LIBRARY OF THE UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN CM r- o> 560.5 BIOLOGY AN EVALUATION OF THE GENUS POTERANTHERA J. J. WURDACK FIELDIANA: BOTANY VOLUME 29, NUMBER 9 Published by CHICAGO NATURAL HISTORY MUSEUM JUNE 10, 1963 ; OGY LIBRARY -v. noil I U&IH BIOLOGY AN EVALUATION OF THE GENUS POTERANTHERA J. J. WURDACK Department of Botany, United States National Museum FIELDIANA: BOTANY VOLUME 29, NUMBER 9 Published by CHICAGO NATURAL HISTORY MUSEUM JUNE 10, 1963 Library of Congress Catalog Card Number: 6S-18911 PRINTED IN THE UNITED STATES OF AMERICA BY CHICAGO NATURAL HISTORY MUSEUM PRESS An Evaluation of the Genus Poteranthera The open grassy swamps of tropical America support an abun- dance of herbaceous annual Melastomataceae, especially of the gen- era Acisanthera, Poteranthera, and Siphanthera. The diminutive habit and small fugaceous flowers have combined to insure both paucity of collections and difficulty of identification. For the present, both Poteranthera and Siphanthera can be treated fairly satisfactorily; the latter genus will be discussed in a future paper. Since several of the Central American taxa to be treated by Dr. Williams in the Flora of Guatemala are involved in the Acisanthera-Poter anther a generic problem the following discussion has been prepared for pub- lication at this time. The three sections of Poteranthera, as treated by Cogniaux, are a generic salmagundi. Sect. Tulasnea was composed of species now referable to Siphanthera; the species of Sect. Onoctonia are all con- gruent within Acisanthera; and the generotype, P. pusilla Bong., seems to be an advanced monotype related to Acisanthera, but hav- ing only five stamens and lacking staminodia. The microlicioid seed characters sharply differentiate Sect. Tulasnea from the other two sections. As a criterion for delimiting the genus Poteranthera, the size of the anther pore has no fundamental significance. The annual species of Acisanthera, with sessile clasping leaves, blunt ventral appendages of the connective, and two-celled ovary, form a closely related group. The differences in stamens are illus- trated in the accompanying plate, prepared by M. L. Solt. Apart from three Brazilian taxa dubiously (if at all) distinct from A. lim- nobios (DC.) Triana, these species may be keyed as follows: Small anther thecae completely sterile. Large anther thecae 2.5-4 times longer than broad, the connective prolongation about as long as the thecae and with a dorsally acute appendage. 5. A. rosulans. Large anther thecae about as wide as long, the connective prolongation only 1/5-1/3 as long as the thecae and with a dorsally rounded appendage. 6. A. genliseoides. Large and small anther thecae both polleniferous. 535 536 FIELDIANA: BOTANY, VOLUME 29 Large and small anther thecae oblong, tapering to a narrower pore. Plant glabrous, the calyx-lobes longer than the hypanthium . . .1. A. bivalvis. Plant glandular-puberulous, the calyx-lobes usually shorter than the hypan- thium 2. A. limnobios. Small (and usually large) anther thecae broadly oval and truncate, the pore equaling the anther width. Large stamen connective prolongation longer than or about equaling the anther thecae, the ventral lobes and thecae horizontal; calyx lobes lan- ceolate-oblong 3. A. nana. Large stamen connective prolongation distinctly shorter than the thecae, the ventral lobes and thecae erect; calyx-lobes lanceolate. A. A. crassipes. 1. Acisanthera bivalvis (Aubl.) Cogn. Fl. Bras. 14, pt. 3: 216. 1885. Melastoma bivalvis Aubl. PI. Guian. 1: 404. 1775. Figure 15, A. British Honduras, northeastern lowland South America to Bahia, Brazil. 2. Acisanthera limnobios (DC.) Triana, Trans. Linn. Soc. Bot. 28: 33. 1871. Microlicia limnobios DC. Prodr. 3: 117. 1828. Fig- ure 15, B. There seems no reason for refusing to accept Triana's synonymi- zation of A. pellucida Wright ex Griseb. under A. limnobios, rather than Cogniaux' maintenance of the West Indian population as a dis- tinct species. Certainly the Cuban material is conspecific with Vene- zuelan and Guianan collections. From the limited southeastern Brazil collections examined, I have been unable to decide whether A. divaricata Cogn., A. inundata (DC.) Triana, and A. fluitans Cogn. are distinct from A. limnobios. There seem to be at. most two species, one having short (up to 1 mm. long) prolongation of the connective from the anther base to the filament insertion and one with prolongation 2-2.5 mm. The agglomeration of species with short prolongations (according to Cogniaux' descrip- tions) includes A. limnobios, A. inundata, and A. fluitans, while the species with long connective is A. divaricata. Generally only one flower per specimen is examinable without dissection; in each of the Gardner collections 2591 and 4615 (K, P, US, W) both long and short prolongations of connectives exist on individual plants, while 4614 (K, W) has prolongations about 1.5 mm. long. Only assem- blage of all duplicates of the type collections, or additional collect- ing coupled with field observations, will resolve the taxonomic im- portance of the connective length. The vegetative distinctions used by Cogniaux will probably prove of no value in specific delimitation. DLT FIG. 15. Large and small stamens, lateral and ventral views, of various Aci- santhera species (all X 15). A, A. bivalvis; B, A. limnobios; C, A. nana; D, A. cras- sipes; E, A. rosulans; F, A. genliseoides. 537 538 FIELDIANA: BOTANY, VOLUME 29 For the time being, it seems best to use the name A. limnobios for all extra-Brazilian material. Among the collections which I have assigned to A. limnobios, Schomburgk 485, Wilson Browne 9, and Wurdack & Monachino 40902 approach A. nana but the pores of the small anthers are distinctly less broad than the thecae. These collections seem worthy of sub- specific distinction, but the transfer will not be made until the holo- types of A. limnobios and A. inundata collected by Martius are ex- amined. SPECIMENS EXAMINED GUATEMALA: Chiquimula: southeast of Conception de las Minas, Steyermark 31159 (F). HONDURAS: Comayagua: El Achote near Siguatepeque, Standley 56122 (F).— Santa Barbara: Llano del Conejo, Molina 3683 (F, US). —El Paraiso: Suyate, Standley 28047 (F, US); Galeras, Standley & Williams 67 (F, US); Giiinope, Standley, Williams, Molina & Pa- dilla 2052 (F, US); El Limonar, Williams 17324 (F); Ojo de Agua, Williams 1731 4a (F); Guinope, Williams & Molina 11513 (F, NY); Galeras, Williams & Molina 12614 (F, NY). — Morazan: San Juan del Rancho, Molina 1737 (F, US), Standley 15157 (F); Las Mesas, Valerio 3704 (F); Las Mesas, Williams, Molina & Williams 23299 (F, US, EAP, 20 other herbaria). PANAMA: Chiriqui: Rio Caldera, Killip 3623 (US). — Panama: Alhajuela, Killip 3220 (US). CUBA: Isle of Pines: Britton, Britton, & Wilson 15620 (NY, US). —Las Villas: Yaguaramas, Ekman 18386 (F, NY, US); Manacas, Leon & Cazanas 6043 (NY). COLOMBIA: Huila: Neiva, Rusby & Pennell 394 (NY), 106 8a (NY) .— Meta: Orocue", Cuatrecasas 4364 (US) . "Plains of La Cruiz," Purdies.n. (K). VENEZUELA: Aragua: Maracay, Pittier 5838 (US), 10111 (US).— Bolivar: Ciudad Piar, Maguire, Wurdack, & Bunting 35904 (NY, US), 36020 (NY); Hato La Vergarena, Wurdack & Guppy 14 (NY, US); base of Cerro Carichana, Wurdack & Monachino 40902 (NY). BRITISH GUIANA: Sand Creek, Rupununi River, Wilson-Browne 9 (NY), 36 (NY); "Roraima," Schomburgk 485 (syntypes of A. inun- data var. pusilla Cogn.; K, NY, W). WURDACK: EVALUATION OF GENUS POTERANTHERA 539 BRAZIL: Rio Branco: Surumu near Serra do Mel, Ule 8249 (K, MG).— Para: Monte Alegre, Lima 53-1421 (IAN), Ducke 16127 (MG); Rio Moju, Black 54-16222 (IAN).— Goias: Couto de Magal- haes, Rio Araguaia, Froes 301 28 A (IAN), 30160 (IAN).— Mato Grosso: Cuyaba, Malme 1570 (US); Lagoinha, Archer & Gehrt 51 (US).— Minas Gerais: Corinto, Mexia 5658 (US). 3. Acisanthera nana Ule, Notizbl. Bot. Gart. Berl. 6: 349. 1915. Figure 15, C. Usually the calyx-lobes are, like those of A. limnobios, narrowly ovate-oblong and obtuse, rather than lanceolate and acute as in A. crassipes and A. rosulans. Hoehne's illustration of Poteranthera pauciflora (Naud.) Triana (Anex. Mem. Butantan 1. Tab. 6A, III. 1922), based on Kuhlmann's Boa Vista (Rio Branco) collections, is surely rather A. nana. The Colombian material cited below varies toward A. limnobios, having thecae of the anthers somewhat larger than in other specimens. SPECIMENS EXAMINED COLOMBIA: Cundinamarca: "Alto del Paramo" northwest of Zipa- guira, S. G. Smith 1379 (US).— Meta: La Serrania, Cuatrecasas 7876 (NY); Villavicencio, Pennell 1624 (NY, US). VENEZUELA: Amazonas: Santa Barbara, Steyermark 58425 (F, NY).— Bolivar: Sta. Elena, Tamayo 2888 (US). BRITISH GUIANA: Frechal, Tate 12 (NY); Wichabai, Rupununi River basin, A. C. Smith 2305 (NY, US). BRAZIL: Rio Branco: Serra Pellada, Ule 7654 (isotypes, F, K, L, US); Igarape" Carana, 6 km. west of Boa Vista, Black 51-13764 (IAN) ; Boa Vista, Kuhlmann 503 (RB). 4. Acisanthera crassipes (Naudin) Wurdack, comb. nov. Onoc- tonia crassipes Naudin, Ann. Sci. Nat. ser. 3, 12: 277, 1. 12, f. 4- 1849. 0. pauciflora Naudin, I.e., 1. 12, f. 3. 0. calcarata Naudin, I.e., 1. 12, f. 2. Poteranthera crassipes (Naudin) Triana, Trans. Linn. Soc. Bot. 28: 33, /. 17c. 1871. P. pauciflora (Naudin) Triana, I.e., /. 17b. P. calcarata (Naudin) Triana, I.e., /. 17a. Acisanthera gracilis Ule, Notizbl. Bot. Gart. Berlin 6: 349. 1915. A. bartlettii Gleason, Pap. Mich. Acad. Sci. 17: 145, t. 19. 1933. Figure 15, D. Acisanthera crassipes is quite variable in size and glandulosity; most of the variability can be seen in Wurdack 213, which was col- 540 FIELDIANA: BOTANY, VOLUME 29 lected in an area of only a few hundred square meters. The smaller plants have few branches and the stem only slightly thickened bas- ally. The development of glandular hairs on the hypanthium and stem varies independently of the plant size, with some specimens nearly glabrous. While Naudin's sketch of Onoctonia calcarata sug- gests a long connective prolongation and completely sterile small anthers like those of A. rosulans, all anthers visible on the three sheets of the Leprieur collection are quite conformable with those of A. crassipes. SPECIMENS EXAMINED BRITISH HONDURAS: Belize: Belize, Bartlett 11260 (holotype of A. bartlettii, NY), Gentle 3420 (NY); Sibun River, Gentle 1793 (NY), Bartlett 11387 (NY).— Stann Creek: All Pines, Schipp 677 (F, NY). —Toledo: Monkey River, Gentle 3784 (F, NY, US). VENEZUELA: Bolivar: Caicara, Sprague 8 (K, US); Hato La Ver- garena, Wurdack & Guppy 181 (NY). — Amazonas: Puerto Aya- cucho, Maguire, Wurdack, & Bunting 36040 (NY), Steyermark 58517 (NY), Wurdack 213 (NY, US); Sabana Huachapana, Maguire, Wur- dack, & Keith 41464 (NY, US).— Without definite locality: "bords de 1'Ore'noque," Chaff anjon 197 (P). BRITISH GUIANA: Rupununi savannahs, J. G. Myers 5517 p.p. (K); Orinduik Falls, Ireng River, Irwin 490 (US); without definite locality, Schomburgk s.n. (holotype of Onoctonia pauciflora, P). FRENCH GUIANA: Cayenne, Santini 2368 (L); Mt. SineYy, Meli- non 23 (holotype of Onoctonia crassipes, P; isotype, L); without defi- nite locality, Leprieur s.n. ann. 1838 (holotype of Onoctonia calcarata, P; isotypes, P). BRAZIL: Rio Branco: Surumu near Serra do Mel, Vie 8248 (syn- types of A. gracilis, F, K, L, US) ; Rio Canta, Black 51-13814 (IAN). -Para: Santar<§m, Spruce 1043 (K, P, RB), Spruce s.n. (F, NY, W). —Amazonas: Manaos, Ule 8939 (K, L, MG). 5. Acisanthera rosulans Huber, Bol. Mus. Paraense 2: 509. 1898. Figure 15, E. Both this and the following species are dubiously distinct from A. crassipes. The large anthers of the Benoist collection conform fairly well with Naudin's drawing for Onoctonia calcarata, but not with those of the Leprieur collections. WURDACK: EVALUATION OF GENUS POTERANTHERA 541 FRENCH GUIANA: "Savane du Rocher, Gourd on ville," Benoist 1501 (P). BRAZIL: Para: Rio Maraca, Guedes 61 8b (holotype, MG); Campo do Cupijo, Cameta, Ducke M. G. 16300 (MG); Rio Camara, Black 50-9998 (IAN).— Maranhao: Carolina, Rio Tocantins, Pires & Black 2189 (US). — Goias: Couto de Maglahaes, Rio Araguaia, Frdes 30159 (IAN). 6. Acisanthera genliseoides (Hoehne) Wurdack, comb. nov. Poteranthera genliseoides Hoehne, Anex. Mem. But. 1: 49, t. 6 A, f. 1. 1922. Figure 15, F. All of the Hoehne collections cited below are syntypes; since one syntype number has not been seen, no lectotype has been chosen. SPECIMENS EXAMINED BRAZIL: Mato Grosso: Casa da Pedra, Chapada, Hoehne 2399 (R), 2100 (R), 2402 (R), 2404 (R), 2405 (US).— Goias: Glaziou 21436 p.p. (RB). Poteranthera pusilla Bong. Mem. Acad. Sc. St-Pe*tersb. ser. 6, 3: 138. 1835. This rarely collected and inconspicuous annual is unlike all spe- cies subsequently placed in the genus. The strong toral constriction of the hypanthium, the 3-celled ovary, the lack of alternisepalous small stamens or staminodia, and the reniform foveolate seeds sug- gest placement in the Tibouchineae near Acisanthera. SPECIMENS EXAMINED BRAZIL: Maranhao: Carolina, Pires & Black 2184 (IAN). — Goias: Serra da Chapada, Riedel s.n. (isotype, K); "entre Colombista et Rio Bananal," Glaziou 21436 (K, RB p.p.). Siphanthera duidae (Gleason) Wurdack, comb. nov. Poteran- thera duidae Gleason, Bull. Torr. Club 58: 415. 1931. The species is known only from the Duida cumbre; two recent collections are Maguire, Cowan, & Wurdack 29516 and 29517. No further nomenclatural adjustment is necessary for the genus Poteran- thera; the only remaining binomial, P. minor Gleason, is a taxonomic synonym of Siphanthera foliosa (Naudin) Wurdack. Publication 969 UNIVERSITY OF ILLINOIS-URBANA