A synopsis of the New World species of Drypetes section Drypetes (Putranjivaceae) with asymmetrical fruits, including description of a new species

Abstract A synopsis of the New World species of Drypetes (Putranjivaceae) with asymmetrical drupes is presented. The group consists of three species: Drypetes alba, with two varieties,from the West Indies, Drypetes gentryi from Mexico, and the newly described Drypetes asymmetricarpafrom Costa Rica. The new species can be distinguished from both its relatives by its longer fruiting pedicels. In addition, the new species differs from Drypetes alba by its larger fruits, and from Drypetes gentryi by having shorter staminate pedicels and stigmas borne on styles (rather than sessile). Lectotypes are designated for Drypetes albavar. latifoliaand Drypetes incurva.


Introduction
Th e genus Drypetes Vahl (Putranjivaceae) contains about 220 species of dioecious trees and shrubs, mostly of the Old World tropics. About 17 known species are found in the Americas, with the greatest diversity in the West Indies. However the Amazonian species are poorly studied and further research undoubtedly will yield many new species. In the classifi cation of Pax and Hoff mann (1922), the most recent comprehensive treatment available for the genus, all but three of the American species belong to the pantropical section Drypetes based on their pistil consisting of a single carpel [although Pax and Hoff mann named this section Hemicyclia (Wight & Arn.) Pax & K. Hoff m., it contains D. glauca Vahl, the type of the genus, and therefore must be called Drypetes , as was pointed out by Airy Shaw (1969)]. Th e other three species have 2-carpellate pistils and belong to section Oligandrae Pax & K. Hoff m. with 3-4(-7) stamens [ D. laterifl ora (Sw.) Krug & Urb.] or section Sphragidia (Th waites)  stamens ( D. brownii Standl. and D. guatemalensis Lundell); both of these sections also are pantropical. It is worth noting Drypetes has not been examined phylogenetically and the classifi cation by Pax and Hoff mann (1922) may not refl ect evolutionary relationships (Levin 1986).
Although no formal groups below the sectional level have been recognized among New World members of section Drypetes , there is a distinctive group of species with strongly asymmetrical drupes. In these species, the young ovary is symmetrical, as in other members of section Drypetes , but as the fruit develops the ovary grows faster on one side than the other, resulting in an oblique fruit apex with the stigma shifted to one side (Fig. 1B). As many as four species commonly have been recognized in this group. Here these are reduced to two species, one with two varieties, and a new species is described. Th e south Asian species Drypetes gardneri (Th waites) Pax & K. Hoff m., D. lanceolata (Th waites) Pax & K. Hoff m., and D. venusta (Wight) Pax & K. Hoff m. have somewhat similar fruits but diff er signifi cantly from the American species in foliar and fl oral characters and probably are not closely related.

Methods
Th e study was based on the examination of specimens from the following herbaria: A, ARIZ, BM, CAS, CM, CR, DAV, F, G, G-DC, GH, GOET, HAJB, ILLS, K, LL, MEXU, MICH, MO, NY, P, SD, TEX, U, UC, UCR, and US. All cited specimens were seen by the author unless otherwise indicated. Identifi cations were made by comparison with the original descriptions and, when available, with the type material or photographs of types. Details of the fl owers and fruits were examined under a stereoscopic microscope. All descriptions and data on fl owering times, habitats, and distribution are based on the herbarium material examined. Countries in the Selected Specimens sections are listed alphabetically.

Taxonomic treatment
Key to the New World asymmetrical-fruited Drypetes Ecology. Forests on limestone and schist, from sea level to 1100 m.
Phenology. Flowering October to March. Fruiting December to July. Conservation status. Least concern. Drypetes alba var. latifolia is widespread and probably secure in Cuba and Jamaica. Its status in the Lesser Antilles is diffi cult to assess because of a paucity of specimens.
Discussion. Th e plants I include in Drypetes alba have been segregated into species or varieties in various ways since the middle of the 19 th century (Grisebach 1857;1865;Howard 1989;León and Alain 1953;Müller 1863;1866;Pax and Hoff mann 1922). Characters that have been used include those of the leaves (petiole length and blade color, shape, and degree of marginal serration), staminate fl owers (size, pedicel length, and stamen exsertion), and pistillate fl owers/fruits (pedicel length relative to fruit length, style presence/absence, and fruit size). Most of the distinctions were based on observations of the one or two specimens available to earlier workers and these disappear when more specimens are examined. For example, Grisebach (1857), Müller (1866), and León and Alain (1953) distinguished D. alba (equivalent to var. alba is this treatment) from D. glomerata or its synonym D. serrata (here synonyms of var. latifolia ) on the basis of the former having staminate fl owers that are more than 1 mm long borne on longer pedicels and with exserted stamens in contrast to the latter having staminate fl owers that are about 1 mm long borne on short pedicels and with included stamens. Th ese diff erences appear to be an artifact of fl ower age: specimens with immature fl owers, generally with the anthers indehiscent, were called D. glomerata or D. serrata , whereas those bearing fully mature fl owers with dehiscent anthers were called D. alba . Pax and Hoff mann (1922) separated the same taxa using petiole length: 5-8 mm for D. alba vs. about 1 cm for D. serrulata , the superfl uous name they used for what Grisebach and Müller called D. glomerata . In their concept, D. serrulata is restricted to the Lesser Antilles whereas D. alba is found throughout the Greater Antilles. Measurement of specimens shows that plants from the Lesser Antilles have slightly longer petioles (7-12 mm vs. 5-10 mm) than those from farther west, but clearly the variation is great and broadly overlapping. As I treat them, the two varieties have completely overlapping petiole lengths (6-10 mm long for var. alba vs. 5-12 mm for var. latifolia ). Howard (1989) reported that D. serrata , which he considered to be restricted to the Lesser Antilles, had larger fruits than D. alba of the Greater Antilles, but he did not provide comparative measurements. Although mature fruits of D. alba are rarely found on herbarium specimens, those I have seen are about 12-13 mm long throughout its range. Diff erences in leaf color, shape, and degree of marginal serration, alone or in combination, were used by Grisebach (1865) and Müller (1866) to describe new varieties based on single specimens, but these characters have been ignored by subsequent authors, presumably because they found, as I have, that these characters vary considerably even within individuals and certainly do not show consistent patterns.
Th e only character that seems consistently to diff erentiate Drypetes alba var. alba from var. latifolia is the presence of a style about 1 mm long in the former and its absence in the latter, the stigma being sessile. Th is character was fi rst observed by Müller (1863) when he described D. incurva having a sessile stigma; he later noted the same condition in D. glomerata (Müller, 1866). I have found that all specimens from an individual island show the same condition and I have seen no intermediate specimens.
Although this character is consistent geographically, in the absence of additional diff erences it seems too minor to support more than a varietal distinction. Th e distribution of the varieties is curious, with var. latifolia found both east and west of var. alba . DNA sequence data might elucidate this interesting distribution and clarify the evolutionary history of D. alba. Grisebach (1865) based Drypetes alba var. latifolia on Wright 1927 . It has long been recognized that this collection, like many of Wright's Drypetes collections, is a mixture of two species, in this case D. alba and D. laterifl ora (Krug and Urban 1892;Pax and Hoff mann 1922). Th e material at GOET includes two sheets, both from the Grisebach Herbarium, and therefore presumably is the original material studied by Grisebach. Th ese have the additional numbers 46 and 47 on the labels. Th e sheet labeled 46 (GOET 7917) consists of staminate and pistillate fl owering branches of D. laterifl ora and the sheet labeled 47 (GOET 3380) consists of a fruiting specimen of D. alba var. latifolia . Grisebach briefl y described the leaves, staminate fl owers, and fruits ("drupa"), thus he must have considered both sheets to be his new variety. In deciding which material best matches the protologue, the staminate fl owers argue for GOET 7917 ( D. laterifl ora ) and the fruits argue for GOET 3380 ( D. alba var. latifolia ). However Grisebach described the leaves as being subentire. Th e leaves of D. laterifl ora on GOET 7917 are completely entire, whereas the leaves of D. alba var. latifolia on GOET 3380 are very shallowly crenulate-serrulate. Th e latter sheet thus better matches the protologue of D. alba var. latifolia and therefore I designate it as the lectotype.
Etymology. Th e specifi c epithet refers to the strongly asymmetrical drupes, which are unique among Central American Drypetes . Distribution. Known only from Costa Rica, where it is found from the north central part of the country to the central west coast. It may also be expected in extreme southern Nicaragua.
Ecology. Forests at elevations from sea level to 750 m. Phenology. Flowering January (possibly longer, but only a single fl owering specimen known). Fruiting March to June (possibly longer as only immature fruits are known).
Conservation status. Probably of Least Concern. Th e range of Drypetes asymmetricarpa spans at least 200 km. Although its habitat is highly fragmented, the species is found in Manuel Antonio National Park and near both Guanacaste and Rincón de la Vieja national parks.
All the pistillate specimens studied have immature fruits, so their full size and color at maturity are unknown. Th e label on Hammel & Trainer 17046 , with fruits 9-11 × 7-8 × 6-7 mm, says "fruits ca. 1/3 full size," but based on my experience with other species I suspect this overestimates the mature size.  Discussion. When Monachino (1948) described Drypetes gentryi , he examined only a single specimen. No other descriptions of the species have been published, so I provide here an expanded description: