less 22 mm. long, 22 mm. wide, glabrous and

1958 | SCHULTES, SYNOPSIS OF HERRANIA Midi

purple within, without yellow-brown and very densely and minutely stellate-
pilose and sparsely and coarsely stellate pilose. Petals 5, sessile, thick,
blood-red, concave, obovate, mostly 9 mm. long, 8 mm. wide, apically
stongly cucullate, within with 5 thick-callused, purple, muricate-papillose
veins, glabrous between the nerves but near the thickened margin densely
muricate-papillose. Ligules linear, 19 mm. long, 2 mm. wide at base, the
base strongly and abruptly contracted, spirally twisted in the bud but in
flower erect, yellowish red, minutely granulose. Staminal tube 5-parted,
the stamens 2- and 4-antheriferous, with short, free, strongly flattened fila-
ments; anthers 2-locular, the locules 1.5 mm. long, 0.6 mm. wide, yellow.
Staminodia thick, conspicuously petaloid, yellowish, strongly deflexed,
hiding the petals and anthers, oblanceolate-elliptic, entire, apically sub-
acute, 20 mm. Iong, 10 mm. wide, densely papillose-granular on both sur-
faces. Ovary sessile, elongate-ovoid, 10-costate and 5-locular, reddish
golden, very densely stellate-pilose, 3 mm. in diameter. Style fleshy-terete,
simple, yellow, apically conspicuously 5-parted into a stigma, 1 mm. long.
Fruits numerous, up to 45 to a tree; elongate-ellipsoid, conspicuously
irregular, mostly 11-14.5 cm. (according to the collector, up to 17 cm.)
long, 4.5—5 cm. (according to the collector up to 8 cm.) in diameter, apically
rotund-obtuse, not constricted near the apex, basally obtuse (not indented)
and pedunculate (with a woody, articulated peduncle up to 10 mm. long,
4 mm. in diameter), with 10 subequal ribs, the 5 primary ribs thick, irregu-
lar, blunt-rounded, 6-8 mm. high, 5 mm. wide, the secondary ones similar
but rather smaller, about 5 mm. high, 5 mm. wide, somewhat striate-fibrous
between the ribs, very sparsely and rather grossly beset with simple, white,
probably stinging hairs up to 1.5 mm. long, lacking a velvety indumentum;
pericarp crassulent-leathery or subligneous, 3-4 mmi. thick; bright yellow
when ripe. Seeds 45, triangular or angular-ovate in outline, flattened, 13
mm. < 10 mm. < 3 mm., in a white pulp, measuring 18 mm. * 15 mm.
x 7 mm. with the pulp.

Colombia. SANTANDER: Municipio de Girén, region of Capitancitos, alt. 695
m., Ortiz Méndez s.n. (Type). NorTE DE SANTANDER: Rio Tibu, above Beltrania,
Anglo- -Colomb. Cacao Coll, Exped. (Bartley & Holliday) 179; Rio Oru, Anglo-
Colomb. Cacao Coll. Exped. (Bartley & Holliday) 180; Rio Nuevo, Anglo-
Colomb. Cacao Coll. Exped. (Bartley & Holliday) 182. Trinidad. Imperial
College of Tropical Agriculture, Baker s. n

When Herrania umbratica was originally described, the resemblance
of its fruit to that of H. nycterodendron led to the suspicion of some rela-
tionship between the two concepts. Further investigation, however, has
indicated that Herrania umbratica and H. albiflora are probably very
closely allied.

The vegetative differences between Herrania umbratica and H. albiflora
are slight. Herrania albiflora f. titanica would seem to represent, in some
respects, a link between the two species. Exact relationship, however,
cannot be established, until complete flowers of Herrania umbratica and
additional fruits of H. albiflora are found. In the present state of our

274 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xxxrx

knowledge, we may say that the fruit of Herrania umbratica, are in
some respects to that of H. nycterodendron, differs from the fruit of

albiflora (as represented in Goudot’s line drawing published =f the
original description of H. albiflora and as known for H. albiflora f.
titanica) in being irregularly contorted (instead of very regular); in having
the primary and secondary ribs nearly equal and so thick that there is
little flat intercostate area (instead of having the primary ribs twice or
more higher and much thicker than the secondaries and extensive flat
areas between the ribs); in being apically extremely thick and blunt (in-
stead of having a slender, somewhat attenuate tip which is but slightly
obtuse) ; and in having the peduncle much stouter than that of H. albiflora.

Further investigation and additional collections may possibly indicate
that Herrania umbratica would better be considered as a variety of H.
albiflora, but a complete understanding of the floral structure of H.
umbratica must be had before any definite decisions can be made in this
respect.

The collector of the type of Herrania umbratica records that this same
species is found in townships in the vicinity of Girén, the type locality:
San Vicente, Lebrija, Zapatoca, and Betulia. There are, however, no
specimens from these localities. Regarding the habitat of Herrania um-

bratica, Ortiz Méndez wrote in his field notes: “. . . loose sandy, sandy-
clay soils . . . the normal growth of the plant occurs in rather wet situa-
tions . . . it grows and develops in the shade of the trees which are known

in the region by the names juanblanco, canalete, guarumo, guamo, anaco,
barba de mono, cocotinajo, qualanday, etc.” Of diseases which attack
Herrania umbratica, he reported: “The trunk and branches are completely
healthy. There are sporadic cases of insect attack to the fruits, but these
attacks do not harm the seeds. Fungal attack is absolutely negative. The
leaves are attacked slightly by crisomélidos and minadores.

PLANTS OF UNCERTAIN POSITION

THEOBROMA MariaE (Mart.) K. Schum. f. minor Diels, Notizbl. 15: 48. 1940.
The possible position of this concept is discussed under Herrania nitida f
Ssphenophylla,

THEOBROMA MONTANA Goudot ex Bernoulli, Neue Denkschr. allg. Schweiz.
Gesell. gesam. Naturw. 24: 15. 1871, nomen nudum. Under the caption: “species
mihi ignotae,” Bernoulli published this name without a description and without
the citation of specimens. It may represent a species of Herrania, for Bernoulli
commented: ‘Vero similiter Herraniae species.

HERRANIA GUIANENSIS Sagot ex K. Schum. Mart. Fl. Bras. 12(3): 75. 1886.
Nomen in syn. French Guiana. “Karouany”, Sagot (?) s. . When K. Schu-
mann published as a synonym of Theobroma ens Willd. ex Spreng. Sagot’s
manuscript name Herrania guianensis, he cited Sagot 1206, a collection from
French Guiana consisting merely of flowers. In the Utrecht herbarium, I found
specimen number 000030 to be a collection of flowers only. They represent
Theobroma speciosum, but on the outside of the packet there is a handwritten

1958] SCHULTES, SYNOPSIS OF HERRANIA 275
nnotation: “Herrania guianensis Sagot”. This is probably part of the Sagot
collection cited by Schumann.
COMMON AND NATIVE NAMES KNOWN FOR THE SPECIES
OF HERRANIA

In compiling the following enumeration of names used for Herrania in Middle and

the country or countries in which the name is employed has been ee d. me of
the names reported are taken from the Indian languages, in which cases it has almost
always been possible to designate the specific tribe.

abare Venezuela: ale a ees H, lemniscata
a-no-kwa Colombia: (Kubeo ns) H. nitida
a-wa-ka-de-ro Colombia: (Xuripake es H. ni
awarivacabariyek Venezuela H. lemniscata
-ay-0 Colombia: (Makuna Indians) 4d. nitida
be-se-o-wa Colombia (Tanimuka Indians) 4H. nitida
boscacao Dutch Guiana Ht. pores
bur-oo-ma British Guiana: (Arawak H, lemniscata
Indians)
cacahuillo Pert H.n
cacahuio Peru H. abs {. sphenophylla
cacaita Venezuela H. albiflora
cacao cahoual Colombia pulcherrima
cacao cahouit Colombia H. pulcherrima
cacao cahousi Colombia H. pulcherrim
cacao caiman Colombia HA. Mariae, H. nitida
cacao canaludo Colombia, Ecuador H. Cuatrecasana
cacao cimarron Costa Rica H. purpurea
cacao cuadrado Colombia H. pulcherrima
cacao de andira Brazil H. Mariae var. putumayonis
cacao de ardilla Panama H. purpurea
cacao de chimbe Colombia, Pert H.nycterodendron
cacao de cintillas Colombia H., laciniifolia
cacao de macaco Brazil H. Camargoana
cacao de mico Costa Rica A. purpurea
Colombia H. Camargoana
cacao de monte Brazil H. Mariae
Colombia H. albiflora & f. titanica, H.
breviligulata, H. Cuatreca-
sana, H. Dugandii, H., lacinii-
folia, H. Mariae, H clero-
dendron, H. tomente
Ecuador H. balaensis, H. pulcherrima
var. see
Panama H. purpure
eru . nitida, i. nycterodendron
cacao de murcielago Colombia, Pert H. nycterodendron
cacao esquinado Colombia H. pulcherrima
cacao mani Panama H. purpurea
cacao montaras Colombia H. albiflora
cacao montaraz Colombia H. albiflora
cacao jacaré Brazil H. Camargoana, H. Mariae, H.

Mariae var. putumayonis

276 JOURNAL OF THE ARNOLD ARBORETUM _ |[voL, XxxIx

cacao quadrado Brazil H. Mariae
Colombia H. pulcherrima
cacao silvestre Colombia H. lacinufolia, H. Mariae var.

putumayonis, H. nitida
H. nitida, H. nycterodendron

cacao simarron Colombia Hf, albiflora
cacao symarron Colombia H. albiflor
cacaoito de monte Colombia H. nitida, i. pulcherrima
cacaorana Brazil A, iis
caca-t Brazil H, Maria
cacau de quina Brazil H. ee dees
acaul Brazil H. Mariae
cacau-jacaré Brazil H. Mariae
cacau-rana Brazil . Mariae
cahouit Colombia 1. pulcherrima
cha-te-ra Colombia, Peru: (Tikuna : nitida
Indians)

chocolatillo Panama HT. purpurea
coco del monte Panama H. purpurea
ee-so-pe-ke Brazil: (Tukano Indians) H, Camargoana
hee-ree-la-na-pee- Colombia: (Yukuna Indians) 4H. nitida

ta-re
he-me-ka- Colombia: (Taiwano Indians) JH. »itida
jo-kee- eee -yo-ke Colombia: (Kubeo Indians) H. nitida
ko-kee-ot-chu Colombia, Ecuador: (Kofan H. Cuatrecasana, H. nitida

Indians

-ra-ta Colombia: (Karijona Indians) JH. nitida
maipoilie doron Dutch Guiana: (Karib Indians) H. kanukuensis

doron

ma-mi-ree Colombia: (Kabuyari Indians) dH. nitida
maripoele kakaoeleo Dutch Guiana: (Karib Indians) H. kRanukuensis

matayaka Venezuela: (Maquiritare H. lemniscata
ndians
mi-to-ro-re Colombia: (Karijona om H. nitida
mu-se-ge-ke ene Pert: (Wito H. Mariae var. putumavonis,
H. nycterodendron
mu-se-na Colom Pert: (Witoto H. Mariae var. putumayonis, H.
India nycterodendro
o-so-pee-ko Brazil: (Tukano Indians H. Camargoana
o-yaw-pee-ka-ye Colombia: (Desano Indians) f nitida
palo de chimbe Colombia, Peru 1. nycterodendron
palo de murciélago Colombia, Peru 7 nycterodendron
pan y cacao Colombia H. albiflora
rus-u Panama: (Bribri Indians) rea
sacha cacao Colombia: (Inga Indians) H. breviligulata, H. Cuatreca-
sana
so-pee-ja-ke Colombia: (Gwanano Indians) H. nitida
tach-ko-au Colombia: (Mirafia Indians) dH. nitida
toot-choo Scape (Yuruti ae H. nitida
wild cacao Panama: Canal Zon Hi. purpurea

wild cacao British Guiana H. lemniscata

1958]

SCHULTES, SYNOPSIS OF HERRANIA

INDEX TO EXSICCATAE

or purposes of facility in consulting material of Herrania in our herbaria, the
following list summarizing the collections which have been consulted in the prepara-

tion of this synopsis is offered.
the collector.

Acosta-Solis 10923 (14a)

Allen 282 (15

Alston 8861 (15)

André K 26 (2)

Anglo-Colombian Cacao Collecting
pedition 78 (3); 35, 39, 45, 56 (4);

DAS) 2996, OF 10, 893

ea

4
164, 169, 170 (15); 179, 180, 182 Cie
arenes 2514 (10)
Aristeguieta 7598 (1)
Bailey s.n. (1)
Bailey & Bailey 31 (15)

5)

Black 47-1916 (11)
Black & Schultes 46-223

(11); 46-331
12a); 46-389 (13)
Bonpland 7580 (1)
Cooper & Slater 72a, 283 (15)
Cruz (de la) 3892 (10)
Cuatrecasas 11168 (5); 10742 (6); 16010,

21337 (14a)
Curran 135 (1)
Dodge & Nevermann 7178 (15)
Ducke 595 (11); s.1., 7618 (12)
Dunlap 448 (15)
Eggers 14362 (2)
Exped. pe Mutisii Novae-Granat.
GL) (9)
Forest ae British ae F 1764 (10)
Froées 21468, 21540, 22673 (3) ; 23003 Civ
20630, 21041 (11); er (12); 21040
(12a); 20578 (?13)
Garcia- Barriga 8375 (9), 14016 (12)
)

3759

Goldman 1974 (15

ee: 2111, 2565, 4101, 4117, 4126
ts.m. (1)3 sm. (9) 3 sm. (14)

Grass 10721 (12)

Hart

aan ee (la)
Hayes 398, 399 (15)
Hulk 26 (7)
Idrobo & Schultes 768, 791, 1325
612, 721, 787, 1192 (16)
n s.n. (10)
Jaramillo 202 (14)

(125;

Numbers in parentheses refer

The list is arranged alphabetically by the last name of

to the corresponding species in the text.

Kalbreyer 2047 (9)
Kenoyer 443 (15)
Killip 34247 (12)
Killip & Smith 27431, 28234 (12)
Klug 1853 (12a) ; 1588, 2069 (13
Krukoff’s 4th Exped. Bras. Amazon 4523

(11)
Krukoff’s

CL2
Lanjouw & Lindeman 2304 (7)
Lawrence 437 (14)
Lucas 2 (15)

Martius s.m2. (11)
0

5th Exped. Bras. Amazon 6085

Maxon 4835 (15)
Maxon, Harvey & Valentine 6804 (15)
Mexia 7328 (3)
Murga Pires 775, 115
Murca Pires & Black on ae 873 (12)
Myers 3371 ou
Ortiz Méndez ; )
Pennell 3799, a 4208 (1)
Pérez-Arbelaez 10303 (9)
Philipson et al 1420, 2199 (16)
Pittier 2574, 2675, 12158 (15)
Pittier & Durand 3926, 6721 (15)
)

)
Rowlee & ciaee 1029 (15)
Ruiz & Pavon (?) s.n. (12)
Rusby & Squires 252 (10)
Schomburgk s.n »)
Schultes 18639 (1); 18638 (2); 3342 (5);

6038 (6); 3478, 3670 (8); 6238, 6461,
6759, 8072 (11); 4010 (lla); 3405,
3698, 3730, 5351, 5359, 5491, 5529, 5685,

6146, 6147, 6149, 6192a, 6304, 6383
6640, 6878, 8129, 11627 (12); 4011,
4012, 6017, 6335, 6777 (13); 7324

(14a); 5754, 5755 (15); 11629, 11821

16).
sce Baker & Cabrera 18439, 18553
ere & Black 8286, 8377 (12)

Schultes & Cabrera 18720, 19082 (3);
18712, 18715, 18976, 19100 (5); 13628,

278

13630, 13632, 14343, 14537, 14880,
14882, 18933, 19284, 19665 (12); 187074
(14a) ; 18652, 18693 (12)

Schultes & Lopez 8758, 8759, 8762, 8763,
8956, 9144, 9162, 9205, 9240, 94166
9619, 9722, 9747, 9869 (4); 10464 (15).

Schultes & Murca Pires 8978, 9130 (4)

Schultes & Smith 2050 (3)

Schunke 45 )

Scolnik, Araque Molina & Barkley 195001

10

Seemann s.71. (15)
Shattuck 198 (15)

Sprague 135 (16)

Spruce 4969 (12)

Stahel & Gonggrijp 3015

Standley 27434, 28416,
31722, 36832, 40911 (15)

28647, 31319,

BoTANIcAL MUSEUM
VARD UNIVERSITY

JOURNAL OF THE ARNOLD ARBORETUM

[VOL. XXXIX

48792

Steyermark 60558 (10)

Tejera 268 (1)

Tessmann 4024 (12) ; 3287 (12a)

Traill 64 (12) ; 65 (11a)

Triana s.n., 5333, sn. (12)

Ule 5031 (11

Univ. Calif. 3rd Bot. Gard. Exped. Andes
30173 (14a)

Valerio 461 (15)

Von Wedel 976, 1112, 1721 (15)

Wetmore & Abbe 73 (15)

L. Williams 11339 (10) ; 2332, 2816, 2843,
3345, 3364 (12)

R.S. Williams 662 (15)

Collector unknown [Hort. Trinidad] s.n.
(1)

Standley & Valerio 48421, 48545, 48584,
(15)

Jour. ARNOLD ArB. VoL. XXXIX PLaTE I

emp ar feeds? Seep ab elaine -
reese UE «ets Sy ar df Hentz Rend of
Be fats lene ORT on 4/ Compe Pervensal ve we ae

con Phas Lif hn at racitn, Cer be la Tee coxkeza (Oe, e bom :

o Len eade & oniferee we cue oA eo alk

qua ire
ypue crifaio Laren oy we
oe a ae Bd eS > , SSS
Mor ner, Alreamas, be & ARS Ghee, pom atavel peapeee:
arin = Cenok: comune event | ales, + at/ch ew Der i

ids bewisccacio. Cectrerity —xe ok Dale leas To aa

| ab? sae nx ao Tx Posen eine ob EO
comm fas impeacon f ch, shee es obtusa! epi
4
clas Cxopenuon - Je <tnxanchan BeBe le read + abxen, er laura
SRinadedS Anchas, ap ae PS pipe Z nat pe Ke 7k 018) 4,
| J,

POON, Gui oe, ee Na "Ee boxdcs ©94_-

q\
* aot & s ie a a. oa ‘
de> Prema ras anche: yap a Bene Agr
‘ fppe . y,

j

Me woe ee va oh Sos ey ee
Vemas: Praratelas reetas ee as doo ) Las ent foleas, ie

| Orirarmter Betla- artes, Jubuta Bas,’ a5 ie a9 4 Sofas
Ke Bees gem Bed falle:; i ay , t,

x

= HL peat ah opines} aig? Deas, ae ¢ ren ans d Las”
Onl & 3% be oe ‘Bel ae Cat iad oy aicitle Ber. -
So Reg ican

BL, ay o. ir em umeo, bbe oa |

~ Sait hoe eee lesvanra?

a
del eens f° ed pre Cor 9 aed & Bve Suinelys beat y
Ga :
f ae ey to Coxtaceo, Colexde: oe Ce PC TA
pene , Omen & Ghhur ey Cennard, “Opa2o, bik Nese Drasede
Page of manuscript of Valenzuela’s Diario de la Expedicién Botdnica al Reino
Nueva Granada in which the detailed description of cacao esquinado (Her-
rania pulcherrima) is set down.

Forms of the leaflets of sundry species of Herrania

Piate III

Jour. ARNOLD ArB. VoL. XXXIX

ee]

deo.Tonv. ofl. EG. 4

!
i

Ann dee Seiene.nat.3* Serie.

HH? Bouliot ve

J Goudot dele.

1. Herrarua albiflora Gt

19 ae OE

Drawing of Herrania albiflora from Goudot’s original description of the genus

and species.

Flower and buds (left) and stipules (right) of Herrania albiflora.

‘wnof

-

“aay GIONAY

XIXXX “TOA

AI aLv1g

Jour. ARNoLD Ars. VoL. XXXIX PLATE V

HERRANIA
brevi ligulata
RAE. Schultes

Ws z
VE

me
AS

| Re RET PLEET EEN HO errr LTTE ae
%

Herrania breviligulata, Fic. 1. Leaf, * 1/3. Fic. 2. Flower, * 1/2. Fic. 3.
Petal, X 2. Fic. 4. Staminode and anthers, * 2. Fic. 5. Ovary and style, x 4.

Jour. ARNOLD Ars. VoL. XXXIX PLaTE VI

Fruits of Herrania breviligulata.

Jour. ARNOLD Ars. VoL. XX XIX PLaTE VII

HERRANIA Camargoane
RE. Schultes

CH
7

Hy

oy
by

Herrania Camargoana. Fic. 1. Leaf, * 1/4. Fic. 2. Flower, & 1/2. Fic. 3.
Petal, X 2. Fic. 4. Staminode fa anthers, xX 2. Fic. 5. Ovary and style, x 4.
Fic. 6. Fruit, & 1/2.

Jour. ARNOLD Ars, VoL. XXXIX Pirate VIII

Inflorescences of Herrania Camargoana.

Jour. ARNOLD Ars. VoL. XXXIX PLaTE IX

\ HERRANIA. Ranzkoensis
s¢ FE Sie

)

v

1

Vee + ne

a a |
a
ss i

Herrania kanukuensis. Fic. 1. Leaf, X 1/6. Fic. 2. Fruit, & 1/2. Fic. 3.
Portion of lower surface of leaflet, showing pilosity, x 4.

Jour. ARNoLD Ars. VoL, XXXIX PLATE X

HERRANIA #ofanorwmn
yj RE. Schuttes

SS 7)
> \ Wa
Ss “= yy Wipe

4, |

Myf

]

i ae

a

Herrania kofanorum. Fic. 1. Leaf, & 1/2. Fic. 2. Flower, X 1/2. Fic. 3.
Petal, X 2. Fic. 4. Staminode and anthers, X 2. Fic. 5. Ovary and style, X 4.

Jour. ARNOLD Ars. VoL. XXXIX PLATE XI

HERRANIA = lemniscata
(Schomburgk) WE. Schultes

co
a if pat) J

CA

\
fee
¥

/
———"—_

We :

fee i [a

Woe ! 1

\\y \ \\ :

NV Ney -
PLN \
\\
\\'
\

e a4 j

|
ON
Qs >

\ WW

c : VAS

—

Herrania lemniscata. Fic. 1. Leaflet, & 1/5. Fic. 2. Flower, X 1/3. Fic. 3.
Petal, & 2. Fic. 4. Staminode and anthers, X 2. Fic. 5. Ovary, X 2. Fic. 6.
Fruit, X 1/2. Fic. 7. Petioles and stipules. Fic. 8. Base of leaflets.

Jour. ARNOLD ArB. VoL, XXXIX PLATE XII

fe 1 7 7 a er, Ane aa oe
Tah, X11, Pepper fice Sea :

CO? MG 2 at >

Herrania lemniscata. Schomburgk’s field painting of Lightia lemniscata, Tab.
XLI in the Schomburgk collection of water-colors in the British Museum (Nat-
ural History)

Jour. ARNOLD ArB. VoL. XXXIX PEATE: ULI

Flowers and buds of Herrania Mariae var. putumayonis.

Jour. ARNOLD Ars. VoL. XXXIX PLATE XIV

Fruits of Herrania nitida.

Jour. ARNoLD Ars. VoL. XXXIX PLaTE XV
|

we
\s

i HERRANIA nycterodendrom ‘ |

— ———
oat = =

——
-———

ae,

a

“ALO

Herrania nycterodendron. Fic. 1. Leaf, X 1/4. Fic. 2. Flower, X 1/2. Fic. 3.
Staminode and anthers, X 2. Fic. 4. Ovary and style, x 4. Fic. 5. Fruit, x
1/2. Fic. 6. Portion of lower surface of leaf, showing pilosity, x 4.

Jour. ARNOLD ArB. VoL. XXXIX PLATE XVI

HERRANIA
pulcherrima

Goudot

Herrania pulcherrima. Fic. 1. Leaf, x 1/4. Fic. 2. Flower, & 1/2. Fic. 3.
Petal, X 2. Fic. 4. Staminode and anthers, * 2.

Jour. ArNotp Ars. VoL. XXXIX Pirate XVII

HERRANIA purpurea
(Pittier) RE Schultes

N\

hi a

Ce

Herrania purpurea. Fic. 1. Leaf, X 1/3. Fic. 2. Flower, X 1/2. Fic. 3.
, & 2, Fic. 4. Staminode and anthers, X 2. Fic. 5. Fruits, * 1/2.

296 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xxxrx

THE GENERA OF THE WOODY RANALES IN THE
SOUTHEASTERN UNITED STATES

CaRROLL E. Woop, Jr.

THE TREATMENTS PRESENTED BELOW of the sixteen genera of seven
woody ranalian families which occur in the southeastern United States
have been prepared as part of a generic study of the seed plants of that
area. This work has been undertaken as a joint project of the Gray Herbar-
ium and the Arnold Arboretum and has been made possible through the
interest and support of George R. Cooley and through a grant from the
National Science Foundation. In view of the co-operation and interest
others have shown in this undertaking it seems worth while to publish our
treatments of at least some of the families in advance of the completed
work. In this way some of the material brought together in the course of
these studies will be made available, and we should hope to have con-
structive criticisms from other botanists as the work progresses.

In attempting a generic treatment of the approximately 1300 genera
of seed plants known to occur within the area bounded by and including
North Carolina and Tennessee, on the north, and Arkansas and Louisiana, on
the west, the objectives are toward a review and reorganization of familial
and generic lines (often obscured in Small’s Manual of the Southeastern
Flora) and, especially, toward bringing together at least a part of the vast
botanical literature which bears upon the plants of this rich area. The
work is being done by taxonomists and is both taxonomic and floristic,
but the approach, as well as the scope, is intended to be somewhat broader
than is usual in a regional manual. The basic scheme is biological with the
intent of including material from all branches of botany and of under-
scoring the biosystematic aspects of each genus and family, insofar as
possible. In such an approach more problems may be raised than are
solved but, in at least some instances, some of the difficulties which must
be resolved before the plants of our area can be understood adequately
become evident when the literature of a particular genus or family is
brought together. Of course, with the existing information such a large
goal is impossible for all genera (or even most) but a biological or bio-
systematic viewpoint is that which we are attempting to maintain through-
out these studies.

The difficulties and weaknesses of this undertaking are apparent to no
one more clearly than to those of us who have planned and worked on this
project. More than a year and a half were spent on the tedious but basic
chore of drawing together a file of more than 50,000 references arranged
by families and genera which provided the starting point for the work.
The large number of references and their wide distribution in the botanical
literature of the world point up the almost incredible amount of informa-
tion which is rather effectively “lost” to most botanists and even to authori-

1958 | WOOD, GENERA OF WOODY RANALES 297

ties on particular groups of plants. The amount of material published
about some genera of plants (cf. Magnolia and Liriodendron) is very large,
while that concerning others is so scattered and fragmentary that it seems
almost impossible to retrace. The matters of locating references, attempt-
ing to deal with information from a number of fields, culling, and trying
to synthesize in a few choice words the “essence” of the genus sometimes
seems to be (and probably is) a nearly hopeless task. The degree of success
undoubtedly varies considerably from one group to another, and both sins
of omission and commission will be evident. However, the work is absorb-
ing and it is difficult to set aside one group for another, leaving behind
many unsolved problems. It is to be hoped that others may take an interest
in some of these problems and that the notes and references included will
prove to be useful to students and researchers. One thing is certain: there
are enough botanical problems in the Southeastern flora to supply all
possible workers for a long time to come.

Although the general scheme being followed throughout this work will
be apparent in the treatments below, a few explanatory comments will be
apropos. It should be noted especially that the descriptions of families
and genera are based mainly upon the species which occur in the south-
eastern United States and are not necessarily wholly applicable to those
beyond this area. However, additional information which may provide a
more balanced concept of the genus may sometimes be included in brackets.
Although each description is regional, the concept of the genus is broader,
with an attempt being made to delimit the group more in terms of all its
species. Such an approach is essential, for many genera are represented
in our area by only a few outlying species which often belong to different
sections or subgenera. Were the viewpoint essentially regional these might
be placed in different genera to the obscuring of true interrelationships
(and to the ensuing confusion of all). This viewpoint is, in general, a con-
servative one which stresses similarities, rather than differences. In many
instances it would seem far better at this stage of our knowledge to point
out a problem than to attempt a solution which may only cause ultimate
confusion.

The notes included vary widely with the group. Not all branches of
botany will be found to be equally well represented: the training and back-
ground of the present investigators leads to the emphasis of taxonomic and
biosystematic materials, although an attempt has been made to include as
much other information as possible. Pathology and palaeobotany, each a
large field with its own special and extensive literature, are likely to be rep-
resented least well. Limitations of space prohibit the inclusion of the source
of each item of information, but a series of selected references will provide
many of these. References included are primarily to journals. Many stand-
ard texts have been consulted but, rather than to cite each of these re-
peatedly, it is planned to include a list in the final work. Many references
have been annotated as to content, especially when the title is not self-
explanatory. Abbreviations used for journals follow the clear and useful
general principles followed by Schwarten and Rickett (Bull. Torrey Bot.

298 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxix

Club 74: 348-356. 1947, and a much amplified list now in press) which
are in accordance with the recommendations of the Madison Botanical
Congress of 1893 and the International Code of Botanical Nomenclature,
1956. Those references which we have not seen are followed by an aster-
isk (*),

The illustrations are by Dorothy H. Marsh (Mrs. Stephen Marsh) who
has worked meticulously in depicting the details upon which the accuracy
of scientific illustrations depends. Since the adequate illustration of more
than a thousand genera is likely to be quite impossible, the drawings have
been planned on the assumption that, even though few in number, illustra-
tions which provide some insight into the details of the plant are far more
desirable in a work at the generic level than a larger number of “recogni-
tion” drawings. These more detailed drawings, which eventually will
represent genera scattered throughout the whole range of families, are being
made mostly from fresh or preserved materials as these become available.
A number of kind individuals have been most helpful in their efforts in
this direction, and the living collections of the Arnold Arboretum, which
include many woody plants of authentic southern origin, have provided
invaluable material.

The project is under the direction of Dr. Reed C. Rollins and the writer.
Dr. Kenneth A. Wilson is working with us at the present time and we have
had as our other excellent collaborators Dr. R. B. Channell, now of Van-
derbilt University, and Dr. C. W. James, now of the University of Georgia.
All three have worked conscientiously on the many tasks directly concerned
with the studies on the Southeastern flora: the preparation of files of refer-
ences, the identification of specimens, the preparation of generic treatments
and the supervision of drawings of groups which they have studied. The
basic plan of the generic treatments has been modified through the ideas,
trials and efforts of all of us. In these studies a great many friends and
colleagues from the staffs of our respective institutions, from many parts
of the United States, and from other countries, as well, have brought us
appreciated assistance. The appropriate time has not yet come, nor would
there be here sufficient space to mention each one individually, but to each
I am personally most grateful for his help.

The order Ranales, including as it does the most primitive of known
living angiosperms, is one of particular interest from an evolutionary point
of view. In recent years, the studies of Professor I. W. Bailey and his
numerous collaborators and students have elucidated many aspects of
the woody members of this group and have stimulated a renewed interest
not only in these plants as a group but in the “complete” approach to
problems of the interrelationships of higher categories through the use of
information from all parts of the plant, rather than that from a single organ
obtained by a single discipline. This approach not only has pointed out the
many remarkable primitive structures in various members of this order
but has emphasized again and again the very different rates at which
structures or organs of a plant may have evolved. Hence, any evolutionary

1958 | WOOD, GENERA OF WOODY RANALES 299

arrangement of families must be an attempt to sum up the degree of rela-
tionship with other groups as well as the over-all level of specialization.

The ranalian families included here are those characterized by their
predominantly woody habit and by the possession of the characteristic
ranalian ethereal oil cells in the tissues of the plant. In our flora this group
includes the Magnoliaceae, Annonaceae, Illiciaceae, Schisandraceae, Canel-
laceae, Calycanthaceae, and Lauraceae. In spite of investigations on these
and other ranalian families by numerous authors, a number of families
remain to be studied carefully. The interrelationships of all the fainilies of
the order are still far from certain, although various related groups have
been pointed out. Thus, the Degeneriaceae, Magnoliaceae, Annonaceae,
and Himantandraceae form a group of related families within the order;
Illiciaceae and Schisandraceae another; Austrobaileyaceae, Trimeniaceae,
Amborellaceae, Monimiaceae, Calycanthaceae, Gomortegaceae, Lauraceae,
and Hernandiaceae yet another. The relationships of the Canellaceae are
undoubtedly with the ranalian complex (presumably with that group of
families having monocolpate pollen and tri-lacunar nodes — perhaps
Myristicaceae), instead of with the Parietales where they are placed in the
Englerian system, but exactly where remains to be seen. I have not at-
tempted to deal at all with the matter of splitting the Ranales (sensu lato)
into other orders, retaining all of these families together, instead, and only
arranging these groups in this rough way pending the outcome of studies
now under way at several institutions.

In connection with these ranalian families Professor I. W. Bailey has
given freely and most helpfully of his knowledge of these more primitive
angiosperms and Dr. C. E. Kobuski has most kindly read the entire manu-
script with a practiced editorial eye. The flowering material used in the
illustration of Ilicium floridanum came through the kindness of Mrs. J.
Norman Henry from plants cultivated at the Henry Foundation, Gladwyne,
Pennsylvania, and the fruiting specimens through that of Dr. R. B. Chan-
nell from Gallman, Mississippi.

MAGNOLIACEAE (Macnorta FAmMity)

Deciduous or evergreen trees or shrubs with simple, alternate, stipulate
leaves with pinnate venation, the stipules inclosing the bud and leaving
conspicuous encircling scars at each node. Flowers solitary, terminal [or
axillary], perfect [except Kmeria], all parts free [in ours]. Perianth of
9-15 green, yellow or white tepals in whorls of threes, the outermost whorl
sometimes partially differentiated as a “calyx.” Semen numerous, spirally
arranged on the elongated receptacle below the carpels, [7—]3(2 or 1)-
veined, often poorly differentiated into “anther” and “filament” with four
microsporangia (often confluent as two), dehiscing longitudinally; pollen
ellipsoidal, monocolpate, the germinal furrow distal. Gynoecium of num-
erous conduplicate carpels, free [in ours], spirally arranged on the upper
part of the receptacle and so closely imbricated and packed as to appear

300 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

syncarpous |[in ours], the style elongate, vascularized, the ovaries 1-loculed,
in both our genera with 2 anatropous 2-integumented ovules at the inner
angle of the locule, back to back with the funiculi adjacent. Fruit cone-
like [in ours], composed of the imbricated follicles or samaras which are
clearly distinct at maturity; seeds 2 in each carpel, each with abundant
endosperm and a minute embryo; embryo sac development normal in all
known cases; basic chromosome-number 19 throughout the family.

A family of about 200 species in 6—10 genera, the generic lines not well
agreed upon, but including Magnolia, Talauma, Michelia, Manglietia and
Liriodendron. The family is bicentric in distribution: all of the genera
occur in eastern or southeastern Asia and Magnolia, Talauma and Lirioden-
dron also occur in the New World. Although many of the 80 fossil “species”
of Magnolia and the 25 of Liriodendron are doubtful, the record is suffi-
cient to show that these two genera were formerly of wide distribution in
the Northern Hemisphere but have become extinct over most of this area.

The family as here considered does not include Schisandra, Kadsura,
Illictum, Drimys and its relatives, Trochodendron, Tetracentron, Austro-
baileya and other genera which the studies of I. W. Bailey, his collaborators
and students have shown clearly to belong elsewhere. Magnoliaceae (sensu
stricto) seems to be most closely related to Degeneria and Himantandra.
(See also Annonaceae.) Canright notes, “. . . the tissues and organs of
the Magnoliaceae reveal many transitions between the more primitive
Degeneriaceae and the slightly more specialized Himantandraceae; yet all
three families undoubtedly form a compact alliance within the woody
Ranales.”’

With their woody habit, wood with some primitive features, mostly
perfect flowers, numerous and mostly free floral parts with hypogynous
and partially spiral insertion, sporophyll-like stamens, monocolpate pollen
and ethereal oil cells, the Magnoliaceae are often considered to be among
the most primitive living angiosperms. Bailey & Nast in concluding their
studies on the Winteraceae (1945) wrote, however, “. . . it is unfortunate
that so much attention has been focused upon the Magnoliaceae (sensu
stricto) in discussions concerning the origin of the angiosperms, for the
seedlings, stems, roots, leaves, stamens, and carpels of these plants all
exhibit a relatively high degree of morphological specialization. More
primitive and significant ranalian structures are retained by such families
as the Winteraceae, Degeneriaceae, Himantandraceae, Trochodendraceae,
etc.” The detailed studies of Canright bear this out and he concludes, in
part, “. . . the evidence in support of the postulated primitiveness of the
Magnoliaceae is in some tropical stamen types, plus the occurrence of distal-
furrowed monocolpate pollen.”

It may be added that Sect. Tasmania of Drimys (Winteraceae) appears
to combine more numerous primitive features than any other known group
of angiosperms. However, in spite of its more advanced position on a
world-wide basis, Magnolia is probably the most primitive genus within
the Southeastern flora.

1958 | WOOD, GENERA OF WOODY RANALES 301

REFERENCES:

Battery, I. W., AND COLLABORATORS. See the very important series of papers
Heating with related groups in Jour. Arnold Arb. 23-38. 1942-1957, espe-
cially Battey & Nast, The aa pa of the Winteraceae
VII. Summary and conclusions. 26: 37-47. 1945.

Canricut, J. E. The comparative ae of the Magnoliaceae. Doctoral
Dissertation. Widener Library, Harvard University. May, 1949. [With
the exception of data on carpels and seedlings, most of the material covered
is included in the three published papers. |

The comparative morphology and relationships i the Magnoliaceae.
I. Trends of specialization in the stamens. Am. Jour. Bot. 39: 484-497.
1952. II. Significance of the pollen. see aeacta Ht 3: 355-365. 1953.
IV. Wood and nodal anatomy. Jour. Arnold Arb. 36: 120-140. 1955. [The
most comprehensive survey; extensive bibliography. |

Danpy, J. E. The genera of Magnolieae. Kew Bull. 1927: 257-264. 1927. [| Key
to the genera of family, sensu stricto.

Magnoliaceae. Inst. Paranaense Bot. Cat. Estat. on Bot. Fan. 11:
1, 2. 1956. [List of genera with synonymy; estimate of no. of spp. |

ey R. D’O. Goop. Magnoliaceae Jaume. Pflanzenareale o 35-38. Maps
41-43.

D’Ippoiito, G. Contributo all’anatomia comparata del caule delle Magnoliacee
in feiarione specialmente alla struttura anatomica del legno secondario.
Malpighia 15: 438-460. 1901. [Includes Magnolia, Michelia, Illicium,

GarRIsSon, R. Studies in the development of axillary buds. Am. Jour. Bot. 42:
257-266. 1955. [Includes ee ie stellata var. rosea, Liriodendron
Tulipifera, Schisandra chinensis, am rs. |

Goop, R. D’O. The past and present cerca of the Magnolieae. Ann. Bot.
39: 409-430. 1925.

Groppter, R. Vergleichende Anatomie des Holzes der Magnoliaceen. Biblioth.
Bot. 31: 1-49. 1894.

Jounson, M. A. Relationship in the Magnoliaceae as determined by the pre-
cipitin reaction. Bull. Torrey Bot. Club 80: 349-350. 1953.

he precipitin reaction as an index of relationship in the Magnoliaceae.
Serol. Mus. Bull. 13: 1-5. 1954. [Magnolia, Michelia, Talauma closely
related; Liriodendron more isolated; Jllicium unrelated. ]

LEMESLE, R. De l’ancienneté des caractéres anatomiques des Magnoliacées.
Revue Gén. Bot. 45: 341-355. 1933.

es caractéres histologiques du bois secondaire des Magnoliales.
Phytomorphology 3: 430-446. 1953. [See erie 19555

McLaucHLin, R. P. Systematic anatomy of the woods of the Magnoliales.
Trop. Woods 34: 3-39. 1933. [Magnoliaceae, Schisandraceae, Winteraceae,
Trochodendraceae, Cercidiphyllaceae, Lactoridaceae, Himantandraceae. |

. Some woods of the Magnolia family. Jour. Forestry 26: 665-677. 1928.
[A key to their structural identification, gross and microscopic; Magnoli-
aceae sensu lato.

ManevaL, W. E. The development of Magnolia and Liriodendron, including
a discussion of the primitiveness of the Magnoliaceae. Bot. Gaz. 57: 1-30.
1914. [M. virginiana, L. Tulipifera.]

Marsupa, S. On the anatomy of Magnoliaceae. ee Coll. Sci. Imp. Univ.
Tokyo 6: 115-149. 1893. [Magnoliaceae, sensu |

302 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxxrx

OzeNbA, P. Structure du noeud foliaire des Magnoliacées et des Anonaccées.
Compt. Rend. Acad. Sci. Paris 224: 1521-1523. 1947.

PARMENTIER, P. Histoire des Magnoliacées. Bull. Biol. France Belgique 27:
159-337. 1895.*

PLouvier, V. Sur la recherche du pinitol chez quelques Caryophyllacées,
Maenoliacées et oo es de familles voisines. Compt. Rend. Acad. Sci.
Paris 244: 382-385. .

U. S. Dep. Acr., at SERV. Woody Plant Seed Manual. U. S. Dep. Agr
Forest Serv. Misc. Publ. 654. 1948. [Liriodendron, 222, 223; Magnolia,

ae
WeEEVERS, T. s Quercetin bei den Magnoliaceae und seine Verbreitung im
Le Atses Proc. Roy. Acad. Amsterdam 33: 778-785. 1930.*

WuitaAker, T. W. Chromosome number and relationship in the Magnoliales.
Jour. Arnold Arb. 14: 376-385. 1933. [Undocumented counts. |

WorsbELL, W. C. A study of the vascular system in certain orders of the
Ranal Ann. Bot. 22: 651-682. 1908. [Mostly petiole; includes Mag-
noliaceae sensu lato; bibliography. |

KEY TO THE GENERA OF MAGNOLIACEAE

Deaves entire, acute to cordate-auriculate at the base; petals white, green or
ow; stamens without distinct filaments, introrse or latrorse; styles decidu-
it e numerous carpels in fruit forming a cone-like follicetum; carpels
opening on ae serous ees the seeds pendulous by threads, with a flesh
scarlet to pink outer coat. ............0 0.0.0.0... cee ee. Magnolia.
Leaves with 4 or 6 an pines ieneansae at the tip; petals erg yellow
with an orange band near the base; stamens with distinct, although thick
filaments, extrorse; styles persistent, flat and wing-like; the numerous carpels
maturing as a cone-like mass of samaras, each of which falls separately.
2. Lirtodendron.

Tribe MAGNOLIEAE DC.
1. Magnolia L. Sp. Pl. 1: 535. 1753; Gen. Pl. ed. 5. 240. 1754.

Trees, or sometimes shrubs, mostly with showy and large flowers, the
leaves deciduous to evergreen, entire or sometimes cordate-auriculate at
the base. Tepals 9-15, in series of 3, white or green to yellow [pink or
purple], similar or the outer 3 sometimes partially differentiated, deciduous.
Stamens with filament and connective hardly differentiated, the latter pro-
duced into a blunt point [in ours] beyond the anther-sacs (sporangia) ;
anther-sacs 4, linear, opening introrsely or latrorsely. Styles recurved,
deciduous, the stigma along the inner face. Fruit a cone-like follicetum of
the more or less fleshy imbricated carpels, the individual follicles at matur-
ity clearly separate, dehiscent along the outer (abaxial side), the two seeds
hanging by a delicate silky thread of unrolled spiral vessels (from the
funiculus and placenta); seeds with a fleshy scarlet to pink outer layer
and a hard bony inner layer, both derived from the outer integument.
2n = 38, 76, 114. (Including Tulipastrum Spach.) Type species: M.
virginiana L. (Named in honor of Pierre Magnol, 1638-1715, Professor

1958 | WOOD, GENERA OF WOODY RANALES 303

of Botany at the botanical garden at Montpellier, France.) — MAGNOLIA,
Bay, CUCUMBER-TREE.

An ancient genus with about 75-80 species, in two centers: about 50 in
the Old World (Japan to the eastern Himalayas, south to Java) and about
25 in the New World (eastern U.S., the Greater Antilles, and Mexico to
southeastern Venezuela). Dandy divides the genus into 2 subgenera and
11 sections; both subgenera and 4 sections are represented by the 8 species,
including 5 varieties, of our area.

Subgenus Macnoria. Anthers dehiscing introrsely; flowers neither
precocious nor with a much reduced outer whorl of tepals; leaves evergreen
or deciduous. Eight sections, 5 entirely Asiatic, 2 entirely American.

Sect. RyTIDOSPERMUM Spach includes 3 species of Asia and 4-6 of
America, all white-flowered and deciduous, with the leaves crowded to-
gether near the tips of the branches. Magnolia tripetala L., M. Fraseri
Walt., M. pyramidata Bartr., M. macrophylla Michx. and M. Asher
Weatherby represent this group in our area. Magnolia macrophylla and
M. Ashei are very similar, differing principally in size of plant and shape
of fruit, and probably are only varietally distinct. It is also notable that
the Mexican Magnolia dealbata Zucc., of the mountains of Veracruz to
Oaxaca, is hardly distinguishable from M. macrophylla and may well be
only disjunct populations of that species. All of the species of the section,
except M. pyramidata and the Mexican plant (as yet uncounted), have
been determined to be diploid (2 = 38)

Sect. Macnoria (Magnoliastrum DC.) includes only M. virginiana, the
Sweet Bay, of wide distribution from eastern Massachusetts to southern-
most Florida, eastern Texas and Arkansas. The species, a diploid (2 =
38), is notable for the adnate stipules, leaves glaucous beneath, and very
fragrant, small white flowers. Two geographical varieties (var. virginiana
and var. australis Sarg.), based primarily on size of plant and pubescence
of branchlets, peduncles and leaves, are currently recognized but need
further study.

Sect. THEORHODON Spach is composed of about 15 American evergreen
species with stipules free from the petioles. All are tropical in distribution,
with the exception of the exceedingly handsome M. grandiflora (2n = 114)
which ranges from southern Florida northward on the coastal plain to
eastern North Carolina and eastern Texas and Arkansas. Magnolia grandt-
flora is widely cultivated throughout our region and has escaped in some
areas; the exact limits of its native occurrence need to be determined more
carefully. Its closest relationships seem to be with the group of species
which includes M. Schiediana Schlecht., also with 114 chromosomes, and
others of Central America southward to the isolated table-top mountains
of southeastern Venezuela. The 8 species of Cuba, Hispaniola and Puerto
Rico are all closely related, on the other hand, and form a separate sub-
section. Magnolia Hamori Howard, of Hispaniola, is a diploid; other
chromosome numbers are unknown.

304 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

Subgenus PLreurocHasMA Dandy. Anthers dehiscing laterally or sub-
laterally; flowers precocious and/or with a much reduced (calyx-like)
outer whorl of tepals; leaves deciduous. Three sections, two entirely
Asiatic.

Sect. TuLtipastrum (Spach) Reichb., with the outer whorl of tepals
reduced to a small “calyx,’”’ includes only the green- or yellow-flowered
M. acuminata L. (sensu lato), of eastern North America, and the purple-
flowered M. liliflora Desrouss., of eastern China. Both are tetraploids,
2n = 76. Magnolia acuminata, the most variable of our species, appears
to be composed of three more or less well defined geographical varieties
(var. acuminata, var. cordata (Michx.) Sarg., and var. ozarkensis Ashe),
but some aspects of its variation deserve further study. (See Hardin.)

No wild hybrids have been found in the genus, although garden hybrids
are common. No hybrids between subgenera have been obtained but a
number of intersectional hybrids are known, including M. * Thompsoni-
ana (Loud.) C. de Vos (M. tripetala X virginiana) and M. virginiana X
grandiflora. The best known hybrid is the intersectional M. x Soulangiana
Soulange-Bodin (M. denudata liliflora) (2n = 95, 114), which is widely
grown in a number of cultivars. All species appear to be proterogynous,
the stigmas being receptive just before the flowers open.

Although Magnolia retains relatively primitive sporophyll-like stamens,
especially in some of the tropical Asiatic species of Sect. Gwellimia, the
genus is advanced within the family in respect to carpellary features, being
surpassed only by Liriodendron which has both more specialized stamens
and carpels. Magnolia is most closely related to Talauma: some of the
tropical Asiatic species are so similar in flower that fruit is necessary for
proper identification,

REFERENCES:

AranasIEV, M. A physiological study of dormancy in seed of Magnolia acuminata.
Cornell Agr. Expt. Sta. Mem. 208: 1-37. 1937.*

Anprews, F. M. Karyokinesis in Magnolia and Liriodendron, with special
reference to behavior of chromosomes. Bot. Centr. Beih. 11: 134-142. 1901.

Brusu, W. D. Southern magnolia (Magnolia grandiflora Linnaeus). Am. Forests
52: 32-33. 1946.

. Our native magnolias. Am. Forests 62: 31-32, 58-59. 1956.

Coker, W. C. Magnolia cordata Michaux. Jour. Elisha Mitchell Sci. Soc, 59:
81-88. 1943. |Taxonomy, distribution. ]

Crum, P. Chelating agents for the control of lime-induced chlorosis in southern
magnolia (Magnolia grandiflora). Proc. Nat. Shade Tree Conf. 30: 267-

Danpy, J. E. A survey of the genus Magnolia together with Michelia and
Manglietia. In Camellias and Magnolias. Roy. Hort. Soc. Conf. Rep
64-81. London, 1950.

. Key to the species of Magnolia. Jour. Roy. Hort. Soc. 52: 260-264.
1927.

DauMANN, E. Das Bliitennektarium von Magnolia und die Futterkérper in der
Blute von Calycanthus. Planta 11: 108-116. 1930.

1958] WOOD, GENERA OF WOODY RANALES 305

Earte, T. T. Embryology of certain Ranales. Bot. Gaz. 100: 257-275. 1938.
[M. grandiflora, Cimicifuga racemosa.
Origin of the seed coats in Magnolia. Am. Jour. Bot. 25: 221, 222.
1938. [M. grandiflora; corroborates Gray’s observations. ]
Evans, C. R. Germination behavior of Magnolia grandiflora. Bot. Gaz. 94:
729-754, 1933.

Farr, C. H. Cell division by furrowing in Magnolia, Am. Jour. Bot. 5: 379-
395. 1918. [M. tripetala and cultivated spp. not specified, Jamaica. ]
FLANpERS, B. C. Reliability of growth rings in Magnolia grandiflora as indi-
cators of age. Quart. Jour. Fla. Acad. 13: 77-109. 1950. [False growth

rings. |

FREEMAN, O. M. A new Magnolia ee Nat. Hort. Mag. 16: 161-162. 1937.
[M. virginiana 2 X grandiflora ¢

New Magnolia hybrids. Loc. ae 30: 132-135. 1951. [M. acuminata
x ‘cordate and reciprocal; cultivated plants. ]

Gigson, H. H. American forest trees — 39. Cucumber tree, Magnolia acuminata
Linn. Hardwood Rec. 23: 16, 17. 1906.

Gray, A. A short exposition on the structure of the ovule on seed-coats of
Magnolia. Jour. Linn. Soc. 2: 106-110. 1858. [M. tripeta

Grier, N. M. Note on fruit of mountain Magnolia. Rhodora 19: 256. 1917,
[Misshapen cones of M. acuminata.

GrRIESEL, W. O. Retardation of maturation in Magnolia aewere by maleic hydra-
zide. Science 119: 843-845. 1954.

Cytological changes accompanying abscission of perianth segments

of Magnolia grandiflora. Phytomorphology 4: 123-132. 1954.

Harpin, J. W. An analysis of variation within Magnolia acuminata L. Jour.
Elisha Mitchell Sci. Soc. 70: 298-312. 1954. [3 vars. recognized. |

Harper, F. Two more available plant names of William Bartram. Bartonia
21: 6-8. 1942. [M. pyramidata oe see also MERRILL, E. D., In defense
of the validity of Bartram’s bino ials. Bartonia 23: 10-35. 1945.]

Havet, A. W. The stem of ie as a laboratory type. Science 61: 469.
1925

Howarp, R. A. The morphology and systematics of the West Indian Magnoli-
aceae. Bull. Torrey Bot. Club fe 335-357. 1948. [Magnolia and Talauma. |

Jounson, H. W., ano C. W. Epcerton. A heart rot of magnolia caused by
Fomes geotropus. Mycologia 28: 292-295. 1936.

Jounson, M. A. The precipitin reaction as an index of relationship in the
Magnoliaceae. Serol. Mus. Bull. 13: 1-5. 1954. [Mostly comparison of
sp. in different sections; M. tripetala and M. obovata show close corre-
spondence. |

JounstTone, G. H. Asiatic Magnolias in cultivation. 160 pp. Roy. Hort. Soc.,
London, 1955.

KENNEDY, G. G. Some historical data regarding the sweet bay and its station
on Cape Ann. Rhodora 18: 205-212. 1916. [M. virginiana at its northern-
most locality, e. Mass. ]

Kurz, H., anD K. WAGNER. ee associations in northern Florida. (Abs.)
Tou ur. Tenn. Acad. 27: 207. 1952.

MEEHAN, T. On the stipules A Magnolia Fraseri. Proc. Acad. Phila. 1887: 155.
1887.

Mitrais, J. G. Magnolias. 251 pp. London, 1927.

306 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. Xxxrx

Ozenpa, P. Sur la vascularisation des carpelles et du pistil chez les Magnolia.
Compt. Rend. Acad. Sci. Paris 217: 31-33. 1943

Piovuvier, V. Sur la presence de rutoside dans les Aeuirs de quelques Magnolia.
Compt. Rend. Acad. Sci. Paris 216: 459-461. 1943.

Sur l’huile des graines de Magnolia macrophylla Michx. (Magnoliacée).
Loc, cit. 222: 1009-1011. 1946.

Reep, R.A. Nuclear phenomena in the tapetum of Magnolia grandiflora. Was-
mann Collect. 7: 1-15. 1947

SANTANTARAI, B. A note on the induction of roots on the twigs of Magnolia
grandiflora L., with the aid of synthetic hormones. Indian Jour. Hort. 12:
32-33. 1955.*

SAwapa, K. Oriental magnolias in the South.. Nat. Hort. Mag. 29: 54-57. 1950.

WEATHERBY, C. A. A new Magnolia from west Florida. Rhodora 28: 35-36.
1926. [M. Ashez; see also SMALL, J. K. A magnolia as a new border plant.
Jour. N. Y. Bot. Gard. 34: 150-152. 1933. |

Tribe LIRIODENDREAE Reichb.

2. Liriodendron L. Sp. Pl. 1: 535. 1753; Gen. Pl. ed. 5. 239. 1754,
“Tiriodendrum.”

Large, deciduous trees with long-petioled leaves with conspicuous stipules
and very characteristic leaf-blades with 2 lateral lobes near the base (and
sometimes 2 smaller above) and 2 at the apex which appears as if cut off
abruptly by a broad, shallow notch. Perianth segments 9, deciduous, the
3 outer ones sepaloid, green, glaucous, reflexed, the 6 inner ones in 2 whorls
making a campanulate, tulip-like corolla, greenish-yellow, each with an
orange band near the base. Stamens numerous (+ 30), the filaments
stout, narrowed to the broader, apiculate. extrorse anther. Gynoecium of
numerous spirally arranged carpels tightly imbricated into a cone-like
column as long as the petals; style elongated, broad, flattened and wing-
like, constricted to a small, recurved stigmatic crest. Fruit a spindle-
shaped cone of closely appressed 2-seeded samaras, these falling separately
at maturity leaving the persistent receptacle; seeds with a thin, dry, and
leathery testa. Type species: L. Tulipifera L. (The name from Greek,
lirion, lily or tulip, and dendron, tree, from the tulip-like flowers.) —
YELLOW-POPLAR, TULIP-POPLAR, TULIP-TREE.

An ancient genus formerly of wide distribution in the Northern Hemis-
phere, now reduced to two very similar species, L. Tulipifera, of eastern
North America, and L. chinense (Hemsl.) Sarg., of a small area in central
China (parts of Kweichow, Chekiang, Hupeh, Kiangsi, and Wushan Prov-

es).

Liriodendron Tulipifera (2n = 38), a very handsome and important
timber tree, reaches its best development in the rich, hardwood forests of
the Appalachians, attaining a maximum height of 200 ft. and a circumference
of almost 35 ft. It is distributed from western Massachusetts and southern
Vermont to southernmost Ontario, southern Michigan, Indiana, south-
eastern Missouri, eastern Arkansas, and Louisiana to central Florida.

1958 | WOOD, GENERA OF WOODY RANALES 307

Lilie
4 yy t
ty Lilie

Fic. 1. Liriodendron. a—-j, L. Tulipifera: a, flowering branchlet, * 1/4; b,
flower-bud with stipular bud-scales, & 1/2; c, stamen, abaxial view, X 2;
d, unopened anther, filament, and anther after anthesis, cross-sections, pollen
omitted, * 6; e, gynoecium, with sepals, petals, and stamens removed, x 1;
f, gynoecium, portion of cross-section, with spirally arranged imbricated carpels,
ovaries adnate to axis to lower right, increasingly flattened styles toward out-
side, the locules and stylar canals in black, X 3; g, carpel at anthesis, ese
section, X 1; h, mature gynoecium with many samaras already shed from axis,
rls samara, X _ j, lower part of samara, vertical section, with eee
ovule TO) and seed with bony coats, abundant endosperm and small embryo,

. f, g, Jj, semi-diagrammatic.

The leaves although always unmistakable are extremely variable and
most of those described for 25 fossil species may be matched from the
existing populations. The faintly fragrant proterogynous flowers of L.
Tulipifera are provided with copious watery nectar at anthesis, but seem
to be visited primarily by bees collecting the abundant pollen.

Liriodendron chinense (2n = 38) seems to be generally a smaller tree,
(ca. 50 ft.) with slightly different leaves, smaller flowers and smaller and
slightly eat fruit. The plant is not nearly so hardy as our native
species. Apparently no crosses between the two species have been made;

isolation of the parental species and in comparison with similar crosses
in Catalpa and Platanus.
Liriodendron is isolated within the Magnoliaceae with no close relatives.

08 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

wW

It is the only genus in the group with a definitely localized stigma, and
the manner of wasoulavization of the ovules is unique in the family.

REFERENCES:

Anprews, F. M. Karyokinesis in Magnolia and Liriodendron, with special
reference to behavior of chromosomes. Bot. Centr. Beih. 11: 134-142.
1901

BARRACHINA Y ALMEDA, J. Tulipero de virginia estudio botanico selvicola e in-
dustrial. Mem. Acad. Cien Artes Barcelona 23: 39-66. 1932.

Berry, E. W. The origin of stipules in Liriodendron. Bull. Torrey Bot. Club 28:
493-498. 1901.

_ Notes on Liriodendron leaves. Torreya 1: 105-107. 1901.

——. Notes on the phylogeny of Liriodendron. Bot. Gaz. 34: 44-63. 1902.

———., Additional notes on Liriodendron leaves. Torreya 2: 33-37. 1902.

———.. Liriodendron notes. Torreya 3: 129-132. 1903

_ Liriodendron in the Miocene of America and Eastern Asia. Torreya

1: 82-84. 1941.

Betts, H. S. Yellow-poplar (Liriodendron tulipifera). U.S. Forest Serv. Am.
Woods. 8 pp. Washington, 1954.

Cuapman, A. G. Some effects of varying amounts of nitrogen on the growth of
tulip poplar seedlings. Ohio Jour. Sci. 33: 164-181. 1933.

Creasy, W. D. Secondary succession and growth of yellow poplar. Castanea 19:
81-87. 1954

Curtis, C. C. Second flowering of the tulip-tree. Jour. N. Y. Bot. Gard. 2:
136-138. 1901.

Driver, C. H. Morphology of mycorrhizae of Liriodendron tulipifera L. (Abs.)
Jour. Tenn. Acad. 25: 224

FRASER, J. Liriodendron ae, Gard, Chron. III. 88: 109. 1930.

Gipson, H. H. American forest a or whitewood, Liriodendron
Tulipifera Linn. Hardwood Rec. 19: 14, 1905.

Guarp, A. T. The development of the seed a Ciiodenon tulipifera L. Proc.
Indiana ge a 75-77. 1943. (1944).

+b

Heptine, G. H., anp G. G. Hepccocx. Decay in ee oak, yellow pop-
lar, and bass “yood i in the Appalachian region. U. S. t. Agr. Tech. Bull.
570: 1-29. 1937.

Ho.ttick, A. Wang. like appendages on the petioles of Liriophyllum populoides
Lesq. and Liriodendron alatum Newb., with description of the latter. Bull.
Torrey Bot. Club 21: 467-471. 1894

. A new fossil Liriodendron from the Laramie at Walsenberg, Colo. and
its significance. (Abs.) Proc. Am. Assoc. Adv. Sci. 43: 225. 1895.

———. Appendages to the petioles of Liriodendra. Bull. Torrey Bot. Club 23:
249- 250. 1896.

. The tulip tree and its ancestors. Proc. Nat. Sci. Assoc. Staten Island 5:
80. 1896.*

Hom, T. On the validity of some fossil species of Liriodendron. Bot. Gaz. 20:
312-316. 1895.

a on the leaves of Liriodendron. Proc. U.S. Nat. Mus. 13: 15-25.

: acne plants of North America. 30. Liriodendron tulipifera L.
Merck’s Rep. 18: 198-201. 1909.*

1958 | WOOD, GENERA OF WOODY RANALES 309

HucKENPAHLER, B. J. Auxins fail to stimulate rooting of yellow-poplar cuttings.
Bot. Gaz. 117: 73-75. 1955

Jounson, T. W., and others. Observations on yellow-poplar (Lirtodendron tulipi-
fera) dieback. Forest Sci. 3(1): 84-89. 1957

Li, H. L., anp J. W. Wricut. The Chinese tuliptree (Liriodendron chinense) in
northeastern United States. Morris Arb. Bull. 5: 34-35. 1954. [Notes on
history, distribution, introduction into U. S., hardiness and chromosome-

number.

Limstrom, G. A., AND R. F. Finn. Seed source and nursery effects on yellow-
poplar plantations. Jour. Forestry 54: 828-831. 1956.

Meyer, B.S. The daily periodicity of transpiration in the tulip poplar, Lirioden-
dron tulipifera L. ee Jour. Sci. 32: 104-114. 1932

Mitutncton, W.-E., AND J. E. GuNKEL. Structure and development of the
vegetative shoot tip - Liriodendron tulipifera L. Am. Jour. Bot. 37: 326-
Shins e150)

Pray, T. R. Foliar venation of angiosperms I. Mature venation of Liriodendron.
Am. Jour. Bot. 41: 663-670. 1954. II. Histogenesis of the venation of
Liriodendron. ioe. cit. 42: 18-27. 1955

SCHWERIN, F. Grar von. Angeblicher Atavismus bei Liriodendron. Mitt.
Deutsch. Dendr. Ges. 28: 135-143. 1919

Voict, J. W. A preliminary investigation of the effect of the descaling of winter
buds on their growth in east central Illinois. Trans. Ill. State Acad. Sci.
35(2): 78-80. 1942. [Incl. Liriodendron, Betula nigra, Tilia europea. |

Welsse, A. Uber die Blattstellung von Liriodendron tulipifera. Ber. Deutsch.

ot. Ges. 20: 488-493. 1902.

ANNONACEAE (CusTarD-APPLE FAMILY)

Trees, shrubs [or vines] with alternate, exstipulate, simple, entire leaves
with pinnate venation; buds naked, the leaves conduplicate; oil glands
present, the plants aromatic. Flowers perfect [in most], hypogynous,
regular, axillary [or terminal], usually nodding. Perianth trimerous, gen-
erally of 3 small sepals and 6 petals in 2 whorls of 3, the inner smaller [or
sometimes lacking]. Stamens numerous, spirally inserted on the recep-
tacle, filament and anther poorly or not at all differentiated, the sterile tip
variously modified; sporangia 4, extrorse, opening longitudinally; pollen
in tetrads [or single], monocolpate [to acolpate], the germinal furrow
proximal. Carpels many—1, usually free but sometimes united by the
ovaries at anthesis [and rarely more completely syncarpous|; stigmas
terminal; ovules many—1, in 1 or 2 rows along the adaxial wall or basal,
anatropous, 2-integumented. Carpels free in fruit and berry-like, or co-
alescent, forming a fleshy syncarp. Seeds arillate (in ours), with ruminate
endosperm and a small embryo. Embryo-sac development normal (Poly-
gonum type) [insofar as investigated], endosperm development cellular.

A tropical family with 75-120 genera and more than 1000 species, many
poorly known, and the classification of the family not yet well agreed upon.
Represented in our area by Asimina, the only extra-tropical genus, and
by Annona, which reaches subtropical Florida.

310 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

The family is distinguished by the exstipulate, simple leaves, the 3 whorls
of 3 perianth segments, the numerous more or less fleshy spirally inserted
stamens (each with a single vein), the usually numerous carpels, the fleshy
fruits, the large seeds with ruminate endosperm, the tri-lacunar nodes with
tripartite median trace, and the monocolpate (or derived) pollen with
proximal germinal area (evident in those species in which the pollen is
shed in tetrads). Anatomical features of wood, stem and leaf are remark-
ably uniform throughout the family.

It has been agreed generally that the affinities of the group are with the
Ranales (sensu lato) but more precise relationships have been a matter of
speculation. The anatomical evidence brought together by Vander Wyk
and Canright strongly supports the view that the Annonaceae should be
most closely allied with Degeneriaceae, Magnoliaceae and Himantandra-
ceae, on a level of specialization above the Magnoliaceae and perhaps
roughly comparable with Himantandraceae.

REFERENCES:

Apatia, R. D., anp D. B. CHoxsur. The chromosome numbers in the family
Anonaceae. Cur. Sci. Bangalore 20: 102. 1951.*

AsANA, J. J., AND R. D. Apatia. The enrpmnsome numbers in the family Anon-
aceae. Cur. Sci. Bangalore 14: 74-75. 1945.

. Contributions to the embryology of a Annonaceae. Jour. Bombay
Univ. B, Biol. Sci. II. 16(3): 7-21. 1947.*

Bowpen, W. M. Chromosome ia in the Annonaceae. Am. Jour. Bot. 35:
337-381. 1948. [7 sp. Asimina 2n = 18; 5 sp. Annona, 2n = 16; and
others.

Corner, E. J. H. The annonaceous seed and its four integuments. New Phytol.
48: 332-364. 1949,

Diets, L. Die Gliederung der Anonaceen und ihre Phylogenie. Sitzungsb.
Preuss. Akad. Wiss. 1932: 77-85. 1932.

Fries, R. E. Einige Gesichtspunkte zur systematischen Gruppierung der ameri-
kanischen Annonaceen- Gattungen. Ark. Bot. 30A(8): 1-31. 1943.

. Revision der Arten eae aera Gattungen. I-V. Acta Horti
Berg. 10: 1-341. 1931; 12: 1-220, 289-577. 1939.

. Verstreute beobachtungen cents Ea familie Annonaceae. Ark.
Bot. II. 3: 35-42. 1955. [Old world genera. |

Garratt, G. A. Bearing i wood anatomy on the relationships of the Myristi-
caceae. Trop. Woods 36: 20-44. 1933. | Many data on Annonaceae. |

Hurcuinson, J. Contributions toward a phylogenetic classification of flowering
plants. Il. The genera of Annonaceae. Kew. Bull. 1923: 241-261. 1923.

Licopo.!, G. Sull anatomia a fisiologia del frutto nell’ enone reticulata e nell’
Asimina triloba. Atti Accad. Napoli II. 1: 1-9. 1887.

OzenNpa, P. Structure du noeud foliaire des Magic et des Anonacées.
Comp t. Rend. Acad. Sci. Paris 224: 1521-1523. 194

PerIAsAMy, Kk. On the cree biology of some members i ine Anonaceae. Jour.
Madras Univ. 24: 1954.*

Russy, H. H. The sees apple capes in Florida. Jour. N. Y. Bot. Gard. 36:
P38-930.. 7035: | Asiina and Annona. |

VANDER Wyk, R. The comparative ae of the Annonaceae. Doctoral
Dissertation. Widener Library, Harvard Univ. 1950. [Stem, wood, node

1958 | WOOD, GENERA OF WOODY RANALES 311

and petiole, leaf, cotyledon, and pollen, especially, in comparison with De-

generiaceae, Magnoliaceae, Himantandraceae, Myristicaceae, Eupomatiaceae,

Canellaceae, Calycanthaceae. |

AND J. E. Canricut. The anatomy and relationships of the Annonaceae.
Trop. Woods 104: 1-24. 1956. [Stem and wood anatomy. ]

Vorct, A. Untersuchungen iiber Bau und Entwicklung von Samen mit ruminiertem
Endosperm aus den Familien der Palmen, Myristicaceen und Anonaceen,
Ann, Jard. Bot. Buitenzorg 7: 151-190. 1888.

Wester, P. J. Annonaceous possibilities for the plant breeder. Philipp. Agr.
Rev. 6: 312-321. 1913.

1. Asimina Adans. Fam. Pl. 2: 365. 1763.

Trees (to 40 ft.), shrubs or subshrubs (ca. 2 ft.). Flowers nodding,
axillary on greatly reduced branches, solitary or in pairs, borne on the
wood of the preceding season or on the growth of the year, ill-scented or
fragrant. Sepals 3 (rarely 4), small, valvate in the bud. Petals usually
6(—8 or sometimes 12), in 2 (or more) series of 3, the outer largest (often
very, much larger), imbricated, brown or purplish, greenish, white or yellow-
ish, often increasing greatly in size during anthesis. Stamens numerous—10,
inserted on the subglobose to nearly flat receptacle; pollen in tetrads.
Carpels 15-1, distinct, the styles short, the stigma small; ovules many-—6,
in two rows or one. Carpels berry-like in fruit (usually one 1-4 maturing)
free, banana-shaped or somewhat torulose to ellipsoid or ovoid, the flesh
aromatic; seeds flattened to round, many-4, inclosed in a pulpy mem-
branaceous aril. (Including Pityothamnus Small and Deeringothamnus
Small.) Type species: A. triloba (L.) Dunal. (The name from asiminier,
an early French-colonial name for. A. triloba, this, in turn, from the Indian
name assimin).— PAWPAW.

Perhaps 10 species, all of eastern North America. Asimina triloba
(northern Florida to Texas, north to New Jersey, western New York,
southern Ontario, Michigan, Illinois, southeastern Iowa and southeastern
Nebraska) and A. parviflora (Michx.) Dunal (Piedmont and Coastal Plain,
northern Florida to Mississippi, north to southeastern Virginia) have the
widest ranges; the others are mostly confined to Florida. The genus in-
cludes the only truly extra-tropical species in the family. According to
Fries, the only close relative is Stenanona Standl., of Panama and Costa
Rica.

Although the species of Asimina range from deciduous trees reaching
40 ft. (A. triloba) to low, fusiform-rooted shrubs with dimorphic stems
and partly persistent leaves and, although there is considerable diversity
in flower-color and petal-shape, -size, and -sculpturing, the group appears
to be a natural one. The differences used by Small in segregating Pztyo-
thamnus and Deeringothamnus seem trivial as generic distinctions. The
chromosome-numbers of species of the former group are the same as those
of A. triloba and A. parviflora (2n = 18) and, even more significantly,
vigorous hybrids have been obtained between A. ¢riloba and at least “A.

312 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

obovata” (= A. grandiflora (Bartr.) Dunal?) and “A. angustifolia” (= A
pygmaea (Bartr.) Dunal?), two very different species assigned to Pityo-
thamnus. The former hybrid is known to be fully fertile. No such data
are available for the two species placed in Deeringothamnus (A. pulchella
(Small) Rehd. & Dayton, of sw. Florida, and A. Rugeli Robins., of ne.
Florida). Although differing in the nearly flat receptacle (correlated with
reduced numbers of stamens and carpels), in the narrow, hardly fleshy
petals of nearly equal size, and in the lack of bracts on the peduncles of
the flowers, these plants appear to be only the most specialized members of
the genus, standing at the opposite extreme from A. triloba but connected
to it through the “Pityothamnus” grou

Although apparently rather stable in its vegetative and floral morphol-
ogy, A. triloba shows a wide variation in size, color and palatability of
fruits. Two general types have been observed: (1) large, yellow-fleshed,
highly flavored, early ripening and (2) small to large, white-fleshed, mild-
flavored, late or very late ripening fruits. A number of selected clones,
propagated by grafting, are in cultivation.

The flowers of all species are proterogynous. The brown or purple flowers
of A. triloba and A. parviflora are reputed to be ill-scented, while those of
the white-flowered species, especially A. pulchella and A. Rugelii, are
fragrant. Beetles appear to be involved in pollination. Most species set
few fruit; hand-pollination seems to be necessary to obtain a good fruit-set
in A. triloba.

REFERENCES:

Bowpen, W. M. Triploid mutants among diploid seedling populations of Asi-
mina triloba. Bull. Torrey Bot. Club 76: 1-6. 1949. [5 triploids among
205 seedlings in 23 collections. |

AND B. MILLer. Distribution of the Papaw, oo triloba (L.) Dunal,
in southern Ontario. Canad. Field-Nat. 65: 27-31. 1951.

Deminc, W. C. Papaw (Asimina triloba) notes. er Nut Growers Assoc.
Ann. Rep. (1945) 36: 98. 1946.*

Exett, A. W. William Bartram and the genus Asimina in North America. Jour.
Bot. 65: 65-69. 1927. [Includes key to species. |

Fries, R. E. Revision der Arten einiger Annonaceen-Gattungen. V. Acta Horti
Berg. 12: 289-577. 1937. [Asimina, see crena vine 546-554. |

Gray, Asa. The genus Asimina. Bot. Gaz. 11: 161- 886.

HeRMs, W. B. Contribution to the life history of pie triloba. Ohio Nat.
8: 211-217. 1907. [Embryo sac; embryology. |

Lampton, R. K. Developmental morphology of ovule and seed in Asimina triloba
Dunal. (Abs.) Jour. Tenn. Acad. 28: 182-183. 1953. [See also Diss. Abs.
13: 290, 291. 1953.*]

Leverett, F. The northern limit of the papaw tree. Science 23: 919, 920. 1906.
See also other authors, loc. cit. 23: 920; 23: 749-751. 1906; 24: 48. 1906.

Locke, F. J. Microsporogenesis and cytokinesis in Asimina triloba. Bot. Gaz.
98: 159-168. 1936

Nasu, G. V. Revision of the genus Asimina in North America. Bull. Torrey
Bot. Club 23: 234-242. 1896.

PopreNoE, P. Where are the best papaws? Jour. Hered. 7: 291-296. 1916.

1958] WOOD, GENERA OF WOODY RANALES Sie

_ The best papaws. Loc. cit. 8: 21-33. 1916. [Includes information on
propagation, fruits, largest trees, etc. |

Reuper. A. AND W. A. Dayton. A new combination in Asimina. Jour. Arnold
Arb. 25: 84. 1944. [A. pulchella; Deeringothamnus discarded. |

SMa, J. K. Deeringothamnus pulchellus. Addisonia 11: 33-34. 1926. See
also Bull. Torrey Bot. Club 51: 390. 1924. Deeringothamnus Rugelii. Loc.
cit. 15: 17-18. 1930.

——. A new pawpaw from Florida. Torreya 26: 56. 1926. [A. tetramera. |

Smit, G. H. Vascular anatomy of Ranalian flowers II. Ranunculaceae (con-
tinued), Menispermaceae, Calycanthaceae, Annonaceae. Bot. Gaz. 85: 152-
177. 1928. [Asimina triloba. |

SopayMa, Y. Studies on the pawpaw (Asimina triloba Dunal). III. On the time
of flower bud differentiation and bearing habit. (In Japanese.) Saikyo Univ.
Facul. Agr. Sci. Rep. 7: 81-86. 1955.* [English summary. ]

AND N. Kunimura. Studies on the pawpaw (Asimina triloba Dunal).
I: (1) Flowering season and microsporogenesis in pawpaw, (2) Insect
visitors on the flowers of pawpaw. (In Japanese.) Saikyo Univ. Facul. Agr.
Sci. Rep. 5: 33-46. 1953.* II. On the fruit setting by the self pollination and
the germination test of pollen. (In Japanese.) Loc. cit. 6: 29-37

Upuor, J. C. T. Die nordamerikanischen Arten der Gattung Asimina. Mitt.
Deutsch. Dendr. Ges. 45: 61-76. 1933.

ZIMMERMAN, G. A. Hybrids of the American pawpaw. Jour. Hered. 32: 83-91.
1941. [Notes on A. ¢riloba and interspecific hybrids. |

2 Annona... (Sp: Pl. 12536. 17os5eGen. Pleed: 9241, 1754.

Trees (deciduous with us) with rather coriaceous 2-ranked leaves, nod-
ding flowers borne on axillary or supra-axillary branches (which may abort
producing the effect of either terminal or axillary flowers), and fleshy
syncarpous fruits. Sepals small, valvate in the bud; petals rather thick and
fleshy, generally whitish or yellowish, the outer whorl larger and alternate
with the sepals and valvate, the inner smaller or very much reduced [or
lacking] and valvate [or imbricate]. Stamens club-shaped, the tip modi-
fied, generally broad and truncate, very numerous and tightly packed,
inserted on the hemispherical receptacle; pollen grains in columns of
tetrads. Carpels sessile, numerous, on the receptacle, the ovaries free or
united at anthesis, the styles generally club-shaped, fleshy, and conspicuous;
ovules solitary in each ovary (rarely 2), erect. Carpels coalescent, forming
a many-seeded syncarp with a smooth, squamulose or muricate surface,
the individual carpels being indicated on the surface by more or less dis-
tinctly outlined areoles. Seeds ovate or elliptical, with a thin outer coat
and a thin aril. Type species: A. muricata L. (The name from anon or
hanon, the native Hispaniolan name for A. muricata, but changed by
Linnaeus to Latin annona, a year’s harvest, in preference to the use of a
“barbarous” name, but so that the sound might be kept.) — Cusrarp-
APPLE,

A genus of about 110 species assigned to 17 sections, most species
tropical American, but about 10 in Africa; several species widely culti-

314 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

vated and naturalized in tropical regions; one species native, one sparingly
naturalized and several eee in subtropical Florida.

Sect. PHELLOXYLON Sa ith ovate petals, the inner valvate, and
ovaries connate at anthesis, en only Annona glabra L. (2n = 28),
the most widespread species in the genus (West Indies, north to the Ba-
hamas and southern Florida, south to southern Brazil, Mexico to Equador;
also in Africa from Senegal to the Belgian Congo). Always at low altitudes
and associated with abundant moisture, the Pond-apple or Alligator-apple
occurs in Florida in the Everglades and coastal areas in shallow ponds,
along the borders of small fresh-water streams or on swampy hummocks;
in other areas it seems sometimes to be associated with mangroves. The
leaves appear in March—April and flowers in April; the edible but not very
palatable fruits ripen in November.

sect. ANNONA, cauliflorous, but also with ovate petals, the inner imbri-
cate, and with free ovaries, is represented in cultivation by A. muricata
L., the Guanabana or Sour-sop, a favorite tropical fruit only precariously
hardy in Florida.

Sect. Atta Mart., with elongate petals, the inner greatly reduced, in-
cludes A. squamosa - (Sweet-sop or Sugar-apple), A. reticulata L. (Bul-
lock’s Heart) and A. Cherimola Mill. (Cherimoya), all in cea in
tropical Florida, i. first naturalized on some of the Florida K

All species are proterogynous. At the time of pollination a oe fluid
exudes from the stigmas, gluing the styles together and providing a recep-
tive medium for the pollen. The members of sect. Atta seem to be cross
pollinated (by beetles?) but A. muricata may be self-pollinated, at least
in some areas, Artificial hybrids have been obtained at least between 4.
squamosa, A, reticulata, and A. Cherimola, and between these species and
A. glabra. The Atemoya (A. squamosa & Cherimola) is regarded as a very
promising tropical fruit combining the better features of both parents. The
chromosome numbers of the 4 cultivated species above have been reported
as both 14 and 16. Numerous cultivars, perhaps involving hybridization,
are known in various areas of the tropics. Other species are being intro-
duced and tested for their horticultural possibilities

According to Fries the genus is most closely related to the American
genera Raimondia, Rollinia and Rolliniopsis.

REFERENCES:

Fries, R. E. Revision der Arten einiger Anonaceen-Gattungen II. Acta Horti
Berg. 10: 129-341. 1931. [American spp. of Annona, 197-31

Istam, A. S. pag aang report on the colchicine- induced tetraploids of Anona
squamosa L. Cur. Sci. Bangalore 22: 118-120. 1953.*

Juttano, J. B. Morphological contribution on the genus Azona. Philipp. Agr.
24: 528-541. pos

—

Kumar, L. S. S., AND K. Ranapive. A cytological study of the genus Anona.
Jour. Univ. Racbay 10B: 1-8. 1941.* [A. muricata, A. reticulata, A. squa-
mosa, A, Cherimola, 2n = 14. |

Leon, Hno., AnD Hno. ALAIN. El genero Annona en Cuba. Revista Soc. Cuba.
Bot. 3: 116-124. 1946. [15 spp. ]

1958] WOOD, GENERA OF WOODY RANALES S15

Merser, T. M. The alkaloids of Anona muricata Linn. Buitenzorg Lab. voor
Scheikundig Onderzoek Meded. 94: 1-8. [undated] ; reprinted from De In-
genieur in Nederlandsch-Indié, no. 6. 1941, Rubriek VIL*

Noonan, J. C. Review of investigations on the Anona species. Proc. Fla. State
Hort. Soc. 66: 205-210. 1953.

Pryt, L. VAN Der. On the flower biology of some plants from Java with general
remarks on fly-traps. Ann. Bogor. 1: 77-99. 1953. [Includes self-pollina-
tion in A. muricata. |

PoreNnogr, W. Anona culture in Florida. Am. Eagle 41(6): 1, 19

Prest, R. L. Custard apples Seen ee Fruit & Ue Ree (8): 29>
30. 1952. See also Queensl. Prod. 33(24): Loses

Rogpyns, W., AND J. GHESQUIERE. eee de révision des espéces africaines du
genre Annona L. Bull. Soc. Roy. Bot. Belg. 67(1): 7-50. 1934. [All ex-
cept A. glabra belong to the otherwise 5. Am. sect. Helogenia. |

Sarrorp, W. E. Classification of the genus ete with descriptions of the new
and imperfectly known species. Contr. U. S. Nat. Herb. 18: i-xii + 1-68.

_ The genus Annona: the derivation of its name and its taxonomic sub-
divisions. Jour. Wash. Acad. 1: 118-120. 1911. See also SPRAGUE, Jour.
Bot. 59: 158. 1921, and Kew Bull. 1928: 114. 1928.

Santos. A. C. Alkaloid from Anona reticulata Linnaeus. Philipp. Jour. Sci. 43:
561-564. 1930. [Anonaine from bark.

SamueLsson, G. Uber die Pollenentwicklung von Anona und Aristolochia und
ihre sy stematische Bedeutung. Sv. Bot. Tidskr. 8: 181-189. 1914. [A
Cherimola. |

ScHROEDER, C. A. Fruit morphology and anatomy of the Cherimoya. Bot. Gaz.
112: 436-446. 1951. [A. Cherimola.]

_ Hand pollination of cherimoya practical method for improving fruit
eet for better yields. Calif. Agr. (Calif. Sta.) 1(8): 2. 1947.*

Srmons, J. S. La guanabana y su ee Puerto Rico Agr. Exp. Sta. Rio
Pedras Agriculture Exp. 2(3): 4

Sturrock, D. Tr Pee ae for ee Florida and Cuba and their uses. 131
pp. Arnold Arb., | Annona, 43-50. |

WESTER, P. iP ee experimen nts with Anonas. Bull. Torrey Bot. Club 37:
529-539. 1910. [A. reticulata, A. squamosa, A. ee A. glabra. |

" Hybridization of Annonas. Philipp. Agr. Rev. 8: 177-181. 1915.*

_ A contribution to the history and vernacular aeerire of the culti-

vated Anonas. Philipp. Jour. Sci. (Bot.) 7: 109-123. 1912.

ILLICIACEAE (ANISE-TREE FAMILY)

A single genus, //licium L., of southeastern Asia and southeastern North
America. Although often placed near Drimys (Winteraceae) or associated
with the Magnoliaceae, its closest relatives are undoubtedly Schisandra
ethereal oil cell, cambiform “mucilage” cells in the phloem, unilacunar
nodes, and similar cuticles, stomata and chromosomes, a combination of
characters unique within the Ranales. The total evidence from morphol-
ogy, anatomy and cytology suggests the inclusion of these three genera
either in a single family or in two closely related families; it further pro-

316 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xxx1x

vides a strong obstacle to the association of these genera with either Mag-
noliaceae or Winteraceae. According to the A. C. Smith, ‘‘The three genera
will probably be treated by future phylogenists as composing a sub-order
of the Ranales, coordinate with suborders composed of the (1) Magno-
liaceae, Himantandraceae, and Degeneriaceae, (2) Winteraceae, (3)
Trochodendraceae and Tetracentraceae, (4) Eupteleaceae, and other com-
binations of families . . . not yet sufficiently investigated.”

/ilictum stands apart from the Schisandraceae in its shrubby or arbor-
escent habit, unmodified receptacle, free stamens, vascularized style, com-
paratively few carpels in a single whorl, single ovule, fruit a single whorl
of follicles, unilacunar nodes with a single trace, pseudo-siphonostelic
arrangement of the primary tissues, lack of crystal-bearing sclerenchyma,
and a specific combination of primitive and specialized anatomical features
in the secondary xylem and phloem,

1. Illicium L. Syst. Nat. ed. 10. 1050. 1759.

Glabrous evergreen shrubs or small trees with thin-coriaceous, exstipu-
late, entire, alternate or distally clustered leaves, the blades decurrent on
the petioles. Flowers solitary or 2 or 3 together in minute glomerules,
axillary or sub-terminal and appearing crowded among leaves toward the
tips of branchlets, sometimes bracteolate, subtended by several caducous
bracts. Flowers perfect, hypogynous, the parts free. Receptacle convex
to short-conical, usually concealed by carpel-bases. Perianth segments
numerous (12-15 or 21-33 in ours), several seriate. Stamens numerous
(6-7 or 30-38, rarely —SO, in ours), 1-seriate or 2- or 3-seriate, erect, with
fleshy filaments and basifixed 4-sporangiate anthers introrsely dehiscent
by longitudinal slits; pollen tricolpate. Carpels 11-15(—20), free, in a
single whorl, each composed of a laterally flattened ellipsoid ovary distally
attenuate into an acute style; style vascularized, conduplicate, stigmatic
along the upper side along all or most of its length; ovary unilocular, with
a single anatropous ovule on the adaxial side near the base. Fruit a fol-
licetum of a single whorl of free, spreading follicles 10-18 mm. long, these
dehiscing along the upper side. Seed with a sub-basal hilum, ellipsoid and
laterally flattened, 5-7 mm. long, brownish, glossy; endosperm copious,
oily; embryo minute, near the hilum. Typr sprectes: J. anisatum L. (The
name Latin, #dicium, an allurement, in reference to the fragrance of the
fruits of J. verum Hook f., an economically important species confused by
Linnaeus and others with /. anisatum.) — ANISE-TREE.

A genus of some 37 species of southeastern Asia and 5 of southeastern
North America. Two species, representing each of the 2 sections, occur
entirely within our area. Our species are easily recognized by the fragrant
(when crushed), thin-coriaceous, more or less evergreen leaves, the red to
yellow relatively small flowers with numerous perianth segments, and the
very characteristic star-shaped fruit, a ring of follicles.

Section ILtictum (Badiana Spach), with 13 species in which the inner

1958] WOOD, GENERA OF WOODY RANALES 317

Fic. 2. Illicitum. a-i, 7. floridanum: a, fruiting branchlet, * 1/3; b, opening
bud with carpels receptive to pollen, X 3; c, flower, later stage at shedding of
pollen, the carpels connivent, < 1; d, stamens, inner, outer, and an unusual
sub-tepaloid form, adaxial views, < 5; e, two carpels on axis, other flower parts
removed, & 3; f, carpel, vertical section’ to show folded, infused sporophyll
with single ovule, X 10; g,. mature carpel, in section, with horny endocarp
(striped) and seed with horny outer seed-coat (black) and endosperm (dotted),
x 2; h, mature carpel, cross-section, the mature embryo at micropylar end of
seed (to right), X 2; 1, seed, X 2; f, g, h, semi-diagrammatic. j, J. parviflorum:
stamens, adaxial view, X 5.
perianth segments are thin, narrowly oblong or ligulate or lanceolate and
somewhat lax at anthesis, is represented with us by J. floridanum Ellis

2n = 28), a very well marked species which ranges from northwestern

Florida to eastern Louisiana and northward to central Alabama at low
elevations, especially in wet areas (e.g., around bay-heads). The only
close relative of this species is the very similar and perhaps conspecific
I. mexicanum A.C. Smith, of Veracruz and Tamaulipas. Both plants differ
from Old World members of the section in their comparatively long pedi-
cels, numerous stamens and brightly colored perianth segments.

The proterogynous flowers of J. floridanum, borne from March to May,
are showy, deep red or purple, with an intensely unpleasant odor, both
features suggestive of pollination by carrion flies. The species is hardy
as far north as Philadelphia where, however, it may lose its usually ever-
green habit. The leaves were used with those of lex Cassine as a tea by
the Indians of western Florida.

318 JOURNAL OF THE ARNOLD ARBORETUM | [volL. xxxrx

Section CyMBOSTEMON (Spach) A. C. Smith, with 29 species in which
the inner perianth segments are fleshy to paper-like, usually ovate to sub-
orbicular and not lax at anthesis, occurs in our area as J. parviflorum
Michx. ex Vent., a shrub or small tree which appears to be restricted to a
few counties (Lake, Marion, Seminole and Volusia) at the headwaters
of the St. Johns River in northeastern Florida. With its obtuse leaves,
small, yellowish flowers produced in May and June and reduced number
of stamens (6 or 7), this species is unlikely to be confused with J. flori-
danum. Beyond our area the section is represented in the New World by
I. cubense A. C. Smith, of eastern Cuba, and J. Ekmaniu A. C. Smith,
of western Hispaniola.

Illicium verum, Star anise of southeastern China and adjacent Indo-
China, provides a volatile oil used as a medicine, a condiment, or in flavor-
ing liqueurs such as absinthe and anisette. The seeds of /. anisatum L.
(2n = 28), of Japan, contain a poisonous alkaloid.

REFERENCES:

BaiLey, I. W., ano C. G. Nast. Morphology and relationships of Jdlicium,
Schisandra and Kadsura. J. Stem and leaf. Jour. Arnold Arb. 29: 77-89.
1948.

Gray, A. Magnoliaceae. Genera Pl. U. S. 1: 52-64. pl. 21-25. 1848. I. flori-
danum,

LemMesLe, R. Les ponctuations aréolées des fibres des genres Schizandra L., Kad-
sura J., [licium L. et leu r rapports avec la phylogenie. Compt. Rend. Acad.
Sci. Paris 221: 113-115.

OswaLp, F., Jr. Beitrage zur Penukniss der sae ala der Friichte des
Steranis, /llicium anisatum. 47 pp. Marburg, 18

SCHLOTTERBECK, J. O., AND C. R. ECKLER. The caer and development of the
fruit of Jlicium foridass, Am. Pharm. Assoc. 49: 485-589. 1901;
Pharm. Arch. 4: 190

SmitH, A. C. The ails raat and Schisandraceae. Sargentia 7: 1-224.

1947. |A complete monographic treatment.

Sze, Y. C. fllicium religiosum Siebold, Mang Tsao, a phytochemical study. Am.
Jour. Pharm. 101: 505-574, 622-637, 687-716. 1929.* [7. anisatume. |
TIEGHEM, P. VAN, Sur les dicotylédones du group des Homoxylees. Jour. de Bot.
(Morot) 14: 259-297, 330-361. 1900. [349-354, anatomical features of

llicium, which should not be included in Winteraceae. |

WHITAKER, T. W. Chromosome number and relationship in the Magnoliales.

nold Arb. 14: 376-385. 1933. [Includes undocumented counts for
Z ies (also as J. aaa ) and J. floridanum. |

SCHISANDRACEAE (ScHISANDRA FAMILY)

A family of two undoubtedly closely related genera of climbing vines,
Schisandra Michx. (ca, 25 species) and Kadsura Kaempf. ex Juss. (ca. 22
species), all of eastern and southeastern Asia and Malaysia, with the ex-
ception of a single species of Schisandra in the southeastern United States.
The two genera have numerous features in common with /llictum (q.v.)
but differ in the scandent habit, unisexual flowers, enlargement of the
torus after anthesis, arrangement of the carpels, absence of a vascularized

1958 | WOOD, GENERA OF WOODY RANALES 319

style, indehiscent drupe-like carpels, unilacunar nodes with three traces,
typically eu-stelic arrangement of the primary vascular tissues, relatively
primitive type of secondary xylem and presence of crystal-bearing scleren-
chyma. Although the group is of great phylogenetic and phytogeographic
interest, its economic importance is restricted to the few Asiatic species
which occasionally are cultivated as ornamentals.

1. Schisandra Michx. Fl. Bor.-Am, 2: 218. 1803, nom. cons.

Monoecious or dioecious (?) woody climbing vines, the leaves simple,
petiolate, exstipulate, alternate on long-shocts or congested on short-shoots,
the blade oblong-elliptic to ovate or lanceolate, pinnate veined, entire to
sinuate or remotely undulate-denticulate. Flowers pedicellate, unisexual,
solitary in the axils of caducous bracts or foliage leaves near the base of
the annual shoots, sometimes subtended by 2 or 3 minute secondary bracts.
é flowers: tepals 9-12 [5-20], free, 2- or 3-seriate, all similar, the largest
elliptic to obovate, 5-8 mm. long; androecium a sessile flattened fleshy
5-cleft pentagonal shield, consisting of 5 radiating stamens with the con-
nectives fused into the shield, the anther-sacs borne on the lower margins
of the anthers: opening longitudinally; [in other species stamens 4—60,
variously aggregated; pollen 3- or 6-colpate.] @ flowers: tepals similar
to 6+: gynoecium consisting of a receptacle distinctly longer than broad
(1.5-3 mm. long), and numerous (12—) 20-30 [-120], 3—5-seriate carpels;
ovary ellipsoid to obovoid, the wall fleshy, with stigmatic crests produced
distally into an acute, unvascularized pseudostyle and proximally into an
irregularly oblong appendage; ovules 2(—3) superposed or obliquely super-
posed, attached to the adaxial wall of the carpel above the base. Fruit
with a greatly elongated receptacle (2-3 cm. long, 2-3 mm. diameter ),
the pedicel remaining slender, the receptacle bearing 7-12 carpels spaced
on its surface; carpels becoming berries, usually ellipsoid or subglobose, the
pericarp red, fleshy; seeds 2, ellipsoid-reniform, more or less rugulose;
endosperm copious, oily; embryo small, near the hilum. (Stellandria
Brickell, 1803, nom. rejic.) Typr spEcies: S. coccinea Michx. = S. glabra
(Brickell) Rehder. (The name from Greek schisis, a cleaving, and andros,
of a man, in reference to the “fissures” between the anthers. )

About 25 species of eastern Asia (Manchuria southward to northern
Indo-China and the Himalayas, Java and Sumatra), a single species, S.
glabra in southeastern United States, entirely endemic to our area. Schis-
andra glabra apparently is a very rare plant, being rather poorly known
from few and widely scattered localities mostly on the coastal plain in
southeastern South Carolina, Georgia, western Florida, Alabama, Missis-

red, and the anther-sacs yellow. The greatly elongated receptacle which
bears the red or scarlet fruits (July-August) is especially noteworthy.

320 JOURNAL OF THE ARNOLD ARBORETUM _ [vo.. xxxrx

The genus has been divided by A. C. Smith into four sections based
primarily on various modifications of the androecium, which in the most
primitive section (Pleiostema) may be composed of relatively numerous
(—60) and essentially free stamens. Our species belongs to section Schis-
andra in which the androecium is highly modified, flattened and _ shield-
like, composed of five united stamens, and presumably representing one
of the end-products of the genus. In addition to S. glabra, the section
includes S. repanda (Sieb. & Zucc.) A. C. Smith, of southern Japan and
southern Korea, and S. bicolor Cheng, of northwestern Chekiang, China.
All three are strikingly similar in characters of perianth-segments, androe-
cium and gynoecium and furnish still another example of the now-familiar
pattern of disjunction between eastern Asia and eastern North America.

The chromosome numbers of S. ($ Maximowiczia) chinensis (Turcz.)
Baill., S. ($ Pleiostema) sphenanthera Rehder & Wilson, and Kadsura
japonica (L.) Dunal have all been reported as 2” = 28.

REFERENCES:

Bartey, I. W. anp G. Nast. Morphology and relationships of Jllicium, Schis-
andra and Kadsura. 1. Stem and leaf. Jour. Arnold Arb. 29: 77-89. 1948.

Gray, A. Ord. ees Genera Pl. U. S. 1: 52-64. pls. 21-25. 1848.
| Schisandra, 57-58, pl. 2

Kozo-PoLyJANsKkI, B. M. ee floral mechanism of Woo-we-dzy, Schizandra ie
ensis (Turcz.) Baill. Compt. Rend. Acad. URSS. IT. 53: 749-751 46.

LeMEsLE, R. Les ponctuations aréolées des fibres des genres Sobiconine is on
sura J., Illicium L. et leur rapports avec la phylogenie. Compt. Rend. Acad.
Sci. Paris 221: 113-115. 1945.

OzeENDA, P. Sur l’anatomie libéroligneuse des Schizandracées. Compt. Rend.
Acad. Sci. Paris 223: 207 -209. 1946.

REHDER, A. Schisandra Michaux, nomen genericum conservandum. Jour. Arnold
Arb. 25: 129-131. 1944,

— . C. The families Illiciaceae and Schisandraceae. Sargentia 7: 1-224,

[A comprehensive monograph.

Wrrnaxen, T. W. Chromosome number and relationship in the Magnoliales.
Jour. Amold Arb. 14: 376-385. 1933. [Includes Schisandra

ZAKORDONETS’, A. I. Propagation of Schisandra chinensis with green cuttings.
(In Ukrainian. Russian summary.) Bot. Zhur. Kyiv 12: 44-50. 1955.*

CANELLACEAE (WiLp CINNAMON FAMILY)

A small family of disjunct distribution with five genera of the West
Indies, Venezuela, Brazil, East Africa, and Madagascar. The family is
notable for the combination of alternate, exstipulate leaves vascularized
by 3 traces from 3 gaps, ethereal oil cells throughout the plant, wood with
a number of primitive anatomical features, 3 sepals, 4-12 petals in one or
more whorls, a monadelphous androecium forming a tube around the ovary
with the anthers extrorse, a single pistil with 2-6 parietal placentae, and
monocolpate pollen. Although placed in the Parietales in the Englerian
system, the relationships to the woody Ranales have been pointed out a
number of times. Within that group an affinity to Myristicaceae has been

1958] WOOD, GENERA OF WOODY RANALES S24

suggested by several investigators. The family has also been placed by
itself in the order Canellales near the Laurales (here included in the woody
Ranales).

REFERENCES:

Bonnet, E. Essai d’une monographie des Canellées. Paris. 64 pp. 1876. [Treats
Canellaceae as a tribe of Magnoliaceae. |

Cronoutst, A. Outline of a new system of families and orders of dicotyledons.
Bull. Jard. Bot. Bruxelles 27: 13-40. 1957. [Canellales. |

Garratt, G. A. Bearing of wood anatomy on the relationships of the Myristi-
caceae. Trop. Woods 36: 20-44. 1933. [Includes Canellaceae. ]

LeMESLE, R. Nouvelles remarques histologiques, microchimiques et phylogén-
étiques sur la famille des Canellacées. Revue Gén. Bot. 58: 193-201. 1951.

Miers, J. On the Canellaceae. Ann. Nat. Hist. III. 1: 349. 1858. [Notes simi-
larities of Canellaceae to Drimys and to Jlicium! |

Occutont, P. Nota sdbre a biologia das Canelaceas brasileiras. Arq. Jard. Bot.
Rio de Janeiro 8: 275-279. 1948. [Density and abundance of Cimnamoden-
dron regulated by the simultaneous action of a seed-attacking fly and a fruit-
eating bird. |

PritzEL, E. Der systematische Wert der Samenanatomie, insbesondere des Endo-
sperms, bei den Parietales. Bot. Jahrb. 24: 348-394. 1898. [Canellaceae,
eae

SoLEREDER, H. Systematische Anatomie der Dicotyledonen. 97-99. Stuttgart,
1899.

TrecHem, P. van. Sur les Canellacées. Jour. Bot. (Morot) 13: 266-276. 1899.

VANDER WYK, R., AND J. E. CanricHt. The anatomy and relationships of the
Annonaceae. Trop. Woods 104: 1-24. 1956. [Includes stem and wood
anatomy of Canellaceae. |

VesTaL, P. A. The significance of comparative anatamy in establishing the rela-
tionship of the Hypericaceae to the Guttiferae and their allies. Philipp.
Jour. Sci. 64: 199-256. 1937. [Canellaceae, 224, 225, 241, 242.]

Warsurc, O. Winteranaceae (Canellaceae). Nat. Pflanzenfam. I1(6): 314-319.
Nachtrage 251, 252. 1895-1897.

1. Canella P. Br. Hist. Jamaica 275. pl. 27. fig. 3. 1756; Swartz, Trans.
Linn. Soc. 1: 96. 1791, nom. cons."

Small tree, to 8-10 m. with gray bark and obovate to oblanceolate,
rounded or emarginate, deep green, lustrous, evergreen leaves. Flowers
small, perfect, regular, in terminal and axillary cymose inflorescences.
Sepals 3, imbricate, persistent, the petals 5, deep red, connate at the base.
Stamens 10, completely united in a tube with the 10 linear extrorse anthers
on its outer surface below the summit. Ovary superior, 1-locular with 2
parietal placentae and about 4 semi-anatropous ovules, the stigma 2-3
lobed. Fruit berry-like, red, with 2-4 shining black seeds. Embryo small
with a large amount of endosperm. Pollen monocolpate. (Winterana L.,
1759, nom. rejic.) Type species: Canella alba Murr. (= C. Winterana
(L.) Gaertn.). (Name from Low Latin, camella, cinnamon, from Latin,

* Conservation unnecessary.

bee JOURNAL OF THE ARNOLD ARBORETUM _[VvoL, xxxIx

canna, a cane or reed, applied to the bark which assumed the form of a
roll or quill in drying, the name given by Browne to the West Indian plant
“the Canella alba of the shops.”’) — Witp CINNAMON, CINNAMON-BARK.

Probably a single species wide-ranging in the West Indies and with
outlying stations in northern South America and in the region of Cape
Sable and the Florida Keys, in subtropical Florida, where the plant occurs
in hammocks with other tropical genera, generally in the shade of larger

ees,

Oil cells are conspicuous in most parts of the plant and the pale inner
bark has a cinnamon-like odor. It has been used as a spice, stimulant
and tonic. The wood is hard and very dense. The maroon flowers, glaucous
on the outside, are borne primarily in June and July (January?) and the
crimson fruit matures from March to April. The ripe fruit is eaten by
birds which probably disperse the plant. It is cultivated as an ornamental
to a limited degree.

The stamens are so closely united that the 10 anthers appear to be a
ring of 20 close-packed anther-sacs, each splitting longitudinally. How-
ever, ten vascular bundles are present in the tube composed of the fila-
ments.

The genus stands apart in the family by the 10 stamens and the 5 petals
coherent only at the base.

REFERENCES:

GREENISH, H. G. Canella bark, a study of its structure. Pharm. Jour. III. 24:
793-797. 1894.*

PLANCHON, G. Note sur la structure des écorces qui portent dans le commerce le

de Canelles. Bull. Soc. Bot. France 20(Session extraord.): xlvi-xlvii.
1873.

SARGENT, C. S. Canella. Silva N. Am. 1: 35-38. 1891; 14: 97. 1902.

SupwortH, G. B. On legitimate authorship of certain binomials with other notes
on nomenclature. Bull. Torrey Bot. Club 20: 40-46. 1893. [Canella Win-
terana vs. C. alba]

Swartz, O. The botanical history of the Canella alba. Trans. Linn. Soc. 1: 96—
102. 1791,

CALYCANTHACEAE (CALYCANTHUS FAMILY) .

Shrubs with opposite, entire, exstipulate leaves, numerous strap-shaped
tepals spirally arranged on a cup-shaped receptacle, the stamens at its
apex and extrorse and the carpels numerous, free, on the inner surface of
the receptacle, the mature receptacle resembling a large, dry rose-hip;
embryo large, the cotyledons convolute, endosperm lacking.

A small family of disjunct distribution including only two very similar
genera (sometimes combined): Calycanthus L. (about 4 species of the
eastern and western U.S.) and Chimonanthus Lindl. (2 or 3 species of
China).

1958 | WOOD, GENERA OF WOODY RANALES o26

The group is notable not only for its odd floral structure (highly modi-
fied receptacle and numerous free tepals, stamens and carpels) but for a
number of anatomical features, including ethereal oil cells, dicolpate pollen
(a modification of the monocolpate type), unilacunar nodes with a funda-
mentally double trace from a single gap, and 4 cortical vascular bundles
(with inverted orientation of xylem and phloem) which extend throughout
the stems of mature plants and which have branches entering the leaves
at the nodes.

Although some authors have placed the family in the Rosales, the total
evidence available clearly indicates that its relationships are with the
Ranales, especially that group of families characterized by monocolpate
and derived types of pollen and double-trace unilacunar nodes (or unilacu-
nar modifications): Austrobaileyaceae, Amborellaceae, Trimeniaceae,
Monimiaceae, Gomortegaceae, Lauraceae, Hernandiaceae, Chloranthaceae,
and Lactoridaceae. Its closest relationships are probably with the Moni-
miaceae, a tropical group, chiefly of the southern hemisphere.

1. Calycanthus L. Syst. Nat. ed. 10. 2: 1066. 1759, nom. cons.

Deciduous shrubs with opposite, entire, exstipulate leaves and red-brown,
purple-brown or greenish solitary flowers terminal on short, leafy axillary
branches of the season. Receptacle cup-shaped, bearing on its outer sur-
face and apex bracts and numerous undifferentiated strap-shaped, free,
rather fleshy tepals. Stamens numerous, on the edge of the receptacle, with
stout filaments, the apex of the connective prolonged, succulent, the anthers
extrorse; inner stamens reduced to staminodia; pollen 2-colpate. Gynoe-
cium of numerous free carpels within the receptacle; style filiform, stigma
terminal, the ovary 1-celled with 2 anatropous, 2-integumented ovules.
Fruit an indehiscent pseudocarp from the accrescent receptacle and tepal-
bases, somewhat resembling a large dry rose-hip, bearing within it the
numerous large achenes with tough exocarp. Seed solitary, large, lacking
endosperm, the embryo large, with convolute cotyledons. 2” = 22. (But-
neria Duham., 1755, nom. rejic.) Typr species: C. floridus L. (The
name from Greek, calyx, a cup, and anthos, flower.) — STRAWBERRY BUSH,
SWEET-SHRUB, BUBBY BLOSSOM, SWEET BUBBY, SWEET BETTIE, SPICEBUSH.

Perhaps three species of the eastern United States and a single well
marked, less closely related species, C. occidentalis Hook. & Arn., of the
North Coast Ranges and Sierra Nevada foothills of California. The 2 or
3 species of the Asiatic Chimonanthus, in which the number of stamens
is reduced to five, are sometimes included in Calycanthus as a separate
section. Calycanthus fertilis Walt., C. floridus L., and C. Mohri (Small)
Pollard are generally recognized in our area, but the range of variation
and the true limits of the species are not well understood. Fruit char-
acteristics have been little used but may provide good taxonomic char-
acters.

324 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

I'iG, 3. Calycanthus. a-h, C. floridus: a, flowering branchlet, * 1/2; b, flower,
vertical section, to show carpels, stamens, staminodia, and cup-shaped “recep-
tacle,” X 2; c, two stamens, lateral view, X 5; d, stamen, abaxial view, < 5;
e, inner, reduced stamen, abaxial view, * 5; f, two staminodia from edge of
cup, lateral view, X 5; g, carpels, X 5; h, mature dry pendulous pseudocarp,
x 1/2; b, g, semi-diagrammatic. i-k, C. fertilis: i, pseudocarp, vertical section,
with mature carpels, some removed, * 1/2; j, mature carpels, lateral and
abaxial views, < 1; k, seedling with unfurling cotyledons, « 1/2.

Agamospermy (which may account for some of the taxonomic difficul-
ties) has been reported in Calycanthus fertilis, C. floridus, C. occidentalis
and Chimonanthus praecox. Embryos seem to be of nucellar origin,
although parthenogenesis has been claimed for the same species. Pseu-
dogamy appears to be the rule, pollination being necessary for the develop-
ment of the endosperm, without which the embryo does not grow. Caly-
canthus floridus var. ovatus (Ait.) DC., presumably of garden origin, has
been reported to be a triploid (2” = 33), with about 50 per cent sterile
pollen.

The flowers are proterogynous. Pollination in Calycanthus occidentalis
has been shown to occur through the agency of Colopterus truncatus (Ran-
dall), a small nitidulid beetle.

Ethereal oil cells occur especially in bark, leaves, and tepals. The
flowers are quite variable in fragrance but some forms have an extremely
pleasant spicy, strawberry-like odor when crushed. The seeds contain an
alkaloid, calycanthine, with a physiological action similar to strychnine;
poisoning of cattle and sheep eating the fruits has been reported in
Tennessee.

1958 | WOOD, GENERA OF WOODY RANALES 325

REFERENCES:

Brorrerio, I. Osservazioni sulla sviluppo delle Calycanthaceae. Ann. Bot. Roma
18: 387-394. 1930. [Pseudogamy, nucellar embryos.

Boureav, E. L’evolution vasculaire du Calycanthus floridus L. (actuel: Cal
canthacées) et l’explication du systéme vasculaire du Zygopteris ee
B. R. (filicale palaeozoique; zygopteridées). Bull. Mus. Hist. Nat. Paris I.
18: 440-447. 46,

CHEADLE, V. I., AND K. Esau. Secondary phloem of Calycanthaceae. Univ. Calif.
Publ. Bot. 29: 397-510. 1958. [C. floridus, C. occidentalis, Chimonanthus
praecox. |

Daumann, E. Das Bliitennektarium von Magnolia und die Fiitterkorper in der
Bliite von Calycanthus. Planta 11: 108-116. 1930. [C. floridus.

Diets, L. Kaferblumen bei den Ranales und ihre Bedeutung fur die Phylogenie
der Angiospermen, Ber. Deutsch. Bot. Ges. 34: 758-774. 1916. [Beetle
pollination in Eupomatia and C. occidentalis. |

Faun, A., AND I. W. Battey. The nodal anatomy and the primary vascular
cylinder of the Calycanthaceae. Jour. Arnold Arb. 38: 107-117. 1957.

Grant, V. ae pollination of Calycanthus occidentalis. Am. Jour. Bot. 37: 294-
297. 5

KEARNEY, ie Ee Jr. The nomenclature of the genus Biittneria Duham. Bull.
Torrey Bot. Club, 21: 173-175. 1894.

Licnrer, O. Recherches sur les massifs libéro-ligneux de la tige des Calycanthées.
Bull. Soc. Bot. France 31: 128-132. 1884

_ Sur la valeur morphologique des massifs libéro-ligneux corticaux des
tiges des Calycanthées. Compt. Rend. Acad. Sci. Paris 98: 700-702. 1884.

erches sur l’anatomie comparée des Calycanthées, des Melastom-
acées et He Myrtacées. Arch. Bot. Nord France 3: 1-455. 1886-1887.

Provvier, V. Contribution 4 l’étude biochimique de quelques Calycanthacees,

Magnoliacées et Rosacées. (Summary.) Ann. Sci. Nat. XI. 7: 237-238

Potxarp, C. L. Calycanthaceae. N. Am. Fl. 22: 237-238. 1908.

Qurntan, C. E. Contributions toward a knowledge of the anatomy of the lower
dicotyledons. III. The anatomy of the stem of the Calycanthaceae. Trans
Roy. Soc. Edinb. 52: 517-530. 1920.

sates F. Seedlings of as woody plants. Minn. Bot. Stud. 2: 69-86.
1 ve floridus, 72—

mparative Fe. of hypocotyl and epicotyl in woody plants. Loc.
a De Ge 136. [C. floridus, 100-102.

Sax, K. Chromosome behavior in Calycanthus. Jour. Arnold Arb. 14: 279-281.
1933. [Undocumented counts of C. fertilis, C. floridus, C. floridus var.

ov atus a

Scuaeppr, H. Morphologische Untersuchungen an den Karpellen der Calycanth-
aceae. Phytomorphology 3: 112-118. 1953. [Carpel conduplicate, 3- veined;
C. floridus, Chimonanthus praecox.]

Scutiruorr, P. N. Zur Apogamie von Calycanthus. Flora 116: 73-84. 1923.
[ Parthenogenesis. |

SmitH, G. H. Vascular anatomy of Ranalian flowers, — II. Ranunculaceae
(continued), Menispermaceae, Calycanthaceae, Annonaceae. Bot. Gaz. 85:
152-177. 1928. [Calycanthus floridus.

Srerns, E. E. The fruit of Calycanthus, L. Bull. Torrey Bot. Club 15: 205-209.
1888. [C. fertilis, eastern Tenn. ]

326 JOURNAL OF THE ARNOLD ARBORETUM _[vot. xxxrx

WituraAMs, J. L. The sieve-tubes of Calycanthus occidentalis. Ann. Bot. 8: 367-
370 4

LAURACEAE (Laure. FAmMIty)

Evergreen (mostly) or deciduous trees or shrubs with alternate [or
sometimes opposite or subopposite], exstipulate, entire or rarely lobed,
pinnately veined, subtripli-veined |tripli-veined or 3-veined] leaves, (ex-
cept Cassytha, a greatly reduced parasite resembling Cuscuta); wood and
leaves usually with ethereal oil cells and often with mucilage cells. In-
florescences usually axillary, basically cymose, paniculate or reduced and
sub-umbellate, racemose, spicate [or capitate]. Flowers bisexual or uni-
sexual (the plants then dioecious or polygamo-dioecious), small, regular,
generally whitish or yellowish, often hairy. Perianth regular, of 6 tepals
in 2 whorls of 3, similar or the outer smaller, free nearly to the base or +
united to form a perianth tube on which the stamens are inserted. Stamens
basically 12, in 4 series of 3, the outermost (series I) opposite the outer
whorl of tepals, the succeeding series alternating; any one or more series
(in all our genera) reduced to staminodia or altogether lacking and the
filaments of one (series III) or more series of fertile stamens flanked —
stalked or sessile ‘‘glands.” Anthers 4- or 2-locular, introrse or extro
upwardly dehiscent by 4 or 2 flap-like valves: pollen sticky, the aa
of each locule raised upward by the valve, grains non-aperturate, single.
Pistil 1, with a single stigma and style and a 1-locular ovary with a single
pendulous, anatropous, 2-integumented ovule; ovary free from the peri-
anth tube, although sometimes + surrounded by it. Fruit a berry or
drupe; perianth lobes persistent [often accrescent| or deciduous in fruit,
the perianth tube and pedicel often greatly enlarged to form a cupule sub-
tending the fruit, in a few (e.g., Cassytha) the perianth tube completely
surrounding the fruit but free from it. Seed lacking endosperm, the embrvo
large with fleshy plano-convex cotyledons, a small plumule and radicle;
germination usually hypogeal. (Including Cassythaceae. )

—

A family of perhaps 30-40 genera and about 2500 species, mostly ever-
green and primarily of the tropics and warm-temperate areas of both hemi-
spheres, especially Central and South America and southern Asia. Repre-
sented with us by about 11 native and 2 more or less naturalized species
in 8 genera.

The family is easily recognized by the small, regular, usually 3-merous
flowers with their curious 4- or 2-locular stamens dehiscent by as many
flap-like valves, the 1-locular ovary with a single pendulous anatropous
ovule, the baccate fruit often subtended by a cupule derived from the
accrescent perianth tube, and the large embryo without endosperm. Equally
well characterized by a unique combination of anatomical characters, the
Lauraceae seem to form a well-marked and natural family.

The relationships of the Lauraceae seem to be with that group of the
woody Ranales (sensu lato) which have secretory cells, unilacunar nodes

1958] WOOD, GENERA OF WOODY RANALES OZ

and monocolpate, dicolpate or derived (in this case non-aperturate) pollen
grains. The relationships are particularly with Monimiaceae (cf. subfam.
Hortonoideae, Atherospermoideae), Hernandiaceae, and Gomortegaceae
(note woody habit, non-aperturate pollen grains, stamens with associated
staminodes, valvular anthers, unilacunar nodes, simple and exstipulate
leaves, related alkaloids).

Although the family is a natural one, the generic (and specific) lines
are very difficult in some groups of Lauraceae and may be artificial in many
instances. Convergent tendencies may be noted again and again. Strong
emphasis has been placed upon the various permutations and combinations
possible within the 4 series of stamens and staminodes and upon the 4-
vs. 2-locular condition of the anthers (although both of the latter are
known to occur within some undoubtedly related groups, e.g., Cinnamo-
mum, Sassafras). Other important characteristics include inflorescence-
type and the development of the perianth tube and lobes and their condi-
tion in fruit (persistent, deciduous, cupules, etc.). Patterns of leaf-venation
and cuticle may be useful at the specific level but may vary widely within
related groups and similar patterns may occur in completely unrelated
species. (As a result, generic determinations of fossil materials based on
vegetative characteristics are, at best, dubious.)

The four trimerous whorls of stamens in the basic lauraceous flower are
designated here as series I-IV (cf. Mez), beginning with the outermost
whorl which is opposite the outer tepals. The stamens of series III are
usually flanked by nectaries (most often stalked and vascularized) which
seem to be staminodial in origin. Series IV, if present at all, is usually
staminodial.

Perfect flowers throughout the family probably are proterogynous, al-
though observations from living plants seem to have been made only on
two species of Persea, in which proterogyny is carried to an extreme (dian-
thesis). Stamens appear to elongate between the time the stigma is recep-
tive and that at which the pollen is shed. The anther valves open from
the base upward, recurving and carrying with them the entire contents of
the locules.

Only a few species have been examined cytologically, but in these 12
is the basic chromosome number.

REFERENCES:

ALLEN, C. K. Studies in the Lauraceae, VI. Preliminary survey of the Mexican
and Central American species. Jour. Arnold Arb. 26: 280-364, 365-434.
1945. [Much information pertaining to our genera. See also loc. cit. 20,
22, 23: 1939, 1941, 1942 for II-V dealing with Asiatic genera. |

. Lauraceae in Woodson & Schery, Flora of Panama. Ann. Missouri Bot.
Gard. 35. 1-68. 1948.

Banpbutska, H. On the cuticles of some fossil and recent Lauraceae. Jour. Linn.
Soc. 47: 383-425. 1925.

Janssonius, H. H. Mucilage cells and oil cells in the woods of the Lauraceae.
Trop. Woods 6: 3-4. 1926.

326 JOURNAL OF THE ARNOLD ARBORETUM _[voL. xxx1x

. De Slijm- en Oliecellen in het hout der Lauraceae in Nene met de

zienswijze van Tschirch. Pharm. Weekbl. 39: 1053-1055. 1927.

KosTERMANS, A. J. G. H. Studies in South American Seana Lauraceae
and Hernandiaceae, especially of Surinam. Meded. Bot. Mus. Utrecht 25:
1-70. 1936

. Revision of the Lauraceae I. Rec. Trav. Bot. Néerl. 33: 719-757. 1936.

[eile age II. 34: 500-609. 1937. [Includes Endlicheria, Cryptocarya

(Am n spp.), Licaria.] III. 35: 56-129. 1938. [Aiouea, Systemono-

pace peer ne Mezilaurus; notes on Licaria, Cryptocarya.| V.

35: 834-931. 1938. [ Anaueria, Beilschmiedia (American spp.), Aniba. |

(Reprinted as Meded. Bot. Mus. Utrecht 37, 42, 46, 48 [1937-1938],

respectively.

. Las Lauraceas Chilenas. Revista Univ. [Chile] 24: 201-232. 1939.

| Apparently Rev. Lauraceae IV.]

A historical survey of the Lauraceae. Jour. Sci. Indus. Res. Indonesia

1: 83-95, 113-127, 141-159. 1952. [Chronological account including taxo-

nomic notes. |

. Lauraceae. Inst. Paranaense Bot. Cat. Est. Gen. Bét. Fan. 34: 1-4.

1957. |List of genera and synonyms, estimate of no. of spp. |

. Lauraceae. Reinwardtia 4: 193-256. 1957. (First publ. as Comm.
Forest Res. Inst. Indonesia 57: 22 Mar. 1957.*) [Review of family, oe
stem, leaf, wood, inflorescences, flower, fruit, pollen, pollination, myrmec
phily, chromosomes, palaeontology, relationships, distribution, size of genera,
classification, key to genera, brief descriptions of genera and synonymy. |

Mez, C. Morphologische Studien tuber die Familie der Lauraceen. Thesis, 32
pp. Berlin, 1888.

Lauraceae Americanae. Jahrb. Konigl. Bot. Gart. Berlin 5: 1-556.
1889. The basic framework for American studies; attention focused on
inflorescence. |

Spicilegium Laureanum. I. Versuch einer pflanzengeographischen
Anordnung der tropisch-amerikanischen Lauraceen. [Survey of distribu-
tion of American Lauraceae, pp. 71-105.] II. Zusatze zu meiner Monogra-
phie der amerikanischen Lauraceen im Jahrbuch des Berliner Gartens Vol.
V. Arbeit. Konigl. Bot. Gart. 1: 71-166. 1892.

Mrrande, M. Sur Jorigine pluricarpellaire du pistil des Lauracées. Compt.
Rend. Acad. Sci. Paris 145: 570-572. 1907. [Cassytha and others: pistil
basically 3-carpellate. |

Money, L. L., I. W. Bartey, anp B. G. L. Swamy. The morphology and rela-
tionships of the Monimiaceae. Jour. Arnold Arb. 31: 372-404. 1950.

NEES VON EsENBECK, C. G. Systema Laurinum. 704 pp. Puc. 1836. [The
basic monograph. |

OnrTant, F. Uber den Bau des Blattrandes bei Myrtaceen, Hamamelidaceen,
Lauraceen, und Monimiaceen. Festschr. A. Tschirch. 299, 1926.* [See
Just, Bot. Jahresb. 54(2): 456. 1926.|

Pax, F. Lauraceae. Nat. Pflanzenfam. III(2): 106-126. 1891. [ Artificial classifi-
cation

PETZOLD, V. Systematische-anatomische Untersuchungen iiber die Laubblatter
der amerikanischen Lauraceen. Bot. Jahrb. 38: 445-474. 1907.

1958 | WOOD, GENERA OF WOODY RANALES 329

Recorp, S. J. AND R. W. Hess. American timbers of the family Lauraceae.
Trop. eG ea 7-33. 1942

SastrI, R. L. Studie es in Nea ueiceace II. Proc. Indian Sci. Congr. Assoc.
42(3, 2) 225- 226. 1925.* [ Morphology. |

STERN, W. L. Comparative anatomy of xylem and phylogeny of Lauraceae.
Trop. Woods 100: 1-72. 1954. [Includes a general review of position of
family; extensive bibliography. |

Tupper, W. W. A comparative study of the lauraceous woods. Am. Jour. Bot.
14: 520-525. 1927.

Key TO THE GENERA OF LAURACEAE

A. Foliose trees and shrubs; not parasitic. Subfam. LAUROIDEA
B. Plants evergreen; inflorescences variously cymose- Beer with no
involucre at the base; flowers bisexual; anthers of at least one series of
stamens (III) extrorse. Tribe PERSEEAE.
C. Anthers 4-loculed; fertile stamens 9, the 2 outer series introrse; fruits
without cupules or the cupules without evident double margin
D. Staminodia of series IV large, cordate-stipitate; locules of anthers
of 2 outer series in 2 planes, one above the other
E. Fruit without a cupule at the base, the not greatly enlarged
perianth persistent (or sometimes completely deciduous) ;
leaves pinnately veined, lacking glands in the axils of the main
veins beneath; perianth lobes unequal to subequal. .. 1. Persea.
. Fruit with a shallow cupule at the base, the perianth lobes de-
ciduous from the enlarged tube; leaves pinnately or subtripli-
veined, with conspicuous glands in the axils of the main veins
beneath; perianth lobes equal. ............ 2. Cinnamomum.
D. Staminodia of series IV small, inconspicuous, stipiform; locules
of anthers of 2 outer series arranged in an arc. .... 3. Nectandra.
C. Anthers 2-loculed; fertile stamens only 3, extrorse; fruits subtended
by thick cupules with evident double margins. .......... 4. Licaria.
. Plants deciduous, the small yellow flowers produced before or with the
unfolding leaves; inflorescences racemose or sub-umbellate in the axils
of enlarged bud-scales (Sassafras) or with an evident involucre of scales
at the base; flowers usually unisexual; fertile stamens 9, the anthers of
all series introrse. Tribe LITsEEAE
F. Inflorescences racemose, at the tips of branches, subtended by the en-
larged, involucre-like bud-scales; leaves unlobed or with 2 or 3 lobes;
fruit a dark blue drupe on a swollen red cupule; anthers 4-loculed;
SHEUDS ROL ELCES 5-2 ces re eee ee ae te te LORIN ious:
F. Inflorescences sub- umbellate, involucrate with 4 or 5 en scales;
fruits bright red on unswollen or slightly swollen pedicels; shrubs.
G. Anthers ghee inflorescences nearly sessile; branchlets not
Evidently: Zip - Zags token rues ate eee tnt dat Wen . Lindera
G. cane ene inflorescences clearly pedicellate; branchlets

tH

w

CAA: Va RON A I ee ae eat ee og on ee
A. Parasitic orange a green twining herbs of tropical Florida; leaves reduced
o scales; anthers 2-loculed; fruit surrounded by the Percent fleshy perianth
tubes 2oubiame <CACGSVTHOIDEARS cae ee ie Ne eee, 8. Cassytha.

330 JOURNAL OF THE ARNOLD ARBORETUM _ [vor. xxx1x

Subfam. LAUROIDEAE Kosterm.

Tribe PERSEEAE Mez

1. Persea Miller, Gard. Dict. Abr. ed. 4. 1754, nom. cons.!

Evergreen trees and shrubs with chartaceous to coriaceous pinnately
veined (usually more or less pubescent) leaves. Inflorescences axillary,
peduncled, cymose or rarely sub-umbellate (usually described as paniculate
but in ours small, with the exception of P. americana). Flowers bisexual,
small. Perianth lobes 6, free nearly to the base, the 3 outer ones usually
shorter than the inner ones, hairy and persistent in fruit. Fertile stamens
9, staminodia 3; stamen filaments slender, hairy, the anthers 4-loculed, the
bases of the 2 upper locules laterally tangential to the apices of the 2 lower:
stamens of series I & II introrse; stamens of series III extrorse (or the
2 upper locules lateral and the lower extrorse), the filaments flanked by
2 glands near the base; series IV sterile, the staminodia stipitate with
cordate-sagittate tips. Ovary subglobose, the style slender, usually pubes-
cent. Fruit baccate, small and globose or [in P. americana and relatives |
large and fleshy and obovoid to pyriform (often obliquely so), borne on
the spreading perianth-lobes and scarcely enlarged pedicel; [perianth
lobes occasionally deciduous in fruit in some forms of P. americana].
(Farnesia Heist., 1763, nom. rejic.; incl. Tamala Raf.) Type species:
Laurus Persea L. = P. americana Miller. (The name an ancient one used
by Theophrastus, transferred by Plumier to one of the tropical American
species of Nectandra and afterward adopted by Linnaeus. )

As usually delimited, a genus primarily of tropical America but reaching
south to Chile and northward in our area to Delaware and Arkansas and
with a single species P. indica (L.) Spreng., in the Canary Islands. Two
or three native species occur in our area; P. americana, in general culti-
vation throughout tropical America and an important crop in southern
Florida, persists after cultivation and has escaped to hammocks. Koster-
mans has extended the limits of Persea to include a number of Asiatic
genera.

Persea Borbonia (L.) Spreng., the red bay, is a handsome tree of the
borders of streams and swamps and ‘‘bayheads,” ranging from Florida to
Texas, northward to southern Arkansas and Delaware. A variant, Persea
palustris (Raf.) Sarg., distinguished primarily on the basis of pubescence
is currently treated as f. pubescens (Pursh) Fern. (2m = 24). A third
plant, P. littoralis Small, of coastal dunes of Florida, is a small tree with
small leaves mostly obtuse at the apex and smooth beneath. The variation
in this group is in need of further study. Persea humilis Nash, scrub bay,
is a distinctive shrub or small tree of the Pinus clausa—Ceratiola “scrub”
of central Florida. Its habit and habitat, reduced inflorescences, somewhat
larger fruits, and small leaves silky with shining golden hairs beneath, are
characteristic,

* Conservation unnecessary.

1958 | WOOD, GENERA OF WOODY RANALES Rppl

The avocado, Persea americana (2n = 24) introduced into Florida at
an early date by the Spaniards and long casually cultivated there, has
become an increasingly important fruit crop in southern Florida in the last
40 years. Commercial plantations are concentrated especially in Dade
County (about 4/5 of the total crop) but extend about as far north as
Cape Canaveral, on the east coast, and Tampa, on the west. Scattered
trees of the hardier cultivars may be found considerably farther north,
however. A large technical and horticultural literature including much
of the information available on the genus has accumulated in connection
with the cultivation of this species.

Persea americana is a complex species of very wide distribution in culti-
vation in tropical America. Its origin presumably lies in Central America
and involves the group composed of P. americana, its var. drymifolia
(Schlecht. & Cham.) Blake, P. Schiedeana Nees, P’. floccosa Mez, P.
nubigena L. O. Williams, and others, all of the uplands of this region. At
the present time three general groups of cultivars are recognized: West
Indian (originally introduced there from Central America by the Span-
iards), Guatemalan, and Mexican. The three groups differ in characters
of foliage, fruit, time of flowering and ripening of fruits, and hardiness.
Persea americana var. drymifolia, of the Mexican uplands, is most inti-
mately involved in the Mexican cultivars which are the important com-
mercial types in California but which are not well adapted to the conditions
of tropical Florida, where Guatemalan and West Indian forms grow best.
In recent years hybrids between members of the three groups have been
made and some of these are quite successful (e.g., the ‘Boorn’ cultivars,
‘Hickson,’ and ‘Luta’).

The flowers of avocados are proterogynous and have two periods of
opening on successive days (dianthesis), the flowers closing in between.
“A” and “B” types are recognized. In the former the stigmas are receptive
in the morning; the flowers close by early afternoon to re-open the follow-
ing afternoon when the pollen is shed. In type “B” cultivars, the first
(pistillate) opening occurs in the afternoon and the second the following
morning. All of the flowers open on a tree at a given time will be in the
same stage: thus self-pollination is virtually impossible and both “A” and
“B” cultivars are necessary for cross-pollination, an important considera-
tion in the planting of commercial groves. In a sub-type of the “B” form
the first period opening is mostly suppressed, with only a few flowers open
in the afternoon and most having only a single fairly long period of opening
in the morning when the stigmas appear to be receptive and the pollen is
shed. Such trees set very few fruit. Beyond this species, observations on
pollination appear to have been made only on P. Skutchu C. Allen, of
Costa Rica and Panama, which also shows dianthesis with equal numbers
of “A” and “B” plants but with the two periods of anthesis between dawn
and late-morning (10-11 a.m.) and late-morning and early afternoon (2-3
P.M.) on successive days.”

2 Since the above was written, it has been possible through the kind hospitality of
Mr. and Mrs. George R. Cooley for Dr. K. A. Wilson and the author to make some

332 JOURNAL OF THE ARNOLD ARBORETUM [voL. xxxrx

Persea is generally considered to be most closely related to Phoebe Nees,
a large genus of the tropics of both hemispheres. These genera are con-
ventionally differentiated on the basis of position of anther-locules and
nature of the perianth in fruit. Kostermans would reject the former as
a generic criterion and would restrict Asiatic Phoebe to those species with
an appressed and indurated perianth. The New World species might be
retained partly in Phoebe or assigned to Persea and perhaps Cinnamomum.
The same author greatly extends the limits of Persea to include the Asiatic
genera Machilus Nees, Nothaphoebe Blume, Alseodaphne Nees, Stem-
matodaphne Gamble, and Caryodaphnopsis Airy-Shaw, the first three of
which would be retained as subgenera.

REFERENCES:

ABRAMSKY, M., AND B. Brae. The pyruvate oxidation system in avocado
fruit panies: (Abs.) Pl. Physiol. 30 (sup.): xxvili-xxix. 1955.

ANDERSON, E. Variation in avocados at the Rodiles Plantation. Ceiba 1: 50-55.

1950. [Near Atlixco, ae Mexico; scatter diagram of mene in leaf-
characters in 132 tree

Brake, §. F. A eit tes of the North eer and West Indian
avocados (Persea spp.). . Wash. Acad. 10: . 1920.

Brincuurst, R. S. Sexual Pai problems of ie — Citrus Leaves
32(11): 26-28. 1952.*

. Breeding tetraploid avocados. Proc. Am. Soc. Hort. Sci. 67: 251-257.
1956.* [Colchicine; see also Genetics 41: 646. 1956.

Burcis, D. S. anp H. S. Wotre. Do avocado roots develop root-hairs? Proc.
Fla. State Hort Soc. (1945) 58: 197-198. 1946.*

Coriins, G. N. The avocado, a salad fruit from the tropics. U. S. Dept. Agr.
Fl. Ind. ae a: 1-52. 1905.

Cummincs, Kk., anp C. A. SCHROEDER. Sau of the avocado fruit. Calif.
Avocado Soc. "Year 27: 56-64. 1942.

FAIRCHILD, D. Two relatives of the — and their reintroduction into
Florida. Proc. Fla. State Hort. Soc. (1945) 58: 170-175. 1946. [P. Schie-
deana and se ae Anay. |

FERNALD, M. otanical specialties of the Seward Forest and adjacent areas

preliminary observations on dianthesis in Persea Borbonia and f, pubescens near
Brooksville, Hernando County, Florida. Four plants (one the glabrous form and three

species, the two eed f anthesis occur in the afternoon. Each of the four plants
held to a slightly different schedule, but the first afternoon period commenced about
noon and ended between 3 and 4:30 p.m. (Eastern Standard Time). The other period
of opening occurred between about 3:30 and 4:45. these flowers remaining open until
dark (about 7:30) or after, but those . three plants closing before 8 and those of the
fourth before 8:30. Flowers in hehe minate Meare anthesis are noticeably less
ee in their final closing aes an ee a few n “B” plants may overlap the opening

type) placed in a polyethylene bag showed later that afternoon and the following day
periods of anthesis which appeared to correspond roughly to those of P. Borbonia.

1958 | WOOD, GENERA OF WOODY RANALES 333

of southeastern Virginia. Rhodora 47: 93-142, 149-182, 191-204. 1945.
ee of status of P. Borbonia and f. pubescens, 149-151.]

Haas, A. R. C. Growth and water relations of the avocado fruit. Pl. Physiol.
te 383- 400. 1936.

. Chemical composition of avocado fruits. Jour. Agr. Res. 54: 669-687.
1937

HarrIis, J. A., AND W. PopeNoe. Freezing-point lowering of the leaf sap of the
horticultural types of Persea americana. Jour. Agr. Res. 7: 261-268. 1916.

Hom, T. The seedlings of Jatropha multifida L. and Persea gratissima Gartn.
Bot. Gaz. 28: 60-64. 1899. [P. americana.

House, H. D. Nomenclatorial notes on certain American plants — m.
Midl. Nat. 8: 61-64. 1922. [Transfers our native species to Borbonia Mill.
1754.]

Krome, W. H. Avocado growing in Dade County [Florida]. Ceiba 4: 339-350.

LEEMANN, A. Das Problem der asl ae Planta 6: 216-233. 1928. [ Develop-
ment of Hager cells in P. ca. |

LESLEY, J. W., Dio: Se eis Environmental conditions affecting
pollination a ar Calif. Avocado Soc. Yearb. 36: 169-173. 1951.*
Lyncu, S. J., AND R. O. NEtson. Current methods of vegetative propagation

of oe mango, lychee and guava in Florida. Ceiba 4: 315-337. 1956.
PoreNoE, W. The avocado in Guatemala. U. S. Dept. Agr. Bull. 743: 1-69.
1919. See also, Avocados as food in Guatemala. Jour. Hered. 9: 98-107.
1918.
Central American fruit culture. Ceiba 1: 269-367. 1952. | Persea, 304—
310.] In Spanish, Ceiba 3: 225-338. 1953 [Persea, 267-274.

Pratt, H. K., R. E. Younc anp J. B. Brae. The identification of ethylene as
a volatile product of ripening avocados. Pl. Physiol. 23: 526- a 1948.
Rapak, F. Wild volatile-oil plants and their economic importan I. Black

sage; 7 Wild sage; ae tae. bay. U. S. Dept. Agr. PI. Ind. ‘Bull. 235:
5-37. 1912. [P. Borbo
REECE, P. a The floral ae of the avocado. Am. Jour. Bot. 26: 429-433.

1939.
. Differentiation of avocado blossom buds in Florida. Bot. Gaz. 104:
323-328. 1943.
Roginson, T. R. Pollen sterility in the Collinson avocado. Jour. Hered. 21:

35-38. 1930.
Rotrs, P. H. The avocado in Florida; its propagation, ad and market-
ing. U. S. Dept. Agr. Bur. Pl. Ind. Bull. 61: 1-33.
Ryerson, K. Avocado culture in California. Part I. ee one varieties,
and Penne. Univ. Calif. Agr. Exp. oe Bull. 365: 1-61.
SARGENT, C. S. Persea. Silva N. Am. 7: 3-8. pls. 301, 302. oe ;
SCHROEDER, C. A. Pollen germination ie the avocado. Proc. Am. Soc. Hort.
Sci. 41: 181, 182. 1942.
. Multiple embryos in the avocado. Jour. Hered. 35: 209, 210. 1944.
. The structure of the skin or rind of the avocado. Calif. Avocado Soc.
Yearb. 35: 169-176. 1950.*
. Flower bud development in the avocado. Calif. Avocado Soc. Yearb.
36: 159-163. 1951.*
. Floral development, sporogenesis, and embryology in the development
of the avocado, Persea americana. Bot. Gaz. 113: 270-278. 1952.

334 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

. Abnormal fruit types in the avocado. Calif. Avocado Soc. Yearb. 38:
121-124. 1953/1954."

Proliferation of mature fruit pericarp tissue slices in vitro. Science
122: 601. 1955. [P. americana. |

: — production in avocado. Calif. Avocado Soc. Yearb. 39: 184-186.
1955

AND P. A. WIELAND. Diurnal fluctuation in size in various parts of the

avocado tree and fruit. Proc. Am. Soc. Hort. Sci. 68: 253-258. 1956.

Sxutcu, A. F. Observations on the flower behavior of the avocado in Panama.
Torreya 32: 85-94. 1932. [A, B, and modified B types in a dooryard plant-
in

. “The behavior of the flowers of the aguacatillo lea ae Tor-
reya 45: 110-116. 1945. | = P. Skutchi C. Allen, Costa R

SMALL, J. K. Tamala littoralis, Addisonia 17: 45-46. pl. 567. 192,

STONEBACK, W. J., AND B. CALvertT. Histology and chemistry of the avocado.
Am. Jour. Pharm. 95: 598-612. 1923.*

Stout, A. B. The flower mechanism of avocados with aha to pollination
and the production of fruit. Jour. N. Y. Bot. Gard. 2 . 1924.

The pollination of avocados. Florida Agr. Exp. .. Bull 257: 1-44.
1933.

Sturrock, D. Tropical fruits for southern Florida and Cuba and their uses.
Publ. Atkins Inst. Arnold Arb. Harvard Univ. 1: 1-131. 1940. [Persea,
67-71.]

Torres, J. P. Some notes on avocado flower. Philipp. Jour. Agr. 7: 207-227.

WALLACE, J. M., p R. J. Drake. Albinism ay Neeser oe of
avocado seedlings. Calif. Avocado Soc. Yearb. 40: 156-164. 1956

Wo re, H. S., L. R. Toy, ano A. L. STAHL. ee production in Florida.
Florida Agr. Exp. Sta. Bull. 272: 1-96. 1934

Wotre, H. S. mpligus varieties in Florida. Fruit Varieties & Hort. Digest
2(1): 14-17. 1947.

ZENTMYER, G. A. Diseases of the avocado. U.S. Dept. Agr. Yearb. 1953: 875-
881. 1953.

2. Cinnamomum Trew, Herb. Blackwell. Cent. 4, signature mM. ¢. 354.
1760; Blume, Bijdr. Fl. Nederl. Indié 568. 1825 [1826].

Evergreen trees with [or without] conspicuous buds with imbricate
scales, the alternate |or opposite] leaves pinnately veined, subtripli-veined,
[tripli-veined, or 3-veined|, with [or without] glands in the axils of the
veins beneath. Inflorescences paniculate, axillary, with deciduous bracts,
produced on the growth of the season. Flowers small, inconspicuous, bi-
sexual [rarely polygamous|. Tepals equal, | persistent or] deciduous from
the perianth tube. Fertile stamens 9, the anthers 4 [rarely 2]-loculed;
staminodia 3: stamens of series I & II introrse, glandless; stamens of series
III flanked by glands at the base of the filament, the anthers [extrorse or |
the 2 upper locules laterally extrorse, the lower extrorse; series IV of con-
spicuous stipitate cordate-tipped glandless staminodes. Stigma discoid or
peltate. Perianth tube accrescent in fruit, growing out into a thin cup
surrounding the base of the fruit; perianth lobes deciduous from the tube

1958] WOOD, GENERA OF WOODY RANALES 334

[or the basal part of the entire lobes persistent on the rim]. Fruit in ours
a black, globose drupe. (Including Camphora Trew, loc. cit. signature L.
t, 347: Nees in Wall. Pl. As. Rar. 2: 61, 72. 1831.) Type species: Laurus

Cinnamomum L. = C. zeylanicum [Garc.| Blume. (The name the Latin
transcription of Greek kinnamomon, derived, in turn, through Hebrew from
the ancient name for cinnamon.) — CINNAMON.

A large genus (100-275 species) of eastern Asia, with the largest number
in India, Indo-China, China, and Japan, but also in the Philippines, Indo-
nesia, New Guinea, Polynesia and Australia. No species are currently
recognized from the western hemisphere, although Kostermans has sug-
gested that some of the American species currently assigned to Phoebe Nees
may well belong to Cinnamomum.

Section CaMPpHoRA (Trew) Meissn., characterized by completely de-
ciduous perianth lobes, conspicuous perulate vegetative buds, and alternate
leaves with pinnate venation (rarely 3-veined or tripli-veined) and with
glands in the axils of the veins, is represented with us by C. Camphora
(L.) Nees and Ebermaier (2n = 24). The wood of this species, native
to the warm-temperate and subtropical rain-forest zone of eastern Asia
from southern Japan to northern Indo-China but now widely planted in
the tropics throughout the world, is the source of camphor, which is re-
moved by distillation. The plant was introduced into Florida as early as
1875 as a shade tree and was later established in large plantations in a
not very successful attempt to promote a camphor industry in competition
with that of Formosa and Japan. At the present time this handsome ever-
green with conspicuous scaly buds, very small glaucous flowers (March—
April) and black, globose drupes is cultivated in our area as an ornamental
tree and for windbreaks from southern Georgia and Florida to southern
Louisiana (also s. Texas and Calif.). It is hardy wherever the temperature
does not fall below 15°F and has become naturalized to varying degrees
throughout this region.

Section CINNAMOoMUM (Malabathrum Meissn.), in which the perianth
lobes are persistent or absciss above the base (leaving the tube crowned
by the truncate lobes), the leaf-buds naked or with obsolete scales, and
the leaves opposite or subopposite, 3- or tripli-veined and without glands,
occurs only in cultivation in the warmer parts of Florida. Cinnamomum
zeylanicum (2n = 24), the cinnamon of commerce, is a tender plant in
sub-tropical Florida, while C. Cassia Blume, of China, is a more hardy
species in that area.

The flowers of Cinnamomum are similar to those of Persea and Phoebe
but the perianth tube usually is deeper and grows out to form a thin cup
in which the fruit sits. Some species have the anther-locules arranged in
an arc (Neocinnamomum Liou), a situation paralleled in Phoebe vs. Persea
and Nectandra vs. Ocotea. Close relationships with Azouea Aubl. and
Phoebe and also with Ocotea have been postulated by Kostermans.

REFERENCES:
Berry, E. W. Primary venation in Cinnamomum. Torreya 4: 10, 11. 1904.

336 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

Brown, E. G. Cinnamon and cassia: sources, production and trade. I. Cinna-
mon. Colonial Plant & Anim. Prod. London 5: 257-280. 1955 [1956].*
CuHoupuHury, J. K., anp J. N. Mirra. Abnormal tricotyledonous embryo and
the morphological structure of normal fruit and ae of Cinnamomum

Camphora F. Nees ae & Cult. 19: 159-160. 1953.

DARLINGTON, R. C., AND B. V. CHRISTENSEN. (eee study of oils of cassia.
Jour. Am. Plann: yee (Sci. Ed.) 32: 118-120. 1943.* | Physical properties
of extracted oils from C. Cassia, C. zeylanicum, “C. saigonicum.” |

Dewey, L. H. The camphor tree (Cinnemomum eee Nees & Eberm.).
U.S. Dep. Agr. Div. Bot. Circ. 18

Fujita, Y. Cinnamomum Seach . and 7 allied species. Their inter-
relationships considered from the view-points of species characteristics,
chemical constituents, geographical distributions and evolution nes

Tokyo 65: 245-250. 1952. (English and Japanese text.) Be Gs
Camphora an var. linaloolifera, C. nominale and var. linalis, C. see
hs es

——— AND R. KA. Biogenesis of the essential oils in camphor trees.
XXIV_XXV. ic Japenese) Jour. Chem. Soc. Japan, Pure Chem. Sect.
77: 328-333. 1956

GRASMANN, E. Der Ka nae Mitt. Deutsch. Ges. Natur- und Volker-
kunde Ostasiens Tokio 6(56): 277-315. 1895.

Hrrota, N. An examination of the camphor tree and its leaf oil. Perfum. &
Essential ier Rec. 44: 4-10. 1953.*

Hoop, S. C. AND R. H. TRUE. pore one in the United States. U. S.
De ep. i Vearb, 1910: 449-460.

MarLIeEr-SPIRLET, M. L. Sur ae as de feuilles de Cinnamomum
L. Bull. Jard. Bot. Bruxelles 17: 266-305. 1945.

LEHMANN, C. Studien uber den Bau und die Entwicklungsgeschichte von
Olzellen. Planta 1: 343-373. 1925. [Lauraceae, 346-356: C. Camphora,
Laurus nobilis. |

Merritt, E. D. Camphorina vs. Si ele Bot. Gaz. 70: 84-85. 1920.

SANTOS, JK . Leaf and bark structure of some cinnamon trees with special
reference to the Philippine eee Philipp. Jour. Sci. 43: 305-365. 1930.
[6 spp. incl. C. seylanicum and assia

SastriI, R. L. N. Studies in the Lauraceae -— if Floral anatomy of pclae
iners Reinw. and Cassytha filiformis L. Jour. Indian Bot. Soc. 31: 246.
1952.

Witson, E. H. Camphor, Cinnamomum Camphora Nees & Ebermaier. Jour.
Arnold Arb. 1: 239-242. 1920. [An account of the plant and the camphor
industry in Formosa. |

YoTuHers, W. W., AND A. C. Mason. The camphor thrips. U.S. Dep. Agr. Bull.
1225: 1-29. 1924. [A serious insect pest in Florida. ]

Youncman, B. Professor Naonori Hirota’s work on camphor trees. Kew
Bull. 1952: jie 65. 1952. [Summary of extensive work on taxonomy, chem-
istry, and eee ak in Japan, Formosa, China.] See Mem. Ehime Univ.
Sect..2. 112) 1051

3. Nectandra Rolander ex Rottboll, Acta Lit. Univ. Hafn. 1: 279. 1778,
nom. cons.

Trees [and shrubs] with alternate coriaceous | to membranaceous |

leaves usually pinnately veined, the reticulation conspicuous [or obscure].

1958] WOOD, GENERA OF WOODY RANALES Sou

Inflorescences usually paniculate, axillary or subterminal, the narrow
bracts deciduous. Flowers bisexual, small, the 6 elliptic perianth lobes
hairy, spreading or reflexed at anthesis, deciduous, the tube conspicuous
or almost entirely lacking. Fertile stamens 9, the staminodia 3 [when
present]: stamens of 2 outer series (I & II) with nearly sessile reniform
anthers [or fleshy and petaloid], the anther-locules 4, arranged in an arc,
introrse. Stamens of series III longer, the squarish anthers on filaments
almost equal their length, with 2 large subreniform nearly sessile glands
at the base of filaments, the 4 anther-locules in two horizontal planes, the
2 upper laterally extrorse, the 2 lower extrorse. Staminodia stipiform (in
ours triquetrous, on slender pubescent filaments). Ovary glabrous; stigma
capitate, conspicuous. Fruit a thin-walled drupe, ellipsoid, globose or
oblong, with a shallow woody cupule (formed by the enlarged perianth
tube) subtended by the enlarged pedicel; cupule margin simple. (Not
Nectandra Berg., 1767.) Type species: NV. sanguinea Rol. ex Rottb. (The
name from Greek, nektar, nectar, and andros, of a man, in reference to the
stamens and anther valves which were mistaken for nectaries and stamens,
respectively.) — LANCEWoop.

A large genus (about 175 sp.) of tropical America, the majority of the
species in South America (especially the Andes), with about 35 in Central
America, and a few in the West Indies. Nectandra coriacea (Sw.) Griseb.
(Ocotea coriacea (Sw.) Britton), which occurs in the West Indies, the
Yucatan Perinsula, British Honduras, and Guatemala, reaches southern
Florida, where it extends as far north as Indian River County, on the east
coast, and Cape Romano (Collier County), on the west.

This species, which may reach 30-40 feet in height, bears small panicles
of small, white, very fragrant jasmine-scented flowers which are followed
by the first green, then dark-blue, then black fruit with green, yellow or
red cupule and enlarged fruiting pedicel. Flowering and fruiting are quite
variable. The leaf-venation is its most outstanding characteristic: there
are 6-8 pairs of lateral nerves which are rather obscured by a very con-
spicuous over-all elevated reticulum (at least in dried specimens). Nect-
andra coriacea is most closely related to N. salicifolia (HBK.) Nees (N.
sanguinea sensu Nees, non Bol, ex Rottb.), of wide distribution in Mexico
and Central America.

Nectandra is distinguished from Ocotea Aubl. (1775) by the arrange-
ment of the anther-locules of the two outer series of stamens, these being
in an arc in the former and in two planes in the latter, a distinction which
does not entirely hold. Further studies in tropical America may well show
that Nectandra should be regarded as a subgenus of Ocotea, a course re-
cently advocated by Kostermans.

It has been suggested, on the basis of the illustration of Ocotea Cates-
byana in Sargent, Silva 7: pl. 303, that both Nectandra and Ocotea occur
in southern Florida. Although this illustration does show the anther-
locule arrangement of Ocotea, rather than that of Nectandra, this would
appear to be an error, for only Nectandra coriacea (also incorrectly drawn

338 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

in Small’s Manual) seems to be represented by herbarium specimens, in-
cluding those studied by Sargent, from southern Florida. The disposition
of Ocotea Catesbyana (at least sensu Sargent) as a synonym of N. coriacea
would appear to be the correct one.

REFERENCES:

ALLEN, C. K. Nectandra coriacea, Addisonia 22: 9, 10, 12. pl. 709. 1943 [1944].

. Studies in the Lauraceae, VI. Preliminary survey of the Mexican and
Central American species. Jour. Arnold Arb. 26: 280-364, 365-434. 1945.
|Ocotea, 330-364; Nectandra, incl. N. coriacea, 365-406.] See also Fl.
Panama in Ann. Wicoui Bot. Gard. 35. 48,

Humpueey, C. J. Tests on the durability of green-heart (Nectandra Rodiaei
Schomb.). Mycologia 7: 204-209. 1915.*

Berry, E. W. The physical conditions and age indicated by the flora of the
Alum Bluff formation. U. S. Dept. Interior Geol. Surv. Prof. Paper 93-E:
41-59. 1916. [Includes ‘‘Nectandra apalachicolensis.” |

KosTERMANS, A. J. G. K. Lauraceae. Reinwardtia 4: 193-256. 1957. [Relation-
ship of Ocotea and Nectandra, 232. |

SARGENT, C. S. Ocotea. Silva N. Am. 7: 9-12. pl. 303. 1905. [N. coriacea, as
O. Catesbyana; details of stamens incorrect. |

4. Licaria Aubl. Hist. Pl. Guiane Franc. 1: 313. pl. 121. 1775.

Evergreen trees [or shrubs] with pinnately-veined leaves and small
flowers in axillary or subterminal panicles near the tips of the branchlets
[or flowers rarely solitary, subumbellate or capitate]. Flowers bisexual:
tepals 6, in two whorls, [spreading or] erect, united below into a distinct
perianth tube; (in ours, flowers obconical, ca. 2—2.5 mm. long, the perianth
tube about half this length). Fertile stamens 3: stamens of series I & IT
[small and staminodial or] abortive; stamens of series III fertile, [entirely
free, partly connate or] united into a staminal tube, their anthers 2-loculed,
extrorse [or introrse], the filaments each with 2 glands (these in ours
flattish, pressed against the staminal tube below the anthers, united in
pairs, touching each other); stamens of series IV abortive [or rarely
staminodial, minute]. Ovary included in perianth tube, free, the style
thick [to slender], stigma inconspicuous. Berry ellipsoid, (in ours black,
to 2 cm. long), smooth, the base partly covered by a thick, hemispherical
double-margined cupule. the inner margin entire, erect, the outer one
spreading, thicker, irregular, the pedicel thickened, merging into the tube.
Cotyledons flat-convex, large, including the minute 2—4-leaved, glabrous
plumule and minute conical radicle. (Acrodiclidium Nees & Mart. 1833:
incl. Misanteca Schlecht. & Cham. 1831, Chanekia Lundell, 1937 and
others.) Type species: L. guianensis Aubl. (The name derived from
licari kanali, the native name of the type species.) — SwEETWoop.

A genus of about 45 species of wide distribution in tropical South and
Central America and the West Indies; a single species, L. triandra, of the
West Indies from Martinque to Cuba, has been known from a single locality
in our area.

1958 | WOOD, GENERA OF WOODY RANALES 339

The handsome Licaria triandra (Sw.) Kostermans (Misanteca triandra
(Sw.) Mez) is one of the rarest plants in our flora, if it still persists at all.
It was first recorded from two trees discovered in 1910 in Brickell Ham-
mock, in Miami, but as many as 25 trees were counted there as recently
as 1946 by the late W. M. Buswell (see Little, Checklist of Native and
Naturalized Trees of U. S.). However, this unique hammock has been
swallowed up and destroyed by the greedy real-estate development of
metropolitan Miami and the species is presumably extinct at that locality.
Apparently a few trees (planted) still survive on the campus of the Uni-
versity of Miami.

Licaria, as used here, includes, among others, Acrodiclidium (character-
ized by the presence of series I and II as staminodes, the stamens of series
III free), Misanteca (in which staminodia are lacking and the stamens of
series III united), and Chanekia (lacking staminodia and with stamens
free). These genera merge with one another and share their “general
facies, the shape of the cupule, ovary and stigma,” characters which also
separate them from Mezilaurus and Endiandra, their nearest allies (Koster-

ans).

Although there is agreement that the group as now constituted is a
natural one, the proper name for it has been a matter of dispute. Licaria
is based upon sterile material which Kostermans identifies positively with
Acrodiclidium; Lundell, however, rejects Licaria in favor of Misanteca
because the genus is based upon a sterile specimen

Our species falls into subgenus MitsanTeEca (Schlecht. & Cham.) Koster-
mans. Apparently L. triandra | West Indies], L. limbosa (Ruiz & Pavon)
Kosterm. [Venezuela to Peru and Bolivia], L. Pittiert (Mez) C. Allen
[Costa Rica], L. Cervantesii (HBK) Kosterm. [southern Mexico, Guate-
mala], L. caudata (Lundell) Kosterm. [Br. Honduras, Guatemala], and
L. Cufodontisii Kosterm. [Costa Rica] form a group of more or less closely
related species.

REFERENCES:

ALLEN, C. K. Studies in the Lauraceae, VI. Preliminary survey of the Mexican
and Central American species. Jour. Arnold Arb. 26: 280-364, 365-434.
1945. [Licaria, incl. relatives of L. triandra, 424-432. ]

Kostermans, A. J. G. H. Revision of the Lauraceae I. Rec. Trav. Bot. Néerl.

: 719-757. 1936. [Acrodiclidium, 719-754]. II. 34: 500-609. 1937.
[| Licaria, 575-604.

Studies in South American Malpighiaceae, Lauraceae and Hernandi-
aceae, especially of Surinam. Meded. Bot. Mus. Utrecht 25: 1-70. 1936.
| Acrodiclidium, ra typification of Licaria, 34-38. |

LUNDELL, C. W. Plants of Mexico and Central ace = Wrightia 1: 145-
160. 1946. eee Licaria in favor of Misanteca. |

SARGENT, C. S. Misanteca. Trees and Shrubs 2: 133-135. pl. 155. 1911.

Tribe LITSEEAE Mez

5. Sassafras Trew, Herb. Blackwell. Cent. 3, signature P. ¢. 267. 1757;
T. F. L. Nees & Ebermaier, Handb. Med.-Pharm. Bot. 2: 418. 1831.

340 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Deciduous trees with elliptic leaves, entire, mitten-shaped or 3-lobed at
the apex, narrowed at the base, involute in the bud; buds with few im-
bricate outer scales. Plants usually dioecious, the flowers usually uni-
sexual [or apparently bisexual, but not often functionally so], in lax,
drooping, few-flowered “racemes,” the upper flowers of the lowest raceme
opening first. Perianth of 6 yellowish tepals, in two whorls of 3. Staminate
flowers with 9 fertile stamens on the margin of the short perianth tube;
anthers 4[or 2]-loculed, introrse (but the lower locules of series IIT la-
trorse), opening by 4[or 2] valves; filaments flattened, elongated, those
of series III with a pair of orange-colored short-stipitate glands at the
base; staminodes and pistillode absent [or 3 staminodes and pistillode
present in the Asiatic species]. Pistillate flowers with 6 rudimentary
stamens, in 2 whorls [or 12 in 4 whorls, similar to stamens and staminodes
in staminate flowers]; ovary ovoid, nearly sessile in the short perianth
tube, the style slender, the stigma enlarged. Fruit a dark blue ovoid berry
supported by the club-shaped enlarged and fleshy pedicel and perianth
base. Seed oblong, pointed; testa thin; embryo subglobose, erect. (In-
cluding Pseudosassafras Lec., Vushunia Kamikoti.) Type species: Laurus
Sassafras L. = S. albidum (Nutt.) C. G. Nees. (The popular name for
the plant, used as early as 1569 by the French in Florida, adopted by

rew.)

As currently delimited, a genus of three species of eastern American-
eastern Asiatic distribution: Sassafras albidum (2n = 48) (including var.
albidum and the more southern var. molle (Raf.) Fern.), of wide distribu-
tion from s. Maine to se. Iowa, s. to Texas and Florida; S. Tzwmu (Hemsl.)
Hemsl., of central China (from Kwantung and Kweichow, to Szechuan,
Hupeh, Anhwei, and Chekiang) ; ith S. randaiense (Hayata) Rehd., of
the central mountain range of Formosa.

The two Asiatic species ao the subgenus PSEUDOSASSAFRAS
(Lecomte) Keng. Both are less specialized than the American and differ
from it in the pubescent tepals, in the presence of 3 staminodes and a
pistillode in the staminate flowers, and in having in the pistillate flowers
12 staminodes similar in appearance to the stamens and staminodes of the
staminate flower, The anthers of S. Tzumu are 4-loculed, those of S.
vandaiense 2-loculed. All three plants are similar, however, in habit, bark,
winter buds, leaves, inflorescence and fruit and certainly constitute a
natural genus (although each of the three has been assigned to a separate
monotypic genus at one time or another). Rehder suggested that the near-
est relative of Sassafras is Lindera (some of the deciduous species of which
have lobed leaves very similar to those of Sassafras) which differs pri-
marily in its 2-celled anthers and in the “‘umbellate” inflorescence. Koster-
mans, however, would ally Sassafras with the evergreen Actinodaphne
Nees, a rather different group, more suggestive of Litsea

Sassafras albidum is with us a very familiar plant, long reputed to have
medicinal properties. It is at most a mild, aromatic stimulant. Gumbo
filé, a powder used to give flavor and consistency to gumbo soup, owes its

1958] WOOD, GENERA OF WOODY RANALES 341

properties to the secretory and mucilage cells of the leaves from which it
is prepared.

Although seldom planted, the species is one of our most handsome native
plants, attractive at all seasons of the year. The tree occasionally reaches
a height of 80-90 feet and a diameter of 6 feet. It is sometimes weedy, for
it tends to sucker from the roots. The attractive small, yellow flowers are
produced in early spring with the first unfolding of the leaves.

REFERENCES:

Anonymous. Dye from sassafras. Sci. News Letter 42(18): 279. 1942. [From
the root bark. |

Bastin, E. S. Structure of Sassafras. Am. Jour. Pharm. 67: 312-318. 1895.*

Berry, E. W. Notes on Sassafras. Bot. Gaz. 34: 426-450. 1902. [In relation
to fossil forms, many of which are referred to Platanus.]

Bake, S. F. Note on the proper name for the sassafras. Rhodora 20: 98, 99.
1918. [.S. officinale. |

Coy, G. V. Morphology of Sassafras in relation to phylogeny of angiosperms.
Bot. Gaz. 86: 149-171. 1928. [Staminate, pistillate flower; development of
embryo sac (Polygonum type) and e 0.

FERNALD, M. L. Nuttall’s white sassafras. Rhodora 15: 14-18. 1913.

Contr. Gray Herb. Harvard Univ.— No. CXII. IV. The nomen-

clature of Sassafras. Rhodora 38: 178, 179. 1936. [S. albidum the correct

me. |

eae B. W. Sassafras shaped history. Nat. Hist. 56: 159-164. 1947.

Hom, T. Medicinal Sa of ACs th America. 23. Sassafras officinale Nees.
Merck’s Rep. 18:

Humpureeys, E. W. ae ane non-lobed Sassafras leaves. Torreya 10:
101-108. sen

An analogy between the development of the plates of crinoids and the

leaves of ae Bull. Torrey Bot. Club 35: 571-576. 1909. [Entire
leaves at base of annual growth followed by 2- or 3-lobed leaves and again
by entire. |

Lioyp, J. U. Laurus Sassafras. Pharm. Rev. 16: 450-459. 1898.*

Kenc, H. A taxonomic revision of Sassafras (Lauraceae). Quart. Jour. Taiwan
Mus. 6: 78-85. 1953.

McMirtav, F. W., G. E. McKIBBen AND W. R. Boccess. Controlling sassafras,
persimmon, and elm with 2, 4, 5-T and eee of 2, 4, 5-T and 2, 4—-D.
Ill. Agr. Exp. Sta. Forestry Note 58: 1-5.

REHDER, A. The cues and Asiatic ee A Be Gs Jour. Arnold Arb.
1: 242-245. 1920

RoBeErRTSON, C. Flowers and insects. XVII. Bot. Gaz. 22: 154-165. 1896. [In-
cludes Sassafras, visited mostly by flies and a few small bees; list of visitors. |

RUSSELL, G. W. Large trees of Sassafras. Gard. Month. 28: 22. 1886. See also
HA.tock, N., Science 23: 51. :

SARGENT, C. S. Sassafras. Silva N. Am. 7: 13-18. pls. 304, 305. 1905.

SCHAFFNER, J. H. The jacket layer in Sassafras. Ohio Nat. 4: 192, 193. 1904.
[ Ovule. ]

SPAULDING, P. Heart rot of Sassafras Sassafras caused by Fomes ribis. Science
26: 479, 480. 1907.

Weiss, H. F. The structure and development of the bark in the sassafras. Bot.
Gaz. 41: 434-444. 1906.

342 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

6. Lindera Thunb. Nov. Gen. Pl. 3: 64. 1783; Blume, Mus. Bot. Lugd.-
Bat. 1: 323. 1851, nom. cons.

Dioecious or polygamo-dioecious shrubs with entire [or 3-lobed]| de-
ciduous [or evergreen] leaves. Flowers small, yellow, short pedicelled, in
almost sessile [in ours] umbel-like cymose clusters of 3-6, each cluster sub-
tended by 2 pairs of decussate deciduous bracts, the clusters 1-4 above
the axils of the preceding year’s leaves on ereatly reduced supra-axillary
branches terminated by a vegetative bud which grows after anthesis.
Tepals glabrous, the two whorls similar, thin, the perianth tube very
short or none; perianth deciduous, only a small disc remaining beneath
the fruit in ours. Staminate flowers with 9 stamens (series IV completely
aborted), the 3 innermost (series IIIT) each with a pair of conspicuous
stalked glands at the base; anthers 2-loculate, all introrse; pistillodium
present. Pistillate flowers with stamens variously developed, the inner-
most series usually reduced to filaments with two glands at the base; some
staminate flowers sometimes present among the pistillate; ovary and style
about equal. Fruit a bright red drupe on the short, hardly or slightly
thickened pedicel topped by the disc-like somewhat accrescent perianth

a (Benzoin Fabr. 1763, non Lindera Adans. 1763, nomina rejicienda.)
Type species: L. umbellata Thunb. (Named for John Linder, 1676-1723,
early Swedish botanist.) — W1Lp ALLSPICE, SPICEBUSH.

A large genus, of about 100 species, both deciduous and evergreen, pri-
marily of eastern Asia, with only two in the western hemisphere, both
occurring in our area, Linders Benzoin (L.) Blume, var. Benzoin is wide-
spread along streams and in damp woods from southwestern Maine to
southern Michigan and Illinois, south to North Carolina, Kentucky, Mis-
souri and southeastern Kansas. Its var. pubescens (Palmer & Steyerm.)
Rehd. is more southern in distribution, reaching Florida and Texas. Lin-
dera melissifolia (Walt.) Blume is apparently exceedingly rare and
local being known from widely scattered localities from Florida to Louisi-
ana, northward to southern Missouri and to eastern North Carolina. The
two species are quite distinct, differing in their ecology and in numerous
morphological features. (See Steyermark.) Both are known to occur close
together but in different habitats in southern Missouri.
indera Benzoin is a handsome shrub, worthy of cultivation, although,

like most native plants, it is seldom grown. The flowers are among the
very earliest to appear in spring. Staminate flowers are somewhat larger
than pistillate and frequently occur in clusters of 3 or 4 umbels, in contrast
to the less conspicuous pistillate inflorescences which are usually either
single or paired. As a result, staminate plants are far more frequent than
pistillate in the flowering condition in herbaria. This species and the
Asiatic L. praecox and L. glauca have all been reported to have 24 somatic
chromosom

Characters of flower and fruit in Lindera are very similar to those of
Litsea, the two genera being distinguished primarily by the 2-loculed (or

1958 | WOOD, GENERA OF WOODY RANALES 343

very rarely partly 4-loculed) anthers of Lindera vs. the 4-loculed anthers
of Litsea.

REFERENCES:

ALLEN, C. K. Studies in the Lauraceae, III. Some critical and new species of
Asiatic Lindera, with occasional notes on Litsea. Jour. Arnold Arb. 22:
1-31. 1941.

Fuyjira, Y. Phylogeny of Lindera membranacea Maxim., L. umbellata Thunb.
and L. citriodora Hemsl. viewed from the constituents of the essential oils.
(In Japanese; English summary.) Jour. Jap. Bot. 31: 188-190. 1956 [“The
chemical constituents show that L. membranacea is the precursor of the
other two species.”’ |

Jensen, H. W. The abnormal meiosis of Benzoin aestivale in relation to the
origin of sex chromosomes. Am. Nat. 76: 109-112. 1942. [L. Benzozn.]

Liovp, C. G. Lindera Benzoin. Drugs & Medicines N. Am. 2: 117-119. 1887.

ME LvINn L. Notes on some rare plants in North Carolina. Jour. Elisha Mitchell
Soc. 70: 312-314. 1954. [Includes L. melissifolia from Bladen Co., N. C.]

Nasu, G. V. Benzoin aestivale, Addisonia 5: 15, 16. pl. 168. 1920.

SCHROEDER, E. M. Dormancy in seeds of Benzoin aestivale L. Contr. Boyce
Thompson Inst. 7: 411-419. 1935. [Best germination following stratifica-
tion at low temperatures. |

STEYERMARK, J. A. Lindera melissaefolia. Rhodora 51: 153-162. 1949. [Dis-
covery in Missouri; clear distinctions between this and L. Benzoin; excel-
lent general notes. |

7. Litsea Lam. Encyc. Méthod. Bot. 3: 574. 1791, nom. cons.

Dioecious, [evergreen or] deciduous shrubs with pinnately veined leaves
and naked [or imbricate-scaled] buds. Flowers unisexual, in small pedun-
culate axillary sub-umbellate 3-5 flowered clusters, each with an involucre
of 4 or 5 decussate deciduous scales, globose before anthesis; in our species
borne singly above the scar of leaves of the preceding year near the tips
of the branches, or 2 or 3 on very short axillary branches. Tepals yellow,
6 [or occasionally fewer or lacking], almost completely free [or united
into an ovoid or campanulate tube], deciduous after anthesis. Staminate
flowers with 9 [or 12] fertile stamens, those of series III [and IV] with
2 stipitate glands at the base; filaments well developed, in ours 2—3 times
as long as the ovate emarginate anthers; anthers all 4-locular, all introrse;
pistillode lacking [or small]; staminodia none in our species. Pistillate
flowers with 9 [or 12] staminodia, those of series I and II usually without
glands, those of III [and IV] flanked by 2 glands at the base of the fila-
ments; ovary attenuate into the style, the stigma dilated. Fruit a more
or less globular berry seated on a minute disc [or on a shallow cupule or
disc on the enlarged pedicel]. (Malapoenna Adans. 1763, Tomex Thunb.
1783, Sebifera Lour. 1790; nomina rejicienda. Glabraria sensu Blume,
1851, not L. = Boschia Korthals [Bombacaceae]). TypE spectres: L.
chinensis Lam. = L. sebifera Pers. (according to list of nomina conser-
vanda) or = L. glutinosa (Lour.) C. B. Robinson (according to Koster-
mans). (The name presumably from a local name of southern China, litsé

344 JOURNAL OF THE ARNOLD ARBORETUM _ [voL, xxxIx

de Chine being given as the common name for the type species.) — PoNnp-
SPICE.

A large genus of perhaps 400 species, mostly evergreen, and primarily
of eastern and southeastern Asia from Japan to the Philippines, India,
New Caledonia, tropical and subtropical Australia and New Zealand, with
5 species in North America. Of the American species, 3 are distinctive
closely related plants of the eastern Sierra Madre of Mexico; the fourth,
L. glaucescens HBK. is highly variable and widespread from northwestern
Mexico east and south to Costa Rica; the fifth, L. aestivalis (L.) Fern.
(L. geniculata (Walt.) B. & H.), is a rare plant of very spotty distribution
on the coastal plain from Florida to Louisiana, north to eastern North
Carolina, southeastern Virginia (at least formerly), and Tennessee, oc-
curring around pond-margins and in swamps.

Litsea aestivalis is a shrub to 2 or 3 m. with characteristic zig-zag
branchlets and narrowly oblong leaves. The pedunculate and involucrate
umbel-like clusters of small yellow flowers are borne in early spring before
the appearance of the leaves (this being the only coe American
species). The bright-red globose fruit is borne in early sum

Like Lindera, presumably a close relative (from which differs pri-
marily in the 4- oiled anthers), the genus is complex and taxonomically
difficult in eastern and southeastern Asia, where the interrelationships of
Litsea, Lindera, Sassafras, Actinodaphne, Neolitsea and others are to be
sought.

REFERENCES:

See under Lauraceae, ALLEN (1945), Mez (1889).

ALLEN, C. K. Studies in the Lauraceae, III. Some critical and new species of
Asiatic Lindera, with occasional notes on Litsea. Jour. Arnold Arb. 22:
1-31. 1941

BARTLETT, H. H . A synopsis of the American species of Litsea. Proc. Am.
Acad. 44: 597-602. 1909.

FERNALD, M. L. Botanical specialties of the Seward Forest and adjacent areas of
southeastern Virginia. Rhodora 47: 93-142, 149- ie i 1945.
| Typification of Laurus aestivalis L., L. Pace L., 140-141.]

ME vIn, L. Notes on some rare plants in North Ca rolina, cee Elisha Mitchell
Soc : 312-314. 1954. [Includes L. aestivalis from Bladen Co., N. C.]

SARGENT, C. Rare American shrubs. Litsea geniculata. Gard. & Forest 8: 374,
375. 1895. [Includes an excellent plate; see also Sims, Bot. Mag. t. 1471.
1812

Subfam. CASSYTHOIDEAE Kosterm.
8. Cassytha L. Sp. Pl. 1: 35. 1753;-Gen, Pl. ed. 5. 22. 1754, “Cassyta.”
Parasitic green to orange twining plants with wiry chlorophyll-bearing
stems and minute, spirally arranged scale-like leaves, the plants super-
ficially resembling Cuscuta and attached to host plants by small haustoria.
Inflorescences indefinite, spicate [or racemose or reduced to heads], the

1958] WOOD, GENERA OF WOODY RANALES 345

minute flowers borne singly at irregularly separated nodes, sessile [or
pedicellate] in the axil of a minute bract, with 2 similar bracteoles close
under the perianth, bisexual. Tepals 6, the outer whorl much smaller and
resembling the bracts, united below to form a shallow tube to which the
stamens are adnate, later accrescent and inclosing the fruit. Fertile sta-
mens 9, 2-loculed, staminodia 3: stamens of series I sub-petaloid, series
I & II introrse, without glands; stamens of series III flanked by nearly
sessile glands, the anthers extrorse; series IV of distinct cordate sessile
[or stipitate] staminodes. Ovary broadly fusiform, the style indistinct,
the stigma capitate. Fruit drupaceous with a hard endocarp, completely
inclosed by, but free from the enlarged and succulent cream-colored peri-
anth tube which has a small opening at the apex surrounded by the
persistent erect perianth lobes. Seed coat membranous or coriaceous;
cotyledons thick, fleshy, often unequal, sometimes more or less consoli-
dated at maturity. Typr species: C. filiformis L. (The name from Greek,
kasytas or kadytas, dodder |Cuscuta|.) —- WOE-VINE, LOVE-VINE.

A curious genus with perhaps 15-20 species, more or less maritime and
mainly Australian but with a few in Africa and a single species, C. fili-
formis, of pantropical distribution and the only species in the Americas.
Cassytha filiformis occurs in the subtropical portions of our area, primarily
in coastal areas as far north as Brevard and Pinellas counties, where it
may easily be mistaken at first for Cuscuta. It is parasitic on a wide range
of herbaceous and woody hosts and can be a destructive pest.

Although Cassytha is an obligate parasite, the plant is at least partly
autotrophic, for chloroplasts with abundant starch are present in the cortical
parenchyma throughout the stem. In addition, numerous stomata (oriented
transversely on the stem), an extensive xylem system, and haustoria with
many well-developed spiral tracheae running to their ends which curve
directly into the wood area of the host, all suggest that the plant is pri-
marily a water-parasite. This habit, the reduction in exposed surface, the
extreme development of mucilage in the plant (reminiscent of the cacti),
and the profuse development of the plant in the brilliant sunlight of coastal
and white-sand scrubs suggest further a special adaptation to a xerophytic
type of habitat unfavorable to most Lauraceae.

Seedlings germinate in nearly pure sand and elongate quickly with
rather rapid nutating movements which bring about contact with a host

lant. The primary root remains rudimentary and only four secondary
roots (which lack root-caps) function during the period prior to attach-
ment to the host

Although the inflorescences of Cassytha are described as ‘“‘indeterminate”’
racemes or spikes or (in a few species) heads, the “inflorescences”? would
not seem to be radical departures from the usual determinate, basically
cymose inflorescences of other Lauraceae, but would merely reflect the
reduction to “simplicity” in this highly modified plant. Each of the soli-
tary axillary flowers with its two minute bracteoles would appear to be
the product (by reduction) of an entire cymose lauraceous inflorescence.

346 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

Thus the short “racemes” or “spikes” of Cassytha would represent a stem
with inflorescences (each reduced to one flower) produced at successive
nodes, just as in many Lauraceae a succession of axillary inflorescences
is produced as the growth of the season proceeds. The even more highly
modified capitate inflorescences of some species of Cassytha would be
derived through shortening of the internodes of ‘‘spicate’’ inflorescence.
Cassytha is sometimes separated from the Lauraceae on account of its
great reduction and parasitic habit but in all anatomical features (in-
cluding rubiaceous stomata, ethereal oil cells and mucilage cells) and in
all floral characters it is very clearly a member of that family. In flower
and fruit it approaches Cryptocarya R. Br., a large pantropical genus, in
which the fruit is completely inclosed in the enlarged perianth tube.

REFERENCES:

BENTHAM, G., AND F. MUELLER. Cassytha. Flora Austral. 5: 308-313. 1870.

Boewic, H. The histology and development of Cassytha filiformis, L. Contr.

ot. Lab. Univ. Pa. 2: 399-416. 1904. [Includes germination of seedling;

material from Florida. |

Hart, T. S. The Victorian species of Cassytha. Victorian Nat. 42: 79-83. 1925.

. Notes on the identification and growth of certain dodder-laurels
(Cassytha). Victorian Nat. 63: 12-16. 1946. [Includes notes on germina-
tion and early growth.

KeINHOLZ, R. An ecological and anatomical study of beach vegetation in the
Philippines Proc. Am. Philos. Soc. 65(Suppl.): 58-100. 1926. [Cassytha,

es C. R., anp L. Cuatk. Anat. Dicotyl. 2: 1152, 1153. 1950. [Sum-
mary of anatomical details of Cassytha; mostly from Keinholz. |

MirANDE, M. Recherches sur le développement et l’anatomie des Cassythacées.
Ann. Sci. Nat. IX. 2: 181-285. 1905. [Seedling and adult plant, C. fili-
formis; see also under Lauraceae. |

Sastr1, R. L. N. Embryo sac haustoria in Cassytha filiformis Linn. Cur. Sci.
Bangalore 25: 401, 402. 1956.*

Studies in the Lauraceae — I. Floral anatomy of Cinnamomum iners
Reinw. and Cassytha filiformis L. Jour. Indian Bot. Soc. 31: 240-246. 1952.

ScuMipT, A. T. Zur Anatomie von Cassytha filiformis L. Oesterr. Bot. Zeitschr.
52: 173-177. 1902.

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 347

STATISTICS OF COMPOSITAE IN RELATION TO
THE FLORA OF CHINA

SHIU-VYING Hu

IN Jury 1953 a project for the preparation of a descriptive flora of
China was initiated in the Arnold Arboretum under the auspices of the
China International Foundation. The first undertaking of the project was
the compilation of a comprehensive index to the species of Phanerogams
of that country. A staff of five persons spent two and a half years in check-
ing through the extensive botanical literature and in making reference cards
of the species of the flowering plants of China. The complete index
consists of one hundred and twenty thousand cards, each including name
and synonyms of a species, the date and place of publication and the dis-
tribution of the taxon as indicated by the citation of various collections,
with special emphasis on the type material. This index not only constitutes
a comprehensive foundation for the floristic studies and taxonomic re-
searches of the Flora of China Project, but it also reflects a much clearer
picture of the vegetation of China than any information we have had before.
In a concise and systematic way it tells both the kind of plants which can
be found in China and the locations in which each species occurs.

The Compositae constitute the largest family of flowering plants in the
world. This statement also holds true for China. The species of Com-
positae are represented in all parts of the country, from the extensive sea-
shore, in the east, to the alpine tundra of Sinkiang and Tibet, in the west,
and from tropical Hainan Island, in the south, to the Mongolian desert,
in the north. With the information given on the above-mentioned index
cards, I have made an enumeration of the Compositae of China, which
gives a total of 219 genera and 3216 species described or recorded from
this area. Due to certain nomenclatural changes some of these genera and
many of the species have been reduced to synonymy. Meanwhile, because
of the treaties made between China and Russia in 1860-64, certain mono-
typic and oligotypic genera, which were known only from the type localities
or from small areas which are no longer within the boundary of China,
have to be excluded. Consequently, only 167 genera and 2029 species are
here recognized. The basis for choosing the valid binomials has been the
recent treatments of various groups by competent taxonomists like Bab-
cock, Chang, Chen, Good, Handel-Mazzetti, Kitamura, Ling, and Stebbins.
There is no doubt that the numbers of the recognized taxa will be changed
when careful studies of available material are made, the generic limits
redefined, and the specific status better determined. Nevertheless, the
changes of details probably will not have an appreciable effect on the
general picture of the nature, distribution and origin of the vegetation,
which an analytic study of the data on the principal elements of the family
may reflect.

348 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXXIX

Up to the present there is no map which shows the phytogeography of
China. In recent years students of economic geography as well as of plant
geography have tried to prepare phytogeographic maps to illustrate vege-
tation types in China, but have failed to produce anything which can give
a true picture of the vegetation. The difficulty has been that they have
not had distributional data of the species that constitute the principal
elements of the vegetation. The present paper is an attempt to analyze
the distributional data for a large natural group and to utilize it to interpret
some of the problems involved in the composition and phylogeny of the
vegetation of China. When similar studies on the Gymnospermae, Gram-
ineae, Cyperaceae, Liliaceae, Orchidaceae, Ranunculaceae, Fagaceae,
Lauraceae, Rosaceae, Leguminosae and Ericaceae have been made, and
with the aid of a few recently monographed families, such as the Mag-
noliaceae, Theaceae, and Araliaceae, we shall be in a much better position
to present a truer phytogeographic map of China.

China is here defined so as to include all the territory covered by W. T.
Ting’s Atlas published in 1934. It has been a general practice among
Chinese botanists to include every taxon published from ‘‘Manchuria,”
“Soongoria” and the ‘““Tien-shan Range” in the flora of China. The reason
for this practice is that in a standard map of China there is a Manchuria
in the northeast (which includes Kirin, Liaoning and Heilungkiang), a
Soongaria Basin in western Sinkiang and a Tien-shan in central Sinkiang.
But when Maximowicz collected in Manchuria in 1856, and Regel made
his Soongaria expedition in 1840 these areas covered much more territory
than they do in the present map of China. Due to the 1860 treaty with
Russia, the northern and eastern half of Manchuria became a part of the
Russian Far East. Likewise, due to the 1864 treaty the western half of
Soongaria and western Tien-shan no longer belong to China. For this
reason, genera like Symphyllocarpus Maxim., from northeastern Manchuria
(in the old sense), and Plagiobasis Schrenk and Acanthocephalus Karelin
and Kirilov, from the region of Soongaria beyond the present Chinese
border, are excluded from this study.

I. THE TRIBES OF COMPOSITAE

Cassini, in 1812-18, established eleven tribes for the Compositae, Hoff-
mann in 1889 arranged 806 genera in two subfamilies and thirteen tribes
in Engler and Prantl’s Die Natirlichen Pflanzenfamilien IV. 5: 118. Dalla
Torre and Harms in 1905 listed 899 known genera in their Genera Siphono-
gamarum. They placed 877 genera in two subfamilies and thirteen tribes
following Hoffman’s system and left twenty-two genera as Genera incertae
sedis at the end. The latest record of the total number of genera of
Compositae is found in Lemée’s Dictionaire Descriptif et Synonymique
des Genres de Plantes Phanérogames 7: 484. 1939, where 1014 genera are
recognized. In this work the genera of Compositae are dispersed alpha-
betically among all the genera of the Phanerogams. It is of little use for

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 349

our purpose of making comparative studies of the principal elements of
the family. For a concise over-all picture of the Compositae and for
systematic comparisons of its tribes and genera, the record in Dalla Torre
and Harms still appears to be most useful, especially since many of the
larger herbaria are arranged according to this scheme. In modern manuals
and handbooks such as those of Fernald, Rehder, Bailey, etc., the names
of these tribes appear repeatedly. Since both the system and the names
are familiar to students of botany, they are adopted in this discussion.

When Genera Siphonogamarum was published, the flora of China was
little known to the botanical world. It therefore contains rather inadequate
information on the Chinese Compositae. The data on the number of genera
in each tribe and the distributional notes of the genera as given in this work
are here taken unchanged to give a general view of the tribes of the family.
To insure clarity the figures on the tribes of Compositae in China as re-
vealed by our studies are given separately.

1. THE TRIBES OF COMPOSITAE IN GENERAL

Dalla Torre and Harm’s information on the sizes, ranges and centers
of generic concentration of the tribes of Compositae are summarized in
the following graph. Abbreviations commonly used in the literature are
adopted for the area column with the following exceptions: Med. Reg. =
Mediterranean Region; Aust. = Australia; Pac. Is. = Pacific Islands;
and Mad. = Madagascar.

With respect to the number of genera most of the tribes are medium-
sized, including forty to sixty-five genera. The large tribes are the Astereae,
Inuleae and Heliantheae, each of which has over one hundred genera. The
small tribes are the Calenduleae and Arctotieae, the former containing 8
genera and the latter 11. The distributional patterns of the tribes are
worthy of notice. The small and medium-sized tribes all have definite areas
of generic concentration. Even in the large tribes, although they occupy
more extensive areas, the centers of the concentration of their genera are
evident.

1. The tribe VERNONIEAE has about 49 genera, 19 of which are from
South America, 16 from South Africa, 5 from tropical America, 3 pan-
tropic, 2 in Madagascar, 2 in India, and 1 each in Australia and North
America. This distributional record indicates that the Vernonieae is essen-
tially a southern tribe with the centers of concentration in South America
and South Africa. Only a few of its genera, such as Elephantopus and
Vernonia, extend to the warm region of the north temperate zone.

2. The tribe EuUpAToRIEAE includes about 51 genera, 16 of which occur
in South America, 16 in Mexico and Central America, 9 in North America,
3 in the West Indies. A few genera such as Eupatorium, Mtkania, etc., are
represented in all warm regions. Thus, this distributional record indicates
that the Eupatorieae are essentially a New World tribe with the generic
concentration in tropical America.

350 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

TRIBES OF COMPOSITAE & THEIR DISTRIBUTION

| VERNONIE AE | 49 oes
2. EUPATORIEAE Sl

3. ASTEREAE 107 Ee

5S. HELIANTHEAE 158

6. HELENIFAE 59
7___ANTHEMIDEAE Sl

| 8, SENECIONEAE $2
-CALENDULEAE | 8 ii ean ees Rs = aes

| 10. ARCTOTIEAE | 11
Jil CYNAREAE | 38} mm bs eee

12. MUTISIEAE =| 65
13. ieee 62 a ae

RIB — mates MED. PAC. | TROP

sis AREAS N. AFR REG. EUR. | ASIA |N. AM.|C.AM.|S. AM. |AUST. 1s. |ASIA MAD. |S. AFR.

Fic. 1. The tribes of Compositae and their distribution in the world.

3. The tribe ASTEREAE includes about 107 genera, 27 of which are from
North America, 19 from South America, 11 from Mexico and tropical
America, 18 from Africa, 3 from Madagascar, 2 endemic to St. Helena,
2 in the Hawaiian Islands, 3 in Europe, 5 in Australia and New Zealand,
and 4 from Asia. In addition, there are many genera that are widely dis-
tributed. For example, Brachycome is represented in Australia, Tasmania,
New Zealand, eastern and tropical Africa; Lagenophora occurs in tropical
and eastern aa. New Zealand, Australia and the Pacific Islands: Myri-
actis is found in New Guinea, oo India and Central Asia; Podocoma
is represented in tropical America and Australia; Vittadinia occurs in New
Guinea, Australia, New Zealand, New Caledonia and South America; and
both Aster and Eiger on have large numbers of species in America, Asia,
Europe and tropical Africa. This distributional record clearly indicates
that the Astereae encompass both widespread and endemic genera. It has
several centers of generic concentration, namely Asia, Africa, and North
and South America. It is especially well-represented in the New World.

4. The tribe INULEAEF is the largest tribe in the Compositae. Of its
approximately 166 genera, 84 are from Africa (especially South Africa),

om Madagascar, 37 from Australia, 12 from South America, 4 from
tropical America. In addition, 12 are common in the Old World tropics
and 4 are pantropical genera. This distributional record indicates that the
Inuleae is a southern tribe’ with the centers of generic concentration in
South Africa, Australia and, to a lesser degree, tropical Asia. It is rather
poorly represented in the northern temperate regions.

1958] HU, STATISTICS OF COMPOSITAE IN CHINA oo

5. The tribe HELIANTHEAE includes about 158 genera, 80 of which are
from Central America, 30 from the United States occurring chiefly in the
southwestern states from Texas to California, 21 from South America, 6
from the Sandwich Islands, 5 from Madagascar and 3 from Africa. In
addition, there are 6 pantropic and 4 cosmopolitan genera. This distribu-
tional record clearly indicates that the Heliantheae are a New World tribe
and, with the exception of the introduced and adventive elements, one
which is poorly represented in the Old World.

6. The tribe HELENIEAE includes 59 genera, 36 of which occur in the
United States and Mexico, 17 in tropical America and 5 in South America.
Tropical Africa has two genera, with Welwitschiella being endemic. This
distributional record indicates that the tribe Helenieae is essentially a New
World group. With the exception of the introduced taxa, it is almost
absent in the Old World

7. The tribe ANTHEMIDEAE includes 51 genera, 18 of which are from
South Africa, 10 from the Mediterranean region and the Canary Islands,
4 from Australia, 6 from central and southern Asia, 5 pantropic or cosmo-
politan, 1 from North America, 2 from South America and 1 from New
Guinea. This record indicates that the Anthemideae are essentially Old
World with the generic concentration in South Africa and, to a lesser de-
gree, in the Mediterranean region. The tribe has a few widespread, as well
as a few endemic, genera. It is very poorly represented in the New World.

8. The tribe SENECIONEAE includes 52 genera with 12 occurring in South
Africa, 10 in western North America, 4 in Central America, 3 in South
America, 2 in the Bourbon Islands, 2 in Australia, 2 in China, 1 endemic
to Juan Fernandez Island, and 1 in the European Alps. In addition, three
genera are represented in a range covering Africa, Persia and Afghanistan,
four in a range covering North Africa, Europe, temperate Asia and North
America, two in North Asia, North America, the West Indies and South
America, and one, Senecio, including 1200 species, occurring through-
out the world. At first sight these data do not seem to indicate any sig-
nificant pattern of distribution for the tribe. But when the Afghanistan-
Persia-Africa, the Europe-North Africa-temperate Asia-North America,
and the North America-West Indies-South America patterns of distribu-
tion are correlated with the continuous chains of high mountains that tie
together the continental masses, a very interesting pattern of distribution
becomes apparent. Evidently the elements in this tribe are predominantly
montane forms, Their distributions correspond with the direction of moun-
tain axes which radiate from western China westward through central Asia
to Europe and Africa, northward through northeastern Asia to the Amer-
icas and southward through the Malayan Peninsula, the Malayan Archi-
pelago to Australia. The Senecioneae, unlike most other tribes which have
obvious centers of generic concentration, are comparatively ea repre-
sented in eastern Asia, western North America and South Afr

9. The tribe CALENDULEAE is the smallest one of the ie It has
eight genera, five of which occur in the south of Africa and the rest in the

one JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

Mediterranean region and western Asia. The Calenduleae are strictly Old
World in distribution.

10. The tribe ArcTorTiraeE is the next smallest tribe in the family. It
comprises eleven genera, ten of which occur in South Africa and one from
Syria to Persia. It is strictly an Old World tribe.

11. The tribe CyNAREAE has 38 genera, 11 of which are confined to
western Asia and Persia, 5 to the Mediterranean region, 12 to a wide range
extending from southern Europe through Asia to Japan, 5 to Europe to
central Asia, 2 common in the temperate and subtropical regions of the
northern hemisphers, 1 endemic to China, 1 to India, and 1 to the Juan
Fernandez Islands. The tribe Cynareae is essentially an Old World group.

12. The tribe Muristear has 65 genera, 36 of which are from South
America, 4 from the West Indies, 3 from Central America, 8 from tropical
and South Africa, 1 endemic to the Hawaiian Islands, 1 from North Amer-
ica, 1 from the European Alps, 2 to the Himalayan region, 2 to China,
2 to Japan and 1 from Afghanistan to Japan. The Mutisieae are pre-
dominantly southern with the centers of generic concentration in South
America and, to a lesser degree, in South Africa. In the northern hemi-
sphere the tribe is represented by a few endemic genera and one wide-spread
genus.

13. The tribe CrcHoRIEAE includes 64 genera, of which 16 are confined
to North America, especially the western United States, 13 to the Medi-
terranean region, 2 to Europe, 2 to North Africa, 2 to Australia, 5 to China,
1 to South America, 1 to the Society Islands, and 1 to Juan Fernandez
Island. In addition, there are many genera with wide ranges. Sonchus and
Taraxacum are cosmopolitan. Launaea ranges from South Africa to central
Asia and temperate Europe. There are six other genera occurring from
Europe to central Asia, six from Europe to temperate Asia and North
America and four in western and central Asia. This distributional record
seems to indicate that the Cichorieae are essentially an Old World group
with an African-Eurasian-American distribution. There are nine genera,
Sonchus, Crepis, Lapsana, Hypochoeris, Mulgedium, Lactuca, Taraxacum,
Prenanthes and Hieracium, which occur on all these continents. It seems
that along this African-Eurasian-American distribution-belt variation oc-
curred particularly in the Mediterranean region, in China, and in the west-
ern United States.

2. THE TRIBES OF COMPOSITAE IN CHINA

All the 13 tribes of Compositae are represented in the flora of China by
either native or introduced species. Their size, as represented by the recog-
nized genera, their effect in the appearance of the general flora and their
prominence in the natural flora, as illustrated by the species/genus ratio,
may be summarized in the. following table.

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA Sis

TABLE I. The Tribes of Compositae in China *

LARGE SMALL
GENERA GENERA

(10 or (9 OR No. oF
No. oF No. oF SPECIES/GENUS MORE FEW NDEMIC

TRIBES GENERA SPECIES RATIO SPECIES) SPECIES) GENERA
Vernonieae 6 40 7- 1 5 0)
Eupatorieae 4 24 6 1 3 0
Astereae 22 203 9+ 3 19 9
Inuleae 19 229 12+ 6 13 2
Heliantheae 24 46 2- 0 24 1
Helenieae 3 5 2- 0) 3 0
Anthemideae 17 Za 16— 3 14 4
Senecioneae 17 358 23- 5 12 7
Calenduleae 1 2 2 0 1 0
Arctotieae 1 1 1 0 1 0
Cynareae 23 439 19+ 6 17 7
Mutisieae 6 71 12— 3 3 2
Cichorieae 24 340 15 10 14 3
TAL 167 2029 12+ (average) 38 129 34

?No distinction is made here between native and introduced species. See text.

As indicated by the number of recognized genera the Heliantheae and
the Cichorieae are the largest tribes of Compositae in China. However,
their positions in the flora of China are very different. In the Heliantheae,
with the exception of Sheareria, all the genera have been introduced through
the intentional or accidental activities of man. Naturally, these intro-
duced genera are small ones in the flora of China and their species/genus
ratio for the tribe in China is less than two. Over two-thirds of these genera
contain only one species each, and the others have two to six species. How-
ever, to a casual traveler who visits only the large cities or coastal areas of
the country the members of this tribe may appear to be the most prominent
feature in the general flora of the region. This is because the most com-
monly cultivated Compositae in the gardens of Chinese metropolises,
(species of Zinnia, Helianthus, Coreopsis, Dahlia and Cosmos) are Heli-
antheae. Likewise, the most widespread weeds of the area, common in
parks, gardens and school-yards (species of Xanthium, Siegesbeckia,
Eclipta, Bidens, Galinsoga and Wedelia) also belong to this tribe. With
the exception of these cultivated and weedy species, the other members
of this tribe have a very limited distribution, being found chiefly in the
warmer regions of the country. They occupy a very minor position as
constituents of the natural vegetation in China.

Unlike the Heliantheae, the tribe Cichorieae is large both in the number
and size of its included genera. The species/genus ratio for this tribe in
China is fifteen. As illustrated in Table II, Lactuca, Taraxacum, Crepis
and Youngia are the large genera in this tribe. Each of the first two genera

354 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

has fifty-seven species in China, and each of the other two genera has over
thirty species. Moreover, the Cichorieae include the largest number of
large genera among all the tribes of the Chinese Compositae. Many of
these large genera, such as Taraxacum, Lactuca, Youngia and Ixeris con-
tain relatively widespread species, some of which have become more or less
weedy. The region on the borders of Yunnan, Szechuan, and Sikang (the
Meridional Ranges) seems to constitute the area of the species concentra-
tion of many genera of this tribe. Maps 22 and 23 indicate that a large
number of the species in the genera Lactuca, Crepis and Youngia occur in
this region. According to Stebbins, this area is also the point of origin and
the center of distribution of three endemic genera, Dubyaea, Soroseris and
Faberia. He also maintains that Dubyaea is the most primitive genus of
the tribe. Because of its large number of genera, its rather high species/
genus ratio and its unique endemism, the Cichorieae contribute important
elements to the composition of the natural vegetation of the country.

The next largest tribes of Compositae in China are the Astereae,
Cynareae, Inuleae, Anthemideae and Senecioneae. Like Cichorieae these
tribes are comparatively large for the number of their included genera.
Senecioneae and Cynareae both have high species/genus ratios, while
Astereae, Senecioneae and Cynareae contain the largest number of endemic
genera. They are important elements in the natural vegetation of the
country. The most striking feature lies with the Senecioneae. This tribe
has the highest species/genus ratio among all the tribes of Compositae in
China, and this high ratio is due to the large number of species of only
four genera, namely, Senecio 160, Ligularia 105, Cacalia 60, and Creman-
thodium 47. The center of concentration of species of these closely related
genera is the Meridional Ranges. This region is not only the home of many
endemic species of these large genera, it is also the site of many monotypic
and oligotypic endemic genera of this and related tribes. Good’s statement
(1929, p. 313) concerning Cremanthodium in this region, “the present
point of highest species concentration happens also to be the generic point
of origin,” could be applied to many genera of the Astereae, Cynareae,
Inuleae, Anthemideae and Senecioneae, including the genus Senecio.

The tribe Mutiseae is a relatively small one in China. Nevertheless, the
tribe is fairly well represented in the natural vegetation of the country.
With the exception of Gerbera which has a South Africa-Madagascar-
tropical Asia distribution with the Chinese species marking the northern
limit of its range, all the other genera of Mutiseae in China are essentially
Chinese. Three of them (Leucomeris, Myripnois and Nouellia) are genera
endemic to China. The other two have ranges extending either from India
to Japan (Ainsliaea) or from Afghanistan to Japan (Pertya) with China
being the center of their distribution. Forty-seven out of a total of fifty
species of Aimsliaea occur in China, and ten of the twelve species of Pertya
are Chinese.

Vernonieae and Eupatorieae are poorly represented in China and those
genera which do occur are essentially widespread tropical taxa. As weedy
species they may produce quite a prominent effect on the general flora in

1958] HU, STATISTICS OF COMPOSITAE IN CHINA Oe)

the warmer regions of China. With the exception of Eupatorium their
distributions are very limited.

The Helenieae, Calenduleae and Arctotieae are represented in China
only as cultigens.

Il. THE GENERA OF COMPOSITAE IN CHINA

Two hundred and nineteen genera of Compositae have been recorded
from China. Fifty-two of them belong to the doubtful and excluded cate-
gory. The sizes and distributions of these genera within China and a
comparison with the figures for the world are summarized in the following
table. In this enumeration the genera are presented in the sequence of
Dalla Torre and Harms. The numerals in the parentheses immediately
following the names are those assigned by those authors. The names which
do not have such numerals in parentheses are either invalid epithets or
valid genera published after 1905. Most of the names in the latter group
represent genera peculiar to the flora in China. The total number of species
in each genus and the general area of distribution are chiefly adopted from
Dalla Torre and Harms. A few of them are summarized from the /ndex
Kewensis.

In presenting the distribution of the genera within China the following
abbreviations are used. A =Anhwei, Cha = Chahar, Che = Chekiang,
Chi = Chinghai, F = Fukien, H =Hainan, He = Heilungkiang, Hn =
Honan, Hp = Hopei, Hun = Hunan, Hup = Hupei, J = Jehol, Kan =
Kansu, Ki = Kiangsi, Kir =Kirin, Ks = Kwangsi, Kt = Kwangtung,
Ku = Kiangsu, Kwe = Kweichow, L = Liaoning, M = Mongolia, N =
Ninghsia, Sa = Shansi, Se = Shensi, Si = Sikang, Sin = Sinkiang, St =
Shantung, Sy = Suiyuan, Sze = Szechuan, T = Taiwan, ibe PIVeL
and Y = Yunnan.

In the remarks column the word “endemic” refers to the genus or species
described from or limited to China.

Genera known only in cultivation or as adventives are in large and
small capitals. Synonyms and other excluded names are in italics.

TABLE II. An Enumeration of the Genera of Compositae in China

TRIBES & GENERA WORLD FIGURES SPECIES IN CHINA
TOTAL GENERAL RECORDED & DISTRIBUTION &
NO. SPP. DISTRIBUTION RECOGNIZED REMAR
Vernonieae
ETHULIA (6) 3 Trop. As, Afr, Mad. T; adventive.

On
ol ed
w
he
wm
oO

Vernonia (23) 450 Am, Afr, Mad, trop. e map 1, 20 spp.
S. endemi
Camchaya 3 Indo-China, Siam 1 1 A
STOKESIA (35 N. Am. 1 1 Cultivated.
ELEPHANTOPUS (47) 16 Pantrop; Am, espec. 5 2 T, Kt, H, Sze;
adventive.
PSEUDOELEPHANTOPUS 1 Pantrop. 1 1 T; adventive.

x Oe OK OK Ok

356 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxx1x

TABLE II. (Continued)

TRIBES & GENERA WORLD FIGURES SPECIES IN CHINA
TOTAL GENERAL RECORDED & DISTRIBUTION &
NO. SPP. DISTRIBUTION RECOGNIZED
Eupato
Aecconeiind (57) 6 Pantrop. 4 2 T, Kt, H, Che, Hup,
Kwe, Y, Sze, Hp
AGERATUM (67) 30 Trop. N. Am. 3 2 T, F, Kt, H, Che,
Y, Sze; cultivated,
naturalized
= vias (88) 400 Am, Eur, As, Afr. 28 «17
Mikania (90) 150 Bante pope. Am, espec. 4 3 Kt, H, T.
* Ok Ok Ck Ok
Astereae
Solidago (121) 80 N. Am. 6 5 Kt, Hun, Kwe, Y,
T, Kir, Ray, Mong,
Sin,
Pteronia (134) 50 S. Afr, 1 0 Chinese sp. trans-
ferred to Vernonia
Dichrocephala (138) 5 Afr, trop. As. 8 3 T, F, Kt, H, Y,
Kwe, Sze, Tib.
lacie Be 2 India to s. China. 1 1 Kt, ye, VY
Grang 3 Trop. As, Afr. 3 2 T, Kt, H
Season es 12 Trop. As, to Austr. 3 1 T, Kt.
hynchospermum (147) 1 India, Malaya. 2 1 T, Y; monotypic.
Myriactis (148) 10 E. Himal. reg. 9 5 Y, Sze, T, Kwe;
ae iral Aistribation,
Betis (151) 10 Medit. reg. 1 1 Cultiva
CaLotis (157) 16 Austr. 1 1 H; nae
Asteromoea (165) 11 E. Asia. 11 10 Kt, Hun, Y, Sze,
Sa, Se, St, Hp, Ku,
F, Che, Ki, Hup.
Kalimeris 7 E. Asia. 7 0 Transferred to As-
lteromoea.
Martinia 1 E. Asia. 1 6) instead to As-
teromoea
Callistephus (170) 2 E. Asia. 2 1 Cha, Tib, Hp, Sa,
ir, He; endemic
Callistemma 2 E. Asia. 2 0) = Callistephus.
Boltonia (164) 5 N. Am. 6 ) Misidentified; Chi-
nese material
Asteromoea
Heteropappus (168) 6 E. Asia. 10 5 St, He, M, Che, T,
Kir, Kw.
Arctogeron 1 Mongolia. 1 1 M; endemic
Wardaster 1 W. China. 1 1 Si; endemic
Aster (172) 200 Am, As, Eur, S. Afr. 204 137 See map 3
Calimeri 9 0) = Aste
Diplopappus 4 0 Aster
Rhinactina 2 0 = Aste)
Turczaninovia 1 e) Aster
Asterothamnus 7 NW. China. 7 7
Pseudolinosyris 2 Centr. Asia. 2 0) Not in Chinese ter-

ritory now.

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 357

TABLE II. (Continued)

TRIBES & GENERA WORLD FIGURES SPECIES IN CHINA
TOTAL GENERAL R DED & DISTRIBUTION &
NO. SPP. DISTRIBUTION RECOGNIZE REMARKS
Galatella 25 Centr. Asia. 17 7 Si, M.
Brachyactis 13 Centr. Asia. 3 3 Hp, Se, ee anne
J, Sy, Si,
Heteroplexis 1 S. China. 1 1 Ks; re
Erigeron (173) 150 Am, As, Eur, Austr. 40 25 See map 4
Microglossa (193) 10 Trop. As, Afr, Mad 3 3 Kwe, Y, Sze
ierhapperia 1 Monotypic. 1 1 Y; endemic
Conyza (198) 50 Pantrop. 36 7 ARS ve Che,

F, ;
Kwe; indicates ex-

Thespis (199) 1 Himal. reg. 1 1
* ok k Ok Ok
Inuleae
Cava 1 W. China. 1 1 Tib, Sze; mono-
Blumea (211) 60 Trop. Afr, As, Austr. 59 30 See map 5; 10 en-
demics.
Bileveillea 3 0 = Blumea.
Leveillea 3 0) = ae
Blumeopsis 2 Indo-China 1 1
Laggera (212) 10 Afr, As, Austr 6 2 1, a Sze, Hup.
Pluchea (213) 30 Pantr 7 3 ct. Hi Sze:
Epaltes (225) 10 Pantrop. 2 2 aH
Poilania 1 Indo-China. 1 0) = oe
Sphaeranthus (227) 17 Afr, As, Austr. 8 3 Te Kits OW.
Pterocaulon (229) 12 Am, Austr, Maurit, 1 1 H.
Mad, India
Evax (238) 15 Med. reg, As, N. Am. 1 1 ;
Filago (241) 12 Eur, N. Afr, As, Am. 3 3 Tib, Sin.
Leontopodium (254) ca. Eur, As, Am, S. Am, 63 57 See map 6.
70 Japa
Anaphalis (255) 70 Eur, As, N. Am 67 51 See nee Ye
Antennaria (250) 15 Eur, As, Austr, Am. 5 Z Sin, Sz
Phagnalon (260) 20 Canary Is, Med. reg, 1 0 Not in Cae ter-
Abyss, Himal. reg. ritory
Gnaphalium (264) 120 Pantrop. 52 20 See map 8.
HELICHRYSUM (278) 300 Eur, Afr, Austr. 2 2 1 cult, 1 Sin; in-
troduced
Tugarinovi 1 N. China. 1 1 Sy; endemic.
Inula (333) 90 Eur, As, Afr. 42 28 See map 9,
vanagees 1 S. China. 1 ) = Inula.
16 Centr. As. to trop. Afr. 1 1 We
ae (350) 30 Med. reg, Eur, As, S. Afr. 4 4 Sin, Tib, Kwe.
Carpesium (353) 30 Eur, As. 30 «18 T, Kt, H, Che, Ku,

Hp, Se, St, Tub,
Kir.
Adenocaulon (354) 4 Himal. Reg, Am, Chile. 4 iD St, Kir, Sa, Cha.

358 JOURNAL OF THE ARNOLD ARBORETUM

TABLE II.

(Continued )

[| VOL. XXXIX

TRIBES & GENERA

WORLD FIGURES

SPECIES IN CHINA

TOTAL GENERAL

DISTRIBUTION

RECORDED &

DISTRIBUTION &
RECOGNIZED REMARKS

Buphthalmum (364)

Anisopappus (368)

Helianthe
Shenae. a6)

AMBROSIA (417)
Xanthium (419)

7 Eur, W. As
3 Trop. Afr.

x Ok
2 Centr. China.

15 Am, Med. reg, Afr.
15? C eamonolian,

ACANTHOSPERMUM (401) 3 Trop.
(409)

PARTHENIUM
ZINNIA (424)
SANVITALIA (427)
HeEviopsis (428)
Siegesbeckia (431)

Enuypra (435)
Eclipta (437)
RupBEckIA (449)
Blainvillea (461)
Wedelia (463)
Wollastoni
TITHONIA (467)
HELIANTHUS (471)

Spilanthes (478)

SYNEDRELLA (495)
Coreopsis (498)

Dania (499)
GLOSSOGYNE (505)

Bidens (508)

Cosmos (509)
GALINSOGA (517)

Tripax (516)
ee ea
TAGETES (582

Anthemideae
ANTHEMIS (601)

. Am.
N. & Centr. Am.
10 Pantrop.

=
“I co

9 Centr. & S. Am, Austr.
4 S. Am, Austr; 1 sp. cos-

rimelean:
30 N. Am, Mex.
9 Pantrop

60 Pantrop.
10 Centr. Am
60 N. & Ce ntr, Am.
3 Am; 2 pantrop.

Trop. Am.
70 Am, trop. Afr, Hawai-
ian Is

9 Mexico,
5 Trop. As, Austr,

90 Cosmopolitan, — espec.
Am.
20 Am.
4 Am

20 Centr. Am.

3 Am
12 Am
Am

100 Eur, Med. reg, Abyss,
As, Am

1 @ Chinese material
transferred to
Chrysanthemum.

| Kt, H

2 2 Che, Kt, a UP

CO mR WH Re NT
Se

=

—

=

e

A

Pp

ot

a
=
5
a 8
O.2
Qu
co
OQ
Oo

blesome weed

1 1 Cultivated.

1 1 :

6 5 T, Kt, Y, H, Sze.

3 0) = Wedelia

1 1 Cultivated.

6 4 Cultivated.

3 2 T, H, Y; 1 en-
demic.

1 T, H, Kt

7 5 Cultivated.

1 1 Cultivated.

2 1 - F, Kt, H; in-

roduced.

17 6 neces many
vars.

1 1 Cultivated.

1 1 , Kt; newly
introduced weed.

1 1 Ts wly — intro-
duc

1 1 Cultivated.

2 2 Cultivated.

2 2 Cultivated.

6 4 North China.

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 359

TABLE II. (Continued)

TRIBES & GENERA WORLD FIGURES SPECIES IN CHINA
TOTAL GENERAL RECORDED & DISTRIBUTION &
NO. SPP. DISTRIBUTION RECOGNIZED REMARKS
Handelia 1 Monotypic. 1 0) Not in Chinese ter-
ritory
Achillea (603) 100 N. hemisphere. 14 10 See map 10.
tarmica 6 0) = Achillea.
Matricaria (610) 50 ae reg, 5. Afr, Eur, 14 4 He, St, Hp, L.
Allardia (614) 2 ate As. 2 2 Tib, Sin; endemic.
Waldheimia 2 y 0 = Allardia
Formania 1 China. 1 1 Y; endemic.
Filifolium 1 N. China. 1 1 , to e; en-
demic.
Brachanthemum 5 Centr. As. 5 5 Sin, M, Kan; arid
regions.
Chrysanthemum (612) 200 Eur, As, Afr, eae, Is. 96 73 See map 11.
Pyrethrum (612b) 50? Conte & W. 18 5 Sin.
Ta 2 130 Centr. a W ‘me 44 0 = Chrysanthemum
Cancrinia (633) 9 Centr. 4 3 Sin, Tib, Y, Kan
Cotula (622) 50 Temp. c ‘subtrop. reg. 4 2 T, Kt, Hup, Sze
SOLIvA (623) 6 S. Am, Austr, N. Am 1 1 T; adventive.
Centipeda (624) 5 Austr. trop. As. Mad, 1 1 T, Kt, H, Che, Ku,
Chile Hun, Sze, Hp, St;
widespread weed
M yriogyne 1 1 0 = Centiped
SPHAEROMORPHAEA 1 India, Siam. 1 1 T; adventive
Crossostephium (630) 1 Luzon, E. As. 1 1 , Kt; monotypic
Stilpnolepis 1 Monotypic 1 1 Sy; endemi
Artemisia (629) 200 Cosmopolitan. 186 156 See map 12.
Pee ees ok
Senecioneae
Stereosanthus (650) 4 China. 5 4 Y, Sze, Si; en-
demic.
Nannoglottis (649) 2 China. 2 2 Kan, Y; endemic.
Tussilago (651) 1 N. Afr, Eur, As, Am. 5 1 Y, Sze, Hp, Sy, Se,
Kan, Sin.
Petasites (652) 14 N. hemisphere. 10 9 T, Ku, Hup, Sze,
St, M, Y, Se, Si.
Nardosmia 19 N. hemisphere. 3 0 = Petasites.
Erechtites (660) 15 Trop. Am, Austr, N. 2 2 T, H, Y, Ks, Kt.
Zeal,
Doronicum (671) 25 N. hemisphere. 6 3 Y, Sze, Si, Se, Kan,
N, M, Sin, Tibet.
Gynura (676) 25 Trop. vs, Austr, Afr. 26 =6:16 T, H, Sze, Si, Kt,
Y, Kwe, Hup, Che,
Se, A.
Emilia (682a) 40 Trop. Afr, Mad, trop. As. 10 3 Kt, Che, westward
to Y.
Cineraria (677) 25 S. Afr, Mad. 9 1 Largely transferred

to Senecio or Ligu-
laria

360 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

TABLE II. (Continued)

TRIBES & GENERA WORLD FIGURES SPECIES IN CHINA
TOTAL GENERAL RECORDED & DISTRIBUTION &
NO. SPP. DISTRIBUTION RECOGNIZED KS
Cacalia (680) 70 As, N. & Centr. Am. 67 60 Widespread; not in
F, Kt, H; endemic
Parasenecio 1 1 0 = Cacalia.
Syneilesis 4 China. 4 4 Kir, He, Che, A, T;
see Senecio
Chlamydites 1 Monotypic. 1 1 Tib; endemic.
Senecio (682) 1200 Cosmopolitan. 363 160 See map 14.
ARNICA (66 20 N. Am, Eur 3 1 Cultivated
Ligularia (683) 150 Eur, As. 125 105 T, Y, Sze, Si, Kan,
Kwe, Sa, , Sin,
Tib, L, Kir, Ki, St,
Hn, ~~ Che, un,
Hup, Hp, Chi, He,
Sy, Cha, J.
Farfugium 3 E. As. 4 1 T, Kt, Che, Hup.
Senecillis 70 Centr. & E. As. 45 @) = Ligularia.
Cremanthodium (684) 60 Himal. reg. 57 47 See map 15
Werneria (686) 30 S. Am, trop. Afr. 2 fe) Transferred to Cre-
manthodium,
KOK Ok Ok Ok
Calendul
eeaneae (694) 15 Med. reg. to Persia. 2 2 Cultivated.
Arcto
Peis (703) 60 S. Afr. 2 1 Cultivated, recent
introduction
Gorteria (704) 4 S. Afr. 1 0) Misidentified.
kk Ok Ok Ok
ynareae
Echinops (713) 60 E. As, S. Eur, Med. reg. 15 11 Lower Yellow Riv-
S. Afr er.
NXeranthemum (716) 6 Med. Le, W. As. 2 0) = Helichrysum or
Blumea.
Atractylis (721) 15 Canary Is, Med. reg., 9 8 North of- Yangtze,
Afr. lower Yellow River.
Atractylodes 8 0) = Atractylis.
j 7 1 ) = Atractylis
723) 6 Eur, - N. Am. 3 2 N. China to Sin.
aes (724) 250 Contr, Ss. 2 2 Tib, Sin; arid re-
gions.
Xanthopappus (726) 2 NW. China. 2 2 Chi, Si, Kan; en-
demic.
Takeikadzuchia d. N. China. 1 1 Cha; endemic.
Olgaea 11 N. China. 11 9 M, Kan, N, Sa, Sin,
ie endemic
Alfredia 11 Centr. As. 3 ¢) Not in Chitiese ter-
ritory.
Carduus (732) 1 Eur, Afr, As 19 #11 See map 16.

00 .
ONoporDON (738) 20 Afr, Eur, W. As. 1 1

1958]

TABLE II.

(Continued )

HU, STATISTICS OF COMPOSITAE IN CHINA

So

TRIBES & GENERA

WORLD FIGURES

SPECIES IN CHINA

TOTA GENERAL RECORDED & DISTRIBUTION &
NO. SPP DISTRIBUTION RECOGNIZED REMARKS
Cnicus (750) 1 Eur, As. 32 ) Transferred to Cir-
sium,
Cirsium (733) 150 N. Afr, Eur, As, N. & 70 59 Widespread.
Centr. Am.
Cephalonoplos 4 E. As 3 3 St, Sa, Se, Kir, He,
L, J, Cha, Sy, Tib;
a segregate from
Cirsium
CyNARA (734) 12 Med. reg. 2 2 Cultivated; recent
introduction.
Hemistepta 1 E. As. 1 1 , Kt, Ku, Y, Sze,
St, Hn, p, Se;
monotypic, a com-
mon wee
Saussurea (728) 125 N. hemisphere, Austr. 338 270 See map 17; larg-
est genus i in hina.
Bolocephalus 1 Monotypic. 1 1 Si; endem
Vladimiria 1 Monotypic. 1 1 We: eee
azzettia 1 1 0) = Vladimiria.
Jurinea (730) 50 Centr.& S. Eur, N. Afr, 18 18 Y, Sze, Tib, Sin.
W. & Centr. As
Tricholepis (744) 12 Himal. reg 1 1 ;
Synurus 8 emp. As. 4 Chey 2Isi,- Hy; St,
, se, Kir, H
SILYBUM (735) 2 ce Is, Med. reg. to 1 1 Introduced.
rsia
Serratula (745) 40 a to Japan 43 19 :
Leuzea 1 0 Transferred to
Rhaponticum,
Rhaponticum 1 1 Hp, Se, Kir, He,
Cha.
Centaurea (747) 500 Med. reg., Eur, As, N. 8 i Sin, Ku, Hn, to M,
& S. Am © Aastr: Tib, Kan
CARTHAMUS (748) 20 Med. ee to Centr. As. 1 1 Cultivated in W.
Chi
* Ok * Ok x
Mutisieae
eucomeris (756) 2 E. Himal. Reg. 1
Pertya (775) 12 Afghan. to Japan 12 10 ¥. S, Kan, to Che,
Myripnois (782) 2 China. 3 2 ae to Kansu, en-
demic.
Ainsliaea (783) 50 India to Japan. 58 47 Ve ae due east to
Taiw
Nouelia (791) 1 Monot 1 1 Ne. ae endemic.
Gerbera (798) 40 S. Afr, Mad, trop. As, 18 10 Kt, Hun, Hup, Y,

Tasmani

? 1 a]
widespread,
others endemic.

362

TABLE II. (Co

JOURNAL OF THE ARNOLD ARBORETUM

ntinued )

[VOL. XXXIX

TRIBES & GENERA

WORLD FIGURES

SPECIES IN CHINA

TOTA GENERAL RECORDED & DISTRIBUTION &
NO. SPP DISTRIBUTION RECOGNIZED EI c

Leibnitzia 5 E. As. 4 4 a ie He, J, Y¥,
Tib; segregate from
Gerbera.

Anandria 6 E. As. 2 0 Misidentified ;

Gerbera
ek Ok Ok Ok
Cichorieae

CicHorIuM. (823) 8 Med. reg. to temp. As. 2 2 Ku, St, Hp, Sa, Se,
Sin; cultivated or
weeds.

Lapsana (825) 9 Eur, As, Am. 3 3 Kt, Che, Ku, Ki,
Hup, Sze, Se; 1
widesptead along
Yangtze, 2 local-
ized, coastal.

Koelpinia (832) 3 ia reg., Afr, Centr. 1 1 Sink

Acanthoce phalus 2 an As. 1 0 Not in Chinese ter-
ritory now.

Rhagadiolus 8 Centr. As. 1 0 Not in Chinese ter-
ritory now.

Hypochoeris (842) 50 Eur, Med. reg, N. As, 4 3 Kir, sai Jehol, St,

N. Am ‘Hp,

Heteracea 1 Centr. As. 1 0 Not in Chin

Tragopogon (849) 40 Eur, Med. reg, W. & 10° 10 See map 20; are

centr. As. Asian element

Scorzonera (851) 100 Eur, Med. reg, W. & 20 = 18 See map 20,

Centr. As.
Picris (845) 40 Eur, Med. Reg, Abyss, 9 8 Kir, He, Y, Sze, St,
temp. As. Kt, Kan; 1
widespread, 2
demic.

Lagoseris 32 centr. & W. As, 1 0 Not in ig

Taraxacum (862) 1200? Cosmopolitan 67 57 See map

Leontodon 123 Eur, centr. & W. As. 8 0 Chinese oo =
Taraxacum

Chondrilla te 18 Eur, As. 3 3 Sin, Kan.

Sonchus (865) 45 Cosmopolita 19 10 Widespread; weedy.

Launaea (863) 30 Med. Reg, - ae , trop. 5 4 Y, Sze

As, E. Indi

Lactuca (866) 150 Cosmopolitan, 112 81 See map 22.

yeelis 5 2 0 = Lactuc
rbita 60 ae As 7 7 Y, Si, Sze

Mulgedium 7 Centr. As. 2 0) Lactuca,.

Soroseris 9 W. China. 9 9 Y, Sze, Kan, Chi,
Sin; endemic.

Prenanthes (876) 30 Eur, a reg, Canary 25 19 Y, ze, Si, Tib,

Isl, Am Hu n, .Kan,

Se, Sa, St, Hp, Cha,
Kir

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 363
TABLE II. (Continued)
TRIBES & GENERA WORLD FIGURES SPECIES IN CHINA
TOTAL GENERAL RECORDED & DISTRIBUTION &
NO. SPP. DISTRIBUTION RECOGNIZED PeMARK
Faberia (872) 6 W. China. 7 6 Y, Sze, Kwe; en-
emic.
Crepis (875) 200 N. hemisphere. 83 31 See map 23
Barkhausia 4 0) = Crepis
Geblera 2 0 = Youngia
Youngia 25 Himal. reg. to Japan. 39 30 See map 22
eris 20 Himal. reg. to Japan. 26 14 See map 23; com-
mon wee
Crepidiastrum 7 E. As. 4 3 T, Che.
Dubyaea 12 E. Himal. reg. 12 7 Y, Sze
Hololeion 3 Japan. 1 1 Ku.
Hieracium (877) 400 Eur, Am, Afr, As. 17. 14 Ku, i, T, Sze,
Kwe, Sin, Hp, M,
e, St, Kir, He,
5) mostly
isolated am known
from one collection.
Crepidiastrixeris 3 Japan, E. China. 1 1 Che; a hybrid genus.

With the exclusion of the doubtful and invalid genera there are only
167 genera of Compositae in China. Most of these genera are small. Fifty-
seven of them have only one species each, and 40 others have 2 or 3 species
each. The largest of all is Saussurea with 279 species.
are Senecio with 160 species, Artemisia with 156, Aster with 137 and

Ligularia with 105 species.

Some

The next in size

botanists, such as Franchet, interpret

Senecio in a broader sense and place Ligularia, Cacalia and Cremanthodium

in it as sections.

In this broader sense, Senecio would have 372 species in

China and thus become the largest genus of Compositae in that country,
as it is also the largest genus of the Compositae in the world.

1. Tuer Larce GENERA AND THEIR DISTRIBUTIONS

nly 38 genera of the Chinese Compositae contain ten or more species.

For convenience of discussion, these are called the “large genera.”
are scatiered in the Vernonieae,
ideae, Senecioneae, Cynareae, Mutisieae Gl

They

Eupatorieae, Astereae, Inuleae, Anthem-
Givers (TABLE I).

By

plotting the occurrence of all the species of a large genus on an outline

map

of China, striking distributional patterns of these large genera are

revealed. In these maps circles represent species occurring only in one
province and the dots denote species recorded from two or more provinces.
The combined number of circles and dots within the confines of a province
gives the total number of species of the genus under discussion in that
province. Thus a distribution map for a given genus tells both its range

and its area of concentration of species.

In a few cases two relatively

364 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xXxxIx

small genera are plotted on one map. Triangles are used to represent the
species of the second genus, with the white ones indicating local endemics
and the solid ones species of wider range.

Vernonia is a genus of wide range with species in the Americas, Africa,
Madagascar, and tropical Asia. In China there are thirty-four species,
with a concentration in Taiwan, Kwangtung, Yunnan and Szechuan. Ten
of these thirty-four have an Indo-Malaysian range, two of them extend to
Java, and twenty of them are endemic to China. Six of these endemics
occur in Yunnan (Map 1). In general Vernonia is a southern genus and its
range does not reach north of the Yangtze River.

Eupatorium is another widespread genus with many species, especially
in Central and South America, Africa, Europe and Asia. In China there
are seventeen species, and most of them are evenly distributed from Hainan
to Heilungkiang (Map 2). Endemism is relatively low. Taiwan seems to
have the largest species concentration. It is interesting to note that this
genus is absent in the northern and western half of the country.

Aster is the third largest genus of Compositae in China. It contains
137 species occurring in every province of the country. The centers of
species concentration are Yunnan (51 species), Szechuan (48 species),
Sikang (30 species), Kansu (20 species) and Hopei (20 species). Taiwan
has 15 species. Considering the small size of the island, the genus is very
well represented there. This genus has many widespread species in China.
For example, A. ageratoides and its varieties occur in twenty-three prov-
inces and A. altaiacus and its varieties occur in nineteen provinces. Many
species share the Yunnan-Szechuan-Kansu range. Local endemism is high
for the genus. Of the endemic species, Yunnan, with its 15 species, has the
highest number, Szechuan has 13, Taiwan 10, Tibet 9, Kansu 6, Sikang 5.
It is interesting to note that the species in Sinkiang are all widespread,
while nearly half of the 21 species in its neighboring province, Tibet, are
local endemics (Map 3)

Erigeron is a widespread genus with species occurring in America,
Australia, Asia and Europe. In China there are twenty-five which are
fairly evenly distributed throughout the country. They are absent from
Ninghsia, Kansu, Chinghai and Tibet (Map 4). There are several weedy
species which occur in extensive areas. For example, E£. acer occurs in ten
provinces from Hupei and Szechuan due north to Kirin and westward to
Sinkiang. EF, canadensis has an even wider range, occurring in fourteen
provinces from Taiwan-Kwangtung northward to Kirin and thence due
west to Sinkiang. This genus has very few endemic species. Half of the
eight endemics are in the Altai region.

Blumea is a genus with an African-Asiatic-Australian range. In China
there are thirty species which concentrate in Taiwan, Kwangtung, Hainan,

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 365

ark
gee wera
jp

4 12
ea &4,
gaits ae

.
ax:

; @ ‘
Juancry

Ms

es
wancnet@

a ae
ANAPHALIS 7 Q GNAPHALIUM §

Maps 1-8. The geographical ranges in China of eight eee Sas of Com-
positae in the tribes Vernonieae, Eupatorieae, Astereae, and Inulea

366 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

Yunnan and Kweichow (Map 5). Most of the species have an Indo-
Malaysian distribution. In China they occur largely in the few border
provinces where there are port cities. The large number of endemics in

.weichow (50%) is evidently due to the careless work of Léveillé and
Vaniot who published too many species from fragmentary collections.
Blumea serves as a good example of the route of migration and the area
of extension of tropical elements in the flora of China

Leontopodium is a discontinuous genus occurring in the high moun-
tains or high latitudes of Europe, Asia and South America. In China there
are 57 species which are concentrated in Yunnan, Szechuan, Sikane and
Tibet, and thence northeastward through Kansu, Shansi, Chahar_ to
Heilungkiang and Mongolia. In the mountains of the Meridional Ranges
there are many endemics and hybrids (Map 6). It is very likely that this
region is both the center of concentration and the place of origin of the
genus. For example, L. kamtschaticum, as is indicated by many recent
collections, is concentrated in Sikang: thence it extends westward to Tibet,
and northeastward to Szechuan, Kansu, Chahar, Mongolia, ialinekiane.
Far Eastern Russia and Kamchatka. It is evident that although the species
was first described from Kamchatka, this peninsula is only on the periphery
of its range. Leontopodium japonicum tells almost the same story. It is
very likely that the species originated in the west, somewhere in the moun-
tains on the Kansu-Shensi-Szechuan border, thence it extended eastward
through Hupei and Anhwei to Japan or through Hopei and Korea to Japan.
These are common routes for the migration of many Sino-Japanese species.

Anaphalis is another genus which has a discontinuous range and which
has its concentration of species in the Meridional Ranges of China. It has
been recorded from Europe, Asia and North America, but the bulk of the
species are in China. There are 51 species, many of iem local endemics,
concentrated in Yunnan, Szechuan, Sikang and Kansu (Map 7)

Gnaphalium is a genus of the warm regions throughout the world. In
China there are twenty species, rather evenly distributed from Hainan to
Heilungkiang. Although there are a few endemics in Yunnan. Sikang and
Tibet, there is no region which can be considered as the center of concen-
tration of species for this genus. There are a few widely spread species.
For example, G. affine extends from Taiwan to Nepal, occurring in fifteen
provinces in China. It is a very tough species and colonizes all sorts of
waste places, even the perpendicular cracks of dry hard city walls. G.
Aypoleucum is another widespread species which extends from Taiwan to
Nepal. It occurs in eight provinces, It is interesting to note that in China
the species occurs only in areas where there are large centers of commer-
cial or political activity (Map 8

Inula is an Old World genus with species occurring in Africa, Europe
and Asia. In China there are twenty-eight species which are evenly dis-

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 367

tributed in the temperate and the mid-high altitudes of the subtropical
regions (Map 9). It seems that Yunnan, Szechuan and Kweichow con-
stitute the center of the species concentration. In Yunnan and Kweichow
almost half of the species are endemics. There are a few widespread species.
For example, J. britannica and its varieties occur in fifteen provinces north
of the Yangtze River and /. cappa occurs in six provinces south of the same
river. Several species of this genus are very good indicators of the types of
vegetation in China. In addition to J. britannica and I. cappa, which have
a northern or a southern distribution, 7. salsoloides expresses a special
floristic relationship between Hopei, Chahar, Suiyuan, Ninghsia, Shans,
Shensi, Kansu, Chinghai and Sinkiang, and J. serrata illustrates the floristic
affinity between Sikang, Yunnan and Kweichow.

Carpesium is a genus of the north temperate or subtropical regions of
the Old World. Its range extends from Europe through Asia to Japan. It
has eighteen species in China. It seems to have two centers of concentra-
tion, the one being Yunnan, Szechuan, Sikang, Hupei, Shensi, Shansi and
Hopei and the other being Taiwan and Kwangtung. There are several wide-
spread species. Carpesium abrotanoides occurs in twelve provinces, from
Taiwan-Kwangtung westward to Yunnan and Sikang and thence due north
to Shensi and Hopei. Carpesium cernuum shares the same range, but ex-
tends even more northward to Kirin. There are relatively fewer local
endemic species in this genus.

Achillea is a widely distributed genus of the northern hemisphere.
There are 10 species in China. With the exception of a southern variety
of A. sibirica, which occurs in Kweichow and Yunnan, and an Eurasian-
American species, which was recorded from Chekiang, the genus is re-
stricted to the north of the Yangtze River (Map 10). Its distribution is
a good illustration of the southern limit of the northern elements in the
flora of China.

Chrysanthemum is a large genus with species occurring in Europe,
Canary Islands, Africa, temperate Asia and America. In China there are
seventy-three species, many of which are local endemics known only from
the type collection. Most of the endemic species are in Sikang, Yunnan,
Szechuan, Kansu, Shensi, Shansi, Mongolia and Sinkiang (Map 11). The
taxonomy of this group is in bad shape, and it is highly possible that many
of the endemics described as species in this genus are merely local variants
of a few species. It is interesting to note that the genus is poorly repre-
sented in low altitudes of South China. Chrysanthemum indicum is a wide-
spread species recorded from thirteen provinces.

Artemisia is a cosmopolitan group which in China is the second largest
genus of Compositae. Its 156 species represent every province of the
country. As far as the number of species in each province is concerned,
Hopei takes the lead with 59 species, 3 of which are local endemics. Yun-

368 JOURNAL OF THE ARNOLD ARBORETUM _[vot. xxx1x

12

4 e
Joanc™

es

e
vs'@
ef

- 5 eso

ih
08

ECKINOPS
3 carDuus [6

Maps 9-16. The geographical ranges in China of eight large esoe of Com-
positae in the tribes Inuleae, Anthemideae, Senecioneae, and Cyna

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 369

nan has 50 species, 11 of which occur in that province alone. Szechuan
has 41 species, 5 of which are local endemics. Taiwan has 25 species, 7 of
which are endemics. Considering the small size of the island, the genus is
very well represented there. In fact, Artemisia is the only genus of Com-
positae that has been recorded from every province (Map 12). This genus
needs revision; many local variants, apomictic or polyploid forms have
been named as species and, consequently, many taxa show anomalous
patterns of distribution. For example, A. dubia var. septentrionalis has
been recorded from Hainan, Kweichow and Hopei, and A. handel-mazzetti
has been recorded from Yunnan and Hopei only. Such disjunction is not
known in any other species of flowering plants in China.

Gynura has an African-Asiatic-Australian distribution. There are 16
species in China. With the exception of G. ovalis var. pinnatifida, which
extends into the southern part of Shensi, all the rest are distributed to the
south of the 30°N. parallel. Thus the species of Gynura serve as good
examples for showing the northern limit of the range of southern elements
in the flora of China.

Cacalia is an Asiatic-American genus. Its species occur in Asia, North
America, Central America and the West Indies. In China there are 60
species which are distributed in high altitudes and mid-latitudes. Yunnan,
Szechuan, Sikang, Kansu, Shensi, Shansi, Hupei, Hopei and Honan seem
to be the center for the concentration of the species.. It has a high per-
centage of endemism (Map 13). Forty per cent of the species in Yunnan
are known only from the type material. All the species in Taiwan are
endemics. Cacalia is morphologically closely related to Senecio and Ligu-
laria. The distributional patterns of these three genera are also similar.
It is worthy of note that this is the only genus that has 5 species in Honan,
a province in which other genera of Compositae are relatively poorly
represented.

Senecio is the largest genus of Compositae, and a very heterogenous
one. Its species occur in all parts of the world. In China, because of the
recognition of Cacalia, Ligularia and Cremanthodium, all of which are
included in this genus by some authors, Senecio becomes the second largest
genus. However, if Franchet or James Small’s interpretation of the genus
were adopted, Senecio would be the largest genus in China. The concen-
tration of species of this genus is in Yunnan, Szechuan, Sikang, Kweichow
and Hupei (Map 14). In Yunnan alone there are 73 species, 41 of which
are known only from that province. Szechuan has 51 species, 12 of which
are local endemics. Sikang has 13 species, 6 of which are local endemics.
Kweichow has 23 species, 7 of which are not known elsewhere.

Ligularia is a genus with European-Asiatic distribution. In China
there are 105 species highly concentrated in Yunnan, Szechuan, Kansu
and Hopei. In Yunnan alone there are 50 species, 27 of which are known

370 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

only in that province. Szechuan has 49 species, 14 of which are local
endemics.

Cremanthodium is an endemic genus of the Meridional Ranges. Its
distribution extends to the Himalayan Region. In China there are 47
species concentrated in Yunnan, Sikang, Szechuan, Tibet and Kansu
(Map 15). There are 38 species in Yunnan, about one-fourth of which are
endemic to that province. Within China this genus is strictly limited to
the Southwest. Its closely related genera Cacalia, Senecio and Ligularia
are all well represented in Taiwan, but the range of Cremanthodium ex-
tends hardly beyond the Long. 110° E.

Echinops is an Old World genus with species occurring in southern
Europe, the Mediterranean region, tropical Africa and eastern Asia. There
are 11 species in China distributed north of the 35° N. parallel (Map 16).
Judging from the specimens in the Gray Herbarium, the center of con-
centration of the species appears to be western and central Asia. China
is only on the periphery of its range. Several species have a considerably
wider range. For example, E. latifolius extends from Dahuria to Honan,
and £. gmelinit covers almost the same area. There are four endemic
species known only from their type collections.

Carduus is another European-African-Asiatic genus. There are eleven
species in China, two of them with wide ranges. Carduus acanthoides
occurs in six provinces from Yunnan-Kweichow northward to Kansu and
Hopei. Carduus crispus has an even wider range, occurring in 16 provinces
from Chekiang westward to Szechuan and northward to Heilungkiang and
Mongolia. The other species are local endemics (Map 16)

Cirsium is a widespread genus with species occurring in North Africa,
Asia and North and Central America. There are 59 species in China. They
are distributed throughout the country. Yunnan, Szechuan, Kweichow
and Taiwan are areas of high endemism. Yunnan has 22 species, 13 of
which are confined to that province. Taiwan has 12 species, 7 of which
are endemics. There are several widespread species. For example, C.
arvense (first recorded from the Canary Islands), and its various varieties
occur in 11 provinces, from Kiangsu to Heilungkiang and thence due west
to Sinkiang. Cirsium chinense is another widespread species. Its range
extends from Taiwan to Yunnan and northward to Shensi and Hopei. In
most places it is a very troublesome weed.

Saussurea is a genus widely distributed throughout the northern hemi-
sphere and the mountains of Australia. In China it constitutes the largest
genus of the Compositae. There are 279 species, especially well repre-
sented in Yunnan, Szechuan, Sikang, Tibet, Hupei, Kansu, Shensi, Shansi,
Hopei, Jehol, Kirin, Mongolia and Sinkiang (Map 17). Yunnan alone
has 94 species, 53 (almost 57%) of which are confined to that province.

1958] HU, STATISTICS OF COMPOSITAE IN CHINA Sl

There are 82 species in Szechuan. Twenty-six (almost 32 per cent) of them
are local endemics. Twenty-two (a little over 43 per cent) of the 51 species
from Sikang are endemic to that province. It is interesting to note that
the genus is poorly represented in eastern and southern China, and it has
never been recorded from Kwangsi, Hainan and Honan. It is evident that
the species of this genus prefer high altitudes or high latitudes.

Jurinea is an Old World genus with species occurring in central and
southern Europe, North Africa, western and central Asia. In China there
are 18 species, all of which are local endemics (Map 18). They are known
only from a few western provinces, and often only through the type collec-
tion.

Serratula is another Old World genus occurring in Europe and North
Africa, thence due east to Japan. In China there are nineteen species
distributed chiefly north of the Yangtze River. Some species also occur
in Kirin, Heilungkiang, Mongolia and Sinkiang (Map 18).

Pertya is an Asiatic genus occurring from Afghanistan to Japan. There
are ten species in China, all with very limited range (Map 19). Almost
half of them are known only through the type material.

Ainsliaea is another Asiatic genus. Yunnan, Szechuan and Hupei con-
stitute its center of distribution, and northern India and Japan are on the
periphery of its range. There are 47 species in China (Map 19). Many
of them are local endemics. For example, there are 22 species in Yunnan,
14 of which are endemic to that province. There are several species which
indicate the relationship between the flora of Taiwan and the mainland of
China. For example A. fragrans occurs in Taiwan, Kwangtung, Chekiang,
Kiangsu, Kiangsi and Hopei. Ainsliaea macroclinidioides has the same
range. In both cases the type localities are on the periphery of the range
of the species. Ainsliaea reflexa and its varieties occur in Taiwan and
also in Yunnan. This distributional pattern is common with many genera
of woody plants.

Gerbera is a southern genus with species occurring in South Africa,
Madagascar, tropical Asia and Tasmania. There are 10 species in China,
and Yunnan and Szechuan again constitute the center of the species con-
centration. Gerbera anandria is a widespread species. It occurs in 15
provinces from Kwangtung northward to Kirin and Mongolia. A little over
45 per cent of the eleven species and varieties in Yunnan are endemic.

Tragopogon is an Old World genus with species occurring in Europe,
the Mediterranean region, and western and central Asia. There are 10
species recorded from China. With the exception of three European species
(one recorded from Nanking as a cultigen, and two from gardens in Peking)
all the rest are localized in Sinkiang, especially the Tien-shan-Altai region

ie JOURNAL OF THE ARNOLD ARBORETUM _ [volL. xxx1x

m A JURINEA

an Smee
saussurea IT = ore @ serratuLca 18

a P ——$—<—_1—
‘ ?
& PERTYA Ln H AX & TRACOPOGON
@ Ainstiaza 19 . ® scorzonera 20

@ Lactuca
a youncia 22

@ CREPIS
@ IXERIS

ALL SMALL GENERA 24

Maps 17-23. The geographical ranges of seven large genera of Compositae in
the tribes Cynareae, teers e and Cichorieae. Map 24. The distributions of
small genera of Compositae in China. Dots = endemics, circles = adventives,
and H = ia seen

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 373

(Map 20). Tragopogon is a good example of the extent of Central Asiatic
elements in the flora of China.

Scorzonera is another Old World genus with species occurring in
Europe, the Mediterranean region, western and central Asia, and thence
due east to China, Korea and Japan. In China, there are eighteen species
distributed in the arid regions of mid-high latitudes and the arid regions
of Szechuan and Tibet (Map 20). The percentage of local endemics is
low. Several species have wide ranges. For example, S. albicaulis occurs
in twelve provinces from Chekiang westward through Hunan and Kwei-
chow to Szechuan and thence due north through Honan, Shantung, Shensi
and Hopei to Chahar, Liaoning and Kirin. Scorzonera austriaca occurs in
eight provinces extending from Kiangsu northward through Honan, Shan-
tung to Kansu and Mongolia. It is worthy of notice that in the distribu-
tion of Senecio, Ligularia, Saussurea and many other genera of Compositae,
Yunnan and Szechuan are twin provinces in respect to high numbers of
species, but this is not so with Scorzonera. Four species of Scorzonera have
been recorded from Szechuan and none from Yunnan.

Taraxacum is a cosmopolitan genus of weedy species. The taxonomy
of this genus is very difficult. When Handel-Mazzetti published his mono-
graph of Taraxacum in 1907, he included 57 species. Index Kewensis lists
in all over 1200 species, but binomials have been assigned to many apo-
mictic forms. It is very hard to decide what is the approximate number
of species of this genus in China. When Dahlstedt published H. Smith’s
collection in 1926 he added one-fourth more binomials to the Chinese
Taraxacum. With the species added by Kitagawa in 1933-38, and Koro-
leva in 1940, 57 have been recognized from China. It seems that a large
number of species are found in Yunnan, Szechuan, Tibet, Kansu, Sinkiang,
Mongolia, Chahar, Liaoning and Kirin (Map 21). It is interesting to note
that this genus is poorly represented in the warmer regions of China. It
has not been recorded from Fukien, Kwangtung, Kwangsi or Hainan.

Sonchus is another cosmopolitan and weedy genus. There are 10

ward to Sinkiang, Mongolia and Kirin and S. oleraceus occurs in fourteen
provinces, from Hainan northward to Kirin and westward to Sinkiang.
Four local endemics have been recorded from Yunnan, Szechuan, Kwei-
chow and Tibet. There seems to be no center of species concentration in
China.

Lactuca is a cosmopolitan genus. There are 57 species in China dis-
tributed from Kwangtung and Hainan northward to Kirin and Heilung-
kiang. It seems that Yunnan, Szechuan and Kweichow form a center of
concentration of species on the mainland, and Taiwan furnishes an area
of diversification off the coast. A high percentage of endemism occurs

374 JOURNAL OF THE ARNOLD ARBORETUM | [VvoL. xxxrx

among species in both regions. In Yunnan there are 27 species, 17 (about
63 per cent) of which are endemics. In Taiwan there are 14 species, 11
(79 per cent) of which are endemics. There are a few widespread species:
I. indicus and its varieties occur in 15 provinces, from Kwangtung north-
ward to Kirin and L. tatarica occurs in 8 provinces, from Honan to Mon-
golia and westward to Sinkiang (Map 22)

Prenanthes is a widespread genus with species occurring in South
Africa, the Canary Islands, the Mediterranean Region, Europe, Asia, and
America. The strongest development of this genus is in central and northern
Europe. There are nineteen species in China. Yunnan, Szechuan and
Kweichow again form the center of concentration of species. High ratios
of endemism occur in Szechuan and Kweichow, where over 60 per cent of
the species are known only from the type localities. There are a few wide-
spread species. Prenanthes brunoniana and its varieties occur from north-
ern India eastward through Yunnan, Kweichow to Hupei and Hainan.
Prenanthes tatarinowii and its varieties occur in eight provinces from
Hupei northward to Chahar and Kirin.

Crepis is a widespread genus with species occurring in the Canary and
Madeira Islands, Europe, Africa, Asia, and North America. In China
there are 31 species, concentrated in Yunnan, Szechuan, Kweichow, Sikang,
Tibet, Sinkiang, and Mongolia (Map 22). Yunnan seems to be the area
of the highest species-diversification. It has not only the largest number
of species but also the highest ratio of endemism (about 30 per cent).
Most species have small ranges, usually limited to two or three provinces.
For example, C. rigescens is limited to Yunnan and Szechuan, C. tibetica
to Yunnan, Sikang and Tibet, C. bodinieri is confined to northern Yunnan
and the adjacent area of Szechuan, and C. chrysantha to the Altai Region.
As suggested by Babcock, Crepis is originally an Asiatic genus and the
Altai region (northwestern Sinkiang and southwestern Mongolia) seems
to be its center of origin. This region is still a part of the center of the
concentration of species for the genus today. It is also worthy of note
that Crepis is not represented in Taiwan, Fukien, ——— Hainan,
Chekiang, Anhwei, Kiangsi, Shantung, Honan or Ki

Youngia is an Asiatic genus with species occurring from the Himalayan
Region eastward to Japan. In China there are thirty species distributed
from Hainan, Kwangtung and Taiwan, thence northward to Heilung-
kiang (Map 22). Like Crepis, it has its center of species concentration in
Yunnan and Szechuan, but it differs from Crepis in that it occurs also in
the tropical regions, Again it differs from Crepis in that it has many wide-
spread species. For example, Y. japonica and its varieties occur in thirteen
provinces from Hainan, Kwangtung and Taiwan northward to Shantung,
Hopei and Shensi. Voungia sonchifolia occurs in thirteen provinces from
Chekiang to Szechuan and thence due north to Kirin and Heilungkiang.

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 375

Ixeris is another Asiatic genus with species distributed from the eastern
Himalayan Region to Japan. In China there are fourteen species ranging
from Hainan and Kwangtung northward to Heilungkiang (Map 23). Most
of them are widespread weeds. For example, J. chinensis occurs in twenty-
one provinces. Actually it is very difficult to determine the approximate
number of species of /xeris in China, since contemporary authors do not
agree on the status of some of the taxa. The criteria for distinguishing
Ixerts, Youngia, Crepis and Lactuca are not sufficiently strong and many
species have been changed back and forth among these genera. For ex-
ample, J. chinensis has been named Prenanthes chinensis, Youngia chinensis
and Lactuca chinensis. Likewise, Ixeris chinensis ssp. graminifolia has been
named /xeris graminifolia, Crepis graminifolia and Lactuca chinensis {.
graminea by outstanding synantherologists of our time.

Hieracium is a widespread genus with species in Europe, North and
South America, North and South Africa, and northern and eastern Asia.
There are fourteen species in China, distributed from Kiangsu, Kiangsi,
Kweichow and Szechuan northward to Kirin, Heilungkiang and Sinkiang.
Most of them have small ranges. Hieracium umbellatum is the only wide-
spread species. It occurs in twelve provinces from Kiangsi-Hupei-Szechuan
northward to Kirin, Heilungkiang, Mongolia and Sinkiang. According to
Stebbins most of the Asiatic species are apomictic ( Babcock, 1947, p. 83).

In conclusion, we may point out that the larger genera of Chinese Com-
positae evidently have four types of distribution. The first type, which is
the most frequent, includes widespread genera with definite centers of
species concentration. Twenty-two of the thirty-eight large genera (almost
57 per cent) have this type of distribution. Aster(T), Leontopodium,
Anaphalis, Artemisia(T), Inula, Chrysanthemum(T), C acalia(T), Senecio
(T), Ligularia(T), Cremanthodium, Cirsium (T), Saussurea(T), Jurinea,
Gerbera, Ainsliaea(T), Prenanthes, Taraxacum, Lactuca(T), Crepis,
Youngia, Pertya and Ixeris(T) all belong here. The center of species
concentration of all these genera is the Meridional Ranges. The genera
marked (T) also have secondary centers of concentration in Taiwan. The
second type comprises the northern genera with ranges limited to the
north of the Yangtze River. Seven of the thirty-eight large genera (about
20 per cent) have this type of distribution. Achillea, Echinops, Trago-
pogon, Scorzonera, Carduus, and possibly Hieracium and Serratula belong
here. Most of these genera also occur in Europe, the Mediterranean
Region and central or western Asia. The third type includes the southern
genera, the ranges of which are limited to the south of the Yangtze River.
Of the thirty-eight large genera only Vernonia, Eupatorium, Blumea, and
Gynura have this type of distribution. The fourth type includes those
widespread genera which have no definite centers of species concentration.
Erigeron, Gnaphalium, Carpesium and Sonchus exemplify this type of
distribution and all include some widespread weedy species.

376 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. xxxix

2. THe SMALL GENERA AND THEIR DISTRIBUTIONS

The small genera are taxa comprising one to nine species. This arbitrary
classification is made merely for the convenience of discussion. There are
128 such small genera of Compositae in China. Their distribution among
the tribes is given in Tastes I and II. Most of these genera include one
to three species, but a few have four or more species. The sizes of these
small genera as indicated by the number of the included species are shown
in the following graph (Fic. 2).

60

NUMBER OF GENERA
ro 8 > a
o 8&6 8 °O

oO

i 2 3 4 5 6 7 #8 9
NUMBER OF SPECIES

Fic. 2. The size and number of small genera of Compositae in China.

Map 24 illustrates the general pattern of distribution and the areas of
concentration of the small genera of Compositae in China. An analysis
of the distributional record of these genera reveals that they may be grouped
into three types. These are the genera with species in China known only
in cultivation, the genera containing only isolated endemics, and the genera
with native species in China and also elsewhere in other floristic regions.

Endemics. Thirty-four of the 128 small genera are endemic to China.
Their occurrence in various provinces is as follows: Yunnan 15, Szechuan
8, Kansu 6, Sikang and Hopei each 4, Heilungkiang, Kirin, Kweichow,
Taiwan and Tibet each 3, Chekiang, Chinghai, Kwangtung, Mongolia,
Shansi and Suiyuan each 2, and Hupei, Kiangsi, Ninghsia, Kwangsi and
Shensi each 1.

A few of these endemic genera were first described from the Himalayan

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA eid)

Region. Recent collections extend their range to Yunnan, Szechuan, Kwei-
chow, Kwangtung and Taiwan. It is evident that the high mountains
bordering Yunnan, Szechuan, Sikang and Kansu (the Meridional Ranges)
constitute a center of aggregation for the small genera.

Genera known in China only as cultivated plants. Twenty-one
of the 128 small genera of Compositae are known in China only in cultiva-
tion. The commonest species belong to the genera Zinnia, Helianthus,
Coreopsis, Cosmos, Ageratum, Gaillardia and Calendula. In the warmer
part of the country, the escaped Ageratum conyzoides is naturalized and
appears weedy in gardens, fields or along the roadside.

Genera with native species in China as well as in other floristic
regions. Seventy-three of the 128 small genera of the Chinese Compositae
occur also in central Asia, tropical Asia, Africa, the Pacific Islands, Aus-
tralia or America. There is no record of their introduction from these
regions to China or vice-versa. They were probably dispersed accidentally
through man’s activities. Five of them have widespread species which are
usually considered as weeds. Eclipta has only one species in China, and
this species has been recorded from thirteen provinces. It is a common
weed in cotton or soybean fields and its occurrence in China can be traced
back to the ancient historical period. Likewise, Xanthium, as represented
by X. strumarium, occurs in seventeen provinces in China. Bidens parvt-
flora occurs in thirteen provinces and B. biternata in ten provinces.

However, the majority of these small genera have limited distributions.
Map 24 indicates that they concentrate in Taiwan, Kwangtung, Hainan,
Chekiang, Hupei, Szechuan, Yunnan, Hopei, Shansi, Sinkiang and, to a
lesser extent, in Mongolia, Liaoning and Heilungkiang. A comparison of
their distributions outside China and their concentration within the country
presents evidence of a correlation between the occurrence of these small
genera of Compositae and the courses of the ancient trade routes or the
ports of the newer waterways. These correlations are shown by the follow-
ing statistics:

(1) Within China the distributions of small genera with tropical Asian
and African range are like this: Hainan 5, Kwangtung 4, Taiwan 4, Sze-
chuan 3, Yunnan 2, Kweichow 2 and Fukien 1.

(2) Within China the distributions of small genera with pantropical
or tropical American range are like this: Taiwan 18, Kwangtung 16,
Hainan 14, Szechuan 10, Yunnan 9, Chekiang 3, Fukien 3, Hupei 2,
Kiangsu, Kweichow, Hunan and Kwangsi 1 each.

(3) Within China the distributions of small genera with tropical Asian,
Pacific Islands and Australian Range are like this: Taiwan 3, Hainan 3,
Kwangtung 2, Yunnan, Szechuan and Hupei | each.

(4) Within China the distributions of small genera with Central Asian-
Medit European-American Range are like this: Sinkiang 20, Tibet
6, Mongolia 5, Kansu, Shantung, Hopei, Szechuan, and Liaoning each 3,
Kiangsu, Kweichow and Heilungkiang each 2, Honan and Anhwei 1 each.

378 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

It is interesting to point out that the largest number of the tropical
Asiatic or African genera occur chiefly in Hainan, Kwangtung and Taiwan.
Their absence from the coastal towns in Chekiang and Kiangsu or the
metropolises along the Yangtze River, and their occurrence in Yunnan,
Kweichow and Szechuan, indicate that these genera were probably intro-
duced through the ancient trade routes connecting Rangoon (Burma) and
Yunnan (Map 31), or those connecting Hanoi (Indo-China) and Yunnan,
thence due north through Kweichow and Szechuan to the ancient Chinese
capital, Sian, in Shensi.

The pantropical genera are also concentrated in Taiwan, Kwanetung,
Hainan, and, to a lesser degree, in Yunnan and Szechuan. Some of them
also occur in Chekiang, Hupei and Kiangsu. It is evident that these genera
were introduced to the interior provinces of Yunnan, Szechuan, Kweichow,
etc., through the Yangtze waterway as well as through the Burma and
Indo-China trade routes.

The most striking facts are centered about the genera with central
Asiatic, Mediterranean and European distribution. Of these genera Sin-
kiang has the largest number, followed by Tibet and Mongolia. Liaoning
and Heilungkiang also have some species. It is evident that the distribu-
tions of these genera follow the ancient trade routes. The most-used trade
routes of ancient China passed through Sinkiang Province, and it is here
that the largest number of genera common to China and Central Asia,
the Mediterranean Region, and Europe are found.

Tibet is on the main ancient trade route that connected the upper
Gangetic Plain, Central Asia and west China. Heilungkiang is on the
eastern end of the great Northern Trade Route which connects Manchuria,
Siberia, Central Asia and Europe. The occurrence of genera of Com-
positae which are predominantly central Asiatic, Mediterranean or Euro-
pean indicates that these small genera are adventives to the flora of China.

Regarding the small genera of Compositae in China we may observe
that (1) most of the small genera are really small, about 70 per cent of
them including only one or two species: (2) slightly over one-fourth of
the small genera are endemics, many of them being aggregated in the
Meridional Ranges; (3) about half of the small genera are probably
adventives introduced to China either through the ancient trade routes or
through the more recent waterways.

(To be concluded)

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXXIX OcToBER 1958 NUMBER 4

STATISTICS OF COMPOSITAE IN RELATION TO THE
FLORA OF CHINA *

SHIU-YING Hu

III. A COMPARISON OF THE COMPOSITAE OF CHINA
WITH THOSE OF NEIGHBORING COUNTRIES

The nature of the Compositae in the flora of China can be understood
better and its significance more fully realized by a comparison of its com-
ponents with those of its neighboring countries, namely, Korea, Japan,
Indo-China, India, Central Asia (Pamir) and Siberia. The numbers of
species in genera common to those areas are given in TaBLe III.

TABLE III. A comparison of the genera of Compositae common to China
and its neighboring countries

NDO-
Tripes & GENERA CuinA Korea JAPAN CHINA INDIA pe SIBERIA

Vernonieae

Vernonia 34 — 1 29 45 — oe
Ethulia 1 a — 1 1 — —-
Elephantopus 2 — — 8 1 aa care
Camchaya 1 — — 1 — — —
Eupatorieae
Adenostemma 2 — 1 1 i — —
Ageratum 2 — 1 1 —- —-
Eupatorium 1 3 1 6 2 — —-
Mikania 3 — 1 1 — —
Astereae
Solidago 5 7 1 it — 1
Dichrocephala 3 — 1 2 = _ —-
Cyathocline 1 — 1 2 —- —
Lagenophora 1 — 1 1 — — —
Grangea 2 — — 1 1 — —
Rhynchospermum il 1 1 1 1 — —
Myriactis 5 1 S — —

* Continued from Volume XXXIX, p. 378.

380 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

TABLE III. (Continued)

N
Tripes & GENERA Cuina Korea JaPpaAN CHINA InptaA PAMIR SIBERIA

Asteromoea (Kalimeris

or Boltonia 10
Heteropappus
Aster

NN mn

Galatella 7
Erigeron 25
Microglossa 3

ra
WwW
~I

| ee | Reres
Ww

Pet oe
rowwal Flu
ws
=

ive)
=
=
iss)
a
ies)
Ww
—
Ww
iar
|

|

Blumeopsis
Laggera
Pluchea
Epaltes ee
Sphaeran
Ptero sani

ilago
Leontopodium

mm

many
Rea BDH DY WY RY CO

Phagnalon (Pr)
Gnaphalium 20
Helichrysum 2
Inula 28
Vicoa 2
Pulicaria 4

elleoll Illi dl

Q
~
Pu
a
6
x.
rar
[e<)
al | os |
(=
oO
— bo
ew OW OO bh se

ne
Q
s
=
Ss
a
2. =
=
bo
io

Siegesbeckia
Eclipta
Enhydra
Blainvillea
W edelia

til ae a OE ee OO, ee)
—

sig tN hea Hace cape hee dy es pg aes et es es
|
|

Helenieae
Tagetes 2 —- aaa o bes = ao

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 381

TABLE III. (Continued)

InDo-
TriBpes & GENERA CuInA KorEA JAPAN CHINA’ INDIA’ PAMiR SIBERIA
Anthemideae
Anthemis 4 — — — — — 1
Achillea 10 4 8 1 1 1 10
Matricaria 4 Z 2 H) — 6
Allardi 2 — —_— — 5 1 2
Chrysanthemum 73 10 30 2 4 7 5
Pyrethrum 5 — — — — Z 15
Brachanthemum 5 — — — — — 2
peda 1 1 1 1 1 — —
Crossostephium il — 1 1 — —
Artemisia 156 oT 40 4 27 20 58
Senecioneae
Tussilago 1 —- —— — 1 — 1
Petasites 9 ps 2 a —- —- |
Doronicum 3 — ) —- 4
Nardosmia (?) — — — -- — 4
Gynura 16 — — — 7 — _—
Emilia 3 — 1 4 5 —- —
Cacalia 60 7 25 — — —_— 1
rnica 1 — 4 — — —- 1
Syneilesis 4 2 2 — — —
Senecio 160 10 1 9 63 3 22
Ligularia 103 7 10 — 1 5
Farfugium 1 1 3 — — —— —
Cremanthodium 47 — a — 7 —_ —
Cynareae
Echinops 11 — 1 — 3 1 7
Atractylis 8 — 1 — — —- -—
Arctium 2 — 1 — 1 —- 2
Cousinia — — — 5 5 if
Carduus 11 1 1 — 2 —— 3
Cirstum 5 7 77 3 8 1 13
Cephalonoplos — 1 — — — —
Hemistepta 1 1 1 1 — — —
Saussurea 270 31 53 ) 39 7 25
urinea 8 —_ —_— — 3 1 4
Tricholepis 2 — 1 10 — —
Serratula 19 — 1 — 1 1 8
Centaurea 7 — — — 6 1 14
Carthamus 1 — — — 3 — —
Mutisieae
Pertya 10 _— 2 1 — — —
Leucomeris 1 — — 2 2 — —
Ainsliaea 47 3 10 — 4 — —
Gerbera 10 — — — 5 ~— 1
Cichorieae
Cichorium 2 — — — 1 — 1

382 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. xXxxIx

TABLE III. (Continued)

InNDO-

Tribes & GENERA Crina Korea Japan. Cuina Inpia PAMIR SIBERIA
~ Lapsana > — — — 1 — 1
Koelpin 1 — — — 1 1 1
Hypochoeris 3 1 1 == 1 —
Tragopogon 10 = — a 3 4 10
Scorzonera 18 2 2 —- 3 1 11
Picris 8 1 5 1 1 mo 2
Taraxacum 57 Z 37 — 2 2 22
Chondrilla 3 = -— —— 2 = 11
Sonchus 10 1 2 2 4 — 5
Launaea 4 — — 2 7 — —
Lactuca 57 4 6 9 22 1 7
Cicerbita 7 — ~ — -— _- 2
Prenanthes 19 — 2 — 6 — ---
Cre pis 31 — 2 6 14 3 14
Y oungia 30 6 6 — _— — —

verily 14 6 14 — — —— ——
Cre pidiastrum 2 ps 6 — — — —
Hololeion 1 a“ 2 — — — —
3 — 5 — 26

Hieracium 14 —

1. CHINA, KOREA AND JAPAN

The close floristic affinity between Korea, Japan and China is well
known, and for this reason plant geographers usually group the three coun-
tries in one phytogeographical province, the Eastern Asiatic Region. The
Compositae of these countries confirm this relationship. Kitamura, in his
monographic work Compositae Japonicaec, and Hara, in his Enumeratio
Spermato phvtarum ete arum, both include Korea as well as Japan. The
figures given in Taste III for these countries are abstracted from these
publications. It hee several interesting features which help us to
understand the flora of China and eastern Asia.
As indicated by the number of species in the genera which they have
in common, China has the more complex Composite flora. China and
Japan have 69 genera in common and in these there are a total of 1751
species in China and 514 species in Japan. The species/genus ratio for
China is 25, and for Japan 7.4. There are 44 genera common to China and
Korea and in these China has 1473 species and Korea has 203 species.
The species/genus ratio is 35.9 for China and 4.6 for Korea. It is inter-
esting to note that the differences of species/genus ratio are due chielly
to the large genera like Vernonia, Aster, Leontopodium, Anaphalis, Senecio,
Ligularia, Echinops, Saussurea, Prenanthes and Crepis. Many small genera
which may be regarded as adventives to the flora of China have the same,
or nearly the same, number of species in China, Korea and Japan. Solidago,
Xanthium, Siegesbeckia, Eclipta, Centipeda, Hemistepta, Hy pochoeris and
Lapsana are only a few examples.

rat
~—
—’

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 383

(2) Thirty-one of the thirty-seven large genera of Chinese Compositae
occur also in Japan. Twenty-five of these are also found in Korea. Numer-
ous species of these large genera are common to all three countries. For
example, Anaphalis sinica Hance, Artemisia annua Linn., A. borealis Pallas,
A. campestris Linn., A. japonica Thunb., Aster fastigiatus Fischer, A. in-
dicus Linn., A. scaber Thunb., Cup abrotanoides Linn., Chrysan-
themum indicum Linn., Cirsium pendulum Fischer, Gnaphalium affine
D. Don, Lactuca sibirica (Linn.) Benth. are only a few of them. In a
way, Korea and Japan are merely additional areas to the continuous range
of the large Chinese genera or of the widespread species. As observed by
Kitamura, many of these genera extend from China to Japan through
Korea a in fact, this is an important route of plant migration between
China and Japan.

(3) Geographically Japan is farther from China than is Korea but floris-
tically there are over one-third more Composite genera which China shares
with Japan than with Korea. This is probably due to the small size of
Korea and the corresponding limitations in the climatic and edaphic fac-
tors that support a more varied vegetation. It is worthy of note that the
genera which are common to China and Japan but absent from Korea are
pantropical taxa (lephantopus, Adenostemma and Emilia), tropical Asiatic
or African elements (Dichrocephala and Crossostephium), Australian ele-
ments (Lagenophora), and Sino-Japanese elements with the range extend-
ing from the eastern Himalayan region to Japan but not reaching Korea
(Myriactis and Pertya). The last named genera are especially interesting
because they illustrate some southern routes of plant migration between
China and Japan. These routes, as illustrated by Pertya, are evidently
from eastern China, especially through Chekiang or Taiwan to Japan.

(4) It is interesting to consider the genera of Chinese Compositae that
are absent from Korea and Japan. Among the large ones there are Blumea,
Gynura, Cremanthodium, Jurinea, Gerbera and Tragopogon. Among the
small ones, with the exclusion of the 34 which are Chinese endemics, there
are still 38 that occur in China but not in Korea or Japan. These are
(a) pantropical genera (Mikania, Grangea, Pluchea, Epaltes, Pulicaria),
(b) tropical Asiatic or African genera (Ethulia, Anisopappus, Sphacran-
thus), (c) the Australian genus, Pterocaulon, (d) American genera (Acan-
thospermum, Tridax, Sanvitalia), and (e) European-African-Central Asi-
atic genera (Allardia, Centaurea, Cousinia, Filago, Tragopogon). Species
of groups (a—d) occur only in the tropical and subtropical areas of China
(Taiwan, Hainan, Kwangtung, and Yunnan) while species of group (e)
occur chiefly in Sinkiang and Tibet.

(5) It is interesting to point out the genera of Compositae that are
represented better in Japan than in China or are absent from China. There
are five genera which are common to China and Japan, but better repre-
sented in Japan. These are Solidago, Heteropappus, Bidens, Cirsium, and
Crepidiastrum. There are 11 genera represented in the flora of Japan but
absent from China. These are Arnica, Dendrocacalia, Diaspanthus, Hetero-
kalimeris, Ixyoungia, Macroclinidium, Macropertya, Miricacalia, Paraixerts

384 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

and Senecillicacalia. Ten out of these 11 genera are segregates from Chi-
nese groups (Youngia, Ainsliaea, Pertya, Senecio and Cacalia). Arnica
is an American genus which, together with Solidago, another American
genus which is better represented in Japan than in China, may be taken
as an indication that in certain respects the Japanese flora contains more
North American elements than the flora of China.

2. CHINA AND INDO-CHINA

Gagnepain in 1924 in the third volume of Lecomte, Flore générale de
VIndo-Chine treated 78 genera and 205 species of Compositae. Sixty-eight
of these genera (87%) and 67 of these species (34%) are common to
China and Indo-China (Tasie III). This large percentage of common
genera and species indicates that there is a close floristic relationship be-
tween these two countries. A careful analysis of the data presented in
TaBLe III reveals that this statement is only partially true. First, many
genera of Compositae important in the natural flora of China are absent
from Indo-China. For example, 15 of the 38 large genera of Chinese Com-
positae, such as Cacalia, Ligularia, Taraxacum, Ainsliaea etc., do not occur
in Indo-China. Secondly, the genera best developed in China, having hun-
dreds of species there, such as Leontopodium, Anaphalis, Chrysanthemum,
Saussurea, Artemisia, etc. are represented in Indo-China by only 1-4
species. In China they occur chiefly in Hainan, Taiwan, Kwangtung and
some in Fukien and Yunnan. Thus we may conclude that there is a close
affinity between the Composite flora of the warmer regions of China and
Indo-China. There is no outstanding geographical barrier between these
two countries and, as might be expected, they have many species in
common.

The genera of Compositae of Indo-China are all small. Of the 69 occur-
ring in both countries there are 1365 species in China and only 199 species
in Indo-China. The species/genus ratio for these genera are 19.8 for China
and 2.88 for Indo-China. If it were not for the two pantropical genera
Vernonia and Blumea, this ratio would be even lower for Indo-China.

Indo-China is not suited to Compositae and its contribution to the Com-
posite flora of China is almost nil. Gagnepain described several new
genera for Indo-China. Two of these, Blumeopsis and Camchaya, have
been recorded from Yunnan. As there is a great chance of mistaking a
localized adventive for an indigenous genus, all of Gagnepain’s Indo-Chi-
nese genera of Compositae await verification through comparative study
with material from other parts of the world.

3. CHINA AND INDIA

Hooker, in 1881, in the third volume of the Flora of British India (in-
cluding Pakistan, Bhutan, Nepal and Burma) covered 127 genera and 591
species of Compositae. Ninety-four of these genera (74%) are common
to China and India (Taste ITI).

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 385

The Composite flora of India is richer than that of Indo-China. There
are more genera in that country and some genera have as many as 60
species. Nevertheless, when the Composite flora of India is compared with
that of China, it reveals that India is better represented at the generic
level and rather poor in species. This is shown by the species/genus ratios
of Compositae in the two countries. In China there are 2029 species for
its 167 genera of Compositae, with an average of 12 species to each genus.
In India there are only 591 species for its 127 genera, with an average of
less than 5 species to each genus. The high generic number and the low
species/genus ratio may be taken as an indication that India is a good
meeting ground for the genera of Compositae characteristic of many of
its neighboring countries, and a poor place for the generation of new
entities in the evolution of the family. In regard to the Composite flora
of China, there are more genera that migrate from China to India than in
the opposite direction.

The high species/genus ratio of the Chinese Compositae is due to the
occurrence of the 38 “large” genera (i.e., with ten or more species) in that
country. With the exception of Pertya, these genera also occur in India,
but only in Vernonia and Blumea are there more species in India than in
China. In the rest of the 35 genera, India has far fewer species. In the
following genera, for example, the numbers of species in China and India
are respectively: Aster 137:14, Leontopodium 57:1, Gnaphalium 20:7,
Artemisia 156:27, Cremanthodium 47:7, Senecio (including Cacalia)
220:63, Saussurea 279:39, Cirsium 59:8, Ainsliaea 47:4, and the Crepts
complex 96:14. Moreover, in India, the number of species belonging to
these genera is largest for the northern provinces, especially the southern
slopes of the Himalayas, and becomes gradually less toward the central
and southern provinces. Obviously, in the distribution of these genera,
India is on the periphery of their range. In these large genera, there are
many species common to China and India, e.g., Aster altaicus, A. tibeticus,
Anaphalis cuneifolia, A. triplinervis, Artemisia glauca, A. desertorum and
Saussurea deltoides. Saussurea is one of India’s largest genera of Com-
positae. Twenty-nine of its thirty-nine species (almost two-thirds) also
occur in China. Many of them are recorded only from the Himalayan
region of India, but are widespread in China.

There are 15 other genera of the 94 common to the two countries which
have more species in India than in China. These are: (1) Blumea,
(2) Cotula, (3) Dichrocephala, (4) Emilia, (5) Laggera, (6) Pluchea,
(7) Sphaeranthus, (8) Vernonia, (9) Allardia, (10) Carthamus, (11) Cous-
inia, (12) Launaea, (13) Pulicaria, (14) Cyathocline, and (15) Tricho-
lepis. Genera 1-8 are pantropical elements, occurring in Africa, Asia,
America and Australia. In the course of their migration, they may have
reached China by way of India or Burma, or they may have been intro-
duced from other tropical regions to tropical China independently. Genera
9-13 are central Asiatic or Mediterranean entities. There are many pos-
sible routes for their migration to China. Again, India, because of her
more numerous chances for communication with the Arabic world, may

386 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

have supplied routes for the migration of these genera to China, especially
through the Gangetic plains. Thus genera 1-13 cannot be considered as
Indian influences on the flora of China. Genera 14 and 15 (Cyathocline
and Tricholepis) are native of India and are Indian contributions to the
Composite flora of China. The number of their species is small, and their
effect on the flora of China is slight, however.

Twenty-eight genera of Indian Compositae are absent from China.
Twelve of these are genera endemic to India, occurring chiefly in the west-
ern peninsula (7 genera, e.g. Centratherum) but also in central India (2
genera, e.g. Lagascea), the western Himalayan region (2 genera, e.g. Cata-
mixis), and northern India (the genus Caesulia). The remaining 16 genera
have a wider distribution. Some of them extend from India westward
through Africa to the Canary Islands (/floga), or to the Mediterranean
region or Europe, (Volaturelia), or to western or central Asia (/pilasia).
Others are tropical genera occurring also in America, Australia and Africa
(e.g., Sclerocarpus and Chrysogonum).

4. CHINA, CENTRAL ASTA AND SIBERIA

Fedtschenko, in 1903 in his Flore du Pamir, covered 25 genera and 61
species of Compositae. All of them, except Kentrophvllum and Pterotheca,
are genera common to China and Pamir (Taser IIT). Tribes Vernonieae,
Eupatorieae, Mutisieae and Heliantheae are absent from Pamir. Senecio:
neae and Inuleae are very poorly represented, the former tribe with 4
species (3 in Senecio and 1 in Ligularia) and the latter tribe with only
two species (1 each in Leontopodium and /nula). Cichorieae and Astereae
are weakly represented in Pamir. The best developed tribes in this region
appear to be the Cynareae and Anthemideae. The largest genus is Arte-
misia, Which has 20 species. Cousinia is the only genus that has more
species in Pamir than in China. For the comparable genera, the species/
genus ratios are 63.5 for China and 2.5 for Pamir. The overahelminely
large number of Chinese species in most of the common genera of these
two regions seems to indicate that Pamir has had very slight influence in
the development of the Composite flora of China.

Krylov, in 1949 in his Florae Sibiriae Occidentalis, treated 68 genera and
399 species of Compositae. Fifty-two of these genera are common to China
and Siberia (Taser III). Tribes Vernonieae and Eupatorieae are absent
from western Siberia. Mutisieae are represented by only one species. Heli-
antheae are represented by only two genera, these famous for their weedy
species. In fact, three out of the five Siberian species of this tribe are
widespread taxa in China.

Twenty-five of the large genera in China also occur in western Siberia.
All of them except Hieracium ee much smaller numbers of species. For
the ae aes able genera there are 1729 species in China, with an average of
33 species to a genus, and nie 397 species in western Siberia, with an
average sf 7 species to a genus.

The only genera that have approximately the same number of species

1958| HU SRATISEICS “OF COMPOSI TAR AIN: CHUN x 387

in the two regions, or slightly more species in Siberia, are Pyrethrum,
Doronicum, Centaurea, Tragopogon, Achillea, Chondrilla and Hieracium.
It is very likely that in these genera the Chinese Composite flora expresses
the influence of the Siberian elements.

IV. A COMPARISON OF THE COMPOSITAE OF CHINA
WITH THOSE OF NORTH AMERICA

The recognition of the identity or close similarity of angiospermous
genera of eastern Asia and eastern North America is as old as the history
of plant taxonomy. This relationship was known before the publication
of Linnaeus’ Species Plantarum. In a proposition prepared for the debate
of a student, J. P. Halen, who was the respondent of the thesis entitled
“Plantae Camschatcenses Rariores,” Linnaeus in 1750 pointed out this
affinity by listing 11 species which were supposed to be common to North
America and Siberia (Linn. Amoen. Acad. 2: 336. 1752). About 1840
Asa Gray became keenly interested in the relationship of the flora of
Japan to that of the temperate part of North America. In a book review
on Siebold and Zuccarini, Flora Japonica (Am. Jour. Sci. Art. 39: 175-176
1840), Gray selected 14 species of the ornamental or otherwise generally
interesting plants of Japan and contrasted them with their closely related
North American forms. His interest in discovering the relationship of the
vegetation of the eastern sides of the two great continental masses in the
northern hemisphere lasted for a long time. By 1856, in his Statistics of
the Flora of the Northern United States, he concluded that there were
more genera characteristic of eastern North America that it shared with
an antipodal region, eastern temperate Asia, than with its neighboring

istrict, western North America.

Among his examples of extra-European genera common to eastern North
America and eastern Asia, Gray listed six genera of Compositae, namely,
Vernonia, Elephantopus, Diplopappus, Pluchea, Eclipta and Cacalia, At
that time the flora of China was practically unknown to the botanical
world. Our knowledge of that rich flora did not commence to develop
until some French missionaries, including David, Farges and Delavay, sent
their collections from Szechuan and Yunnan to Paris. Our summary of
the data scattered in publications on the Chinese flora yields 72 genera
of Compositae which are common to China and North America. They
belong to 10 tribes, namely Vernonieae, Eupatorieae, Astereae, Inuleae,
Heliantheae, Helenieae, Anthemideae, Senecioneae, Cynareae, and Cicho-
rieae.

An analysis of the numerous articles dealing with the floristic relation-
ships of eastern Asia and eastern North America reveals that the authors
have built up their evidence and accordingly drawn their conclusions from
areas that are not comparable in size, climate, degree of rainfall or other
factors which directly and indirectly affect the development and composi-
tion of vegetation. Actually, these authors were comparing a rather limited

388 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

area of the eastern United States with the entire “China Proper,’ which
includes all the regions commonly known as “North China,” “South China,”
“East China,” “Central China,” “West China,” “the Southwest” and “the
Northwest.” Some authors also included the Northeast, Korea and Japan.
These names, to the people in China who use them daily, do not refer to
regions with defined boundaries. Rather, they signify territories radiating
from the better-known metropolises over a distance of one to three hun-
dred miles. These centers are Peking and Tensin in the North, Nanking
and Shanghai in the East, Canton and Foochow in the South, Wu-Han
(Wuchang-Hanchow) in central China, Chengtu and Chungking in the
West, Kunming in the Southwest, Sian and Lanchow in the Northwest,
and Mukden and Harbin in the Northeast. Any general physical atlas
(for example, Plate 3 of Bartholomew’s The Times Survey Atlas of the
World) would show that eastern Asia, as such, includes a much greater
area, a more varied physiography, and more greatly diversified climate
and related ecological conditions, than does eastern North America. For
this reason, some of the conclusions drawn by eminent phytogeographers
on the floristic interrelations of these regions are rather misleading. When
the vegetation of comparable areas of eastern China and eastern United
States is analyzed, and when the known cultigens and adventives are ex-
cluded, the corresponding area of eastern China does not have two or three
times as many genera as that of eastern United States, as some authors
have claimed. In the case of Compositae, for example, the contrary is true.
There are far fewer genera and species in the corresponding part of China.
Moreover, the same analysis of the genera and species of Compositae gives
no evidence to confirm Gray’s well-known conclusion that eastern North
America shares more genera with its antipodal region in eastern Asia than
with its neighboring district of western North America

In outlining comparable areas of eastern China and eastern North Amer-
ica the extreme low temperature and the amount of annual rainfall are
considered as the determining factors. For obvious reasons the area covered
by Fernald’s eighth edition of Gray’s Manual is chosen as the basis for a
comparison. Map 25 shows that the January temperature of this area
varies from 10°F. in the north to 40°F. in the south, and the annual
mean temperature is approximately 60°F. The corresponding area in
China has a more southerly limit in the coastal area which reaches the
30th parallel. Fernald’s area on the North American mainland extends
approximately over Long. 65°—96°W. and Lat. 26°-50°N. (shaded area
of map). The area in China with the same annual isotherms and approxi-
mately the same length of growing season falls on Long. 104°-135°E. and
Lat. 30°-47°N. (shaded). In China this area covers parts of Kirin and
Heilungkiang, Liaoning, Jehol, Chahar, Suiyuan, Shensi, Shansi, Shantung,

opei, Honan, Kansu, northern Chekiang, Anhwei, Hupei and eastern
Szechuan, T he physical features of the land-mass are of very ancient
formation. The eastern portion of this area constitutes the alluvial plains
of the Lower Yangtze River, the Yellow River, the Liao-ho and the Sungari
River. The western portion of the area is an old plateau. The area in-

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 389

cludes many famous ranges where classical botanical collections were
made in the last hundred years. Tsingling (including Tai-po Shan, 2000-
4000 m.), Alashan (2000 m.), Lu-pan Shan, Taihang Shan, Ta-ching Shan,
Wu-tai Shan (3040 m.), Chang-po Shan, Tai Shan and Tien-mu Shan are
the well-known explored mountains. In rainfall and general climate this
area is comparable with Fernald’s area. Both of them have warm or hot
summers and cold winters, and both have a mean annual rainfall of 20-40
inches.

25. World map showing comparable areas of eastern China and eastern
North America (shaded). Solid lines mark the area with temperatures varying
from 10° F. in the north to 40° F. in the south during January. The broken
lines mark the area with an annual mean temperature of 60° F.

Regarding the Compositae, Fernald treated 115 genera, 82 of which are
native and 33 introduced. In the comparable area of China there are 94
genera of Compositae, 27 of which are known to have been introduced as
cultigens or weeds. A comparison of Fernald’s 115 genera with the Com-
positae of China indicates that 47 of them occur in China at large, and
only 37 of them occur in the corresponding area. Evidently of the genera
of Compositae common to the two countries, only about four-fifths are
shared by the corresponding area in China. A detailed comparison of these
common genera is presented in Taste IV. The total number of species for
China at large and that for the northwestern United States, as included
in Hitchcock’s Vascular Plants of the Pacific Northwest, are also given
for reference. The abbreviations in the column for Fernald’s area are:
E = naturalized from Europe, T Am = adventives of tropical America,
A = naturalized from Asia. For the corresponding area in China, such an
accurate record is lacking. The relatively short botanical history and
the prolonged period of human activities in the area make it forever im-

390

JOURNAL OF THE ARNOLD ARBORETUM

| VOL. XXXIX

possible to ascertain the origin of certain of its extra-Chinese elements of

Compositae.

These genera, the center of species concentration of which

are known to be in the Americas, the Mediterranean region, or western
or central Asia, are marked with (?).

TABLE IV.

of China and North Americ

Wien OF SPP.

GENUS

D ernonia

Elephantopus

Eupatorium

Erigeron
Anaphalis

Antennaria
aad
Pulic
ee
Nanthium

4 rs

Acanthospermum
Parthenium

Eclipta
Rudbeckia
Spilanthes
Bidens
Galinsoga

Achillea

Chrysanthemum

Cotula
Matricaria
Artemisia
Tussilago
Petasites
Erechtites

Lapsana

IN CHINA

AT LARGE

34

a
Sia n>)

i" bo Ww

i)

Oo

reny
Leal
NMOrR OPN ATOR DN BPR HR HNN FO NH H~INY

A Comparison of Genera of i hae Common to Limited Areas

No. OF SPP.

IN COMPARABLE
AREAS OF CHINA

o™
os)

tb
PN

ay

a

os wae ie
wv aa) =
Ww WY Ww

Dw

FPHOWNODDVOKFYOOFNHKHNINONNYOBOWOF

(oe)
o™~
Ww
oe

hd
Mr BwTW Ow HE

a
=
etal

No. OF SPP.

IN FERNALD’S
RANGE

Ro |
ot ae

~I

15
1(T Am)
3(T Am)

1
12(11 E)
1(E)
2
1

No. OF SPP.

IN HitCHCOcK’s
AREA

D we
NY HOONOO

bo
nn

i)
Or ODF DWN FH WO WNH BPHYHP WH HEF DROW OOONFH OC

ios)

—

—

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 391

TABLE IV. (Continued)

Tragopogon 10 3 KI@09) 5
ICris 8 2 2(E) 0
Taraxacum of Hy) 11(some E) 5
Sonchus 10 4 4(E) 4
Lactuca 57 23 16 6
Prenanthes 19 5 10 2
31 2 S(some E,A) 11
Hieracium 14 4

ioe)

|
|
|
|

i
1]

|
|

Several noteworthy points on the floristic relationship as expressed by
the Compositae of the two areas can be drawn from the above data. First,
in regard to the number of native genera, Fernald’s area has 82, while
the corresponding area in China has 67, which is one-fifth less than in
Fernald’s area.

Secondly, regarding the genera common to the two areas, in Fernald’s
area 22 out of the 47 genera which also occur in China are naturalized or
adventive. The sevenves of tropical American origin, such as Parthenium,
Galinsoga, Acanth

tc., occur in the warmer regions of China,
but they are Mbecnt coral ine corresponding area under discussion, This
phenomenon may indicate one of the two or both measures: that north-
eastern North America supports more elements of the warmer regions than
the corresponding area in Asia, and that due to the longer distance and
the shorter period of communications between this part of Asia and trop-
ical America these elements have not yet been introduced or established.
Thirdly, among the common genera native to both areas there are nine
genera which have far more species in Fernald’s area (Eupatorium, Soli-
dago, Aster, Erigeron, Gnaphalium, Xanthium, Bidens, Prenanthes and
Hieracium), while there are only three which have more species in the
corresponding area in China (Artemisia, Lactuca and Taraxacum).
astly, among the native genera, only Cacalia is common to the north-
eastern North America and the corresponding area of China while absent
from the Pacific Northwest. Meanwhile, there are three genera, Anten-
naria, Rudbeckia and Erechitites, which are common to the areas of
Fernald and of Hitchcock and are absent from the corresponding area of
China. In the Composite flora, there is no evidence that the northeastern
North America shares a larger number of genera with a comparable area
of eastern Asia than with her neighboring area in western North America.

V. THE REFLECTION OF COMPOSITAE ON THE
VEGETATION OF CHINA

In the foregoing analyses I have presented the general features of the
Chinese Compositae. Special emphasis was given to the constituent genera
of the family and their distribution. What light can the knowledge of this
overall picture of the largest family of flowering plants of China cast on

392 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxtx

the understanding of the vegetation of that country? The following dis-
cussion will be centered around this subject.

The Compositae reflect the very uneven floristic composition of the
vegetation of China. If the distributional maps of the Compositae (Maps
1-24) were superimposed on a single map, the resulting picture would
have various shades of darkness, with the darkest area, representing the
region richest in Compositae, falling over Yunnan, Szechuan and parts
of their neighboring provinces, and the lightest area, representing the poor-
est region, falling over Honan, Kiangsu, Shantung and parts of their neigh-
boring provinces. What is true of Compositae is also true of the general
vegetation. The region with the largest number of genera and species of
Compositae is also the richest floristic region in China, and the region
poorly represented with Compositae is also poor in natural vegetational
coverage.

1. THe AreA RICHEST IN COMPOSITAE SUPPORTS THE RICHEST
VEGETATION

The area of greatest floristic richness is very limited, being formed by
the Meridional Ranges which extend from western Yunnan northward
to eastern Sikang, western Szechuan and the adjacent territory in Kansu
and Shensi. Here, in order to present the vertical distribution of the pre-
dominant genera of the Compositae in these mountains, a brief account
of the complex physiography, which results in sharp changes of elevation
and climatic differentiation, is also given. The consequent diversified types
of vegetation are then described more fully.

This region consists of very ancient formations characterized by high
mountains and deep gorges. The mountains are formed mainly from mud-
shales and granitic rocks. Occasionally limestones have been forced up
through the older rocks to form bold peaks and stupendous precipices. At
the bottom of the deep gorges flow torrential tributaries of seven large
rivers which are, from west to east, the Chiukiang, Salween, Mekong, Chin-
shakiang, Yalungkiang, Tatuho and Min Rivers. Most of the tributaries
drain from mountains capped with perpetual snow. The principal courses
of the seven rivers all run from north to south, parallel to the meridian
and hence the mountains are known collectively as the Meridional Ranges.

At lower altitudes, the valleys of the large rivers are bordered by deeply
eroded treeless mountains. The climate here is hotter and drier than the
altitude warrants and barren areas and desert-like vegetation are common,
with Artemisia and Inula predominant among the Compositae. As one as-
cends the mountains along the tributaries of the rivers, the change in
topography and vegetation is sudden. Gentle slopes are inhabited by
various tribes, such as the Lolo, Miao, Chiarung, Ch’iang, etc., and up to
about 7500 ft. all the arable land is cultivated, the natural vegetation
being greatly disturbed. The slopes which are too steep to reach, and the
areas which are too far from human dwellings are covered with mesophytic
forest. Trees, shrubs and herbaceous undergrowth flourish. The much-

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 393

travelled and experienced collector, E. H. Wilson, considered this zone to
have one of the world’s richest vegetations. Along the roads skirting the
banks of the streams, on the edges of the forests, or on drier grassy slopes
are many species of Artemisia, Crepis, Aster, Eupatorium, Gnaphalium,
Arctium, Carpesium, Erigeron and Tussilago. The slopes from 7500-—
9000 ft, are covered by mixed forests of many species of deciduous trees
and conifers. On the southern flanks of the mountains bamboo forests,
mixed with some deciduous or evergreen trees are common. In the forests
or on the flood plains species of Senecio, Anaphalis, Cirsium, Ainsliaea
and Gerbera are common. From 9000 ft. to timber line the slopes are cov-
ered with virgin coniferous forests. The Composite family is rather poorly
represented in this zone. Above timber line, in the alpine meadows, with
the extraordinarily rich assemblage of herbaceous types, the family has
its best development. Many species of Artemisia, Saussurea, Dubyaea,
Soroseris, Cremanthodium, Ligularia, Aster and Jurinea form pure colonies.
Many others growing mixed with grasses, sedges, Aconitum, Saxifraga,
Delphinium, etc. are found. It is not an exaggeration to call the alpine
meadows of the Meridional Ranges the Land of Compositae. On the razor-
like ridges species of Leontopodium and Anaphalis form colonies. In fact,
species of Compositae can be found in all kinds of habitats in the alpine
region of the Meridional Ranges. In places where no other flowering plant
thrives, species of Compositae grow. Thus, immediately below the per-
petual snow line, in the rock cracks where there is a thin veneer of wind-
blown soil, one may find different species of Saussurea and Soroseris.

In considering the floristic richness of this area it must be remembered
that the seven rivers have hundreds of tributaries. Thus, in a very limited
region, the complicated habitats and the diversified vegetation are re-
peated several hundred times. The proximity of subtropical swamps, semi-
desert scrubs, mesophytic forests, grassy slopes, bamboo woods, coniferous
forests, alpine meadows and high-altitude tundra provide unusual oppor-
tunities for the close contact of many species. This brings about unique
chances for the hybridization of related forms. In many places these di-
versified habitats can be found within five miles of one another. The
heterogeneity of external conditions induces mutation and accelerates spe-
ciation. Frequent landslides after the annual monsoon storms, or occasional
earthquakes, provide new habitats for the colonization of new forms. All
these conditions are contributing factors to the rich Composite populations
of the Meridional Ranges.

As stated before, the region that is richest in Compositae is also the
land with the richest vegetation in China. It is rich in gymnosperms and
there is no comparable region in the world that has so many species of
Taxus, Cephalotaxus, Larix, Abies, Picea, Tsuga and Juniperus as the
Meridional Ranges in China. It is rich in broad-leaved trees and shrubs.
In fact, this region is the homeland of many garden specialties, especially
those in the genera Acer, Akebia, Berberis, Camellia, Clematis, Cotoneaster,
Deutzia, Euonymus, Hydrangea, Ilex, Jasminum, Kerria, Lonicera, Nan-
dina, Paeonia, Rhododendron, Rosa, Syringa, Viburnum, etc. It is par-

394 JOURNAL OF THE ARNOLD ARBORETUM - [vot. xxxix

ticularly rich in herbaceous types. Species of Aconitum, Allium, Anemone,

Cor vdalis, Cypripedium, Delphinium, Dianthus, Fritillaria, Lilium, Poly-

gonum, Potentilla, Rheum, Ranunculus, Simiried: Sedum, Pidicbum. etc.
are very numerous. There is no place in the world that can surpass this
region in the number of species of Primula, Gentiana, Meconopsis and
Pedicularis. It is also rich in monotypic or oligotypic families and genera.
Aucuba, Alangium, Coriaria, Delavayia, Dipteronia, Eucommia, Euptelea,

Euscaphis, Helwingia, Siac mrus and Tetracentron are only a few exam-
ples. They are important constituents of the mesophytic forests of the
region, As the mountains of this region generally reach 12,000— OOO Ths
alpine vegetation reaches the peak of its development. Grasses and sedges
are numerous both in kind and in individuals. The southern portion of
the region reaches the subtropics. The north-south direction of the valleys
of the main rivers favors the movement of tropical monsoon rainfall farther
north than the latitudes warrant. Species characteristic of the warmer
regions are well developed in the lower elevations. Cycas, Podocarpus,
palms, bamboos, lauraceous trees and shrubs, and large woody leguminous
vines characteristic of the tropical rain forests are abundant in the jungles
at low elevations. The experienced explorer F. K. Ward, after visiting
the southwestern corner of this region, called it the “Plant Hunter’s Para-
dise.” E. H. Wilson, after going through the northern portion of the
region, commented upon it as being the richest area in vegetation on earth.

The entire area is highly significant in the vegetation of China.

2, THe AREA POOREST IN COMPOSITAE HAS THE Poorest VEGETATION

This area is called the Central Plain (Chung-yuan) in ancient Chinese
literature. It is the plain on which the ancient history of China was built.
“ven in modern times it is still the focal point in the struggle for power
among the war lords, for whatever party gets this area gets control of
the national government. In modern geography and in current news this
area is called the North China Plain. In size it is approximately equal
to the region which has the richest vegetation, but in its physiography it
is very different.

Geographically this plain is the alluvial fan formed by the Yellow River.
It covers central and eastern Honan, southeastern Hopei, western Shan-
tung, northern Kiangsu and northern Anhwei. Throughout the territory
no elevation exceeds 200 ft. above sea level, excepting in northern Kiangsu,
where there are a few low, barren hills up to 600 ft. high. On the Kiangsu-
Anhwei border in the south there is a swampy lake, Hung-tse-hu, and on
the Kiangsu-Shantung border there is a similar lake, the Wei-shan-hu.
The changeable courses of the lower Yellow River radiate like the ribs
of a fan with Loh-yuan at the pivot, Tiensin on a rib to the north. and
Suchow on another rib to the south.

Geologically this area is a new land. It is so new that the configuration
of the surface has undergone noticeable changes in the last fifty years.
Around my home village (Kiangsu Province), the heavy deposits of the

—

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 395

frequent floods, so frequent that thirteen floods occurred in the summer
of 1924, have elevated the land up 3 ft., and the repeated erosions have
converted a former road into a river.

Historically this area is one of the earliest inhabited spots of the world.
The discovery of the Peking man, Sinanthropus pekinensis, in 1928, ma-
terially proves this proposition. The rich alluvial land, the temperate cli-
mate of this latitude and the timely rainfall, the maximum of which occurs
in July, all favor the development of agriculture. For four thousand or even
more years, the people in this area have been farmers. It is estimated that
at present every square mile of the cultivated land in this region supports
1479 people. Ninety per cent of them are farmers, who obtain their entire
livelihood from the products of the soil. For at least four thousand years
men have reshaped every inch of the land in this area, and there is no
spot with natural vegetation. All land surfaces that can possibly be culti-
vated are utilized. The small lots of farm land are carefully tilled. There
are no farms in the world that have so few weeds as the farms in this area.
Natural resources are utilized to the limit. After the planting of winter
wheat, and the harvesting of sweet potatoes and carrots, for hundreds of
miles at a stretch the land is of one brown color. The fields are turned
over. The fallen leaves are collected for fuel. The herbs on the roadsides,
or along the banks of canals are carefully scraped off with a thin veneer
of the top soil. When dry, the mixture of plants and earth is collected and
used for spreading over the floor of the animal house in winter. Eventually
this becomes the fertilizer for the fields the next spring.

Under such intense utilization of land, the species of Compositae occur-
ring in the area, with the exception of the cultivated forms, are no more
than can be counted on the fingers of two hands. These are all wide-
spread weeds. Eclipta alba is common in gardens and cotton or soybean
fields. Cirsium lineare is common in kao-liang (Sorghum) fields. Lactuca
tartarica and Artemisia campestris occur in alkaline soil. Saussurea affinis
and Ixeris chinensis occur in gardens and graveyards. Artemisia annua
and Xanthium strumarium occur on village commons. Inula britannica
occurs in protected woods on the outskirts of villages. Bidens chinensis
and Taraxacum officinale occur in graveyards. Scorzonera albicaulis is
limited to the arid region of the barren hills. With the exception of Eclipta
and Cirsium, the occurrence of all the above-mentioned species is occa-
sional. Thus the Composite flora of this area is poor in the number of
individuals as well as poor in kinds.

The region poor in Compositae is also poor in other vegetation in China.
Forests are unknown to this area. Woods protected by temples or well-
to-do families are rather rare. The component species are few. Thuja
orientalis, Salix babylonica, Populus alba, Juglans regia, Castanea mol-
lissima, Ulmus pumila, Ulmus parvifolia, Morus alba, Sophora japonica,
Ailanthus altissima, Melia azedarach and Euonymus bungeanus are the
common species. Lycium chinensis and Tamarix chinensis are often found
in long stretches of sand which may mark the course of a former river.
Herbaceous species are also few. Cynodon dactylon, Eleusine indica, Im-

396 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

perata cylindrica, Miscanthus sacchariflorus, Eremochloa colonum, Digi-
taria sanguinalis, Chenopodium album, Amaranthus viridis, Celosia argen-
tea, Acalypha australis, Apocynum sibiricum, Galium aparine and Mazus
rugosus are the common species.

3. THE CHARACTERISTIC RANGES OF SOME SUBGENERIC TAXA IN CHINA

Maps 26 and 27 represent the linear distribution of several species of
Leontopodium and Taraxacum. This is done by drawing a line through
the provinces where the species under discussion has been recorded. It is
interesting to note that the lines representing the distribution of each
species form a more or less modified S-shape. Looking at these distribu-
tional lines by provinces they connect Tibet, Sikang, Yunnan, sometimes
Kweichow, Szechuan, Kansu, Shansi, Shensi, Hopei or Suiyuan and Mon-
golia or Chahar, Heilungkiang, Kirin and Liaoning. Looking at them by
topography, they link up the Tibetan Plateau, the Meridional Ranges, the
Tsingling, the Taihang Shan or the Yin Shan, the Great Khingan, the

26, 27. Linear ranges of some species of two genera showing S-shaped
distributional patterns: 26, Species of Leontopodium; 27, species of Taraxacum.

Little Khingan and the Chang-po Shan. Although the length and the
shape of the lines representing different species may vary, they all pass
the Tsingling. Evidently, with regard to the distribution of these species,
the Tsingling is a bridge for their northeastward extension and not a
barrier to their dispersion.

Geographers in describing the physiognomy of China have emphasized
the ill effect of the Tsingling on the unity of the country. They have also
created a misleading impression that the Tsingling Range has been a bar-
rier to the distribution of animals and plants. For example, one author
(Cressey, 1934. p. 14-15) wrote, “Greatest of all the mountains of China
is the eastward extension of the Kuen Lun, known in China collectively as
the Tsingling Shan. . . . The mountains divide China into two major geo-

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 397

graphical regions, characterized by striking contrasts in climate, agriculture
and human activities. . . . There are two Chinas, each with distinct char-
acteristics in sharp contrast to those of the others. . . . One China is in
the South, a land of abundant rainfall. . . . This is the land of . . . rice
and bamboo . . . the people are shorter in stature. . . . The other China
is in the North, a land of limited and uncertain rainfall. . . . The stand-
ard crops are millet and Kaoliang and beans. . . . The people are taller.
. . . The South tends to be radical and revolutionary, while the north is
stolid and conservative. . . . The boundary between the North and the
South is transitional . . . it coincides with the crest of the Tsingling Shan.”
Another author (Lee, 1939, p. 2) in a more concise manner maintained,
“The Tsingling Range forms the natural divide between northern China
and the Yangtze Valley . . . these ranges that have naturally sharpened
the climatic contrasts and regional differences in other geographical con-
ditions, against which the Chinese have struggled for their unity during
historical time.”

Actually whether Tsingling does have such far-reaching influence on the
physical conditions of the land, on the distribution of plants and animals,
and on the life of the people is questionable. First, the mountains consti-
tuting the Tsingling Range are of unequal heights. As the range extends
from the border of Kansu-Szechuan-Shensi eastward to Hupei, Honan and
Anhwei, the elevation is gradually reduced. In the west the mountains are
continuous and often snow-capped, but to the east they seldom reach
600-1200 ft. in altitude. Moreover, there are many broad gaps from
Hupei eastward, and in Kiangsu Province there are only plains and hills.
The differences in climatic conditions and human activities described in
the foregoing quotations are found in the low land where the tail-end of
Tsingling is not high enough to be a climatic barrier. Actually the people
living to north and south in the mountains of the western end of Tsingling
have much more in common than those living in the plains in the east
where there are no mountain barriers. For example, the people in the
mountains of Szechuan, south of Tsingling, and those of Labrang in south-
western Kansu, north of the Tsingling, have more in common than the
people of Suchow and Shanghai, both in Kiangsu Province. Likewise in
plant distribution the high mountains constituting the western portion of
Tsingling are bridges over which the montane and alpine elements of
the south extend to the higher latitudes of the north and the boreal ele-
ments migrate to the high altitudes of the south. To regard Tsingling as a
bridge and not a barrier in the distribution of plants of the subgeneric level
is essential in the understanding of the vegetation of China.

In the Compositae many species in the genera Aster, Chrysanthemum,
Artemisia, Senecio, Cacalia, Leontopodium, Taraxacum, Ligularia, Saus-
surea and Lactuca extend from Tibet, Yunnan, Szechuan or Sikang in the
south, over the Tsingling Range to Shensi, Shansi, Kansu and even Hopei,
Manchuria and Korea in the north. Linear connections of these distributions
reveal an S-shaped pattern with Tsingling falling at the northern half be-
yond the middle. Maps 26 and 27 illustrate the S-shaped distribution of

398 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

the species of Leontopodium and Taraxacum which may be taken as exam-
ples of this type of distribution.

This is a pattern of distribution not limited to the Compositae alone,
but a general pattern of distribution of many taxa at the subgeneric level.
There are numerous examples in the Coniferae. Ostenfeld and Syrach-
Larsen in The Species of the Genus Larix (p. 18) give a perfect S-shaped
distribution of L. potaniniz. In fact this species is closely related to L. grif-
fithiana and L. mastersiana. Morphologically there is no clear-cut division
between them. Their status as species depends largely on the temperament
of the taxonomist. They can certainly be regarded as geographical variants.
Larix griffithiana occurs also in Yunnan. Looking at the distribution of
the group as a whole, the S-shaped range can be extended to the southern
Himalayas. Abies delavayt, A. georgei, A. forrestii, A. faberi, A. faxoniana
and A. chensiensis (French spelling for Shensi) present the same taxonomic
problem, and form the same S-shaped range. These species are fairly dis-
tinct, but there are some intermediate collections.

Maps 28-30. Distribution of three monotypic genera showing S-shaped ranges:
28, Ostryopsis davidiana; 29, Tetracentron sinense; 30, Euptelea pleiosperma.
Each dot represents a collection in the herbarium of the Arnold Arboretum.

Another example of the S-shaped range is found in Betula platyphylla.
This species is known to some botanists as B. japonica and to others as
B. mandshurica. In this case the taxonomists are more conservative and
the geographical variants are regarded as varieties. The five varieties of
this species, Betula platyphylla vars. rockii (from Yunnan), szechuanica
(Szechuan-Kansu-Shansi), mandshurica, kamtschatica and japonica form
an elongated S-shaped range extending from Yunnan through the Merid-
ional Ranges, the Tsingling, the Taihung Shan, the Yin Shan, the Khingan,
the Changpo Shan to Kamchatka and Japan.

Many monotypic genera also have similar patterns of distribution. To
cite a few examples, the two varieties of Ostryopsis davidiana (var. cin-
erascens and var. nobilis |Betulaceae|) form an S-shaped distributional
pattern over northwestern Yunnan, western Szechuan, Kansu, Shensi,
Shansi, Hopei and Chahar (Map 28). Tetracentron sinense (Pétracen-
traceae) forms a shorter curve but with the same pattern over Yunnan,

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 399

Szechuan, Shensi and Hupei (Map 29). Euptelea pleiosperma (Euptelea-
ceae) forms an even more symmetrical S-shaped distributional pattern
covering Sikang, Yunnan, Szechuan, Shensi, Honan, Hupei and Kweichow
(Map 30).

Many more cases can be cited from Gentiana, Potentilla and Saxifraga
to illustrate the S-shaped distribution of subgeneric entities. In short
it is conclusive that the mountain ranges including the Tsingling are bridges
for the distribution of plants. It should also be noted that all the examples
cited have small seeds which can be dispersed through adhesion to animals
and man as well as by wind. In many cases closely related forms also
occur in the northern Rockies. For example, Larix occidentalis and L.
potaninii, Philadelphus lewisti and P. tenuifolia are morphologically twin
species of the two continents.

4, ENDEMIC ELEMENTS

About one-third of the 167 recognized genera of Compositae in China
are endemic to that country. Most of them are monotypic or oligotypic
genera but several of them, such as Cremanthodium and Youngia, have
from 30 to 60 species. The occurrence of the endemic genera and species
of Compositae is illustrated in Map 32. In this map the lines represent

wees
m5
“PAS FRENCH
SF Py INDOCHINA
“y

ich’ ™.

Map 32. The occurrence of endemic genera (in lines) and species (in dots) of
Compositae in China.

400 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

monotypic or oligotypic genera and the dots represent species known only
from the type localities. Evidently the occurrence of the endemic genera
and species of Compositae is a widespread phenomenon in the country.
There seems to be no restriction on where endemism may occur, but it
appears clear that there are definite areas of concentration of the endemics.
The Compositae of various regions of the country have been studied by
botanists of different nationalities. The Japanese botanists have published
very liberally on the flora of Taiwan and on that of northern or north-
eastern China. The Russian botanists have described many new genera
and species from Mongolia and Sinkiang. British botanists have published
new taxa from Tibet and Yunnan. French, Austrian, Italian and German
botanists have published voluminously on the material collected from
central China. Individual differences in background, technique and tem-
perament on the part of these botanists naturally affect the quality of
the new genera or species they described. Some of them “split” more than
the others. Consequently the areas they work on appear to support more
endemics. In addition to these inevitable human defects, there is still
another difficulty which affects the value of the map showing endemism of
Compositae in China. This difficulty is that some areas are better explored
and more endemic elements have become known than in other areas. For
these reasons questions may be posed as to the validity of specific cases
on the occurrences of narrow endemics. Nevertheless, for the general
trend of endemism in the flora of China, the endemic genera and species
of Compositae present remarkable examples.

The Meridional Ranges, especially in northwestern Yunnan, have the
highest degree of endemism in the Compositae, both on the generic and
on the specific level. Many monotypic genera like Vierrhapperta, Formania,
Viadimiria and Leucomeris occur only in Yunnan and others like Wardaster
and Bolocephalus occur only in Sikang. Many oligotypic genera occur
throughout the region forming S-patterns in Yunnan, Sikang, Szechuan and
Kansu or Shensi. Mvyriactis, Stereosanthus, Nouellia and Dubyaea are a
few examples. Some of them such as Faberia extend the range to Kweichow
while others such as Soroseris, Tricholepis and Myriactis extend the range
to Chinghai and Sinkiang or Tibet and North India. The high degree of
endemism in the Meridional Ranges and the extension of some genera to
the Himalayas, to Tibet or to Kweichow represent common patterns of
the occurrence of monotypic or oligotypic endemic genera in the flora of
China. For example, Docynia delavayi (Rosaceae), occurs only in Yunnan,
while Tetracentron sinense (Tetracentraceae) is very numerous at mid:
high altitudes both in Yunnan and Szechuan. LEuptelea pletosperma
(Eupteleaceae), Decaisnea fargesti (Lardizabalaceae), and Sibiraea lae-
vigata var. angustifolia (Rosaceae) are abundant in this region but with
their ranges extended to India, Shensi, Hupei, or to Kansu and Sinkiang.

Taiwan is another area with a high degree of endemism, largely on the
specific level. There are, however, a few oligotypic endemic genera which
Taiwan shares with the mainland, especially with Yunnan. For example,
there are 5 valid species of Myriactis on the mainland. These are concen-

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 401

trated in Yunnan and its adjacent regions, one as an endemic, one ex-
tending to Szechuan, two extending to Nepal and Kweichow. There are
also one species and two varieties of the same genus in Taiwan. Rhyn-
chospermum verticillatum is another species which occurs in Yunnan,
Szechuan and Taiwan. There is a close tie between the flora of Yunnan
and Taiwan. There are many subgeneric entities that are common to both
regions and absent in areas between them. This tie is best expressed in
some small endemic genera.

North China has several endemic genera of Compositae. Some of them
are restricted only to one province. Tugarinovi and Stilpnolepis are known
only from Suiyuan. Takeikadzuchia is restricted to Chahar. Others have
wider ranges. Myripnois was first described from Peking. Additional col-
lections have extended its range to Shansi and Kansu. Filifolium is known
from Hopei to Heilungkiang. Probably many of these genera are due to
the splitting activities of some Japanese and Russian botanists. The g
eral flora evidently does not have a proportionate number of isolated
endemics on the generic level. In general there are fewer endemics in
North China. The genera characteristic of the flora of the region often ex-
tend to East China. Xanthoceras sorbifolia was first published from Peking.
Additional collections extend its range to Shansi and Kansu. Hemiptelea
davidii was also described from Peking but material in the herbarium of
the Arnold Arboretum shows that it occurs in the Lower Yangtze Valley in
the south, and in Manchuria and Korea in the north.

Northwestern China also has several endemic genera of Compositae.
Brachanthemum is recorded from Kansu, Sinkiang and Mongolia. Xan-
thopappus from Kansu and Chinghai, Olgaea from Ninghsia, Shansi,
Kansu, Sinkiang and Mongolia, and Asterothamnus is recorded from
Sinkiang and Mongolia. The occurrence of these endemic genera of Com-
positae reflects the special character of the vegetation of the prevailing
desert condition of this region. In the general flora endemic monotypic
or oligotypic genera are not uncommon, especially in Caryophyllaceae,
Tamaricaceae, Cruciferae and Zygophyllaceae. For example, Acantho-
phyllum spinos (Caryophyllaceae) occurs only in Sinkiang and the adjacent
area of Mongolia. The monotypic genus Tetradena (T. mongolica, Zygo-
phyllaceae), is endemic to the Kusuptschi desert of southwestern Suiyuan,
and the oligotypic genus Reaumuria (Tamaricaceae) is represented in
Kansu by R. ¢trigyna and in Sinkiang and northwestern Mongolia by
R. soongorica.

The region drained by the Middle Yangtze, that is, the Hupei-Szechuan
border and the adjacent area of Shensi, marks the eastern end of Tsingling.
Like the Meridional Ranges this is also a river-gorge country. But here
the mountains walling the gorges of Yangtze, Hanshui and their numerous
tributaries are only of moderate height. In general the river beds are about
600 ft. above sea level, and the altitudes of the mountains vary from
1800 to 7500 ft. The area is thickly populated and the vegetation is greatly
disturbed. Some forests are preserved in the less accessible areas. Botan-
ically this is the best known area of the country, for much has been pub-

402 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

lished on the extensive collections made by A. Henry, E. H. Wilson, P. C.
Silvestri and P. Farges from this region. In this region the endemism
among Compositae is not as striking as that in the Meridional Ranges.
On the generic level there is no genus which is limited to this region in
particular. The genera that are endemic to China and also occur here
seem to have their centers of distribution elsewhere. For example, Ainsliaea
and Pertya both occur here but the center of their species concentration
is evidently to the west of this area, in the Meridional Ranges. Skeareria
also occurs here but the core of its range seems to be to the east of this
area in the Lower Yangtze Valley. The endemics of this area are largely
of the specific or subspecific levels, and they are relatively fewer than
those of the Meridional Ranges. Senecio doryotus, Ligularia fargesti,
Ainsliaea henryi, and Pertya sinensis are a few examples of the endemic
species. Ligularia hodgsonii var. pulchella and Anaphalis sinica var.
calvescens are some examples of the infraspecific endemics.

It seems that in this area endemism in Compositae does not reflect the
picture of endemism in the general flora. It is true that this area seems
to be on the periphery of the ranges of some endemic genera, for example,
in the distribution of Tetracentron sinense (Tetracentraceae), which forms
an S-shaped range over Yunnan, western Szechuan, southern Shensi and
western Hupei (Map 29). As remarked by E. H. Wilson, this monotypic
genus is common in western Szechuan and rare in this area. Apparently
western Hupei marks the eastern limit of the range of the genus. For-
tunearia sinensis (Hamamelidaceae) is common in the lower Yangtze re-
gion, that is, the Anhwei-Chekiang-Kiangsu border. One collection from
Chikungshan of southern Honan and one collection from Franghsien of
western Hupei mark the western limit of the range of the species. Loro-
petalum chinensis (Hamamelidaceae) is widespread in the warmer region
of China. It occurs in woods along the lower Yangtze valley and thence
extends southward to Kwangtung and Kwangsi. Apparently its occurrence
in the mid-Yangtze marks the northwestern limit of the range of this
monotypic genus which is characteristic of the mesophytic forest in the
warmer part of the country. On the other hand, this region constitutes

genera, For example, Decaisnea fargesii (Lardizabalaceae) and Sino-
menium acutum (Menispermaceae) both have S-shaped distributions in
Yunnan, Szechuan and Hupei. With D. fargesii the range extends bilater-
ally sail with evident disjunction to Sikkim, on the west, and Huang Shan
in Anhwei, on the east (Map 33). With S. acutum the range extends east-
ward as a narrow band to Japan (Map 34). In both genera the mid-
Yangtze region is a part of the central core of their distribution. It is
noteworthy that Sinowilsonia henryi (Hamamelidaceae) is restricted to
the mountains of northwestern Hupei. Cercidiphyllum japonicum var.
sinense (Cercidiphyllaceae) has an equiformal distribution from this region
westward to western Szechuan and eastward to southern Anhwei. The
occurrence of woody endemics in the mid-Yangtze and in the Meridional
Ranges (which are sometimes called the Upper Yangtze) seems to tie the

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 403

flora of the mid-high altitudes together. The relative paucity of en-
demics of Compositae in the mid-Yangtze region is apparently due to the
absence of high mountains and alpine vegetation in the area.

Maes 33, 34. Mid-Yangtze region as the center of two monotypic genera: 33,
Decaisnea fargesit; 34, Sinomenium acutum and its variety.

The region generally called the lower Yangtze valley is another area
characterized by a moderate degree of endemism. It covers northern
Kiangsi, southern Anhwei, southern Kiangsu and the adjacent Chekiang.
This area is a land of numerous hills. The botanically better-known ones
are Lu Shan, Huang Shan, Tien-mu Shan, and the Nanking Hills, includ-
ing the Ox Head Hills. Their altitudes vary from 900 to 4500 ft. The
area is thickly populated and the vegetation is greatly disturbed. Endemism
in the Compositae is at the specific or infraspecific level. For example,
Pertya desmocephala is a narrow endemic which is known only from the
type locality in Chekiang. Youngia japonica ssp. elstonii is known only
from Chekiang and Kiangsu. At the generic level there are no Compositae
limited to this area. Sheareria was first described from Kuling in northern
Kiangsi. Additional collections extend its range to Chekiang, Hupei, Hunan
and even northern Kwangtung. Synurus occurs in Kiangsi and Chekiang,
but its range extends northward to Heilungkiang and to Japan. There are
certain widespread endemic Chinese genera which are exceedingly abun-
dant in this area. Asteromoea, Hemistepta, Youngia, and Ixeris are a few
examples. They generally occur as weeds. Their origin is obscure and
their ranges can hardly be regarded as illustrating certain distributional
patterns. Endemism in the general flora of this area is more evident at
the generic level than it is in the Compositae. The monotypic genera
Fortunearia (F. sinensis, Hamamelidaceae) and Fontanesia (F. fortunei,
Oleaceae) evidently have the center of their ranges in this area. The latter
species apparently occurs more or less as a cultigen. The occurrence of
several woody endemics in this area illustrates the same principle as ex-
pressed by the Composite genera. There is a close tie between the flora
of this region and that of North, Central and South China. For example,
Hemiptelea davidii (Ulmaceae) ranges from this area northward to Hopei,
Shansi, Heilungkiang and Korea. The occurrence of Decaisnea fargesii

404 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

(Lardizabalaceae) and Stephandra chinensis (Rosaceae) in Huang Shan
of southern Anhwei ties the flora of this area to that of Central and West
China. Fortunella hindsii and Ilex lohfauensis are characteristic elements
of the Wu-yi Range of southeastern China, and their northern limits are
in southern Anhwei or Chekiang. The occurrence of Cercidiphyllum
japonicum var. sinensis (Cercidiphyllaceae), Simomenium acutum var.
cinereum (Lardizabalaceae) in Huang Shan, Anhwei, and of Ilex lati-
folia in southern Kiangsu, Anhwei and Chekiang shows the very close
relationship between the floras of this area and Japan.

The region drained by the Pearl River and its tributaries is generally
known as South China. This area is a land of hills and mountains which
are collectively called Nanling. The Nanling Range extends along with the
tropic of Cancer from the border of Yunnan and Kwangsi eastward to the
Fukien-Kiangsi-Chekiang border. The mountains are about 2000 to 4000
ft, above sea level. The vegetation is subtropical. In the last 50 years ex-
tensive botanical explorations have been made in the area largely through
the cooperation of local Chinese universities and the Arnold Arboretum of
Harvard University. Although much of the accumulated material awaits
careful study, what. has already been published is sufficient to indicate that
this region possesses a relatively high degree of endemism. In the Composi-
tae, at the generic level, there is the monotypic genus Heteroplexis (H.
vernonioides) from Kwangsi. At the specific level, there are Vernonia
chingiana from Kwangsi, V. solanifolia from Kwangtung, Ainsliaea cleisto-
gama and A. parvifolia from Kwangtung, and A. plantaginifolia from south-
ern Hunan. The general flora of this region shows the same pattern of
endemism as reflected by the Compositae. In the north, the monotypic
Handeliodendron (H. bodinieri, Sapindaceae) is restricted to the border re-
gion of Kweichow and Kwangsi. In the south, the monotypic genus Myti-
laria (M. laosensis, Hamamelidaceae) is restricted to the Yunnan-Kwangsi
and Indo-China border. Concentrated in this area but with wider distribu-
tion are Bretschneidera sinensis (Bretschneideraceae, a monotypic family)
and Eustigma oblongifolium (Hamamelidaceae). The former species radi-
ates in an equiformal area covering Kwangsi to northern Kwangtung on the
east, eastern Yunnan on the west, southern Kweichow and Hunan in the
north, and northern Indo-China in the south. The latter covers almost the
same range and it extends even to Taiwan.

There are a few genera of Compositae which appear to be endemic to the
flora of China as widespread weeds or as cultigens. They may be mono-
typic, as Hemistepta and Callistephus, or they may occur as oligotypic
genera, each having one widespread variable species and a few isolated en-
demic species, such as Asteromoea, Youngia and Ixeris. This condition of
endemism associated with man’s activities is a common phenomenon in the
general flora of China. For example, Ginkgo (G. biloba, Ginkgoaceae),
Metasequoia (M. glyptostroboides, Pinaceae), Broussonnetia (B. papyri-
fera, Moraceae), Platycarya (P. strobilacea, Juglandaceae), Pteroceltis

. tatarinowti, Ulmaceae), Nandina (N. domestica, Berberidaceae),
Chimonanthus (C. praecox, Calycanthaceae), Melia (M. azedarach, Meli-

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 405

aceae), and Ailanthus (A. altissima, Simarubaceae) are all Chinese species
or monotypic genera, the wild state of which is obscure, and their existence
is associated with man. Probably this condition is brought about largely
through man’s continuous destruction of the natural vegetation. Since the
period when the angiosperms became dominant features of the world’s flora,
there is no geological evidence that catastrophic changes have annihilated
the plants of any extensive area in China as did the Pleistocene glaciers in
Europe and America. Yet with the exception of the less accessible areas of
the Meridional Ranges, there is hardly any area covered with natural vege-
tation. In most places poverty is a striking feature of the flora. Unless pro-
tected by the Buddhists as temple property, woods are rare. China is an
old country, and it has long been extremely overpopulated. There has been
a constantly greater demand for food than the arable land can produce.
The conversion of forested areas into temporary farms by burning the hill-
sides is a common practice. Repeated intentional forest fires have denuded
the mountains throughout the country, exterminated many species and left
many others to represent isolated endemic genera or families.

In conclusion, the endemics in the Chinese Compositae reflect a very fair
picture of endemism in the general flora of the country except in the mid-
Yangtze region where there are proportionately more isolated woody en-
demics at the generic level. The lack of alpine vegetation in this area and
the protection from human destruction which the steep gorges afford the
vegetation of limited areas are probably the chief contributing factors of
this situation.

5. ExtTrA-CHINESE ELEMENTS

About 48% of the genera of Compositae in China are probably not of
Chinese origin. Some of them have as many as 30 species, while others have
only one or two or a few species. Their distribution in China is widespread
either as weeds or cultigens or localized in coastal regions or port areas as
adventives. In general, the tropical elements such as Vernonia, Gynura,
Blumea, Emilia, Erechtites, etc. are concentrated in the south, especially
in Yunnan, Kwangtung, Hainan and Taiwan. The Mediterranean, Central
Asiatic and European elements such as Tragopogon, Achillea, Cousinia,
Echinops, Arctium, Matricaria, etc. are limited to the west and the north,
especially to Sinkiang, Tibet, Mongolia and Hopei. It is noteworthy that
most of the widespread weeds like Erigeron canadensis, Xanthium stru-
marium, Eclipta prostrata, Bidens pilosa, etc. appear to be of New World
origin.

As America is a melting pot of the modern world, so was China in the
ancient historical times, and the flora of China is the result of this min-
gling. In 126 B.C. a general of the Chinese Empire subdued the people
and annexed the country of the Iaxartes and Oxus rivers in Central Asia
to the Han Dynasty. Since then, for approximately sixteen centuries, China
has had more influence on land and sea in the Old World than most people
realize. With the conquest of Central Asia began the silk trade. the most

406 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. xxxix

far-reaching large-scale overland commerce of the ancient time. It reached
from the Pacific coast on the east to the shores of Britain on the west. Al-
though the transactions were carried on through intermediate merchants,
large-scale movement of men and animals provided the opportunity for the
introduction of plants accidentally. Caravans of hundreds of horses, carts,
yaks or camels passed back and forth over the great highways of Central
Asia. The diary of a Taoist monk recorded that it took that 73-year-old
man (Kiu Ch’ang-chung) and his 16 disciples 10 months to travel from
Peking to northeastern Mongolia and thence westward to the Altai and Tien
Shan mountains, and across the Pamir Plateau to reach Samarkand in 1222
A.D. (Bretschneider, 1875, p. 15-56). On their return trip it took them
only three months and ten days to reach Peking through southern Mongolia,
the regular postal route. The military, diplomatic, commercial and religious
intercourse with Persia, India and Arabia overland (Map 31), and through
them with Europe and the Mediterranean world explain the presence of
western Asiatic, Mediterranean and European elements in western and
northern China.

Meanwhile, China was a sea power in the Pacific and Indian oceans for
eleven centuries. There were regular communications between the mother
country and the overseas Chinese in Malaysia. The commercial centers in
Indo-China, Siam, India, Ceylon, Persia, Arabia and the East Coast of
Africa were frequently visited by Chinese fleets (Map 31). For example, in
1405 A.D.a fleet of 62 ships and 27,800 men, under the command of Cheng
Huo went on regular patrol duties. The expedition took two and a half

ears. During his lifetime Cheng Huo made seven such expeditions and
reached as far as Arabia and East Africa. With the periodical return of the
overseas Chinese from the Malaysian islands and important ports of trop-
ical Asia, and with the expeditions of the Chinese fleets, many tropical plants
were either accidentally or intentionally introduced to the coastal regions of
China. For example, Chrysanthemum segetum was brought in from the
Arabian world to China for its edible young shoot. Sphaeranthus africanus
has been introduced to Hainan Island and Taiwan, and a very closely re-
lated species, S. senegalensis, to Yunnan.

Unawareness of the interchange of plants as associated with human activi-
ties in the ancient historical times has created many unnecessary problems
in plant taxonomy. Chrysanthemum segetum was brought in by a colony
of Arabians who came to Canton about the fourth century. It has been
adopted by the Cantonese as a vegetable and has spread with them to all
warmer parts of China. Chinese plants in most herbaria have been named
C. coronarium. Bailey in 1917 saw the plant in Chinese gardens and named
it C. coronarium var, spatiosum in 1920. The same cultivar has been named
by Loureiro as Buphthalmum oleraceum from Canton in 1790. Likewise,
material of Sphaeranthus africanus from South China has been named S.
cochinchinensis by Loureiro, S. suberiflorus by Hayata.

This confusion is not limited to the nomenclature of species of Composi-
tae. It is a common disorder in the taxonomy of many Chinese plants.
Angraecum falcatum was published from a plant supposedly of Chinese

HU, STATISTICS OF COMPOSITAE IN CHINA 407

1958]

[east conerruve

MEST LONGITUDE

[or Sf or

XV

os i eae oc 7 ORL

ou 9 os or Ot

SNOILO3S SIH 3LvNSIS3G 22-1 ‘MOODEVa AB : fea it |
ry \
| GasodOud S¥ SIda¥d JO SALNCY NOLIWYSIN: —— NOILD3CONd V34Y-IWNDF WHLANIZY S.Lu36NvI y \\. k
BINA NVWOM 40003 yh
$31N0¥ Vas ----- t

(Q31LIWO AVZ9YV I B¥idWw3 NYWOY 3H1 30 3SOHL) is ate
NIDINO OLWISY JO 3VLISOdNOD JO” —
NOILNGIYLSIO SHLNI ONINV3SS MISHL

8 QV OOI ‘v9 SaLnou sovYL YyOrvN |
\ \ ;

me

as

The major trade routes ca. 100 A.D. and their bearing on the dis-

tribution of Compositae of Asiatic origin. (Base map Goode’s No. 205.)

Map 31.

408 JOURNAL OF THE ARNOLD ARBORETUM - [vor. xxxrx

origin. In the last 170 years orchidologists have placed it in nine different
genera and a long list of synonyms has been created on its account. In
China it has only been collected from the vicinity of Ningpo in Chekiang,
a center of Chinese Buddhism. There is no other Chinese orchid that is
closely related to it. In the Ames Orchid Herbarium of Harvard University
I found its nearest relative to. be an African species, A. pusillum. Perhaps
this isolated Chinese cultigen was one of the ancient introductions from
Africa.

The migration of plants accompanying the large body of human move-
ment is always reciprocal. There were Chinese plants introduced and estab-
lished in Africa long before Linnaeus’ time. Myrsine africana Linn. was
originally described from Ethiopia. The wide distribution of the species in
the interior of China and concentration of related species of the genus in
eastern Asia indicate that the natural origin of M. africana is Chinese.

VI. THE REFLECTION OF CHINESE COMPOSITAE ON THE
ORIGIN AND ROUTES OF MIGRATION OF SOME
MEMBERS OF THE FAMILY

It is well known that China is one of the most unique phytogeographical
regions of the world. Before the data on the Compositae of this important
phytogeographical region were available, James Small had proposed a hypo-
thetical scheme for the origin and development of the Compositae of the
world. Although his scheme has been questioned and disproved by compe-
tent monographers of special groups, such as Babcock and Stebbins, it has
also been followed by synantherologists who work on the entire family of
certain regions, like Kitamura. Now, what advancement can a better
knowledge of the Chinese Compositae offer to the understanding of the
origin of and routes of migration of some genera of Compositae?

Small’s long article on the Origin and Development of the Compositae
was published in thirteen issues of The New Phytologist. His main thesis
“is that Senecio was the first genus of the Compositae to come into existence
and that it has directly or indirectly given rise to all the other genera of the
family.” He considered “the origin of one definite kind of living organism
from another definite kind of living organism as a normal, natural result
of the actual diving of the parent organism in a particular region,’ and,
upon many suppositions, he derived Senecio from the Siphocampylus-Cen-
tropogon group of Lobelioideae. As these genera are natives of tropical
America, he proposed that Senecio and, in turn, Compositae originated in or
about the Bolivian region of South America. He suggested further that
from tropical South America the genus migrated along the Andes north-
ward to Central America and then along the Cordilleran system to Alaska,
thence to the Old World. He maintained that throughout the world the
path of migration of the genus is commonly along the mountain ranges,
usually about 3000 ft. and frequently above 6000 ft. To illustrate his scheme
of the development of the subdivisions of the family he made the large

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 409

number of species of Senecio to represent the substantial trunk and con-
structed a family tree for the Compositae, employing this to show the evo-
lution of the principal subdivisions through certain modern genera in time
and space.

Although Small’s article contains informative summaries of former works
on Compositae and helpful observations which advance our understanding
of the family, his conclusion on the origin and development of Compositae
is unsupportable. It represents too much ‘‘mental effort,’ in Small’s own
words, and little truth. First, the origin of an inclusive family like the
Compositae is not so traceable as to have a living parent, as Small put it.
Recent researches in smaller categories of Compositae, for example Bab-
cock’s Crepis, Keck’s Artemisia and Stebbins’ Dubyaea, Soroseris, etc., all
indicate that present evidences are not sufficient even to trace the precise
progenitors of a genus, a section or a subtribe of Compositae. No one can
pretend to know the exact origin and development of the Compositae. Sec-
ondly, Senecio is not a primitive genus in the Compositae. It is large, in-
clusive and heterogenous. It is inevitable that such an artificial genus is
polyphyletic. Thirdly, in constructing the family tree to show the evolution
of the Compositae in time and space, a very important geobotanical region,
central and eastern Asia, including China, was not included. This omission
of a region which is vital in the origin, development and distribution of
plants, animals, men and culture, naturally shifted the actual points of ori-
gin of many genera or even tribes of Compositae to some assumed areas.

1. ON THE ORIGIN OF COMPOSITAE

The origin of Compositae is obscure and it may never be elucidated.
There are many gaps in the fabric of the evolution of the angiosperms. The
missing progenitor of the Compositae is but one of them. Paleobotanical
evidences in the Cromerian Beds in England, in the Teglian and the Reu-
verian Beds in central and southern Europe and in the Wilcox Beds in
North America indicate that members of Compositae were widespread
throughout the northern hemisphere in the Oligocene era. Fossil remains
of achenes distinctly resemble those of modern species including Tussilago
farfara, Lapsana communis, Picris hieracioides, Crepis fuscipappus, Cardu-
us nutans, Cnicus palustris, Cirsium heterophyllum, and Eupatorium
japonicum, indicating that members of the Senecioneae, Cichorieae, Cynar-
eae and Astereae were common in the Old World in the Pliocene era. Evi-
dence for a progenitor of Compositae is lacking, and the home of the origin
of the family is obscure, as is that of the angiosperms. Large genera with
hundreds or thousands of species, like Senecio and Aster are heterogenous.
Their subgenera or sections may possibly represent taxa of entirely different
origins.

2. ON THE ASIATIC ORIGIN OF SOME WIDESPREAD GENERA

Babcock’s classical treatment of the genus Crepis elucidated the origin
and routes of migration of a complicated, widespread genus of Compositae.

410 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

His conclusions throw light upon our understanding of the origin and dis-
tribution of many widespread genera which have distributional patterns
similar to that of Crepis (e.g., Aster, Leontopodium, Artemisia, Chrysan-
themum including Tanacetum, Senecio including Cacalia, Ligularia and
Cremanthodium, Saussurea, Jurinea, Taraxacum and Lactuca). These gen-
era were formerly regarded as of European or American origin. Additional
evidence favors the suggestion that they are actually of Asiatic origin. As
illustrated in Tasie V, China has the largest number of species in Asia in
each of these genera. In China the Meridional Ranges appear to be the cen-
ter of their species concentration. One may safely conclude that Merid-
ional Ranges are the home of these large genera of Compositae.

TABLE V. Widespread genera with an unusually large number of species in China

InpDO-CHINA INDIA CHINA Japan Korea PAMIR SIBERIA

Aster 1 14 137 35 9 2 9
Leontopodium 1 1 57 6 3 1 2
Gnaphalinum 3 7 20 a 4 0 4
Anaphalis @ 0 51 5 1 @) 0
Carpesium y 2 18 10 5 0 0
Artemisia 4 zy 156 40 a7 20 58
Chrysanthemum 2 4 73 30 10 7 3
Cacalia 0 0 60 25 7 0 1
Senecio 9 63 160 13 10 3 22
Ligularia 0) 0 105 10 7 1 5
Cremanthodium 0 7 47 @ 0 0 0
Saussurea 2 39 270 53 ail 7 25
Jurinea 0 3 18 0 0 1 4
Ainsliaea 0 4 47 10 3 0 0)
Taraxacum 0 2 57 37 2 22 2
Lactuca ) 22 57 6 a if 7

"re pis 6 14 31 2 0 3 14
Youngia 0) 0 30 6 6 0 O

|
|
|

In our discussion on the area richest in Compositae we have pointed out
the climatic conditions and the floristic character of the Meridional Ranges.
The distribution maps of the large genera of Compositae of China, (Maps
3, 6, 11-15, 17-19, 21-23), all indicate that these genera actually agegre-
gate in this area. It is highly possible that the area richest in Compositae
may also be the home of many of the large genera of Chinese Compositae.
Geological, geographical, and floristic evidences seem to give support to
this proposition.

Geological evidences seem to favor the suggestion that the Meridional
Ranges are the home of montane and alpine genera such as Leontopodium,
Jurinea, Aster, Artemisia, Senecio, Ligularia, Cremanthodium, etc. This
region is geologically much older than the surrounding area. While the
present Szechuan Basin and the Tibetan Plateau were still beneath the

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 411

Tethys Sea throughout the Cretaceous, rising land and mountains were
already in existence in this region. Montane and alpine flora existed here
before the Himalayan uplift in the late Eocene. Authorities on the flora of
Tibet generally regarded the alpine elements of that region and the Hima-
layas as being derived from that of this river-gorge region (Ward, 1935,
p. 264). Here, localized glaciation may have occurred at higher elevations,
but the lower valleys, especially those nearer to the equator, have never
been touched by the Pleistocene glaciers. Thus the Meridional Ranges
have become a haven as well as a ae of many modern genera of angio-
sperms, including some Composita

Geographical evidences justify ie consideration that the Meridional
Ranges are the home of some widespread genera which occur in all im-
portant mountain ranges of the world. Map 35 represents a polar equal-
area projection of the world with the continental land-masses developed
radially from the north pole and the major mountain ranges plotted in
black. It is worthy of note that the various continental masses are tied to-
gether by more or less continuous chains of high mountains, with the center
of the tie falling in central and southeastern Asia. This phenomenon is of
profound significance in the understanding of widespread genera or species
of plants and animals. It is of special interest to us in our consideration of
the possible origin and routes of migration of genera of Compositae of
Asiatic origin. First, among all the living species of animals and plants
there is one cosmopolitan species the origin of which is somewhere in this
core and which is known to have immigrated gradually from here to all
the major land-masses of the world in prehistoric times, without the aid of
a land-bridge or continental drift. This species is Homo sapiens. Secondly,
the Meridional Ranges are also a portion of the same core of the world’s
land-masses. It is highly possible that such montane and alpine genera as
Senecio, Aster, Artemisia, Leontopodium, etc., have migrated throughout
the world in the same direction as man did, but with better speed because
of the moving forces contributed by wind, animals and, to some extent, by
water, ice, landslides and other natural mechanisms of plant dispersal.

Finally, floristic evidences testify to the antiquity and prove the original-
ity of the present flora of the Meridional Ranges and stamp the region as
the home of the genera which have their species concentration there.
Paleobotanical evidences indicate that the early flowering plants which be-
came predominant during the Cretaceous period were mostly woody plants
with generalized distributions. It is generally accepted that they lived for
millions of years under fairly moist and rather warm conditions. The
herbaceous forms represent later developments which came about with the
climatic and altitudinal changes. The mesophytic forests of the middle alti-
tudes of the Meridional Ranges and the alpine flora of the region were de-
veloped in accordance with these paleobotanical principles. The most strik-
ing fact is that the alpine flora of this region is actually situated as if it were
on an island surrounded by mesophytic forest. The majority of its com-
ponents are found only as fossils in other parts of the world. The earliest
known angiosperm is believed to be Homoxylon, known through a piece of

412 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1tx

CHAINS OF HIGH MOUNTAINS f
TYING THE CONTINENTAL Py
LAND MASSES & THEIR XY

y : al i, \, o «)
BEARING IN THE DISTRIBU reas ni a
TION OF COMPOSITAE OF ee “=?
CHINESE ORIGIN We <
\ 4 ae; . 2
(ca
\ Wea A As
Wana” a5
Map 35. Polar equal area projection of the world showing the continental
land-masses tied together by more or less continuous chains of high mountains

(dark), with the center of the core in central and southeastern Asia. (Base map
Goode’s No. 201.)

es

SON

a

ys
ie

es

SCALE
sp 1000100 Oe.
i 1000 0 3000 LOMO ER

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 413

fossil wood of Jurassic origin from the Rajmahal Hills of India (Sahni,
1932, p. 4). In the mesophytic forest of the Meridional Ranges this primi-
tive homoxylous structure reappears in Tetracentron sinense. In the same
forest at different levels grow Sassafras, Liriodendron, Cercidiphyllum,
Magnolia, Cinnamomum, Machilus, Ailanthus, Cedrela, Dipteronia, Aleur-
ites, Mallotus, Liquidambar, Paliurus, Grewia, Actinidia, Mahonia, Dios-
pyros and Zizyphus. All these are well-known fossil genera to botanists
in other parts of the world. Together with numerous species of Acer, Pop-
ulus, Salix, Viburnum, and species of Ericaceae, Menispermaceae and
Coniferae, they form the mesophytic forest of the middle-high altitudes
surrounding the alpine flora which is rich in Compositae. The antiquity
and the originality of the flora of the Meridional Ranges is indisputable.
This is true of the many Composite genera as well as of the general flora.

3, ON THE MIGRATION OF COMPOSITAE

The Compositae is one of the most ubiquitous families of the flowering
plants. In the extent of areas covered this family is surpassed only by the
Gramineae. Actually there are more genera and species in the Compositae
than in the Gramineae, but unlike the latter, members of this family do not
form continuous stretches and consequently they do not become so promi-
nent in the vegetation of any area as do the Gramineae. General ubiquity
is an expression of the possibility of wide dispersal of the disseminules of
the members of the family. Ridley in his monumental work on the Dis-
persal of Plants throughout the World has cited many species of Composi-
tae as his examples of dispersals by the action of wind, the force of water
or ice and by the transport of animals, including fish, reptiles, insects, birds,
mammals and man. There is no need to repeat his findings here. Neverthe-
less, there seems to be a necessity to supplement his summaries by pointing
out how man’s movement over the earth has accelerated the natural proc-
esses of plant dispersal, especially in connection with the migration of
Compositae.

In preparing the enumeration of the Compositae of China and in trying
to apply the findings of the enumeration to the advancement of a better
understanding of the vegetation of China, I have referred to hundreds of
articles and books. Most of them are involved in the taxonomy of the
family and its subordinate groups. Some of them are reports of botanical
explorations and lists of local floras. Not a few of them are on phytogeo-
graphical principles and the geography or geology of the country. This
variety of literature gives me the impression that there is an evident lag on
on the part of most taxonomists and many phytogeographers in giving due
emphasis to the human effect on the distribution of plants. Although it is
a well-known fact that man in five thousand years has altered the surface
of the earth more than has Nature in five hundred million years, when
phytogeographers attempt to explain the widespread or discontinuous gen-
era they rather rely on hypotheses of land-bridges, continental drift, and
nunataks, and pay no attention to what the prehistorical and historical

414 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

man might have done to bring about the present picture of natural distribu-
tion. Their writings impress the reader with their belief that the world has
been left to the operation of natural processes alone. They maintain that
various floras have developed in situ and that they have not been disturbed
until the arrival of the white man. They consider the plant specimens col-
lected by the early European explorers, and described by post-Linnaean
authors as indigenous to the particular area. They believe that those plants
have been left there since the Tertiary time. Such a concept is incorrect.

From the standpoint of the colonization of the world by the Europeans,
there is an Old World and a New World, but from the standpoint of the in-
habitation of the land surface by man, the major land-masses of the world
are all old. With the exception of a few islands where active volcanoes ex-
plode periodically, there is no land surface too young to have been inhabited
by man. It is true that where the land-masses are connected, as in the case
of Asia-Europe-Africa, man has traveled freely to and fro. It was not un-
common for people like Abraham of Ur to wander on foot with large com-
panies of herds and herdsmen from Iraq through Jordan to Cairo and then
back to get settled in Israel. Nor was it strange for a monarch such as the
Queen of Sheba to travel through the tropical heat from Ethiopia to Jeru-
salem for a visit. In prehistorical and historical time man has traveled far
and wide. Accompanying every human movement is the deliberate or acci-
dental introduction of plants. To overcome the barrier created by large
bodies of water prehistorical man made boats long before the invention of
a written language. The Neolithic culture of the tribes inhabiting the
Pacific Islands and of some of the Indians in the Americas seems to indi-
cate that these land-masses have been occupied and their vegetation dis-
turbed for from seven to ten thousand years. The migratory races and their
domesticated animals carry plants far and wide. In Asia and Polynesia the
origin of many of them is obscure. This type of introduction extends over
a long period of time. The early stages are beyond elucidation. In regard
to the distribution of Compositae, the achenes of fifteen modern species of
the family have been unearthed with the Neolithic remains in Europe. This
indicated that man has been associated with these species for a long time
and it is inevitable that he became instrumental in their migration.

There are some botanists who have an excellent range of knowledge of
plants of certain areas, but who because of their prejudice against what
they call ‘“‘closet botanists” have made some hasty conclusions concerning
the vegetation of isolated areas. Upon finding some colonies of plants
which are closely related to a geographically remote ally, they concluded
that the taxon is a relict species, implicitly ruling out the possibility that
it was introduced (accidentally) by man and that what they have inter-
preted as an ancient survivor may actually be a recent arrival. The area
that appears undisturbed to a twentieth century botanist may have been
visited many times by indigenous peoples and an isolated colony may be an
accidental introduction. In exchanging experiences of exploration in China,
an entomologist who had surveyed the Tsingling from an airplane, and
caught some ants on one of the peaks, told me that the area appeared to be

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 415

uninhabitable, but my association with the hunters and medicine diggers
and the collectors of wild edible plants leads me to a different conclusion,
It is amazing how far-reaching is the influence of the village dwellers on
the vegetation of the region, and how exact is their knowledge of where
certain things grow. For the reward of a couple of sewing needles a woman
had walked with me into the deep wilderness of a mountain of the Merid-
ional Ranges to show me a special kind of bamboo. For a few pennies a
fifteen-year-old medicine digger led me to an apparently virgin forest of
Tsuga, Picea, Betula and Acer, to collect a species of Cimicifuga which
yields certain medicine. What appears to a city dweller as vast wilderness
is to tribal people like a playground is to school children. They seem to
have visited everywhere, and know exactly where to find certain things.
This is true with tribal people everywhere, in Africa and America as well
as in Asia. Thus what appears to be relict species may be an accidental in-
troduction of tribal people.

It does not take long to change vegetation through man’s unintentional
activities. Allan in 1940 in A Handbook of the Naturalized Flora of New
Zealand reported that in less than two and a half centuries over 1000 species
of alien plants have been recorded from that country. Many of them have
firmly established themselves, and have even become dominant features in
the vegetation of New Zealand. On the basis of 500 well-established aliens,
Allan estimated that approximately 56% of them have entered the country
in seed mixtures and been further spread by sowing. About 15% were intro-
duced through adhesion to animals. Among the common aliens which have
become abundant throughout that country are these Compositae: Arctium
lappa, Cirsium arvense, C. vulgare, Chrysanthemum leucanthemum, Sen-
ecio vulgaris, S. jacobaea, Hvpochoeris radicata, Crepis capillaris, and
Sonchus arvensis. If in 200 years man can unwittingly create “a new flora
and a new vegetation” in a distant land, how much more could he have done
in intermingling the flora of the land he has occupied for many thousands
of years? The modern man has speed, whereas the prehistoric, ancient and
medieval man had time.

To illustrate man’s instrumentality both directly and indirectly in the
distribution of some Composite genera, I redraw Babcock’s diagrammatic
representation of the principal migration routes of Crepis, and superimpose
it upon a map of the ancient trade routes (Map 31). In so doing a striking
correlation between the principal routes of migration of Crepis and the
major ancient highways is revealed. It is undeniable that the dispersal of
Crepis is accelerated by man’s activities. As the species of Crepis are nor-
mally of little or no economic importance to man, one may conclude that
this distribution has been unintentional. It is very possible that the dis-
persal has been effected through the attachment of the seeds to man and
his animals.

It is customarily accepted among botanists that the fruits of Xanthium,
Arctium, Bidens, etc., which have specially modified mechanisms for ad-
hesion, are dispersed through attachment to animals, and that the fruits of
Taraxacum, Artemisia, Crepis, Senecio, etc., which are small and plumed,

416 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

are specially modified for wind dispersal. Small had demonstrated by a
specially designed wind dispersal apparatus that a light breeze of 1.97
m.p.h. is sufficient to carry the seed of Taraxacum to indefinite distances,
and that a moderate breeze of 4.4 m.p.h. can transport the seed of Leonto-
podium. Few people realize, however, that the seeds of many Compositae
which have small size, light weight and achenial hairs can be transported
more effectively through adhesion than by wind. When man and animal
travel in nature, their feet and clothes or fur are usually wet because of
dew, rain or wet ground. When the momentum of the walking feet or
running animal knocks against the partially disintegrated ripe head of
Saussurea, Senecio, Crepis, Taraxacum, Leontopodium, Artemisia, Lactuca,
etc., hundreds of small, light, hairy achenes naturally adhere to the wet sur-
face of the feet of man or to the body of an animal, and are thus carried
away from the parent plant. The distance may be long or short. The Ger-
man botanists have attributed the migration of a Chinese species, Lactuca
tatarica, to Europe through its adherence to a Steppenkuhn, Syrrhaptes
paradoxus (Hegi, 1928, p. 1133, fig. 809). This is a widespread species in
China. As man js the only species that has spread over the land surfaces
of the earth in an era when there is no known continental drift or land-
bridge, his effect on the migration of small-seeded species with disjunct or
widespread distribution, especially those with small, light seeds like the
Compositae, should be given serious consideration.

SUMMARY

The Compositae is the largest family of flowering plants in China.
It contains 167 genera and 2027 species.

A large number of the genera are small. Thirty-two per cent of the
167 genera have only one species each. Twenty-two per cent contain two
or three species each. This is due to the presence of large numbers of en-
demic genera in the interior and to the numerous introduced adventives in
the border areas.

The Composite flora of China is characterized by an unusually large
number of endemic species in the genera Aster, Leontopodium, Anaphalis,
Senecio, Ligularia, Cremanthodium, Cirsium, Saussurea, Jurinea, Ainsliaea,
Taraxacum, Lactuca, Crepis, Youngia, Pertya and Ixeris. Like Dubyaea,
Soroseris, Callistephus and other local endemics, these genera should be
considered to be Chinese in origin, Although some of them occur in China’s
neighboring countries, their numbers there are much smaller.

The distribution of the Chinese Compositae is very uneven. Many
of them concentrate at the river-gorge area on the Yunnan-Szechuan-
Sikang-Kansu borders, known also as the Meridional Ranges. The distribu-
tion of many species forms an S-shaped range over these mountains and
thence extends north-eastward or westward.

5. About 48% of the Chinese Composite genera are extra-Chinese ele-
ments. Species of these genera generally are concentrated in the bordering
areas. The bordering provinces in adjacent Central Asia and Siberia, that

1958] HU, STATISTICS OF COMPOSITAE IN CHINA 417

is, Sinkiang, Mongolia, Heilungkiang and Tibet, have a considerable num-
ber of central or western Asiatic, European or Mediterranean genera such
as Echinops, Carduus, Achillea, etc. Their distributions seldom extend
south to the Yangtze River. The Composite flora of the coastal provinces,
namely, Kwangtung, Fukien, Taiwan and Hainan contains a considerable
number of pantropic elements such as Blumea, Vernonia, Elephantopus,
Sphaeranthus, Spilanthus, etc. The distribution of these genera seldom ex-
tends north of the Yangtze River.

6. The immigration and emigration of Composite genera in and out of
China correlate fairly well with the ancient trade routes and modern water-
ways. Evidently the distribution of many widespread genera as well as
some localized adventives are associated with man’s activities. The more
widespread genera have had longer periods of association with man and the
localized adventives are recent arrivals.

The origin of the Compositae is obscure. The Meridional Ranges of
China constitute an area of origin of widespread large genera as well as of
small narrow-endemic ones. Geologic, geographic and floristic evidence all
favor this conclusion.

The degree of endemism is relatively high in Chinese Compositae.
About 30% of the genera are endemic. There is no regional limit of en-
demism at the species level. The endemic genera are concentrated largely
in the Meridional Ranges.

LITERATURE CITED

ALLAN, H. H. 1940. A eon of the naturalized flora of New Zealand. Dep.
Sci. Ind. Res. Bull. 83:
Bascock, E. B. 1947. The pa ‘Crepis I. The oes ates distribu-
tion ‘and evolution of Crepis. Univ. Calif. Publ. Bot. 21: 1-198.
BARTHOLOMEW, J. G. 1922. The Time Survey Atlas of me Wor if apes
BRETSCHNEIDER, E. 1875. Notes on Chinese Mediaeval Travellers to ae West.
6)

DALLA Torre, C. G. DE, AND H. Harms. 1900-1907. Genera Siphonogamarum.
[ Compositae, p. 523-583. ]

Cressey, G. B. 1934. China’s Geographic Foundations. xvii + 436

FERNALD, M. L. 1926. The antiquity and dispersal of vascular rid “Quart.
Rev. Biol. 1: 212-245.

. 1929. Some relationships of the floras of the northern hemisphere.

Proc. Internat. Congr. Plant Sci. Ithaca 2: 1487-1507.

. 1931, Specific segregations and identities in some floras of eastern North

America and the Old World. Rhodora 33: 25-63.

. 1931. The home of Kerria japonica. Rhodora 33: 199-200.

. 1944, ae drift and plant distribution (a review). Rhodora

46: 249-251.

—. 1950. Gray’s Manual of Botany, ed. 8. Compositae: 1357-1567.

FEDTSCHENKO, O. A. 1905. Flore du Pamir. Compositae, 108-137.

GAGNEPAIN, F. 1924. Composées. Flore Générale Indo-Chine 3: 448-663.

Goop . The taxonomy and geography of the Sino-Himalayan genus
Cremanthodium Benth. Jour. Linn. Soc. Bot. 38: 259-316.

. 1953. The Geography of the Flowering Plants, ed. 2. xiv + 452 pp.

418 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

GoopsPEED, T. H. 1936. Essays in Geobotany. xxv + 319 pp.

Gray, A. 1840. Bibliographical Notices, 7. Dr. Siebold, Flora deo sectio

Prima, Plantae ornatui vel usui inservientes, digessit Dr. J. G. Zuccarini.

Am. Jour. Sci. Arts 39: 175-176.

. 1856-1857. Statistics of the flora of the northern es States. Am.

Jour. Sci. Arts 22(2): 1-30. 1856; 23: 62-84, 369-403, 185

1884. ergs Flora of Noxth America 1(2): hoe.

1750. ont Camschatcenses Rariores. Linnaeus,
Loe

Hatentius, J. P. Amoeni-
tates een 2 33
Hecr, G. 1928. Illustrierte ae von Mittel-Europa 6(2): 394-1386.
a ak A. 1935. Historical and aaa Atlas of China, Harvard-
nching Inst. Monogr
ieee i; Ls, EP AL. 1955.
Compositae. 343 pp.
Hooker, J. D. 1881. Flora of British India 3: 219-419.
James, P. E. 1943. An Outline of Geography, xvi + 475 pp. pl. 1-2
Keck, D. D. 1946. A revision of the Artemisia vulgaris complex in North Amer-
ica. Proc. Calif. Acad. IV. 25: 421-468. fig. 1-19.
KITAMURA, 5S. 1937-55. Compositae Japonica. Mem. Coll. Sci. Kyoto Univ.
B. 13: 1-421. 1937; 15: 285-446. 1940; 16: 155-292. 1942; 22: 76-

ane Plants of the Pacific Northwest. Pt. 5

126. 1955.
sano P. 1949. Flora Zapadnoi Sibir [Florae Sibiriae Occidentalis] 11:
20649-3070.

Ler, J. S. 1939. The Geology of China. xv + 528 pp.

Lr, C. C. 1228. Kiu Ch’ang-chun’s Travels. (In chinese also see Bretschneider,
Notes on Chinese Mediaeval Travellers. 15-56.)

Li, H. L. 1952. Floristic relationships between eistetn Asia and eastern North
sED IED: Trans. Am. Phil. Soc. II. 42: 371-429. 56 maps.

: : Endemism in the ligneous flora of eastern Asia.

Congress 5: 1-5 (reprint).

LINNAEUS, C. oe Plantae Camschatcenses Rariores. Publico Examini sub-
misis Jonas P. Halenius. Amoenitates Academicae 2: 332.

LypeE, L. W. 1933. The Continent of Asia. xxl + 777 pp.

MarsH, G. P. 1874. The Earth . Modified by Human Action, a new edition of
Man and Nature. XVli + 6

OSTENFELD, C. H. ann C. ae ee 1930. The species of the genus Larix
‘a es geographical distribution. Biol. Medd. Dansk. Vid. 9(2): 1-10

. maps I-S,

oo = < 1930. The Dispersal of Plants throughout the World. xx + 744

7th Pac. Sci.

S

Pp.
Ross-Craic, S. 1954. A revision of the genus Sphaeranthus. Hook. Ic. Pl. 36:

1-90. f. 3501- 3525

SAHNI, B. 1932. Hemowilon rajmahalense, gen. A fossil angio-
spermous Ww devoid of vessel, from the Rajmahal Hills, Behar. Palaeont.
India II. on

SMALL, J. 1917- ‘en origin and ees of the Compositae. New Phytol.
16: 157- i: oe 21, 253-276. 17: 13-40, 69-94, 114-142, 200-230,
1918; 18: 1-35, 65-89, 129-176, ae 234. 1919.

se in AND J. K. JoHNsToN. 1937. oe Evolution in Compositae.

c. Roy. Soc. Edinb. 57(3): 26-

1958 | HU, STATISTICS OF COMPOSITAE IN CHINA 419

STEBBINS, G. L., Jr. 1933. A new maui of the tribe Cichorieae, Family
Compositae. Madrono 12: 65-8
. 1940. Studies in the See Mem. Torrey Bot. Club. 19: 1-76.
Warp, F. K. 1935. A sketch of the geography and botany of Tibet, being ma-
terials for a flora of that country. Jour. Linn. Soc. Bot. 50: 239-265.
— . 1937. Plant Hunter’s Paradise. maps 1-347.
Wutrr, E. V. 1950. An Introduction to Historical Plant Geography. (Transl.
by E. Brissenden.) xv + 223 pp.

420 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

A NEW SPECIES OF PHALERIA (THYMELAEACEAE)
FROM NEW GUINEA *

Liry M. Prerry

THE COLLECTIONS of the Fourth Archbold Expedition to New Guinea
include two numbers of Phaleria with flowers having an unusually broad
perianth-tube. I have searched both herbaria and literature but up to now
have found nothing to match it. For a long time it has been a recognized
fact that the species of this thymelaeaceous genus are difficult to separate.
Mr. C. T. White (1919) drew up a key to the species of Queensland, based
on the position of the flower-heads, the pubescence (or the lack of it) on
the outside of the perianth-tube, the relative length of the latter, and the
relative length of the leaves. Dr. A. C. Smith (1942) presented a key to
the Fijian species with the following introductory paragraph: “The Fijian
species are difficult to separate; one finds that such characters as leaf-tex-
ture, shape, and size, position of inflorescence, length of peduncle, size and
internal pubescence of perianth, shape of faucial scales, length of filaments,
and pubescence of ovary are variable within a species and not very depend-
able. The only specific characters which are more or less constant appear
to be the 4- or 5-merous condition of the flower, the external pubescence of
the perianth (in one species), the degree of persistence of the floral bracts
and their size, and, in some cases, the length of petioles.”

I was unable to find a key to the Malaysian species of the genus, and
perhaps a look at the synonymy given in critical works such as those of
Valeton (1913) and Hallier f. (1922) might suggest a reason. To draw up
a workable key, it would be necessary not only to make a complete study
of the material designated by the various epithets but also supplementary
collections from the type-localities with both flowers and fruits. This is not
feasible at present. However, the two collections under consideration have
flowers with perianth-tubes so much more broadly infundibular than any
which I have seen, or found indicated in the literature, that I believe this
difference to be of specific importance. Perhaps among the species described
from Malaysia, the fruits are nearest in size to those of P. macrocarpa
(Scheff.) Boerl., a species apparently not included in Hallier’s list of
Malayan representatives with critical synonymy, and, other than the orig-
inal publication as Drymispermum (1876) and the nomenclatural trans-
fer to Phaleria (1900), I have not found it mentioned in the literature. To
facilitate the comparison of P. macrocarpa with the two Brass collections,
I am indebted to Mr. J. Leandri, Muséum National d’Histoire Naturelle,
Paris, for negatives of the type and three duplicates which he has on loan
from the Bogor Herbarium. Of these, two are reproduced in the plate ac-

* Botanical Results of the Richard Archbold Expeditions.

1958] PERRY, NEW SPECIES OF PHALERIA 421

seROEE SURENSTEM, HARVARD ENICREOTT
awe of AFCA
Mir Bey Share

pibtgits Oak Beet. sie 3”

Be eanlesaoegs (aaah mot

srt 18: tomo a oe Bi ett Pagel #6 Bore St
"Sema Sesomen 158

VESOLIS ARROMETEM, MANY ANE UNIVERSITY
Finare of PAPC
Mites Mag Dearset

Ph bie shoes LY

le treet ADS 1880
deed eet ren
fae cases

Fics. a, b. Type and duplicate of Phaleria macrocarpa (Scheff.) Boerl. (coll.
Teysmann, near Doré, New Guinea). Fic. c. Type of P. elegans L. M. Perry.
Fic. d. Paratype of P. elegans.

422 JOURNAL OF THE ARNOLD ARBORETUM _ [vor. xxxrx

companying this note: the type, (fig. a@) bearing Teysmann’s field label and
another label with the name Drymispermum macrocarpum Scheff. in
Scheffer’s handwriting, consists of separate leaves and fruits (young, fide
Mr. Leandri) ; the other (fig. 6) shows the tip of a branchlet with the leaves
attached and axes of inflorescences in the lower axils. Mr. Leandri af-
firmed that Scheffer’s “pedicels” are actually very short peduncles, each
bearing many fruit-scars. The reproductions of the two Brass collections
(figs. c, d) include flower-buds, flowers and fruit. Of the three flowers at-
tached to the sheet, the uppermost is laid open to show the inside of the
flower; the leaves are elongately narrowed toward both base and apex,
whereas those of Scheffer’s species are obtuse with an abrupt acumen, and
with obviously shorter petioles. The similarity in the shape of the fruits
is probably a sectional rather than a specific character,

Phaleria elegans sp. nov.

Frutex arborescens 2—3 m. altus, sparsim ramosus; foliis magnis 18—30
cm. longis, 5.5—-9 cm. latis, lanceolatis vel oblanceolatis, utrinque angus-
tatis, apice acutis vel breviter acuminatis, basi cuneatis, glabris, crasse
membranaceis vel subcoriaceis, nervis lateralibus utrinsecus 8-12 patent-
ibus ante marginem arcuatis, venis costalibus pluribus, reticulo laxo, petiolo
1.3-1.7 cm. longo in sicco nigrescente; inflorescentiis terminalibus et axil-
laribus subsessilibus vel breviter pedunculatis, pedunculo 5—9 mm. longo
(in fructu), apice bracteis circiter 8 instructo, interioribus majoribus ovatis
vel oblongis acutis vel obtusis 2-3 cm. longis 6-14 flores involucrantibus;
floribus sessilibus extus puberulis; perianthio (post compresso sicco) late
infundibulari 3.84.2 cm. longo, ima basi 3-4 mm. juxta medium ca. 1 cm.
sub lobis 1.3 cm. lato, tubo 2.9-3.3 cm. longo, intus infra medium puberulo
superne fere glabro, lobis 5, ca. 8-9 mm. longis rotundatis intus praecipue
prope marginem puberulis; staminibus 10, filamentis perianthii lobis sub-
aequantibus; disco cupulari ca. 2 mm. alto crenulato; stylo filamentis
breviore; ovario glabro biloculari, loculo uniovulato; fructibus in sicco late
fusiformibus, 2.7-3.3 & 1.6-1.9 & 1.3-1.4 cm., utrinque acuminatis,
spermis duobus.

Papua. GOoDENOUGH IsLAND: sparsely branched tree 2 m. tall, flowers white,
fruit in dense clusters, undergrowth of an oak forest, east slopes, alt. 1750 m,,
Oct. 1953, Brass 24484 (a, type; LAE); tall shrub, sparsely branched, up to 3
m. high, leaves somewhat fleshy, petioles red, flowers and bracts cream-colored,
occasional in forest gullies, east slopes, alt. 1600 m., Oct. 1953, Brass 24883 (A,
LAE),

The characters of this species might be indicated as follows: lanceolate
or oblanceolate leaves elongately narrowed at both apex and base; flower-
cluster terminal or axillary; large ovate cream-colored bracts surrounding
the flower-cluster about to open, but missing on the cluster of fairly large
but unripe fruits; the broadly infundibular perianth pubescent outside,
with only very narrow scales in the throat.

1958 | PERRY, NEW SPECIES OF PHALERIA 423

LITERATURE CITED

Wuirte, C. T. Queensland Department of Agriculture Botany Bulletin 19: 21.
1919, (Phaleria, 19-21.)

SMITH, C. Fijian Plant Studies II. Sargentia 1: 1-148. 1942. (Phaleria,
67-73.)

ee Tu. Thymelaeaceae. Icones Bogorienses 4: pls. 368-371. 1913.

Hauer, H. Beitrage zur Kenntnis der Thymelaeaceen und ihrer natiirlichen
Umgrenzung. Mededeelingen van ’s Rijks Herbarium Leiden no. 44: 1-31.
1922. (Phaleria, 20-24.)

SCHEFFER, R. H. C. C. Enumeration des plantes de la Nouvelle-Guinée, avec
déscription des espéces nouvelles. Annales du Jardin Botanique de Buiten-
zorg 1: 1-60. 1876. (Drymispermum, 46.)

ARNOLD ARBORETUM
HARVARD UNIVERSITY

424 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

A TAXONOMIC REVISION OF PODOCARPUS, XI
THE SOUTH PACIFIC SPECIES OF SECTION PODOCARPUS,
SUBSECTION B

NettTa E. Gray *

THE SPECIES OF PopocarpPus in subsection B (2) of section Podocarpus
(Eupodocarpus of previous papers in this series) are trees and shrubs of
southeastern Asia, Australia, and the islands of the western Pacific Ocean,
the area covering almost 100° in longitude and 80° in latitude. At present
there are twenty-nine species in this group, and in some varieties have been
recognized. This subsection is about as large as subsection C (2) of South
and Central America. The extent of the southerly range is about the same
as subsection C but native specimens of subsection B are found about 10°
farther north than are those of the former. Of the twenty-nine species,
only three (Podocarpus neriifolius, P. glaucus, and P. polystachyus) have
been found both north and south of the Equator. Podocarpus neriifolius is
found, usually in mountainous regions, throughout the entire area except
Australia; P. polystachyus occurs at lower altitudes and in the coastal re-
gions of both large and small islands and also in the Malay Peninsula of
Asia; P. glaucus is of very limited distribution in the mountains of the
Philippine Islands and New Guinea.

The observations on leaf anatomy included here were made from trans-
verse sections taken from near or below the middle of the leaf. Sections
from other parts of the leaf were also made but it was not found that these
added pertinent information. Observations were most easily made from un-
stained sections dehydrated in glycerine; for permanence some sections
were mounted in glycerine jelly. Both the upper and the lower epidermis
were examined in flat mounts stained with safranin O in examples of all
species.

As in all other subsections of sect. Podocarpus, the leaves are hyposto-
matic and have a single vascular bundle flanked by winglike areas of trans-
fusion tissue. There is always at least one resin canal below the phloem;
in all species of this section except Podocarpus drouynianus and P. spinu-
losus there are three. In several species a pair of resin canals has also been
observed below the vascular bundle near, or even imbedded in, the transfu-
sion tissue (P. ridleyi |Fig. 1], P. deflexus, usually in P. salomoniensis,
rarely in P. elatus, P. neriifolius and P. archboldii var. crassiramosus).
These additional vascular resin canals have not previously been described
for this subsection of Podocarpus.

* The author wishes to express her appreciation to Prof. M. Y. Orr of the Edinburgh
Botanical Garden and to Dr. Rudolph Florin of Stockholm for fragments of critical
specimens. She is also grateful for a research grant from the Georgia Academy of

cience.

1958] GRAY, REVISION OF PODOCARPUS, XI 425

The well-developed accessory transfusion tissue (Fig. 1) extends from
the transfusion tissue proper to the margins of the leaf, between the layers
of mesophyll. The full extent of lignification is observed only in the most
mature leaves, but the horizontal orientation of the elongated cells of this
tissue may be seen in leaves which are still quite young.

A layer of hypodermis of variable pattern is found between the meso-
phyll and the epidermis of the leaves. In many species this pattern is con-
sistent and can be used as a reliable taxonomic character. The upper hypo-
dermis is a continuous layer from the midrib to the margin only in Podo-
carpus elatus (Fig. 3), P. philippinensis, P. novae-caledoniae, P. gibbsi
(Fig. 2), P. koordersii, and P. ridleyi (Fig. 1). The use of this diagnostic
character was strikingly successful in the identification of a large number
of cultivated specimens of P. elatus from many parts of the world. In P.
novae-caledoniae the hypodermis is rarely interrupted. In all the others, it
is interrupted between the margin and the midrib, sometimes being repre-
sented by only a few fibers. In some cases, the diameter of the fibers can
be relied on for specific determinations; for example, they are always large
(50 or more) in P. polystachyus (Fig. 4) and always small (20, or less) in
P. macrophyllus (Fig. 5). Lower hypodermal fibers are not found between
the stomatal rows in most species. They have been found consistently in
P. novae-caledoniae, P. philippinensis, P. ridleyi, P. elatus, P. drouynianus,
P. spinulosus, P. archboldii var. crassiramosus, P. dispermus, P. ledermannit
from New Guinea, P. sylvestris, and P. affinis. They were found occasion-
ally in P. rumphii, P. neriifolius, P. idenburgensis, P. polystachyus var.
rigidus, P. pilgeri and P. archboldii; in some of these species, only a few
specimens showed these fibers, while in others only a few fibers were found
in most of the specimens.

The presence or absence of auxiliary sclereids in the mesophyll was rarely
found to be consistent enough for use as a distinguishing character. How-
ever, they were always absent in Podocarpus philippinensis, P. glaucus, P.
brassii, P. drouynianus, P. spinulosus, P. idenburgensis, P. dispermus, P.
macrophyllus vars. maki (Fig. 5) and chingit, P. neriifolius vars. atjehensis,
degeneri, and polyanthus, P. forrestii, P. thevetiifolius, P. costalis, P. pil-
geri in the Philippines, and P. nakaii. They were present in only one speci-
men of P. elatus, two specimens each of P. rumphii and P. archboldi, and
several specimens of P. neriifolius.

The presence of vascular fibers or sclereids either below or above the
bundle was not consistent enough to be diagnostic.

The combination of characters consistent for this subsection as observed
in transverse sections of the leaves is: (1) no marginal resin canals (as
found in subsection A); (2) no resin canals above the vascular bundle (as
found in subsection F); (3) organized accessory transfusion tissue (absent
in subsection D); (4) three resin canals below the vascular bundle (with
the exception of P. drowynianus and P. spinulosus, not found in subsections
Cand D). Subsections B and F are the only subsections of sect. Podocarpus
with a pair of small, subulate bracts beneath the receptacle which bears
the ovule.

Figures 1-5. TRANSVERSE SECTIONS OF LEAVES OF Popocarpus. Fic. 1. P.
ridleyi (Wasscher) Gray [Eyma 4911], showing relative thickness of cuticle
(C), continuous upper hypodermal layer (UH) and extent of lower hypodermal
fibers (LH), location of vascular resin canals (RC) with two lateral canals em-
bedded in the transfusion tissues (TT), accessory transfusion tissue (ATT), and
auxiliary sclereids (AS) in the mesophyll. Frc. 2. P. gibbsii Gray | Clemens
32021], with upper hypoderm of alternating thick- and thin-walled fibers and
auxiliary sclereids only below the accessory transfusion tissue. Frc. 3. P. elatus

. Br. [ Laseron 435], showing upper hypoderm continuous and auxiliary sclereids
lacking. Frc. 4. P. polystachyus R. Br. [G. L. Smith s.n.], with scattered large
upper hypodermal fibers. Fic. 5. P. macrophyllus var. maki Endlicher, illustrat-
ing thickness of leaf and small, scattered upper hypodermal fibers. All figures
ca. & 20.

1958] GRAY, REVISION OF PODOCARPUS, XI 427

Within subsection B, anatomical characters of the leaves alone are not
sufficient to determine the species. Nor have leaf-size and -shape, terminal
bud shape, buds or cones of male strobili, length of peduncle, or size and
shape of seed, proved to be reliable when used individually. Combinations
of some or all of these characters are needed to differentiate the various
species. Because so many of the specimens encountered are sterile, vege-
tative characters have been used insofar as possible in the preparation of
the following key.

Key to Sect. Popocarpus (SouTH PAcIFIC SPECIES), SUBSECTION B

A. Leaves with continuous upper hypoderm (Figs. 1, 2, 3
Ba eeaves rarely Over S.-C. 1Ong; -BOMe0 3 ory, 354 ee ake ei0 0st,
BB. Leaves over 5 cm. long
C. Leaves not more than 5 mm. wide, the midrib A or concave above;
INewmeCaledoiia oars os weer oo an tien, Bence: . P. novae-caledoniae.
CC. Leaves over 5 mm. wide, the midrib ae above.
. Leaves without fone hypodermal fibers between oe eae

Ows: Java c& Welebesqmene mete perce sss 3. P. koordersit.

DD. Leaves aa lower hypodermal fibers between the est rows
Ciigss 1.3).

E. Leaves 13-23 cm. long; terminal buds globose; Philippine

Tslands. --2:o eevee eee eee cy 4. P. philippinensis.

EE. Leaves usually less than 15 cm. long; terminal buds ovate.
F. Apex of leaves usually long-acute or -acuminate; leaves
with 3-5 vascular resin ainda upper hypodermal fibers
large; Malay Peninsula & Borneo........... 5. P. ridleyt.
. Apex of leaves obtuse, often Sa eere leaves with 1-3
vascular resin canals, hypodermal fibers not over 22 » in
diameters oAlictial Weve seat ra oe 4 6. P. elatus.
AA. Leaves with interrupted upper hypoderm (/*igs. 4, 5).
G. Leaves not over 6.5 mm. broad.
H. Leaves not over 2.5 cm. long, rarely over 5 mm. wide.
I. Terminal buds large, globose; pollen cones thick; New Guinea,
EN OLORIA eR O08 Ino emer rt Ret t= ead ee ae ee elo she OP OSstt,
II. Terminal buds small; leaves eee when young; pollen cones
slender; Mindoro & New Guinea 8. P. glauc
HH. Leaves over 2.5 cm. lon
J. Leaves not over 5.5 cm. long, erect and adpressed to twig;
Philippines; Borneo Gina we ferent eo: . P. brevifolius.
JJ. Leaves usually over 5.5 cm. lon
K. Leaves pungent; seed eae beaked; southeastern Australia.
eae 1h ee vs . Spinulosus.
KK. Leaves rarely pungent; seed with rounded a
L. Leaves 5-12 cm. long, sessile, sie. vith 1 vascular
resin canal; western Australia. ..... 11. P. drouynianus.
LL. Leaves 6-— 9 cm. mae petiolate, with 3 vascular resin
Canals; New sCaled OMmidamrewne scree sic okra a os
Fc teen ee 2a. P. novae-caledoniae var. colliculatus.

br
|

428 JOURNAL OF THE ARNOLD ARBORETUM _ [VOoL. XxxIx
GG. Leaves ws least some of them) over 6.5 mm. broad.
s with lower hypodermal fibers ee the stomatal rows.
N. ae mostly less than 10 cm. lon
O. Leaves less than : 2 cm. long; ijl ppiueean glee . tins
OO. Leaves usually over 5.2 cm. long
P. Midrib of eae — below: Borneo. ...........
beh ee one cae eed 8a. P. polystachyus var. rigidus.
PP. Midrib of leaves not pea below.
Q. Twigs very stout; leaves with acute or acuminate tips;
New Guinea. 20a. P. archboldit var. crassiramosus.
QQ. Twigs slender; leaves with rounded oe e apex; New
CRPUON Ie. 4 45-2 annie RSs Roaned ee . P. sylvestris.

NN. Leaves over 8 cm. oe
R. Leaves 9-15 mm. pn New Guinea. . 14. P. idenburgensis.
RR. Leaves 17-30 mm
S. Seeds solitary, re mm. long; New wees ihe Beas
Lith hd Ot ae ee ee ae eee Pz ledermannii,
SS. Seeds usually in pairs, 2.5 cm. long; ssa,
P. dispermiis.

16
MM. Leaves mostly without lower hypodermal fibers between the stomatal
rows.
T. Leaves with auxiliary sclereids.
U. Leaves narrow, 6.5-9 mm. wide; Solomon Islands. ........
17. P. salomoniensis.

UU. Leaves wider, over

. Leaves strongly anced with 5 vascular oe canals;
Malay Peninsula. .................... 18. P. deflexus.
VV. Leaves not deflexed, with 3 vascular resin ena Annam,
Indochina, Hainan. 19. P. annamiensis.

TT. Leaves usually without auxiliary ‘sclereids,
W. Leaves for the most part deflexed; Sumatra. ............
eee 278. £, neriifolius var. atjehensis.
Leaves erect, “spreading or divaricate.
X. Terminal buds globose or ovate,
Y. Margins of leaves revolute; New ney
20 archboldii,

nf

WwW.

Yi. seein of leaves not revolute.
s 8-11 mm. broad; ean 21. P. nakait.
ce ae mostly more than 10 mm broa
a. Leaves with parallel margins, the sex usually
short-angustate; Borneo, New ges etc
Seay ee foie a
aa. Leaves lanceolate, aie apex eee acuminate;
Sumatra, etc. Wye gp ante ease sees wees
eee 27d. P. neriifolin var. teysmannii.
XX. Terminal buds ovate, conical,
. Upper midrib of leaf re | ar or flat.
c. Peduncles of female cones 2 mm. long; Philippines,
Formosa, shores. ........... 23:2. Costaliy,
cc. Peduncles of fewale-« cones over 2 mm. :
d. Upper hypoderm of large aes New Guin
Be ae Neneh olie.

1958 | GRAY, REVISION OF PODOCARPUS, XI 429

dd. Upper hypoderm of small sae China.
ee en a earl Sroka! sora eT . P. forrestit,
bb. Upper midrib of leaf ae seer
Male cones usually solit
f. Leaves usually less ave 8 cm. long, apex usually
obtuse, apiculate; Philippine o ee Is-
TAGS? tc Sey at once tee ae eee P. pilgeri.
ff. Leaves usually 8 cm. long or more, a apex usu-
ally acute to acuminate; entire ee except Aus-
ay sre me UN ae . P. nertifolius.
g. Leaves usually less than 10 mm. wide; Fiji.
PATA ORES ee eee var. doeeiere
gg. Leaves ie over 10 mm. wide; female
cones numerous all over youngest growth;
Sumatra & New Guinea. ................
ae a neriifolin var. polyanthus.
Male ¢ cones usually cluster
h. Leaf with upper hy eas on ae large fibers;
Philippines to Bangka. ... . P. polystachys
bh. ms with upper hypoderm 2 ‘small se
. Leaves 9-11 mm. broad; China -
29, B er line
ii. Leaves less than 9 mm. broad.
j. Leaves more than 4.5 mm. broad.
k. Leaves 5.5-12 cm. long ~ a mm.
broad; China, Japan & Bu
29a, P. pea var. Be ae
kk. Leaves 3.5-7 cm. long & 4.5-8 mm.
wide; China, tan Formosa & Burma.
29c. P. macrophyllus var. maki.
sie leaves: less than 4.5 mm. broad;
29b. P. macrophyllus var. chingii.

ia")
fe")

1. Podocarpus gibbsii, sp. nov.

Arbor 10—20 m. alta, ad 20 cm. diametro; ramulis erectis; alabastris per-
minutis ovatis, 1.5 mm. diametro, squamis paucibus exterioribus lanceo-
latis, carinatis, rectis, ad 6 mm. longis, apiculatis longi-acuminatis, ali-
arene foliaceis; foliis 1.3—2.1 cm. longis, 4-6 mm. latis, spathulatis, inter-
dum ellipticis vel lanceolatis, coriaceis, planis, apice obtusis vel acutis, basi
in petiolum brevem crassum angustatis; costa supra manifesta, subtus plana
et inconspicue; strobilis masculis axillaribus, sessilibus, solitartis vel duobus,
squamis exterioribus confertim imbricatis, deltoideis, apice longe acumi-
natis; strobilis femineis et seminibus ignotis.

DIsTRIBUTION: Edges of low jungle, 1600-2300 m. altitude, Mt. Kina-
balu, British North Borneo.
British North Borneo. Marai Parai, near camp, Clemens 32021 (TYPE, A; tBM,

7 This symbol preceding the abbreviation of an herbarium ee that the details
of the leaves of this specimen have been examined in transverse section

430 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. XxxIx

+ny, tuc);* Marai Parai spur, Gibbs 4092 (+3m); Colombon Basin, Clemens
40001 or tny, tuc).

This small tree of apparently very limited range may be described in
somewhat further detail. The inner scales of the very small vegetative
terminal buds are thinner than the outer ones and have merely acuminate
tips. The extreme margins of the short, spatulate leaves are slightly re-
curved and the sharply prominent midrib above becomes flat toward the
apex. Juvenile leaves may be up to 5.2 cm. long and 9 mm. wide. Trans-
verse sections of the leaves (Fig. 2) show three vascular resin canals,

midrib (some fibers with large lumen), lower hypodermal fibers absent
between stomatal rows, vascular sclereids abundant above and rare below
the bundle, the single layer of palisade mesophyll composed of large cells
often almost as broad as long, mesophyll cell walls, expecially the lower,
lignified but not pitted and thickened as much as is usual for most aux-
iliary sclereids

Podocarpus gibbsit differs from the other species in having continuous
upper hypoderm in its very short, broad, spatulate leaves. The two Clemens
specimens were used by Wasscher (11) for his discription of Podocarpus
glaucus but these differ from Foxworthy’s description of that species in
that the young foliage is not glaucous and, in addition, the leaves have a
continuous upper hypoderm. Only Clemens 40001 is fertile, bearing male
cone buds and immature cones. I do not understand the description “tiny
yellow pistils” written on the label of this specimen, for the species is
dioecious and I cannot find any structures to which the note might apply.
The upper hypodermal fibers in the leaves of the Clemens specimens show
clearly the alternation of groups of fibers with open and closed lumina.
This was erroneously described by Orr (9) for Podocarpus glaucus. Orr
did not observe the lignification of cell walls in the mesophyll, particularly
the lower, where an almost continuous layer is formed. This is very
apparent in cleared and stained whole mounts of leaves. Gibbs 4092 has
more scattered, larger, acute-tipped leaves which are a normal juvenile
variation in most podocarps.

This species is named in honor of Dr. L. S. Gibbs, who first collected
it on Mt. Kinabalu.

* The following symbols indicate the location of the specimens cited: Academy of
Natural Sciences of Philadelphia (pH); Arnold Arboretum (A); Bernice P. Bishop
Museum, Honolulu (sisu); Brisbane Herbarium of Agriculture (srr) ; British Museum
of Natural History (pm); Brussels Botanical Garden (Br); University of California
at Berkeley (uc) and at Los Angeles (ucLA); California Academy of Science (CAs) ;
Chicago Natural History Museum (Field Museum) (Ff); Cornell University (cv) ;

Herbarium (Gu); University of Ilinois pee (iLL) ; Royal Botanic Gardens, Kew
(ax); McGill College Herbarium, Montreal (mrmc) ; Missouri Botanical Garden (Mo) ;
Department of Forests of New Guinea and Papua (LAE); New Jersey College for
Women (Rutgers) (rut); New York Botanical Garden (Nv); Paris Muséum National
d’Histoire Naturelle (Pp); Pennsylvania University Herbarium (PENN); United States
National Herbarium (us); Victoria National Herbarium, Melbourne (MEL).

1958 | GRAY, REVISION OF PODOCARPUS, XI 431

Stapf, when determining and describing the specimens for Gibbs (6),
described but did not identify two specimens of Podocarpus (Gibbs 4089,
4092). Gibbs 4092 (sm), which I examined, resembles Stapf’s description
of Gibbs 4089. Either the numbers were reversed in attaching the labels
to the specimens or Stapf reversed the descriptions. Stapf’s description of
Gibbs 4092, “a very striking species with fairly crowded leaves, oblong-
linear, obtuse or sub-obtuse at the apex, attenuated at the base, 1.5-2 cm.
long, 3.5-4.5 mm. broad, with the margins recurved, the midrib raised
above and rather broad and flat beneath,” shows that this plant is like
the two Clemens specimens and is, therefore, Podocarpus gibbsu, the new
species. The other specimen, Gibbs 4089, is said to have “leaves, in shape
and size, intermediate between those of P. brevifolius and P. polystachyus
but thinner and much more loosely arranged than either.” This is like the
British Museum’s specimen of Gibbs 4092 and, therefore, is probably also
P. gibbsii. Wasscher (11, p. 471) lists these specimens under doubtful
species. His suggestions that Gibbs 4089 is P. pilgeri and Gibbs 4092
is P. glaucus are untenable.

Lee)

Podocarpus novae-caledoniae Vieillard ex Brongn. & Gris, Bull.
Soc. Bot, France 13: 425. 1866; Parlatore, DC. Prodr. 16: 513. 1868,
Pilger, Pflanzenreich IV. 5(Heft 18): 76. 1903, Nat. Pflanzenfam.
ed. 2. 13: 248. 1926; Compton, Jour. Linn. Soc. Bot. 45: 425. 1922;
Dallimore & Jackson, Handb. Conif. 53. 1923, 1931, 78. 1948; Florin,
Svenska Vet.-Akad. Handl. III. 10: 279. 1931.

P. ensifolia Carriére, Traité Conif. ed. 2. 655. 1867 (non R. Bi:

A shrub (seldom a small tree up to 3 m. high) with erect crowded
branches. Terminal buds ovate, 2-3 mm., the long triangular scales with
acuminate apex. Leaves linear, crowded, more or less erect, leathery,
4-9 cm. long, 3-5 mm. broad, the apex short-acuminate or rarely long-
attenuate and sometimes pungent, gradually narrowing at the base to a
short obscure petiole; margins blunt with sharply recurved edges but not
revolute; midrib shallowly concave or flat above. Transverse sections of
the leaves with three vascular resin canals (rarely only one); upper
hypoderm usually in a single continuous layer except at the margin and
midrib where it is doubled, lower hypodermal fibers not abundant be-
tween stomatal rows, fibers averaging 20-30 » in diameter; vascular fibers
few above but numerous between the vascular bundle and central resin
canal; auxiliary sclereids in the mesophyll so abundant as to form a dis-
tinct layer between the palisade mesophyll and accessory transfusion tis-
sue; cuticle thick. Male cones axillary, in pairs, sessile, surrounded at the
base by stiff obtuse scales. Microsporophylls densely imbricate, each with
a short rotundate apiculus. Female cones axillary and solitary on peduncles
7-10 mm. long, the fleshy receptacle subtended by two minute scales and
bearing 1—2 ovules. Seeds ellipsoid, shiny, 7-8 mm. long, 4 mm. wide.

DIsTRIBUTION: Stream banks, at low altitudes, New Caledonia.

432 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

New Caledonia. River Dumbea, Vieillard 1266 (IsoTyPE, A, BM, P),
Compton 402, 419 eee Buckhols 1793, 2137,. 1199, [255,347 a. 1759
(+1LL), Franc 96 (BM, NY, UC), Anon. s.n. (nu), Pancher (P), Balansa 189 (P),
Godefroy 552 (Pr), LeRat 603, 1587 (Pe); Rio des Pirogues, White 2231 (a-2
sheets, BRI, DS, K, P, UC-2 2 sheets, US), Moe 1340 (ILL); Riviére Camboui,
Compan 2017, 2169 Ce Couvélée, Franc 2408 (A-2 sheets, BRI, +F, K, NY, P,
uc, us); Table Unio, Lecard s.n. (Ba, P), a 92 (Pp); Port Bouquet, Balansa
704 (Ny), 2504 (x, P); Bourail, Pennel sn. (p); Mt. Mou, Franc 107 (x),
Glandoger in 1906 (+mo); Mt. Dore, Vieillard & Pancher 396 (Pp); Mt. Koghi,
LeRat 2372 (ep); Rive gauche de la Toutouta, Virot 369 (a); Prony, Franc 94A
(BM), 96A (BM, K—2 sheets, US), 5.2. (BM), Franc in 1913 (A-7 sheets, F, GH,
NY, UC—2 sheets), MacDaniels 2540 (tcv), Juillet 96 (BRI, US). iapalieesn
known: Pancher ex Hennecart in 1879 (kK, NY), Pancher in 1870 (+R, nr
sheets, tUCLA), Franc 90 (tuc), Baudouin 635 (P)

Podocarpus novae-caledoniae is a low, rounded shrub rarely forming a
small tree when growing in a thicket with other vegetation. The young
leaves are glaucous, becoming dark green when older (3). According to
Buchholz (ms.), there seem to be differences in the thickness of the leaves,
as specimens found in Plaine des Lacs and near Prony have leaves more
coriaceous than those found in the Dumbea valley. This is not apparent
in transverse sections of the leaves. Compton (3) described the fleshy
receptacle as becoming bright scarlet or purple, soft, translucent, and
sweet to the taste, whereas Buchholz (ms.) says that it turns yellow or
brick red. These variations may merely be due to different stages of
maturity of the fruit.

Orr (9) examined the leaf anatomy of Podocarpus novae-caledoniae and
described most of the above characters. He does not seem to have ob-
served the lower hypodermal fibers between the stomatal rows since he
did not include this species in his list of those showing this character.

2a. Podocarpus novae-caledoniae var. colliculatus, var. nov.

Frutex vel arbuscula, 1 m. alta; foliis coriaceis, linearibus, saepe falcatis,
6—9 cm. longis, 5—6.5 mm. latis, apice coarctis acutis vel rare acuminatis,
basi in petiolum 1-4 mm. ieeun sensim angustatis; strobilis maseulis
alabastris immaturis, globosis, cum squamis brevibus adpressis; strobilis
femineis seminibusané ignotis.

DistTR1BuTION: Isle of Pines, New Caledonia, at 80 m. altitude, and
possibly on New Caledonia itself.

New Caledonia. Germain in 1874-76 (+p). IsLE or Pines: Pic N’Ga, Virot
1053 (Type, +ILL; NOUMEA),

The description of Podocarpus novae-caledoniae var. colliculatus was
prepared only from Virot 1053. I am in doubt as to whether the Germain
specimen really came from New Caledonia. The terminal buds of the
twigs of this variety have short, thin, appressed scales very similar to those
of the species. The midrib of the leaf is flat or broadly rounded above and
broadly prominent or ridged below. Transverse sections of the leaves of

1958 | GRAY, REVISION OF PODOCARPUS, XI 433

both collections cited above are very similar. These show three vascular
resin canals, an interrupted upper hypoderm, rare lower hypodermal fibers,
vascular sclereids above and below the bundle and auxiliary sclereids in
the mesophyll. That the upper hypoderm is occasionally interrupted in-
stead of being a continuous layer is in contrast to the main species. The
low, shrubby habit is in agreement but the broader leaves with a definite
petiole and the occasionally interrupted upper hypoderm differentiate the
variety. In these latter characters it is like P. sylvestris but this is a tree
and the leaves are mostly broader. The vegetative buds of all three of
these taxa are similar.

3. Podocarpus koordersii Pilger ex Koorders & Valeton, Meded.
Lands Plant. [Batavia] 68: 268. 1904; Nat. Pflanzenfam. ed. 2.
13: 248. 1926; Florin, Svenska Vet.-Akad. Handl. IIT. 10: 279. 1931;
Wasscher, Blumea 4: 433. 1941.

A small tree, to 14 m. tall, with stout twigs. Vegetative buds globose
with thick, broad, obtuse or roundish scales, the outer ones sometimes
acuminate. Leaves spreading, thickly coriaceous, rigid, glossy above,
straight or somewhat falcate, linear or linear-lanceolate, 9-21 cm. long,
6-13 mm. wide (larger on sterile twigs), gradually narrowing to a short
thick petiole and acute apex; midrib broad and usually roundly prom-
inent on both surfaces. Transverse sections of the leaves with 3 vascular
resin canals below the bundle, with vascular fibers always above and some-
times below, isolated auxiliary sclereids in the mesophyll both above and
below the accessory transfusion tissue, upper hypoderm fibers in a con-
tinuous layer, doubled at the margin and midrib, absent below between
the stomatal rows, small, 16-24 » in diameter. Male cones axillary, sessile,
in fascicles of 2-8, the buds globose with thick rounded scales; mature
cones 4.5 cm. long, 2 mm. in diameter. Microsporophylls broadly trian-
gular, each with a small apiculus. Female cones unknown.

DIsTRIBUTION: From sea level to 50 m. in Java and up to 300 m. in
Celebes.

Java. Koorders 39480 (Type, +via Florin ex Leiden Herbarium); Anon. (tvia
Florin ex Berlin Herbarium); Banjumas, Koorders 40251 B (+a); Horsfield s.n.
(+pm), Horsfield in’ 1859 (+MEL). Celebes. Malili, Oesoe, Cel/II-285 (A),
Cel/II-287 (+prt), Cel/IJ-288 (t+p1sH). San Cristobal. BSIP on ridge, Logie
BSIP 356 (+LAE).

Podocarpus koordersii differs from P. nertifolius in having leaves with
margins parallel over a greater length, a broader, less prominent midrib
and a continuous upper hypoderm (as seen in transverse sections). The
specimens from Java all have rounded or obtuse vegetative bud scales
but those from Celebes have acuminate tips on the outer bud scales. In
most of the leaves of P. koordersii the upper hypodermal fibers show alter-
nate groups of thick-walled cells with small lumina and thinner-walled
cells with large lumina. This is especially apparent in the specimens from

434 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

Celebes, in which I find also that the mesophyll above and below the
accessory transfusion tissue is an almost complete layer of thin-walled
sclerenchymatous cells with simple pits, very large and empty of cell con-
tents. Wasscher (11) referred to P. rumphii a group of specimens from
Celebes with related collection numbers, including those I have examined,
but he says that they have very small, more lanceolate leaves. This and the
continuous upper hypoderm readily refer them to P. koordersii, Orr (9)
also examined P. rumphii and agrees that this species has interrupted upper
hypoderm while in P. koordersi it is continuous.

4. Podocarpus philippinensis Foxworthy, Philip. Jour. Sci. Bot. 6:
163. 1911; Pilger, Nat. Pflanzenfam. ed. 2. 13: 248. 1926; Florin,
Svenska Vet.-Akad. Handl. III. 10: 280. 1931.

Large, straight trees, 20-30 m. tall, with smooth reddish-brown flaky
bark and erect branchlets. Terminal buds globose with thick ovate scales,
the tips rather acute but sometimes very blunt. Leaves erect or erect-
spreading, straight to subfalcate, linear-lanceolate, 13-23 cm. long,
9-18 mm. wide (juvenile to 31 cm. long, 25 mm. wide), the apex acute to
acuminate, gradually narrowing at the base to a short petiole; midrib
mostly flat or somewhat depressed, broadly prominent below. Transverse
sections of the leaves show three vascular resin canals, vascular sclereids
or fibers rare above and absent below the bundle, no auxiliary sclereids,
upper hypoderm. in a continuous single layer except at the margin and
midrib where it is doubled, lower hypodermal fibers abundant between
stomatal rows, the small fibers 15-22 » in diameter. Male cones unknown.
Female cones axillary, on peduncles 5—8 mm. long; receptacle fleshy, sub-
tended by two small bracts, bright red. Seeds ellipsoid-globose, 12-15 mm.
long, 11 mm. wide, the crest suppressed.

DIsTRIBUTION: On mountain slopes in the Philippine and Selangan
Islands.

Philippine Islands. Luzon. Bataan Prov.: Lamao Forest Preserve, For. Bur.
6326, Curran (BISH, BRI, (MO, NY); For. Bur. 8987, Curran (+mo, Ny); For. Bur.
17523, Curran (+BR), For. Bur. 17592, Curran (Ny), For. Bur. 17594, Curran
(+pm); Mt. Mariveles, For. Bur. 2743, Borden (Ny). Pampanga Prov.: Mt.
Arayat, For. Bur. 17664, Curran (tNny, tp), Merrill 3917 (+ny). Tabayas Prov.:
Merrill 1992 (mo); Alikad, Jones 63 (+¥F-2 sheets). Locality unknown: Anon.
sn. (+uc). Selangan Islands. Cons. For. N. Borneo s.n. (+E).

Podocarpus philippinensis was distinguished by Foxworthy from P.
rumphii by the smooth reddish-brown flaky bark, the leaves narrowing
more gradually toward the petiole and the seeds with shorter peduncles.
With the exception of Jones 63 and Merrill 1992, all the collections listed
by Foxworthy were examined and all are either female or sterile. Fox-
worthy (5) left under P. rumphii several specimens, one of which (Merrill
3917) is certainly P. philippinensis, for an examination of a transverse sec-
tion of a leaf showed the same continuous upper hypoderm characteristic

1958 | GRAY, REVISION OF PODOCARPUS, XI 435

of this species. I have not seen the other specimens, but they are prob-
ably also P. philippinensis as the larger leaves and longer peduncles which
they may show seem to be natural variations in most species of Podocarpus.
Wasscher (11), who does not recognize P. philippinensis, saw many of the
same collections and included them under P. rumphii without comment.
The upper hypoderm is so consistently continuous in Podocarpus philip-
pinensis that his treatment of these specimens is not tenable.

Orr (9) agrees that the leaves of Podocarpus philippinensis differ from
those of P. rumphii by the continuous hypoderm, an evident but not very
prominent upper midrib, the absence of sclereids in the mesophyll, and no
auxiliary sclereids. This description of the leaf anatomy resembles that
of P. elatus, an Australian species. It differs from P. koordersii only in
the lower hypodermal fibers between the stomatal rows and the absence
of auxiliary sclereids. I do not find that the hypodermal fibers are as large
compared with those of P. koordersii as described by Orr (9).

Orolfo 3919, from the Selangan Islands, agrees with Podocarpus philip-
pinensis. It had been identified by Orr (9) as P. teysmanni but, on ex-
amination of authentic material of this latter species, I find that it is quite
different.

5. Podocarpus ridleyi (Wasscher), stat. nov.
Podocarpus neriifolius var. y ridleyi Wasscher, Blumea 4: 453. 1941.

A small tree with stout branches but small verticillate twigs. Terminal
buds ovate with long-acuminate outer scales up to 13 mm. long. Leaves
crowded on new growth, spreading, stiffly coriaceous, linear-lanceolate,
Ae ag or falcate, 10-15 cm. long. 9-12 mm. wide (rarely to 17 cm. long,
19 mm. wide), narrowing at the base to a short petiole, gradually narrow-
ing to. an acute or acuminate apex; midrib flat or broadly prominent
above, more prominent below. Transverse sections of the leaves show 3-5
vascular resin canals, the two outermost often located within the transfu-
sion tissue on either side of the bundle, vascular fibers above the bundle
and usually below, upper hypodermal fibers in a continuous layer, usually
large, 22-50 » in diameter, lower hypodermal fibers also large and scat-
tered or abundant between the stomatal rows, auxiliary sclereids often
abundant in the palisade layer and mesophyll both above and below the
accessory transfusion tissue. Male cones axillary, sessile or on peduncles
up to 2 mm. long, fascicled 1-3, buds small, ovate, the outer scales keeled,
acute, 2-4 mm. long, with erose scarious margins; mature cones large,
evlmdnct up to 8 cm. long, 4 mm. wide. Female cones solitary, axillary;
peduncles to 9 mm. long; immature receptacle subtended by two subulate
bracts 2.5 mm. long. Mature seed not seen.

DISTRIBUTION: On mountains, up to 1000 m. altitude, Malay Penin-
sula, Borneo, and New Guinea.

Malay Peninsula: Johore, Gunong Blumut, Holttwm 10720 (TypE, ta);
Malacca, Padang Bata, Ridley 10016 (A). Borneo: western part, G. Semedoen,

436 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

Hallier 720 (+Ny, uc). Netherlands New Guinea: Seroei, Japen, Boschpr.
bb30484, (+A), 0630650 (+A); Japen, Soemberbaba near Seroi, Van Dijk 830
(+A); Wissel Lake region, environs of Post, on foot of Mt. Boebeiro, yma 4911
(ta). Cultivated. Java: Buitenzorg Botanic Garden, V.F.31 (+mo-3 sheets).

Podocarpus ridleyi is a common tree on Gunong Blumut and, accord-
ing to Holttum in Wasscher (11), it is a very striking plant with its yellow-
green leaves. Leaves of these specimens are superficially similar to those
of P. koordersii, P. philippinensis, P. rumphii and P. neriifolius, but can
be distinguished from each of these species by the anatomy of the leaf as
observed in transverse sections (Fig. 1). The most striking feature is the
presence of five vascular resin canals in the leaves of all specimens except
Hallier 720, Five vascular resin canals have been found elsewhere only in
P. deflexus (Malay Peninsula), P. salomoniensis (Solomon Islands), some
specimens of P. neriifolius and P. decipiens (Fiji Islands), and P. ulu-
gurensis (Africa). Only the first three of these are in this subsection of
sect. Podocarpus. The continuous upper hypoderm is similar to that of
P. koordersti and P. philippinensis, but P. koordersii does not have lower
hypodermal fibers between the stomatal rows and P. philippinensis does
not have auxiliary sclereids; the hypodermal fibers are larger in P. rid-
leyi than in either of these species. In most of the specimens of P. ridleyi
the upper hypoderm consists of alternating groups of thick- and thin-walled
fibers as has been described for some specimens of P. koordersii.

The specimens of Podocarpus ridleyi have had various identifications.
Both Holttum 10720 and Hallier 720 have been identified with P. neri-
folius and Wasscher lists them in his new variety ridleyi under this species.
Both the continuous upper hypoderm and hypodermal fibers between the
stomatal rows differ from P. neriifolius and the specimens are an entirely
distinct species. Two of the specimens from Java had been named as
P. macrophyllus, a species having smaller leaves with interrupted upper
hypoderm of very small fibers and a very prominent midrib, The other
specimen was called P. rumphii which I have found to be a species also
with interrupted upper hypoderm.

In Hallier 720, from Borneo, the upper hypodermal fibers are smaller
in diameter, usually not over 18 », and only three vascular resin canals
have been observed, Rare interruptions were found in the upper hypoderm
of Holttum 10720. There seem to be no vascular fibers below the bundle
in the three cultivated specimens (two male, one sterile) from Buitenzorg
collected by Sargent in 1903.

6. Podocarpus elatus R. Brown ex Mirb. Geogr. Conif. in Mém. Mus.
13: 75. 1825 (nomen!); Endlicher, Syn. Conif. 213. 1847; Gordon,
Pinetum, ed. 1. 273. 1858, ed. 2. 334. 1875; Carriére, Traité Conif.
656. 1867; Parlatore in DC. Prodr. 16: 517. 1868; Bentham, Fl. Aus-
tral. 6: 246. 1873; Van Tieghem, Bull. Soc. Bot. France 38: 169. 1891;
Hemsley, Jour. Linn. Soc. Bot. 30: 199. 1894; Bailey, Queensl. Fl. 5:
498. 1902; Pilger, Pflanzenreich IV. 5(Heft 18): 75. 1903, Nat.

1958] GRAY, REVISION OF PODOCARPUS, XI 437

Pflanzenfam. ed. 2. 13: 248. 1926; Gibbs, Jour. Linn. Soc. Bot. 39:
183. 1909; Baker & Smith, Res. Pines Austr. 435. 1910; Dallimore &
Jackson, Handb. Conif. 43. 1923, 1931, 66. 1948; Florin, Svenska
Vet.-Akad. Hand]. III. 10: 279. 1931.

P. ensifolia R. Br. ex Gordon, Pinetum ed. 1. 275. 1858; ed. 2. 335.1875
(non Carr.).

P. acicularis Van Houtte ex Gordon, Pinetum ed. 1. 275. 1858.

P. falcata A. Cunn. Herb. fide Baker & Smith, Res. Pines Austr. 435. 1910.

Nageia elata F. Mull. Cens. 109. 1882

An erect tree up to 30 m. high with solitary or verticillate branches
bearing spreading, rather lax leaves. Vegetative buds small, ovate or sub-
globose with short, more or less sharply acute scales. Leaves light green
and shiny above, coriaceous, straight or subfalcate, linear-lanceolate, the
acute to obtuse apex often mucronate, gradually tapering at the base to a
short petiole, 5-11 cm. long, 7.5—10 mm. wide (occasionally up to 20 cm.
long, 16 mm. wide). Midrib broad and slightly prominent above, scarcely
evident below; leaf margins sometimes slightly thickened, never revolute.
Transverse sections of the leaves (Fig. 3) usually show three vascular resin
canals (only one resin canal in the leaf examined of the type, Brown 3117,
one to three in Laseron 435, one in Anon. 10545 from Mt. Spergeon, and
one to three in two cultivated specimens from Rome and California) ,
vascular fibers rare above the bundle but usually abundant below, auxiliary
sclereids absent in the mesophyll, palisade mesophyll thick but of a single
Jayer of cells, upper hypoderm in a neat single continuous layer except
at the margin and midrib where it is doubled, lower hypodermal fibers
abundant between stomatal rows, both upper and lower fibers averag-
ing 15-22 » in diameter. Pollen cones axillary and sessile, in groups of
2 or 3, 5 cm. long, 4-5 mm. wide, surrounded at the base by numerous
thick broad obtuse scales. Microsporophylls densely imbricate, broadly
triangular with a short apiculus. Female cones solitary in the axils of
the leaves, on stout peduncles 3-10 mm. long; receptacle of two fused
thickened equal bluish scales scarcely free at the tips and subtended by
a pair of small subulate bracts. Seeds globose, 12-14 mm. wide (22 mm.
long on Mt. Spergeon), with very blunt inconspicuous crest.

DisTRIBUTION: Native to southeastern Australia (Queensland, New
South Wales) and Tasmania. Cultivated as an ornamental around the
world.

Australia. QUEENSLAND: Cape York Peninsula, 100 mi. south of Cape York,
Hartmann in 1886 (+MEL); Fitzroy Island, Moore 39 (+meL); Rockingham Bay,
Mueller in 1864 (2 sheets, 6 and °, tmeL); Barron River, Kenevunga, Cowley
52 (BRI); Mt. Spergeon, Anon. 10545 (+Brr); South Kennedy Dist., Dalrymple
Heights, Sclemers in 1947 (prt); Bundaberg, Keep in 1898 (BRI); Gympie
Dist., Kenny in 1907 (srr); Beech Mountain, White in 1923 (srt, +BR), Wilson
in 1921 (+A, seedling); Brisbane River, Ithaca Creek, Bailey in 1875 (BRI),
Moreton Bay, Mueller in 1855 (MEL); Pioneer River, Griffiths in 1889 (MEL).
NEw SoutH WateEs: Moore Park near Kyogle, White 12579 (mo, uc); Rich-

438 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xXxxIx

mond River, Henderson 106 (MeL); Casino, McLean in 1918 (srt); Ballina,
Baker in 1892 (+pr); Clarence River, Beckler sn. (MEL); Hastings River,
Beckler s.n. (MEL); Gloucester River, Betche in 1882; Tiona, south of Forster,
Garden 19004 (A, Mo); Hunder and Paterson Rivers, "Brown in 1802-5 (MEL),
Brown 3117 (Type, +8m); Ourimbah, Laseron 435 (+a); Hlawarra Distr., near
Bulli, Kirton 8 (mex), Anon. 152 (ert); Anon. (MEL). LOCALITY UNKNOWN:
Dundowran, Nichenbach, Sandy Flat, Tryon in 1928 (prt); Bennett’s River,
ania in 1943 (MEL); Ash Island, Leichhardt sn. (MEL). Tasmania: Van
Diemen’s Land, Caley in 1805 (+BM).

Clava ted. AUSTRALIA. Queensland: Brisbane Botanic Garden, White 2337
(A, BRI), Everist 75 (mo). New South Wales: Sydney, The Do main, Boyce
19005 (a, Mo, uC), Botanic Garden, Camfield in 1896, 2 sheets, 6 and 2 (tuc),
Eames s.n. (cu). Asta: Nepal (2), Villa — caekas in 1889 (A), Wallich
sn. (tp); Singapore, pees sn. (+BR). RICA: Tunis, Anon. in 1868
+mo, 2 sheets), Anon. in 1870 (+mo); Natal, ae Durham Botanic Garden,
Wilson i in 1922 (tA); Cape Colony, Capetown State Nursery, Bowman in 1921

—— Los Zooras, Jardin Suizo, Harshberger 1031 (PENN, tus); site un-

nown, Anon. S.n. (pw), NortH America. United States. New York: New
| Botanical Garden, 3740 ex DPM in 1900, Nash in 1905 (+Nny), Hartling in
1919 (+ny), 3750 ex DPM in 1900, Nash in 1905 (+Ny), 4788 ex DPM in 1900,
Taylor in 1906 (+NY), 8766 ex MBG in 1901, Nash in 1901 (+Ny), 11800 ex
DPB in 1902, Taylor in 1905 (ny), 14415 ex Edinburgh in 1902, Taylor in 1905
(xy); Manhattan, Dept. of Parks 3740, Hartling in 1900 (t1LL). New Jersey:
cult, in greenhouse from CU56 from Atkins Garden, Cuba, Johnson in 1952
(ruT). District of Columbia: Washington, Foxworthy in 1903 (+cu, Ny), Anon.
sm. (ILL). California: Stanford University, Dudley in 1909 (+ps); San Fran-
cisco, Golden Gate State Park, Walther in 1936 (cas), Buchholz in 1942 (+1LL);
Mill Brae, Mills’ Place, Walther in 1921 (cas); Berkeley, Eastwood in 1913
(cas), Curran 23 (A); Goleta, Sexton’s place near Santa Barbara, Eastwood in
1917 (tcas), Van Rensselaer 1716 (1LL), Buchholz in 1941 (}1LL); Santa Bar-
bara, Eastwood in 1926 (a, +cas); Hueneme, Bord’s place (Ventura Co.),
Walther in 1921 (cas), Vijos in 1917 (cas); San Diego, Wongenheim in 1941
(i Jefferson’s, Walther in 1926 (cas), Wilson in 1937 (cas), Hawaiian Is-
lands: Oahu, Curran 122 (tus).

Podocarpus elatus is not represented in our herbaria by many specimens
collected directly from native habitats in Australia, but there are a large
number of specimens from cultivation both here and abroad. The culti-
vated specimens have masqueraded under a number of different names:
P. macrophyllus (China), P. neriifolius (Nepal), P. coriaceus (West In-
dies), P. polystachyus (Singapore), P. latifolius (Africa), P. purdieanus
(Jamaica) and P. spinulosus (Australia). These specimens were first rec-
ognized as being identical by the striking uniformity in general appearance
and by the transverse sections of the leaves (Fig. 3) in which the single
layer of upper hypoderm and three vascular resin canals proved to be dis-
tinguishing characters. Later it was found that the broad, straight, often
mucronate leaves also served to identify this plant

An undetermined specimen collected in Tasmania by Caley in 1805

1958] GRAY, REVISION OF PODOCARPUS, XI 439

(BM) proves to be Podocarpus elatus. The early date of this collection pre-
cludes the possibility that this species could have been planted by Euro-
peans and grown to a reproductive size by the time of its collection. A
leaf from a specimen (P) supposedly collected by Wallich in Nepal and
determined by Lindley, first as P. macrophyllus, later as P. neriifolius, was
examined. This specimen is also P. elatus, but it is very doubtful that
this species is really to be found in Nepal. Either this specimen was picked
up by Wallich en route and not given clear collection data or it has been
mislabelled subsequently. Gibbs (6) lists 819 frem Fiji as P. elatus. This
is also very unlikely, as I have examined numerous specimens from Fiji
and never found this species. Orr (9) included P. elatus in Fiji and also
suggests that it is in New Caledonia and the islands of Polynesia, but it
has not appeared among any specimens I ,have examined from these places.

Both Stiles (10) and Orr (9) examined the anatomy of the leaves of
Podocarpus elatus and are in agreement as far as they report their observa-
tions. However, Stiles includes it with P. polystachyus, attributing the
differences in hypoderm and auxiliary sclereids to external conditions.
This does not seem to be the case since both the native collections and
the large number of cultivated specimens from other climates exhibit
unexpected uniformity in the leaf anatomy.

Baker and Smith (1) included photographs of the mature tree (p. 432),
male cones and mature seeds on branches (p. 434), photomicrographs of
transverse and longitudinal sections through the leaves (pp. 436-37) and
wood and bark anatomy (p. 438).

Orr (9) stresses that the three vascular resin canals found in this species
are exceptional among the Australian members of sect. Podocarpus. It
is interesting to note that there are rare specimens in which only one
vascular resin canal is developed, and also that one specimen has as
many as five, two being in the transfusion tissue; however, the regular
number is three. Orr observed auxiliary sclereids in the mesophyll, but
I found a few in only a single specimen, Anon. 10545 from Mt. Spergeon.
This specimen also differed from the type in the large seed, 22 mm. long
and 13 mm. wide, with three distinct ridges on the back and a strong
beak.

Other specimens from northern Queensland differ from the type in hav-
ing long leaves (up to 18 cm.), with apices not mucronate but long-acute
(Mueller in 1864, Rockingham Bay, and Hartmann in 1886, Cape York).
The leaves of Moore 39 from Fitzroy Island are up to 30 cm. long and
22 mm. wide, thus reaching the lower limits of Podocarpus dispermus,
but the leaves in this latter species have interrupted hypoderm. The
specimen also shows five vascular resin canals, three below the vascular
bundle and two in the wings of transfusion tissue, and a flat or depressed
upper midrib. Its general appearance is that of a juvenile specimen.

Allen Cunningham, a well-known Australian explorer of early days, in
the diary (7) of his travels mentions seeing a Podocarpus several times
in the vicinity of Paramatta and Liverpool, very near Sydney, New South
Wales. We now know that this species was P. elatus.

440 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

7. Podocarpus brassii Pilger, Engler Bot. Jahrb. 68: 246. 1938; Was-
scher, Blumea 4: 469. 1941

An erect tree, 10-12 m. tall, densely branched, with thick, stiff, verticil-
late, ridged branchlets. Terminal buds large, globose, the scales often
spreading, rigid, ovate-lanceolate, keeled, with membranous margins, up

ong, apex stiffly acuminate. Leaves crowded, erect to patent,
thick, very stiffly coriaceous, broadly lanceolate, 1-1.8 cm. long, 3—7 mm.
broad (juvenile 2.5-4 cm. long), the apex acute or obtusely apiculate,
narrowing at base to a short broad petiole, shiny above, duller beneath
with the exception of a broad marginal band; margins revolute; midrib
narrow and prominent or scarcely prominent above, broader below. Trans-
verse sections of the leaves show three (rarely one) vascular resin canals,
upper hypoderm of scattered fibers, rare isolated lower hypodermal fibers
between the stomatal rows, rare vascular sclereids and no auxiliary sclereids.
Leaves quite thick with more than one layer of palisade mesophyll. Pollen
cones axillary, solitary, sessile, surrounded at the base by ovate-triangular
scales, cylindric, 2.5-3 cm. long, 3—7 mm. thick. Microsporophylls densely
imbricate, each with long triangular acute apiculus. Female cones solitary
in upper leaf axils; peduncles thick, 2-9 mm. long; receptacle of 2—3 fused
fleshy scales, 5-9 mm. long, 2.5-7 mm. broad, subtended by 2 narrowly
triangular bracts, 3 mm. long; ovule usually solitary. Seeds elliptical-
globose, rounded at the base and apex, 7-10 mm. long, 5—6 mm. broad.

DistRIBUTION: High mountains, above 3000 m. altitude, New Guinea.

Guinea. NETHERLANDS NEw GuINEA: Lake Habbema, 3225 m. Camp,

Bras 9341 (ta), 9342 (+a), 9560 (+LAE), Brass & Meyer-Drees 10435 (A),

436 (A). TERRITORY OF NEw GuINEA. Eastern Highlands District: Mt. Wil-

a Lake Piunde, Womersley 8866 (LAE). Papua. Central Division: Mt. Al-

bert Edward, Brass 4295 (1soTYPE, +A—2 sheets; +BRI), Brass 4395a (prt), Brass
4396 (BRI).

The leaves of Podocarpus brassii are smaller than those of P. brevifolius,
which are usually over 2 cm. long. It also differs from P. glaucus in that
its young foliage is not glaucous and the pollen cones are thick instead
of slender. Orr (9) agrees with the above description of the leaf anatomy.
I have observed that the upper hypodermal fibers are larger in Brass 9342
than Brass 4395.

8. Podocarpus glaucus Foxworthy, Philipp. Jour. Sci. 2: 258. 1907,
6: 159. 1911; Pilger, Nat. Pflanzenfam. ed. 2. 13: 248. 1926; Florin,
Svenska Vet.-Akad. Handl. II]. 10: 262, 266. 1931; Wasscher, Blumea
4: 468. 1941

A small bushy tree, 5-6 m. tall, with branchlets crowded toward tips
of branches, light colored. Terminal buds very small, the outer scales
spreading, up to 2 mm. long, with acuminate apex. Leaves densely crowded,
erect-patent, coriaceous, glossy, pale beneath and glaucous when young,
oblong, elliptic-oblong or spatulate, 0.9-1.7 cm. long, 3.5—-5.5 mm. wide,

1958] GRAY, REVISION OF PODOCARPUS, XI 441

apex obtuse or rounded, the base gradually narrowed to 1-2 mm. petiole
and somewhat decurrent; margins thickened; midrib prominent near base
on upper side, becoming flat and indistinct toward the apex, very prom-
inent beneath. Transverse sections of the leaves show three resin canals
below the vascular bundle, upper hypodermal fibers large and scattered,
no lower hypodermal fibers between the stomatal rows, no vascular fibers
nor auxiliary sclereids. Pollen cones solitary in upper leaf axils, scattered,
slender cylindric, 1-1.5 cm. long, 2-3 mm. thick. Microsporophylls densely
imbricate. Female cones and seeds not known.

DIsTRIBUTION: Mt. Halcon, 2400 m. altitude, on Mindoro in the Philip-
pine Islands and the slopes and summit of Mt. Moetaro in New Guinea.

Philippine Islands: Mrnporo: Mt. Halcon, Merrill 5672 (+ex Florin, Berlin
Herbarium). Netherlands New Guinea: Wissel Lake region, Mt. Moetaro, Eyma
5208 (tA).

Podocarpus glaucus is a little-known species and its position in the
genus has been doubtful until the present time. The examination of the
transverse sections of the leaves shows three resin canals under the phloem
of the vascular bundle, transfusion tissue, accessory transfusion tissue,
and hypodermal fibers on both sides of the leaf. These characters definitely
place it in subsection B of section Podocarpus; it seems to be most closely
related to P. brevifolius, which has larger leaves with parallel margins, and
P. brassii, which has very thick pollen cones.

Florin (4) listed Podocarpus glaucus as a hypostomatic species of section
Stachycarpus. The solitary pollen cones and the leaf anatomy show that
it could not belong to this section. Wasscher (11) concluded correctly that
it belongs in section Podocarpus, but he did not see Merrill 5672, the speci-
men on which the species is based. Clemens 32032 and 40001, which he
lists, have continuous upper hypoderm and belong elsewhere, invalidating
his description for this species. I have not seen the other two specimens
which he lists, but I do not believe that P. glaucus has yet been collected
in Borneo.

Orr (9) followed Florin in including Podocarpus glaucus in section
Stachycarpus and he described the leaf anatomy correctly, but unaccount-
ably treated the species along with P. amarus and P. rostratus as aberrant
species in this group. His suggestion that the three species comprise a
separate group is not tenable as they differ greatly from each other and
each species must have separate consideration. By not placing this species
in section Podocarpus with others from the Pacific islands, he missed see-
ing its similarity to P. brassw and P. brevifolius.

9. Podocarpus brevifolius Foxworthy, Philipp. Jour. Sci. 6: 160. 1911;
Pilger, Bot. Jahrb. 54: 40. 1916; Florin, Svenska Vet.-Akad. Hand.
III. 10: 279. 1931; Wasscher, Blumea 4: 466. 1941; Metcalf, FI.
Fukien 1: 21. 1942.

P. nerifolius var. brevifolius Stapf, Trans. Linn. Soc. Bot. I]. 4: 249. 1894;
Pilger, Pflanzenreich IV. 5(Heft 18): 93. 1903.

442 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

P. wangii C. C. Chang, Sunyatsenia 6: 26. 1941; Metcalf, Fl. Fukien 1: 21.
1942.

A small tree or shrub with verticillate, spreading, stout branchlets. Ter-
minal buds small, 2.5 mm. wide, with outer scales keeled, 3-7 mm. long,
tardily deciduous, the tips acute to long-acuminate sometimes spreading.
Leaves densely crowded, erect, thick-coriaceous, rigid, quite flat, straight
or slightly falcate, elliptic to lanceolate, sessile or subsessile, 1.5—5.5 cm.
ong, 4-6 mm. wide, shining above, duller beneath, apex acute or some-
what rounded; upper midrib prominent and narrow, broader below; mar-
gins flat or revolute. Transverse sections of the leaves show three vascular
resin canals below the phloem (rarely only one), interrupted upper hypo-
derm of small or often quite large fibers, no hypodermal fibers between
the stomatal rows below, no auxiliary sclereids and no vascular fibers above
the xylem, rarely below the phloem. Male cones solitary in the upper
leaf axils, sessile, cylindric, thick, 2-3 cm. long, 4-5.5 mm. in diameter,
scales at the base triangular. Microsporophylls imbricate, triangular, with
obtuse apiculus. Female cones solitary, axillary near the tip of the branch-
let; peduncles 2-4 mm. long, broad and flattened; receptacle of two fused
fleshy scales, free at the tips, subcylindrical, 5—6.5 mm. long, 2.5—-4 mm
thick, subtended by a pair of subulate to narrow-triangular, keeled, acute
bracts, 3-5 mm. long. Seed ovoid, 1.2 cm, long, apex obtuse

DistRIBUTION: High mountain slopes in the Philippine Islands, on
Mt. Kinabalu in British North Borneo and on Hainan island.

Philippine Islands. Luzon: Zambales, Bur. Sci. 5002, Ramos (a, Ny). For. Bur.
9511, Curran & Merritt (+mo, Ny, tus). Hainan. Wang 36533 (A, Ny). British
North Borneo. Mt. Kinabalu: Paka Cave to Low’s Peak, Clemens 10657 (A,
GH, tUC); upper Kinabalu, Clemens 27825-27103 (4 tNy); above Paka, Clemens
28901 (a, ty); Gurulau Spur, Clemens 50825 (a, +uc); Penibukan, Clemens in

[936-(A, WY UG).

Podocarpus brevifolius was first described as a variety of P. neritfolius
by Stapf from an early collection from Mt. Kinabalu. He included it under
this species only because Hooker had combined P. neriifolius and P. poly-
stachyus and he believed it to be more closely related to the latter. Fox-
worthy (5) suggested that it is more closely related to P. pilgert, with
which I would agree. According to Wasscher (11), it is “a very distinct
species in its adpressed, small, lanceolate, thick-coriaceous leaves.”

This examination of a number of specimens of Podocarpus brevifolius
shows that the layer of upper hypodermal fibers is always interrupted.
I believe that Orr (9), who is not in agreement with this, examined only
Clemens 32021 from the British Museum since his description fits this
specimen perfectly. It is here included in the new species, P. gibbsii.

Metcalf (8) who lists this taxon from China, assigns to it three speci-
mens: Wang 36533 from Hainan and Wang 39578 and 40196 from
Kwangsi. I agree on the specimen from Hainan but J] question the two
from Kwangsi on the mainland as I have not seen them and they may be
Podocarpus macrophyilus var. chingii.

1958 | GRAY, REVISION OF PODOCARPUS, XI 443

10. a aaa paar (Smith) R. Brown ex Mirb. Geogr. Conif.
n Mém. Mus. 13: 75. 1825; Endlicher, Syn. Conif. 213. 1847; Car-
ns Traité ne 653: 1867 (in mare Parlatore in DC. Prodr. 16:
513. 1868; Bentham, Fl. Austral. 6: 247. 1873; Mahlert, Bot. Cen-
tralbl. 24: 281. 1885; Pilger, Pflanzenreich IV. 5(Heft 18): 78. 1903,
Nat. Pflanzenfam. ed. 2. 13: 247. 1926; Brooks & Stiles, Ann. Bot. 24:
305. 1910; Baker & Smith, Res. Pines Austr. 443. 1910; Dallimore &
Jackson, Handb. Conif. 56. 1931, 81, 1948; Florin, Svenska Vet.-Akad.
Handl. III. 10: 280. 1931.

P. bidwillit Hoibrenk ex Endlicher, Syn. Conif. 213. 1847.

P. excelsa Loddiges ex Endlicher, Syn. Conif. 213. 1847.

P. pungens Don ex Lamb. Pinus, ed. Le 2s 21s 824ed 2. 2A le 2S.
Nageia spinulosa F. Mill. Cens. 109. 1882.

Taxus spinulosa Smith in Rees Cyclop. 35. 1819.

A small tree, sometimes shrublike, with scattered or subverticillate
branches. Terminal buds small, ovate with erect triangular, narrowly
acuminate or attenuate scales up to 8 mm. long. Leaves scattered, coria-
ceous, erect to patent, linear, subsessile, 2—-6.5 cm. long, 2-4.5 mm. wide,
shiny eboro) apex long-acuminate and often pungent, ae tapering abruptly
to a very short petiole, or sessile; midrib narrow and evident above,
broader and less prominent below. Transverse sections of the leaves show
only one vascular resin canal, interrupted upper hypoderm often scarce
between margin and midrib, no hypodermal fibers between the stomatal
rows below, often two layers of palisade parenchyma, and regular trans-
fusion tissue, upper vascular fibers rare. Male cones often very abundant,
solitary or in clusters of 3-5 at the tips of very short, 1-3 mm., axillary
peduncles subtended by 5—6 mm. long bracts, apiculate, simulgne leaves;
each cylindric cone, 4-8 mm. long, surrounded at the base by a few stiff
scales. Microsporophylls sce imbricate, obtuse, apiculate. Female cones
axillary to prophylls or true leaves in Baal part of young shoot; peduncle
slender, up to 1 cm. long; receptacle of 2—3 fused fleshy unequal scales,
free at the tips, subtended by 2 subulate bracts 1-4 mm. long, glaucous,
deep purple and edible when mature, bearing one ovule. Seed broadly
ovoid, 1-2 cm. long, 7-10 mm. wide, often with a stout beak.

DistRIBuUTION: In light forest, usually in sandy soil; southeastern New
South Wales near Sydney.

Australia. New SoutH WaALEs: Sydney, Anon. s.n. (+cAs), Anon. in 1897
(a), Anderson 34 (NY), Wright in 1853-56 (Ny, y), Prajerus in 1889 (mMTMG),
Kenny s.n. (BRI-2 sheets); Port Jackson, Mueller in 1855 (MEL), Manly Beach,
Dawson s.n. (MEL); Lake Narrabeen, beyond Manly, Wheeler’s place, Maiden
in 1887 (MEL); Paramatta, North Rocks, Anon. s.n. (MEL); Loftus, Camfield
in 1897 (tmo); Collarez, Burges in 1930 (Ny); Mosman, Cannon 443 (a);
Blackdown Tableland, Simmons 76 (prt); Lake Tabourie, Baker in 1892 (Bri);
Beerwah, White 12858 (sr1-2 sheets, 6 and 2); R. Brown in 1802-5 (+E);
Jervis Bay Naval Reserve, Willis in 1954 (MEL). STRADBROOKE ISLAND: White
1665 (Bri); White 1708 (A, BRI). Locality unknown: Anon. s.n, (BRI).

444 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Podocarpus spinulosus resembles P. drouynianus in having long, slender
crowded leaves, but in P. spinulosus the leaves taper to a definite short
petiole and the midrib is prominent. In the transverse sections of the
leaves the cuticle is a little thinner, the accessory transfusion tissue is
more regular and the palisade layer is thicker. Orr (9) agrees with the
description of the leaf anatomy and the specimen, R. Brown in 1802-05
(£), is probably a duplicate of the one he used.

Lt,

—

Podocarpus drouynianus F. Miiller, Fragm. 4: 86, ¢. 31. 1863-64;
Bentham, Fl. Austral. 6: 274. 1873; Bertrand, Ann. Sci. Nat. V. 20:
64.1874; Van Tieghem, Bull. Soc. Bot. France 38: 169. 1891; Pilger,
Pflanzenreich IV. 5(Heft 18): 77. 1903, Nat. Pflanzenfam. ed. 2
13: 247. 1926; Baker & Smith, Res. Pines Austral. 443. 1910; Dal-
limore & Jackson, Handb. Conif. 48. 1931, 66. 1948; Florin, Svenska
Vet.-Akad. Hand]. III. 10: 279. 1931.

P. brownii Bertrand, Ann. Sci. Nat. V. 20: 64. 1874 (nomen).

Nageia drouyniana F, Mull. Cens. 104. 1882.

A small tree or shrub, 1 m. high, with erect branches. Terminal buds
ovate, with narrow, acute or attenuate scales up to 5 mm. long. Leaves
spirally arranged, erect to patent, thin, leathery, linear, straight, subsessile,
2-12 cm. long, 2-5 mm. broad, green above, glaucous beneath, apex acute-
acuminate, rarely pungent, short angustate at the base; margins some-
what revolute and thickened; midrib scarcely prominent or flat above,
broad and prominent beneath. Transverse sections of the leaves with only
one resin canal below the vascular bundle, the upper hypoderm inter-
rupted, no hypodermal fibers between the stomatal rows, often very little
transfusion tissue, accessory transfusion tissue usually of very irregularly
elongate cells which are quite thick, vascular fibers very abundant below
the phloem and forming a tissue several cells thick; cutin very thick
and rounded over each epidermal cell; epidermis of very large cells, up
to 207 » long and 11-46 p» wide, pitted, with somewhat wavy walls. Pollen
cones solitary or clustered at the ends of slender axillary peduncles 1—2.5
cm. long, or widely scattered, up to six on special axillary branches, each
cone on a separate peduncle with a subulate bract at the base, sometimes
with sterile bracts; mature cones short, thick, cylindric, 4-10 mm. long, 4
mm. wide. Microsporophylls densely imbricate, minutely apiculate. Female
cones solitary in the axils of prophylls or the lowermost true leaves of new
growth; peduncles variable, up to 2.0 cm. long; receptacle of 2 or 3 fleshy
fused unequal scales with the tips free, 2.5 cm. long, subtended by 2 narrow
bracts 2-3 mm. or more long, sometimes foliaceous and unequal (up to
11 mm. long, 2 mm. wide on Gilbert 45), waxy-coated, purple at maturity,
sometimes bearing 2 ovules when young but only one maturing. Seeds
broadly ovoid, 1-1.7 cm. or more long, not crested.

Australia. WESTERN AUSTRALIA: Busselton, Pries in 1870 (MEL); Warren Dist.,

Manjimup, Koch 2499 (BRI, MEL, t}MO), Warren River, Mueller in 1877 (MEL) ;
Tom River, Mueller s.n. (MEL); Dape River, Oldfield s.n. (MEL); Preston Gor-

1958 | GRAY, REVISION OF PODOCARPUS, XI 445

don River, Mueller in 1877 (mEL); Greenbushes, Cook in 1939 (MEL); Bow
River, Jackson in 1912 (pri); Blackwell River, Mueller sn. (pH); Fly River,
Wilson 263 (+a); Vasse River, Dept. Agr. in 1898 (A); Denmark, Jarrah For-
est, White 5377 (A, BRI, tNy), Dept. Agr. 443 (+a); southwest, Eames in 1937
(cu); Baker in 1904 (+a); Gilbert 45 (+8M, Mo); Drummond 154 (MEL).

Podocarpus drouynianus has grass-like leaves which differ from those
of P. spinulosus in having broad bases and in being sessile or very nearly
so. The upper surface of the leaf is nearly flat at the midrib. The leaf
anatomy differs little except for the very thick cuticle and the loose con-
struction of the accessory transfusion tissue.

Orr (9) examined the leaf anatomy of this species and reported the single
vascular resin canal; the thick cutin was also a distinguishing character.

Mueller in 1877 (MEL) from the Warren River is an excellent example
of the reduction from normal vegetative leaves to the scale-like prophylls of
the reproductive portion of the stem. The specimen twig is about 45 cm.
long and the basal portion, 5-6 cm. long, is bare, the old leaves having
fallen away. The next 20 cm. bears short, 2.5 cm. long, spreading leaves
which decrease upward to narrow triangular scales 5 mm. long with spread-
ing tips. In the following 10 cm. the prophylls are up to 8 mm. long and in
the axils are solitary peduncled fruit. The remaining 14 cm. has large
linear, mucronate leaves, 4-7 cm. long and 2-3 mm. wide, terminating in a
bud which has just burst with new growth without having yet elongated.

Podocarpus affinis Seem. Fl. Vitiens. 266. 1865-73; Parlatore in
DC. Prodr. 16: 517. 1868; Warburg, Monsunia 1: 193. 1900; ae
Pflanzenreich IV. 5(Heft 18): 78. 1903, Nat. Pflanzenfam. Se

248. 1926; Florin, Svenska Vet.-Akad. Handl. III. 10: 279. oe

Dallimore & Jackson, Handb. Conif. 38. 1931, 60. 1948.

A tree with short, spreading branchlets, densely leafy toward the tips.
Leaves spirally arranged, patent, linear- elliptic, 3—5.2 cm. long, 6-9.5 mm
wide, reddish-brown underneath when dry (young leaves eaneouSe the
apex obtuse with pointed tip, the base narrowed into a short distinct
petiole; midrib narrow and often very prominent above, narrow and some-
times with shallow groves on either side below. Transverse sections of the
leaves show three vascular resin canals with vascular fibers usually present
both above and below the bundle, an interrupted hypoderm of fibers 18-30
uw in diameter, usually quite abundant on the upper side, but very rare on
the lower between the stomatal rows, auxiliary sclereids in the mesophyll
between the palisade mesophyll and the accessory transfusion tissue and
also a few in the palisade making that layer discontinuous, palisade tissue
occasionally doubled in thicker leaves, the cuticle sometimes thick. Male
and female cones unknown.

—
bo

Fiji. Vitr Levu: Namosi, Voma Peak, Seemann 574 (Type, +BM, +cH, XK),
Gillespie 2721 (tuc), Horne 769 (kK), 973 (+GH, K).

Orr (9) examined the leaf anatomy of Podocarpus affinis and his account

446 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

agrees with the above description. The short leaves distinguish this species
from the others in Fiji but only the vegetative characters are known.

13. Podocarpus sylvestris Buchholz, Bull. Mus. Paris II. 21: 285.
1949

A tree becoming 15-18 m. high with the trunk up to 2 m. in diameter;
outer bark gray and furrowed between longitudinal plates, the inner bark
reddish brown; twigs very slender, subverticillate. Vegetative buds
spherical or ovoid, solitary or in groups of three, with thin, slightly cari-
nate, ovate, apiculate or obtuse, appressed scales. Leaves aggregated toward
the ends of twigs, linear-lanceolate, 5—9 cm. long, 7-9.5 mm. wide (some-
times up to 17 cm. long and 15.5 mm. wide), narrowed above to an ob-
tusely rounded apex, narrowed at the base to a short petiole, bright shiny
green becoming gray-green above, lighter dull green beneath; midrib flat
on both surfaces or broadly prominent. Transverse sections of the leaves
show 3 vascular resin canals, hypoderm interrupted above with small fibers
rarely more than 20 » in diameter, few and scattered fibers between the
stomatal rows below, vascular sclereids usually above and sometimes below,
auxiliary sclereids present in the palisade mesophyll and abundant in the
mesophyll above the accessory transfusion tissue. Pollen cones axillary,
in clusters of 3, sessile or the peduncle 1 mm. long, cylindrical, 8-16 mm
long, 2—2.5 mm. in diameter when fully expanded; surrounded at base by
7—9 thin imbricated bud scales, the outer acute, the inner obtuse. Micro-
sporophylls with a scarious obtuse apiculus. Female cones axillary on
slender peduncles 5-9 mm. long; receptacle of 2 fused fleshy unequal
scales 6-7 mm. long, subtended by a pair of small deciduous bracts and
bearing a single ovule. Seed elongated, becoming 13 mm. long and 8-9 mm.
wide, with only a minute or suppressed crest.

DISTRIBUTION: Scattered in mixed angiosperm forests of southern New
Caledonia at 150-400 m. altitude.

w Caledonia: Plaine des Lacs, foret du Mois de Mai, Buchholz 1696 (Type,
ha. a 13518 (+111, P),. 13522 ne *); 1354 (ILL), 1360, 1372 — juwenile
(ILL, P) 1392 (ILL), Bernier 158 (P), 219 (e), Bernier in 1947 (iLL); Baie
des Pirogues, White 2118 (+a, BRI, P), White sn, (K-2 sheets); north a bs
Louis, Thy River, Buchholz 1233 (4111); Mt. Canala, Compton 1273 (+BM);
North Prony, Cribs 4718 (te); Mt. Balade, Vieillard 1265 (A, P); Wayap,
Vieillard 1265 (Pp); Pic du Pin, Virot 598 (a); Mont Mi, Virot oss (A).
Locality unknown: Kay 34 (+p), Vieillard 1265 (Pp).

It is likely that Podocarpus sylvestris has been confused in the past
with P. longifoliolatus Pilger in subsection F, a tree which occurs on Mt.
Mou in high coniferous rain-forests above 1000 m. It differs from the latter
in its more slender twigs and globose vegetative buds with closely appressed
ovate or obtuse, apiculate scales. Transverse sections of the leaves show
only three vascular resin canals and smaller hypodermal fibers. Podo-
carpus longifoliolatus has, in addition, two or more resin canals above the
vascular bundle. The female cones of P. sylvestris bear solitary ovules.

1958] GRAY, REVISION OF PODOCARPUS, XI 447

According to the manuscript notes of Prof. Buchholz, the vegetative
buds of this species are more like those of P. novae-caledoniae. It differs
from the latter in the much wider leaves with interrupted hypoderm.
Podocarpus novae-caledoniae is the only other species in New Caledonia
belonging in this subsection and its shrubby habit distinguishes it readily.

The wood is light, reddish, of excellent quality and is used for lumber.
It is called ‘False Kauri” where it is marketed. One specimen, Buchholz
1392, has white wood but does not seem to differ otherwise.

14. Podocarpus idenburgensis, sp. nov.

Arbor 12-33 m. alta; ramulis tenuibus gemmis terminalibus parvis,
ovatis vel globosis; foliis apice ramulorum congestis, lineari-lanceolatis vel
lineari-ellipticis, 8-15 cm. longis et 9-15 mm. latis, apicibus acutis, petiolis
6-10 mm. longis, strobilis masculis alabastris axillaris cicatricorum foli-
orum oppositis positis, pedunculo tenue, 3-12 mm. longo, strobilis maturis
4 cm. longis, 5 mm. crassis; strobilis foemineis solitariis, pedunculo tenue,
8 mm. longo; semine globoso, 9 mm. longo, apice tenue rotundato.

A tree 31-57 cm. in diameter, the outer scales of the terminal buds on
the twigs narrowly triangular, erect or spreading, up to 6 mm. long, the tips
long acute or acuminate. Leaves straight or falcate, coriaceous, gradually
or more abruptly narrowing to a definite petiole and gradually narrowing
to an acute apex; midrib usually narrowly prominent above, almost flat
below. Male cone bud scales acutely triangular, erect; microsporophylls
imbricate, tips broadly obtuse, scarious, somewhat erose. Female cones
axillary; receptacle very fleshy of 3 fused scales almost 1 cm. long.

Distripution: In high mountain rain-forests in Netherlands New
Guinea.

New Guinea. NETHERLANDS New GUINEA: 6 km. sw of Bernhard Camp,
Idenburg River, Brass & Versteegh 12581 (Typr, +LAE); 2 km. sw of Bernhard
Camp, Idenburg River, Brass & Versteegh 13530 (+a, LAE); Bele River, 18 km.
NE of Lake Habbema, Brass & Versteegh 11133 (+A, LAE).

This species must be distinguished from both Podocarpus neriifolius and
P. rumphii in New Guinea. It differs from these species in its large male
cones, usually solitary or rarely paired, on peduncles up to 12 mm. long.
The shape of the leaves is like that of both of the other species, while the
terminal buds are more like those of P. neriifolius. Transverse sections of
the leaves did not show any auxiliary sclereids and there were no lower
hypodermal fibers in Brass & Versteegh 11133.

The name refers to the Idenburg River which flows adjacent to the region
in which the species is found most abundantly.

15. Podocarpus ledermannii Pilger, Bot. Jahrb. 54: 210. 1916, Nat.
Pflanzenfam. ed. 2. 13: 248. 1926; Florin, Svenska Vet.-Akad. Hand.
III. 10: 279, 283. 1931; Wasscher, Blumea 4: 456. 1941.

A tree up to 20 m. high with gray bark and slender spreading branch-

448 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxrx

Jets; terminal buds small and ovate with long-triangular apiculate bud
scales. Leaves scattered or crowded at the tips of the branches, thinly
coriaceous, flexible, elliptic-lanceolate or narrow-elliptic, broadly cuneate
or cuneate-rounded at the base, obtuse or abruptly narrowed and acute at
the apex, often almost caudate at the tip, scarcely shiny above, dull below,
8—12 cm. long, 12-27 mm. wide; midrib abruptly prominent above, broader
below. Transverse sections of the leaves show 3 vascular resin canals, in-
terrupted hypoderm of large scattered or isolated fibers, some hypodermal
fibers scattered below between the stomatal rows, vascular sclereids both
above and below the bundle, auxiliary sclereids in the mesophyll both
above and below the accessory transfusion tissue. Pollen cone buds in
groups of 2 or 3 on very short peduncles; scales ovate-triangular, acute,
keeled, 2 mm. long, inner ones more obtuse. Microsporophylls ovate-tri-
angular obtuse or with acute apiculus. Female cones solitary, axillary, on
peduncles 6.5—9 mm. long; receptacle of 2 or 3 fused fleshy scales with
free tips, subtended by 2 subulate bracts 2.5 mm. long. Seed elliptic or
globose, about 8 mm. long, crested.

DistTRIBUTION: In mountain forests of New Guinea, Java and Borneo.

New Guinea. NETHERLANDS NEW GUINEA: Bele River, 18 km. NE of Lake
Habbema, Brass 11058 (+a); Idenburgh River, 6 km. SW of Bernhard Camp,
Brass 12749 (+A). Terr1ToRY oF New Gutnea. Sepik District: Lordberg,
Ledermann s.n., ex Berlin Herbarium (tex Florin). Eastern Highlands District:
Chimbu, Cavanaugh NGF3336 (+LaE). Morobo District: Yungaing, Clemens
2352 (+A). Java. Horsfield sn. (+GH); Preanger, Warburg 2678 (+Nv). Borneo.
SARAWAK: Foxworthy 377 (+us); Mt. Poi, Clemens 20348 (+a, Ny).

Podocarpus ledermannii differs from P. neriifolius in the more oval or
oblong shape of the leaves and from P. rumphii in its abruptly or narrowly
prominent midrib. Transverse sections of the leaves differ from P. nerii-
folius and P. rumphii in the large hypodermal fibers, hypodermal fibers be-
tween the stomatal rows and more abundant auxiliary sclereids. Orr (9)
examined this species and found the same differences. The leaves of the
specimens from Java show five vascular resin canals below the bundle and
those from Borneo show three to five and there are few if any hypodermal
fibers between the stomatal rows.

16. Podocarpus dispermus White, Contrib. Arnold Arb. 4: 10. 1933.

A small tree up to 17 m. high with light brownish-gray, slightly flaky
bark. Vegetative buds small; the scales narrow, acute, stiff and erect, up to
6 mm. long, longer than bud. Leaves dark glossy green, straight, broad-
linear or narrow-lanceolate, apex acute, often pungent, gradually narrow-
ing at base to a short petiole, 10-20 cm. long, 20-30 mm. wide; midrib
broadly prominent above and below. Transverse sections of the leaves
show usually 3 vascular resin canals (rarely one or five), an interrupted
upper hypoderm of fibers 18-33 » in diameter, lower hypodermal fibers
between the stomatal rows, vascular sclereids present above and below,

1958] GRAY, REVISION OF PODOCARPUS, XI 449

and auxiliary sclereids absent in palisade and spongy mesophyll. Pollen
cones sessile, in clusters of 1-3, 3 cm. long, 3 mm. broad, surrounded by
acute scales at the base. Microsporophyl crowded, imbricate, with the
apiculus short, broad, and a Female cones axillary, on thick pe-
duncles 5-15 mm. lone: ce ee fleshy receptacle subtended by two
small deciduous scales, turning scarlet when the fruit is ripe. Seeds usually
in pairs, ellipsoid, 2.5 cm. long, 1.7 cm. broad.

DIsTRIBUTION: In rain forests in northeast Queensland.

Australia. QUEENSLAND: Atherton Tableland, Gadgarra Reserve, Kajewski
1192 (Typr, +A, BRI, ILL, NY), Kajewski 1107 (+A, BRI, NY-3 sheets); Millea
Millea, Tardent in 1930 (a, Br); Johnstone River, Michael 97 (srt).

Podocarpus dispermus differs from P. elatus in having broader and usu-
ally longer leaves with interrupted hypoderm. The receptacle and seeds
are larger and the ripe receptacle of P. dispermus is red instead of blue-
black. This species is not very abundant and has a restricted range in the
region of Atherton in northern Queensland. Orr (9) did not see any speci-
mens of P. dispermus, but he correctly included it in section Podocarpus.

The pollen cones on Michael 97 are on short (1.5 cm.) branches arising
from stems more than one year old, which also bear leaves. This is the only
specimen I have seen with pollen cones.

17. Podocarpus salomoniensis Wasscher, Blumea 4: 430. 1941.

Tree up to 20 m. tall with pale yellowish-brown, thin bark, usually
smooth but sometimes fissured and flaking, the scattered drooping branches
with stout twigs. Terminal buds large, ovoid, with long-attenuate scales up
to 11 mm. long, the outer spreading and keeled, the inner erect. Leaves
spirally arranged, spreading or crowded, linear to linear-lanceolate, some-
times falcate, 12-18 cm. long, 6.5-9 mm. wide, shiny above, dull beneath,
long-tapering to an acuminate apex or sharp point, very gradually narrowed
at base to a short, more or less distinct petiole; midrib sharply prominent
above (in dried leaves sometimes located in a broad channel or fold),
broader below; margins revolute. Transverse sections of the leaves usually
show five vascular resin canals, the outer two of which are commonly found
in the transfusion tissue, interrupted upper hypoderm of small fibers, lower
hypodermal fibers absent between stomatal rows, vascular fibers above but
not below the midvein, auxiliary sclereids in the mesophyll usually quite
rare. Pollen cone buds solitary in the axils of upper leaves of new growth,
large, 4-5 mm. diameter, connate; scales broadly triangular, apex acute, as
long as bud. Female cones solitary | in the leaf axils; mermnelee divaricate,
11-15 mm. long; receptacle fleshy, 8-9 mm. long, of four decussate bracts,
of which only the two lower are fertile, subtended by two narrow bracts, 4
mm. long, just beneath its base. Seed elongated, rounded at the apex and
somewhat narrowed toward the base, 11 mm. long, 8 mm. broad.

DISTRIBUTION: On slopes in rain forests at 400-900 m. altitude in the
Solomon Islands.

450 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Solomon Islands. SAN CristoBAL IsLAND: Hinuahaoro, Brass 2881 (Type,
BM, BRI, tex Florin, Leiden Herbarium); Waimasi River, Walker BSIP. 254
(A, BRI, +K

Podocarpus salomoniensis is quite distinct from P. neriifolius which is
also found in the Solomon Islands. It has long, much narrower leaves
which are usually quite crowded, the receptacle has four fleshy scales in-
stead of two and the bracts are very close to its base. Transverse sections
of the leaves show five vascular resin canals. Wasscher (11) recognized a
relationship between P. deflexus Ridley and P. salomoniensis in the leaf
and fruit, but he distinguished the latter by its ““non-deflexed leaves, with
the midrib usually not channelled beneath.” In the leaf transverse sections
I find these species similar, except for the smaller hypodermal fibers, 10—26
uw in diameter, in P. salomoniensis.

18. Podocarpus deflexus Ridley, Fl. Malay Peninsula 5: 283. 1925;
Florin, Svenska Vet.-Akad. Handl. III. 10: 279. 1931; Wasscher,
Blumea 4: 427. 1941.

A small tree, 5-8 m. tall, the divaricate branches with stout branchlets
showing numerous leaf scars. Terminal buds large, globose, with the outer
scales narrow-triangular and reflexed, the inner scales almost triangular
and adpressed. Leaves densely crowded at the ends of the branches,
strongly deflexed, thick-coriaceous, rigid, linear or linear-lanceolate, 10-27
cm. long, 7-12 mm. broad, long-tapering to a shortly rounded (rarely
acute) apex, gradually narrowing at the base to a very short petiole; mid-
rib prominent above, broad and flat or deeply and broadly channelled be-
low; margins of the blade sometimes recurved. Transverse sections of the
leaves show 3—5 vascular resin canals, the center one often with the lumen
closed due to the channel below and the outer pair located in the trans-
fusion tissue; upper hypoderm of small groups of large fibers, 30-70 p in
diameter, and usually two fibers deep, no lower hypodermal fibers between
the stomatal rows, vascular sclereids above and more rarely below the
bundle, auxiliary sclereids in both palisade and mesophyll. Male cone buds
1—3 in the upper leaf axils, globose; mature cones unknown. Female cones
solitary in the upper leaf axils; peduncle thick, 9-15 mm. long; receptacle
large and fleshy, up to 8 mm. thick. Seeds obovoid, narrowed at the base,
11-12 mm. long, 8-9 mm. broad.

DIsTRIBUTION: On rocky slopes, 1650-2300 m. altitude, in Gunong
Tahan, Pahang.

Malay Peninsula. PAHANG: Gunong Tahan, Ridley 16024 (+k).

The leaves of both Podocarpus deflexus and P. salomoniensis are large,
crowded, and have five vascular resin canals, but there are several striking
differences. The former species has strongly deflexed leaves, which may be-
come longer, and are twice as wide, with the midrib channelled below. In
transverse section, the small groups of very large hypodermal fibers con-
trast with the larger groups of much smaller fibers in a neat single layer in

1958 | GRAY, REVISION OF PODOCARPUS, XI 451

P. salomoniensis. Podocarpus deflexus differs from P. neriifolius in its re-
flexed leaves, five vascular resin canals and the large upper hypodermal
fibers. These large hypodermal fibers of P. deflexus are most like those of
P. polystachyus and its relatives but the latter species has only three
vascular resin canals. Orr (9) also observed the interrupted upper hypo-
derm of large fibers of P. deflexus, the absence of lower hypodermal fibers
and the presence of auxiliary sclereids. He apparently did not see the five
vascular resin canals as he reported no such condition for any species of
Podocarpus. Florin (4) was impressed by the thick cuticle in this species.

19. Podocarpus annamiensis, sp. nov.

Arbor media, 5-12 m. alta; ramulis plerumque crassis, oppositis vel
verticillatis; gemmis terminalibus ovatis vel saepe globosis, 2.5-4 mm.
diametro, squamis exterioribus saepe quam alabastris longis, ad 5 mm.
longis; foliis apicibus ramulorum congestis, erectis vel tantum interdum
divaricatis, 4-10.5 cm. longis, 5-10 mm. latis, late linearibus vel lineari-
lanceolatis, apicibus abrupte angustatis, obtusis, orbiculatis vel acutis,
petiolo gradatim angustato, 2-6 mm. longo, marginibus interdum revo-
lutis: strobilis masculis ex alabastris solitariis, vel 2—3-fasciculatis, ses-
silibus, subglobosis, 1.5—3.0 mm. diametro, strobilis maturis ignotis; stro-
bilis femineis solitariis, pedunculo 2-10 mm. longo; semine ovoideo in-
ferne haud angustato, apice obtuso, apiculato vel orbiculare, 8-10 mm.
longo, 6 mm. lato.

DisTRIBUTION: Frequent summit tree of mountains in the coast range
of Annam and Cochin China in Indochina and nearby Hainan.

Indochina. ANNAM: Mt. Bana, Poilane 1561 (Type, a-2 sheets, +p—2 sheets) ;
Mt. Bana, 25 km. from Tourane, Clemens 3475 (+A, NY, TP, uc); Nhatrang Prov.,
Poilane 3541 (+p); Nhatrang Prov., Hoi Li, Chevalier 38692 (tp); Quang-tri
Prov., massif de Dong-co-pah, Poilane 3541 (a). CocHIN CHINA: in cacumine
montis Dink propé Baria, Pierre 354 (A-2 sheets, Ny—3 sheets) ; Biuh Biuh, Pierre
354 (+p—5 sheets); Chin’a Chiang, prov. Bienhoa, Pierre 5532 (+a, +p-2 sheets).
Burma: site unknown. Brandis 38 (+mMEL). Hainan: in mountain forests, Liang
63510 (a, +ny), Liang 65091 (a, tny, P), Liang 65554 (+Ny—juvenile), Liang
65555 (NY), C. Wang 35031 (A, NY); Ng Chi Leng, Fan Yah, Chun & Tso 44217
(a, try); Lingshiu, How 73776 (ta).

The trunks of the trees are usually 0.5-1 m. in diameter with dark
brown bark. The outer scales of the terminal bud are broadly triangular,
keeled, stiff, erect with apex acute and apiculate or acuminate; the inner
scales are shorter with rounded or minutely apiculate apex. The leaves
are usually straight but are sometimes falcate, and rarely they may be up
to 18 cm. long and 20 mm. wide; the midrib is from broadly prominent to
flat above, rarely narrowly prominent, and below it may be broadly prom-
inent, flat or even channelled. The pollen cone buds are found in the axils
of the lower leaves of the new growth and the bud scales are tightly ap-
pressed, broadly triangular, stiff, with obtuse to acute apices and scarious
margins, strongly keeled. The young cones are greenish white. The female

452 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

cones are axillary and widely scattered; receptacle 4-8 mm. long, of two
fused fleshy almost equal scales with free tips.

Podocarpus annamiensis differs from P. neriifolius, which is found in
nearby regions, in the smaller stiff leaves which are crowded toward the
ends of the twigs and in the abruptly obtuse or acute apex of these leaves.
The habit and shape of the leaves is much like that of P. macrophyllus
but it differs from this species (see no. 29) in the leaf anatomy. The
stiff leaves also suggest P. polystachyus, but this is a tree of the strand
rather than of the mountains, and the pollen cones occur in clusters of as
many as five.

Transverse sections of the leaves of Podocarpus annamiensis show three
vascular resin canals with the median one usually very small and the
lateral very large and conspicuous, vascular fibers both above and below
the bundle, upper hypodermal fibers 20-40 » in diameter, very rare or in
small groups and always interrupted at the midrib, hypodermal fibers
absent between the stomatal rows, auxiliary sclereids in the mesophyll not
abundant, cuticle thick. The two large lateral vascular resin canals dis-
tinguish transverse sections of this species from the three mentioned above.
It is also distinguished from P. macrophyllus by the auxiliary sclereids
and usually only one layer of palisade mesophyll.

The specimens listed from Annam and Cochin China are all quite sim-
ilar. Poilane 11121 with falcate narrower leaves deviates the most from
the type specimen. Upper hypodermal fibers are quite rare in Pierre 354
and abundant in Poilane 3541. Auxiliary sclereids have not been ob-
served in the mesophyll of leaves from Pierre 354. Arnold Arboretum and
New York Botanical Garden herbarium specimens of Pierre 354 bear
labels giving the site as Mt. Dink near Baria but the Paris Museum speci-
mens give the site as Mt. Biuh Biuh.

The smaller-leaved specimens from Hainan have a more sharply acute
apex to the leaves. Transverse sections of the leaves show abundant hypo-
derm and thick palisade which is often in more than one layer. The large-
leaved specimens from the same locality have auxiliary sclereids only on
the lower side of the leaf. Herbarium labels of these specimens indicate
larger trees than those in Annam, reaching 16 m. in height and 3 m. in
diameter.

20. Podocarpus archboldii, sp. nov.

Ramuli aliquando crassi, verticillati; alabastris terminalibus magnis
globosis, squamis exterioribus late triangularibus ad 6 mm. longis, apicibus
obtusis, recurvatis, marginibus tenuibus scariosis; foliis 3-9.5 cm. longis,
5—11 mm. latis, apieibus acutis vel acuminatis, marginibus revolutis; stro-
bilis masculis 3-4 mm. latis; semine sphaeroideo, sine crista distincta.

A tree 6-38 m. high, 60-90 cm. in diameter with gray scaly bark 2-4 mm.
thick and shallowly fissured. Leaves crowded, erect to patent, coriaceous,
narrowly lanceolate, almost sessile or narrowed to a petiole 2 mm. long,
green above, paler and brownish beneath; midrib narrowly prominent

1958 | GRAY, REVISION OF PODOCARPUS, XI 453

above, broader below. Pollen cone buds axillary, solitary, sessile or on short
2-3 mm. peduncles, large, globose, with obtuse scales. Female cones
on stout peduncles 3-14 mm. long, solitary, axillary; receptacle of several
thick fleshy fused scales, 7-8 mm. long, subtended by 2 deciduous bracts
2-3 mm. long. Seeds sometimes up to 15 mm. long, 13 mm. wide.

DIstRIBUTION: Occasional trees in high primary rain-forests of New
Guinea, 1820-3000 m. altitude, more common at the higher altitudes.

New Guinea. NETHERLANDS New Guinea: Idenburg River, 4 km. SW of
Bernhard Camp, Brass 13121 (Type, 2 +A, LAE); 15 km. SW of Bernhard Camp,
Brass 11946 (@ +a, LAE); 18 km. SW of Bernhard Camp, Brass 11946A (Sa,
+LAE); 9 km. NE of Lake Habbema, 2800 m. Camp, Brass & Versteegh 10452

+a). TERRITORY OF NEw GuINEA. Morobe District: Samanging, Clemens
9503 (+a), 9525 (+a), Mt. Sarawaket, Clemens 10073 (+a). Eastern Highlands
District: above Goroka, J. Leahey’s logging area, Womersley & Floyd NGF6137
(A, +LAE); Wai, High country, Jacobs NGF53 (+LazE). Papua. Central Division:
Wharton Range, Murray Pass, Brass 4605 (+a, BRI), Mt. Tafa, Brass 5118
(tny).

In New Guinea this species is distinguished from Podocarpus neriifolius
by smaller leaves with revolute margins and from P. pilgeri by larger
leaves with gradually acute or acuminate tips. The large globose terminal
buds distinguish it from both species.

Transverse sections of the leaves show three vascular resin canals, an
interrupted upper hypoderm of fibers of medium diameter, hypodermal
fibers rarely between the stomatal rows, vascular fibers both above and
below the bundle, and usually no auxiliary sclereids in the mesophyll. Only
Jacobs NGF53 and Brass & Versteegh 10452 show rare auxiliary sclereids.

Clemens indicated that his specimen 10073 from Mt. Sarawaket had red
fruit. Brass & Versteegh 10452 had a seed of larger size, 15 mm. long and
13 mm. wide.

This species is dedicated to Mr. Richard Archbold who sponsored the
expedition to New Guinea on which this species was first found.

20a. Podocarpus archboldii var. crassiramosus, var. nov.

Arbor gracilis, 8-20 m. alta; ramulis crassissimis, alabastris terminalibus
magnis, globosis vel globosi-ovatis, supra 5 mm. diametro; foliis congestis
tenue lanceolatis, apicibus longe acutis, interdum acuminatis 6-13 cm.
longis, 7-11.5 mm. latis; strobilis masculis ignotis; strobilis femineis soli-
tariis, pedunculo 6-11 mm. longo; semine globoso-elliptico, ad 10 mm.
longo et 9 mm. lato, apice rotundato vel obtuse apiculato.

Outer scales of terminal buds triangular, keeled, erect, to 9 mm. long,
with long acute or acuminate tips which are sometimes recurved. Leaves
divaricate, straight or falcate, gradually narrowing to short, thick petioles
at the base and with long, acute tips; midrib narrowly prominent above,
flat or even channelled below. Female cones axillary among lower leaves
of new growth; receptacle of 2—3 fused fleshy scales with free tips, 6-8 mm.
long, and subtended by 2 small subulate bracts.

454 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

DistTrIBuTION: In high mountain forests of New Guinea, 1850-2650 m.
altitude

New Guinea. NETHERLANDS NEW GuINEA: 9 km. NE of Lake Habbema,
Brass 10874 (Type, +A, LAE); Wissel Lake Region, s. of border of Lake Paniae,
foot of Mt. Poti, Eyma 4538 (+a). Territory or New GuINeA. Western High-
lands District: Al River mountains, Nondugl, Womersley NGF5361 (+A, LAE).
Eastern Highlands District: Chimbu, Cavanaugh NGF3329 (+tar). Morobe
District: Oberamnang, Clemens 5434 (+a); Sattelberg, Clemens 2276 (a); Mongi
valley, below Salawaket, Lane-Poole 524 (+Br1

This variety differs from the species in the very thick twigs and widely
spreading leaves which do not have revolute margins. The hypodermal
fibers are somewhat larger. Transverse sections of the leaves of Cavanaugh
NGF3329 and Lane-Poole 529 have three very large vascular canals and
the latter specimen has two smaller ones in addition. Auxiliary sclereids
were seen only in Clemens 5434 and Eyma 4538.

21. Podocarpus nakaii Hayata, Ic. Pl. Formosa 6: 66. 1916; Pilger,
Nat. Pflanzenfam. ed. 2. 13: 248. 1926; Florin, Svenska Vet.-Akad.
Handl. IIT. 10: 279. 1931.

A tree with terete, glabrous branchlets. Terminal buds globose, outer

Leaves crowded at the ends of twigs, spreading, linear-lanceolate or
linear, 5-10 cm. long, 8-11 mm. broad (sometimes up to 14 mm. broad),
straight or falcate, shiny above, paler beneath, the apex acute or shortly
attenuate, at the base gradually narrowing to petioles 5 mm. long; midrib
broadly prominent or flat above, rarely narrow, very broad below. Trans-
verse sections of the leaves show 3 vascular resin canals (sometimes ob-
scure), upper hypoderm of few and scattered small fibers 13-28 p in
diameter interrupted at the midrib, no hypodermal fibers between the
stomatal rows, vascular sclereids absent above and rare below the bundle,
auxiliary sclereids absent. Pollen cones unknown. Ovulate cones solitary
in the leaf axils; peduncles 2-7 mm. long; receptacle fleshy and subtended
by two thick triangular bracts 1.5 mm. long, 1 mm. wide. Seed oblique-
globose, crested, 1 cm. long, 8 mm. thick.

DistRIBUTION: Known only from Formosa on mountain slopes.

Formosa: Hassen-zan, Kanehira 21184 (a, tuc, UCLA); Sui-sya, Kanehira 740
(tuc); Rengeti, Hayashi 21208 (a, Ny, tuc), Kanehira 316 (a); Nanto, Tahiken,
Wilson 9935 (a-5 sheets, tus), Wilson in 1918 (p); above Shushu, Wilson 10023
(a-3 sheets); around Honsha, Wilson 9937 (a-3 sheets); Shinchiku Prov.,
forests beyond Nanoaherzan, Wilson 10319 (+A).

Podocarpus nakaii, according to the description, differs from P. nerii-
folius by the very short (2-4 mm.) peduncle of the fruit and the short,
thick triangular bracts subtending the receptacle; and I find the terminal
buds have shorter acute scales. The peduncles on Kanehira 740, however,
are 7 mm. long. Transverse sections of the leaves show thicker palisade

1958] GRAY, REVISION OF PODOCARPUS, XI 455

than is found in P. neriifolius. As Orr (9) suggests, the species is not
readily distinguished from P. macrophyllus by means of leaf anatomy.

22

Bo

Podocarpus rumphii Blume, Rumphia 3: 214. 1847; Gordon, Pine-
tum 282. 1858, ed. 2. 346. 1875; Carriére, Traité Conif. ed. 2. 663
1867; Parlatore, DC. Prodr. 16: 515. 1868; Bertrand, Ann. Sci. Nat.
V. 20: 59. 1874; Van Tieghem, Bull. Soc. Bot. France 38: 169. 1891;
Warburg, Monsunia 1: 193. 1900; Pilger, Pflanzenreich IV. 5(Heft
18): 81. 1903, Bot. Jahrb. 54: 210. 1916, Nat. Pflanzenfam. ed. 2.
13: 248. 1926; Foxworthy, Philip. Jour. Sci. 6: 164. 1911; Dallimore
& Jackson, Handb. Conif. 55. 1923, 1931, 80. 1948; Wasscher,
Blumea 4: 432. 1941,

A tree 20-33 m. high with straight branches and subverticillate twigs.
Terminal buds on the twigs globose or ovate, with thick, keeled outer scales
which are acute or acuminate, rarely obtuse. Leaves erect or patent,
coriaceous, linear-lanceolate, straight or subfalcate, more or less abruptly
narrowing to the short-angustate (rarely caudate-acuminate) apex, nar-
rowing abniptly to the short thick petiole, 6-25 cm. long, 10-29 mm. broad,
margins parallel or nearly so; midrib broadly prominent above or scarcely
evident, less so below. Transverse sections of the leaves show 3 vascular
resin canals, interrupted upper hypoderm with fairly small fibers, no hypo-
dermal fibers between the stomatal rows below, almost always vascular
fibers but rarely auxiliary sclereids. Male cones axillary, solitary or clus-
tered 1-3, sessile, or on short 1.5—-3 mm. peduncles, strobili to 4 cm. long,
3 mm, in diameter. Microsporophylls nearly ovate-triangular, apiculate.
Female strobili solitary, axillary; peduncle 2-16 mm. long; receptacle of
2-4 fused fleshy scales with narrow, obtuse free apex, subtended by 2 sub-
ulate bracts, 6-10 mm. long, 3.5-8 mm. in diameter. Seeds 1 or 2, globose-
ellipsoid, grayish when ripe, 10-13 mm. long, not crested.

DISTRIBUTION: Usually a mountain tree in forests of Borneo, New
Guinea and some of the smaller islands in this region.

Borneo. British NortH Borneo. Mt. Kinabalu: Gurulau Spur, Clemens
50691 (a, tBm, tuc); Penibukan Ridge, Clemens 50051 (a, uc), For. Dept.
2174, Apostal sn. (uc). Sarawak: Bidi Cave, near top of mountain, Clemens
20656 (+Nv¥). SouTHEAST BorNEO: Sampit, Buwalda 7793 (+A), peak of Balik-
papan, Kostermans 7408 (+LAE). JAMBONGAN IsLAND: Sanakan, Cabiling, For.
Dept. 3710 (uc).

Moluccas. Weda, Weda, Anon. Boschpr. bb24924 (a). Mororat: G. Pare 2,
Kostermans 1210 (a, LAE). ARoE ISLANDS: Dosinamalaoe, P. Kobroor, Boschpr.
bb 25289 (A), Buwalda 4988 (+a); Selibatabat, P. Wokam, Boschpr. bb 25415
(+a), Buwalda 5271 (+A); Wakatoebi, P. Oedjir, Boschpr. bb25438 (+a). Am-
BOINA: Robinson 309 (Ny).

New Guinea. TERRITORY oF New GurneA. Sepik District: Aitape, Smth
NGF1241 (prt, +LAE—2 sheets). Eastern Highlands District: Aiyura, Smith
NGF1102 (sri, LAE); Tumoma River, Masters 1341 (+prr); Chimbu, Cava-
naugh NGF3336 (+a). Morobe District: Matap, Clemens 11133 (+a). Papua.
Central Division: Rona, Brass 6208 (+A); Koitaki, Carr 12842 (+a). Gulf

456 JOURNAL OF THE ARNOLD ARBORETUM | [voL. xxxrx

Division: Kikori River delta, Hart NGF4545 eg 2 sheets). Milne Bay Dis-
trict: Misima Island, Schacht NGF2762 (+BRI, L

Solomon Islands. SAN CrIsToBAL: ridge forest, “Lovie BolPsot -CUae),

The description for this species was taken from the specimens examined
and from Pilger (1903). These specimens all show leaves with a broad,
scarcely prominent midrib and abruptly acute apex which distinguish
Podocarpus rumphii from P. neriifolius; I find that these characters also
distinguish it from P. ledermannii. Wasscher (11) included in P. rumphii
specimens of P. philippinensis and some specimens from Celebes which I
am excluding from P. rumphit.

The male specimen, Buwalda 7795, has solitary sessile, globose male
cone buds. The seeds on the female specimens are very immature.
Wasscher includes Clemens 50691 and 50051 under P. neriifolius, but the
broad midrib and the leaf anatomy are so like Buwalda 7795 that these
should both be considered as P. rumphii. Cabiling s.n. has a narrowly
prominent midrib and caudate-acuminate apex on the leaves, but their very
large size, 19.3 cm, long and 24 mm. wide, suggest that it is a juvenile
example of this species.

Orr (9) examined the leaf anatomy of Podocarpus rumphii in transverse
section and described lower hypodermal fibers between the stomatal rows.
I do not know what specimens he used, but I found a few fibers only in
Smith NGF1102.

Wasscher’s description excluded caudate-acuminate apices of the leaves,
but I find that some specimens which have such leaves must be included
in P. rumphii.

The specimens listed from the Moluccas have large, broad leaves with
nearly parallel margins and tips abruptly acute or acuminate. The mid-
rib is broadly or scarcely prominent above. One of the specimens with
seeds, Buwalda 4988, has peduncles 6-16 mm., longer than the type, and
the seed is up to 13 mm. long. The male cones are solitary or clustered
and sessile or short peduncles in Kostermans 1210 and Buwalda 5271.

Most of the specimens from New Guinea had been tentatively ascribed
to Podocarpus neriifolius, but the very broadly prominent upper midrib
and the parallel margins of the leaves proved to be satisfactory criteria for
their inclusion in P. rumphii. The foliage on these specimens is usually
much larger.

Smith 1102, from Aiyura, is a sterile specimen with very large drooping
leaves clustered at the tips of the twigs and reminds one of Podocarpus
deflexus from the Malay peninsula. Transverse sections of the leaves
show 5 resin canals below the vascular bundle but the hypodermal fibers
are smaller, there are no auxiliary sclereids and the lower hypodermal
fibers are fairly numerous; these are not characteristics of P. deflexus.

23. Podocarpus costalis C. Presl, Epimel. Bot. 236. 1849; Pilger, Pflan-
zenreich IV. 5(Heft 18): 78. 1903, Pflanzenfam. ed. 2. 13: 248. 1926;
Florin, Svenska Vet.-Akad. Handl. III. 10: 279. 1931; Dallimore &
Jackson, Handb. Conif. 42. 1931, 64. 1948.

1958 | GRAY, REVISION OF PODOCARPUS, XI 457

Tree or shrub (?) with thick, short, densely leafy spreading branch-
lets. Terminal buds short, 2—-2.5 mm. long, ovate with thick, stiff, keeled,
closely appressed scales with obtuse apex, as long as the bud. Leaves
erect to patent, crowded, coriaceous, oblanceolate, sometimes rounded, ob-
tuse, or even emarginate at the apex, narrowing rather gradually from
above the middle to a short, thick petiole, 3.8-7.0 cm. long, 7-10 mm.
broad; midrib broadly prominent above, rarely embedded in shallow
grooves, below broad and scarcely evident or shallowly impressed. Trans-
verse sections of the leaves show uniformly three vascular resin canals,
interrupted upper hypoderm, no hypodermal fibers between the stomatal
ows and no auxiliary sclereids. Male cone buds subglobose, 2—2.5 mm.,
solitary, sessile, with obtuse-rotundate outer scales; mature cone thick
cylindric, 3 cm. long. Female cones solitary, axillary; peduncles 2 mm.
long; receptacle of two fused fleshy equal scales with rounded tips, 7 mm.
long, subtended by two very minute bracts. Seed elliptical, obtusely
crested at the tip, dark when dry, 7-9 mm. long.

DIsTRIBUTION: Shores of several of the Philippine Islands and Formosa.

Philippine Islands. BapuyANn IsLanps: Dalupiri, Bartlett 15138 (A), 15192
(+a). Batanes IsLanps: Mahatow, For. Bur. 80397, Ramos (tny). POLILLo
IsLAND: For. Bur. 29682, Salvoza (+Ny, Uc). Bucas Istanp: For. Bur. 5268,
Merrill (+Nv). Formosa: Isl. Koto-syo, Mori 315 (+a).

The description of Podocarpus costalis was drawn up from the specimens
cited and from Pilger’s description (1903). The only likely suggestion as
to the locality of the original collection by Haenke is Luzon, but I have seen
no other collections of this species from there. The only specimen jee
here which has been examined previously is For. Bur. 5268, Merrill;
was placed in P. polystachyus by both Foxworthy and W econ in ce
of the spatulate shape of the leaves. Foxworthy’s description (5) under
P. costalis is referred to P. pilgeri since all of the specimens he used are
cited under that species, no doubt rightly, by Wasscher (11).

Podocarpus costalis differs from P. polystachyus in the thick, spatulate
leaves which are usually erect on the twigs. Transverse sections of the
leaves show more abundant upper hypodermal fibers of smaller diameter
in P. costalis. Orr (9) included P. costalis with those species having lower
hypodermal fibers but I did not ae these in any case. Both species are
found at sea level in coastal are

The Formosan specimen (Mori 315) differs from the others in having
vascular fibers.

24. Podocarpus thevetiifolius Zippel, Flora 12: 287. 1829 Cee.
Blume, Rumphia 3: 213. 1847; Carriére, Traité Conif. ed. 2.
1867; Parlatore, DC. Prodr. 16: 518. 1868; Gordon, Pinetum Pf a
349, 1875: Warburg, Monsunia 1: 192. 1900: Pilger, Pflanzenreich
IV. 5 (Heft 18): 79. 1903, Bot. Jahrb. 54: 210. 1916, Nat. Pflanzen-
fam. ed. 2, 13: 248. 1926; Dallimore & Jackson, Handb. Conif. 56.

458 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

1923, 1931, 83. 1948; Florin, rae eg -Akad. Handl. III. 10:
280. 1931; ‘Wasscher, Blanen: 4: 462.

A small tree, 10-23 m. tall, with numerous, — sometimes oppo-
site spreading branches. Vegetative buds small, ovate-acute; bud scales
adpressed, acute, 1.5 mm. long. Leaves scattered below, but closely
crowded near ends of twigs, thin-coriaceous, lanceolate, acute or obtuse
at the apex, rarely mucronate, gradually narrowing to a short petiole;
midrib broad, flat, not prominent above, scarcely prominent below; mar-
gins of blades not revolute, flat and [fide Wasscher (11)] with a distinct
narrow shining line along the margins, 2.5-8 cm. long by 5—9 mm. wide.
Transverse sections of the leaves show an interrupted upper hypoderm of
very large fibers, no hypoderm fibers between the stomatal rows, vascular
fibers usually present above the bundle, auxiliary sclereids absent in the
mesophyll. Pollen cones unknown. Female cones axillary, solitary; pedun-
cles slender, 3-8 mm. long; receptacle | fide Blume (1847) ] twice as thick
as the seed. Seed ellipsoid, 10 mm. long.

DISTRIBUTION: New Guinea.

New Guinea. NETHERLANDS New GuINEA: Lobo, Zippel s.n. (+ex Florin).
Papua. Northern Division: Isuarava, Carr 15395 (+a \

This species may be distinguished from those closely related to it by
the small, thinner, flat leaves without prominent midrib above and the
small terminal buds with very short scales. The margins of the leaves are
not revolute as in Podocarpus archboldii. In both specimens the diameter
of the hypodermal fibers is large, sometimes up to 70 p.

25. Podocarpus forrestii Craib & W. W. Smith, Notes Bot. Gard. Edin-
burgh 12: 219, 1920; Dallimore & Jackson, Handb. Conif. 46. 1923,
1931, 69. 1948; Florin, Svenska Vet.-Akad. Handl. IIT. 10: 279. 1931.

A shrub up to 3.5 m. high with fairly stout branches. Vegetative buds
small, ovate. Leaves 3.5-5.4 cm. long, 6.5—9 mm. wide, oblong or oblong-
lanceolate, obtuse or rounded at the apex, gradually narrowing at the
base into a short, winged petiole, dark green above, pale beneath; midrib
broadly prominent above. Pollen cones unknown. Female cones usually
solitary in the leaf axils, pedicels 8 mm. long, ovule on a short fleshy re-
ceptacle. Mature seeds unknown.

DisTRIBUTION: Eastern and western sides of the Tali Range in western
ina

China. YUNNAN: Tali Range, Forrest 6852 (+).

Podocarpus forrestii is distinguished from P. macrophyllus var. maki by
its dwarf habit and shorter, broader leaves. The upper midrib is not as
abruptly prominent. I have examined the transverse section of leaves from
only one specimen and I find that the leaf anatomy is very similar to
that of P. macrophyllus, its var. maki, and P. nakaii. The interrupted

1958] GRAY, REVISION OF PODOCARPUS, XI 459

upper hypoderm is of small fibers; rarely are there any vascular fibers,
and I found no auxiliary sclereids. As the few differences seem to be
only a matter of degree, it may well be that these are all forms of
P. macrophyllus.

26. Podocarpus pilgeri Foxworthy, Philip. Jour. Sci. Bot. 2: 259. 1907,
6: 149. 1911; Pilger, Nat. Pflanzenfam. ed. 2. 13: 248. 1926; Florin,
Svenska Vet.- "Akad. Handl. III. 10: 280. 1931; Dallimore & Jackson,
Handb. Conif. 54. 1931, 79. 1948; Wasscher, Blumea 4: 463. 1941.

Podocarpus celebica Warburg, Monsunia 1: os 1900, non Hemsley 1896;
Pilger, Pflanzenreich IV. 5 (Heft 18): 78. cE
BOC CUE Se ee Pilger, Bot. Jahrb. a 209. 1916, Nat. Pflanzenfam.
: . 1926; Florin, Svenska Vet.-Akad. Handl. III. 10: 280. 1931;
Wascehon ae 4: 463. 1941
Podocarpus costalis auct. non Pilger, Foxworthy, Philip. Jour. Sci. Bot. 6:
161. 1911.

A small tree or shrub, 2-15 m. tall, rarely taller, with scattered or
verticillate branches. Terminal buds ovate-acute with narrowly triangu-
lar scales, sometimes long acuminate, keeled, to 4.5 mm. long. Leaves
scattered or crowded near the tips of twigs, spreading, usually thick coria-
ceous, rigid, usually flat, linear-lanceolate to oblong, cuneately or more
gradually narrowed to the short petiole, abruptly or rather gradually nar-
rowed to an acute apex, sometimes apiculate, 1.5-8 cm. long, 4-13 mm.
broad; midrib narrow and sharply prominent above, keeled, flat, or even
channelled below. Transverse sections of the leaves show 1—3 vascular
resin canals, the central one sometimes being larger; the upper hypoderm
is usually of large isolated fibers, sometimes several in a group and rare
fibers on the lower side between the stomatal rows; vascular sclereids or
fibers are usually present. Male cones solitary, peillany. subsessile, scales
ovate-triangular and acute, cylindrical, 1.5—5 cm. long, 2-4 mm. in diam-
eter; microsporophylls broadly triangular, apiculate with a scarious mar-
gin. Female cones solitary, axillary; peduncles 3-12 mm. long; receptacle
of two oe fleshy bracts, obtuse and free at the tips, 5— 12 mm. long,
3-7 mm. thick, subtended by a pair of subulate bracts 1.5—2 mm. long!
Seed Reece obtuse, 8-8.5 mm. long, 7 mm. broad.

DistRIBUTION: Mountains between 1400-3000 m. altitude on islands
from the Philippines to the Solomons, and one collection from Siam.

Philippine Islands. Luzon: Tayabas, Mt. pee: meee Bur. Sci. 2393
(tny), Gates 7254 (+F), Copeland s.n. (uC); ban, Elmer 7778 (a, to,
NY); ues ae in 1914 (+uc). Murnporo: ae "Haleon, Merrill 5754 (+tNy,
isotype). Gros: Canlaon Volcano, Merrill 241 (+a). Mrinpanao: Misamis
Prov., Mt. cere Mearns & Hutchinson, Bur. Sci. 4673 (+BR, ¢NY) ; Agusan
Prov., Cababaran, Mt. Urdaneta, Elmer 14086 (A, F, GH, ny, tuc); Bukid-
non Subprov., Mt. Lipa, Ramos é& Edand, Bur. Sci. 38500

Celebes. Gowa, Lembaja, Boroe, Boschpr. bb20437 (A).

New Guinea. NETHERLANDS New GuINEA: Idenburg River, Bernhard Camp,

460 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxix

Brass & Versteegh 13519A (+A-2 sheets); NE Lake Habbema, Bele River,
Brass 11341 (+a, LAE); Subdistr. Manokwari, summit of Arfak Mountains,
Vogelkop, Angi gita Lake, Kostermans 2161 (+A), 2236 (+A), 2519 (+a).
TERRITORY OF NEw GUINEA. Western Highlands District: Wahgi-Sepik Divide,
Womersley & Millar NGF6980 (a, LAE); Wahgi-Jimmi Divide, Womersley
NGF5316 (A, +LaE); Mt. Hagen, Cavanaugh NGF3324 (+LAE). Bi riaile Moun-
tains, Schlechter 18780 (+Br, tuc). Morobe District: Wau-Mubo Road near
Skindewai, Womersley & Millar NGF8341 (a, LAE); NE Oberamnang, Clemens
4569 (A), 4696 (ta); Edie Creek, Womersley NGF5373 (A, +LAE); above Wau,
McAdam NGF440 (srt, +LAE). Papua. Central Division: Mt. Tafa, Brass 4034
(+Ny); above the Gap, Carr 13721 (+a); Boridi, Carr 14563 (a), 14556 (ta);
Mt. Obree, Lane-Poole 357A (+Br1). Milne Bay District: Maneau Range, north
slopes of Mt. Dayman, Brass 22811 (a).

Solomon Islands. SANTA IsABEL IsLaAnp: Brass 3265 (+A). SAN CRISTOBAL
IsLAND: on ridge, Logie NGF354 (Lag).

Siam. Kao Knap, Krat, Kerr 17809 (+m),

This description has been limited to those specimens obtained from the
higher altitudes. Therefore it includes all of the specimens from the
Philippine Islands listed by Foxworthy (5) as P. costalis and which were
later referred to P. pilgeri Foxw. by Wasscher (11). The name P. schlech-
tert Pilger is still being retained by some workers for specimens with
small, linear-lanceolate leaves which are pointed at both ends. There is so
much variation, however, in the foliage of P. pilgeri, even within the
same specimen, that these must be included here.

Kerr 17809, from Siam, has exceedingly immature fruit, but the foliage
is so like that of many specimens of Podocarpus pilgeri that I refer it to
that species. It is said to be a small tree, 1.8 m. high, not uncommon in
high evergreen forest. Transverse sections of the leaves show more abun-
dant upper hypoderm fibers than most other specimens. This specimen
and those from New Guinea often show a few auxiliary sclereids in the
mesophyll, but these are lacking in the Philippine specimens.

27. Podocarpus neriifolius D. Don ex Lamb. Pinus ed. 1. 1: 21, 1824,
ed. 2. 2: 122. 1828 (in part); Endlicher, Syn. Conif. 215, 1847;
Gordon, Pinetum ed. 1. 279. 1858, ed. 2. 343. 1875; Carriére, Traité
Conif. ed. 2. 661. 1867; Parlatore, DC. Prodr. 16: 514. 1868; Bert-
rand, Ann. Sci. Nat. V. 20: 59. 1874; Van Tieghem, Bull. Soc. Bot.
France 38: 169. 1891; Kent in Veitch, Man. Conif. 152. 1900; Pilger,
Pflanzenreich IV. 5(Heft 18): 80. 1903, Bot. Jahrb. 54: 210. 1916.
Nat. Pflanzenfam. ed. 2.13: 248. 1926; Bernard, Beih. Bot. Centralbl.
17: 293. 1904; Foxworthy, Philip. Jour. Sci. 2: 258. 1907, 6: 162.
1911 . Dallimore & Jackson, Handb. Conif. 52. 1923, 1931, 77. 1948;
Wilson, Jour. Arn. Arb. 7: 41. 1926; Florin, Sienska Vet. -Akad.
Handl. II. 10: 279. 1931; Merrill, Contr. Arnold Arb. 8: 15. 1934;
Wasscher, Blumea 4: 437. 1941

Myrica esquirolii Leveillé in Fedde, Rep. Sp. Noy. 12: 537. 1913; Rehder,
Jour. Arnold Arb, 10: 108. 1936.

1958] GRAY, REVISION OF PODOCARPUS, XI 461

Podocarpus etter eee Enum. Pl. Javae 88. 1827-8; Van Tieghem, Bull.
Soc ance 38: 169. 1891.

Podocarpus aie Blume, Rumphia 3: 213

Podocarpus junghuhniana ee Pl. are ‘i 2. 1851; Van Tieghem,
Bull. Soc. Bot. France 38: 169. 1891.

Podocarpus te Blume, Bai 3: 214. 1847; Van Tieghem, Bull.
Soc : nce 38: 169. 1891

Podocarpus Eat var. perenne Pritzl, Bot. Jahrb. 29: 213. 1900.

erie nerifolius var. brevipes Pilger, Pianzenreich IV. 5(Heft 18): 81.

avi neglecta Blume, Rumphia 3: 213. 1847.

A medium sized tree up to 40 m. tall (rarely larger) with very spreading
branches and numerous branchlets. Terminal buds ovate with ovate-trian-
gular to long-subulate outer scales usually as long as or longer than the
bud, 5-7 mm. long. Leaves scattered, usually large, 7-15 cm. long,
9-13 mm. wide (sometimes longer and up to 19 mm. wide), straight or
falcate, spreading, usually very gradually narrowing to an acute apex, and
less gradually to a short petiole; midrib narrowly prominent above, broad
and prominent below. Transverse sections of the leaves show 3 vascular
resin canals below the bundle, vascular sclereids or fibers usually present
above the bundle and more rarely below; no auxiliary sclereids, the inter-
rupted upper hypoderm in small groups of from 1-7 fibers, 20-40 p» in
diameter, no hypodermal fibers between stomatal rows, palisade mesophyll
of a pele layer. Pollen cones solitary, sessile, axillary: ovate, large, with
usually thick, ae acute outer scales, the inner scales thinner and
scarious, mature cones 2—8.5 cm. long, 2.5—4.5 mm. in diameter. Micro-
sporophylls narrow, short, acute or obtuse, often apiculate. Female cones
solitary in leaf axils, usually remote, peduncles 3-24 mm. long; the fleshy
receptacle subtended by two subulate bracts 2-6 mm. long. Seeds 9-16 mm.
ee narrow ovoid, sometimes globose but gradually narrowed toward
the

DISTRIBUTION: Subtropical evergreen forest, 650-1300 m. altitude in
the Himalayas, eastward into China, and south into the Malay Peninsula;
on insular lands from Japan south into Java, west to Sumatra and the
Andaman Islands, eastward as far as the Fiji Islands.

India. NepaL: Wallich 6052A (Type, tpM, tBr, tny, +P), Wallich in 1818
(MEL, MO), Anon. 1819 (BR), Scheidweiler sn. (+BR), Lambert 45 (+r),
ee in 1819 (+Br). AssAM: Khasia, Jowae, Clarke 18362 (+a); Jaiutea
Hills, Mann s.n. (A); tropical region, Hooker & Thomson in 1856 (+MEL, PH);
Sibebiun (?), Thomson sn. (GH); Masters sn. (+P, tpH); oe St. (TP).
East BENGAL: Griffith 5006 (GH), Clarke 19721 (MEL). North Burma: Ngaw-
chang Valley, N of Htawgaw, Ward 173 (+a, NY); between aoe and
Lanyang, Ward 311 (Ny).

ey Western China, Wee ous (A). SzEcCHWAN: Mt. Omei, Fang 2346
(A, tny, P), Faber 985 (Ny). EICHOW: Tsunyi, Cheng 5317 (+Ny); Lo,
ees 3463 (4a), 3465 (+P), pete 3223 (ta, te). CHEKIANG: south of
Ping Yung, R. C. Ching 1982 (+a, GH, tuc). YUNNAN: Seyemeo Ulo, Henry

462 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. XxxIx

12919 (a, +Mo, Ny-2 sheets). Kwancsi: Me-kon, Seh-feng Dar Shan, 5. Nan
ning, R. C. Ching 8417 aa sheets, GH, tUC); Shang- sze Dist., Shap Man Taai
Shan, as 24761 (Mo, N

Siam: ngkok, Kerr sm, (+pmM); Kao Sem, Kouate, Kerr 9907 (+BM) ;
Klawry - Kerr 14586 (+3m); Kao Knap, Krat, Kerr 17833 (+BM); Garrett
538 (+Bm). Indochina. ANNAM: Tourane, Clemens in 1927 (tuc); Bordeneuri

Pierre 5532 (A).

Malaya. Perak: Wray 2922 (a). Pawanc: Fraser Hill, Anon, 11024 (+a).
PENANG: Government Hill, Haniff 334 (+stsu, Ny), Curtis 3079 (A); Balik
Pulau, Hu 9422 (a); Tulo, Glandoger in 1906 (+mo

Japan. Kyusuu: Urumai Prov., Higashikirishima, Wilson 6210 (A-2 sheets) ;
no locality, Sargent in 1892 (A). Tukin Islands: Mt. Genka, Kunchon, Wilson
8153 (A). Formosa: Shinchiku Prov. beyond Nauvaheizan, Wilson 10319 (A);
Lake Candidius, Kanehira 21309 (A, uc). Philippine Islands. Luzon: Benguet
Mt. St. Tomas., For. Bur. 31479, Esquerra (+Ny); Bontoc, Mt. Data, Clemens
16251a (tcas, uC), For. Bur. 10894, Curran (Ny), For. Bur. 14422, Darling
(xy); Tayabas, Mt. Binuang, For. Bur. 28635, Ramos & Edano (A). PoLtLto:
McGregor 10779 (+Ny).

South Andaman: Anderson 26, Kerr s.n. (+P); Dr. King 208 (A). Sumatra:
NW Lake Toba near Piso Piso, Bangham 1116 (+A, NY); road from east coast
to Tapanoeli, Bangham 1128 (A, +NY); Tapanoeli, Sipirok, Panobasan, Dk.
Poehoeten Lajan, van Steenis Boschpr. bb 30986 (Mo). Mentawei Islands: Pulau
Siberoet, Sebai-bai, Anon. Boschpr. bb 17444 (A). Borneo: southern part, Kor-
thals s.m. (+MEL).

Java. Prov. ag ties above Tjibodas, Christopherson 156 (+BIsH), Koor-
ders 12438 (a); Pangentjongan, forest Pasir Kajoejoetan, Koorders 265538
(+a-3 sheets), Kore 12608 (A); Tjitjalengka, Koorders 142068 (+a);
Parakansalak, G. Poetri, Tjikramat, Koorders 394058 (A); Tjilaki, southeast
Java, Forbes 924 (A, +MEL); Semarang Oengaran, Koorders 12238 (BRI); Mt.
Kaukuban, Prau, i eee 69 (tMEL); Tjikramat, Warburg 2678 (4nv) :
G. Tiloe, Pengalengan, Warburg 11118 (xy): Boerangrang, Goenseng Soenda,
Bakhuizen Van den Brink 4586 (uc). Without locality: Martius s.n. (+prR),
Blume s.n. (xy), Junghuhn 2 (+cu) Zollinger 2019 (BM, +MEL), Anon. (+MEL-3
sheets).

New Guinea. NETHERLANDS NEW GUINEA: Japen, Seroel, go i bb30698
(+a), bb30699 (a), bb30727 (+m0), bb30803 (A), bb30903 (a); Dalman, Nabire,
Kanehira & Hatusima 12266 (a). Terrirory oF New GuINEA. Sepik District:
Yellow River hills near Sepik River, Womersley NGF3937 (Lar-2 sheets),
Womersley NGF39 919 (+Lar—2 sheets). Eastern Highlands District: Aiyura
Range, Womersley NGF3374 (tar). Madang District: Kani Gebirge, Schlechter
16740 (uc). Morobe District: Boana vicinity, Clemens 8158 (+A); Morobe,
Womersley NGF3128 (+BRI, LAE). Papua. Western Division: Palmer River
below junction of Black River, Brass 7299 (A, +tLAE); Oriomo River, Wurio,
Brass 5907 (+A, BRI, CAS, MO, NY), 5908 (+A, BRI, NY); above sawmill, Hart
NGF5019 (+Lae). Gulf District: Murua River, Brass 1344 (A). Northern
Division: foothills of Hydrographer Range near new Inoto village, Hoogland
3845 (+Lar); foothills of Hydrographer & Owen Stanley Range, Lane-Poole 238
(A, +BRI), 275 (A, +BRr); Ioma, Manbare River, Allen & Martin NGF3283
(A, +BRI, LAE) } Dobodura Plain, near Embi Lakes, Cavanaugh & Fryar NGF2087

1958 | GRAY, REVISION OF PODOCARPUS, XI 463

(prI-2 sheets, LAE-2 sheets). Central District: Kokoda Track Plantation near
Sogeri, Womersley NGF4158 (ta, eee Sogeri, Smith NGF132 (+sr1); Uberi,
Eilogo Mill, 7 Aust. Ore. NGF 4 (pri), Anon. NGF48 (LAE); Hombron
Bluff, near Pt. Moresby, Gray & are NGF7137 (+tLae). Milne Bay Dis-
trict: Milne Bay area, Dawa Dawa River, Smith NGF1322 ({LAE). Goodenough
Island: east slopes, Brass 25023 (aA). New Britain: Mavalu River, Anon.
NGF2897 (tA, tLAE). Solomon Isiands. Maxarta: Dingali, interior from Quoi-
monapu, Kajewski 2370 (A, +BISH, BM, BRI). NEw Georcia: Waterhouse 209

+K).

Fiji Islands. Virt Levu: Mba, Singatoka River, Gillespie 3866 (ps, Ny, tuc,
us); hills between Nandala and Nukumuku creeks, Smith 6167 (a, t1LL); Naita-
siri, vicinity of Nandarivatu, Gillespie 4033 (+BISH), 4281 (+BIsH); Sovuta-
wambu, Degener 14670 (+A, MO, NY, US); woods near road part Tamavua Village,
Gillespie 2143 (+BISH); Prince’s Road forest, Parham 805 (+A); Namosi, Voma
Mountain, Gillespie 2910 (+p1sH), Naitarandamu Mt., Gillespie 3363 (+BIs#).
Vanua Levu: Mathuata, Seanggangga Plateau in drainage of Korovuli River,
vicinity of Natua, Smith 6721 (a, t1LL); south slopes of Mt. Numbiuloa, cs

of Lambasa, Smith 6385 (A, HILL), Smith 6570 (A, tILL). VANUA MBALAN
Bryan 575 (BIsH, ILL). Tavrunrt: Somo Somo, Gillespie 4840 ({BISH, NY,
uc). OVALAU: near summit of range west of Levuka, Gillespie 4433 (+BISH,
tuc). Neauv: hills inland from Sawaieke, Smith 7783 (tus). Exact locality not
indicated: Anon. U.S. South Pacific Expl. Exped. 1838-43 (tus).

Without locality: Horsfield Herb. sn. (GH), Anon. (cH), Kurz (ex Herb.
Sulp. Kurz) (A).

Cultivated. JAPAN: Oldham in 1861 (cH), Hort. Grieb., Anon. in 1849 (mo).
CEYLON: Peradeniya Gardens, Galston 2486 (tuc). JAVA: Hort. Bogor. V.F.33
(GH, Ny), Bot. Hard. Buitenzorg, Sargent in 1903 (A). ENGLAND: Kew Royal
Botanic Gardens, Cook in 1937 (ILL

The preceding description of Podocarpus neriifolius D. Don is limited
to specimens from Nepal (the type locality), Bengal, Assam and northern
Burma. This species has such a wide geographical range and varies so
much from the type that it can be best understood by consideration of
the specimens from separate regions. Podocarpus neriifolius occurs on the
continent of Asia in China, to the east, and in Siam, Indochina and the
Malay Peninsula, to the south, and on islands from Japan in the north,
southward to Sumatra and Java, and eastward through Malaysia to the
Fiji Islands.

Transverse sections of leaves of specimens from Nepal show leaves thin,
often to the point that some, or all three of the vascular resin canals
are obscured beyond recognition. There are no true auxiliary sclereids in
the mesophyll but the cell walls are frequently thickened, without pits, and
in many leaves the cells are large with little cytoplasm. Sometimes the
upper hypoderm, between the margin and the midrib, exists as only a
few isolated fibers of medium or even small diameter.

In China, the foliage is usually large and like that of the trees in Nepal.
However, Fang 2346 and Esquirol 3223 have smaller leaves, not over
8 cm. long and 7.5 mm. wide, but the long-angustate apex is like that of
the type, thus placing them in this species. Transverse sections of the
leaves show essentially similar anatomy. The upper hypoderm is more

464 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxiIx

abundant and in larger groups of fibers (up to 24) and vascular fibers
are rarer above the vascular bundle. Fang 2346 shows a few hypodermal
fibers between the stomatal rows and this specimen and Cheng 5319 have
rare auxiliary sclereids in the upper and lower mesophyll. The male cones
are sessile and are usually in fascicles of two or three.

The terminal buds of the Siamese specimens are larger than the twig,
globose, with broad triangular scales, the outer of which are shorter than
the bud. The leaves are large, those of Kerr 17833, Kerr 9907 and Garrett
538 being over 14 cm. long with a gradually narrowing acute apex. Kerr
17833 has the narrowest, not over 10 mm. wide. Kerr 14586 has wide
leaves which narrow abruptly to a caudate acuminate apex. In all of
them, the upper midrib is quite broadly prominent and the leaves are
thicker as shown in the transverse sections by the often doubled palisade
parenchyma. The upper hypodermal fibers are in small groups, sometimes
up to 55 » in diameter, and are not interrupted at the midrib. In Kerr
14586 the hypodermal fibers interrupt the palisade which is then replaced
by large cells with little cytoplasm. Auxiliary sclereids are found in the
lower mesophyll.

Podocarpus neriifolius also occurs in Indochina and Cambodia. ~
Clemens specimen is said to represent “scattered trees along the river,’
and has leaves only 10 mm. wide. The leaves of all specimens are scat-
tered and have the long-tapering acute apex typical of the species. Trans-
verse sections of the leaves do not always show vascular fibers and only
the Clemens specimen and one of Pierre 5532 have auxiliary sclereids be-
low the accessory transfusion tissue. Pierre 5532 needs special comment,
for two taxa are included under this number and the labels record two dif-
ferent collection sites, both questioned on the specimens. There are a num-
ber of sheets bearing this number in the herbaria of the Arnold Arboretum,
the New York Botanical Garden and the Paris Museum, and, of these,
only the specimens with large, scattered leaves having a long-tapering apex
are P. neriifolius. The others are P. annamiensis, a new taxon in which the
leaves are straighter, stiffer, somewhat more crowded, much smaller with
more abruptly acute apices and the transverse leaf sections show two
very large lateral vascular resin canals instead of three of almost uniform

Of the specimens from Malaya, Anon. 11024 has very wide leaves with
the apex abruptly narrowing to a caudate-acuminate tip and long acum-
inate-attenuate terminal bud scales. The upper hypodermal fibers are
solitary or in very small groups and are rarely interrupted at the midrib.
In Anon. 11024 the palisade mesophyll is undeveloped but cells with
dense cytoplasm lie between the upper hypodermal fibers. Leaves of the
Glandoger specimen usually show two additional lateral vascular resin
canals which are associated with the transfusion tissue and the upper
hypodermal fibers are scant.

Kanehira 21309 from Formosa is not Podocarpus nakaii. It most closely
resembles the specimens of P. neriifolius from the Fiji Islands.

On Luzon, in the Philippines, the leaves have very abundant upper hypo-

1958] GRAY, REVISION OF PODOCARPUS, XI 465

dermal fibers with few or short interruptions in Clemens 16251a and For.
Bur. 31479, but on Polillo, McGregor 10779, they are rare and in small
groups of 1-4 fibers. There are rare lower hypodermal fibers between
the stomatal rows in For. Bur. 31479. In leaves of Clemens 16251a the
palisade is often a double layer, rare in this species.

West of the Malay Peninsula, Podocarpus nertifolius has been collected
from South Andaman Island. The leaves are thin and wide with a caudate-
acuminate apex. Transverse sections of the leaves of Kerr s.n. show upper
hypodermal fibers in small groups, also interrupted at the midrib, and
a palisade of short dense cells between the fibers.

Transverse sections of leaves of plants of this species from Sumatra
usually have solitary scattered upper hypodermal fibers and the hypoder-
mal layer is interrupted at the midrib. Leaves of Bangham 1128 show a
few auxiliary sclereids below the accessory transfusion tissue. The leaves
of this specimen also have abruptly caudate-acuminate apices.

Most of the Javan specimens are like those from Nepal. Some have
an abruptly acuminate or caudate apex to the leaf, especially if the foliage
is quite large. The Martius specimen has leaves with the two lateral vas-
cular resin canals larger than the central one. Two specimens, Reinwardt
s.n. and Anderson 69, have five vascular resin canals, the two extra lateral
ones being very near or in the transfusion tissue. Transverse sections of
the leaves show the interrupted hypoderm with fibers averaging a little
larger than those of the type, often scattered or even isolated; there are
lower hypodermal fibers between the stomatal rows only in the Reinwardt
specimen; auxiliary sclereids only in Koorders 265538.

Only one of the specimens which I have seen from Borneo, Korthals s.n.,
has the narrowly prominent midrib on the upper side of the leaf which
is characteristic of Podocarpus neriifolius. This specimen has also terminal
bud scales with long-acuminate or even foliaceous tips, to 1 cm. long.
Blume cited this as one of the specimens in his P. leptostachyus, which
has been placed in the synonymy of P. neriifolius by Pilger (1903). The
Melbourne herbarium specimen, which I examined, has a solitary, ex-
panded but very narrow male cone.

This species is a fairly common tree in New Guinea and seems to reach
higher altitudes there than in other areas. Characteristics which were
found most useful in delineating the species were the slender branches,
the leaves with margins not parallel and with long-tapering acute or acumi-
nate apices, a sharply prominent upper midrib, transverse sections show-
ing three vascular resin canals, interrupted upper hypodermal fibers be-
tween the stomatal rows, and the lack of auxiliary sclereids in the meso-
phyll. The terminal bud varied from ovate to globose and the length of
the bud scales did not seem to be critical. Male cone buds are small and
usually in sessile clusters of one to three.

Other species of this group in New Guinea are P. thevetiifolius, P. rum-
phi, P. ledermannii, P. archboldii and P. idenburgensis. Reference to the
key will indicate the essential differences.

The leaves of the specimens of Podocarpus neriifolius from the Solo-

466 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

mon Islands are thin, broad and have a very acute or caudate-acuminate
apex and the seeds are large, up to 14 mm. long. Transverse sections of the
leaves show upper hypoderm and palisade as in the Borneo specimen.

Some of the specimens from the Fiji Islands cannot be clearly differen-
tiated from the type of Podocarpus neriifolius in Nepal and thus are in-
cluded in this species. With this inclusion the geographical range for
P. neriifolius is greater than that of any other species. Podocarpus
oleifolius in South and Central America extends for an almost equal dis-
tance north and south of the equator but there is little range of longitude.
Here, in the Fiji Islands, P. neriifolius is a tree found in either open or
dense forests at altitudes usually of less than 500 m. The leaves are
quite thick and coriaceous, sometimes only 4 cm. long and 4 mm. wide on
reproductive branches, and the margins tend to be revolute. The broad
lower midrib may have an open groove, and transverse sections always
show auxiliary sclereids. The pollen cones are usually clustered and ses-
sile. Many of these specimens have been marked as P. elatus, a species
which differs distinctly in having abruptly acute or obtuse, often mucronate
tips to the leaves and continuous upper hypoderm is seen in transverse
sections. One specimen, Gillespie 2143, has the immature seed elongate
with a conspicuous crest and an elongated beak at the micropyle.

27a. Podocarpus neriifolius var. atjehensis Wasscher in Blumea 3:
450. 1941.

A small tree, 15 m. tall, with stout branches. Terminal buds large, the
outer scales long-acuminate, to 10 mm. long, tips spreading. Leaves
crowded, deflexed or drooping, linear-lanceolate, 7-18 cm. long, 5—8.5 mm.
wide, very gradually narrowed to an acute, sometimes mucronate, apex,
gradually narrowing to a short petiole at the base; midrib narrowly prom-
inent above, sometimes limited by shallow grooves on either side, broadly
prominent below. Transverse sections of the leaves show three vascular
resin canals with the lateral sometimes larger and the median obscured,
interrupted upper hypoderm of isolated fibers 23-37 pw in diameter, the
layers interrupted at the midrib both above and below, rare vascular fibers
only above the bundle, cuticle thick. Male cones solitary, axillary, sessile;
scales acute-acuminate, up to 5 mm. long; mature cones 2—3 cm. long,
4—4.5 mm. thick. Female cones solitary in the leaf axils, crowded in the
lower part of new growth; peduncles slender, 8-16 mm. long; receptacle
narrowly fleshy, 7-9 mm. long, subtended by long bracts, up to 6 mm.
Seed sub-elliptical, 9-10 mm. long, 7-8 mm. broad, apex obtuse.

DISTRIBUTION: In Atjeh in Sumatra, at elevations of 2250-3300 m.
Sumatra: Atjeh, Gajolanden, G. Kemiri, Van Steenis 9614 (+a).

The above description was prepared from the type specimen, the only
example seen.

1958] GRAY, REVISION OF PODOCARPUS, XI 467

27b. Podocarpus neriifolius var. degeneri, var. nov.

Frutex vel arbor parva; ramulis sparsis, subverticillatis; gemnis termi-
nalibus magnis, ovoideis, squamis exterioribus attenuatis, e basi lato
crescentibus. A specie differt foliis parvis maturis, plerumque minusquam
10 mm. latis.

Shrub or small tree; branches few, subverticillate; terminal vegetative
buds large, ovoid, the bud scales with apices acuminate to long-attenuate
or abruptly narrowed to a long apiculus. Leaves patent to spreading, thin-
coriaceous, 6-12 cm. long, 6-10 mm. wide, linear-angustate, shiny above,
dull-rusty beneath, the apices narrowly acute, the bases narrowed into
short petioles; leaves on young shoots to 18 cm. long, 17 mm. wide. Trans-
verse leaf sections show 3 vascular resin canals, hypodermal fibers of
somewhat smaller diameter than those of the species, vascular fibers rarely
present and auxiliary sclereids lacking in the mesophyll. Pollen cones
axillary on growth of the previous year, sessile, cylindrical, up to 3 cm.
long, 3—3.5 mm. in diameter, surrounded at the base by numerous thin,
carinate, broadly ovate scales which are sometimes apiculate; microsporo-
phylls with small, narrow, up-turned apiculi. Female cones axillary, soli-
tary, on slender peduncles 7-12 mm. long, the receptacle of several fused
scales, 8 mm. long, with small, spreading free tips, subtended by a pair
of slender, thin, attenuate bracts up to 5 mm. long, and bearing 1—2 ovules.
Seeds 9-12 mm. long, elongate, 5 mm. wide, bluntly crested; immature
seeds narrowed at the base.

DIsTRIBUTION: along streams in forests, 40-800 m. altitude, on Viti
Levu in the Fiji Islands.

Fiji Islands. Virr Levu: Mba, Nandarivatu, Degener 14272 (TYPE, +A, MO,
NY, US); Unidawa Road near Nandala River, Gillespie 4137 (+BISH, +DS, NY,
uc); hills between Nggaliwana and Nandala creeks, south of Nauwange,
ae 5665 (A, FILL), 5666 (A, tILL); 3 miles south in valley of Nandala Creek,

. C. Smith 6254 (A, tILL); Gillespie 4129 (}+BIsH); Singatoka River, Gillespie
oe (+BIsH, US); forest at headwaters of stream which runs to Navua, Gillespie
4250 (+BISH, NY, UC); Lautoka Mts., Greenwood 50A (+A, BRI, NY), 45A
(A, BRI); Namosi, above Waikava, Parham 1701 (+a); Nandina River, Gillespie
2531 (+BISH, UC); Waikava, Parham 2154 (+a); Serua, banks of Navua River,
Gillespie 3382 (+BISH, UC). Locality not indicated: Seeman 575 (+cH). Horne
792 (+GH). Cultivated. Vitr Levu: Exp. Sta. 17 m. east of Suva, Buchholz
sm, (FILL)

27c. Podocarpus neriifolius var. polyanthus Wasscher, Blumea 4:
455. 1941

Tree to 40 m. tall, slightly fluted toward base. Terminal buds large,
conical; outer scales acuminate or only acute, erect, stout, often shorter
than the bud. Leaves spreading, more or less coriaceous, straight or some-
what falcate, lanceolate, gradually narrowing to a short petiole, abruptly
or more gradually narrowing to the acute apex, 6-16 cm. long, 13-20 mm.
broad (only 6-9 cm. long, 7-8.5 mm. broad in New Guinea); midrib

468 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

narrowly prominent, broader and sometimes furrowed below. Male cones
unknown. Female cones numerous, all over the new growth, in the axils
of bracts as well as in those of the leaves and above the leaf scars; the
few bracts sessile with a broad base, acute, to 1.5 cm. long, 2.5 mm. broad;
peduncles thick, 1.5—5 mm. long; receptacle subtended by 2 subulate bracts
to 3 mm. long, composed of 2—4 fused fleshy scales, only 1—2 fertile, short
and thick cylindrical, 6-7 mm. long, 4-5 mm. broad. Seeds elliptical-ovate,
10 mm. long, 6 mm, in diameter.

DISTRIBUTION: In Sumatra (fide Wasscher) and New Guinea in rain
forests, alt. LOO—600 m.

New Guinea. Papua: Milne Bay District, Smith NGF1322 (+pr1).

The specimen cited above has abundant very young ovules on the new
growth at the twig tips. Dormant buds are connate to globose with short
triangular obtuse scales. The leaves are small for Podocarpus neritfolius
but they have the general shape of the species. Transverse leaf sections
show the anatomy typical of P. neriifolius with interrupted upper hypo-
derm of small fibers and no hypodermal fibers between the stomatal rows
of the lower side. One of the sectioned leaves shows a loosely organized
palisade layer on the lower side of the leaf. This has not been found in any
other specimen in Section Podocarpus.

27d. Podocarpus neriifolius var. teysmannii Wasscher, Blumea 4:
453. 1941.

a Teysmannii Miquel, Fl. Nederl. Indié 2: 1072. 1859; Parlatore,
De. dr. 16: 516. 1868; Gordon, Pinetum ed. 2. 348. 1875: Pilger,
ee IV. 5(Heft 18): 81. 1903; Dallimore & Jackson, Handb.
Conif. 56. 1923, 1931; Florin, Svenska Vet.-Akad. Handl. III. 10: 280. 1931.

A small tree with stout twigs. Terminal buds globose, flattened, with
short, outer scales 2-3 mm. long, these closely appressed, rigid, broadly
triangular, with obtuse-acute apex. Leaves scattered, spreading, broadly
lanceolate, 8.5-17 cm. long, 16-26 mm. wide, abruptly narrowing to
a short, thick petiole at the base, also abruptly narrowing to the caudate-
acuminate apex; midrib very broadly prominent above, ridged, flat, prom-
inent or even channelled below. Male cone buds solitary in the leaf axils,
globose, large, 3 mm. in diameter; mature cones unknown. Female cones
and seeds unknown.

DISTRIBUTION: At sea level or low altitudes on Sumatra and nearby
islands.

Sumatra: Teysmann 513 (tex Florin, Berlin Herbarium) ; sea coast, Teysmann
sn. (MEL), Teysmann s.n. ({MEL); Tapanoeli, Angkola en Sipirok, Panobasan,
Dk. Poehoetan Lajan, Boschpr. bb30986 (mo). Bangka: ne sembliang, Teys-
mann s.n. (}MEL). Karimata: Teysmann s.n. ({MEL-2 sheets).

According to Wasscher (11), who combined this taxon with Podocarpus

1958] GRAY, REVISION OF PODOCARPUS, XI 469

nertifolius, var. teysmannii differs from the species chiefly in the globose
terminal buds, the broadly lanceolate leaves with shortly acuminate apex,
and the large male cone buds.

Transverse sections of the leaves show anatomy in agreement with the
species except for very rare auxiliary sclereids in the mesophyll between
the accessory transfusion tissue and palisade layer. Orr’s description (9)
of the leaf anatomy of Podocarpus teysmannii must be referred to P. philip-
pinensis, since the specimen he used (Orolfo 3919 from the Selangen
Islands) is here identified as the latter species.

Large, broad leaves with a caudate-acuminate apex are also found in
Siam, the Malay Peninsula, the Andaman Islands, the Solomon Islands,
etc. Such specimens may belong in this variety, but at present are placed
in Podocarpus neriifolius because the terminal bud is either ovate or can-
not be observed.

Wasscher (11) listed four varieties of Podocarpus neriifolius which must
be recognized, although I have seen no specimens of any of these. They
could not be included in the key to the species and varieties in this paper,
therefore, but for completeness I quote the critical characters given by
Wasscher and indicate the geographical location of each variety. Podo-
carpus nertifolius var. bracteata Wasscher, from Java, “differs . . . in the
large, ovate-acute, male flower buds, with squarrose, ovate-triangular,
acute bud scales.” Podocarpus neriifolius var. linearis Wasscher, cited from
four collections from Java, “differs . . . in the narrower leaves with mar-
gins parallel, the obtuse terminal buds and the large, ovate, male flower
buds.” Podocarpus neriifolius var. membranacea Wasscher, described from
two collections from Celebes, “differs . . . in the membranaceous scales
of the terminal buds and the ovate-acute male flower buds.” Podocarpus
neritfolius var. timorensis Wasscher, described from a single collection from
Timor, “differs . . . in the indistinct, not or hardly prominent midrib,
and the leaves shorter-narrowed toward the apex and often with a mucro,
whereas the male flowers are thicker.”

28. Podocarpus polystachyus R. Brown ex Mirb. Mém. Mus. 13: 75.
1825 (nomen! ), ex Bennet in Horsf. Pl. Jav. Rar. 40. 1838 (nomen!) ;
Endlicher, Syn. Conif. 215. 1847; Gordon, Pinetum ed. 2: 345. 1865;
Carriere, Traité Conif. ed. 2, 662. 1867; Parlatore, DC. Prodr. 16:
515. 1868; Bertrand, Ann. Sci. Nat. V. 20: 59. 1874; Warburg, Mon-
sunia 1: 192. 1900; Pilger, Pflanzenreich IV. 5(Heft 18): 79. 1903,
Nat. Pflanzenfam. ed. 2. 13: 247. 1926; Merrill, Philip. Jour. Sci. 3:
394. 1908; Foxworthy, Sit, Jour. Sci. 6: 161. 1911; Stiles, Ann.
Bot. 26: 455. 1912; Dallimore & Jackson, Handb. Conif. 54. 1923,
1931, 79. 1948; Florin, Svenska Vet.-Akad. Handl. III. 10: 280.
1931; Wasscher, Blumea 4: 456. 1941.

Podocarpus littoralis Teysmann, Nat. Tijdschr. Ned. Ind. 36: 237. 1876.

P. littoralis on ex Biswas, Jour. & Proc. Asiatic Soc. Bengal eee II.
28: 374. of

fen Sea D. Don in Lamb. Pinus ed. 2. 2: 122. 1828.

470 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxx1x

A small tree up to 20 m. tall with a trunk up to 45 cm. in diameter and
numerous subverticillate spreading twigs. Terminal buds ovate with keeled,
narrow, acuminate, stiff scales sometimes becoming 10 mm. long. Leaves
scattered, usually crowded toward the tips of the branches, coriaceous,
shiny above, usually lanceolate to linear-lanceolate, flat, 4-8 cm. long,
6-10 mm. wide (rarely somewhat wider), abruptly narrowed to the mostly
obtuse apex, sometimes mucronate, gradually narrowing at the base to a
short petiole; midrib prominent above, broader and sometimes shallowly
channelled below. Cross sections of the leaves show 3 vascular resin canals,
interrupted upper hypoderm of very large fibers up to 70 » in diameter,
lower hypoderm absent between stomatal rows, vascular sclereids usually
present above or below the bundle (mostly absent in Philippine specimens.
Pollen cones fascicled 3—5 in leaf axils, sessile, cylindrical, 2—4.5 cm. long,
2.53 mm. in diameter, surrounded at the base by small, broad, spreading
scales. Microsporophylls broadly ovate-triangular, acute to somewhat
obtuse. Female cones axillary, solitary on peduncles 3-6 mm. long; re-
ceptacle of two fused fleshy scales, one somewhat shorter than the other,
subtended by 2 subulate, early deciduous bracts. Seed globose or elliptical,
about 10 cm. long.

DistTRIBUTION: Low elevations, often coasts, of islands south and east
of the South China Sea, and the Malay Peninsula.

Philippine Islands. Luzon: Ilocos Norte, Burgos, Bur. Sci. 27146 Ramos
(+A); Tayabas, Bur. Sci. 27325 Quieb Nie ny), Bur. Sci. 26902 Edano (srr);
For. Bur. 24264 Cagayan (aA). Patawan: For. Bur. 3854 Curran (tNy), For.
Bur. 904 Foxworthy (NY). NEGROS: nee part, Masias, Sareno & Torrible
in 1925 (tuc). Mindanao: Davao, Pendesan, Kanehira 2623 (ny).

Malaya. PAHANG: Henderson 18420 (4+uc): Kioala, Ridley 1441 (+Bm).
PuLau Troman: Telok Paya, Anon. s.n. (+uUc). SINGAPORE: Barnes’ Island,
G. L. Smith sn. (Type, tpm); Forest de Kranjo, Langlasse 72 (+P); Wallich
6052B (mo, te); Mangrove swamp at New John’s Street, Sargent in 1903 (A-2
sheets); Serangoon, Ridley 3367 (+mEL); Hooker & Thomson s.n. (GH).

Borneo. Teysmann ce (}mMEL). BritisH NortH Borneo: Kuala Penyu, For.
Dept. 1798, Apostal 35 (t+uc); Jesselton, Clemens 51171 (+uc), 9659 (A, BRI,
GH, tuc-2 sheets), 9568 (A); Native collector 2353 (a); Foxworthy sn. (tus).

Ban gka: Teysmann s.n. (+MEL-2 sheets).

Cultivated. MALAy PENINSULA: Singapore, Bot. Gard., Furtado in 1931 (+srr).
Java: Buitenzorg Bot. Gard., Warburg 1210 (+Nv), VF. 1 , from Lingga, Sar-
gent in 1903 (ny, tuc-3 she ets), Anon. s.n. (tps), V. F. 17a, from Lingga,
Sargent in 1917 (ny, uc), Sargent in 1903 ae sheets, +MOo-— me sheets); Bohn’s
Garden at Buitenzorg, Sargent in 1903 (A- eets). Aprtca: Kisantu, Jardin
Agronomique, Vanderyst 31936 (+BR), Poe 31949 (+BR), Vandervat 36983
(+8R); Belgian Congo, Eala, Corbisier-Baland 1136 (+BR, K, MO, tNY), Cor-
bisier-Baland 1353 (+BR, K, Mo),

Podocarpus polystachyus is quite distinct in having often isolated, ex-
tremely large upper hypodermal fibers seen in transverse leaf sections.
This character, together with the linear-lanceolate, rigid leaves and the
male cones in bundles of 3-5, make possible the positive identification of

1958] GRAY, REVISION OF PODOCARPUS, XI 471

the Corbisier-Baland and Vanderyst specimens from cultivation in Africa.
Determination of the former had been questioned and of the latter un-
known

Transverse sections of the leaves do not show auxiliary sclereids in
upper or lower mesophyll. However, some cells of the lower mesophyll
have thickened and pitted walls but are not devoid of cell contents (except
Henderson 18420).

The description of Podocarpus polystachyus by Wasscher (11) was the
first to include specimens from the Philippine Islands. Foxworthy (5)
took his description from that of Pilger (1903) which was, in turn, from
specimens from the Malay Peninsula and Borneo only. Why Foxworthy
was not certain that this species was distinct from P. elatus R. Br. is not
clear except that the leaves of both are often mucronate. The leaves of
P. elatus are larger, have parallel margins and the upper hypoderm of much
smaller fibers is continuous.

28a. Podocarpus polystachyus var. rigidus Wasscher, Blumea 4: 460.

Leaves thick-coriaceous, very rigid, more lanceolate and broader with
margins not parallel, 3-7.5 cm. long, 8-14 mm. broad; midrib strongly
prominent above, often distinctly, broadly and shallowly channelled be-
neath.

DIsTRIBUTION: Riouw Archipelago and Borneo, on mountain slopes and
summits.

Borneo: western part, G. Kelam, Hallier 2373 (+Ny).

Wasscher (11) described this variety as differing from the species in
its leaf shape, the species having narrower, linear-lanceolate leaves. I
do not find this difference so striking; many specimens having some leaves
of the described shape must be assigned to the species, the leaves of which
may be up to 13 mm. broad. However, in an examination of the transverse
section of a leaf fragment from Hallier 2373, I find abundant large sclereids
in the upper mesophyll which are entirely absent in the species. Some of
the lower mesophyll cells show the same thickening of the walls as ob-
served in the species proper. The shallow channel of the lower midrib
of this specimen is so constricted as to eliminate the central vascular
resin canal.

That Podocarpus polystachyus var. rigidus is found on mountain slopes
and summits is more striking than the small differences in leaf anatomy.
The habitat of the species is limited to the low elevations of sea coasts
and estuaries.

29. aa dey macrophyllus (Thunb.) Don in Lamb. Pinetum ed.
1.2: 22. 1824, ed. 2.2: 123. 1828; Endlicher, Syn. Conif. 216. 1847;
au Rum phia S02 NS Saale Contre. Traité Conif. 664. 1867:
Parlatore, DC. Prodr. 16: 517. 1868; Sieb. & Zucc. in Miquel, FI.

472 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxix

Japon. 2: 70. 1870; Bertrand, Ann. Sci. Nat. V. 20: 59. 1874; Mah-
lert, Bot. Centralbl. 24: 281. 1885; Van Tieghem, Bull. Soc. Bot.
France 38: 169. 1891; Shirasawa, Essenc. Forest. Japon. 1: 31. 1899;
Warburg, Monsunia 1: 192. 1900; Kent, Vietch’s Man. Conif. 150.
1900; Pilger, Pflanzenreich IV. 5(Heft 18): 79. 1903, Nat. Pflanzen-
fam. ed. 2. 13. 1926; Bailey, Cult. Evergreens 179. 1923; Dallimore
& Jackson, Handb. Conif. 49. 1923, 1931, 73. 1948; Florin, Svenska
Vet.-Akad. Handl. III. 10: 279. 1931; Metcalf, Fl. Fukien 1: 20.
1942.

P. chinensis Blume, Rumphia 3: 216. 1947 non Wall.; oe te fsa

24; 281. 1885; Florin, Svenska Vet.-Akad. Handl. une 10: 3
Podocarpus chinensis Wallich ex Parlatore, DC. Prodr. 16: - pity
Taxus macrophylla Thunberg, Fl. Japon. 176. 1784.

A tree 8-15 m. tall, with horizontal branches and numerous, crowded,
leafy twigs. Terminal buds ovate, small, the scales ovate with acuminate
tips. Leaves crowded, straight, green above, paler beneath, linear-lanceo-
late, apex long-angustate, acute or obtuse, gradually narrowing at the
base to a short petiole, 5.3-10.3 cm. (usually 8-10) long, 6-10 mm. (usu-
ally 9-10) broad; midrib prominent above and rather broad, fairly prom-
inent and broader below. Transverse sections of the leaves show 3 resin
canals below the vascular bundle, interrupted upper hypoderm of very
small fibers (16-20 » in diameter), no lower hypodermal fibers between
the stomatal rows, palisade of 2-3 layers of cells, vascular fibers often
present below the vascular bundle and auxiliary sclereids usually absent
from the mesophyll. Pollen cones fascicled 3-5, sessile, very narrow-cylin-
dric, up to 3 cm. long, surrounded at the base by broad triangular, stiffly
coriaceous scales. Microsporophylls obtuse with distinct apiculus. Female
cones solitary, peduncles 16-17 mm. long; receptacle of 2—3 fused fleshy
scales with free tips, subtended by 2 small, narrow, subulate scales 3-4 mm.
long. Seeds ovoid, 10-12 mm. long.

China. Krancsu: Soochow, Dang Wei, Ching & Tso 785 (A); Kunshan,
Mrs. Tsiang Ying 1 (Ny). SzecHwan: Fang 1963 (A). CHEKIANG: 20-40
miles west of Wenchow, Ching 1892 (tuc); Wenchow, Ling Kan, Univ. Nan-
king 7308 (tuc); Yen Tang Shan, Chiao, Univ. Nanking 4737 (a, Ny, tuc);
Nan-Hoo, H. H. Hu 191 (cH); King Yuan, Ching 2439 (Ny, tuc). CHUSAN:
Home s.n. (mo). ANWErI: Chemen, Chung 3124 (A). Krancsr: south of Nan-
chong, Sang-su-ling, Chung 2 (A); Lu Shan, Chiao 18760 (A). FuKIEN: Amoy,
Kulangsu, Chung 5775 (A); Puchen, Chung 3865 (a). Kwanctunc: Koo Long
Ue, McClure B-2 (ny); Canton, Chun 7731 (Ny). YUNNAN: between Muang
Hai and Keng Hung, Rock 2495 (a), Bonati 1802 (tuc). KwetcHow: Tsunyi,
Tsiang 5317 (A, NY Bs

Japan. HonsH Mt. Maya, Settsu, Arimoto in 1903 (tmo); Idsu (ex
Herb. Sokurai), Wilson in 1910 (A). KyusHu: Tairashima, Kawanabe Island,
Ushio in 1917 (+a); Nagasaki, Anon. (ex Herb. Th. Porter) (pH), Maximowicz
in 1863 (cH, tny-2 sheets). Loo CHoo Istanps, Wright 310 (GH, NY).

Cultivated. CHINA: Kwangtung, Tak-Hing, Heung Shan, C.C.C. 13193,
McClure (+uc). Japan: Hondo, Wilson in 1911 (ex Herb. Sakurai) (A); Mino

1958] GRAY, REVISION OF PODOCARPUS, XI 473

Prov., Shioto 8743 (a), Ugai Kichigoro 4440 (a). PHILIPPINE IsLanps: Luzon,
Manila, Fenix 161 (A). NepaL: Netla Thuret 1889 (A). SiInGAporRE: G.T. (ex
Herb. Hook. & Thomson) (GH). SwiTzERLAND: Locarno, Bot. Gard., Baenitz in
1905 (1LL). ENGLAND: Kew, Royal Bot. Gard., Cook in 1937 (1LL). UNitTEep
STATES: New York Bot. Gard. 3750, from C.P. 1900, Hartling in 1919 (111);
California, Goleta, Anon. sn. (f1LL). West InpIEs: Jamaica, Chinchona Gar-
dens, Buchholz in 1946 (111); Trinidad, Roy. Bot. Gard., Ewan 17014 (+o);
Tobago, Castleton Gardens, Harris 45, 50, Broadway 2934 (x).

Transverse sections of the leaves of Podocarpus macrophyllus and its
varieties are ordinarily quite easily recognized by the thick, usually doubled,
layer of palisade mesophyll and the interrupted upper hypoderm of small
fibers. This is very useful as the species is highly successful in cultivation
and the original source of the material has frequently long been forgotten.
Often it seems to be misidentified as P. elongatus, an African species,
which has distinct marginal resin canals.

Stiles (10) selected the leaf of Podocarpus macrophyllus for his de-
scription of the anatomy most typical of Podocarpus. That the plant
he used was probably the variety maki does not invalidate the description,
for the leaves have the same anatomy and are merely shorter and narrower.
He indicated that he observed “slight” differences in the quantity of trans-
fusion tissue, accessory transfusion tissue and sclerenchyma in P. elatus
and P. polystachyus, but attributed these differences to external conditions.
I find, however, on examination of a larger number of specimens, that
continuous upper hypoderm is a consistent character of P. elatus, and the
hypodermal fibers are much larger in diameter in P. polystachyus. Orr (9)
found that the above characters provided distinguishing differences for
these species.

29a. Podocarpus macrophyllus var. angustifolius Blume, Rumphia
3: 215. 1847; Pilger, Pflanzenreich IV. 5(Heft 18): 80. 1903.

This variety differs from the species in having narrower leaves which
are quite variable in length, being both shorter and longer on a single plant
than the extremes for the species. The linear-lanceolate leaves have also
a more tapering apex.

DISTRIBUTION: A few specimens from both China and Japan but best
known from cultivation in the United States.

China. Kwanctunc: Canton Christian College Herb. 13193 (+tuc); Canton,
Read (pH). Krancsu: Tungtin, Soochow, Tai Hu Lake, Ching & Tso 730 (a).
North Burma: Nmai Hka Valley, Ward 521 (A). Japan. Honsuu: Yamato
River in Nara Park, Wilson 7855 (A); Kanagawa-Ken, between Kamakura &
Zushi, Beattie & Kwihara 10449 (a). Hacuiyo Istanps: Wilson 8391 (A).
Locality not indicated: Anon. (ex Herb. Lugd. Batav.) (+BR—2 sheets, Ny).

Cultivated. Japan: LiuKiu Islands, Yokohama Nursery, Wilson s.n. (a).
JAvA: Zollinger s.n. (+BR). AUSTRALIA: Brisbane Bot. Gard., White 9283 (a).
Europe: Germany, Hannover Bot. Garden #6, Zabel in 1892 (a). WEsT
InpiEs: Tobago Bot. Station, Broadway 2934 (F, +Mo-2 sheets, Ny); Trinidad,

474 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. XxxIx

Port of Spain, Broadway in 1933 (A); Trinidad, St. Augustine, Baker 14554
(TRIN); Martinique, Basse Pointe, Duss 2096 ({Ny). CuBA: Prov. Santa Clara,
Harvard Trop. Gard., Jack 8353 (a, NY, US). Unirep States. California: Los
Angeles, Buchholz in 1940 (A, +1LL-3 sheets), pane in 1942 (4+ILL), van
Rensselaer 1713 (iLL), Brown in 1938 (cas), Rollins & Chambers in 1938
(cH), Turner in 1926 (cas); San Francisco, Golden Gate Park Arboretum,
Buchholz in 1942 (+1LL), Walther 4 in 1942 (aA); Santa Barbara, Orpets Nursery,
van Rensselaer 1718 (t1LL). Texas: Houston, Fisher 44301 (fF). New Jersey:
New Brunswick, Rutgers Greenhouse 1B, Johnston in 1952 (truT), from CU1,
Atkins Garden, Cuba, Johnston in 1952 (trut). New York: Botanical Garden
Greenhouse #3635 from D.P.M. 1900, Nash in 1905, #3750 from D.P.M. 1900,
Taylor in 1906, #15022 from Edinb. 1902, Taylor in 1905 (Ny).

Podocarpus macro phyllus var. angustifolius 1 is widely cultivated in south-
western and southern United States where it is used mostly in trimmed
hedges, sometimes as a shrub or small tree. It is also extensively used as
a green foliage in floral arrangements. It frequently passes under the names
P. chinensis or P. sinensis, and is commonly called the “Southern Yew.”

29b. Podocarpus macrophyllus var. chingii, var. nov.

Arbor columnaris, 8 m. alta, ramulis adscendentibus; foliis crassis rigidis,
oblanceolatis, 1.2—3 cm. longis, 3-4 mm. latis, valde gradatim attenuatis
in petiolo brevissimo, apicibus plerumque obtusis, marginibus interdum
revolutis: strobilis masculis alabastris globosis, 1.5 mm. diametro, 1—3-
fasciculatis, squamis tenuibus, late triangularibus, obtusis, imbricatis, stro-
bilis maturis tenuibus, ad 2.5 cm. longis; strobilis femineis ignotis.

‘ree with trunk up to 10 cm. in, diameter; bark fibrous and brown;
young twigs green. Terminal buds very all conical; outer scales ie
up to 3 mm, long, carinate, somewhat tapered to the thick apex, the tips
occasionally spreading. Leaves small, densely crowded, erect, linear-oblong
to narrowly spatulate, the apex acute to rounded, the midrib very sharply
prominent above, broader and flat below.

DisTRIBUTION: Known from only the type locality, in Chekiang, China.

China. CHEeKIANG: 55 Chinese miles (li) west of Lung-sien, alt. 100 m.,
Ching 2477 (Type, +A).

The striking columnar habit and the very small leaves distinguish this
variety. It most closely resembles Podocarpus macrophyllus var. maki, the
foliage being a miniature of that variety. The foliage does not differ greatly
from that of P. brevifolius which has the leaves mostly larger and taper-
ing more quickly toward the base. The latter species, however, is a tree
with stout, spreading branches and solitary male cones.

29c. Podocarpus macrophyllus var. maki Endlicher, Syn. Conif. 216.
1847: Sieb. et Zucc. in Miquel, Fl. Japon. 2: 70, 1870; Dallimore &
Jackson, Handb. Conif. 49, 1923, 1931, 74. 1948.

Podocarpus macrophyllus ssp. maki Sieb. Naamlijst 35. 1844; Pilger, Pflan-

1958 | GRAY, REVISION OF PODOCARPUS, XI 475

zenreich IV. 5 (Heft 18): 80. 1903, Nat. Pflanzenfam. ed. 2. 13: 248. 1926;
Wasscher, Blumea 4: 461. 1941.

Podocarpus chinensis Wall. in Endl. Syn. Conif. 215. 1847; Blume, Rumphia
3: 216. 1847; Bertrand, Ann. Sci. Nat. V. 20: 59. 1847; List N.6051 ex
ee Traité Conif. ed. 2. 658. 1867; Parlatore, DC. Prodr. 16: 516.

868; urg, Monsunia 1: 192. 1900.

en ae Sieb. Ann. Soc. Hort. ne Bas 35. 1844.

Podocarpus makoyi Blume, Rumphia 3: 215. 1847.

Podocarpus miquelia Hort. ex Parlatore, DC. Se 16: 516. 1868.

Podocarpus sinensis (?) Favre in Ann. Sci: NataVe 323/97 1365.

Podocarpus vrieseana Hort. ex Parlatore, DC. Prodr. 16: 516. 1868.

Small tree or shrub with erect branches and thickly leafy twigs. Leaves
straight, erect to spreading, linear-lanceolate, obtuse or shortly angustate
at the tip, gradually narrowing to the short petiole, 3.5-7 cm. long,
4.5-7 mm. wide; midrib narrowly prominent above. Male cones sessile,
in clusters of 3—5 in upper leaf axils, strobili narrowly cylindrical, nearly
filiform, 3—4.5 cm. long, 2-3 mm. in diameter; microsporophylls triangular,
acute, apiculate. Female cones solitary in leaf axils on peduncles 5—11 mm.
long; the fleshy receptacles of 2-3 fused scales subtended by 2 small
subulate bracts. Seeds elliptical, 8-10 mm. long, 6-7 mm. in diameter.

DIsTRIBUTION: A few specimens of this variety from China, North
Burma, Japan and Formosa seem to have come from wild habitats. It is
best known from its successful wide cultivation not only in Japan, its
probable source, but throughout the world in the milder climates.

China. CHEKIANG: southern part, between Ping Yung and Tai Suan, Ching
2168 (A, NY, tuc); Tsing-Yun Dist., Peach Mountain, Keng 438 (A); Chusan
& Ningpo, Capt. Home in 1892 (Bm), Ching 2439 (tuc). KwanTuNG: Wat
Shui Shan, Chun 7343 (a), Wang & Ling 7343 (tuc); Su-liu-kwan, Lofou
Shan, Tsiang 1750 (A), Heunghsan, Chun 97 (ny); Canton, Anon. in 1889
+mo). Locality unknown: ex Herb. Prager, Anon, sn. (tcas), Bonati & ide
1802 (+uc), ex Univ. Nanking, Chiao in 1925 (tuc). North Burma: ’N
Hka Valley, Kingdon Ward 521 (Ny). Japan. HonsHu: Oshima, Prov. i Zu,
Mizushima 815 (A); Hondo, Shioto in ae (tA); Hakone and Tokyo, Hart-
i L . KyusHu: Kagosima, Masamune
in 1922 (ny). No specific locality in Japan: Thunberg s.n. (tups), Lejeuni
sm. (+BR), ex Herb. Lugd.-Batav., Anon. s.n. (+BR, NY), ex Scheidweiler Herb.,
Anon, s.n. (+BR), 6051A Wall. ce Bene sn. (NY), ex Herb. College S. S.
Trin., Harvey (GH), ex Herb. Lu ed.-B atav., T. sn. (GH), Anon. 2 (MEL), ex
Hernhard Herb., Goring 181 (mo). Fonnoss: Tait6ty6, Suibotei, Ee ati
19678 (A); Urajiro-maki Taichu Prefecture, Hayata & Mori 7147 (a). Origin
unknown: part of specimens on Wall. 6052A, not P. neriifolius (+BR

Cultivated. CHtna. Kiangsu: Nanking, Chen & Teng 4104 (uc); Maan Shan
Inin Shan, Tso 1634 (A); Tungting, Tai Hu Lake, Soochow, Ching & Tso 731
(A). Hup ene Ho-Ch’ang Chow 1972 (A, NY). Bulan: Foochow, West Lake
Park, Chang 4360 (+mMo); Wooshihshau, Chung 2491 (a, uc); Electric Factory
Garden, Chung 2700 (A, uc); Hinghwa, via 1304 (A), 7515 (a); Amoy region,
Kulangser, Chung 1635 (A, UC). SINGAP Bot. Gard., Sargent in 1903 (a),
Bohn’s Gard., Sargent in 1903 (A). Tee ex Hort. Cantab. Anon. in 184-
(xy), Tokyo Gardens, Faurie 59 (mo). CEYLON: Hakgala Bot. Gard., Wilson

io)
a
i)
5
Pas
a
j=)
pan
oo
\o
rs
a
as]
q
toe
A
5
N
9
=
3
oe
iS
x
be
a
me
aoe
>
Nt

476 JOURNAL OF THE ARNOLD ARBORETUM _ [VOL. XxxIx

in 1921 (+a). Monaco: Hort. Monac., Anon. in 1824 (+BR). SWITZERLAND:
Locarno, ex Schneider Herb., Anon. s.n. (+A). FRANCE: Riviera, Schneider s.n.
(ta). ENcLanp: Lucas 1777 (+BM), Kew, Anon. s.n. (mo). Ivaty: ex Herb. R.
Bot. Hort. Neap., Amon. in 1868. (+Mo), ex Bernhardi Herb., Anon. s.n. (+Mo),
ex Horti Thenensis Herb. Bossche 1375 (+Br). Arrica: Cape of Good Hope,
Anon. in 1836 (tmo), Hort. Daudin, Anon. in 1851 (+ps); Grootvaderbosch,
Esai sm. (tups). Mauritius or Mapacascar: Blackburn in 1873 (+cvu).

UELA: Caracas, Pittier in 1924 (+vEN), Orozco 324 (+F). BraziL: Rio de
eee Bot. Gard., Curran in 1915 (tus), deLaubenfels in 1952 (+1LL). WeEsT
Inpies: Broadway 2934 (+r); Trinidad, Port of Spain, Broadway s.n. (BM);
Jamaica, Chrysler 1914 (NJuU), Anon, sn. (t1LL). Unitep States. New York:
New York Bot. Gard. 3750A from DPM 1900, Taylor in 1906 (ny), Hartling
in 1914 (Ny), Nash in 1905 (ny), Anon. in 1920 (1LL). New Jersey: Greenhouse
at Rutgers, Johnson in 1952 (tnyu). District of Columbia: Washington, Con-
gressional Garden, Parry in 1918 (us). South Carolina: Columbia, Hough in
1916 (cas); Charleston, Sargent in 1914 (A). Georgia: Augusta, Cuthbert in
1896 (nyu); Cairo, Wight in 1948 (+1LL). Florida: Gotha, Nehrling’s Place,
Rehder in 1920 (a); Glen St. Mary, Hume in 1926 (A); Winter Haven, McFar-
lin 4944 (micH); Panama City, Tyndall Field Hospital area, Gray in 1945 (+1LL) ;
Jacksonville, Huger in 1922 (ny), Dahlgren in 1945 (F). Missouri: St. Louis,
Bot. Gard., Woodson 321 (Mo), Irish in 1896 (mo), G. £. in 1873 (mo), Engle-
mann in 1877 (mo). Louisiana: New Orleans, Kock in 1936 (cu). California:
Hales place, Santa Barbara, Walther in 1921 (+cas). Java: Hort. Bot. Gard
iia VF 18, Anon. 1903 (Ny—2 sheets), VF 16, Anon. in 1903 (Ny). NEW
GUIN Aiyura, Womersley NGF3397 (LAE). AUSTRALIA: Brisbane, private
oa White B11 (A),

This variety is distinguished from the species by the narrower and
shorter leaves, not over 7 cm. long nor 7 mm. wide. The anatomy of the
leaf in transverse section does not differ from that of the species.

LITERATURE CITED

1. Baker, R. T., anp H. G. SmiruH. A research on the pines of Australia.
Sydney.

2. BUCHHOLZ, J. T., AND N. E. Gray. A taxonomic revision of Podocarpus.
I. The sections of the genus and their subdivisions with special reference to
leaf anatomy. Jour. Arnold Arb. 29: 49-63. 1948; IV. The rigs species
of section Eupodocarpus, subsections C and D. Jbid. 123-

3. Compton, R. H. Gymnosperms. Jour. Linn. Soc. Bot. 45: ean 1922.

4. FLorIN, R. Untersuchungen zur Stammesgeschichte der Coniferales u.
Cordaitales. Erster Teil. Morphologie und Epidermisstrukture der Assimi-
lations-organe bei den rezenten Koniferen. Svenska Vet.-Akad. Handl. III.
10. 1931.

5. Foxwortuy, F. W. Philippine gymnosperms. Philip. Jour. Sci. Bot. 6:
149-— ae 1911.

6. Grpps, L. S. A contribution to the flora and plant formation of Mt. Kina-

balu nd the highlands of British North Borneo. Jour. Linn. Soc. Bot. 42:
1-240.

1958] GRAY, REVISION OF PODOCARPUS, XI 477

7, Lee, I. Early explorers in Australia. From the log-books and journals, in-
cluding the diary of Allan Cunningham, botanist, from Mar. 1, 1817 to Nov.
19, 1818. London. 1925

. Metcatr, F. P. Flora of Fukien and floristic notes on southeastern China.

Lingnan University: 1942.

Orr, M. Y. The leaf anatomy of Podocarpus. Trans. Proc. Bot. Soc. Edin-

burgh 34(1): 1-54. 1944.

10. Stites, W. The Podocarpeae. Ann. Bot. 26: 442-514. 1912.

11. WasscHER, J. The genus Podocar pus in the Netherlands Indies. Blumea

4: 359-481. 1941.

ee)

Ke)

AGNES ScOTT COLLEGE
DECATUR, GEORGIA

478 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

ONTOGENY OF THE SPORANGIA IN XIPHOPTERIS
SERRULATA AND PYRROSIA NUDA}!?

KENNETH A. WILSON

IN SPITE OF ALL the recent discussion of relationships and classification
of such well-known genera as Vittaria and Polypodium, and such of their
presumed allies as Dipteris, Cheiropleuria, and Platycerium, no attempts
appear to have been made to correlate the details of sporangial structure
with cytological and other morphological evidence (e.g., the nature of the
gametophytes) which has recently been accumulating. It is surprising
that, in view of the emphasis which is placed on the reproductive parts of
flowering plants, the sporangial structure has for so long escaped detailed
investigation. Undoubtedly one of the main factors in causing this delay
in their study is the opinion that the sporangia of the “higher” or poly-
podiaceous ferns are so simple that they do not by themselves give ade-
quate evidence of relationship (Holttum, 1954). It is true that these
minute sporangia are deceptively similar in appearance. However, before
one can determine accurately whether sporangial structures are comparable
it is important to know the precise details of leptosporangial ontogeny.
Ontogenetic knowledge should lead to (1) clearer concepts of homology
between parts of sporangia of different species, and (2) demonstrations
that some apparent homologies do not exist. Only then can the details of
the mature sporangium be clearly understood, and only then can there be
a basis for the interpretation of presumably homologous parts of a large
number of mature sporangia.

In an earlier paper (Wilson, 1958), the ontogenetic steps in the forma-
tion of the mature sporangium of Phlebodium aureum (L.) J. Sm. (Poly-
podiaceae sensu strictu) were presented. It was shown that the sporangial
stalk results from cells intercalated in the first segments of the sporangial
primordium rather than by the activity of an apical cell as has been gen-
erally believed. Furthermore, it was found that the stalk of the sporan-
gium and the jacket of the capsule are produced by the subdivision of 5
initials or “segments.” Segment O contributes only to the formation of
the stalk, segment I to a portion of the stalk and part of the proximal face
of the capsule, segment II to the stomial region, the stalk, and to the
proximal and distal faces of the capsule, and segments ITI a IV to the
rest of the annulus and to both the proximal and distal faces of the capsule.

However, the study of sporangial ontogeny in Phlebodium still left
several questions unanswered. Phlebodium has a two-rowed stalk; how
is a one-rowed stalk produced? Do the sporangia of another genus in
the same family develop in the same manner? What difference, if any,

A portion of a dissertation submitted in partial fulfillment ek the requirements for
the degree of Doctor of Philosophy in the University of Michig

1958] WILSON, ONTOGENY OF SPORANGIA 479

would be found in the ontogeny of the sporangia in a species of a different
family?

In an attempt to answer these questions two fern species were chosen for
study: Xiphopteris serrulata (Sw.) Kaulf. (Grammitidaceae) and Pyrrosia
nuda (Gies.) Ching (Polypodiaceae). In the present paper, the develop-
mental sequence in the formation of the sporangia in the two species is de-
scribed and a comparison is made with the ontogeny of the sporangia in
Phiebodium aureum.

ACKNOWLEDGMENTS

I am particularly indebted to Dr. Warren H. Wagner, Jr. for his patient
and helpful directions and criticisms during the entire course of this in-
vestigation. The material of Xiphopteris serrulata used in this study was
collected by me during an expedition to Jamaica, The West Indies. I am
deeply grateful to Dr. Grady L. Webster and to Professor H. H. Bartlett
for making this field study possible.

MATERIAL AND METHODS

The plant material of Xiphopteris serrulata was collected in Portland
Parish, Jamaica, in 1954 (K. A. Wilson & W. Murray 588). Pyrrosta
nuda is growing at the Botanical canes of the University of Michigan,
and the material was collected there (UMBG 19980; from Assam, India,
W. Koelz 11716A). Specimens of both species have been deposited in the
Herbarium of the University of Michigan, and duplicates of X7phopteris
serrulata have been distributed to several other herbaria.

Young sori which had been preserved in F.A.A. were selected, embedded
in paraffin, and sectioned. Xiphopteris was sectioned at 12 microns, while
Pyrrosia, because of the larger size of the young sporangia, was sectioned
at 15 microns. The sections were stained in Conant’s quadruple stain
(Johansen, 1940). Sori in various stages of development were also cleared
by the sodium hydroxide technique (Foster, 1949) and then stained in 3
per cent tannic acid in 50 per cent alcohol and 3 per cent ferric chloride in
50 per cent alcohol.- After dehydration in alcohol, the sporangia were teased
out and mounted in Diaphane. Because of the thickness of the frond of
P-yrrosia and the deeply sunken sori, the above method, which was success-
ful for Xiphopteris, proved to be of value only for the sporangia in the
later stages of development. In order to study the earlier stages, 75-micron
sections of fixed Pyrrosia material were cut on the freezing microtome. These
sections were cleared and stained as outlined above and then mounted in
Diaphane without further dissection. When bleaching was necessary, a 50
per cent aqueous solution of Clorox was used. The mature sporangia were
studied in water mounts since dehydration led to the dehiscence of the
capsule. All illustrations were made with the aid of a camera lucida.

480 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

MORPHOLOGICAL OBSERVATIONS

1. Xiphopteris serrulata

The swollen sporangial initial of Xiphopteris serrulata is first divided by
a cell-wall which is slightly inclined but which never reaches the level of the
receptacular cells (figs. 2, 3). This division is soon followed by the forma-
tion of a second wall which is transverse and intercalated in the lower cell
of the initial (figs. 4, 5). This wall is produced on a level with the surface
of the adjacent receptacular cells, and separates the sporangial primordium
from a basal cell. The basal cell undergoes no further division and takes
no part in the subsequent development of the sporangium. The sporangial
initial at this stage consists of two cells: a proximal cell, segment 0, and a
distal cell, the mother initial.

The mother initial divides and cuts off, by a series of three oblique walls,
segments I, IT and III, which contribute to the jacket layer of the capsule
and the distal, three-rowed portion of the stalk (figs. 6-9). In the mean-
time, horizontal intercalary divisions take place in segment O (figs. 8, 9).
After the formation of segments I, II, and III, the mother initial, now
shaped like an inverted three-sided pyramid, divides horizontally and pro-
duces a transverse wall which cuts off segment ITV and thereby becomes
completely enclosed by its daughter cells (fig. 10).

The enclosed tetrahedral mother initial divides in the same manner and
in the same order as it did in producing segments I, II, III, and IV, so that
it becomes enclosed by four cells, the tapetal initials, which separate it from
the cells of the sporangial wall (figs. 12-14). Each of the tapetal initials
becomes partitioned into four cells by means of a vertical and a horizon-
tal anticlinal wall (figs. 13-16), and all sixteen cells produced by these divi-
sions then divide periclinally to form a two-layered tapetum, which soon
disorganizes, surrounding the inner sporogenous cells (fig. 17)

After the mother initial divides to produce the four tapetal initials, the
activity of the central cell changes and it divides equally to form the
sporocytes (fig. 17).

Fics. 1-37. SPORANGIAL ONTOGENY IN Xiphopteris serrulata AND Pyrrosia nuda:
INTERNAL SEGMENTATION. Fics. 1-17, Xiphopteris serrulata: 1, Protruding
initial; 2, 3, Formation of first wall; 4, 5, Formation of segment 0; 6, 7, For-
mation of segment I; 8, Intercalary division in segment 0; 9, Formation
of segment II or segment III, intercalary division in segment I; 10, For-
mation of segment IV; 41, Intercalary division in segment IV; 12-15, For-
mation of tapetal initials; 16, Cross section of primordial capsule; 17, Forma-
tion of sporocytes. Fics. 18+37, Pyrrosia nuda: 18, Protruding initial; 19,
Formation of first wall; 20, Formation of segments 0 and I; 21, Intercalary
divisions in segments 0 and I; 22, 23, Formation of segment II or segment II;
24, Formation of segment IV; 25-28, Formation of tapetal initials; 29-31, For-
mation of sporocytes; 32-37, Various stages in the ontogeny of the paraphyses.
Arrows point to newly formed walls. Roman numerals identify sporangial seg-
ments.

1958 | WILSON, ONTOGENY OF SPORANGIA

481
$

RD B® ED EP
we) SE? BRN) came
Ey ey ee Ne é
0 No Ss SY

g 14 15

oh BS
. (alos

Fics 1-37. SPORANGIAL ONTOGENY IN Xiphopteris serrulata AND Pyrrosta nuda.

482 JOURNAL OF THE ARNOLD ARBORETUM _ [VoL. xxxIx

All divisions that take place in segment 0 are horizontal; there are never
any vertical walls formed in the segment. As a result of the position and
inclination of the first wall formed in the sporangial initial, segment I never
reaches the level of the cells of the receptacle. Therefore, since only seg-
ment O contributes to the base of the sporangial stalk, and since the divi-
sions within this segment are only transverse, the stalk of the mature
sporangium consists of a single row of cells at its base (figs. 59-61). On the
other hand, the upper portion of the stalk beneath the capsule is three-
rowed. The cells in the lower portion of segments I, II, and III each con-
tribute to the formation of one of the three rows of cells which subtend the
capsule.

While the tapetal initials and sporogenous cells are dividing, the jacket
initials are becoming partitioned anticlinally to produce the exterior layer
of the capsule. The first two divisions in segment I are transverse (figs.
41-45), and these are followed by the formation of still another cross wall
in the distal portion of the segment so that segment I becomes composed
of a linear series of four cells (figs. 46, 47).

Both segment IIT and segment III become similarly partitioned by cross
walls; however, segment IT is divided by two transverse walls, while seg-
ment III becomes partitioned by only one such wall (figs. 42, 43a).

The cleavages in the first three capsular segments are accompanied, or
soon followed, by a division in the cap cell, segment IV, which produces a
wall, difficult to see in whole cleared sporangia, that extends from segment
II to segment III (fig. 45a; figs. 62, 63, wall v). Following the first division
of segment IV into two cells, each one of the cells undergoes a division by a
wall perpendicular to the first one (fig. 46); figs. 62, 63, walls h).

The uppermost of the four cells of segment I divides, by the formation of
a vertical wall, into two cells of approximately the same size (fig. 48);
fig. 63, wall v). This vertical division is at times followed by a horizontal
division in one of the two cells (figs. 54a, 57a); more frequently, however,
these cells remain undivided (figs. 53a, 58a). The distal cells of segment
I are immediately recognizable in mature sporangia and may be seen to
form a portion of the capsular jacket on one side, the proximal face, of the
sporangium (figs. 59, 61). The three cells below them, on the other hand,
produce one of the three rows of cells at the base of the capsule.

The division pattern of the other capsular segments is more complex and
may be understood more easily by considering the sequence of division in
each of the segments separately. As the developing sporangium continues
to enlarge, the uppermost of the three cells of segment II becomes divided
by a vertical wall (fig. 456; fig. 63, wall v), and each of the cells thereby
produced now divides by forming a horizontal wall (fig. 45a; fig. 63, walls
h). The cells next to segment III do not become further divided, and they
can be recognized in the mature sporangia as forming part of one of the
lateral walls of the capsule (fig. 60). These first divisions in segment II are
essentially identical to those of segment I, and, as will be shown later, a
similar division sequence occurs in segment ITI.

Additional divisions in segment II all contribute to the formation of the

<1
SS>

as =

me,
ee

es

bog
Setaee’
eeess

S77

Fics. 38-58. SPORANGIAL ONTOGENY IN Xiphopteris serrulata: SUPERFICIAL
SEGMENTATION. Both sides of each sporangial primordium are illustrated and
are designated by the letters “a” and “b.” For explanation of shading see Fics.

8.

484 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxtx

stomium and its associated cells. The cell which undergoes the first divi-
sion is directly related to the position of segment II with respect to segment
I. It is always the uppermost cell contiguous to segment I which be-
comes partitioned by a vertical wall. It must be pointed out here that, in
contrast to Phlebodium aureum, segment II is very rarely produced on the
left side of segment I; in the great majority of sporangia segment II will
be found to be lying to the right of segment I.

The distal cell of segment IT which is contiguous to segment I becomes
partitioned by a vertical wall (not shown in its first appearance; see wall r,
fig. 63). The uppermost daughter cell which borders on segment I under-
goes no further divisions and contributes to the formation of part of a
lateral wall. However, the other daughter cell becomes partitioned by the
intercalation of a horizontal wall (shown by arrow, fig. 53a; fig. 63, wall
e). An additional horizontal wall partitions each one of these last-formed
cells and produces the stomial cells (fig. 566; fig. 63, walls s). Asa rule
the distal cell next to segment III remains undivided; however, on occasion
it may become partitioned by a vertical wall (fig. 58b). Before the forma-
tion of the stomium, a horizontal wall is usually intercalated in the cell di-
rectly beneath the one which is divided by the vertical wall (i.e., the distal
cell contiguous to segment I; shown by arrow, fig. 57); fig. 63, wall f).
Following the divisions which produce the stomium, no additional cleavages
occur in segment II. The two lowermost cells of segment II form part of
the three-rowed stalk at the base of the capsule (fig. 60).

Occasional failure of wall formation in segment II may lead to the for-
mation of unusually large cells in the stomial region, as may be seen in
figure 61. This condition is a result of the absence of the vertical wall
which delimits the annular cells in segment II. A similar situation has been
observed in segment IIT, which would result in the formation of large bow
cells (fig. 55a).

As was stated above, the distal cell of segment III becomes dissected
first by a vertical wall (fig. 63, wall v) and then by two horizontal walls
(fig. 63, walls #) in much the same manner as segment II. The next divi-
sions mirror those of segment IT except for the formation of an additional
wall. Both of the distal cells become divided by the formation of a vertical
wall (figs. 50a, 51a; fig. 63, walls r), after which the cell adjacent to the one
which borders on segment I is partitioned by a horizontal wall (fig. 56a).
One wall forms in the cell below (see wall f, fig. 63), and with the produc-
tion of this series of cells the subdivision of segment III is completed. As
the sporangium enlarges and matures, the walls of the smaller cells in seg-
ment III, which are aligned in a vertical series, become thickened to form
the cells of the bow, while the lowermost cell of this segment contributes to
the formation of the three-rowed stalk (fig. 59). The cells between the
bow cells and the cells of segment II do not become further subdivided
after the division of the distal cell by the vertical wall. These cells all con-
tribute to the formation of part of the lateral face (fig. 60), while the cell
on the opposite side of the annular ring and in contact with segment I
contributes to the other side of the capsular face (fig. 59).

1958] WILSON, ONTOGENY OF SPORANGIA 485

S
gy
S
W

Fics. 59-63. MATURE SPORANGIUM OF Xiphopteris serrulata. 59, Proximal
face; 60, Distal face; 61, Proximal face: segment II with abnormally large
stomial region; 62, 63, Diagrammatic analysis of segment derivatives in the
mature sporangium; 62, Top view: segment IV (adapted from Kiindig); 63,
segments I, II, III (adapted from Miiller). Roman numerals identify sporangial
segments. For explanation see text.

486 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

The first divisions of segment IV into four cells are followed by a division
of the two cells contiguous to segments IIT and III, which produces in each
cell a wall that is parallel to the first-formed wall. In other words, these
walls are intercalated in the sector of segment IV on that side of the first-
formed wall which is away from segment I (see walls 7, fig. 63). This
division sets the boundaries for the formation of the annular cells in the
cap segment. A series of walls perpendicular to the first-formed wall and
to the wall parallel to it produces a series of cells which differentiate into
the top of the annulus. The greatest variation in the division pattern of
Xiphopteris serrulata is seen in the partitioning of segment IV. Generally
this segment contributes seven cells to the annulus; however, it may at
times be observed that only six cells are produced in this series (fig. 60).
The portion of segment IV which contributes to the proximal face of the
sporangium usually becomes divided by walls formed in different planes:
one wall is formed vertically while the other is variously inclined (figs.
59, 61; walls k, figs. 62, 63). The placement of the last cell wall in the
sector of segment IV in the distal face of the sporangium also varies (wall
HWE 02 F285, 520, $7 0)x

2. Pyrrosia nuda

One of the most striking differences in the young sporangia of Pyrrosia
nuda from those of Xiphopteris, as well as from those of Phlebodium, is
in their size. The sporangial initials are themselves almost twice as large
as those of Phlebodium, and well over twice the size of those of Xiphopteris
(fig. 18). However, in spite of the size of the sporangial initial, the first
two divisions are the same as those of Phlebodium. The first wall formed is
oblique and extends to the level of the receptacular cells or below it (fig.
19). The second wall is horizontal and in line with the surface of the cells
of the receptacle (fig. 20).

There seems to be some variation in the orientation of the wall which
produces segment I. It may be noted that at times this wall is formed in
such a way that it does not bisect the wall of segment 0, as it does in Phle-
bodium, but rather is so oriented that segment I fails to include the entire
basal portion of the initial which is not occupied by segment 0. Conse-
quently, the lower portion of the mother initial reaches the base of the
primordium (fig. 65). In such circumstances, when segment II is formed
it will extend to the very base of the sporangial primordium (figs. 67a,
70a)! This is a rare occurrence. Usually the divisions are as in Phlebodium
and the base of the primordium is occupied by segment O and segment I
only (fig. 69), resulting in a two-rowed stalk. When, however, the wall of
segment I is so placed that segment II reaches the base of the sporangium,
the entire stalk of the mature sporangium will be three-rowed. Although
such stalks are rare, they have been observed in mature sporangia of
Pyrrosia.

The details of the further development of the sporangium do not vary
from those of either Xiphopteris or Phlebodium. Of some significance may

1958] WILSON, ONTOGENY OF SPORANGIA 487

be the fact that a larger number of cells form the bow in Pyrrosia than in
either of the other two genera (segment IV contributes twelve cells to the
annulus), but they are formed in the same manner as has been described for
Xiphopteris (see figs. 18-31, 64-84).

Stellate paraphyses occur intermixed with the sporangia of Pyrrosia
(fig. 85). These arise from superficial cells of the receptacle which be-
come divided by a transverse wall (fig. 32). Following the initial trans-
verse divisions, the terminal cell of the developing paraphysis becomes bi-
sected by a vertical wall (fig. 33). These two terminal cells continue to
elongate, and soon, as a result of a protrusion and expansion of the cell
wall beneath them, a third cell is initiated (fig. 34). Elongation of these
cells apparently does not take place at the same rate; the newly formed
cells appear to elongate more rapidly. This process is repeated many times
(see fig. 37) and gives rise to the many-celled paraphyses. Vertical divi-
sions in the stalk may be seen to accompany the formation of the elongated
cells. On the other hand, no walls are formed in the terminal cells of the
paraphysis. The elongation of the terminal cells ceases soon after the walls
at their bases become thickened.

DISCUSSION

The orientation of the first division of the sporangial initial is not always
the same in different species. In Pyrrosia nuda, as well as in Phlebodium
aureum, the first wall formed is oblique and reaches the level of the recep-
tacular cells. On the other hand, the first wall in Xiphopteris serrulata is
essentially transverse and well above the cell surface of the receptacle.

Wagner (1952) has shown the same variation in the species he examined.
The first division of the sporangial initial of Diellia and Asplenium is trans-
verse, while in Nephrolepis and Davallia it is oblique. Kiindig (1888) was
the first to notice the two different orientations of this cell wall. Accord-
ing to him, the first wall is oblique in Asplenium and all the other species
he investigated, except Polypodium vulgare, where the first wall is trans-
verse. The observations on Asplenium and Polypodium are no doubt in
error, and the correct interpretation would be the reverse of that given by
Kiindig. As will be explained below, the structure of the mature stalk di-
rectly reflects the orientation of the first division of the sporangial initial.

Although no division figures were seen in the stalk-forming segments of
Xiphopteris serrulata and Pyrrosia nuda, there is no evidence to indicate
that the walls are not intercalated to produce the sporangial stalk in these
species as they are in Phlebodium aureum. There is certainly no indication
that the stalk is produced by the activity of an apical cell.

Wagner (1952) postulated that the two-rowed stalk is produced entirely
by the oblique orientation of the first wall, and the one-rowed stalk results
from the transverse position of this first-formed wall. The sporangial stalk
of Xiphopteris is one-rowed and is indeed a direct result of the horizontal
orientation of the first wall. All cell walls intercalated are transverse, never

488 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxx1x

>

C

ee a Ss Ve
ray p>
Rptpony

SEGMENTS AND

THEIR DERIVATIVES

Fics. 64-78, SPORANGIAL ONTOGENY IN Pyrrosia nuda: SUPERFICIAL SEGMENTA-
TION, EARLIER STAGES.

1953]

Ny)

\f | Sy 2 “iN
\ y eo \ AN

Se
by BH
=f

J

‘
ae} ) iy
I) d

ad

Fics. 79-85. SPORANGIAL ONTOGENY IN Pyrrosia nuda. 79-82, Superficial seg-
mentation, later stages; 83, 84, Mature sporangium; 83, Proximal face; 84
Distal face; 85, Mature paraphysis.

490 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. XxxIx

longitudinal. Stalk formation in Xiphopteris agrees with that in Diellia.
However, one may question Wagner’s conclusion with regard to the develop-
ment of the two-rowed stalk. In Phlebodium aureum the sporangial stalk
is always two-rowed and the first division is oblique. The first wall formed
in the sporangial initial of Pyrrosia is also oblique, but the stalk is not al-
ways two-rowed: it may be three-rowed instead. As has been pointed out in
Pyrrosia, the two-rowed condition is not solely the result of the orientation
of the first wall, but depends also on the placement of the wall which forms
segment I. When segments 0 and I do not include the entire basal portion
of the initial, the stalk will be three-rowed as a result of the cells con-
tributed to it by segment II. The significance of the first wall in the struc-
ture of the stalk may then be summarized in this manner: The one-rowed
stalk is produced entirely by the orientation of the first division,” while
the two- and three-rowed stalks depend on the orientation of both the first
division and that which produces segment I. The distal portion of the
stalk is always three-rowed and is formed by the lower cells of segments I,
II, and III.

Bower (1915) reported on the development of the sporangia of Dipterts
and Cheiropleuria. The sporangial stalks of these ferns are four-rowed, and,
according to him, the segmentation of the young sporangium “appears to
show a regular cleavage of the segments in two opposite rows,” and the
“subdivision of the two rows of segments of the stalk by walls in the plane
of the drawings has given rise to the four rows of cells of the stalk, as seen
in later stages.” In view of the preceding discussion which points out that
there is no apical cell activity in the production of the stalk in the species
examined, I suggest that a reinvestigation of the sporangia of both Cheiro-
pleuria and Dipteris should be made, stressing stalk formation and, because
of its potential morphological significance, the inner and outer segmenta-
tions of the capsule. For the present, Bower’s conclusions must be accepted
with reservation.

The internal divisions following the formation of segment IV are the same
in Xiphopteris serrulata, Pyrrosia nuda and Phlebodium aureum. The four
tapetal initials are produced by the mother initial in the same sequence and
in the same manner as are the four capsular segments (segments I-IV).
Only after the enclosed mother initial has produced the tapetal initial does
it undergo a change in activity and begin to divide equally to produce the
sporocytes. Moreover, as in Phlebodium aureum, there is no evidence to
indicate that the derivatives of the tapetal initials are potentially sporo-
genous, nor is there any to support the view that any of the derivatives of
the inner cell behave in the manner of a tapetal cell. Thus, in view of the
manner in which the tapetal initials are formed, the opinion that they are
better considered as inner wall cells, at least ontogenetically, is strongly
supported. The central cell may be considered sporogenous only after these
initials have been cut off from the mother initial.

There has been much disagreement in the application of the term “arche-

*A possible exception may be found in the sporangial stalk of Dictymia, which
possesses a single cell at its base.

1958 | WILSON, ONTOGENY OF SPORANGIA 491

sporium.” Kiindig (1888), Miller (1893), Campbell (1905), Bower
(1923), and Smith (1938) used it to refer to the central tetrahedral cell
before the tapetal initials are formed. However, Kny (1895) and more
recently Troll (1954) have limited its usage to the central spore-producing
cell. Since the term “‘archesporium” is intended to refer to the first cell
generation of the sporogenous tissue, it should be used only to designate
the truly sporogenous initial. A more comprehensive survey of the use, or
perhaps misuse, of this term was made by Bower (1908), who suggested
that it be retained ‘merely in a descriptive sense, in those cases where
the cell or cells which give rise to the sporogenous group are obvious, but in
a descriptive sense only.” It might be better still to avoid the use of the
term and prevent confusion. This has been done by Eames (1936) and
Wagner (1952), although Eames considered the central cell, prior to the
formation of the tapetal initials, to represent the “primary sporogenous
tissue.”

Eames also reported that from the ‘‘primary sporogenous tissue” are cut
off by “‘periclinal divisions one or two layers of thin cells which become
the tapetum.” J have not seen any evidence to support this statement. All
of my material indicates that only one layer of tapetal initials is cut off,
which, as a result of periclinal and anticlinal divisions, develops into a two-
layered tapetum. It may be possible that in some ferns two series of tapetal
initials are produced, but there is no evidence at present to support this
statement.

The subdivision of the capsular segments is remarkably similar in
Xiphopteris, Pyrrosia and Phlebodium. In all three, segment I contributes
to a portion of the stalk and part of the proximal face of the capsule, seg-
ment II to the stomial region and the stalk, and segments III and IV to
the rest of the annulus. The proximal face is formed from cells of seg-
ments I, II, III, and IV, and the distal face from those of segments IT, IIT,
and IV.

As has been pointed out by Wagner, a comparison of the ‘‘eusporangiate
method” of sporangial development with the “‘leptosporangiate method”’ is
highly desirable. Bower (1923) described a series showing gradual steps
from the segmentation typical of eusporangiate ferns to that of leptospo-
rangiate ferns. Similar series are also described by him for the sporangial
stalk and for the capsule. His conclusions indicate that there is little funda-
mental difference between the two types and that the facts ‘‘appear to estab-
lish the general sequence of forms from the Eusporangiate to the Leptospo-
rangiate, as a valid evolutionary progression.”

I have shown that there has been a great deal of misunderstanding of
the ontogeny of the polypodioid leptosporangium, and, in view of this,
comparisons based on data of uncertain validity may be seriously ques-
tioned. A review of the literature on the development of the sporangium
of the Marattiaceae and the Ophioglossaceae shows that little is known
about the process and that the various authors do not agree. (For a sum-
mary of this information see Campbell, 1911, and Bower, 1923).

Before any detailed comparison can profitably be made between the two

492 JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

types of development, new investigations should be made of the ontogeny
of the sporangium of the dui ferns, and of the so-called ‘“‘primi-
tive’ leptosporangiate fern

Pirard (1947) ee the stellate paraphyses of Niphobolus
|Pyrrosia| lingua, but these do not agree in their details with those of
Pyrrosia nuda. The paraphyses of P. nuda do not have the thickenings in
the cell walls of the stalk as do those of P. lingua, nor are the elongated
cells borne on a swollen cell, but rather on one of approximately the same
dimensions as the cells of the stalk. Since the various types of stellate pa-
raphyses have been used as taxonomic characters in the genus (Giesenhagen,
1901), it is hardly surprising that those of these two species should be dif-
ferent.

SUMMARY

In a study of the ontogeny of the sporangia of Xiphopteris serrulata and
Pyrrosia nuda, it is shown that the sporangial stalk is produced by the inter-
calation of cell walls in the first-formed segments of the sporangial pri-
mordium. A one-rowed stalk results directly from the horizontal orientation
of the first division of the sporangial initial; two- and three-rowed stalks
are produced by an oblique first division, and the number of rows depends
upon the orientation of the wall forming segment I. The stalk is produced
from the division products of the initial designated as segment 0 and the
lowermost cells formed from segments I, II, and III. The stomial region
develops in segment IT, and the remainder of the annulus forms in segments
III and IV. It is again suggested that the tapetal initials are better inter-
preted as inner wall cells of the capsule, and that the term ‘“‘archesporium”
be limited to designate the cell which directly gives rise to the sporocytes.
Comparison of the leptosporangium with the eusporangium must be de-
ferred until more information on both sporangial types is available. The
paraphyses of Pyrrosia nuda are also described.

ARNOLD ARBORETUM AND GRAY HERBARIUM
HARVARD UNIVERSITY

LITERATURE CITED

Bower, . O. 1908. The Origin of a Land Flora. 727 pp. MacMillan, London.
——. 1915. Studies in the phylogeny of the Filicales. V. Cheiropleuria bicus pis
ay ts and certain other related ferns. Ann. Bot. 29: 495-529.
The Ferns. Vol. I. Analytical Beaminaiton of the Criteria of
a 359 pp. University Press, Cambridge, England.
CAMPBELL, D. H. 1905. The Structure and Development of Mosses and Ferns.
7 pp. MacMillan, London
. The Rusporangiataé. The comparative morphology of the
Ophioglossaceae and Marattiaceae. Carnegie Inst. Wash. Publ. 140. 229
Pp.

1958] WILSON, ONTOGENY OF SPORANGIA 493

Eames, A. J. 1936. Morphology of Vascular Plants. Lower Groups. 433 pp.
McGraw-Hill, New York.

Foster, A. S. 1949. Practical Plant Anatomy. Ed. 2. 228 pp. D. Van Nos-
trand, New York.

GIESENHAGEN, K. 1901. Die Farngattung Niphobolus. 223 pp. Gustav Fisher,
Jen

a.

Hotttrum, R. E. 1954. The classification of ferns: The present position and
some thoughts on future deve a ents. Huitiéme Congr. Intern. Bot. Rap-
ports & Communications, Sect. 2, 4, 5, 6: 5-8.

JouanseNn, D. A. 1940. Plant Microtechnique. 523 pp. McGraw-Hill, New
York.

Kny, L. 1895. Entwickelung von Aspidium Filix mas Sw. posers Wandta-
feln mit erlauterndem Text, Abt. 9. 411-438. Paul Parey, Ber

ee J. 1888. Beitrage zur Entwicklungsgeschichte des Se ee

giums. Hedwigia 27: 1-11

Miuien. C. 1893. Zur Kenntniss der Entwickelungsgeschichte des Polypodiaceen-
sporangiums. Ber. Deutsch. Bot. Ges. 11: 54-72.

Prrarp, N. 1947. Sporanges, paraphyses, et organes connexes chez les fougéres.
Cellule 51: 155-184.

SmitH, G. M. 1938. Cryptogamic Botany. Vol. II. Bryophytes and Pterido-
phytes. 380 pp. McGraw-Hill. New York.

Trott, W. 1954. Allgemeine Botanik. Ein Lehrbuch auf vergleichend-biologischer
Grundlage. 749 pp. F. Enke, Stuttgart.

ane W. Hz, Jr. 1952. The fern genus nee Its structure, affinities and

nomy. Gain Calif. Publ. Bot. 26:

Watson, K. A. 1958. Ontogeny of the eaten ‘of Phlebodium (Polypodium)

aureum. Am. Jour. Bot. 45: 483-491.

494 JOURNAL OF THE ARNOLD ARBORETUM - [vot. xxxrx

THE DIRECTOR’S REPORT

THE ARNOLD ARBORETUM DuRING THE FiscaAL YEAR ENDED
JUNE 30, 1958

IT Is WITH PLEASURE that I report another year of excellent progress in
the care and use of the living collections, the improvement of the herbarium
and library and the contributions of the various staff members. A dry
summer, followed by a wet winter marked the climatic year; nevertheless,
the fine display of flowering shrubs, trees and vines seen in the Arboretum
during the spring gave evidence of the care they had received throughout
the months of difficult weather conditions. An increasing number of visitors
has been recorded and, of these, the registration book in the Administration
Building indicates visitors from thirty-eight states and seventeen foreign
countries. In addition to the regular work assignments of the staff, the
answering of all types of horticultural questions by letter, by phone, or
for individuals calling at the office in person consumes a considerable pro-
portion of the staff working hours and is an indication of the confidence
which the public places in the Arnold Arboretum

Among the highlights of the year were the visits of distinguished col-
leagues from this hemisphere and abroad. In all, directors or representa-
tives of twenty-three botanic gardens and arboreta visited the Arnold
Arboretum during the year. The visit of Dr. George Taylor, director of the
Royal Botanic Gardens, Kew, a newly appointed member of the Board of
Overseers Committee to Visit the Arnold Arboretum, was of special interest
and pleasure,

The Staff:

I am pleased to report the approval of the Harvard Corporation for
the promotion and appointment without limit of time of Dr. Carroll E.
Wood, Jr., as associate curator of the Arnold Arboretum. This appoint-
ment is well deserved and is indicative of the contribution which he has
made to the Arboretum.

Mr. Roger Coggeshall, who has been chief propagator at the Arboretum
since 1954, resigned to accept a position with a private nursery in Massa-
chusetts. Mr. Coggeshall has done an excellent job in increasing the living
collections, in research on the propagation of woody plants and in teaching
popular classes in the education program conducted by the Arboretum. We
regret his leaving the staff, but wish him success in his newly chosen field.
He will continue, however, to assist in the instruction program in plant
ges at the Arboretum.

Alfred Fordham has been reassigned and promoted to the position
of a propagator. Mr. Fordham began his work at the Arnold Arboretum

1958 | THE DIRECTOR’S REPORT 495

as assistant to Mr. William Judd in plant propagation. He received fur-
ther training in this field as an exchange student from the Arboretum to the
Royal Botanic Gardens, Kew, England, before the beginning of World
War II. Upon his return to the Arboretum, following a leave of absence for
war service, he was appointed assistant superintendent of buildings and
grounds. We are pleased that Mr. Fordham is again serving in a capacity
which will permit him to use his specialized training.

During the year Dr. Frans Verdoorn, holder of the honorary appoint-
ment as research associate at the Arnold Arboretum since 1942, accepted a
new position as director of the Biohistorical Institute at the University of
Utrecht, Holland. Dr. Verdoorn was born in Holland and thus that coun-
try regains not only a citizen but a remarkable botanical historian, as well.

Three holders of joint annual appointments with the Gray Herbarium
accepted other positions at the end of 1957. Dr. Howard F. L. Rock, who
was on the curatorial staff in the herbarium, accepted a position at the
University of Tennessee. Drs. Robert Bennie Channell and Charles W.
James, who, with Dr. Wood, were working on a flora of the southeastern
states, accepted teaching appointments at Vanderbilt and the University of
Georgia, respectively. To fill one of the latter positions, the joint appoint-
ment was made for Dr. Kenneth Wilson to work with Dr. Wood on this
project.

Horticulture:

During the summer of 1957 the Arboretum experienced the worst drought
on record. All but one of the ponds on the grounds dried up completely and
the remaining one retained but a few inches of water. As the drought
lengthened and the available -vater in the soil was exhausted, wilting be-
came a familiar sign in all shrub and tree collections. Emergency steps
were taken to water all recently planted trees and shrubs and those others
growing in especially dry areas. A pressure-driven deep-watering system
was employed to supply water directly at the roots of some of the larger
trees, especially those in the conifer collection. Although injury resulted
to many of the plants in spite of our efforts, it now appears that only a few
plants were killed completely.

In the late fall and winter the situation was reversed, with rains so unu-
sually heavy that the ponds filled and overflowed with floodings that
blocked the meadow road for a period of several days. With the assistance
of maps from the City of Boston Parks Department, some long-unused
drainage channels were relocated and opened to lower the water level. It
would seem that this excessive rainfall during a period of mild temperatures
partly offset the dry summer season. Flowering was affected in many plants,
particularly the azaleas, and much branch damage was noted, which re-
quired heavy pruning during the spring.

A number of improvements have been made on the grounds during the

ear. The spectacular collection of torch azaleas on the South Street
bank was thinned and weeded, a retaining wall was built around the

496 JOURNAL OF THE ARNOLD ARBORETUM - [voL. xxx1x

Kalopanax pictus bordering one of the ponds and in the lilac collection
many older varieties between the path and the road were removed to loca-
tions farther up the bank, allowing newer varieties to be placed in the more
conspicuous locations. It is our aim to bring together in a central location
the Lilac Committee’s “100 best lilacs.’ About half of these, all young
plants, are now in their new locations.

A major adjustment was attempted in the Carya collection where a very
few species and varieties dominated the area. A number of duplicate trees
were removed, these to be replaced with taxa not currently in our collection.
The area was cleaned to allow access to the newly opened woodland along
Center Street, which is the area proposed for an expanded planting of /lex
pedunculosa, its relatives and other rarer species.

The Boston Parks Department continued cleaning the cobblestone gutters.
The main portion of the Arboretum between the Arborway and Walter
Street has been completed and work has started on the Peters Hill tracts.
This work, neglected for many years, makes a striking difference in the ap-
pearance of the grounds. Unfortunately, it also calls attention to the poor
condition of the roads in many places, a condition which we hope can be
handled by spot repairs immediately and eventually by retarring or re-
surfacing throughout.

Increased police protection was also granted by the Police Department,
which assigned two mounted patrolmen to the Arboretum grounds during
the spring season. Men on horseback are better able to patrol the remote
areas of the Arboretum and can reach areas not accessible by patrol cars.
The impressive riders and steeds have attracted considerable attention and
have been photographed almost as much as the famed Davidia.

One new pest, the dogwood twig borer, was discovered in our collections
this year. A heavy infestation of this insect was found in the Viburnum
collections, apparently attacking all species, which may be the reason why
some of the Viburnums have not appeared to be doing well. The location
of the pest is extremely difficult. We now spray for this borer, combining
the spray program with that for borers in the locust and mountain ash col-
lections.

Dr. Donald Wyman has continued to act as co-ordinator for a group of
botanic gardens and arboreta working in conjunction with the Horticul-
tural Crops Research Branch of the U. S. Department of Agriculture.
This group is responsible for bringing into the United States for trial
growth under quarantine conditions those plants whose entry previously
has been prohibited by law. Such plants are grown in special houses at
Glen Dale, Maryland, for two or more years before being turned over to the
arboretum or garden responsible for their introduction. During the year the
Arnold Arboretum received a quarantine release on twenty plants previ-
ously held at Glen Dale. These plants originated in England, Germany and
the Netherlands.

In continuation of the program of plant distribution established at the
Arnold Arboretum seventeen years ago, fourteen new species and varieties
were made available to cooperating nurserymen for propagation. In all 650

1958] THE DIRECTOR’S REPORT 497

Above: “Up the hill.” Photo and print by Mr. Henry Soron awarded a prize
as the best black-and-white picture taken in the Arboretum.

Below: Officers Donahue and McNeil, two of the mounted policemen of the
Boston Police Department now patrolling the Arboretum.

498 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxix

plants were distributed to thirty-four domestic nurseries, four foreign nur-
series and ten arboreta and botanic gardens.

The number of new plantings on the ground during the year was slightly
less than average. Individual plants added to the collection totaled 463,
representing 221 species and varieties. Forty-one are new to our collections
and eight are probably new to the United States.

Experimental work in the field of plant propagation, involving 646 species
and varieties, continued in the greenhouses. One program concerned the
propagation of native hybrid azaleas which are considered impractical to
root on a commercial scale. A similar program for rooting Ghent azaleas
under plastic was attempted. Various hormone concentrations were tried,
with Hormodin No. 3 giving the best results. Once the cuttings were rooted
the plants were placed under continuous light. Daylight was supplemented
with artificial light for continuous illumination for a period of three months.
The extended period of light caused a break in bud dormancy and the
plants grew continuously, This appears to be a method for establishing
hardier plants capable of being transplanted at an earlier age.

In a continuing experimental program of rooting Asiatic maples from
cuttings, it was established that wounding the cuttings was definitely ad-
vantageous and that Hormodin No. 3 proved to be the best rooting hor-

ne.

We were able to overcome the common double dormancy in seeds of
Cotoneaster with the use of concentrated sulphuric acid. Seeds were soaked
for two hours in the acid, washed well and stratified in plastic bags for
three months at forty degrees Fahrenheit. This treatment produced a per-
centage of germination at the end of one year comparable to that normally
obtained in the standard two-year treatment for seeds of this genus.

During the year 136 shipments, totalling 1137 species and varieties of
seeds, living specimens, and propagating material, were received from in-
stitutions in the United States and ten other countries. We sent 26 ship-
ments of seeds of 141 species and varieties and 320 shipments of propagat-
ing material or plants comprising 1328 taxa to sixteen countries besides
the United States.

Two pieces of property owned by the Arboretum were recommended to
the Corporation for sale and both were sold during the fiscal year. The tract
of land at 310 South Street consisting of a house and barn with two acres
of land was the gift of Mrs. Andrew J. Peters in 1953. After expending a
considerable sum to fence the property and tear down a rambling frame
addition to the basic brick house, it was discovered that the house was not
structurally sound and could not be used as a staff residence. The property
was on the Boston tax rolls at an inflated valuation, the taxes exceeding
$1500 a year. Protecting the property from vandalism proved difficult and
its sale was recommended. The receipts from the sale of this property will
be added to the Arboretum endowment.

The second property sold was the Butler property on Center Street, ac-
quired as a residence for Dr. Merrill while he was director of the Arboretum
and occupied by him until his death in 1955. Since the cost of repairing

1958 | THE DIRECTOR’S REPORT 499

this house was prohibitive, it was sold, together with the land on which it
stands, and the proceeds of this sale will also be added to the endowment.
By the disposition of these two pieces of property the Arboretum has been
freed of two costly maintenance items.

Case Estates:

The plantings at the Case Estates include demonstration plots, nursery
stock and plantings for hardiness testing. This area is attracting an increas-
ing number of visitors each year. Particularly noticeable is the increase in
requests from garden clubs and other horticultural organizations for guided
tours through the area. During the year portions of the land were made
available to staff members and graduate students of the Cabot Foundation,
the Bussey Institution and the Biology Department for their individual
projects, in addition to the land used for Arboretum purposes.

The desire of the Town of Weston to acquire portions of the Case Estates
for schools was mentioned in the report for 1956-57. Both the director of
the Arboretum and the Corporation made it clear throughout that the land
would not be sold and that the town must take this tract by eminent do-
main. In a special town meeting in 1957 it was voted to take approxi-
mately seventy of the one hundred and forty-five acres then in the prop-
erty. However, further recommendations of the School Site Committee
resulted in a reappraisal of the town’s school needs. After much discusssion
by the committees, officials and residents, the town finally voted to take a
smaller tract of land, which proved to contain thirty-two and one-half
acres, but at the end of the fiscal year the formalities involved in taking
this land by eminent domain had not been completed.

The loss of this area will necessarily reduce our opportunity for expand-
ing the work carried on at Weston, and will also reduce the usefulness of
the Case Estates to Harvard students, the Arboretum adult education
classes and groups from the Weston schools. Any further inroads on the
Arboretum land in Weston would raise serious problems affecting our abilty
to continue the important work now being carried on there.

During the early town meetings called for discussing the Weston school
needs, there was a group which favored locating a new high school on the
Case Estates in addition to the elementary school previously proposed.
However, this proposal did not receive the necessary support and the town
voted to acquire a privately held property some distance from the Case
Estates for this purpose.

Education Program:

Two series of classes consisting of lectures, field trips and demonstrations
were presented in the regular fall and spring adult education program this
year. The field trips conducted by Dr. Wyman at Jamaica Plain remain
extremely popular and were supplemented by a series of field trips called
“Field Botany,” conducted by Drs. Howard and Wood at the Case Estates
in Weston. All classes were open to the public, with advance registration

500 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXXIXx

essential, since in nearly all cases the class size was restricted. The courses
are stimulating to students and instructors alike, but, regrettably, can not
be expanded beyond the present effort. There seems to be an increasing
demand for education in horticulture and botany through more popular,
non-credit courses such as those offered in this program. Fortunately, five
other organizations in the Boston area are now offering similar courses.

For nearly two months (from the last part of April until the middle of
June), numerous groups visit the Arboretum for guided tours through the
collections. Such tours are scheduled in advance and a maximum of six
tours in one day has been accommodated. During other seasons the re-
quests for tours are less frequent.

For the New England Zone meeting of the Garden Club of America
which was held at the Arboretum in May, the staff assisted with a tour and
the presentation of a display of tree peonies arranged by Mrs. George L.
Batchelder, Jr. and Mrs. Ellery Sedgewick. The Librarian’s Club of Bos-
ton also met at the Arboretum with the supervision of our librarian, Mrs.
Schwarten. Tours of the grounds and a discussion of the services of the
Arboretum staff were arranged. All available staff members were required
during the Annual Field Day of the Massachusetts Horticultural Society
held at the Arboretum. Seven buses were used for transportation during a
two-hour conducted tour of the grounds. Since the buses could not seat
all those who wished to take advantage of the tour, a separate caravan of
cars supplemented the buses.

The Arboretum sponsored a special colloquium open to the public at the
Biological Laboratories when Dr. George Taylor of the Arboretum Visiting
Committee spoke on his work in Tibet. Colored slides illustrated his talk
for the audience which nearly filled the hall.

Staff members represented the Arboretum at various meetings during the
year. Dr. Howard gave the general lecture at the banquet of the Nursery-
man’s Short Course at the Waltham Field Station, Dr. Wyman took part
in the refresher short course for arborists, also held at the Waltham Field
Station. Dr, Sax attended the rootstock conference of the American Society
for Horticultural Science and lectured on the production of dwarf fruit
trees to the Plant Propagators Society and the New England Association of
Nurserymen at their annual meetings. Drs. Howard, Kobuski and Sax at-
tended the American Institute of Biological Sciences meetings in Palo
Alto, California. Dr. Sax was reappointed as an American Institute of Bio-
logical Sciences lecturer and during the year visited Reed College, Lewis &
Clark College and Whitworth College in Oregon and Washington, where
he gave a series of lectures on world agriculture and population growth.

Exhibits and Displays:

The Arboretum exhibit at the Spring Flower Show of the Massachusetts
Horticultural Society again won a gold medal as well as two special awards.
One of the latter, from the Massachusetts State Department of Agriculture,
was a new award at the show this year. The exhibit theme, ‘““A Hobbyist’s

1958] THE DIRECTOR’S REPORT 501

Two views of the exhibit of the Arnold Arboretum at the Spring Flower Show
held at Mechanics Hall. The exhibit was titled “A Hobbyist’s Rhododendron
Garden.”

502 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxix

Rhododendron Garden,” attempted to show the full program of care and
selection necessary in producing fine rhododendrons. The Beatrix Far-
rand Silver Bowl for the best display of rhododendrons was won by this
exhibit.

The open house held at Weston in early May was announced in the
Weston and Wellesley papers and was well attended by people from these
areas. We trust that those who attended gained a better appreciation of the
function of the Case Estates in the program of the Arboretum. It is pleas-
ing to learn that many garden clubs now carry the open house date on their
schedule of coming events.

In co-operation with the Boston Camera Club, a contest was held for the
best black-and-white photographs and the best kodachrome slides taken in
the Arboretum during the past year. It was interesting to the staff, who
acted as judges in the contest, to get a glimpse of those scenes and those
individual plants which caught the eye and the imagination of the outsider.
The best black-and-white print was made by Mr. Henry Soron of Arling-
ton and showed a small boy walking on Bussey Hill in the elm collection.
The best kodachrome slide of a general scene was taken by Robert Spooner
of Brookline and showed the Sargent cherries near the Administration
Building and the open field of daffodils and narcissi. Mrs. Karin Eng-
strom of Buzzards Bay won two prizes for the best close-up and the best
pattern shot. The former was of Magnolia soulangiana var. rubra showing
a mature flower and a pubescent bud. The latter was a beautiful picture
of the trunk and adventitious shoots of Cercidiphyllum japonicum enhanced
by backlighting. The contest for the best slide showing the horticultural
value of a single plant had the most entries and the judges concluded that
the outstanding slide was of a Pseudolarix amabilis in fall foliage taken
by Mr. George Taloumis of Salem. The contest served to bring photog-
raphers to the Arboretum at all seasons of the year, judging from the
entries.

Library:

The year 1957-58 brought to a completion the first stage in combin-
ing the non-horticultural collections of the Arboretum with those of the
Gray Herbarium and several other selected collections. Mrs. Schwarten
and her staff have successfully united, recatalogued and reconditioned these
books in the Harvard University Herbarium Building. This action was
voted by the Harvard Corporation on January 19, 1953, and progress
toward this end has been indicated in previous annual reports. The physical
properties of air conditioning, providing controlled humidity and tempera-
ture in a dust-free atmosphere, ensure a longer life for these books. The
new arrangement has already demonstrated that increased efficiency is pos-
sible in the use, as well as in the care of these volumes. Only the pamphlet
collection remains in need of reorganization.

In the course of this project, a relatively small number of duplicate

1958] THE DIRECTOR’S REPORT 503

books were found and set apart as reserves; these are shelved in a separate
section of the library. These include volumes which are not presently
needed and those held as replacements for volumes which receive much
use. Duplicates of monographic treatises and special family studies have
been placed on special shelves in the herbarium adjacent to the herbarium
specimens of the families concerned. The unification program also allows
greater office use or long-term use of these duplicate volumes by staff mem-
bers.

In the course of the unification work in the library, the director and
the librarian of the Museum of Comparative Zoology made available for
our use a large room for work and storage. We are particularly indebted
to them for this kindness.

The bound volumes added to the library during the year numbered 465.
These included 212 volumes transferred to our library from the libraries at
Widener, the Biological Laboratories and the special paleobotanical library.
Also included are 145 volumes dealing with horticulture which were added
to the library in Jamaica Plain. The total number of bound volumes ac-
cessioned now is 50,204. Pamphlets totalling 305 were added to the collec-
tion which numbers 16,573.

The horticultural library in the Administration Building in Jamaica
Plain was also reclassified and portions of it rearranged for more convenient
use by the staff. Additional library shelving was erected on the second
floor to house special collections. As previously indicated, 145 bound vol-
umes were added to this collection. The renumbering of all volumes in the
horticultural library has been initiated.

A set of the Gray Herbarium Index to American Plants is maintained at
Jamaica Plain and 3,000 cards in four issues were received and are being
added to this collection.

Again this year the number of requests for books on interlibrary loan
remains inconveniently high. Much as we wish to aid other workers, it is
impossible to honor all requests for books to be sent on loan to other institu-
tions. Where possible, requests for descriptions or passages are met by
sending typed, photocopied or microfilmed copies of the pages needed and
many other requests are referred to libraries within the home state of the
individual requesting the loan. In addition, no book over one hundred years
of age is ever sent from our library on interlibrary loan. Nevertheless, 217
volumes were sent out. The postage on such loans is usually paid by the
borrowing library, but the general wear and often the real damage to the
books is an expense we must bear. In contrast to the herbarium, where
loans sent to staff members in other institutions may be somewhat offset
by those received for the use of our own staff members, the interlibrary
loan requests made by Arboretum staff members rarely exceed a dozen in
the course of a year.

I wish to express my appreciation to Mrs. Schwarten and her staff for
the concentrated effort they have made and the results they have obtained
in bringing the library to its present excellent condition.

504 JOURNAL OF THE ARNOLD ARBORETUM | [vot. xxxrx

Herbarium:

It is now possible to report significant progress in the unification of the
non-horticultural herbarium of the Arboretum with that of the Gray Her-
barium. This work should be completed within the ensuing year. The ad-
ditional funds made available for this task by the Harvard Corporation
have enabled us to employ additional botanists who, with the herbarium
staff, have made excellent progress. The details of this work have been
discussed in previous reports. During 1955-56, the first year of this work,
103 families were integrated. In 1956-57, an additional 66 families were
completed. During the last fiscal year 94 families were organized, leaving
only 43 from a total of 306 families represented in our herbarium to be in-
tegrated. However, these 43 families include such large woody groups as
the Rosaceae and the Leguminosae, as well as primarily herbaceous fam-
ilies such as the Gramineae, Cyperaceae, Compositae, Liliaceae and Ran-
unculaceae. The progress that has been made is satisfactory and the results
encourage an even greater effort. The ease in using the reorganized and
combined collections has demonstrated that the integration is well worth
the time and effort expended.

As the work proceeds, excessive duplication in some areas within the
collections has become evident. For example, in an experimental project
it was found possible to remove 1058 sheets (either Arnold or Gray) from
the Pinaceae and 23 sheets from the genus Asimina. Such sheets are either
exact duplicates or sterile specimens represented by better material from
the same geographical location. The unification procedure also has re-
vealed a large number of herbarium sheets needing repairs. During the
course of the year our mounter repaired 2302 sheets in addition to mounting
for insertion 5535 sheets. The total number of specimens in the Arboretum
herbarium as of July 1, 1958, therefore, is 699,452.

During the year 11,516 specimens were received by the herbarium. We
received by exchange 6970, by subsidy 2708, by gift 1055, by special ex-
change 682 and for identification 101 specimens. The collection obtained
by subsidy represented a set of plants obtained by Mr. Leonard Brass dur-
ing the Fifth Archbold Expedition to New Guinea and adjacent islands.
While the Arnold Arboretum had subsidized and handled all the botanical
collections of the earlier expeditions, it was not able to do so in this case.
A special collection of plants from Hong Kong and the vicinity was ob-
tained in exchange for publications. In total, we received 993 specimens
from the Western Hemisphere and 10,523 from institutions in or workin
on the floras of the Old World. These figures reflect the Arboretum’s his-
torical interest in the flora of Asia.

As a service to our colleagues, specimens from the herbarium of the
Arnold Arboretum are sent on loan to recognized institutions. In turn, our
staff borrows specimens from other herbaria to assist in their studies. The
shipment of specimens on loan is necessarily an involved bookkeeping pro-
cedure. If the request is specific it may be necessary to send only a single
specimen. Often, however, the request will be for a group of species, a

1958] THE DIRECTOR’S REPORT 505

genus, specimens from a certain geographic area, or representatives of sev-
eral families. Each loan request must therefore be handled individually
and may involve a few minutes or a full week of work. When the specimens
have been withdrawn from the herbarium, a record of the loan is made in
the herbarium and in special files. Thus we know at all times the location
of our herbarium material. The specimens, carefully packed to prevent
breakage, are shipped by express collect within the United States, but
must be prepaid and insured when shipped outside of the United States.
Notice of the loan and the total count is sent in a separate letter with a
return slip to indicate the safe arrival of the loan. When the borrower has
finished his studies the specimens should be packed with equal care and
returned promptly. All specimens returned are immediately fumigated,
counted and checked against the original record to insure the receipt of
the material originally sent. The herbarium sheets are then examined
to determine the conclusions of the study and, if the sheets are properly
annotated, they are filed according to the classification suggested. Speci-
mens of cultivated plants from the horticultural herbarium in Jamaica
Plain are added to the loans of spontaneous plants or sent separately for
special study as required. Thus a great deal of time is consumed in the
preparation and return of a simple loan request.

During the past year the staff sent from the combined herbaria 95 loans
averaging 141 sheets and totalling 13,411 sheets. Sixty of these loans went
to institutions within the United States and 36 to institutions outside of
the country. In contrast, during the same period our thirteen staff members
and students made 71 requests for loans from 44 American and 27 foreign
herbaria, for a total of 8,614 specimens borrowed.

Routine work continues on the herbarium of cultivated plants in Jamaica
Plain. A start has been made on the project of locating the type specimens
of horticultural taxa described by Professors Alfred Rehder and Charles
Sargent. During the year there was an increase in the requests for repre-
sentative specimens, kodachromes or black-and-white pictures of culti-
vated plants.

In preparation for his “Manual of Cultivated Trees and Shrubs” and
the “Bibliography of Cultivated Trees and Shrubs,”’ Professor Rehder
maintained a card file of references to original publications and to illustra-
tions of cultivated plants. After Rehder’s death the administrative deci-
sion was made to curtail that work. This decision was unfortunate, since
the reference file soon became of limited value. With the recent adoption
of an international code of nomenclature for cultivated plants, the
need for an up-to-date index of published cultivar names, together with
references to illustrations, became imperative. Therefore, during the past
year Miss Ethel Upham, formerly of the Arboretum staff, began the work
of bringing up to date the “Rehder Index.’”’ Over six hundred cards have
been added to the file in the past year and it is expected that in two years
the index will be current.

The staff members continued to spend a small part of their time on
individual research projects. Unfortunately, this time has been limited in

506 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

the past three years and will continue to be for the ensuing year or until
the herbarium integration is complete.

Dr. Howard continued his several projects supported by special research
grants. Progress was made in the survey of the anatomical structure of
the petioles of flowering plants and a report on this work was given at
the International Horticultural Congress. His work on the West Indian
vegetation was restricted to the vegetation in relation to bauxite subsoils,
and the flora of the Lesser Antilles. A monographic study of the genus
Coccoloba continues with final treatments prepared for the Lesser Antilles
and Trinidad and the record of the genus in cultivation.

Dr. Perry returned from her trip to Europe, having profitably studied
the New Guinea collections at herbaria in Great Britain, the Netherlands
and Switzerland. In addition to herbarium work, Miss Perry continues her
study of the materials from the 4th Archbold Expedition collected by
Leonard Brass in New Guinea.

Dr. Hu had previously been working on a Flora of China supported by
grants for that specific purpose. When the financial support was exhausted,
Dr. Hu devoted her attention to regular Arboretum projects and has under-
taken to identify the large accumulation of Chinese plants acquired by
Dr. Merrill shortly after the end of the war. As sizable numbers of dupli-
cates of these collections occupy our storage areas, the identification of
the main set will enable the staff to distribute these collections within the
next few years. Studies of the Compositae in China and miscellaneous
monographic studies for the Flora of China have been completed and await
publication.

Dr. Jarrett continued her work on the Moraceae and prepared several
papers on Artocarpus for publication. Miss Jarrett has also developed an
interest in the flora of the Philippines and is continuing that work which
has been a tradition at the Arboretum.

Dr. Johnston reports that his work concerns the Boraginaceae of Texas
and that work is in progress on the Chenopodiaceae and on the gypsum and
serpentine floras.

Dr. Kobuski’s heavy curatorial responsibilities which include supervision
of the herbarium integration, have left him little time for his own re-
search. The Theaceae continue to be his interest with special studies in
progress on the genus Ternstroemia.

Dr. Wood assumed the responsibilities of editor of the Journal of the
Arnold Arboretum with the January number of the current volume. He
continued during the year his studies on the genera of seed plants of the
southeastern United States; the genera comprising the woody families of
the Ranales as they occur in that area were prepared for publication.

Mrs. Weber and Dr. Wilson, both of whom are on special projects, de-
vote their full time to these. As a hobby of botanical interest, Mrs.
Weber has prepared for publication a check list of the plants and plant
families represented in philately, Dr. Wilson has readied several of his
studies on ferns for publication and devotes his current research to further
studies on this group of plants.

1958] THE DIRECTOR’S REPORT 50

~I

Comparative Morphology:

The assistance of I. W. Bailey, Professor of Plant Anatomy, emeritus,
as curator of the wood and pollen slide collections is greatly appreciated
by the staff. A number of wood samples and slides were added to the collec-
tions during the year. Requests for wood samples for study were mainly
from scientists wishing to add such data to their taxonomic monographs.
Fourteen such requests were filled from the wood collection and the
herbarium.

During the year Professor Bailey received a grant from the American
Philosophical Society in support of his research. His interest in the use of
pollen in morphological studies centers now on the variations between
fresh and processed pollen grains from herbarium material or fossil de-
posits. Professor Bailey’s study of the anatomical structure of the stem in
the Cactaceae continues to develop and he is accumulating samples for
further work.

Cytogenetics:

Dr. Karl Sax, his assistant and students have reported the following
contributions in the field of cytogenetics:

The mature trees of Malus sikkimensis, M. toringoides and an M. sar-
gentii hybrid (AA 33340) usually breed true from seed when open pol-
linated, but occasionally they produce some segregates which differ from
the typical in having larger flowers and fruits. Cytological studies show
that the present trees are triploids.

The maternal-type progeny are also triploids, but the large-fruited types
are tetraploids. Apparently the triploid parents are facultative apomicts
which usually produce triploid progeny, but occasionally produce sexual
segregates arising from a triploid egg cell and a haploid pollen grain from
a nearby diploid species. Such progeny would resemble the maternal parent
because of the excess maternal chromosomes.

Both Malus sargentii and M. sargenti rosea are facultative apomicts.
Pollinated with pollen from a diploid Malus species, the progeny of the
triploid M. sargenti rosea range from nearly diploid to approximately
tetraploid, while the sexual progeny of the tetraploid M. sargentu are
triploids.

Apparently the triploid parent produces egg cells ranging from haploid
to triploid, although the apomictic progeny are triploids.

An issue of Arnoldia was published during the year reporting on the
juvenile characters of trees and shrubs. Cuttings from suckers from the
base of a seedling apple tree were found to root much more readily than
cuttings from the fruiting branches of the same tree. The basal suckers
are also juvenile in their leaf characters.

Instruction:
Although no formal courses were offered by staff members during the
past year, both Dr. Howard and Dr. Sax continued to work with graduate

508 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

students on special problems. Mr. Claud Brown completed his work with
Dr. Sax on pine hybrids and, after receiving his Ph.D. degree, accepted a
position in forest genetics at Texas A. & M. College.

A series of luncheon seminars was conducted throughout the year in
the lecture room at the Harvard University Herbarium Building. These
seminars, in which the herbarium staff take part, provide an opportunity
for the staff and graduate students to meet informally to hear papers by
representatives of each group. The seminars also provide a means for visit-
ing scientists to meet and talk with the staff without the usual disruption
of work schedules.

Travel and Exploration:

Travel by staff members during the past year was divided between
attendance at professional meetings, the presentation of invitational
lectures and specialized field work connected with research programs.

The annual meeting of the American Institute of Biological Sciences is
held each fall at a college campus in the United States. As such meetings
offer the opportunity of observing other biology departments, gardens and
herbaria, as well as of meeting with professional colleagues, staff members
are encouraged to attend when possible. The 1957 meetings were held on
the campus of Stanford University at Palo Alto, California, and Drs.
Howard, Kobuski and Sax made the special trip. The California flora,
both native and introduced, is strikingly different from that seen in New
England and scheduled field trips for interested registrants afforded the
opportunity of seeing the coastal range vegetation as well as a bit of the
redwood forests. Special trips were made by Drs. Howard and Kobuski
to the California Academy of Sciences to confer with colleagues there and
to the Strybing Arboretum in Golden Gate Park. Mr. Coggeshall attended
the Plant Propagators Annual Meeting at Cleveland and Dr. Sax attended
the American Society for Horticultural Science rootstock conference at
Geneva, New York.

Every three years an International Horticultural Congress is held which
is attended by persons interested in horticulture from all over the world.
The Fifteenth International Horticultural Congress was held in Nice,
France, April 11-18, 1958. Dr. Howard represented the Arnold Arbore-
tum at these meetings, presenting one paper and reading another for Dr.
Kobuski. Prior to the meetings Dr. Howard visited botanic gardens in
Denmark and Switzerland. During the Congress field trips were conducted
to allow the registrants an opportunity of seeing the natural vegetation of
the area and the commercial and private aspects of horticulture.

Following the meetings, Dr. Howard visited Paris for some work at the
Muséum National d’Histoire Naturelle and also visited many of the fa-
mous parks and gardens of Paris. In and around London a schedule
arranged by Dr. George Taylor allowed Dr. Howard to see several famous
arboreta such as Westonbirt and many gardens, both public and private.

Dr. Hu attended the 9th Pacific Science Congress held in Bangkok,

1958] THE DIRECTOR’S REPORT 509

bove: A group of the participants at a plenary session of the XVI Interna-
tional Aor Congress held in Nice, France.

Below: Members of the Congress during a visit to the Exotic Gardens at
Monaco

510 JOURNAL OF THE ARNOLD ARBORETUM © [voL. xxxIx

Thailand, and presented four papers at these meetings. Arranging her trip
eastward and circling the globe, she was able to visit botanic gardens and
herbaria in England, France, Italy, Switzerland, India, and several coun-
tries of Asia. Following the congress Dr. Hu visited Hong Kong and
Macao for field work and study furthering her work on the flora of China.
An exchange of specimens was also arranged by Dr. Hu between the
Arnold Arboretum and both the Gardens Department, Hong Kong, and
the Department of Botany, National Taiwan University.

Field work associated with individual research projects was carried on
by several staff members. Drs. Wood and Wilson collected materials for
their work on the flora of the southeastern states in Florida, where they
were the guests of Mr. and Mrs. George R. Cooley.

Dr. Johnston spent the summer months doing field work in Texas.

Dr. Howard made another trip to Jamaica to continue his study of the
vegetation on bauxite soils when several additional areas in St. Elizabeth
Parish became available. The results of earlier studies showing the nature
of plantings desirable for the rehabilitation of mined-out bauxite pits is
now becoming evident as the mining scars are gradually being covered
with forage grasses, crop plants and forest trees. A survey trip was also
made to Hawaii to examine areas being considered for mining operations
there. Following this trip Dr. Howard accepted an invitation to serve as
advisor to the Commissioner of Public Lands for the Territory of Hawaii.
He is to give special consideration to the problems of vegetation loss in
strip-mining operations and to advise regarding rehabilitation procedures
if mining is conducted in the Territory.

Gifts and Grants:

The response of the Friends of the Arnold Arboretum to the annual ap-
peal made during the spring was most generous. Most gifts received
were unrestricted and were assigned to the ‘“‘gifts for cultural purposes”
fund which is used to employ additional summer labor on the grounds where
needed. This fund also supports a research assistant for Dr. Sax and an
assistant in the greenhouses to help on plant propagation.

A special grant was awarded to Dr. Sax from the Massachusetts Society
for Promoting Agriculture for work on rootstocks and for methods for
promoting the growth and fruiting of fruit trees in New England. This
project is being carried on in conjunction with the Bussey Institution.

Dr. Howard received a three-year grant from the National Science Foun-
dation to continue work leading to a flora of the Lesser Antilles. Mr.
George Proctor of the Institute of Jamaica, Kingston, Jamaica is co-worker
on this project and will do the field work. During the last year Mr. Proctor
visted Grenada and St. Lucia for major field work and spent a few days
on other islands. Mrs. Dudley Hall assists Dr. Howard in the herbarium
work of this project and is also supported by this grant.

A number of outstanding plant species were given to the Arboretum,
either as living plants or as propagating material. The plants received

1958] THE DIRECTOR’S REPORT onal

from Mrs. J. Norman Henry, and Mr. Orlando Pride were particularly note-
worthy. Mr. A. G. Rotch presented to the Arboretum a portrait of the
Arnold family painted by an unknown artist. He accompanied this with a
genealogical summary which is a valuable addition to our historical records.
The portrait, of considerable value beyond its special interest to us, is at
present under the care of the staff of the Fogg Art Museum.

Publications:

Dr. Wyman edited the twelve numbers of Arnoldia, a bulletin of popu-
lar information which appeared during the year and Dr. Kobuski edited
the Journal of the Arnold Arboretum, the quarterly issues of which serve
the staff as an organ for the publication of scientific articles. Staff mem-
bers have priority in placing manuscripts in the Journal and contributions
from other individuals are by invitation or are papers based largely on the
plant materials in the living collections or the herbarium and wood collec-
tions of the Arboretum. The Journal serves also as a medium of exchange
for similar publications from other scientific institutions. The standards
of the Journal have always been high and the excellence of the editorial
work, together with the quality of the printing, has earned for it a good
reputation in the field of scientific publications. For the last nine years
the editor of this publication has been Dr. C. E. Kobuski, who has long
been on the editorial board associated with the Journal. In fact, excepting
only his leave of absence from the staff for war service, he has served con-
tinuously as associate editor or editor of the Journal since his appointment
to the Arboretum staff in 1927. To him the Journal owes its direct, lucid
style which, with characteristic regular date of issue, has become a hall-
mark of this publication. Dr. Kobuski has now asked to be relieved of
his editorial duties in order to devote more time to his curatorial respon-
sibilities in the herbarium and to his own research. Thus, with this volume
the Journal becomes the editorial responsibility of Dr. C. E. Wood, Jr.
Special acknowledgment to Dr. Kobuski was made in the January num-
ber of Volume 39.

One special publication was issued during the year. This was a photo-
reprint of “The Genus Pinus” by George Russell Shaw, issued previously
as “Publication of the Arnold Arboretum, No. 5,” in 1914. This publi-
cation has been out of print for many years but sufficient back orders had
accumulated to merit a reprinting. To date, over 150 copies of this still
important work have been sold.

It is with considerable pleasure that we note the presentation to Mrs.
Susan D. McKelvey of the Sara Gildersleeve Fife Memorial Award, which
was given in recognition of Mrs. McKelvey’s contributions to botanical
and horticultural literature. Mrs. McKelvey’s works include “The Lilacs,”
“Vuccas of the Southwestern United States’ and the recently published
“Botanical Exploration of the Trans-Mississippi West 1790-1850.” A
full citation of the award was given in the Garden Journal of the New York
Botanical Garden.

blz JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxxrx

Bibliography of the Published Writings of the Staff and Students
July 1, 1957 —June 30, 1958

Baitey, Irnvinc Wiper. Additional notes on the vesselless dicotyledon, Am-
borella trichopoda Baill. Jour. Arnold Arb. 38: 374-380. pl. 1, 2.
. Aggregations of microfibrils and their sper in - secondary wall
of ‘Coniferous tracheids. Am. Jour. Bot. 44: 415-4
. The potentialities and limitations of wood Since in ce study of the
phylogeny and classification of angiosperms. Jour. Arnold Arb. 38: 243-254,
1957.

. Die Struktur der Tiipfelmembranen bei a Tracheiden der Koniferen.

Holz als Roh- und Werkstoff 15: 210-213. 1957.

. Need for a broadened outlook in cell wall terminologies. Phytomorphol-

ogy 7: 136-138.

. The structure of tracheids i in relation to the movement of liquids, sus-

dendions, and undissolved gases. J: The physiology of forest trees. p. 71-
58.

CoccEsHALt, Rocer Grsss. Asiatic maples, their propagation from softwood
uttings. Arnoldia 17: 45-56. pl. 11, 12. o,
The propagation of some little- used shade and street trees. Proc. 33rd
Nat. Shade Tree Conf. 37-43. 1957.
Howarp, RicHarp ALDEN. The Director’s Report. The Arnold Arboretum dur-
ing the fiscal year ended June 30, 1957. Jour. Arnold Arb. 38: 389-403.
957.

. Studies in the genus Coccoloba, IV. The species from Puerto Rico and
the Virgin Islands and from the Balama Islands. Jour. Arnold Arb. 38:
211-242. 1957.
. Studies in the genus Coccoloba, V. The genus in Haiti and the Domin-
ican Republic. Jour. Arnold Arb. 39: 1-48. eae
———. The meadow. Arnoldia 18: 17-24. pl. 3-5.
(with GrorcE R. Proctor). New records 4 | aa flowering plants,
I. Jour. Arnold Arb. 39: 101-106. 1958.
Hu, Suru-yinc. Oriental Hollies. Nat. Hort. Mag. 36: 31-64. illus. 11-20. 1957.
JARRETT, Frances M. A note on the identity of the genus Balanostreblus (Mo-
raceae). Jour. Arnold Arb. 39: 107-110. 1958.
JounstToN, Ivan M. Borraginaceae. /m: Steyermark, Julian, et oe en
to the flora of Venezuela. Fieldiana. Bot. 28: 1080-1082. 19
Studies in the Boraginaceae, XXIX. Echiochilon and te genera.
Jour. Arnold Arb. 38: 255-293. 1957.
ith C. R. McCuLtoucu & J. M. Parker). Terrain study of the Pan-
ama Canal zone with specific reference to the Fort Sherman area and
vicinity. North Carolina State College, Department Engineering Research.
1-267. 6.
KoBuskI, CLARENCE E. The horticultural herbarium. Arnoldia 18: 25-28. 1958.
Sax, Kari. The control of vegetative growth and the induction of early fruiting
in apple trees. Proc. Am. Soc. Hort. Sci. 69: 68-74. 1957.
The effect of ionizing radiation on chromosomes. Quart. Rev. Biol. 32:
15-26, 1957.

1958] THE DIRECTOR’S REPORT obs

. Experimental control of tree growth and reproduction. /x: The physiol-

ogy of forest trees. p. 601-610. 1958.

. The genetic future of man. Ju: The Population Ahead. 1958.

. The juvenile characters of trees and shrubs. Arnoldia 18: 1-6. pl. 1.
1958,

ScCHWARTEN, LAzELLA. Index to American Botanical Literature. Bull. Torrey
Bot. Club 84: 324-336, 398-411, 459-471. 1957; 85: 78-88, 146-155. 1958.

WEBER, CLAUDE. Les mauvaises herbes d’origine Européenne en Nouvelle-Angle-
terre. Bull. Cercle Vaudois Bot. Lausanne 8: 34-39. 1957 (1958

—. Etude phytosociologique des prairies du Canton de Geneve et de ses
environs immédiats. Trav. Soc. Bot. Genéve 4: 20-38. 1956-57 (1958).

_ Note sur la colonisation des digues artificielles de galets en aval de

l’Usine de Verbois (Genéve). Trav. Soc. Bot. Genéve 4: 39-41. 1956-57

(1958).

. La végétation de la Pointe 4 la Bise (Genéve). Trav. Soc. Bot. Genéve
4: 47-53. 1956-57 (1958).

Witson, KENNETH ALLEN. Ontogeny of the sporangium of Phlebodium (Poly-

podium) aureum. Am. Jour. Bot. 45: 483-491. 1958.

Woop, CarroLt E., Jr. (with R. K. Goprrey). Pinguicula (Lentibulariaceae) in
the southeastern United States. Rhodora 59: 217-230. 1957

Wyman, Donatp. Broad-leaved evergreens in the Arnold Arboretum. Arnoldia
17: 61-76. pl. 15, 16. 1957.

—. Hollies for hedges, screens, and barriers. Nat. Hort. Mag. 36: 139-146.
illus. 47-50. 1957.

he new horticultural color chart. Arnoldia 17: 57- 60. Din Gay NO 5c

—— . Winter injury — 1957. Arnoldia 17: 37-44. pl. 10

A spring walk through the Arnold Arboretum. eae 18: 13-16.

1958.

. Two new Mahoberberis hybrids. Arnoldia 18: 9-12. pl. 2. 1958.

A new color chart for horticulture. Nat. Hort. Mag. 37: 47-49. 1958.

RicHarp A. Howarp, Director

514 JOURNAL OF THE ARNOLD ARBORETUM _ [VvoL. xxxIx

Staff of the Arnold Arboretum
1957-1958

RicHARD ALDEN Howarp, Ph.D., Arnold Professor of Botany, Professor of
Dendrology, and Director.

IrvING WIDMER BaltLey, S.D., Professor of Plant Anatomy, Emeritus.

JosepH Horace Fautt, Ph.D., Professor of Forest Pathology, Emeritus.

MicuarL, ANTHONY CaAnoso, M.S., Curatorial Assistant.*

RoGER GIBBS COGGESHALL, Propagator (Resigned April 15, 1958).
ALFRED JAMES ForpHAM, Propagator (Appointed April 15, 1958).
HremMan ArtHur Howarp, Assistant Horticulturist.

SHIU-vING Hu, Ph.D., Botanist, Flora of China Project.

FRANCES MAry JARRETT, Ph.D., Botanist and Assistant Editor.
IvAN Murray Jounston, Ph.D., Associate Professor of Botany.
CLARENCE EMMEREN KosuskI, Ph.D., Curator.*

MARGARET CATHERINE LEFAVouR, Herbarium Secretary.

Susan DreLano McKeE vey, A.B., Research Associate.

Lity May Perry, Ph.D., Botanist.

Kar SAX, S.D., Professor of Botany.

LAZELLA SCHWARTEN, Librarian.*

CLAUDE WEBER, Botanist, Special Project.*

STELLA MABEL WHITEHOUSE, Business Secretary.

KENNETH ALLEN WiItson, Ph.D., Botanist, Southeastern Flora Project.*
Rosert GERow WILLIAMS, B.S., Superintendent.

CARROLL Emory Woop, Jr., Ph.D., Associate Curator and Editor.
DonaLtp Wyman, Ph.D., Horticulturist.

* Appointed jointly with the Gray Herbarium.

1958] INDEX 515
INDEX

Abare, 275 mnie sata (Moraceae), - Note on

Abroma ee 253 e Identity of the Genus, 1

— nitida, rane ilicifolius, ee

one ne 362
Acanthospermum, 358, 390
Achillea, 359, 367, 381, 390
Acrodiclidium, 338

Acrosynanthus jamaicensis, 101-102

Ainsliaea, ae 371, 381, 410

Alfredia, 3

Allardia, ae 381

Alvaradoa ae 102-103

Ambrosia, 358

Anandri

Anakal 357, 366, 380, ee 410
Andrachne cuneifolia, 20

— sect. Atta,

— sect. Bee ee 314
Annonaceae, 309-31
A-no-kwa, 275

Anica aan: 357, 380, 390

3
Arnold eee during the Fiscal Year
foes ie ie 1958, The. The Direc-
r’s Re
nie 8 se, 381, 390, 410
Asimina, 311-3
Aster, 356, 364, a 390, 410

Awarivacabariyek, 275

Barkhausia, 363

Benzoin, 214, 342

Be-se-o-wa, 275

Rea of the Published Writings of
and ete July 1, 1957-

Bubby blossom, 323
Buphthalmum, 358
Bur-oo-ma, 2
Butneria, 323

Cacahuillo, 275

Cacahuio, 275

Cacaita, 275

Cacalia, 360, 369, 381, 390, 410
Cacao cahoua 7

— de monte, 275
— de murcielago, 275

516 JOURNAL OF THE ARNOLD ARBORETUM _ [vot. xxx1x

Cacao esquinado, 275 Cinnamonum sect. Camphora, 215, 335
— jacaré, 275 — sect. Cinnamomum, 335

— wee — sect. Malabathrum, 335

— montaras, 275 — camphora,

— montaraz, 275 — zeylanicum, 215

— quadrado, 276 Cinnamon, 335

n
Cacaoito de monte, 276
Cacaorana, 276

Citation of Some Genera of the Lau-

raceae, The, 21
Clusia portlandiana, 103

Caca-u, 276 Cnicus, 3
Cacau ae quina, 276 Cocesioba, Studies in the Genus, V. The
Cacaut, 276

Cacau-jacaré, 276

Pe enn 323-326

Carduus, 360, 370, 381, 390
Carpesium, 357, 367, 380, 410
Carthamus, 361, 381
Cassytha, 329, 344—346

a, 357

Centaurea, 361, i 390

Centipeda, 359

Cephalonoplos, i.

Chaenotheca ee 208

Chanekia, 338

Chasepthera neopeltandra, 208

Cha-te-ra, 276

China, Statistics of Compositae in Rela-
tion t o the Flora of, 347, 379

Chrysanthemum, om 359, 367, 381, 390

410
Cicca antillana, 208

3
innamomum, 215, 329, 334-336

Genus in Haiti and the Dominican Re-

public,
Coccoloba albicans, 6, 8—9
— bar rar
— barkeri, 19, 24
rear 25, 26
— — ovato- faiieeolaves 26
— buchii, 7, 9-12
— ceibensis, 6, es
— ciferriana, 9,
— costata, 7, 13- ee
—xC. wvitera 15
oo aes nsis,
= tiveaitoli 8, 17, 42

3
——™ C. uvifera, 23-24
— incrassata, 5, 24-25
— eiachiane 31
— krugii, 7, 8, 25-26
—— C. uvifera, 26-28

— laurifolia, 16
— leoganensis, : ‘id 30
6

— nalgensis, 34, 35

= nevroph i 34, 35
— nivea, 46

—no ee 6, 32 i

— obtusifolia,

— pungens, ‘18, 19

1958]

Coccoloba revoluta, 9
— rotundifolia, 28, 29

13
— samanensis, 6, 39-40
— samuelssonii, 13, 14
— saxicola, 4
— scandens, 26
— ee 47

-— ee 5, 40-42

—— eee 28

355;
Composites) of
of Neighboring

Compositae tribe Anthemideae, 351, 353,
1

368, 38
— tribe Arctotieae, 352, 353
— tribe Astereae, 350, 353, 365, 379
— tribe Calenduleae, 351, 353
— tribe Cichorieae, 352} 353, 372, 381
— tribe Cynareae, 352, 353, 368, 372, 381
— tribe Eupatorieae, 349, 353, 365, 379
— tribe Helenieae, 351, 353, 380
— tribe Heliantheae, 351, 353, 380
— tribe Inuleae, 350, 353, 365, 368, 380
— tribe Mutisieae, 352, 353, 372, 381
— tribe Senecioneae, 351, 353, 368, 381
— tribe Vernonieae, 349, 353, 365, 379
Conami ovalifolia, 95
— portoricensis, 208
Conyza, 357, 380

INDEX

517

Coreopsis, 358

Crepis, 363, 374, 382, 391, 410
Crossostephium, 359, 381
Croton glabellus, 208

— lucidus, 208
Cucumber-tree, 303
Custard-apple, 313
Cyathocline, 356, 379

Cynara, 361

Dahlia, 358

Decaisnea fargesii, 403

Deeringothamnus, 311

Diasperus angustifolius, 193
15

color, 7
— epiphyllanthus, 199
— erythrinus, 125
— grisebachianus, 61
— latifolius, 185
— mimosoides, 174
—montanus, 183
— myrtilloides, 126
— nummularioides, 137
— nutans, 56
— orbicularis, 118
— ovatus
— portoricensis, 208
— pruinosus,
— spathulifolius, 131
— speciosus, 188
— subcarnosus, 143
— williamioides, 83
Dichrocephala, 356, 379
Dimorphocladium, 111
—- formosum, 114
Dipholis dope 103-104
Diplopappus,
Director’s ee The. The Arnold Arbo-
retum during the Fiscal Year Ended
June 30, 1958, 494
Dominican Republic, The Genus in Haiti
and the. Studies in the Genus Cocco-
loba, V, 1
Doronicum, 359, 381
Dubyaea,
Duhaldea, 357

518

Echinops, 360, ee 381, 390

Eclipta, 358, 380, 390

Ee-so-pe-ke,

Shae ite ee 379, 390
381

Erigeron, 357, 364, 380, 390
Erythroxylon subcordatum, 40, 41
Ethulia, 355, 379

Eupatorium, 356, 364, 379, 390
Euptelea pleiosperma, 398

Evax,

Exocarpus epiphyllanthus, 199

Faberia, 363

Flueggea vices. 208
Formania, 359

Gaillardia, 358
Galatella, 357, 380
Galinsoga, 358, 380, 390

the Woody Ranales in the
Southeastern United States, The, 296
paces a

— speci

re eee von a 371, 381

enna: sect. nn 163

— bot

iat

ona 358, 3

Gnaphalium, 357, ae 380, 390, 410

Gorteria, 360

Grangea, 356, 379

Gray, Netra E. A Taxonomic Revision
of Podocarpus, XI. The South Pacific
Species of Section Podocarpus, Sub-
section

aaron uvitera 43

a. ehh 369, 381

Haiti and the Dominican Republic, The

JOURNAL OF THE ARNOLD ARBORETUM

[VOL. XXXIX

Genus in. Studies in the Genus Cocco-
loba V, :

Handelia,

Hee-ree- . na-pee-ta-re, 276

Helenium

Hehanthus: 3

Helichrysum, ae 360, 380

be,
, 381
Herrania, , Somos of the Genus, 217
Herrania, 2 29

sect. 229, 230
— sect. Swhevnbicdinn 229, 230
— albiflora, 227, 229, 230, 232-236

36-238

— breviligulata, 230, 238-239

— camargoana, 231, 239-242

— cuatrecasana, 231, 243-244

— dugandii, 232, 244-245

— guianensis, 274
kanukuensis, 231, 245-247

— kofanorum, 231, 247-248

— laciniifolia, 230, 248-250

— lemniscata, 230, 250-253

— mariae, 232, 253-2

— — putumayonis, 256-257

— nitida, 229, 231, 257-261

— — aspera, 258, 270

— — sphenophylla, 261-262

—nycterodendron, 232, 262-264

— pacifica, 267

— pulcherrima, 231, 264-267

— — pacifica, 267

— purpurea, 230, 267-270

— tomentella, 228, 232, ee

eboney 230, 272-
eters a, 362
Heteropappus, 259) 380

pane

e Director’s Re-
Arnold arr during the

Fiscal Year en June 30, 1958, 494
Howarp, Ricuarp A. Studies
eee V.
d the Dominican Republic, 1
RicHArp A. and GEORGE
Proctor. New Records of Jamaican
Flowering Plants, I, 101

1958]

Hu, Suru-yinc. Statistics of Compositae
in Relation to the Flora of China, 347,

Hypochoeris, 362, 382, 390

Tlliciaceae, 315-316
18

Ixeris, 363, 375, 382

ie pares Plants, I, New Rec-
ords of, :
JARRE TT, ee rs M.

A he
of the Genus Balanostreblus
7

Jurinea, 361, 371, 381, 410

Kalimeris, 356, 380
Koelpinia, 362, 382
Ko-kee-ot-chu, 276
Ku-ra-ta, 276

Lactuca, 362, 373, 382, 391, 410
Lagenophora, 356, 379

Lapsana, 362, 382, 390
Launaea, 362, 382
Lauraceae, The Citation of Some Genera
of the, 213
Lauraceae, 326-32
subfam. Cassy aoe 329
— Se Lauroideae, 3

— camphora, 215
cinnamomum, 215
— persea, 215

362
ae tae ih ee 380, 396, 410
Leucomeris, 361,
ae

SON
Licaria, oe 338-339
Lightia, 227

— lemniscata, 250

Ligularia, 359, 360, 369, ‘ee 410
Lindera, 215, 329, 342-34

— benzoin, 214

INDEX

519

Lindera umbellata, 215
Liriodendron, 302, 306— 309

Love-vine, 345

ae 302-306

Magnoliaceae, 299-302
Maipoilie doron doron, 276
Maivi

— hotteana,
— nobilis, 208
Tee oe

— tetraco
areree ome 276
Martinia, 3

Matayaka, 276
Nan 359, 381, 390
Mazzet 6

artes 357, 380

Mikania, 356, 379, 390

TMicentecas 338

Mi-to-ro-re, 276

Monographic Study of the West Indian
Species of Phyllanthus, A, 49, 111

Myripnois, 361

Nageia drouyniana, 444
—elata 7

Acree 329, 336-338

New Guinea, A New Species of Phaleria

(Taymeleacee from,

New Rec Jamaican Flowering
Plants, a ae

520

New Species of Phaleria or anaes
from New Guinea, 420

Note on the Identity of ‘tie Genus Bala-
nostreblus (Moraceae), A, 10

Nouelia, 361

Olgaea, 360
Omphalea eS 153, 159
157

Ontogeny of the Sporangia in Xiphopteris
serrulata and Pyrrosia nuda, 478

Ostryopsis davidiana 398
O-yaw-pee-ka-ye, 276

Palo de chimbe, 276
Palo de murciélago, 276
276

Parasenecio, 360
Parthenium, te 380, 390

. A New Species of Pha-
or aman from New

Petasites, 359, 381, 390

Phag 57, 380

Phaleria (Thymelaeaceae)
inea, A New Species

Phaleria elegans, 422

Phialanthus apne, 104-105

from New
of, 420

— sect. Apolepis, 207
— sect. !
— sect.
— sect.
— sect.
— sect.
— sect.
— sect.
— sect.
— sect.
— sect.
— sect.

Epistylium, és, 154, 206
Eriococcus,
Floribundi, 20
ennai 68, 160-161

JOURNAL OF THE ARNOLD ARBORETUM

[ VOL, XXXIX

Phyllanthus sect. Hemiphyllanthus, 69,
163-164, 207
—sect. Loxopodium, 206

— sect. Nothoclema, 207

— sect. Omphacodes, 68, 142-143, 207
— sect. Orbicularia, 68, 111-113

— sect.
— Sect.
— sect. Typophyllanthus, 179

—sect. Xylophylla, 69, 19 182, 207
— subg. Botryanthus, 4
— subg. Cicca, 206, 207

— subg. Phyllanthus, 207

— subg. Xylophylla, 66-68, 207

— acacioides, 165, 177-179
meri . . proferens, 198

— anderssonii, 51, 53-55

— angustifolius, 182, 193-195

—— genuinus, 193

—antillanus, 208

—apiculatus, 133, 134

— arbuscula, 182, 188-192

— axillaris, 153, 154, 159-160

— cauliflorus, 154, 155, 157-159
— chamaecristoides, 113, 132-133
— baracoensis, 133, 134-135
— cham eae ae: 133-134
— acidhivsens

— — palli

1958] INDEX 521

mae seas ence es comptus, 75 Phyllanthus nutans, 51, 56-57
—ekm ,» 97-1 —_—— pretachionie: 61-62
— eee a 182, 197-198, 199 — — nutans, 57-61
— dilatatus, 203-205 — — purdiaeana, 57, 58
— — domingensis, 202-203 — — trojanus, 57, 58
——epiphyllanthus, 198-202 — orbicularis, 111, 113, 118-122
—— genuinus, 199, 202, 203 — — ellipticus, 118
— Sr euehy 159 —_—— ee 118
— erythrinus, 125 — — obovatus, 11
—_ eee 139, 140 — ovatus, 163, 164, 169-171
— euwensii, 51, 52 — pachystylus, 51, 62-65
— excisus, 71, 85-87 — pallidus, 75
— falcatus, 199 — phlebocarpus, 114, 139-142
—foliis . . . floriferis, 199 — portoricensis, 208
— foliis Pr onbeen 85 — proctoris, 182, 195-197
— foliis . . . longioribus, 193 — pruinosus, 71, 73
— formosus, 113, 114-1 — pubigerus, 208
— foveolatus, 1 — purpureus, 125
— glabellus, 208 — quinquefidus, 148
— glaucescens, 50 0) ifolius, 118, 119
— grandifolius, 50, 146, 169 — sagraeanus, 75
— cornifolius, 151 — scandens, 208
— — genuinus, 148, 151 — scopulorum, 113, 135-137
— — salzmanni, 151 — shaferi, 130
— grisebachianus, 61 — spathulifolius, a
— hotteanus, 208 — speciosus, 188,
—— indequalifiorus, ae cae 191 — subcarnosus, es fea 146
— incrustatus, 71, — swartzii, a 190
— isolepis, 185, ea — trem
— fdelanedire lis: 146-148 = urbanians, 51, 65-66
— — cornifolius, 146, 148, 151-153 — vir
— juglandifolius, 148-151 — eee 70, 83-85
— latifolius, 182, 185-188 — xylophylla, 1
— laurifolius, 208 Picris, 362, 382, 391
— leonis, 143 Pityothamnus, 311
— linearis, 188, 195 Pluchea, 357, 380
— — cognatus, 193 Podocarpus, A Taxonomic Revision of, XI.
— genuinus, 188 The South Pacific Species of Section
— ae aes 164, 165-167 Podocarpus, Subsection B, 424
— mega apodus, 165, 171-174 Podocarpus acicularis, 437
— Fao decL s, 125 — affinis, 428, 445-446
ict) an 70, 80-81 —annamiensis, 428, 451-452
— mimosoides, 165, 174-177 — archboldii, 428, 452-453
—-—macrophyllus, 171 — — crassiramosus, 428, 453-454
— mirificus, 71, 89-91 — bidwillii, 443
montanus, 182, 183-185 — bracteata, is
— myriophyllus, 164, 168-169 — brassii,
— myrtilloides, 113, 123-124, 126 — brevifolius, oe 441-442
— — alainii, 124, 129-130 — brownii, 444
— —erythrinus, 124, 125-126 — celebica, 459
—— myrtilloides, 124, 126-127 — chinensis, 472, 475
— — shaferi, 124, 130-131 — costalis, 428, 456-457, 459
— pee oli: 125, 131-132 — deflexus, 428, 450-451
— eee aeran 208 — discolor, 461
— nobili — dispermus, 428, 448-449
— ee — drouynianus, 427, 444-445

eae 113, 137-139 — elatus, 426, 427, 436-439

Be? JOURNAL OF THE ARNOLD ARBORETUM _ [voL. xxxIx

Podocarpus ensifolia, 431, 437
— falcata, 437
— forrestii, 429, 458-459

—glaucus, 427, 441
ie snlonigr 428, 447
— japonica, 4

— ledermannii, 428, 447-448
— leptostachya, 461
— littoralis, 469
— macrophylla acuminatissima, 461
— macrophyllus, 429, 471-473
— — angustifolius, 429, 473-474
— — chingii, 429, 474
— maki, 426, 429, 474-476
— ee ee 475

— miquelia, 475
—nakaii, 428, 454-455
ae ata 461

9, 467

— — polyanthus, 429, 467-468
— — ridleyi, 435
RE 428, 468-469
— novae-caledoniae, 427, 431-432
— — colliculatus, 427, 432-433
Wi are 427, 434-435

pilgeri, 429, 459-460
a er 426, 429, 469-471
po see) i 471

— pungen
— ridleyi, pe on 435-436
— rumphii, 428, 455-456
Rapa 428, 449-450
— schlec 59
— sinen a “4S
— spinulosus, 427, 443-444

— sylvestris, 428, 446-447
— teysmannui, 468
ei alaeags 428, 457-458
— vrieseana, 475

—wangil, 442
Poilania, 357, 380
Polygonum uvifera, 43
Pondspice, 344
Prenanthes, 362, ng 382, 391

Proctor, GEOR R. and Ricuarp A.
Howarp. New a of Jamaican

Flowering Plants, I, 1
Pseudoelephantopus, oe
Pseudolinosyris, 356

Sia pa 340

Ptar

eeateae a 380

Pteronia, 356

Pulicaria, 357, 380, 390

Pyrethrum, 359, 381

Pyrrosia nuda, Ontogeny of the Sporan-
gia in Xiphopteris serrulata and, 478

Raisin la mer, 44

, 87

Ranales in ‘he Southeastern United States,
The Genera of the Woody, 296

Resinier, 16

Rhagadiolus, 362

Rhinactina, 356
Rhynchospermum, 356, 379
Roigia,

— comosa, 116
seats vn 390
Rus-ub, 2

Sacha cacao, a

Sanvitalia, 3

Sassafras, 329, 339-341

—albidum, 215

— officinale, 215

Saussurea, 361, 370, 381, 410
8

Simaniaces 318-31
Scuuttes, Ricuarp Evans. A Synopsis of
the Genus Herrania, 217
Scortea arbor .. . refertis, 37
Scorzonera, 362, 373, 382
Sebifera, 343
Securinega neopeltandra, 208
Senecillis, 3
Senecio, 359, 360, 369, 381, 390, 410
Serratula, sa 371, 381
Sheareria, 3
Siegesbeckia, 358, 380
Silybum, 361, 390
Sinomenium acutum, 403
Solidago, 356, 379, 390

Soliva, 3
Sonchus, 362, 373 382, 391
So-pee-ja-ke, 2

Sorocea, 107, 10 ‘
guilleminiana, 109, 110
Soroseris, 362
South Pacific Species of Section Podo-

1958]

carpus, Subsection B, The. Taxo-
nomic Revision of Bodocarsus XI, 424
Southeastern United States, es Ps
of the Woody Ranales in the,
Southern Yew, 474

Spilanthes, 358, 380
Staff of the incnoide nae 1957-1958,

514
Statistics of Compositae in Relation to
the Flora of China, 347, 379

Stilpnolepis, 359
Stokesia, 355
Strawberry bush, 323
Studies in the Genus Coccoloba, V. The
enus in Haiti and the Dominican Re-
public, 1
Sweet bettie, 323

WM t

ynopsis of the Genus Herrania, A, 217
Synurus, 361

Tach-ko-au, 276
Tagetes, 358, 380
pemieedcr 360
Tamala, 3
ae i
Taraxacum, 362, 373, 382, 391, 396, 410

tion Podocarpus, Subsection B, 424
Taxus macrophylla, 472
— spinulosa, 44
Tetracentron sinense, 398
Theobroma, 22
— albiflorum, 232

— purpureum, 267

INDEX

Theobroma speciosum, 274
Thespis, 357, 380

Tithonia, 358

Tomex

Toot-choo

Trane 362, ae 382, 391

ulip-poplar, 306
Tulip-tree, 06
Turczaninovi

35
Tussilago, 359, "381, 390

Uva caleta, 44
Uva cimarrona, 16
Uva de mar, 4
Uva de ee 44
Uvero,

Uvifera Ae aoa a aes 46
Uvifera nodosa, 32
Uvilla, 16, 29

Vernonia, 355, 356, 364, 379, 390

Vladimiria, 361
Waldheimia, 359

nographic

of the West nae ote of
iets 49, 11

Wedelia, 3

eae portianciana: 106

Werneria, 360

West Indian Species of Phyllanthus, A
Monographic Study of the, 49, 111

Wild Allspice, 342

Wild cacao, 276

9, 71
WILSON, hes A. Ontogeny of the
Sporangia in Xiphopteris serrulata and
yrrosia nuda, 4
Winterana, 321
Woe-vine, 345
Wollastonia, 358
Woop, Carrott E., Jr. The Citation of
Some Seer oe ae Lauraceae, 21
Woop, Carr rR. The Genera of
the ae ee in the Southeastern
United States, 296

524 JOURNAL OF THE ARNOLD ARBORETUM

Xanthium, 358, 380, 390

Xiphopteris serrulata and Pyrrosia nuda,
Ontogeny of the Sporangia in, 478
Xylophylla, 66, 179

—contorta, 193, 194
— epiphyllanthus, 199

Xylophylla falcata, 199
— latifolia, 180, 185

eciosa, 188
Yellow-Poplar, 306
Youngia, 363, 374, 382, 410
Yushunia, 340

Zamon maron, 16
Zinnia, 358

[ VOL. XXXIX