19601 HOWARD, STUDIES IN THE GENUS COCCOLOA. IX
STUDIES IN THE GENUS COCCOLOBA, IX.
A CRITIQUE OF TlHE SOUTHT AMERICAN SPECIES
RICHARD A. HOWARD
THE FIRST SIX PAPERS of this series treated the West Indian species
of the polygonaceous genus (occoloba, with the work based largely upon
field studies in this area. The seventh paper dealt with the species of
Central America and Mexico; the writer has only a slight knowledge of
these species in the field, but fortunately had available an unusually large
number of collections as well as extensive field notes made by recent
collectors. This paper will treat the species of South America. The limita-
tions of this study are greater both from lack of field knowledge and the
fewer specimens available. For these reasons, this treatment is probably
more conservative than future workers will deem necessary. The goals of
this study have been 1) to bring together all records of taxa of Coccoloba
for South America: 2) to correct existing errors of literature citation:
3) to evaluate and compare published species where possible: 4) to locate
the existing type collections; and 5) to select lectotypes of many species.
It is hoped that these notes will encourage collectors and botanists in
the area to gather complete material of the species and to record adequately
the variations so as to allow a better understanding of certain morphologi-
cal problems. Only then can the monographer prepare complete descrip-
tions and keys for the species in South America.
The earliest comprehensive treatment of the species of Coccoloba in
South America was prepared by Mleisner for the Flora Brasilicnsis (5(1):
23-44. 1855). Although some species were illustrated, the descriptions in
general were inadequate and no keys were given. Specimens cited previ-
ously from the Martius herbarium have since been found in Brussels,
Munich, and Leningrad, and a few numbers have been widely scattered.
Representative specimens of several species attributed to Martius could
not be located for this study and in some other species the selection of
a lectotype has required very careful comparisons between specimens
from several herbaria and published descriptions.
In 1856 Meisner provided a synopsis of the genus for Ie('andolle's
Prodromus (14: 150-171). One might expect that many of the speci-
mens which Meisner cited would be in the Prodromus herbarium or in the
Delessert herbarium at Geneva. but, unfortunately, very few are repre-
sented in either collection. Meisner's personal herbarium, which has been
acquired by the New York Botanical (arden, includes specimens of most
species of Coccoloba. A few of these are small fragments, yet they repre-
sent the only existing type material for certain species.
The only complete monograph of Coccoloba was prepared by Lindau


JO) RNAL OF THE ARNOLD ARBORETL' M
and published in 1890 in Engler's Botanischcr Jahrbiichcr (13: 100 229).
Mianv adjustments are necessary therefore to bring the specific names
into conformity with the present rules of nomenclature. Lindau's de-
scriptions are not models of accuracy, and the key published in the mono-
graph contains so many errors of fact that it is almost unusable. Fol-
lowing the description of a plant. Lindau cited the collector's name and
collection number and at the end gave a list of the herbaria in which
he had seen the material. In the selection of a lectotype. therefore, it
has been necessary to associate the collection with a particular herbarium.
The information supplied in this paper should make easier the work of
the next monographer who wishes to obtain material for study.
Lindau frequently cited specimens to be found only in the Berlin
herbarium. A study of these, annotated by Lindau, has shown them to
be fragments of specimens from other herbaria. There is a question
whether the holotype is really the Berlin fragment and only specimen
the author cited, or the more ample specimen in some distant herbarium
which the author may not have seen and did not annotate.
The specimens of Coccoloba in the Vienna herbarium were lost or de-
stroyed during World War II. According to Lindau's citations some of
the Jacquin material was at Vienna and is no longer available. A very
few of the Vienna specimens were photographed before 1940, and prints
of these in the herbarium of the Chicago Natural History Museum are the
basis for discussion of these names here.
Most existing treatments of the genus have assumed the flowers to be
perfect. During field study of the species in the West Indies the author
determined that the flowers are functionally unisexual and that the plants
are dioecious. Recently Buchinger and Sanchez examined the species
found in Argentina and concluded (Bol. Soc. Argent. Bot. 7: 251 255.
1905) that one species of ('occoloba in that country is truly dioecious. but
that the remainder have perfect flowers or are monoecious. Very few col-
lections are cited in their work and none of those was available to me. I
have examined many specimens which I assign to these same species
from Argentina and find no evidence in the herbarium specimens to sup-
port their conclusions. In the specimens which I examined the ilowers
with conspicuous pistils have small or rudimentary stamens which do not
produce pollen. Flowers with large stamens which produce pollen appear
to have undeveloped or rudimentary pistils. There were no specimens
with pollen-producing flowers and fruits on the same branches. I do not
doubt their decision, since it was based on field study of carefully selected
material. but I do poin t out the difficulty of stating definitely the sexual
condition in herbarium specimens and the futility of using in a key such
contrasting characteristics as plants monoecious or plants dioecious.
The species in the \est Indies were studied in the field to determine
leaf variation. I found that vigorous terminal shoots, short branches.
branches of average growth. and adventitious shoots might all have dif-
ferent sizes andt shapes of leaves. This leads one to question Lindau's
V\L. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 215
descriptions and keys, since he seems to have relied much on the size and
shape of leaves and particularly on the shape of the leaf base.
The habit of the plant also was found to be variable in West Indian
species. Some are shrubby or tree-like at the base, but with scrambling
branches having the appearance of lianas. Thus herbarium specimens
consisting of flowering or fruiting branches without adequate habit notes
may not be indicative of the true character of the species.
Relatively few of the South American species considered in the follow-
ing notes are represented by both flowers and fruit. In some cases it
has been possible to associate as one taxon certain flowering and fruiting
specimens originally described as separate species. When peculiarities of
growth can be properly associated with flowering and fruiting specimens
additional species will undoubtedly be combined. Throughout the fol-
lowing text I have been conservative in maintaining as distinct species
all specimens concerning which there may be some question. When fur-
ther evidence becomes available, many of these species may be united.
I have been fortunate to have on loan for study the important historical
collections of Coccoloba from the Botanisches Museum, Berlin. which
formed the basis for Lindau's work; the Jardin Botanique de l'E:tat,
Brussels, which include many Martius collections; the Botanische Staats-
sammlung, Munich, also rich in Martius material; the New York Botanical
Garden which include recent collections, as well as Meisner's personal
herbarium; and the Herbarium of the Department of Systematics and
Plant Geography of the Botanical Institute of the Academy of Sciences
of the U.S.S.R., Leningrad, which contains so many duplicates of classic
collections. I am most grateful to the directors and curators of these
institutions for making their collections available for an unusually long
period of time. Selected collections have been borrowed from other insti-
tutions and I have visited still more herbaria for study. The many cour-
tesies which made this publication possible are much appreciated. The
standard abbreviations of Index Herbariorum have been used in this paper.
Species of Coccoloba have been reported from every country in South
America, with the sole exception of Chile. In the citation of specimens
the countries and their subdivisions, as well as the specific localities of
collections, are listed in alphabetical order.
Coccoloba acrostichoides Chamisso, Linnaea 8: 132-133. 1833: Meis-
ner, Fl. Bras. 5(1): 33. pl. 13, fig. 2. 1855; Lindau, Bot. Jahrb. 13:
138. 1890.
Coccoloba rubiginosa Martius ex Meisner, Fl. Bras. 5(1): 33. 1855, in syn.
Chamisso did not cite a specimen in the original description, and pre-
vious monographers have not indicated a type. There is a specimen in
the Leningrad herbarium collected by Sellow and bearing on the label
the name "Coccoloba acrostichoides N" and the annotation "Hb. Cham."
This specimen is probably authentic. A more ample specimen in the Berlin
herbarium has a tag which bears the number "1393" attached but lacks


JOURNAL OF THE ARNOLD ARBORETUM
the reference to the Chamisso herbarium. Lindau did not see or cite either
of these collections, yet one must be considered the lectotype and I so
designate the specimen at Berlin. All other Sellow collections, including
those cited by Lindau, are annotated "Corcoloba acrostichoidcs (ham. et
Schl."
The illustration in Flora Brasiliensis well represents this species. A
printer's proof of this drawing is in the Brussels herbarium and the speci-
men Claussen 330 is mounted on the same sheet. There is little similarity
between the two, and the illustration appears to be a compilation.
Coccoloba acrostichoides is one of the most distinctive species of the
genus, with copious dark-brown pubescence on the lower leaf surface.
In leaf shape and in the nature of the inflorescence it is similar to C.
brasili/nsis. The species is not known in fruit and I have not seen any
collections more recent than those of Glaziou.
'occoloba rubiginosa Martius ex Meisner was published in synonymy
and has no validity. A specimen at Munich from the Martius herbarium
bears this name. but no location or collector is indicated.
Brazil. MINAS GERAIS: Ouro Preto. (Glaiou 15356 (BlR. LE : without specific
location. Clausscn 330 (BR). RIo DE JANEIRO: San Antonio, Sellow "B 1393.
c 429" (ln. LOCATION NOT KNOWN: Herb. 3lartins s.n. (M-type of C. rubi-
ginosa). Sellouw 1251 (13), 1393 (n-lectotype: LE').
Coccoloba acuminata HBK. Nov. Gen. 2: 176. 1817.
Coccoloba acumninta var. pubescens Lindau, Bot. Jahrb. 13: 193. 1890.
Coccoloba acuminata var. glabra Lindau, ibid. 194.
Coccoloba strobilulifera Meisner, Fl. Bras. 5(1): 25. 1855.
I discussed this species in an earlier paper (Jour. Arnold Arb. 40: 185.
1959) when I concluded that the amount of pubescence on the plant
varied with the age and the vigor of the specimen. I cannot find any
value or significance in the two varieties Lindau published.
A specimen from the Brazilian state of Pari, Huber 4393, bears an
unpublished herbarium name attributed to Huber. The specimen is
clearly referable to synonymy here.
A collection in the herbarium of the University of Wisconsin numbered
"45" but without collector was supposedly made in La Lima. Chile. The
authenticity is to be questioned as the specimen compares favorably with
material from Colombia and Peru.
In addition to the specimens cited below. I have seen material from
Guatemala. Honduras, Nicaragua. Costa Rica and Panama (Howard,
loc. cit.).
Brazil. ACRE: Rio Macauhan. Krukoff 5666 (A, LE, M, NY). AMAZONAS:
Rio Embira, Krukoff 4715 (A, LE, M , NY); Yururi Miry. .'l 5723 (B). PARA:
Para, Alto Purus, Ponto Alegre. Huber 4393 (F, u). British Guiana. Courantyne.
Schomburgk 1600 (B). Colombia. ANTIOQ-IA: Rio Magdalena, Brazuela de
Ierales, Pennell 3699 (GH). BOLIVAR: San Martin de Loba, Curran 40 (cH),
100 (v), 211 (v). MAGDALENA: Rio Magdalena near Mompos, Humboldt 1479
VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 217
(Herb. Willd.-type of C. acuminata); Rio Sevilla, Record 9 (A, GH, Y). SAN-
TANDER: Rio Magdalena, Kalbreyer 1280 (B). LOCALITY NOT SPECIFIED: Goudot
3 (B), Moritz s.n. (BR, LE, P-type of C. strobilulifera and C. acuminata var.
glabra), Triana s.n. (B). Ecuador. Balao, Eggers 14239 (A, BR, LE, M). Peru.
LORETO: lower Rio Huallaga, Llewelyn Williams 4804 (A). Venezuela. YARACUY:
Los Cafiizos, plains of Yaracuy River, Pittier 8753 (GH). ZULIA: Perija,
Tejera 253 (GH), Mocquerys 845 (P).
Coccoloba alagoensis Weddell, Ann. Sci. Nat. III. 13: 260. 1850.
Weddell based this species on Gardner 1389 from the state of Alagoas,
Brazil. At present I am unable to determine the relationship of this species
to Coccoloba obtusifolia Jacq. and C. peruviana Lindau. A full discussion
will be found under C. obtusifolia.
Coccoloba alnifolia Casaretto, Nov. Stirp. Bras. 71. 1844.
Coccoloba populifolia Weddell, Ann. Sci. Nat. III. 13: 257. 1850; Meisner, Fl.
Bras. 5(1): 40, pl. 18. 1855; Lindau, Bot. Jahrb. 13: 198. 1899.
Lindau accepted the Weddell name for this species, although he cited
in synonymy the epithet Casaretto published several years earlier. Casa-
retto did not cite a specimen in the original publication, but it seems
clear to me that the description and data given are based on his collection
numbered 1194. This collection consists of two sheets in the Turino her-
barium and I have designated one of them as the lectotype.
There is a possibility that the older epithet Coccoloba firma Martius ex
Colla (1836) is the correct one for this species. The current difficulty in
determining the type of C. firma will be discussed under that name.
Brazil. BAHIA: Blanchet 1486 (c, G-type collection of C. populifolia); Sel-
low 1137 (B), 793 (B). PERNAMBUCO: Tapera, Pickel 3681 (GH, NY). RIO DE
JANEIRO: Tijuca, Liitzelburg 343 (M); without specific locality, Casaretto 1270
(TO), Gaudichaud 420 (B, G, p), Luschnath 835 (LE), s.n. (M), Martius s.n.
(M), Riedel 7 (BR, LE), 673 (LE, M), Widgren 719 (BR). STATE NOT SPECIFIED:
Copacabana, Casaretto 1194 (To-lectotype of C. alnifolia), Nadeaud s.n. (p).
CULTIVATED: Sao Paulo Botanical Garden, Hoehne 28527 (A).
Coccoloba arborescens (Vellozo) Howard, Jour. Arnold Arb. 41: 44.
1960.
Polygonum arborescens Vellozo, Flor. Flum. 162. 1825, Icones 4: t. 43. 1827.
Coccoloba crescentiifolia Chamisso, Linnaea 8: 134. 1833, "crescentiaefolia."
Coccoloba vellosiana Casaretto, Nov. Stirp. Bras. 70. 1844.
Coccoloba fasciculata Weddell, Ann. Sci. Nat. III. 13: 258. 1849.
Coccoloba crescentiifolia var. obtusata Meisner, Fl. Bras. 5(1): 26. 1855.
The previous paper in this series (Howard, loc. cit.) contains a dis-
cussion of the nomenclature of this species. For the present, the type is
considered to be the Vellozo illustration. After a careful study of this
species in the field it may be desirable for some future monographer to
select a modern specimen as neotype.


JOURNAL OF THE ARNOLD) ARBORETUM
Brazil. IHAIII: Blanciet 796 (V. NY. p-type of C. fasciculata). Rio DE
JANEIRo: Prayva Grande. Sellow s.n. (It, LE. M-type collection of C. crcscentii-
folia); Glazioi~ 143 (1 R); Riedel 674 (A, BR, p). STA. CATARINA: Hooker Herb.
without collector or number (K). STATE NOT SPECIFIED: Copacabana, Luscknath
s.n. ( R); betw een San Juan and Rio dos Oistres. Prince MaximN. Neuwied. s.n.
BR, NY). LOCALITY NOT SPECIFIED: (Clauisst 2013 (0. N x-type collection of
C. cr ce' tiifoli vat r. obtusata ).
Coccoloba argentinensis Spegazzini. Physis 3: 176. 1917: Buchinger
and Sanchez, Hol. Soc. Arnent. Hot. 7: 251. 1959.
Coccoloba pr ecox Herter. Rev. Sudam. Hot. Montevideo 10: 3S. 1952.
Splegazzini failed to select a type in the original descriltion. He re-
ferred to material collected by MI. Rigou in 1902 from the station "Mar-
gherita." Prov. Santa Fe. and to material collected the following year in
Reconquista. (orrespondence with several Argentine botanists indicates
that the Spegazzini herbarium has been neglected in the past and that
some material has been lost. I)r. Maria Buchinger reported that she
aw the -type" 'several years ago but currently there are no specimens of
the species in the herbarium of Spegazzini at La Plata and no other her-
barium contains such material. As long as there is a possibility that the
original material may be located, a neotype should not be selected. In their
recent paper Huchinger and Sanchez cite only one collection. Schulz 240.
which I have not seen. They further consider the species to be endemic
to the (Chaco territory and to be found only in the provinces of Salta and
Chaco. thereby eliminating the locality of the type collection.
The material cited below appears to me to correspond with the original
description. (Coccoloba argcntinensis is easily recognized by the small
obovate-elliptic leaves, the long fruiting pedicels, and the flowering of the
plant before the leaves are fully developed.
(occioloba pracco. Herter was distinguished on these same characteris-
tics but without comparison with the present species. I have examined
an isotype and conclude that the reduction of C. praccox to the synonymy
of C. argentinensis is necessary. (occoloba praccox Herter is also a later
homonym of (. praricocx Wright ex Lindau (Bot. Jahrb. 13: 142. 1890).
Argentina. C('HACr : Fiontana. M1eyer 2517 (F); Las Paulas, JorgCensen 2103
('(t; , is i 21O l i I) ,): Resistencia. Colonia Benitez, P tersen .31So (A). J jy'y:
Este Hacienda, . s ot Jujuy. Eycrdarnm 'r lBectle 22432 (i i). SALTA: Oran.
Manuelta Pedraza, Evrdarn & BRcetle 22047. 22790 (c, on i; Pichanal, Rodri-
i/u': 1121 (i;H. Ni: Rosario de la Inrontera. Los Bafnos. entulri 9402 (ca).
'I C tIM Capiita:l. Harranca Colorada. ''nturi 955 ((;in). 5Schrciter 1516 (cni:
lrancas. Tapia a Rica, Schrciter (Heirib. Lillo 84971 (Is) . ipos. Venturi 9792
(A., Il. LE):. liruguay. Salto, Arapey, llcrter (Herb. Herter 50852 -type col-
lection of C. pr aer x; . p); Isla G(aspar. Berro 3334 ((;).
Coccoloba ascendens Duss ex Lindau, Bot. Jahrb. 13: 156. 1890.
In a previous discussion of this species (Jour. Arnold Arb. 40: 72, 73.
1959), I selected a specimen from Martinique. Hahn 1005, in the Berlin
[VOL. XLI


1960] HOWARD. STUDIES IN THE GENUS COCCOLOBA, IX 219
herbarium as the lectotype. The species is well represented in the Lesser
Antilles from Guadeloupe south to Trinidad. Lindau (Bot. Jahrb. 13: 156.
1890) cited specimens only from Guadeloupe, Martinique and Brazil.
Eyma (Meded. Bot. Mus. Utrecht 4: 110. 1932) mentioned the dis-
tribution of the species as "Guyana. Brazil, Guadeloupe, Martinique." I
have seen neither the specimens cited from the Guianas (B.W. 6560.
6600, and possibly B.W. 6490 and 6550b) nor those Lindau cited from
the states of Para and Govaz in Brazil (Burchell 8034, 9345 and Gardner
3966). On the basis of an unnatural range I question particularly the
Brazilian determinations given by Lindau. Future workers with addi-
tional material at their disposal doubtless will check the relationship of
Coccoloba ascendcns and C. sparsifolia.
Coccoloba barbeyana Lindau, Bot. Jahrb. 13: 185. 1890.
I have examined the single collection cited by Lindau (Ruiz & Pavon
s.n., Peru, without definite locality, in the Barbey-Boissier herbarium)
and failed to find the holotype or the species distinct in any way from
Coccoloba densifrons Martius ex Meisner. Although Lindau accepted
C. densifrons (based on a Martius collection from Ega in the Brazilian
Amazon) he did not list the epithet in the key to the species.
Coccoloba billbergii Lindau, Bot. Jahrb. 13: 219. 1890.
Lindau cited "Coccoloba obtusifolia Meissn. (non Jacq.)" in synonymy
when he described C. billbergii based on Billberg 204 and 204a from
Carthagena, implying that Meisner was in error. However, Meisner noted
both the similarity of these Billberg collections to C. obtusifolia, as de-
scribed and illustrated by Jacquin, and the discrepancy in the specimens
so labeled in the Willdenow herbarium and the reference in the Willdenow
catalogue. Lindau chose to interpret C. obtusifolia on the basis of the
specimen in the Willdenow herbarium and incorrectly to consider C. ob-
tusifolia as a species from the West Indies. I have examined the Billberg
collections cited by Lindau from the Berlin and Stockholm herbaria.
These are correctly associated with C. obtusifolia Jacq. as a South Ameri-
can species. Coccoloba billbergii Lindau must be considered a synonym
of C. obtusifolia Jacq.
"Coccoloba obtusifolia" as used by Lindau (loc. cit. 146, 147) is cor-
rectly assigned to synonymy under C. microstachya Willd. (Howard,
Jour. Arnold Arb. 38: 217-219. 1957).
Coccoloba blanchetiana Weddell, Ann. Sci. Nat. III. 13: 257. 1850.
Weddell cited Blanchet 3561, from near Jacobina, Bahia, Brazil (G, P).
This species is to be referred to the synonymy of C. ochreolata Weddell
and will be discussed there.


JOURNAL OF THE ARNOLD ARBORETUM
Coccoloba bolivarana Llewelyn Williams, Trop. Woods 68: 39. 1941.
In a list of the forest trees of Venezuela, Williams refers to "Brusquillo
Blanco, Coccoloba bolivarana (a new species), a tree about 8 m. high,
growing in rocky areas, with a trunk ramified from the base, and bearing
a juicy, lustrous black fruit." The species has not been published validly
with a Latin description, to my knowledge, and the brief description
given above is scarcely to be considered diagnostic in this difficult genus.
Specimens in the herbaria of the Chicago Natural History Museum and
the Royal Botanic Gardens, Kew, bear this name but attribute it to Stand-
ley. The specimens (Williams 13374) were collected at El Tigre, Los
Garzones, on Rio Cuchivero, Bolivar, Venezuela. They cannot be assigned
to any recognized species and will be described in a later paper.
Coccoloba bracteolosa Meisner, Fl. Bras. 5(1): 30. 1855.
No type was selected in the original description, but a Martius collec-
tion without number from Joazeiro. on the Rio San Francisco, in Bahia.
Brazil, has been photographed by other workers in the Munich herbarium
and distributed as the type. I am referring this species to the synonymy
of Coccoloba ochreolata Weddell and will discuss the problem under that
name.
Coccoloba brasiliensis Nees & Martius, Nov. Act. Acad. Nat. Cur. 11:
30. 1823; Meisner, Fl. Bras. 5(1): 32. t. 13, f. 1. 1855.
Coccoloba senaei Lindau ex Glaziou, Bull. Soc. Bot. France 58: 571. 1911.
The original description cites first the locality of "X Valos" and then
a general littoral distribution in Bahia, "Maragnanum" and Para, the
latter attributed to Martius. I have seen no Martius specimens which
fit the original description among the material from the Martius herbarium
now at Brussels, Leningrad, or Munich. In his treatment of the genus for
Flora Brasiliensis, Meisner redescribed the species and cited collections
by Prince Maximilian and Martius. The illustration in Flora Brasiliensis
is based on the specimen collected by Prince Maximilian and I select this
as the lectotype.
A previous worker, possibly Meisner, has annotated some sheets to in-
dicate a similarity to or a relationship with Coccoloba rigida Meisner.
(Lindau, not following the homonym rule, accepted C. parvifolia Schott
1827, not Poiret 1804. and included in synonymy C. rigida Meisner.)
Lindau distinguished between C. brasiliensis and "C. parvifolia" on the
puberulent inflorescence axis in the latter species and the glabrous one
in C. brasiliensis. An examination of the specimens cited below will show
that the stem, ocreae, petioles, and frequently the base of the leaf blade
are puberulent to short pilose in C. brasiliensis. The bracts and ocreolae
of the inflorescence are likewise slightly puberulent to glabrous. On the
characteristic of pubescence there is scant reason for separating C. brasi-
[VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 221
liensis and "C. parvifolia." However, the leaves of C. brasiliensis are vari-
able in size and shape. In the type, and in the most recent collections,
the leaves are thicker in texture and the bases are rounded-cordate, in
contrast to those of "C. parvifolia." The inflorescence of the type is as
compact as that of "C. parvifolia" but recent collections show longer, more
lax inflorescences in which the ocreolae do not expand with the developing
flower buds, and are, in fact, shorter than the bracts in all stages of de-
velopment. Until these characteristics can be evaluated by field studies
it seems desirable to retain C. brasiliensis as a species distinct from C.
rigida.
The several sheets of the collections by Glaziou cited below suggest
that the existing descriptions of Coccoloba brasiliensis are unsatisfactory
and must be changed to accommodate the narrow, lanceolate-oblong
leaves of the vigorous shoots and the large, oblong, but sessile, leaves of
adventitious shoots. The Glaziou collection without number in the Paris
herbarium has some leaves with blades 18 cm. long and 9 cm. wide. The
largest leaf on the lectotype is 4 cm. long and 2.3 cm. wide, yet the col-
lections cited below show all intermediates between these extremes.
The phrase "Coccoloba senaei Lindau n. sp." is used by Glaziou in a
list of determinations of his collections. The brief description "Arbuste
sarmenteux, fl. blanchatres" can scarcely be considered diagnostic in this
genus and I consider C. senaei to be a nomen nudum.
Brazil. MATO GROSSO: Barao de Capanema, Rio Cravary, Baldwin 3129 (us).
MINAS GERAIS: Conselheiro Mata, Brade 13815 (B), 13816 (B); "in deserto
Minarum," Martius s.n. (M); Rio dos Pedras, Valu, Glaziou 19762 (B, LE),
19763 (B, K, LE); Serra do Cip6, Schwacke 8005 (s), Brade 14841 (B). WITH-
OUT SPECIFIC LOCALITY: Valos, Prince Maximilian 88 (B, BR-lectotype; LE);
Glaziou s.n. (p).
Coccoloba brasiliensis Sprengel, Syst. Veg. 2: 252. 1825.
This epithet is a later homonym of Coccoloba brasiliensis Nees &
Martius. The species was based on a Sellow collection from Brazil which
I have not seen. The description is not that of a species of Coccoloba.
It was referred to the genus Hedyosmum of the Chloranthaceae by Cha-
misso (Linnaea 4: 36. 1833). I have not determined whether this is the
same as Hedyosmum brasiliense Mart., recognized by Occhioni in his
recent monograph of the genus (Contribuccao ao estudo de familia
Chloranthaceae com especial referencia ao genero Hedyosmum Sw. Rio
de Janeiro, 1954).
Coccoloba candolleana Meisner, Fl. Bras. 5(1) : 41. 1855.
Meisner based this species on the collection Blanchet 1818 from Bahia,
Brazil. He noted its similarity to Coccoloba cordata Chamisso. I am
unable to distinguish the two and so refer C. candolleana to the synonymy
of the older name, C. cordata.
Lindau (Bot. Jahrb. 13: 201. 1890) cited Goudot 4, from Colombia, in


JOURNAL OF THE ARNOLD ARBORETUM
his treatment of Coccoloba candollcana. Several recent collections from
Colombia have also been identified as this species. The Goudot specimen
in the Berlin herbarium cited by Lindau is a fragment of a larger speci-
men in the Paris herbarium. A study of the latter sheet clearly indicates
that this collection is to be referred to C. padiformis Meisner, the type of
which is from Caracas, Venezuela.
Coccoloba caracasana Meisner, )('. Prodr. 14: 157. 1856; Howard.
Jour. Arnold Arb. 40: 193. 1959.
Coccoloba caracasana f. glabra Lindau. Bot. Jahrb. 13: 211. 1890.
Coccoloba cyclophylla Blake. Contr. I.S. Natl. Herb. 20: 238. 1919.
In an earlier paper I discussed the significance of pubescence in this
species and cited the collections seen from Mexico and Central America.
Coccoloba cyclophylla Blake is now added to the synonymy of this species.
Blake compared his species, based on Curran 47, from Colombia. with one
from Hispaniola and concluded that they were distinct. A proper com-
parison would have been with (. caracasana from Venezuela, which is
clearly the same.
Colombia. ATLANTICO: La Plaza. Juan Mina. Dugand 643 (v) Molinero,
Dugand 568 (v). BOLIVAR: Castillo Maldonado. Sessc & c ocino 5430 (F);
San Martin de Loba. Curran 47 (rs-type of C. cyclophylla: oH, '). Magdalena,
Santa Marta, Smith 1702 (A. cai); Tucurinca, Romero 1388 (us). LOCALITY
NOT KNOWN: Moritz 1109 (BR, LE). Venezuela. API'RE: San Fernando de Apure,
Grisol s.n. (A, P). ARAGL'A: Maracay, l'ogl 100 (M '; San Juan de los Morros,
Alston 6023 (BM). CARABOBO: Between San Joaquin and Mariara, Pitticr
12111. 12112 (c. LE. NM); Valencia, Pitticr 8705 (ci' . D)ISTRITO FEDERAL:
Caracas, I'arras 30 (c); Humboldt 732 (Herb. Willd.); Bonpland 732 (n holo-
type). GI'ANRICO: El Sombrero. Pitticr 11460 (A. c. GH. l), 12367 (Mt, Nv):
La Rubiera. Pittier 12328 (G. M)i. LARA: Between Varitagua and Duaca. Pitti'r
343 (A. uM). MERIDA: Tovar, Fendler 2(53 (c;i'. VARACI"V: Bruzual. Curran
642 (N).
Coccoloba carinata Ruiz ex Meisner, l)('. Prodr. 14: 150. 1856.
Meisner published this epithet in synonymy under :MuehlElnbeckia
leptobotrys Meisner and stated that it was a manuscript name found in
the herbarium at Berlin. I have not seen authentic material.
Coccoloba caurana Standley. Field Mus. Publ. Bot. 22: 73. 1940; L.
Williams. Explor. Bot. Venez. 189. 1942.
Although the description of this species was published twice as new.
the taxon belongs in the synonymy of Coccoloba fallax Lindau and is dis-
cussed in more detail under that name. The type, in the herbarium of
the Chicago Natural History Museum, is Wlilliams 11366, from Venezuela.
VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 223
Coccoloba cerifera Schwacke, P1. Nov. Mineir. 1: 7. 1898.
I have not seen the original publication to check the validity of this
name and so, for the present, must interpret the species from the speci-
men indicated as the type in the Berlin herbarium. The plant is unlike
any other species found in southern Brazil. It resembles Coccoloba schom-
burgkii in the shape and thickness of the leaves, but differs from it in
having long pedicels from which the apparently staminate flowers have
fallen. The Pires and Black collection cited below compares well with
the type specimen.
Brazil. MINAS GERAIS: Serra do Cip6, Schwacke 11780 (B-holotype), Pires
& Black 2816 (NY).
Coccoloba charitostachya Standley in A. C. Smith, Lloydia 2: 176.
1939.
The type of this species is A. C. Smith 2356, collected near the mouth
of the Charwair Creek, in the basin of the Rupununi River, British Guiana.
The holotype is in the herbarium of the Chicago Natural History Museum.
This species is not well defined in the original description and has not
been recollected. Additional material is badly needed for an understanding
of the species. At present the species may be characterized by the slightly
rugose leaf blades which are elliptic-ovate in shape, rounded at the apex
and lighter in color on the lower surface. The fruits are all insect-infested
and abnormal and of no diagnostic value. Flowers are not known.
Coccoloba chacoensis Standley, Field Mus. Publ. Bot. 17: 239. 1937.
This species is to be referred to the synonymy of Coccoloba spinescens.
Coccoloba chacoensis was based on Cardenas 2529, from Bolivia; the
holotype is in the Chicago Natural History Museum. Buchinger and
Sanchez (Bol. Soc. Argent. Bot. 7: 253. 1959) recognized this species and
maintained it as distinct from C. spinescens Morong because of the ab-
sence of lateral branches terminating in spines and the presence of "glands"
on the lower leaf surface. Although the Argentine botanists undoubtedly
know these species in the field, the characteristics they have chosen to
separate the species are not reliable in herbarium specimens. The holotype
compares favorably with such collections as Morong 882 (the type of C.
spinescens) and Hassler 11476 and 2486.
Coccoloba confusa Howard, nom. nov.
Coccoloba declinata Martius, Beibl. Flora 20: 90. 1837; Meisner, Fl. Bras.
5(1): 29. 1855, as to plants, not as to name.
Coccoloba declinata var. major Meisner. Fl. Bras. 5(1): 30. 1855.
Coccoloba declinata var. minor Meisner, ibid.
Polygonum declinatum Vellozo (Flor. Flum. 162. 1825, Icon. 4: t. 41.
1827) was described briefly and in general terms. The illustration given


224 JOURNAL OF THE ARNOLD ARBORETUM [IvL. XLI
in the Icones is a poor one showing scarcely any characteristic of value
in defining the taxa in Coccoloba. It was suggested by Dr. I. MI. John-
ston in his library notes that Pharmacopolis, the locality given, might be
near the mouth of the Rio Taquari in the western part of the state of Rio
near Paraty.
The transfer of the Vellozo name to Coccoloba made by Martins in an
obscure paper is seldom correctly cited. Martius refers to two specimens,
one from "Montem Talegraphi," in "Prov. Sebastionopolitana," and the
other near Pendamonhangaba and Taubate, in Sao Paulo. Martius' brief
description of Coccoloba declinata is obviously based on the specimens
cited and could scarcely be drawn from either the description or the plate
given by Vellozo.
In 1855 when Meisner prepared the treatment of the genus for the
Flora Brasiliensis, he described the species and established three new
varieties: a minor, /3 Vclloziana and y major. For Coccoloba declinata
var. minor, Meisner cited in reference "Coccoloba declinata Mart. Herb.
propr." and a specimen from the Rio Doce collected April 1816 by Prince
Maximilian. I have not seen this specimen in material from Brussels,
Leningrad, Munich, or New York, the usual places for Martins and Meis-
ner specimens. Meisner's variety, Velloziana, is based on the epithet and
reference "Polygonum declinatum Vellozo Flor. Flum. IV. t. 41" and in
place of a specimen Meisner cites "in prov. Rio de Janeiro."
Neither Meisner nor more recent authors cite the specimen from "Mon-
tem Talegraphi" which Martius mentioned in the original transfer of the
Vellozo name. The specimen from Iendamonhangaba is referred by Meis-
ner to his var. major.
Meisner prepared the treatment for Coccoloba in DeCandolle's I'rodro-
mus, using the same broad concept of varieties comprising the species
C. declinata. In this treatment, however, the assignment of var. major
to the species is questioned.
In his monograph of the genus Lindau reduced to synonymy Meisner's
varieties minor and Velloziana. Lindau recognized the var. major, but
commented on the poor condition of the type specimen and suggested
that it might be a new species.
There appear to be no Vellozo collections available: thus the inter-
pretation of the species must be made from the inadequate original de-
scription and the rather poor drawing. In comparing the drawing with the
specimens cited by Martius, Meisner and Lindau, I am convinced that
these authors have misinterpreted Vellozo's species. \ellozo's drawing
shows that a scrambling plant is intended, but the material cited by
previous authors for Coccoloba declinata is shrubby, with geniculate
branches. Again, the drawing shows a short, regular inflorescence, while
the voung inflorescences of the specimens cited by earlier authors are
geniculate, with the older inflorescences elongate. The most obvious match
for the Vellozo species is the Martius specimen cited as the type of C.
racemulosa. This is a liana with short lateral branches and comparable


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 225
inflorescences. The venation of the type of C. racemulosa compares favor-
ably with that illustrated by Vellozo, while the venation of specimens
cited by Martius, Meisner and Lindau does not. It appears to me that
the drawing of Polygonum declinatum Vellozo most nearly compares to
Coccoloba racemulosa Meisner.
The material which Martius, Meisner and Lindau cited and described
as Coccoloba declinata therefore requires a new name. I propose the name
C. confusa, to call attention to this problem. Lindau's description (Bot.
Jahrb. 13: 166. 1890) applies to the specimens cited, including those
originally mentioned by Martius in transferring Vellozo's specific epithet.
The inclusion of the Krukoff and the Archer collections from Amazonas
and Para creates a sizeable gap in the known range of this species. On
the basis of available material, these specimens must be assigned here.
Brazil. AMAZONAS: Humayta near Livramento, Krukoff 6669 (NY).
PARA: Belem, Archer 7830 (F, K). Rio DE JANEIRO: Caju, Riedel s.n. (LE);
Copacabana, Luschnath s.n. (LE); Rio de Janeiro, Glaziou 3089 (BR), Riedel
675 (A, BR, LE, M), 676 (A, BR, LE, M); Telegraphenberg, Luschnath, Oct. 1833
(BR). SAO PAULO: Pedamhongaba, Taubate, Martius "66" (BR, LE, M-type of
C. declinata var. major).
Coccoloba conduplicata Maguire, Bull. Torrey Bot. Club 75: 304.
1948.
A clear-cut species, but known only from the material cited in the origi-
nal description. The folded leaf blade noted in the specific name appears
to be of less significance than the author indicates, since the majority of
the leaves are not conduplicate.
Surinam. Tafelberg, Maquire 24437 (A, BR; NY-holotype), 27205 (A, NY).
Coccoloba cordata Chamisso, Linnaea 8: 133. 1833.
Coccoloba candolleana Meisner, Fl. Bras. 5(1 : 41. 1855.
Coccoloba cordata var. praecox Hassler, Repert. Nov. Sp. 14: 162. 1915.
Both Meisner and Lindau have indicated the similarity of Coccoloba
cordata Chamisso and C. candolleana Meisner. Lindau distinguished be-
tween these in his key by placing C. candolleana in a group of species
having leaves glabrous on both surfaces and C. cordata in one having
leaves pubescent along the midrib and nerves on the lower surface. Lin-
dau has placed the two species adjacent to one another in the text. Among
the major characters of these species, Lindau described C. candolleana as
having tomentose ocreae while those of C. cordata are glabrous.
The type collection of Coccoloba cordata is Scllow s.n. from the Rio
Pardo area of Brazil. I have seen three sheets of this collection. One
in the herbarium at Berlin bears the full data given by Chamisso and is
selected as the lectotype. The other two were annotated only "C. cor-
data N."
The descriptions given by Chamisso and Lindau are not entirely accu-


JOURNAL OF THE ARNOLD ARBORETLI 'M
rate. Various branches of the specimens comprising the type collection
have puberulent or tomentose ocreae, although other portions are glabrous.
Most of the leaves are coarsely pubescent along the midrib at the base of
the leaf, but some are completely glabrous.
Coccoloba candolleana was based on Blan cht IS IS. A single specimen
is cited as being at Geneva while a smaller specimen from the MIeisner
herbarium, now at the New York Botanical Garden, bears several of
IMeisner's annotations: (1) "(. cordata Chain. ?" which he declared to be
different on the basis of the shorter petiole and more coriaceous bladtes;
(2) a variety of ('. cordata named for Blanchet, a name which was not
published, and finally (3) "Coccoloba candolleana n." The Blanchet col-
lection has broader leaves and shorter petioles, but can be compared favor-
ably with portions of the Sellow collection which is the type of (. cordata.
In his monograph Lindau included a collection from (olombia (Goudot
4) in the description of Coccoloba candolleana. This collection is better
referred to ('. padijormis, a species known from Venezuela and Central
America (Howard. Jour. Arnold Arb. 40: 210. 1959).
Coccoloba cordata var. praccox Hassler was based on a specimen,
Iassler 7279, from Paraguay which is staminate, in contrast to the pis-
tillate type of C. cordata. The deciduous characteristic of the plant is not
significant, and the variety is not worthy of recognition.
Argentina. CHACtc: harranrqueras. Isl de Temores. Curran 402 (I s. FOtR-
M(IsA : Coman. Jorginson 2004 ((;Hi, s): irane. .orel 432 (HR). Jr J Y:
San Pedro. Di Lclla and Ga(rcia 220) (A). 1MIsi :s: Ap6stoles on Rio (himi-
ray, Ibarrola 1150 (v). SALTA: Embaracacion. Everdam t& Bcetle 22916. 22028
(cii): )ran. 'Tahacal. S5chreiter 5472. 8455 (c(i) ; Oran. Vesipucio a las Iablillas,
.Schreiter 11402 iF : Rio Blanco,. I 'eturi 55.5 ( F, (n, Li:, M : Rio Piedras,
Rodrim'vez 7,3 O(i. T cu'MNix: C'apital. Barranca Colorada, Venturi V55 IF);
Raco. Sclirciter 1516. 8443 ((;c i. Brazil. BAIIIA: lanchet 1818 ((;. xvN type
collection of C. candolle'ana)l. Rio GRANDE DO S L.: Rio Pardo. S'ellowl s.in.
(n-lectotype; iM. ST PAUL(): (Gaudichaud 140 (P). Paraguay. CHACO': Villa
Rica, Balan.si 3277 (c). Concepcion. IIasslcr 7279 (i(;-type of C. cordata var.
praccaox ).
Coccoloba cordifolia Meisner, Fl. Bras. 5(1) : 37. 1855.
Meisner cited several specimens but did not indicate a type. A sheet in
the I)elessert herbarium has been cited as the type by previous workers,
although it bears two branches and two labels (Blanchet 100 and Blan chct
3528). Meisner (DC. Prodr. 15: 155. 1856) suggested that Coccoloba
lacvis might be identical with C. cordifolia. Lindau (Bot. Jahrb. 13: 186.
1890) combined the two. accepting the older C. la vis (asaretto, a step
which I believe to be correct.
Coccoloba coronata Jacquin, Enum. Pl. Carib. 19. 1760. Select. Stirp.
Am. Hist. 114. t. 77. 1763; 1)ugand, Caldasia 4: 427. 1947: Howard,
Jour. Arnold Arb. 41: 40. 1960, not Lindau, Symb. Ant. 1: 228.
1899.
VIl. XL1


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 227
Coccoloba virens Lindley, Bot. Reg. 21: t. 1816. 1835.
Coccoloba novogranatensis Lindau, Bot. Jahrb. 13: 192. 1890; Howard, Jour.
Arnold Arb. 40: 85-87, 208, 209. 1959.
Coccoloba dioica Karsten ex Lindau. Bot. Jahrb. 13: 170. 1890.
Coccoloba caribaea Urban. Svmb. Ant. 5: 337. 1907.
Coccoloba waittii Johnston, Sargentia 8: 122. 1949.
The correct name of this species has evolved through the series of my
papers cited above with the final correction suggested by one of Dugand's
papers which I had overlooked. It is possible that Coccoloba jagifolia
Jacq. should be assigned to the synonymy of this species as well. The
identity of that plant will be discussed under that epithet.
I have not seen any material collected by Jacquin, and the species is
regarded as typified by an illustration. Specimens from St. Vincent, south
to Trinidad and Tobago, and from Guatemala and Panama are cited in
earlier papers (Jour. Arnold Arb. 40: 85, 209. 1959). I have also ex-
amined the following specimens from South America.
Colombia. BoAc'A: Between Anapoima & Apulo, Triana 978 (B, p-type of
C. novogranatensis) ; El Humbo, Lawrance 760 (A) ; Santa Marta, H. H. Smith
2421 (A, BR). BOLIVAR: Sabanilla, Karsten s.n. (1, LE). CAQUETA: Hetucha on
Rio Orteguaza. W'oronow & Juzepczuk 6100 (F). MAGDALENA: Molino, Haught
4148 (F, NY). DEPT. NOT KNOWN: Pozo del Higuer6n. Usiacuri, Dugand 788
(v). Ecuador. El Oro. Arenillas. Little 6721 (V). Peru. LORETO: Rio Putumayo.
Klug 2240 (A, B, GH). Venezuela. ANZOATEctIl: Soledad. Gines 3913 (us).
BOLIVAR: Between Upata & Altagracia, Steyermark 57681 (F). DISTRITO FED-
ERAL: Barrancas, Tamayo 1315 (us); Caracas, Karsten s.n. (LE-holotype of
C. dioica). SUCRE: Crist6bal Colon, Broadway 143, 144, 491 (oH, NY, us).
ZULIA: Perija, Tejera 91 (us). DEPT. NOT KNOWN: Sta. Ana, Paraguana,
Tamayo 854 (us).
Coccoloba corrientina Rojas, Bull. Geogr. Bot. 28: 162. 1918.
The original description of this species is brief and generalized. No
specimens are cited and, in fact, a broad general distribution is attributed
to the species. Through personal correspondence I have learned from Dr.
Maria Buchinger and other Argentine botanists that no Rojas collections
can be attributed to this species. The species appears to be impossible to
typify.
Buchinger and Sanchez (Bol. Soc. Argent. Bot. 7: 255. 1959) list Cocco-
loba corrientina as a questionable species, but suggest its similarity to
C. morongii, the correct name of which is C. paraguariensis. I concur with
this suggestion.
Coccoloba crescentiifolia Chamisso, Linnaea 8: 134-136. 1833, "cres-
centiacfolia."
The correct name of this species is Coccoloba arborescens (Vellozo)
Howard based on Polygonum arborescens Vellozo. For a discussion of the
nomenclature and additional synonymy see the previous paper in this


JOURNAL OF THE ARNOLD ARBORETUM
series (Jour. Arnold Arb. 41: 43-45. 1960). Specimens previously re-
ferred to this species are listed under C. arborescens above.
Coccoloba cruegeri Lindau, Bot. Jahrb. 13: 209. 1890.
Coccoloba ernstii Johnston. Proc. Am. Acad. 40: 6S5. 1905.
Coccoloba oblonga Lindau, Bot. Jahrb. 13: 136. 1S90.
This species was based on a Crueger collection from Trinidad. Lindau
described the fruits and assigned the species to the section Campdcria.
Currently available collections from Berlin and (;Gttingen are fragmentary
and sterile. The specimen from the Trinidad herbarium appears to have
had flowering parts at one time. There is some doubt as to what Lindau
actually saw and described, for there is no evidence in the vegetative
parts that the species is related to the others which he assigned to section
Campderia.
Coccoloba obla onga Lindau was described at the same time as ('. crucgcri
and was based on Ricdcl 614, reported to be in ''herb. l'etrop." I have
now seen that specimen and the others cited below, and there is no ques-
tion that (. oblonga is the same as C. cruegeri. The Riedel collection from
Ilheos, Bahia, represents a sizeable, but not unlikely, extention of range.
As Lindau indicated, the material is for all practical purposes sterile,
although the specimens do have immature and poo rly preserved inflores-
cences and flowers.
The range of this species is now from the islands of Trinidad and Mar-
garita to British Guiana and Bahia in Brazil. Collections from Trinidad
and Margarita were cited in an earlier paper (Jour. Arnold Arb. 40: 75.
76. 1959).
British Guiana. Ituribisi Lake. Essequebo Coast. Forest Dept. B.G. 5474
(A, K). Brazil. BAHIA: Ilheos, Riedel 614 (B, LE, P-type of C. oblonga); with-
out specific location, Blanchet 3160A (P). Venezuela. SUCRE: Crist6bal Colon,
Broadway 697 (GH.).
Coccoloba cujabensis Weddell, Ann. Sci. Nat. III. 13: 259. 1850.
Coccoloba longiochreata Hassler, Repert. Spec. Nov. Reg. \eg. 14: 162. 1915.
One specimen of Martius 1241 in the Munich herbarium has been
designated as the lectotype of this species. Coccoloba cujabensis is sug-
gestive of C. ruiziana Lindau. but field study or additional collections are
necessary to determine their correct relationship.
Hassler compared his new species, Coccoloba longiochreata, with C.
alagornsis Weddell and C. floribunda Lindau. He distinguished it on the
basis of the larger ocreae. The type selected by Hassler is a vigorous
staminate flowering specimen. The type of C. cujabensis is a fruiting col-
lection. On the basis of vegetative characteristics they are similar, and
any question is resolved by a study of the Werdermann and Kuntze col-
lections cited below. These show the intermediate conditions of ocreae
and leaf venation between the type specimens of C. cujabensis and C.
(VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 229
longiochreata. Coccoloba cujabensis is not at all comparable to C. ala-
goensis or "C. floribunda."
Bolivia. Beni, Trinidad, Missiones Guarayos, l'erdermann 2366 (Mu). Brazil.
MATO GROSSO: CorumbA, Kuntze s.n. (NY); Cuyaba at Patricio da Silva Manso,
Martins 1241 (B, BR, LE; M-lectotype). Paraguay. Chaco. Fiebrig 1284 (Mi-type
collection of C. longiochreata).
Coccoloba cyclophylla Blake, Contr. U. S. Natl. Herb. 20: 238. 1919.
The type is Curran 47, from San Martin de Loba, Dept. Bolivar, Colom-
bia (US 537207). Blake compared his new species with "Coccoloba
rotundifolia Meisner," now known as Coccoloba leoganensis Jacq., from
Hispaniola. There is no question that C. cyclophylla is the same as C.
caracasana Meisner, which was based on material from Caracas, Venezuela.
Coccoloba cylindrostachya Lindau. Bot. Jahrb. 13: 163. 1890.
Lindau described consecutively Coccoloba cylindrostachya (based on
Glaziou 13135) and C. glaziovii (based on Glaziou 8089) in his mono-
graph of the genus. Both collections were made in the vicinity of Rio de
Janeiro. An examination of the specimens fails to support the differences
which Lindau felt would separate the two species. Therefore, C. cylindro-
stachya has been referred to the synonymy of C. glaziovii.
STo be continued]



JOURNAL
OF THE
ARNOLD ARBORETUM
VOL. XLI JULY 1960 NUMBER 3
STUDIES IN THE GENUS COCCOLOBA, IX.
A CRITIQUE ON THE SOUTIH AMERICAN SPECIES *
RICHARD A. HOWARD
Coccoloba declinata (Vellozo) Martius. Beibl. Flora 20: 90. 1837;
Meisner, Fl. Bras. 5(1): 29. 1855, as to name only.
Polygonum declinatum Vellozo, Flor. Flum. 162. 1825; Icon. 4: 41. 1827.
Coccoloba racemulosa Meisner, Fl. Bras. 5(1): 30. 1855.
Coccoloba declinata var. velloziana Meisner, ibid.
A full discussion of the identity of Coccoloba declinata has been given
under the name C. confusa.
In the original description of Coccoloba racemulosa Meisner cited only an
unnumbered specimen, now in the herbarium at Munich. which was col-
lected in September, 1818, along the river San Francisco near Carinhanha
in Minas Geraes. Brazil. Lindau placed a fragment of this Martius collec-
tion in the Berlin herbarium, but in his monograph he also cited Perrottet
83 from British Guiana. I have seen the latter collection and have referred
it to C. lucidula. Both C. declinata and C. lu:ciula are poorly represented
by herbarium material. When additional collections are available for
study the relationship of these two species should be re-examined.
Coccoloba declinata, as represented by the Martius collection, is not
well defined. The type specimen has partially mature leaves and func-
tionally staminate flowers. The distinguishing characteristic, given by
both Meisner and Lindau, is the origin of the ocreolae at the apex of the
pedicel. However, careful dissection proves that this is an effect of drying
and that while the ocreolae are fused to the pedicels in the type specimen,
they may be free in other collections. The species is similar to C. ilhecnsis,
C. bractcolosa and C. glaziovii, differing in the more scandent habit, the
shorter inflorescences and the smaller leaves, though this characteristic
may be due to immaturity. Field studies or additional collections may
clarify the relationship of these species.
* Continued from volume XLI, p. 229.


JOURNAL OF THE ARNOLD ARBORETUM
Coccoloba densifrons Martius ex Meisner. Fl. Bras. 5(1): 26. pl. 7.
1855.
Coccoloba barbeyana Lindau. Bot. Jahrb. 13: 185. 1890.
Coccoloba pichuana Huber, Bol. Mus. Goeldi 5: 342. 1909.
Coccoloba douradensis Glaziou, Bull. Soc. Bot. Fr. IV. II (Mem. 3f): 571.
1911 (provisional name with mixed type).
Although Lindau recognized ('occoloba densifrons in his monograph
(Bot. Jahrb. 13: 177. 1890), citing the holotype (Martius s.n., from Ega
in Brazil), he did not include the species in his key to the genus.
Coccoloba barbrvana is based on a Ruiz and I'avon collection from
Peru, but I fail to find any reliable differences between these species, either
in Lindau's descriptions or in the specimens he annotated.
Coccoloba pichuana Huber was based on a I)ucke collection from
Obidos, Brazil. In the original description it was compared with C.
padiformis, a very similar species occurring on the north coast of South
America and in Central America. At present the two species are dis-
tinguished on the basis of leaf size and shape and the nature of the
venation, but these differences may prove to be unreliable when both
species are represented by more adequate material. For the present,
therefore, the authentic material of C. pichuana is best referred to the
synonymy of C. densifrons.
I have indicated previously (Jour. Arnold Arb. 41: 46. 1960) that
Coccoloba douradensis, published as a provisional name with a short, four-
word description, is best considered as a nomen nudum. Furthermore, the
collection cited by Glaziou is a mixture, parts of which are Coccoloba
marginata Benth. while the remainder is more correctly assigned here.
Another specimen which may be referred to (occoloba densifrons is
Haenke s.n., collected in Peru on May 24. 1897, and now in the Berlin
herbarium. The sheet, bearing Gross' annotation label, has been given an
unpublished name referring to the city of Guayaquil.
Coccoloba densifrons is similar to Coccoloba sphacroccca (C. padi-
formis), the former being known only in flower and the latter in fruit.
Additional collections may prove these to be the same. On the basis of the
specimens I have seen, they are presently to be distinguished by the finely
reticulate upper leaf surface of C. sphacrococca, in which the primary
veins are scarcely evident. In C. densifrons the primary veins are arcuate
and clearly evident, impressed above but sharply ridged when dry.
Brazil. ACRE: Il'e 9348 (c, K). AMAZONAS: Borba, Rio Madeira, Ducke 466
(A, F, K, NY, Uts): Ega. on Rio Negro, Martius s.n. (M-lectotype . n, x);
Humayta near Tres Casas, Krukof 6120, 6228 (A. BR, LE, NY); Rio Embira,
Krukoff 4667 ((:. L. ENY): Sao Paulo de Olivenca near Palmares, Krukoff 8331
(A, BR, F, LE, NY). (GOYAZ: Serra Dourada, Glaziou 21978 (p-in part. type of C.
douradensis. PARiA: Obidos. Ducke 4866 (BM-type of C. pichuana). Colombia.
ANTIOQcIA: Rio Carepa. Haught 4722 (NY, us). MACDALENA: Santa Marta,
Espina 87 (v). META: Sierra de la Macarena, Philipson, Idrobo & Jaramillo
2104, 2265 (ci). PUI'TMAYo: Rio Putumayo, Cuatrecasas 10820 (us). Dept.
VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 233
unknown: Aguaviva. Dugand 850 (Y); Arroyo de Piedras and Luruaco, Dugand
985 (v); Molinero. Dugand 573 (Y); Rio Toribio, Espina and Giacometto A95
(r, Y); Rio Tucurinca, Dugand 1012 (v). Ecuador. Guayaquil. Haenke 2288
(NY), s.n. (B). Peru. LORETO: Florida, Rio Putumayo. Klug 2260 (A. F, GH, NY);
Gamitana Cocha, Rio Mazan, Schunke 76 (A, F, NY); Iquitos, Mishuyacu, Klug
1077 (r, NY); Iquitos, Rancho Indiana, Mexia 6426 (F. GH); Puerto Arturo,
Yurimaguas. L. Williams 5138 (F); Ucayali, Tessmann 3399 (G, NY, us). SAN
MARTIN: Chazuta, Rio Huallaga, Klug 4127 (F, GH, NY). Without definite
locality: Ruiz & Pavon s.n. (G-type of C. barbeyana).
Coccoloba dioica Karsten ex Lindau, Bot. Jahrb. 13: 170. 1890.
A single fragmentary specimen in the Leningrad herbarium, to which is
attached Karsten's embossed label, bears this epithet. I have previously
referred this species, the type of which was collected near Caracas, Vene-
zuela, to the synonymy of Coccoloba coronata Jacq. (Jour. Arnold Arb.
41: 40. 1960).
Coccoloba dioica Steudel, Nomen. Bot. 390. 1841.
Lindau (Bot. Jahrb. 13: 220. 1890) refers this epithet to Muhlenbcckia
sagittifolia Meisner. The name is used in a list without description by
Steudel, and was reported to apply to a species from Chile. I have not seen
any specimens bearing this name.
Coccoloba diversifolia Jacquin, Enum. Pl. 19. 1760; Hist. Stirp. Am.
114. pl. 76. 1763; Howard, Jour. Arnold Arb. 30: 421-424. 1949, 40:
195-196. 1959.
Although this species is listed in several floristic treatments of South
America, I have not seen correctly identified specimens from the area.
Jacquin attributed the species to Hispaniola in his second treatment and I
have previously cited specimens from the Greater Antilles, Antigua,
Mexico, British Honduras, Guatemala and Florida.
Coccoloba dugandiana A. Fernandez, Mutisia 5: 1. 1952.
There is no question that this is a distinct species, but unfortunately
Fernandez' description has been based on what must be considered
anomalous material. Furthermore, he did not indicate that the species was
dioecious, or that the flowers he described were functionally staminate.
The abnormally emarginate leaves of the type are not characteristic of the
species. Fernandez' statement that the immature achenes are pubescent
is in error, for the fruiting perianth, not the achene, is pubescent.
Additional collections by Cardona and Llewelyn Williams from the
Bolivar and Amazonas regions of Venezuela represent the pistillate plants
to be assigned to this species. Whether these are typical I cannot deter-
mine. The leaves of these collections are comparable to the type in the
thickness of the blade, the denseness of the vein reticulations and the


JOURNAL OF THE ARNOLD ARBORETUM
pubescence. However, all the leaves are smaller and of contrasting shapes,
the blades varying in shape from ovate, obovate or oblong to obovate-
elliptic, and in size from 12 X 10 or 14 X 7 to 15 X 11 cm. long and
broad. The leaf base is rounded or truncate and slightly cordate. The
apex is obtuse to acute or slightly apiculate. The petioles of the mature
leaves are approximately 1.8 cm. long. The inflorescences of these speci-
mens are at most 10 cm. long and 7 cm. wide. Individual branches of the
inflorescence are densely flowered and all parts of the flower, including the
rachis, are densely tomentose. The tomentum on the perianth characterizes
this species. Only three fruits remain attached to the specimens and all
are crushed, indicating that the perianth is fleshy. The hypanthium sur-
rounding the mature achene is only slightly vascular. The mature fruit is
distinctly pubescent, a most unusual condition in this genus.
It is not clear how the leaf shape of this species will finally be defined.
Certainly the pubescent leaves, the compact pistillate inflorescence, and
the pubescent fruit clearly distinguish it at present. Additional collections
of this species from the interior of Colombia and Venezuela are needed.
Colombia. BOVACA: Cairo Guira near mouth of Rio Meta, Ilaught 2624
(us-isotypel. Venezuela. AMAZONAS: Rio Orinoco between San Fernando de
Atabapo and San Antonio, J. Silverio Level 41 (A); Sanariapo. I'illiams 15961
(A, F), 16042 (F). BOLIVAR: Alto Rio Paragua, Cardona 1183 (xvs. 's).
Coccoloba erecta Glaziou, Bull. Soc. Bot. Fr. IV. 11(Mem. 3f): 572.
1911.
This specific name must be considered a nomcn nudunm, for Glaziou's
description states only. "Arbuste droit, fl. blanc jaunatre." He cited his
own collection number 14220 which, unfortunately. is a mixture. Glaziou
apparently mixed his field tags and numbers for the sheet of Glazion 14220
in the Paris herbarium bearing the annotation "Coccoloba erecta Glaz.
n. sp." is actually C. schwarkcana, while a comparable specimen in the
Berlin herbarium, Glazion 14219, is one of two specimens cited by Lindau
in describing C. schwauckcana.
Coccoloba ernstii Johnston. Proc. Am. Acad. 40: 685. 1905.
This species has been referred to the synonymy of Coccoloba crucgeri
Lindau in earlier papers (Jour. Arnold Arb. 40: 74. 1959. 41: 228. 1960).
Coccoloba crrnstii was based on Johnston 2,50 (cGi) from El Valle, South
Hill, Margarita Island, Venezuela.
Coccoloba excelsa Bentham in Hooker. London Jour. Bot. 4: 624. 1845.
C. micropunctata Eyma. MIeded. Bot. Mus. Utrecht 4: 1. 1932.
C. parinmensis var. ihostmrinni Meisner. Fl. Bras. 5(1): 35. 1855.
Although I have examined the material cited by Eyma, as well as addi-
tional specimens, I cannot distinguish between Coccoloba cxcclsa and C.
[VOL. XLI


1960] HOWARD. STUDIES IN THE GENUS COCCOLOBA. IX 235
micropunctata. Eyma believed that C. micropunctata could be distinguished
on the basis of its narrow, tubular, membranaceous ocreolae and its non-
punctulate leaves and inflorescences. The ocreolar distinction appears to
be weak, varying with the age and the sex of the flowering material. In
the description of C. micropunctata Eyma reported his material as "sub
lente dense glanduloso(?)-punctulatae." although an examination of his
material indicated broken hairs and clear or black hair bases which are
comparable to structures often called punctations in C. excelsa. Coccoloba
micropunctata is based on material collected from vines, but the writer's
experience in the West Indies indicates that leaf shape and size, as well as
texture and pubescence, varies widely in lianas.
In the original description of Coccoloba parimensis, Bentham based the
species on Schomburgk material from Rio Parime. He reported further
that "Hostmann 245 from Surinam may possibly be a variety of the same
species." Meisner (Fl. Bras. 5(1): 35. 1855) described this Hostmann
collection as C. parimensis var. hostmanni. Lindau (Bot. Jahrb. 13: 170.
1891) assigned this same material to the synonymy of C. excelsa.
British Guiana. Bartica, Linder 50, 151 (cu, NY); Berbice, Schomburgk 178
(B). 400 (K-holotype, NY); Malali, Demerara River, De la Cruz 2737 (GH, NY).
French Guiana. Acarouany, Sagot s.n. (A, P); Caroni. Melinon s.n. (GH, P).
Surinam. Goddo. Stahel 77 (v-type of C. micropunctata); Gonini, B.W. 3741
(A, NYi; Kwatta hede, Saramacca River headquarters. Maguire 23927, 23929
(A. NY) ; without specific location. Hostmann 245 (B-type of C. parimensis var.
hostmanni), Wlullschlagel 804 (GOET).
Coccoloba excoriata Linnaeus, Syst. Nat. ed. 10. 1007. 1759.
Lindau's treatment of Coccoloba excoriata (Bot. Jahrb. 13: 211-212.
1891) is confusing. Of the many specimens which he cited, the ones I have
seen should be referred to C. venosa. I have not seen a Widgren specimen
from Rio de Janeiro to which he referred, nor the one which Schomburgk
referred to as "Fl. & Faun. Br. Guy. 821." This reference is to "Coccoloba
nivea," a synonym of C. venosa which applies to a plant under cultivation.
Coccoloba venosa is represented in the native flora of South America and
will be discussed later. Coccoloba excoriata L., as correctly applied, is a
synonym of C. tenuifolia L. (Howard, Jour. Arnold Arb. 38: 93. 1957), a
native species of the Bahamas. Cuba and Jamaica, and has not been
reported as either a cultivated or an indigenous plant of South America.
Coccoloba fagifolia Jacq. Hort. Schonbr. 3: 55. pl. 352. 1798.
Although Jacquin published an excellent plate of this species and at-
tributed the plant to the vicinity of Caracas, the species has not been
collected or recognized in recent years. I find Jacquin's species difficult
to place. In the Meisner herbarium, now at the New York Botanical
Garden, there is a packet with three detached leaves bearing the legend,
"Coccoloba fagifolia Jacq. H.S. (mann tremula, verisim. ipsins Jacquini
patris script.) folia speciminis sterilis Herb. Jacquini in Hb. Mus. Vindo-


JOURNAL OF THE ARNOLD ARBORETUM
bon." This appears to represent one of the specimens cited by Meisner in
his treatment of the species for DeCandolle's Prodromus (14: 165. 1856-
"v.s.c. ex Schoenbr. in herb. Jacq. patr. et fil."). These three leaves have
long, thin petioles and blades of thin texture. They do not represent any
species of Coccoloba known to me, and there is no evidence that they even
belong to the genus. Moreover, these detached leaves do not compare
favorably with the original illustration.
Lindau (Bot. Jahrb. 13: 162. 1890) recognized the species and cited
a specimen collected by Boos which was in the Vienna herbarium. How-
ever, since the Coccoloba material in that herbarium was destroyed during
World War II, the material which Lindau cited cannot be identified.
Presumably the Boos specimen was a small one, for Lindau accumulated
for the Berlin herbarium a large number of fragments from important
species, but C. fagifolia is not represented in the material which I have
on loan from that herbarium. No recent material has been assigned to
Coccoloha fagifolia. so its interpretation must rest on the original illustra-
tion, although a comparison of Jacquin's plate with that of C. vircns
Lindley (Bot. Reg. 21: pl. 1816. 1836), which is C. coronata. shows many
similarities. If Jacquin's plate may be accepted as representing an accurate
description, then it seems clear that C. fagifolia should be assigned to the
synonymy of C. coronata Jacq. The proper identification of the Jacquin
material in the Meisner herbarium remains unsolved.
Coccoloba fallax Lindau, Bot. Jahrb. 13: 172. 1890.
Coccoloba caurana Standley, Publ. Field Mus. Bot. 22: 73. 1040: Llewelyn
Williams. Explorac. Bot. Guayana Venez. 189. 1942.
Coccoloba fallax seems to be distinguished by the fascicled inflorescence
of simple racemes, the conspicuously large ocreae, especially those clustered
at the apex of the stem and around the racemes, and the strongly keeled
midrib. The type collection of Coccoloba caurana has immature in-
florescences, but comparable mature material has been found in Trinidad.
There is no question that C. caurana is the same as C. fallax.
The Trinidad specimens have been cited in a previous paper, in which
the species was also recognized as occurring in Venezuela (Jour. Arnold
Arb. 40: 79. 1959). The type of C'. fallax is Crueger 114. The lectotype
selected is the specimen in the Gottingen Herbarium. The holotype of
C. caurana is Llewelyn Williams 11366 (F).
Coccoloba fasciculata Meisner in Warming, Symbol. Floram Bras. 6:
128. 1870.
Lindau cited this epithet and reference in the synonymy of "C. longi-
pendula" in his monograph (Bot. Jahrb. 13: 177. 1890). The original
publication is a report on Warming's collections from Brazil. One entry
is "Coccoloba fasciculata Wedd.. Meissner in DC l.c. 166 and Mart. I.c.
27," with a citation of "Lagoa Santa: . . . Serra da Piedade legit Warm-
[VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 237
ing Jan.-Febr. florentem." This is obviously a misidentification and not a
new entity, as Lindau implied. Lindau referred the epithet to the synonymy
of C. longependula, which I cannot distinguish from the older C. sticticaulis
(q.v.). Lindau also cited the Warming collections as follows: "Rio de
Janeiro in silvis ad Lagoa Santa: Warming n. 129; ad radices montis
Serra da Piedada: Warming n. 126." I have seen specimens of each and
have referred them to C. sticticaulis Weddell.
Coccoloba fasciculata Weddell, Ann. Sci. Nat. III. 13: 258. 1849.
I have discussed this species in an earlier paper (Jour. Arnold Arb. 41:
44. 1960) and have referred it to the synonymy of Coccoloba arborescens
(Vellozo) Howard. Coccoloba fasciculata Weddell was based on Blanchet
796 collected in Bahia Province, Brazil. The type is in Geneva.
Coccoloba fastigiata Meisner, Fl. Bras. 5(1): 34. 1855.
Coccoloba goudotiana Meisner. Fl. Bras. 5(1): 35. pl. 13. fig. 1. 1855, not
Weddell.
In the original description Meisner described the species and a variety
"f3 glabrata," and cited specimens to illustrate both taxa. This is con-
trary to his usual practice, in the same volume and treatment, of con-
sidering the species to consist of varieties for which specimens are cited.
The variety glabrata Meisner, represented by "Schott 5537 (914)" and
part of Schott 5540 (in Vienna), is better assigned to Coccoloba mosenii.
For the type variety of the species, Meisner cited Schott 5540, collected
in Rio de Janeiro Province, Brazil, and Schomburgk 1262, in the Berlin
herbarium. The latter collection was supposedly made in British Guiana.
Lindau, who worked with the Berlin material around the year 1890,
mentioned the Schomburgk collection under the name Coccoloba fastigiata,
but attributed it to Meisner, "pr. p. ex. Meissn." I cannot determine
whether Lindau actually saw this material, but I suspect that he did not,
for in his citation of specimens seen (Bot. Jahrb. 13: 224. 1890) he listed
Schomburgk 1262 as C. marginata. I have on loan the Coccoloba material
from the Berlin herbarium and I fail to find the Schomburgk collection
represented. I suspect that Meisner was in error in including the Schom-
burgk specimen and that C. fastigiata should be typified only by Schott
5540. Meisner stated that the specimen he saw was in the Vienna herba-
rium. Unfortunately, the Coccoloba material in this herbarium was
destroyed during World War II. A photograph and a fragment of the
holotype is at the Chicago Natural History Museum, and a more complete
specimen is in Brussels. Therefore it seems wise to designate the specimen
at Brussels as the lectotype.
Coccoloba goudotiana Weddell was based on a Goudot specimen col-
lected near San Luis, Colombia, and now in the Paris herbarium. Meisner
elaborated on the original description and illustrated the species. In addi-
tion to the Goudot specimen, Meisner also cited a collection by Pohl in


JOURNAL OF THE ARNOLD ARBORETUM
the Zuccarini herbarium at Munich. I have seen both of these specimens
and it seems clear that the illustration in Flora Brasiliensis was made by
combining features present in both specimens. Coccoloba goudotiana
Weddell is C. obovata, while C. goudotiana as described and illustrated
by Meisner is C. fastigiata.
Brazil. Rio DE JANEIRO: Aguacu. Schott 5540 (BR-lectotype. F-photo and
fragment of Vienna material); Mana, Glazion 788< (,. c. p>; Realengo near
Sapopemba, Glaziou 11441 (LE, P); without specific locality. Jan: s.n. (,LE).
Without locality: Pohl s.n., in Herb. Zuccarini as cited by Meisner (i ).
Coccoloba fastigiata var. glabrata Meisner, Fl. Bras. 5(1): 34. 1855.
Meisner cited "Schott 5537 (914)" and, in part, Schott 5540 as repre-
sentatives of his new variety with specimens in the Vienna herbarium.
Lindau recognized the variety (Bot. Jahrb. 13: 172. 1890), citing Schott
5537 and Pohl 914 with specimens at Vienna and Berlin. The Vienna
material has been destroyed and material from these collections is not
included in the specimens of Coccoloba from the Berlin herbarium. A
photograph of Schott 5540 and a fragment of this specimen from Vienna
are preserved in the herbarium of the Chicago Natural History Museum
and clearly identify the specimen as C. jastigiata. It appears that the
fragment of Schott 5537 (F), all that remains of the collection, should be
referred to C. moscnii Lindau (q.v.).
Coccoloba ferruginea Endlicher, Catalogus Hort. Acad. Vind. 1: 274.
1842; Ettingshausen, Blattskel. d. Apetalen 91. pl. 26. fig. 2. 1858.
Lindau referred this epithet to the synonymy of his (occoloba excoriata
(q.v.). Endlicher cited the name without description as a plant which had
been obtained from the Berlin botanic gardens and was cultivated in
Vienna gardens. Ettingshausen illustrated a leaf skeleton, using the name
Coccoloba ferruginea. He acknowledged that his material was from a
cultivated specimen, but did not discuss the illustration or the species.
The leaf skeleton is inadequate for positive identification, although it is
probably C. vcnosa L.: none of the specimens I have seen hears the name
C. Jcrruginca.
Coccoloba filipes Standley, Trop. Woods 40: 14-15. 1934.
In the original description Standley wrote, "the distinctive characters
of the species are the greatly elongate, many-flowered, lax racemes, with
flowers on conspicuously long and slender pedicels." The type. Dugand
380, was collected in flower at Santa Rosa. west of Barranquilla, Colombia,
on March 13, 1933. The flowers, although borne singly at each nodule on
the inflorescence axis, are apparently functionally staminate, producing
pollen grains. The specimen, now at the U.S. National Herbarium, bears
two terminal, thin-textured, darkened leaves, suggesting that the plant is
deciduous. The leaves arise from above the base of the ocreae.
VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX
No additional material which can be referred to this species has been
seen. It appears to be distinct among the Colombian species of the genus
and, on the basis of the material available for study, is not closely related
to or even similar to any other South American species thus far described.
Although several species from Cuba or Puerto Rico have similar long
pedicels, none are so tenuous.
Coccoloba firma Martius ex Colla, Herb. Pedem. 5: 48. 1836; Eyma,
Meded. Bot. Mus. Utrecht 4: 3. 1932.
The original description, while brief, is explicit and clearly intended to
be that of a new species. In a list of species Colla attributed the name to
Martius. No more definite location than Brazil was given. Eyma also
used the name attributed to Martius in comparing it with his proposed
new species, Coccoloba micropunctata. Eyma cited specimens at Kew
and the Natal Herbarium, Durban, which he said were listed under the
name Coccoloba populifolia Weddell. The correct name for the latter
species is C. alnifolia Casaretto. Eyma apparently recognized both C.
firma and "C. populifolia." I have not seen the specimen from the Natal
Herbarium, South Africa, nor have I been able to locate a specimen so
named at Kew. There is a Martius sheet at the British Museum bearing
the name C. firma, which I believe is properly C. alnifolia. No material
of C. firma was found in the collections of Martius which I have on loan
from Munich, Brussels and Leningrad. I am therefore reluctant to accept
this unsupported name until the material which Colla saw, or at least that
which Eyma cited, has been located. If this should prove to be the same
as C. alnifolia Casaretto (1844), as I suspect, the older C. firma Martius
ex Colla should be accepted as the correct name.
Coccoloba gardneri Meisner, Fl. Bras. 5(1): 36. pl. 15. 1855; Howard,
Jour. Arnold Arb. 41: 44. 1960.
This species is based on Gardner 600 from Serra dos Orgaos, Brazil.
In the original description, which is clearly based on the Gardner collec-
tion, Meisner cited Polygonum frutescens Vellozo in synonymy and also
Coccoloba vellosiana Casaretto, the latter with some doubt. I have pre-
viously considered the identity of species described by Vellozo and Casa-
retto and have also concluded that C. gardneri Meisner represents still a
third species. Meisner's species was placed in the synonymy of C.
crescentiifolia by Lindau, but clearly it is not the same as C. arborescens
(Vell.) Howard, loc. cit. I cannot adequately define the species on the
basis of the single collection available and wish to call this problem to the
attention of future collectors who may visit the area of Rio de Janeiro.
Coccoloba glaziovii Lindau, Bot. Jahrb. 13: 163. 1890.
Coccoloba cylindrostachya Lindau, Bot. Jahrb. 13: 163. 1890.
239


JOURNAL OF THE ARNOLD ARBORETUM
Coccoloba sublobata Heimerl, Denkschr. Akad. Wien. Math.-Nat. Klasse 79:
244. 1908.
Glaziou 8089 is the only collection cited by Lindau in the original
description. Of the specimens cited from four herbaria, I have designated
that at Berlin as the lectotype.
I am unable to distinguish between Coccoloba glaziovii and C. cylin-
drostachya which Lindau based on Glaziou 8089 and 13135 respectively.
Lindau distinguished between them by placing C. glaziovii in a group of
species with the rachis glabrous and C. cylindrostachya in a contrasting
group having the "rachis vario modo pilis instructa." His species descrip-
tions amplify this supposed difference and suggest others, e.g.. the position
of the petiole in relation to the base of the ocrea, which, however, is not
borne out in a re-examination of the type specimens cited. I have seen all
of the specimens which Lindau cited from both collections and can only
conclude that one species is represented.
Glaziou 3087 was annotated by Lindau as a possible new species. It is
a vigorous shoot, probably adventitious in nature, with longer internodes,
larger leaf blades and more pubescent parts. Lindau never published the
name written on the sheet. I believe the specimen should be included in
Coccoloba glaziovii. Another unpublished name honoring Lindau is found
on the collection Schwacke 13673 with the author given as Schwacke.
Although the specimens seen have more delicate branches, more tenuous
and pendant inflorescences and more oblong leaves, I feel that this mate-
rial is properly assigned to C. glaziovii.
The Chicago Natural History Museum has a photograph of the holo-
type of Coccoloba sublobata which was in the herbarium at Vienna. The
type was lost during World War II and the only duplicate of this collection
which I have seen is in the Berlin Herbarium. This Berlin specimen bears
an unpublished name, attributed to Heimerl and referring to the subundu-
late leaves, which is more appropriate than the one actually published.
Although the type of C. glaziovii is a staminate plant, there is no question
that Wacket's collection represents the same species in fruit. Heimerl's
original description does not refer to the type collection by number. The
number 12 cited below is legible in the photograph of the type and on
the isotype.
In his monographic treatment of Coccoloba striata (Bot. Jahrb. 13: 164-
165. 1890), Lindau cited numerous specimens from the Brazilian states of
Bahia, S1ao Paulo, Rio and Pernambuco. 1 have seen only one of these,
Mosen 3664, which I believe to be more properly associated with C.
glaziovii. Coccoloba striata, based on a Schomburgk collection from the
Roraima area, appears to be a northern species. The remainder of the
collections cited by Lindau should be examined for their correct relation-
ship here.
Brazil. MINAS GERAES: Itabira do Matto Ventro, Schwacke 13673 (B, P).
PARANA: Volta Grande. Dusin 11966 (cr, NY). RIO DE JANEIRO: Rio de Janeiro,
Glaziou 3087 (BR), 8089 (B-lectotype, BR , C, , GH, LE, s), 13135 (type collection
[VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 241
of C. cylindrostachya, B, BR, G, LE). Slo PAULO: Serra do Cubatao, near Santos,
Wacket 12 (B-isotype of C. sublobata); Sorocaba, Mosen 3664 (p).
Coccoloba goudotiana Weddell, Ann. Sci. Nat. III. 13: 260. 1849.
This species is based on Goudot s.n. from San Luis, Colombia. The
type specimen is in the Paris herbarium and the species is clearly synony-
mous with Coccoloba obovata HBK. (1817).
Coccoloba goudotiana Weddell sensu Meisner, Fl. Bras. 5(1): 35. pl.
13. fig. 1. 1855.
In the original publication Meisner did not intend to describe a new
species, but only to list the Weddell species. However, he cited the
Goudot material used by Weddell, as well as a Pohl collection from the
Zuccarini herbarium which is not the same species. Meisner's description
and illustration combined features of both plants. I have seen the Pohl
specimen, now in the Munich herbarium, and refer Meisner's interpreta-
tion of Coccoloba goudotiana to the synonymy of C. fastigiata Meisner.
Campderia gracilis Meisner, Fl. Bras. 5(1) : 23. pl. 6. 1855; DC. Prodr.
14: 170. 1856.
Meisner described and illustrated this species in 1855. He cited a
Spruce collection without giving a collector's number from around Barra
in the "Prov. Rio Negro." The following year in the Prodromus he cited
Spruce 958, reporting specimens to be in the DeCandolle Herbarium and
the Herbarium Monacense (Munich). The illustration in Flora Brasiliensis
appears to be a compilation of the two specimens. The original pencil
sketch of the illustration of flowers and fruits is attached to the sheet at
Munich and this sheet should be designated as the lectotype. However,
I cannot determine the source of the fruiting material which Meisner
illustrated, since comparable achenes do not appear on either sheet at
the present time. Lindau referred Campderia gracilis to the synonymy of
Coccoloba ovata Bentham. This appears to be the correct specific place-
ment on the basis of the material I have examined, but there is a question
as to whether C. ovata belongs in the genus Coccoloba. This matter will
be discussed further under C. ovata. I have seen specimens of Spruce 958
from Munich (lectotype), Berlin, Geneva, the Gray Herbarium, Leningrad
and Paris.
Coccoloba gracilis HBK. Nov. Gen. 2: 176. 1818.
Two specimens, one in the Willdenow Herbarium and one at Paris, can
be considered to be authentic for this species. Both are relatively small
and are characterized by slender, nearly geniculate inflorescences of func-
tionally staminate flowers. The original description has information appli-
cable to labels on both specimens but does not exactly match either. It


J(OURNAL OF THE ARNOLD ARBORETUM
appears desirable to consider the Paris specimen as the lectotype since
the only label on this sheet bears the number 3498 of Humboldt and the
geographic location Rio Cachivacu given at the time of publication.
These two specimens are inadequate for accurate determination as to
genus. The floral structure and that of the ocreolae and the bracts appear
to bridge the few weak characteristics used to distinguish Coccoloba and
Ruprchtia in staminate flowering condition. With the material available,
the only possible course to follow is that taken by the previous authors
and monographer who placed the species in Coccoloba. Additional material
is needed to determine its correct position.
Macbride (Publ. Field Mus. Bot. 13: 460. 1937) assigned a fruiting
collection. Wcbcrbaucr 6982, to this species in his treatment for the Flora
of Peru. The Weberbauer collection bears no data regarding the plant
or the place of collection beyond "Peru." The old inflorescence axes are
4-8 cm. long and are erect. The fruiting peduncles are 2-2.5 mm. long.
Fruits, one of which retains the fruiting calyx, are preserved in a packet.
The perianth lobes are free nearly to the base in the fruit. The achene
is only slightly triangular in outline, dark brown and shiny. This
specimen is clearly a Coccoloba, though not necessarily the same as the
authentic material of C. gracilis HBK.
Peru. Rio Cachiyacu, Humboldt 3498 (p-lectotype; without number or loca-
tion fHerb. Willd. 77011).
Coccoloba grandiflora Lindau, Bot. Jahrb. 13: 175. 1890.
Lindau cited two collections in the original description, Glaziou 14217
and Micrs 4657. The former should be selected as the type collection and
the specimen at Copenhagen designated as the lectotype. This is the only
sheet among those cited below on which the label states. "Coccoloba
grandiflora Lindau n. sp." Lindau's work was based upon the material
in the Berlin herbarium - the fragment of a branch together with one
inflorescence having one detached and two attached leaves. I have not
located the Miers collection.
Superficially, this species resembles Coccoloba ctnuiflora Lindau which
is based on cultivated material of unknown origin. The type of C. tenui-
flora was from a greenhouse plant, while C. grandiflora is from native
material. I cannot determine whether cultivation (and, specifically. green-
house conditions) could create the differences seen in these specimens.
Coccoloba grandiflora has more conspicuous ocreolae which are membra-
naceous, split longitudinally and flaring. In all other characteristics the
species are similar. Both species are known only from flowering material.
Brazil. MINAS GERAES: Glaziou 20439 (13, LE, NY). Rio DE JANEIRO): NuOVO
Freiburg, ;laziou 14217 (C-lectotype. 1, K. LE, U ).
Coccoloba grandis Bentham in Hooker. London Jour. Bot. 4: 624. 1845.
This species is based on Schomburgk '25 collected on the Rio Branco,
[VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 243
British Guiana. Lindau placed the species in the synonymy of Coccoloba
latifolia Lam. After having examined the type in the British Museum,
I agree with this placement.
Coccoloba guaranitica Hassler, Repert. Sp. Nov. 14: 161. 1915.
Coccoloba guaranitica var. opaca Hassler, ibid. 162.
Hassler compared his new species with Coccoloba schomburgkii, but on
the basis of the few specimens I have seen, it seems more nearly comparable
to C. padiformis from Venezuela, although the leaves of C. guaranitica
are smaller and more rounded at either end. The fruits of both species
are known only from immature specimens, but in both the perianth lobes
are conspicuous, imbricate and appear to surround only the upper half
of the achene.
Hassler did not select a type in the original description in which he
mentioned two collections, Fiebrig 1429 and 1440. I have not seen the
latter, but the former is a shoot of vigorous and rapid growth.
The variety which Hassler described differs from the species in having
smaller leaves, the margins of which are undulate-crenate. It also has
a shorter inflorescence. The type of the variety, Fiebrig 4305, is a mature
shoot system with many lateral flowering branches. It seems quite ap-
parent that the specific name has been applied to younger and more
vigorous specimens and that of the variety to the more mature branching
specimens and thus the variety is not worthy of recognition.
A specimen of Fiebrig 4305, the type of C. guaranitica var. opaca, was
studied by Gross and annotated with both a specific name honoring Fiebrig
and a varietal name referring to the crisp leaf margin. Neither name
has been published, to my knowledge, although Gross published other
epithets in the genus in small notes, often in obscure publications.
Paraguay. Boquer6n, Puerto Casado, Pedersen 4043 (A, c); Chaco, Fiebrig
1429 (M-isotype); between Rio Apa and Rio Aquidaban, Fiebrig 4305 (type
collection of C. guaranitica var. opaca, B, GH, M, P); Bahia Negra. Rojas 13708
(BR, P, w).
Coccoloba guianensis Meisner, Linnaea 21: 264. 1848.
Several specimens were cited in the original description, but no type
was selected. The specimens are obviously the same as the slightly anom-
alous material described earlier by Bentham as Coccoloba marginata.
The variation in material called C. guianensis and the nomenclature of the
complex has been discussed in other papers (Howard, Jour. Arnold Arb.
40: 84, 85. 1959; 41: 45, 46. 1960).
Coccoloba gymnorrhachis Sandwith, Kew Bull. 1932: 221. 1932;
Eyma, Polygonaceae, Guttiferae & Lecythidaceae of Surinam, 111.
1932.
Sandwith based the original description of this species on a flowering


JOURNAL OF THE ARNOLD ARBORETUM
specimen. Sandwith 168, from British Guiana. The plant was described
as a "bush-rope" and the type shows strong characteristics of a liana. The
leaves, broadest above the middle and narrowed to the base, are rigidly
coriaceous with the arcuate venation impressed above. The flowers are
borne in clusters with tightly appressed ocreolae. Eyma supplemented
the original description with a fruiting specimen from Surinam. The
Richard collection cited below is also in fruit and is the same as the
material which Evma cited, but these two fruiting collections are not
good matches for the type collection and may not belong here. The fruits
are warty, as though insect-infested, but no evidence of insect larvae
could be found.
Additional collections are needed before the morphological characteristics
of Coccoloba gymnorrhachis are fully understood and the species clearly
defined.
British Guiana. Essequebo River, Moraballi Creek near Bartica, Sandwith 168
(K-holotype. Nx). Surinam. Brownsberg, B.W'. 6773 (K). Location unspecified:
L.C. Richard s.n. (p).
Coccoloba ilheensis Weddell. Ann. Sci. Nat. Il1. 13: 258. 1849.
Coccoloba membranacea Klotsch, Linnaea 14: 289. 1840, nomcnn nudum.
This is a poorly defined species requiring both field study and many
more collections for an accurate interpretation. This species is similar
to Coccoloba glaziovii, C. ochrcolata and C. confusa. While collections
representing the type of each of these species can be distinguished on
sight, I cannot find any reliable morphological characteristics useful in
separating them in a key.
Brazil. BAHIA: Ilheos. Martius 1240 (p-holotype. BR. G. GH. LE. M): Luschnath
s.n. "1839" (LE). Without location: Luschnath 42 (LE .
Coccoloba japurana Meisner, Fl. Bras. 5(1): 25. 1855.1
This species is based on a Martius specimen from Ega in the Rio Negro
of Brazil. Authentic material is in the Munich herbarium and is repre-
sented in the collections of the Gray Herbarium by a photograph. The
photograph is a montage of two herbarium sheets and two labels. One
label has the hand-written annotations, "Coccoloba acuminata" and "Cocco-
loba japurana" and on the same sheet is the annotation, "Alsodeia japu-
rana Radlk." One of the sheets has specimens with immature inflorescences.
This almost completely overlaps the other sheet from which an infructes-
cence protrudes. The fruit on this axis is clearly not that of a Coccoloba.
The name Alsodcia japurana Radlk. (Sitz-ber. Math.-Phys. Klasse
Akad. Miunchen 20: 182. 1891) is recorded in the first supplement of
Index Kewensis as an observation. There is no reference to the name
Coccoloba japurana Meisner in the article cited. However, Lindau at-
Rinorea japurana (Meisner) Howard, comb. nov. Basionym: Coccoloba
japurana Meisner in Martius, Fl. Bras. 5(1): 25. 1855.
I VOL. XLI


1960] HOWARD. STUDIES IN THE GENUS COCCOLOBA, IX 245
tributed the transfer to Radlkofer in a list of excluded species and in a
footnote in his monograph i.e., "C. japurana MEISSN. = Alsodeia japurana
RADLK." (Bot. Jahrb. 13: 220. 1890).
Alsodeia is a genus of the Violaceae for which most modern writers
use the name Rinorea. Blake, who monographed the American species
of Rinorea (Contr. U. S. Natl. Herb. 20: 317. 1924), listed Alsodeia japu-
rana as a "doubtful species" with the comment, "This was described from
specimens with very young flowers. It is said by Radlkofer to be allied
to A. racemosa." So many aspects of this misplaced epithet have been
overlooked that I have made the new combination in Rinorea primarily
to place the name in indices for the aid of future workers. An adequate
interpretation of the photograph is impossible. Meisner's original descrip-
tion could well be a Coccoloba; Lindau saw the material now in the
Munich herbarium and would surely have recognized a Coccoloba as to
genus; Radlkofer worked on Alsodeia and would certainly have recognized
that genus; Blake did not accept the species, but it is not clear what
material he saw or to which description he referred. The fruit in a photo-
graph of a properly labelled specimen, supposedly the type, is not a
Coccoloba, but appears to be a Rinorea. The writer cannot solve the
puzzle and may not have placed the specific epithet in its proper niche.
Coccoloba laevis Casaretto, Nov. Stirp. Bras. 71. 1844; Lindau, Bot.
Jahrb. 13: 186. fig. 40. 1890.
Coccoloba cordifolia Meisner, Fl. Bras. 5(1): 37. 1855.
Casaretto did not cite a specimen in the original description, but in the
same publication he described other species based on his own collections.
One sheet, Casaretto 2264, in the Turino herbarium, fits the description
of Coccoloba laevis in all details and should be considered the holotype.
I assume that Casaretto unintentionally omitted the citation of a specimen.
In the original description of Coccoloba cordifolia, Meisner cited several
specimens without selecting a type. He indicated the affinities of his new
species with C. nivea, C. cordata, and C. candolleana and cited in synonymy
"Coccoloba uvifera Salzmann Mss. in Herb. (non Linn.)." In his treat-
ment for the Prodromus (14: 155. 1856) he cited the same specimens,
but he did not repeat the suggested synonymy or the possible relation-
ship. Instead he placed his species next to C. laevis, which he suggested
might be identical with C. cordifolia or C. candolleana. Lindau was the
first to place C. cordifolia Meisner in the synonymy of C. laevis Casaretto,
a decision with which I agree.
The Salzmann collection from Bahia which Meisner cited is represented
in several herbaria and the collection at Kew shows the full range of
variation from the small-leaved type of Casaretto's species to the broader
and more cordate leaved types found in Blanchet 3528.
Lindau described and illustrated the fruit of Coccoloba laevis, but
although I have seen most of the material he cited, I have not found a
fruiting specimen, or even a single fruit, among them. If the figure pub-


JOURNAL OF THE ARNOLD ARBORETUM
lished by Lindau can be trusted, the possibility of a hybrid origin of this
species should be examined. Coccoloba marginata or C. uvifera would
seem quite likely as parents. At present C. laevis is an extremely variable
species which is not clearly delineated.
Brazil. BAHIA: Amaralina, San Salvador. Dahlgren s.n. (F ; Ilha de Cal.
Curran 106 (TGi, NY, Y); Itaparica Island, Casaretto 2264 i To-holotype);
Jacobina, Moritiba, Blanchet 100 (c, NY), 3528 (C, LE, P . PERNAMBUCO:
Pernambuco. Guillamin s.n. (F). Rio DE JANEIRo: Rio de Janeiro. Glaziou 11445
(B, K, P), Salzmannn 476 (c). S.n. (K, LE. P). Locality uncertain: Maceio.
Gardner 1391 (K).
Coccoloba lanceolata Lindau ex (laziou, Bull. Soc. Bot. Fr. IV. 11
(Mem. 3f) : 573. 1911, nomen nudum.
The collection Glaziou 19764 was cited by the author in the original
publication with the brief description, "liane. f1. blanchatres, fruit noir."
The specimens seen are obviously from climbing plants. The leaves and
infructescence are borne on short lateral shoots. While the epithet has
no acceptable standing at the present time. I do not wish either to
describe the plant more fully or to place the name in synonymy until
further material from southern Brazil is available for study. It is probable
that this collection should be assigned to C. salicifolia. The leaves of the
Glaziou specimen, however, are more lanceolate-oblong in shape, less acu-
minate at the apex and thicker in texture. The fruits match the illustra-
tion given by Lindau for C. salicifolia.
Brazil. MINAS GERAES: Riacho das Varas, Glaziou 19764 I B. c, K).
Coccoloba latifolia Lamarck, Dict. Encycl. 6: 61. ill. 316, f. 4. 1804.
Coccoloba grandis Bentham in Hooker, London Jour. Bot. 4: 624. 1845.
Lamarck described this species from material cultivated in the Jardin
des Plantes, Paris. I have not seen authentic material, but his illustra-
tion is of a single detached leaf which does not represent well the species
as currently accepted. The description, although somewhat vague, seems
applicable, but, since Coccoloba latijolia is similar to C. mollis, comparable
field observations would be helpful. It differs in an almost complete
lack of puberulence. in its much stouter and generally hollow stems and
in having strongly bullate leaves. I have seen C. latifolia in Trinidad
where it is a characteristic plant of savanna areas. Its habit is distinctive
and this, together with the presence of many biting ants in the large
ocreae. makes it long remembered by collectors.
Coccoloba grandis Bentham is based on Schomburgk 825. Lindau
placed the species in the synonymy of C. latifolia, where it clearly belongs.
Lindau referred three collections by Burchell from Sio Paulo and Para
to this species. I have seen one sterile sheet of Burchell 3982 in the her-
barium at Kew and feel that this sheet, at least, should be considered
the adventitious leaf form of Coccoloba warmingii Meisner.
I VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 247
In addition to the localities listed below, the species is also known from
Trinidad and its adjacent islands (Jour. Arnold Arb. 40: 81. 1959).
Brazil. CEARA: Ceara, Curran 36 (cH). MARANHAO: Campo de Boa Esperanca,
Froes 1817 (A, NY). British Guiana. Mapenna, Courantyne River, B.G. Forest
Dept. 2601 (A); Rio Branco, Schomburgk 825 (BM-type of C. grandis). French
Guiana. Cayenne, savannahs along St. Madeleine Rd., Broadway 750 (eH, NY);
without locality, Barbier s.n. (A), Sagot 486 (A). Surinam. Koboerie, Herb. B.lW.
5929 (A); without specific locality, Hostmann 682 (GH), s.n. (BR), IWHllschagel
s.n. (M). Venezuela. DELTA AMACURO: Serrania Imataca, N. of Rio Guanamo,
IWurdack & Monachino 39724 (A).
Coccoloba laurifolia Jacquin, Hort. Schoenbr. 3: 9. pl. 267. 1798.
This remains a troublesome name which I cannot place satisfactorily.
Meisner recognized the species (DC. Prodr. 14: 165. 1856), noting that
the type locality was Caracas, Venezuela, and he cited one specimen
(7699) in the Willdenow herbarium. This specimen consists of two
sterile shoots obtained from a plant cultivated in a botanic garden and
certainly is not the Jacquin type. It is properly referred to Coccoloba
diversifolia Jacq.
In his monograph (Bot. Jahrb. 13: 158. 1890) Lindau also accepted
Jacquin's name and cited two specimens (without known collectors and
from Caracas, Venezuela) to be found in the Delessert and Vienna her-
baria. Lindau cited many additional collections from Florida, the Ba-
hamas, Cuba, Hispaniola, Puerto Rico and the Virgin Islands. A specimen
in the Prodromus herbarium at Geneva which Lindau saw, and the one
I believe he cited, was probably collected by Bertero in Hispaniola. The
Antillean and Florida material cited by Lindau has been referred to
Coccoloba diversifolia Jacq.
Although I have suggested that Coccoloba laurifolia Jacq. and C.
diversifolia Jacq. may be the same (Jour. Arnold Arb. 40: 195-196.
1959), I am not entirely convinced of it. A re-examination of all material
available to me from Venezuela has failed to reveal any plants which
can be compared satisfactorily with the description and illustration supplied
by Jacquin. The closest comparison in Venezuela would be with C. padi-
formis Meisner based on the collection Moritz 377 from Caracas. Material
from Central America which I have cited for C. padijormis (loc. cit. 210-
211) and additional collections to be cited in this study are not exactly
comparable to Jacquin's description and illustration. These differences at
present are primarily in the venation as related to the texture of the leaf
blade and in the shape of the leaf apex. A field study of Coccoloba plants
in the vicinity of Caracas will be necessary to determine what species
Jacquin had as a basis for his description and illustration of C. laurifolia.
Coccoloba laxiflora Lindau, Bot. Jahrb. 13: 191. 1890.
The holotype in the Berlin herbarium is Glaziou 11444 from Rio de


JOURNAL OF THE ARNOLD ARBORETUM
Janeiro, Brazil. This species is to be referred to the synonymy of
Coccoloba ramosissima Weddell.
Coccoloba lehmannii Lindau, Bot. Jahrb. 29(Beibl. 49): 7. 1895:
Howard, Jour. Arnold Arb. 40: 200. 1959.
Coccoloba lehmanni Lindau. Repert. Sp. Nov. 1: 156. 1905.
Coccoloba williamsii Standley, Publ. Field Mus. Bot. 11: 148. 1936.
This species has been discussed in an earlier paper in which its range
was extended to Central America and additional species from that area
placed in its synonymy. The selection of a lectotype was also discussed
at that time. Another similar species is Coccoloba lepidota A. C. Smith
(q.v.). Additional material may show that this, too, should be placed in
synonymy here. Two collections from Peru, Tessman 3896 and 5258 from
Iquitos along the Amazon have been tentatively referred to this species.
Several herbarium specimens have been seen of a collection made in 1871
from a cultivated plant in the Calcutta Botanic Garden. No collector or
data are given on the sheets, which have carried the name "Coccoloba
excoriata." These are clearly to be referred to the present species.
Colombia. ANTIOQUIA: Villa Arteaga. Lopez & Sanchez 40 (us): Cauca.
Lehmann 7560 (B-lectotype). META: Puerto Lopez. E.L. & R.R. Little 8294
(NY). Venezuela. ANZOATEGVI: NE. of Bergantin. Stevermark 61217 (F);
BARINAS: Barinitas. Aristeguieta 1702 (us). MERIDA: Between Hacienda Agua
Blanca, above La Azulita and Rio Capaz. Stevermark 56127 (F).
Coccoloba lepidota A. C. Smith, Brittonia 2: 150. 1936.
This species was distinguished by Smith by the "characteristic scales
of the petioles and young stem parts." Smith compared Coccoloba lepidota
with C. ovata. which is clearly distinct. The separation of C. lepidota and
C. lehmannii is more difficult and additional material may show that C.
lepidota should be another synonym of that species. The lectotype of
C. lehmannii. Lehmann 7560 (B), has shorter, more elliptic leaves and
a predominance of simple pubescence. Coccoloba lepidota, as represented
by the type collection, Krukoff 5660, has larger obovate-oblong leaves,
broadest above the middle and tapering to an obtuse or truncate base.
The young stems, petioles and ocreae are covered with lepidote scales
and bear lesser amounts of simple hairs and resinous excretions. A tend-
ency towards this development is found in the type collection of C.
lehmannii and in the other collections cited below. It is probable that
C. lepidota is an extreme variation of C. lehmannii.
Brazil. ACRE: Near mouth of Rio Macauhan. Krukoff 5660 (NY-holotype,
A, LE, M, W), 5659 (A, LE. M, NY).
Coccoloba leptostachya Bentham, Bot. Sulph. 59. 1856; Meisner,
DC. Prodr. 14: 163. 1856.
As has been pointed out in an earlier paper in this series (Jour. Arnold
VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 249
Arb. 40: 188. 1959), Bentham described this species, citing the type
locality as "Libertad in Colombia." The type is a Barclay specimen
at Kew. I have studied this and have concluded that the specimen was
collected in Central America. Recent collections from Libertad in El
Salvador proved to match the Barclay collection well. I have seen no
comparable material from Colombia. Coccoloba leptostachya Bentham is
referred to the synonymy of C. barbadensis Jacquin (1760), which is
known from Mexico, Guatemala and El Salvador. The species need no
longer be considered in the South American flora.
Coccoloba longependula Martius ex Meisner, Fl. Bras. 5(1): 27. pl. 9.
1855; Lindau, Bot. Jahrb. 13: 177. 1890.
After an examination of the type of this species (Martius 759 from Minas
Geraes, Brazil IM ), this has been referred to the synonymy of Coccoloba
sticticaulis Weddell (q.v.).
Coccoloba longiochreata Hassler, Repert. Sp. Nov. 14: 162. 1915.
This species is clearly the same as Coccoloba cujabensis (q.v.) and has
been referred to synonymy there. Hassler cited two collections, Fiebrig
1284 and 1443, in his own herbarium. The collections were made in the
Gran Chaco at Puerto Talavera, Paraguay.
Coccoloba longipes S. Moore, Trans. Linn. Soc. II. 4: 446. 1895.
Coccolobis padifolia Rusby, Mem. N. Y. Bot. Gard. 7: 235. 1927.
In the original description Moore compared his new species with Cocco-
loba laxiflora Lindau, which I now regard as the same as C. ramosissima
Weddell. There is a striking similarity between these two species in the
delicate inflorescences and the long peduncles. There are differences in
the leaf size which should be re-examined when additional materials be-
come available from southern Brazil. At present I distinguish C. longi-
pes by the ovate-oblong leaf blades which taper from the middle to
a blunt apex. Coccolobis padifolia Rusby was described without any
comparison of other species. A study of the type collection indicates that
it should be placed in the synonymy of C. longipes.
Bolivia. Rurrenabaque, Rusby, Mulford Ex. No. 848 (NY-holotype of C.
padifolia, GH). Brazil. Matto Grosso. S. Moore 577 (BM-holotype. B, NY).
Coccoloba lucidula Bentham in Hooker, London Jour. Bot. 4: 627.
1845.
Coccoloba sagotii Lindau, Bot. Jahrb. 13: 184. 1890.
The two species treated here fall into widely separated portions of
Lindau's monographic treatment of Coccoloba, yet it seems clear to the
writer that they are identical. Coccoloba lucidula was based on flowering


JO)LRNAL OF THE ARNOLD ARBORETUM
specimens with delicate, membranaceous, immature leaves which crinkled
in drying, turned black and became lustrous on the upper surface. Ben-
tham cited only one specimen collected by Schomburgk. "2nd Coll. 947
(1262)." The species has not been collected again and no modern col-
lections have been assigned to it. The mature foliage and fruiting speci-
mens of C. sauiotii would appear to belong in synonymv here, but since
at present there is no comparable material for C. lucidula. there is need
for additional mature specimens of the latter.
Coccoloba lucidula is described as a woody vine by collectors of the
specimens cited below. P'rrrottct 1S'20 (P) is a delicate vine tapering to
a tenuous apex. On this specimen the immature condition of the leaves is
clearly shown, from minute to fully expanded, though membranaceous,
forms. Many of the mature leaves of other collections are :olded, indi-
cating that when fresh the midrib is sharply curved downward. The fruit
is distinctive, being nearly spherical and smooth. A small stalk is dis-
tinguishable at the base of the fruit and the apex is more or less obtuse,
with very small, imbricate perianth lobes. P'Trrottet 83 from British Guiana
which is referred here was cited by Lindau (loc. cit. 168) as Coccoloba
racemulosa and, thus identified, was an im)portant example of his Guiana-
northern Brazil distribution (loc. cit. 116).
Coccol/oba saiotfii was described by Lindau and was based on an unnum-
bered Sagot collection from "(Guana gallica" near (avenne. This is a fully
matured branch of scrambling halit. The infructescence is old but fruits
have beern preserved. The leaves are coriaceous and shiny above. Lindau
distinguished this species from C. lucidula by the glabrous branchlets
and infructescence rachis, but close examination shows that in all reliable
characteristics the type collections are similar. The pubescence present
on material of ('. s<,otii was overlooked 1b Lindau. Lindau also referred
to specimens in the herbaria at Berlin and Stockholm, but both of these
are merely fragments. The most complete specimen of this collection
is in the Paris herbarium.
British Guiana. C'overden. Persaiud 13i( F, K NY) D)emerara River. Jcnmian
6300 )! K I); Ituni. south of Mackenzie, Cowan .V255 I K, NY ) Roraima. /cliom-
bur;k 417 (1 / 2 i K -lyp collection, U1 ). Without specific location: Perrottet
S., ((; , s.n. (i : S ,c i burnrgk S I , ,. French Guiana. ('AYENNE: .Martiu s.n.
(K . Poitc/:n .. ( K c, LC. Richiard s.n. (P), .5' o ) s. I. Ittype collection of (.
svgotii. / , p, s), Talbot s.. (K); Montagne de Kaw, Cowan ,37 8 (Nx ).
Venezuela. Bolivar. Tumeremo. Ste'v'rmark a -I42 ( F.
Coccoloba marginata Bentham, in Hooker, Lonldon Jour. Bot. 4: 62(o.
1845.
C'occoloba g<uianensis Meisner. Linnaea 21: 264. 184S.
Coccoloba martii Meisner, Fl. Bras. 5(1): 37. 18 5.
Coccoloba martii var. major Meisner, ibid. 38.
Coccoloba martii var. minor Meisner. ibid.
Coccoloba nitida var. cordata Meisner. ibid.
Coccoloba mntida var. rotundata Meisner. ibid.
250
[VOL. XLI


1960] HOWARD. STUDIES IN THE GENUS COCCOLOBA, IX 251
Coccoloba trinitatis Lindau, Bot. Jahrb. 13: 182. 1890.
Coccoloba douradensis Glaziou, Bull. Soc. Bot. Fr. IV. 11(Mem. 3f): 571.
1911 (provisional name with mixed type; see also C. densifrons).
A discussion of this species is given in two earlier papers (Jour. Arnold
Arb. 40: 84-85. 1959, and 41: 45-46. 1960). I have seen additional
material (cited below) which extends the range of this species to Vene-
zuela and possibly to the Brazilian states of Minas Geraes, Goyas, Bahia,
Acre and Santa Catarina. Additional field study is needed to determine
the variations in individual plants as these occur in South America.
While I am following Lindau in considering Coccoloba martii a synonym
of C. marginata, I wish to point out the possibility that C. martii more
properly may be" assigned to C. peltata Schott. Certainly the Salzmann
collections from Bahia previously identified as "C. pendula" or C. nitida
var. cordata are intermediate between material of C. peltata from Rio de
Janeiro and material of C. marginata from the Guianas. At present Cocco-
loba peltata may be represented only by anomalous material and there-
fore the species may be incorrectly interpreted.
Brazil. ACRE: Rio Macauhan, Krukoff 5479 (c, K, M, NY, w). AMAZONAS:
Sao Paulo de Olivenca. Krukoff 9048 (F. K, NY); without specific locality, Ule
9347 (c, K). BAHIA: Chapada do Rio das Femmeas, Carrasco, Ilheos, Riedel
244 (LE, P), Blanchet 3049 (LE); Liitzelburg 516 (M); without specific locality,
Salzmann 475 (p). GOYAS: Chapada do Rio Preto, Liitzelburg 1304 (M);
Patavidado, Macedo 3.859 (K); without specific locality, Burchell 7768 (P).
MINAS GERAES: Caraca, Tavares 316 (M). British Guiana. Kaieteur Plateau,
Maguire & Fanshawe 23316 (A, NY); Kaieteur Savannah, Potaro River, Jen-
mann 831 (K); basin of Kuyuwini River, A.C. Smith 3030 (A); Waini River,
De la Cruz 3712 (GH); without locality, Poiteau 179 (LE). Dutch Guiana.
Paramaribo. Kappler 1620 (p), Wullschlagel 882 (BR); without specific locality,
Hostmann 506 (p), Wullschlagel 992 (BR). French Guiana. Cayenne, Broadway
307 (cGH); without specific locality, Leprieur 187 (A), Lequillon s.n. (p),
Melinon 252 (A). Venezuela. AMAZONAS: Tamatama, Upper Orinoco, Llewelyn
Williams 15233 (F); without specific locality, Gines 5105 (us). BOLIVAR:
Raudal Guaiquinima, Cardona 474 (us), 475 (us), Maguire 33134 (A, NY).
Coccoloba martii Meisner, Fl. Bras. 5(1): 37. 1855.
Coccoloba martii var. major Meisner, ibid. 38. 1855.
Coccoloba martii var. minor Meisner, ibid.
No type had been selected, but the species was considered to consist
of its two varieties. An examination of the material cited led to the con-
clusion that these taxa may be referred to the synonymy of Coccoloba
marginata Bentham. There is a possibility, as was pointed out in the dis-
cussion of C. marginata, that C. martii and C. peltata are the same.
Coccoloba meissneriana (Britton) K. Schum. in Just, Bot. Jahresber.
28(1): 451. 1902.
Uvifera meissneriana Britton in Rusby, Bull. Torrey Club 27: 129. 1900.


JOURNAL OF THE ARNOLD ARBORETUM
This species is known from but two collections from the same area. It
is similar to Coccoloba peruviana and eventually both may be included in
C. obtusifolia (q.v.). At present it can be distinguished by the tomentum
on the young stems and petioles, on the entire lower leaf surface and on
the midrib of the upper leaf surface. The inflorescence is copiously pu-
bescent, as well. The fruits are comparable to those of C. obtusifolia, hav-
ing the achene surrounded by the imbricated lobes of the perianth. The
Rusby collection is from a staminate plant and was in flower in May.
The Bang collection, made in July. is in fruit.
Bolivia. Guanai, Rusby 1918 (Nv-holotype, c, ii), Bang 1595 (A, GH. K, LE,
M, NY).
Coccoloba membranacea Klotzsch, Linnaea 14: 289. 1840.
This species is apparently based on a Luschnath collection from Bahia,
Brazil. The original description is brief, "Arborescens, floribus lutescenti-
viridibus." Lindau (Bot. Jahrb. 13: 165. 1890) regarded the original
epithet as a nomen nudum and referred it to the synonymy of Coccoloba
ilheensis. At that time he cited "Luschnath 42," a specimen of which is
in the Leningrad herbarium. I have seen that sheet, but there is no anno-
tation to indicate that it is the type of C. membranacea. The specimen is
properly referred to C. ilheensis.
Coccoloba microneura Meisner. I)C. Prodr. 14: 163. 1856; Howard.
Jour. Arnold Arb. 41: 42. 1960.
This species has been discussed in the earlier paper cited above and
referred to the synonymy of Coccoloba nitida HBK. The type was Purdie
s.n., collected in the vicinity of Santa Marta, Colombia. Meisner reported
the type specimen to be in the Arnott herbarium, but such a specimen
cannot be found, although there is a specimen in the herbarium of the
Royal Botanic Garden at Kew.
Coccoloba microphylla Morong in Morong & Britton, Enum. PI. Coll.
Parag. 212. 1892; Ann. N. Y. Acad. 7: 213. 1893, not Griseb. 1866.
This species was based on Morong 899, made along the Rio Pilcomayo
in Paraguay. Because the epithet is a later homonym of Coccoloba micro-
phylla Grisebach, Hassler renamed it C. morongii. An examination of
the type collections shows that it should be referred to the synonymy of
C. paraguariensis Lindau.
Coccoloba micropunctata Eyma. Meded. Bot. Mus. Utrecht 4: 1. 1932.
I am unable to accept Eyma's criteria for distinguishing the material
he cited as a species distinct from Coccoloba excelsa (q.v.), and so have
referred his species to synonymy there. The type selected was Stahel 77
from Dutch Guiana.
I VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 253
Coccoloba mollis Casaretto, Nov. Stirp. Bras. 72. 1844.
Coccoloba polystachya Weddell. Ann. Sci. Nat. III. 13: 261. 1850.
Coccoloba paniculata Meisner. Fl. Bras. 5(1): 43. pls. 20, 21. 1855.
Coccoloba polystachya var. mollis Meisner, DC. Prodr. 14: 151. 1856.
Coccoloba polystachya var. glabra Lindau, Bot. Jahrb. 13: 133. 1890.
Coccoloba polystachya var. pubescens Lindau. ibid.
Casaretto cited no collection by number or name of collector in the
original publication, so one must assume that he was referring to his own
collection. Such a specimen, now in the Turino herbarium, was made on
the island of Itaparica, near Bahia, Brazil, and the data on the label
agrees in description and location with that published by Casaretto. The
label on the specimen also stated the number of the collection as 2218
and the catalogue number as 80. Lindau cited "Casaretto 2218" and
"Meisner 80." These are one and the same sheet. This single sheet in
the Turino herbarium, the holotype of this species, is a sterile specimen
in poor condition consisting of two leafless twigs and five detached leaves,
probably coming from an adventitious shoot since one twig is extremely
pubescent. The conspicuous development of the pubescence can be matched
on the collection Liitzelburg 295 in the Munich herbarium. The label
states that this collection came from a tree 6 meters tall; however, the
very large leaves, long internodes and copious pubescence all suggest ab-
normal or adventitious growth. Other flowering collections by Liitzelburg
cited below from the state of Ceara appear to be transitional in pubescence,
size and shape of leaves and length of internodes.
Weddell described Coccoloba polystachya, being unaware of Casaretto's
name. He saw a living specimen which was given a catalogue number,
and he also cited the collection "Martius 1242." No specimens of the
living plant appear to have been preserved, so the Martius collection may
be taken as the type of C. polystachya. The specimens I have seen of
Martius 1242 bear female flowers and a sheet from Leningrad has the
fragment of a sterile shoot, but there is no doubt that the Leningrad collec-
tion is the same as the more pubescent type of C. mollis Casaretto.
In 1855 in the Flora Brasiliensis Meisner published a treatment of the
genus Coccoloba. He accepted C. polystachya Weddell and referred C.
mollis Casaretto to the synonymy of Weddell's species, with a question.
He cited specimens collected by Salzmann, Spruce and Weddell. The illus-
tration given for this species is of a staminate plant and does not repre-
sent the type of either C. mollis or C. polystachya.
At the same time Meisner described a new species, Coccoloba paniculata,
illustrated in two plates by staminate and pistillate plants. He cited an
unnumbered collection by Pohl, the collection Poeppig 2649, and also
Martius 1242, which Weddell had cited in the original description of
C. polystachya. Meisner distinguished between C. polystachya and C.
paniculata by the more abundant pubescence of the former.
In 1856 Meisner repeated for the Prodromus the description of the two
species and, in addition, listed Coccoloba polystachya var. mollis, based


254 JOURNAL OF TIHE ARNOLD ARBORETUM [VOL. XLI
only on C. mollis (asaretto and the collection "Casarctto 80." It is of
interest to note that Meisner cited Martius 1242 under both species, but
indicated (with an exclamation point) only the citation of this collection
as C. polystachya.
In 1890 Lindau recognized Coccoloba polystachya. with two varieties.
Coccoloba polystachya var. glabra Lindau is based on C. paniculata Meis-
ner, while C. polystachya var. pubcsccns Lindau is based on C. mollis
Casaretto. Lindau noted that the varieties were based on the amount of
pubescence and that intermediates were to be found.
Evma (Mleded. Bot. Mus. Utrecht 4: 4. 1932) recognized that the
oldest name was Coccoloba mollis Casaretto and accepted this, including
with it ('. polystachya Weddell and C. polystachya var. pubcscens Lindau.
Eyma did not treat C. paniculata Meisner or C. polystachya var. glabra,
thus implying his acceptance of them. Macbride's treatment of the genus
for the Flora of Peru (I'ubl. Field Mlus. Bot. 13: 460. 1937) appears to
be based on the work of Evma, although no reference is given. I have
seen only a few of the specimens cited by Eyma. In general the specimens
from French and Dutch Guiana have a different aspect in the texture of
the leaves and the color of the pubescence. Moreover, the petioles and
branches of the intlorescence tend to be longer. The plants from this area
may represent a geographic race, or perhaps even a distinct species. Ad-
ditional material is needed for an understanding of the conditions seen
in these plants.
One collection of Krukoff from the Basin of the Rio Solimoes, also,
is difficult to fit into the general pattern of Coccoloba mollis. This col-
lection, Krukoff 8841, has leaves of still different texture and in this case
the branches of the inflorescence are short, resembling those of Coccoloba
dugandiana. The collection is in fruit and the samples opened, all sterile
and hollow, are strongly triangular in outline and section. The lobes of
the perianth are appressed against the apex, rather than coronate, as in
the few fruits seen of typical C'occoloba mollis. At present the collection
does not merit description as a new species.
I have not accepted the glabrous variety created by Lindau. since ad-
ditional field study of this species is needed to understand the variation
in pubescence with the age and habit of the plant. The species seems
clearly dioecious, the pistillate plants appearing to be more pubescent
than the staminate plants. Collections made from the coastal areas are
also more pubescent than those from inland stations in South America.
The shape of the leaf. particularly the base, and the length of the petiole
are extremely variable in the specimens cited. The species is easily recog-
nized, since so few species of Coccoloba have paniculate inflorescences:
however, no existing description is adequate. I sincerely hope that some
botanist in an area where this plant grows can make the necessary study
of variations in C. mollis.
Bolivia. SANTA CRUIZ: Sara, Bosques de Buenavista, Stcinbachi 6563 (A).
Without specific locality: Yungas, Bang 299 (c,, CI. K, LE, M. NY).


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 255
Brazil. ACRE: Rio Macauhan, Krukoff 5550 (A, M, NY); Seringoel Auristella,
Ule 9346 (G, K). AMAZONAS: Ega, Poeppig 2649 (B, LE); Sao Paulo de Oli-
venca, Palmares, Krukoff 8314, 8337 (A, BR, LE, NY), 8841 (A, BR, NY) ; Humayta
near Tres Casas, Krukoff 5550 (A, M, NY). BAHIA: Camapuan, Riedel 628 (A,
LE); Itaparica, Casaretto s.n. (To-holotype); Sao Bento das Lages Liitzelburg
295 (M). CEARA: Barxa d'Anta, Liitzelburg 26278 (M); Grangeiro, Liitzelburg
25800 (M, w), 25838 (M); Soure, Drouet 2377 (GH); without specific locality,
Gardner 1828 (NY). GOYAZ: Rio dos Alnas, Glaziou 21980 (A, LE); Tocanti-
nopolis, Pires & Black 1650a (us); without specific locality, Burchell 7351-2
(GH). MARANHAO: Loreto, Snethlage 656 (r); Maracassume River, Froes 1811
(A, NY). MATTO GROSSO: Cuyaba, Martius 1242 (M-holotype of C. polystachya,
BR, LE, NY). MINAS GERAES: Paracatu, Riedel s.n. (LE). PARA: Barra do Rio
Negro, Spruce s.n. (Oct. 1850) (B, GH, LE, M, NY); Cassipa, Tapajos River re-
gion, Krukoff 1246 (A, NY). PERNAMBUCO: Tapera, Pickel 2483 (GH). PIAUHY:
Urussuhy, Snethlage 633 (F). Rio DE JANEIRO: Without specific locality, Burchell
5912 (GH). Locality not specified: Pohl s.n. (BR, M, NY). Dutch Guiana. Bradi-
lifi, Matoela, Stahel 189 (A); Zanderij I., Herb. Surinam 189 (NY), 4903 (A,
NY). Ecuador, MANABI: El Recreo, Balao, Eggers 14497 (A, B, LE, M), 15675
(GH, K, LE, M, NY). French Guiana. Godebert, Wachenheim s.n. (A); without
locality, Melinon 106 (A). Peru. LORETO: Florida, Rio Putumayo at mouth of
Rio Zubineta, Klug 1991 (A, GH, NY); Rio Santiago, Tessmann 4372 (NY);
Middle Ucayale, Tessmann 3195 (NY).
Coccoloba monoica Ruiz ex Meisner, DC. Prodr. 14: 149. 1856.
Meisner cited this name in synonymy as "Coccoloba monoica fl. peruv.
Ruiz" and recorded seeing a specimen in the Berlin herbarium. Lindau
(Bot. Jahrb. 13: 220. 1890) apparently saw the authentic material, since
he referred the epithet to synonymy under Muhlenbeckia tamnifolia var.
laxiflora Meisner. The type specimen was not located during a brief
search in the Berlin herbarium several years ago.
Coccoloba morongii Hassler, Repert. Sp. Nov. 14: 162. 1915.
This was a new name, provided by Hassler for Coccoloba microphylla
Morong (1893), not C. microphylla Griseb. (1866). The species is to be
placed in the synonymy of C. paraguariensis Lindau.
Coccoloba moritziana Klotzsch ex Meisner, Fl. Bras. 5(1): 28. 1955.
Meisner published this epithet in the synonymy of his new Coccoloba
moritzii var. opaca and reported that he found the manuscript name in
the Berlin herbarium. I have not been able to locate such a specimen,
although Lindau (Bot. Jahrb. 13: 216. 1890) referred the name to the
synonymy of C. ovata.
Coccoloba moritzii Meisner, Fl. Bras. 5(1): 28. 1855.
Coccoloba moritzii var. opaca Meisner, ibid.
Coccoloba moritzii var. lucida Meisner, ibid.
Lindau (Bot. Jahrb. 13: 216. 1890) has referred this species and its


JOURNAL OF THE ARNOLD ARBORETUM
varieties to the synonymy of Coccoloba ovata. The difficulty in typifying
these names will be discussed under C. ovata. On the basis of the material
I have seen. I believe Lindau's action to be correct. In ihe original publi-
cation Meisner attributed the name Coccoloba moritzii to Klotzsch and
cited in the synonymy of C. moritzii var. opaca the manuscript name
Coccoloba moritziana which he found in the Berlin herbarium. I have not
been able to find the name published in any of Klotzsch's writings.
Coccoloba mosenii Lindau, Bot. Jahrb. 13: 173. 1890, "Moseni."
Coccoloba fastigiata var. glabrata Meisner, Fl. Bras. 5(1): 34. 1855.
This species represents a climbing plant with the leaves borne on short
lateral branches. The leaf blades have a characteristic shape, oblong-
obovate to nearly lanceolate-obovate. The range of variation in habit,
as well as in shape of leaf cannot be determined from the few specimens
on hand. Additional material is needed for further study. The basis for
the proper assignment of Coccoloba fastigiata var. glabrata to synonymy
here has been discussed under that epithet.
Brazil. S.o PAULO. Santos Lorosocaba. Mosen 3458 (s-lectotype. B, C, ,
Locfgrenn 10432 (M). Without definite locality, Burchell 3844 (p).
Coccoloba nigra Fawcett & Rendle, Jour. Bot. 51: 123. 1913; Fl.
Jamaica 3: 120. 1914; Howard, Jour. Arnold Arb. 38: 106. 1957.
As I have discussed in an earlier paper, Fawcett and Rendle based this
species on a collection annotated "Jamaica." but without the collector's
name or number. The type in the Edinburgh herbarium has been studied,
and it is certainly a fragment of Schomburgk 531, the type of Coccoloba
ovata, from British Guiana. A specimen of the Schomburgk collection is
also in the Edinburgh herbarium and the two sheets match, even to the
lichens on the branches. The name Coccoloba nigra must therefore be
assigned to the synonymy of C. ovata Bentham.
Coccoloba nitida HBK. Nov. Gen. 2: 176. 1818.
Coccoloba microneura Meisner. DC. Prodr. 14: 163. 1856.
In the eighth paper of this series (Jour. Arnold Arb. 41: 41-42. 1960),
I corrected an earlier mistake and correctly defined Coccoloba nitida as
a species currently known only from Colombia. A lectotype (Humboldt
1627) was designated in the Paris herbarium. It was collected at San
Bartholome on the Rio Magdalena.
Coccoloba microncura is clearly the same species and was described
from the Purdie collection, without number, from Santa Marta, Colombia.
Coccoloba nivea Jacquin, Hist. Stirp. Am. 115. pl. 78. 1763.
Several modern writers on South American vegetation have used this
[VOL. XLI


1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 257
epithet without citing specimens which can be identified. Coccoloba nivea
Jacq. is a synonym of C. venosa L. (1759), and specimens from South
America will be cited under that name. Schomburgk referred to Coccoloba
nivea under cultivation in his Flora and Fauna of British Guiana.
Coccoloba novogranatensis Lindau, Bot. Jahrb. 13: 192. 1890; Howard,
Jour. Arnold Arb. 41: 40. 1960.
I have referred this species to the synonymy of Coccoloba coronata
Jacq. (q.v.). In an earlier paper (Jour. Arnold Arb. 40: 85-86. 1959) I
selected as a lectotype one sheet of the collection Triana 978 in the herbar-
ium of the Museum d'Histoire Naturelle, Paris. The Triana collection
was made between Anapoima and Apulo, Prov. Bogota, Colombia.
Coccoloba nutans HBK. Nov. Gen. 2: 175. 1818.
Authentic material of this species indicates that the specific name was
proposed by Kunth. The original description is brief and reflects the in-
adequate and immature condition of the specimens. Ocreae are not
present on the stems and the inflorescence, described as nutant, is im-
mature or abortive. Meisner (DC. Prodr. 14: 155. 1856.) repeated the
original description, with minor changes, and reported, "Species non satis
nota, nec in herb. Kunth, nec in Willdenowiano extans." Lindau attributed
the collection to Bonpland and cited from the Berlin herbarium a speci-
men which consists of one detached leaf and the fragment of an inflores-
cence 2 cm. in length. These fragments were obtained from the Paris
herbarium. I have been able to examine the original collection in Paris
which is, in turn, from the Bonpland herbarium. It currently consists of a
short stem, without ocreae, and is broken at both ends. A single recurved
lateral inflorescence having very immature flower buds is present. Poor
though it is, this collection must be designated as the holotype.
No recent collections have been assigned to this species. Macbride,
who treated the genus for the Flora of Peru, saw no material. However, I
believe that two of the collections he cited under Coccoloba sphaerococca
are more properly placed in the present species. A collection, Killip and
Smith 29027 (NY), made at Yurimaguas on the lower Rio Huallaga, Dept.
Loreto, Peru, consists of the stem and leaves of a woody vine. One de-
tached leaf is comparable to that of the Bonpland collection. Additional
smaller leaves are present, but resemble the leaves of Coccoloba ascendens.
Although Kunth did not record the height or habit of Coccoloba nutans
in the original description, Meisner, Lindau and Macbride have referred
to the plant as a tree. I believe that they are in error and that Coccoloba
nutans is a woody vine usually with coriaceous, elliptical leaves rounded
at the base, but that occasionally on vigorous shoots the oblong-obovate
leaves with subcordate bases are produced. The species is similar to
Coccoloba ascendens, but additional material is required for a proper
understanding of the species.


258 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI
Killip and Smith 27958 and 29027 and that of Llewclvn Williams 3805,
assigned in various herbaria to Coccoloba pcltata Schott and Coccoloba
sphacrocarpa Lindau, appear to be C. nutans.
Coccoloba nymphaeifolia Schenk in Zittel. Handb. Palaeont. 2: 491.
1887; Lindau. Bot. Jahrb. 13: 181. 1890.
Schenk used the name (occoloba nymnphacifolia in comparing fossil
leaf material to living species. At that time a plant called Coccoloba
nymphacifolia was under cultivation in the Leipzig botanical garden. A
single leaf is preserved in the Berlin herbarium and bears the annotation,
"Coccoloba nvymphacifolia de Jonge. H. Lips. Brazil." Lindau correctly
referred this specimen and name to the synonymy of C. pcltata Schott.
The epithet (C. nt ymphacifolia is a nomen nudum.
[ To be concluded