Selbyana 25(2): 182-209. 2005. A MONOGRAPH OF ALLOPLECTUS (GESNERIACEAE) JOHN L. CLARK Department of Botany, MRC-166, U.S. National Museum of Natural History, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012; Current address: Department of Biological Sciences, George Washington University, 2023 G Street, NW, Washington, DC 20052 USA. Email: [email protected] ABSTRACT. Alloplectus (Gesneriaceae-Episcieae), whose range extends from Costa Rica to the Andes of Colombia, Ecuador, and Peru, comprises five species. The plants are obligate epiphytes with tubular non­ resupinate flowers. The generic circumscription of Alloplectus is based on phylogenetic analyses of mor­ phological and molecular data from the ribosomal (nrDNA) internal transcribed spacer region (ITS) and the intergenic (tmH-psbA) chloroplast spacer region (details of the analyses are published separately). Alloplectus has a confusing taxonomic history with more than 150 names attributed to this poorly defined group that until recently had not been phylogenetic ally evaluated with material representing the generic type species, Alloplectus hispidus (Kunth) Mart. Many of the species traditionally recognized as Alloplectus are here transferred to Columnea, Nematanthus, Drymonia, Glossoloma, and Crantzia, of which the latter two are resurrected generic names to accommodate well-supported clades. Replacing the traditional poly­ phyletic concept of Alloplectus with a monophyletic Alloplectus presented here requires 29 new combi­ nations, two lectotypifications, and six neotypifications. A nomenclator of all excluded or uncertain species names attributed to Alloplectus is included with currently accepted names. Key words: Gesneriaceae, taxonomy, Alloplectus, Glossoloma, Crantzia, Columnea, Drymonia INTRODUCTION come a collection of species that lacked unifying synapomorphies. The genus Alloplectus Mart. has been a catch­ The taxonomic treatment here recognizes the all group for taxa that do not fit into other genera genus Alloplectus as a well-supported clade (bs in the tribe Episcieae. The traditionally recog­ = 100% in Clark et al. in press) united by the nized species of Alloplectus constitute a poly­ following unambiguous morphological synapo­ phyletic group of 40 species that requires the morphies: indistinguishable dry secondary leaf transfer of 29 names from Alloplectus into Ne­ venation, glabrous internal corolla surface, and matanthus (1 sp.) and Drymonia (4 spp.) and corollas constricted apically. Other characters of recognition of Glossoloma (21 spp.) and Crant­ Alloplectus that are useful in distinguishing it zia (3 spp.) for two well-supported clades. Thus, from Glossoloma are the subwoody perennial epiphytic habit, non-resupinate flowers, and pro­ the revised Alloplectus is here treated as a genus fusely branched stems. Characters that are help­ of five species. This paper provides a taxonomic ful for differentiating Alloplectus from other revision of Alloplectus based on recent phylo­ closely related genera of Episcieae are listed in genetic analyses of molecular and morphologi­ TABLE 1. cal data (Clark et al. in press). The five accepted species of Alloplectus range Most characters traditionally used to define from Costa Rica to Peru and are most abundant Alloplectus, such as a fleshy bivalved dehiscent in the northern Andes in Colombia. Many of the capsule, a pendent inflorescence of a reduced former species of Alloplectus, which form a pair-flowered cyme, and a haploid chromosome well-supported clade now recognized as the ge­ number of N = 9, are symplesiomorphic char­ nus Glossoloma, are easily differentiated in the acters shared with other genera such as Dry­ field by the presence of resupinate flowers. The monia and Paradrymonia. Other characters pre­ majority of species traditionally recognized in viously used to define Alloplectus, such as a tu­ Alloplectus will be treated in a revision of Glos­ bular or pouched corolla tube, are convergent soloma (Clark unpubl. data). with Nematanthus. To a certain extent, Alloplec­ tus was "defined" by what it lacked. For ex­ PHYLOGENY ample, species with poricidal anther dehiscence were treated as Drymonia, species with translu­ The neotropical Gesneriaceae comprise mem­ cent berries were treated as Corytoplectus, spe­ bers of two subfamilies, Gesnerioideae and part cies with non-translucent berries were treated as of Coronantheroideae (Burtt & Wiehler 1995). Columnea, and species with stolons were treated The subfamily Coronantheroideae contains nine as Alsobia or Episcia. Thus Alloplectus had be- genera and 20 species and extends from Chile 182 CLARK: ALLOPLECTUS MONOGRAPH 183 TABLE 1. Comparison of Alloplectus, Glossoloma, Coiumnea, Drymonia, and Crantzia, Characteristic Alloplectus Glossoloma Columnea Drymonia Crantzia Leaves Isophyllous Isophyllous Anisophy lIous Isophyllous Isophyllous (rarely isophyl- lous) Flower orienta- Non-resupinate Resupinate Non-resupinate Non-resupinate Resupinate or tion non-resupi- nate Habit Obligate epi- Terrestrial Facultative epi- Epiphytic or ter- Epiphytic or phyte (rarely epi- phyte or terres- restrial terrestrial phytic) trial Fruit Fleshy capsule Fleshy capsule Berry (rarely Fleshy capsule Fleshy capsule fleshy capsule) (rarely fleshy berry) Anther dehis- Longitudinal Longitudinal Longitudinal Poricidal and Longitudinal cence longitutinal Corolla pouch Terminal when Terminal when Medial when Medial when Terminal when present present present present present to the South Pacific Islands and Australia (Wieh­ well supported (bs = 100% in Clark et al. in ler 1983), The Gesnerioideae contains five press) by both molecular and morphological tribes, 56 genera, and more than 1800 species data. The molecular data are from the ribosomal and represents nearly half the worldwide diver­ (nrDNA) internal transcribed spacer region sity of the plant family Gesneriaceae (Wiehler (ITS), a noncoding region between 18S and 26S 1983, Burtt & Wiehler 1995). The Episcieae is of the nrDNA and the intergenic spacer, trnH­ a member of the Gesnerioideae and ranks as the psbA of the cpDNA. Details of the ITS analysis most diverse tribe in the family with 22 genera are described by Clark and Zimmer (2003) and and an estimated 784 species, or roughly 21 % by Clark et al. (in press). of all Gesneriaceae. Episcieae is also the least The phylogenetic analysis (Clark et al. in studied, and generic concepts remain poorly de­ press) samples 21 of the 22 traditionally recog­ fined, partly because of a simplistic use of fruit nized genera with 155 species and 16 generic structure by earlier workers in delimiting genera. type species. To date, it is the most extensive A special need exists for more generic level phy­ taxon sampling within Episcieae. Although still logenetic analyses with greater sampling of spe­ greater taxon and character sampling is needed, cies diversity. the analysis is adequate to provide a framework Of the 22 genera in Episcieae, CoZumnea L. for the revision of AZZopZectus. The most recent is the only genus that has been consistently taxonomic treatment of AlZoplectus is well over shown to be monophyletic (with the inclusion 100 years old (Hanstein 1865). Hanstein's pub­ here of CoZumnea dielsii Mansf.) using morpho­ lication focused on the Gesneriaceae at the bo­ logical and molecular data (Smith 1994, Smith tanical garden in Berlin and provided an over­ & Sytsma 1994a, 1994b, 1994c). The lack of view of the family. Alloplectus was only a part monophyly for large genera (i.e., >20 spp.) in of Hanstein's (1865) entire monograph of the Episcieae has been suggested for Drymonia Gesneriaceae known at that time, but he recog­ Mart. (Smith 2000, Clark & Zimmer 2003), Epi­ nized 30 species of AZZopZectus in his treatment. scia Mart. (Smith et al. 1997), AlZoplectus Mart. More recent estimates range from 60 (Skog (Clark & Zimmer 2003), Paradrymonia Hanst. 1978 [1979]) to 75 species (Burtt & Wiehler (Smith & Carroll 1997, Smith 2000, Clark & 1995). Zimmer 2003), Codonanthe (Mart.) Hanst. Various sister taxa relationships for AlZoplec­ (Clark & Zimmer 2003), and Nematanthus tus have been proposed in the literature (Smith Schrad. (Clark & Zimmer 2003). NautiZocaZyx & Sytsma 1994a, 1994b, 1994c; Smith & Car­ Hanst. has 70-80 species (Burtt & Wiehler roll 1997; Smith 2000; Zimmer et al. 2002). 1995) and is one of the few large (i.e., >20 spp.) Only one publication, however, included the episcioid genera that has not been tested ade­ type species, AZZoplectus hispidus (Clark & Zim­ quately. mer 2003). Previous analyses were focused on The recognition of AZZopZectus is based on a resolving tribal-level phylogeny of the family total evidence analysis of two molecular datasets (FIGURE 1). The plastid markers such as ndhF, for 155 species and .a morphological matrix of trnL-F, and trnE-T used in previous analyses 120 species and 99 characters. AlZoplectus is were appropriate for addressing tribal level phy- 184 SELBYANA Volume 25(2) 2005 - =80-100% Bootstrap values only one previous analysis (Clark & Zimmer = 60-80% Bootstrap values 2003) included a species that represents the cur­ = <60% Bootstrap values rent clade-based concept of Allopleetus. Previ­ ous studies (FIGURE lA-E) can be summarized "..... Glossoloma I"'" Glossoloma r<]ColUmnta as follows. FIGURE lA. Drymonia and Glossoloma (as :, .. ~. <] ! .. ~ /I "Allopleetus") form a clade that is sister to Col­ .• :. , Drymonia Drymonia ~ :"""'-J Columnta .. ,L<] umnea. Sequence data from the chloroplast gene ndhF (Smith & Carroll 1997,