Revision of the Thalia geniculata complex (Marantaceae)

Lennart Andersson

Andersson, L. 1981: Revision of the Thalia geniculata complex (Marantaceae). Nord. J. Bot. 1: 48-56.
Copenhagen. ISSN

Within the Thalia geniculata complex, which corresponds to subgen. Arthrothalia Schum., three species have been
recognized in recent taxonomic literature, viz. T. geniculata L., T. trichocalyx Gagnepain, and T. welwitschii Ridley, The
variation pattern within this complex has been analyzed with the aid of pictorialized scatter diagrams and it is concluded that
the entire complex is made up of a single polymorphic species, of which no further subdivision is justified. T. geniculata is
redescribed and full synonymy is given.

L. Andersson, Dept of Systematic Botany, Univ. of Gothenburg, Carl Skottsbergs Gata 22, S-413 19 Gothenburg, Sweden.

Introduction
The Thalia geniculata complex corresponds to Schumann's (1902) subgen. Arthrothalia. It comprises tall
plants with an ample, panicle-like and richly branched synflorescence (Fig. IA) and comparatively long
florescence internodes (inflorescence terminology according to Andersson 1976). Schumann's subgenera
Euthalia and Sarothalia differ in having conspicuously congested synflorescences and shorter florescence
internodes (<5 mm long). The fourth subgenus, Anomothalia, comprises one or two small species with a very
sparsely branched synflorescence. It should be noted that Schumann's subgeneric names are contrary to the
code. T. geniculata is the generic type, as it was the only species cited in the protologue, and the name
Arthrothalia is thus illegitimate and Euthalia (i.e. subgen. Thalia) is misapplied.
In Schumann's treatment (1902), Arthrothalia comprised one species only, T. geniculata, and both American
and African specimens were included. All previously proposed species belonging here and based on American
material were synonymized. Two species, T. caerulea and T. welwitschii, which had been described by Ridley
(1887) and were based on African material, were disregarded by Schumann, possibly unknown to him. Also
Hutchinson and Dalziel (1936) treated the African plants as conspecific with T. geniculata. In 1904,
Gagnepain described T. trichocalyx from South and Middle America and T. dipetala from Africa. The

O NORDIC JOURNAL OF
BOTANY
NORD. J. BOT- 1: 48-56. TROP
C06
latter was never accepted, but the former was treated as a separate species by Jonker-Verhoef and Jonker
(1955, 1957). In 1964, Koechlin reestablished T. welwitschii for the African members of the complex and the
same position was later taken by Hepper (1968). Within this complex, three species have thus been recognized
during the last few decades. Their distinctive features, as stated by Jonker-Verhoef and Jonker, and Hepper,
are compiled in Tab. 1.
Although the distinctive features of T. geniculata, T. trichocalyx and T. welwitschii are clearly stated in the
literature and the descriptions are readily understood, the plants rarely match the descriptions and are
consequently notoriously difficult to identify. A planned treatment of the Marantaceae of Ecuador made it
necessary for me to reexamine the complex and the results are presented here.

Material and methods

This is not intended as a monographic treatment; the material used is a fraction only of the vast collections
available in the herbaria. All American specimens from BM, GB, K, NY and S (abbreviations according to
Holmgren and Keuken 1974) have been examined. In addition a representative selection of African specimens
from K and NY was studied and all types that I could locate were examined. About 160 specimens were
selected for measuring and form the basis of Figs 3 and

48 Nord. J. Bot. 1(1) 1981

Fig. l . Thalia geniculata. — A. Synflorescence. — B. Extensive Thalia stands on grazed, seasonally inundated ground. —
Between Durän and Boliche, Guayas province, SW Ecuador.

5—6. Specimens were selected on the basis of state of preservation, such that it should be possible to observe
as many characters as possible in the same specimen. Only one specimen from each collection was used, viz.
the first suitable one encountered. All distinctive characters discussed in the literature were studied.
Length of florescence internodes were measured on the main florescence and for the purposes of Figs 3 and
5—6 the arithmetic mean of the three proximal internodes was calculated. The first internode is often
markedly longer (20—50%) than the subsequent ones, and the figures are therefore not quite comparable to
those given by Jonker-Verhoef and Jonker (1955, 1957). The

4 Bot-
figures for spathe ("bract") lengths used in the diagrams are arithmetic means from five arbitrarily chosen
spathes from a single specimen.
In order to economize with the often very scanty herbarium material, only a single flower and fruit was
examined from each specimen. In ample material collected from two clones in south-western Ecuador, it was
observed, however, that intraclonal variation is very small. Most observations on flowers were made on
Nord. J. 1(1) 1981
rehydrated herbarium material, but in a few cases pickled material was used (Harling collections from
Ecuador).

Analysis
Leafshape and colour. Leaf shape is very variable in the complex and there is a more or less continuous series
from broadly triangular to sublinear or even linear (Fig. 3). Although skewed, the distribution of length/width
ratios is unimodal. Extremely narrow leaves occur in a number of character combinations and no correlation
with the other characters studied has been observed. Leaf shape is obviously not a good diagnostic character.
The colour of various organs varies considerably, from green to an intense brownish-purple. This variation
is partly modificative in nature and depends on exposure of the plant, but it is in all probability also due to
genetical differences (cf. Shuey 1975). Colour is, however, not correlated with other characters such as length
of rachis internodes and spathes, sepal indumentum or aril shape. It is true that local populations may be
homogeneous in this character, as well as in others, or they may consist of two distinct forms, but this is
probably due to the formation of large clones or pure lines (cf. discussion of T. trichocalyx).

Length of rachis internodes and spathes are to a certain extent correlated characters (Fig. 5). As rachis and
spathes are developmentally closely interdependent this is not very surprising. The distributions of
measurements are shown in Fig. 3. If there were two species represented in the material, one would expect the
distributions to be bimodal. This is not the case, however. The distributions are more or less skewed, but this
is probably due to diagram construction (the results of an exponential function, growth, being plotted against a
linear scale).

Corolla lengths show a bimodal distribution (Fig. 3) due to different means of the American and the African
population. This is further discussed below. When measurements are considered for each continent separately,
distributions turn out to be unimodal. Corolla length is not closely correlated with spathe and internode length
(Fig. 5). Although long-spathed specimens generally tend to have larger flowers, the Argentina— Paraguay
population is characterized by a combination of short (and wide) spathes and large flowers (Fig. 6).

49

50 Nord. J. Bot. 1(1) 1981

 5 mm c

Fig. 2. Thalia geniculata.— A. Flower from an African specimen with one of the petals markedly smaller than the two
other ones and with comparatively large sepals. — B. Flower from an Ecuadorean specimen with subequal petals and
comparatively small sepals. — C. Distal part of a flowering florescence. —Abbreviations: ap staminal appendage, o st
outer staminode, p petal, s sepal, st ca callose staminode, st cu cucullate staminode, th theca. — A drawn from rehydrated
material, B from a pickled flower and C from a photograph. — Vouchers: A Richards 12558 (K), B Harling and
Andersson 14611 (GB), C Harling et al. 15669 (GB).
Aril shape. The variation in aril shape is illustrated in Fig. 4. There is a continuous series from large
conspicuous structures ("horse-shoe-shaped with involute ends") to small stunted ones ("semi-circular"). The
large ones (Fig. 4E) are by far the most common ones and occur in combination with all other character states.
The small aril (Fig. 4A—B) is characteristic of a few local populations from the area around the Caribbean
Sea and the east coast of South America (Fig. 6). Most, but not all, of the plants with a small aril are long-
spathed

African plants are characterized by a large aril with evolute ends (Fig. 4F). Within the African population,
there is virtually no variation at all in this respect.

Nord. J. 1(1) 1981

Taxonomical conclusions
From the above analysis, it can be concluded, that none of the characters used to distinguish between T.
geniculata and T. trichocalyx (Tab. 1) serves such a purpose. Fig. 5 shows conclusively that there are no
constant character combinations that justify the recognition of more than one species in the material examined.
The variation pattern is much more intricate than the taxonomical structure presently used to depict it.
Fig. 6 shows the geographical distribution of character combinations in the Americas. Although local and
regional populations are rarely homogeneous they are characterized by different frequencies of particular
character combinations. The differentiation has not reached a level, however, where taxonomic recognition is
possible.

Thalia trichocalyx. Plants well conforming with the original description of T. trichocalyx are found in
northern South America, in the Antilles, in southern Florida and in Mexico (Fig. 6). They intergrade without
sharp limit into typical T. geniculata and various intermediates are frequent (Fig. 5). Nevertheless, typical
geniculata and trichocalyx populations may coexist in some areas (e.g. on Trinidad, Jonker-Verhoef and
Jonker 1955: 176, N.W. Simmonds in sched.) and in these areas they apparently imitate good sympatric
species. These patterns break down, however, when material from a larger area is taken into consideration.
The local variation pattern can therefore not be taken as evidence for

Fig. 3. Distribution of measurements in four parameters which have been considered to be of taxonomic importance.
Ordinate represent number of observations as specified in the text. Solid part of the bars represents American material,
open part represents African.

52 Nord. J. Bot. 1(1) 1981

Fig. 5. Character combinations in the Thalia geniculata complex. Symbols are chosen such that typical T. geniculata are
represented by a circle and typical T. trichocalyx by a dot with three or four flags. Symbols corresponding to type
specimens are indicated by labelled arrows: I Richard s. n. , 2 Weddell 2117, 3 Melinon 23, 4 Melinon s. n., 5 Gouin s. n.,
syntypes of T. trichocalyx; 6 Welwitsch 6443, 7 Welwitsch 6445, syntypes of T. welwitschii; 8 Welwitsch 6444, holotype
of T. caerulea; 9 Wright 3279, lectotype of T. angustifolia; 10 Kiener s. n. , holotype of T. dipetala.

Tab. 1, Allegedly distinctive characters in the Thalia geniculata complex.

Character State
T. geniculata 1 T. trichocalyx l T. welwitschii 2

Leaf-like bracts Red or purple beneath Green

Florescence internodes Lower intern. up to 1.5 cm, Internodes 1 and 0.4 cm
upper ones ca. 1 cm. respectively
Spathes 3—3.5 cm 1—2 cm 1.5—2 cm,
Sepals White without green centre, White with green centre, glabrous
glabrous with long slender, white Glabrous
Petals hairs (Smaller)
Ca. 11 mm long Semicircular
Horse-shoe-shaped with
involute ends
1. According to Jonker-Verhoef and Jonker 1955, 1957. 2. According to Hepper 1968.

Nord. J. 1(1) 1981

54 Nord. J. Bot. 1(1) 1981
Fig. 6. Geographical distribution of character combinations in the Thalia geniculata complex. Symbols are chosen such
that typical T. geniculata are represented by a circle without flags and typical T. trichocalyx by dots with four flags. The
question mark stands for Weddell 21 17, which is labelled "Brésil central (Salinas)" , but in all morphological characters
agrees with specimens from French Guiana. I suspect it is erroneously labelled.
Bot- 53
taxonomic recognition. The locally distinct forms in all probability represent widely distributed clones or pure
lines, or, perhaps, different ecotypes. The marantaceous flower is for structural reasons obligately cross
pollinated, but the comparatively high proportion of fertilized flowers in isolated clones (own field
observations) indicate that the plants are selfcompatible. In species forming large and richly flowering clones,
this would lead to extensive geitonogamy.

Thalia welwitschii. Hepper (1968) states that T. welwitschii differs from T. geniculata in having glabrous
bracts and calyx. Koechlin (1964, 1965), says more authoritatively that "Cette espéce (i.e. T. welwitschii) a
souvent eté confondue avec une espece Américaine: T. geniculata L. qui s'en distingue en réalité par un
certain nombre de characores floreaux et inflorescentiels". The reestablishment of T. welwitschii was
obviously founded on an insufficient knowledge of the variation in American populations (cf. Fig. 5).
In fact, the African populations have a number of distinguishing characteristics although these are not
mentioned by Hepper. As pointed out by Gagnepain (1904, describing T. dipetala), one of the petals, viz. the
posterior one, is more or less stunted, often less than half as long as the two others (Fig. 2). Furthermore, the
sepals are unusually long (2—3 mm) and the large and conspicuous aril has the ends evolute rather than
involute (Fig. 4F). In American plants there is a certain variation in petal size, from three equal to a markedly
stunted posterior one. The difference in size is rarely so pronounced as in African plants, however. Plants with
unequal petals are most common in the Greater Antilles, where also plants with large sepals are found (Puerto
Rico, several collections). Arils with evolute lobes have not been encountered in American material. As
compared to the American population, the African is very homogeneous (cf. Fig. 5) and in spite of its very
large distribution area it exhibits a very limited set of character combinations. Collections from Zambia may
be recognized by the longer spathes and more markedly unequal petals, while all other collections studied are
virtually identical.
Although it is generally possible to recognize African plants on the characters given above, it is hard to
believe that Thalia has a very long history in Africa. The very limited range of variation found in the African
population strongly suggests that it has originated from the recent introduction of a very limited number of
diaspores. If these came from an aberrant population, subsequent propagation and homozygotization might
produce a variation pattern of the kind present in the African population ("bottle neck effect").
Homozygotization which is the inevitable effect of predominant geitonogamy, would lead to some characters
being more pronounced in the new population than in the ancestral one. A different selection pressure may
contribute to the divergence of the new population, but is not necessary.
Although it is possible from a technical taxonomic point of view to maintain the African population as a
separate taxon, I do not consider it justified biologically. Biologically, the African population seems, as far as
there is evidence, to represent an "over-propagated" local population and is comparable to other such
populations in the Americas. All three species proposed for African specimens are therefore taken into
synonymy.
It is not possible, of course, to state with any certainty when and how Thalia came to Africa. It was
apparently very wide spread at the middle of the 19 th century. Possibly, it was introduced by vessels of the
slave trade, a trade which had begun already at the turn of the 15 th century. Vessels of this trade often returned
from South America and the Caribbean with ballast in the form of sand taken from the vicinity of the harbors.
As T. geniculata is often abundant in wet, disturbed habitats it might easily have been transported with the
ballast. Judging from the general appearance of African plants, the Greater Antilles seem to be a probable
source of origin. No plants completely matching them have been encountered, however.

Thalia angustifolia was based on an extremely narrow-leaved specimen, As pointed out above, there is a
continuous range of variation in leaf shape, and as the original collection shows no other distinctive features
(cf. Fig. 5, arrow numbered 9), the species cannot be maintained.

Thalia geniculata var. pubescens Koern. was said to be characterized by villous spathes. Three specimens (not
seen) were cited in the protologue. They were of very different origin and the variety was obviously
heterogeneous. Villous-spathed specimens occur in all parts of the distribution area (Fig. 6) and in

Nord. J. 1(1) 1981

combination with a wide range of character states, It is not a biologically meaningful taxon and was rejected
earlier by Schumann (1902). Nevertheless, it turns up now and then on herbarium labels.

Thalia divaricata Chapm. There are four Chapman specimens labelled T. divaricata. All conform well with T.
geniculata. Although a type cannot be chosen until the entire extant material of Thalia in American herbaria
has been revised, the specimens examined in this study (the Chapman collections at NY), offer sufficient
evidence for taking T. divaricata into synonymy.

Thalia geniculata f. rheumoides Shuey is a red-petioled form of T. geniculata. To judge from the data
profferred by Shuey (1975), the colour is genetically fixed and is maintained through vigorous vegetative
propagation and probably also through geitonogamy (including crossing between gametes of the same clone).
In my opinion the recognition of such biotypes is not sound taxonomy, even at a very low taxonomic level.

56 Nord. J. Bot. 1(1) 1981

Maranta flexuosa Presl. I have not been able to locate the type of this taxon, which Schumann (1902) treated
as a synonym of T. geniculata. The type was said to come from Guayaquil (W Ecuador) and the description
("rachi flexuosa pubescente ad nodos villosa") is in good accordance with the west Ecuadorean population.

Synonymy and redescription

Thalia genicu18ta Linné

175.3:3. - Marama geniculata (L.) Lamarck 1783:1 193.

Type: Plumier 1755: tab. 58, fig. 1 (lectotype, cf. below).
Thalia erecta Vellozo 1828 (cited from Schumann 19()2:173) Marama flexuosa Presl — Type: Not seen.
Thalia altissima Klotzsch in Schomburgk 1848:917, nom. nud. — Orig. mat.: Schomburgk 1327B, Guyana, outskirts of
Serra Pacaraima (K).
Thalia geniculata var. a pubescens Koernicke 1862:74. — T. geniculata var. "villosa Koern." of Schumann 1902:173.
Types: Karwinsky s.n. (LE), Sellow s.n. (B), Poiteau s.n. (LE); syntypes, not seen, cited from the protologue. Thalia
divaricata Chapman 1860:465. — Orig. mat. : Not indicated, four Chapman specimens at NY.
Thalia angustifolia Wright in Grisebach 1866:256. — Type: Wright 3279 (BM lectotype, chosen here; NY).
Thalia caerulea Ridley 1887:132. — Type: Welwitsch 6444 (BM holotype, only element cited in the protologue).
Thalia welwitschii Ridley 1887:132. — Types: Welwitsch 6443, 6445 (BM syntypes).
Thalia trichocalyx Gagnepain 1904: — Types: Gouin s.n., Leprieur 241, Leschenault s.n., Melinon s.n., Richard s.n.,
Richard 23, Weddell 21 17 (syntypes, all in P; to judge from sketches and notes attached to the sheet, Richard s.n. was
obviously the specimen contributing most substantially to the description and it is consequently chosen as lectotype).
Thalia dipetala Gagnepain — Type: Kiener s.n.
(P holotype, only element cited in the protologue). Thalia geniculata f. rheumoides Shuey 1975: 2 10—212. — Type:
Shuey (USF holotype, not seen).

Rhizomatous herbs, usually 1—3 m tali. Aerial shoots with a few blade-less sheaths and about 2—5 expanded
leaves at base; above the uppermost of these with a single-internode stem 0.5—2.5 m tall, the stem then
branching to form the synflorescence. Basal leaves usually with a petiole proper, cauline leaves ( = main
bracts) usually without; pulvinus 0.3—2.5 (—3.8) cm; blades usually subtriangular to ovate, rarely elliptical,
oblong or sublinear, glabrous, often glaucous beneath, (12—) 19—63 x (1.9—) 4.0—26 cm, length/width ratio
( l .8—) 2.0—6.2 (—10.3). Inflorescence a much branched, panicle-like synflorescence with spreading
branches; florescences before flowering compact and spiciform with overlapping spathes; spathes caducous,
glabrous to densely hirsute or villous, 1 .0—3.4 cm; rachis flexuose with very conspicuous scars, glabrous to
rather densely villous or hirsute, first internode 5—16 mm, subsequent ones gradually shorter. Ovary glabrous
or sparsely pilose; sepals scale-like, hyaline or greenish with a broad scarious margin, glabrous to densely
setose, 0.5—2 mm; petals glabrous or, rarely, sparsely pilose,
6—13 mm; outer staminode petaloid and showy, 15—20

Bot, 10)
x 5—10 mm; callose staminode largely firm and fleshy but with a narrow, reflexed petaloid rim. Fruit
indehiscent, ellipsoid to subglobose, pericarp very thin, papyraceous when dry. Seed ellipsoid to subglobose,
smooth, brownish to greyish or black, often marbled, 5—1() x 3—8 mm; aril rudimentary to conspicuous.

Chromosome numbers. 2n - — 18 (Miege 1960), 2n = 26 (Mahanty 1970).

Distribution. The Americas from S Florida and Central Mexico to the province of Buenos Aires in Argentina
and to S Ecuador. It seems to be absent from the larger part of the Amzaon Basin and the Central Brasilian
Plateau. It is widespread also in tropical Africa (Senegal, Guinea, Sierra Leone, Liberia, Ivory Coast, Ghana,
Dahomey, Nigeria, Gabon, Camerun, Zaire, Central African Republic, Zambia, Zimbabwe and An-
gola), but is probably introduced there.
T. geniculata is often very frequent in swampy, open habitats in regions with a pronouncedly seasonal
climate. It seems to be very rare or completely absent in regions with a humid, non-seasonal climate.

Typification. As no collector is cited, and as no material is preserved at LINN, the original description of T.
geniculata seems to have been drawn from literary sources only. In addition to some herbals of little interest
for the typification, Linné (1753) cites Plumier 1 i.e. Cortusa arundinacea, amplis cannacori foliis. Plumier's
description and the plate are too meagre, however, to have allowed Linné to make the rather detailed
Nord. J. 1981
description. It is probable, therefore, that it was made from Plumier's manuscript, which was not published
until 1755 (by Burmannus), but to which Linné probably had access during his stay in Holland. All
information contained in Linné's description is found in the Burmannus edition, the plate of which is
consequently chosen to typify the species.
Acknowledgements — I am much indebted to Dr N. Slack for checking the English. The study was supported by grants from the Swedish
Natural Science Research Council.

References
Andersson, L. 1976. The synflorescence of the Marantaceae. Organization and descriptive terminology. — Bot. Notiser
129: 39—48.
Chapman, A. W. 1860. Flora of the southern United States. Ed. l . — New York.
Gagnepain, M. F. 1904. Zingibéracées et Marantacées nouvelles de l'herbier du muséum. — Bull. Soc. Bot. France 51:
164-182.
Grisebach, A. 1866. Catalogus plantarum cubensium. — Leipzig.
Hepper, F. N. 1968. Flora of West Tropical Africa, ed. 2, 3(1). — London.
Holmgren, P. K. and Keuken, W. 1974. Index Herbariorum 1. The herbaria of the world, ed. 6. — Regnum veg. 92.

55
Hutchinson, J. and Dalziel, J. M. 1936. Flora of West Tropical Africa. ed. 1, 2. — London.
Jonker-Verhoef, A. M. E. and Jonker, F. P. 1955. Notes on the Marantaceae of Suriname. — Acta Bot. Neerlandica 4:
172-182.
— and Jonker, F. P. 1957. (repr. 1968). Marantaceae. — In: Pulle, A. A. and Lanjouw, J. (eds.), Flora of Suriname 1(2):
149-208. - Leiden.
Koechlin, J. 1964. Scitaminales. — In: Aubreville, A. (ed.), Flore du Gabon 9. — Paris.
— 1965. Scitaminales. — In: Aubréville, A. (ed.), Flore du Cameroun 4. — Paris.
Koernicke, F. 1862. Monographiae Marantearum prodromus Pars altera. — Bull. Soc. Impér. Naturalistes Moscou 35: 1-
147.
Lamarck, M. le Chevalier de 1783. Encyclopedie metodique.
Botanique. — Paris.
Linné, C. von 1753. Species plantarum 2 (ed. 1). — Stockholm.
Mahanty, H. K. 1970. A cytological study of the Zingiberales with special reference to their taxonomy. — Cytologia 35:
Miege, J. 1960. Troisiéme liste de nombres chromosomiques d'especes d' Afrique occidentale. — Ann. Faculté Sci. Univ.
Dakar 5: 75-85.
Plumier, C. 1703. Nova plantarum Americanarum Genera. — Paris.
— 1755. Plantarum Americanarum etc. (ed. J. Burmannus). — Leiden.
Presl, C. B. 1830. Reliquiae Haenkeanae etc. I. — Prague.
Ridley, H. N. 1887. Angolan Scitaminae. - J. Bot. 25: 129-135.
Schomburgk, R. 1848. Reisen in Britisch-Guiana in den Jahren 1840—1844 etc. — Leipzig.
Schumann, K. 1902. Marantaceae. In: Engler A. (ed.), Das Pflanzenreich 4.48. — Leipzig.
Shuey, A. G. 1975. A red-petioled form of Thalia geniculata L. from central Florida. — Rhodora 77: 210—212,
Vellozo, J. M. da Conceigao 1825—1835. Florae Fluminensis etc. — Rio de Janeiro.

Appendix: Specimens studied

America
USA: Florida. Chapman s. n. (NY, 4 sheets), Corell 47694 (NY), Curtiss 2827 (NY), Hitchcock 377 (NY), Small 8839
(NY), Small and Carter 2913 (NY). — Mexico: Jalisco. Barkley et al. 7554 (NY). Nayarit. Mexfa 977 (NY). Guerrero.
Hinton 6596, Hinton et al. 9536 (NY). Veracruz, Gouin s. n. (P), Calderon 1023 (NY), Rovirosa 554 (NY). Chiapas.
Breedlove and Thorne 20919 (NY), Dunn et al. 21940 (NY). — Cuba: Pinar del Rio. Britton et al. 1579, 7026 (NY), Earl
and Nilsson 1569 (NY), Ekman 1 1207 (S), Hermann 3292 (S), Shafer
11053 (NY). Habana. Ekman 1074, 11509 (S), Ekman 13186
(NY), Leon 717 (NY), Shafer 422 (NY). Las Villas. Leon
9161 (NY). Oriente. Ekman 2416, 14973 Jack 6363 (NY). Prov. unknown. Wright 3279 (BM, NY). — Jamaica: Harris
11822 (NY), Proctor 18455 (NY). - Hispaniola: Haiti. Ekman 3369 (S), Leonard 10()69 (NY), Nash 918 (NY), Picarda
1584 (S). Dominican Republic. Liogier and Liogier 18861 (NY), Mareano and Jiménez 4954 (NY). —Puerto Rico:
Britton 2365 (NY), Hioram s. n. (NY), Johnston and Cowgill 308 (NY), Sintenis 5602 (NY), - Trinidad: Aitken 27 (K),
Baker and Simmonds TRIN 15102 (K), TRIN 15103 (K, NY), TRIN 15104 (NY), TRIN 15108 (K), Britton et al. 1145

58 Nord. J. Bot. 1(1) 1981

(NY), Broadway 9107 (BM), Kuntze 1156 (NY), Simmonds 29 (K). Guatemala: Petén. Cook and Martin 71 (NY). Izabal.
Le Doux et al. 2117 (NY). - Belize: Croat 23991 (NY), Gentle 2662 (NY), Lundell 4294 (S), schipp 420 (NY). -
Honduras: Atlåntida. Yuncker et al. 8407 (NY). —Nicaragua: Matagalpa. Atwood 1969 (NY). Granada. Baker 2383 (K,
NY), Lévy 89 (K). Managua. Nichols 25 (NY). Zelaya. Harmon and Fuentes 5101 (NY). — Costa Rica: Guanacaste.
Jiménez 2694 (NY), Williams et al. 26678 (NY). Alajuela. Brenes 22057 (NY).
Panamå: Herrera. Allen 789 (NY). Canal Zone. Cowell 157 (NY). Darien. Allen 958 (NY). — Colombia: Magdalena.
André 178 (NY). Santander. Killip and Smith 14787 (NY). Valle del Cauca. von Sneidern 1134 (NY). Venezuela:
Carabobo. Alston 5673 (NY), Vogl 43 (S). Cojedes. Koyama and Koyama 13665 (NY). Bolivar. Wurdack and Wright 404
(NY). - Guyana: Goodland 980 (NY), Cook 201 (NY), Gigliole 6 (K), Harris TP513 (K), Jenman 5137 (NY), Schomburgk
1327B (K). — Suriname: Lanjouw and Lindeman 1171 (NY), Maguire and Stahel 23593 (NY). — Guiane Franqaise:
Broadway 751 (NY), Cowan 38895 (NY), Leprieur 241 (P), Leschenault s. n. (P), Melinon s. n., 23 (P), Richard s. n. (P),
Sagot 755 (P). — Brazil: Acre. Prance et al. 12021 (NY). Mato Grosso. Lindman A 2821 (S). Cearå. Druouet 2640, 2645
(NY). Paranå. Dusén 18045 (S). Rio Grande do Sul. Malme 902 (S). — Ecuador: Manabf. Jåtiva and Epling 1004 (S).
Guayas. Asplund 5862, 15913 (S), Fagerlind and Wibom 436 (S), Harling and Andersson 14611 (GB), Harling et al.
15669
(GB), Hitchcock 20089 (NY), Holm-Nielsen and Jeppesen 108 (S), Rose and Rose 22470 (NY). - Bolivia: Beni. cardenas
1414 (NY), Werdermann 2331 (S), White 1529 (NY). Santa Cruz. Kuntze s. n. (NY), Steinbach 5213 bis (NY). ProVince
unknown. Kuntze s. n. (NY). Paraguay: Anisitis 2475
(S), Hassler 2750, 3975, 8012, 12250 (NY), Kuntze s. n.
(NY), Morong 555 (NY), Woolston 792 (NY). - Argentina: Jujuy. Fries 1696 (S). Tucumån. Venturi 8171 (S). Formosa.
Kurtz 1870 (NY). Corrientes. Petersen 1691 (NY). Santa Fé. Birabén 250 (NY). Buenos Aires. Cabrera 5813 (NY),
Venturi 55 (S).

Africa
Senegal: Heudelot 665 (K). — Guinea Bissau: Gossweiler s, n. (K). - Sierra Leone: Deighton 3224, 6010 (K), Jaeger 2020
(K), Jordan 977 (K), Morton SL 264 (K), Morton and Jarr SL
2296 (K), Small 260, 460 (K). - Guinea: Baldwin jr. 13309
(K), —Liberia: Baldwin 9161 (K). —Ivory Coast: de Wilde 937
(K). - Ghana: Lloyd Williams 561 (K), Morton GC 25079 (K), de Wit and Morton A 2916 (K). — Nigeria: Ariwaodo in
herb. Tuley 948 (K), Barter 1024 (K), Dalziel 228 (K), Gbile and Daramola FHI 63573 (K), Jones 2857 (K), Latilo FHI
73552 (K), Moloney s. n. (K). —Cameroon: Biholong 308 (K), Latilo and Daramola FHI 28942 (K), Letouzey 1482 (K),
Lowe 3096 (K), Zenker and Staudt 198 (K). — Gabon: Le Testu 1130 (K). — Central African Republic: Schweinfurt
1791
(K), Spinage 218 (K). —Sudan: Andrews 1110 (K), Brown and Brown 1727 (K), Lia 127 (K). - Zaire: Devred 3403
Evrard 1003 (K), Gerard 2316 (NY), Germain 167 (NY), Robyns 3076 (K), Taton 1426 (K), Troupin 1209 (K). -Zambia:
Fanshawe 4414 (K), Kiener s.n. (P), Richards 12558, 22376 (K), Stevenson 267/31 (K), Symoens 9912 (NY), Verboom
1213 (K), Watmough 273 (K). — Angola: Welwitsch 6443, 6444, 6445 (BM).

Nord. J. 1981