Dioscorea cordata

Raz, Lauren, 2016, Untangling the West Indian Dioscoreaceae: New combinations, lectotypification and synonymy, Phytotaxa 258 (1), pp. 26-48 : 31-34

6. Dioscorea cordata (L.) Raz, comb. nov. Rajania cordata L. Sp. Pl. : 1032. 1753. Lectotype (designated by Acevedo-Rodriguez & Strong 2005: 90):— DOMINICAN REPUBLIC (“Hispaniola”). Plate 491 of the Boerhaave set of Plumier drawings in the library of Rijksuniversiteit, Groningen (used to create the composite drawing in Tab. 155 of Plumier 1758; Fig. 1 )

Rajania pleioneura Grisebach (1864: 588) . Lectotype (designated here):— DOMINICA. Without locality, no date, Imray 122 (lectotype K! (♀ )). Knuth (1924, p. 335) cited a staminate sheet of the same collection number, but did not cite the herbarium. There is a second Imray sheet at K, received in 1867 (the same year that Imray 122 was accessioned into the Hooker herbarium): it is a staminate specimen of the same species, but it has no number; an unnumbered staminate sheet is also at GH. In his protologue, Grisebach did not cite a collection number, but since he described staminate inflorescences and samaras, he must have had access to material of both sexes.

Rajania sintenisii Uline in Urban (1902: 281). Type:— PUERTO RICO. Mayaguez: Monte Mesa, 24 Oct. 1884, Sintenis 109 ♂ (holotype B! not destroyed, as reported in Acevedo-Rodríguez and Strong 2005). There is a mixed staminate and pistillate sheet at K. The pistillate element is a syntype of R. cordata var. microcarpa ; the staminate element at K has cymules pedunculate (born on hypopodia, sensu Weberling, 1992), versus cymules subsessile in the specimen at B. It is unlikely that the two staminate specimens were prepared from the same individual, and the K sheet is therefore not considered an isolectotype.

Dioscorea cyclophylla Urban (1909: 4) . Rajania cyclophylla (Urban) Knuth (1917: 218) . Type:— JAMAICA. Tyre near Troy 660 m, Harris 9402 ♀ (holotype B!, isotypes BM!, K!, NY!, UCWI!, US!).

Rajania cordata var. microcarpa Uline ex Knuth (1917: 219) . Lectotype (designated here):—U.S. VIRGIN ISLANDS. St. Thomas: Signal Hill 470 m, September 1880, Eggers 184 (lectotype US! (♂), isolectotypes: FTG! (♀ ♂), GH! (♀ ♂), JE “1”! (♀ ♂), JE “2”! (♀), K! (♀ ♂), M (♀ ♂). The US sheet was annotated by Uline. Remaining former syntypes: CUBA. Pinar del Rio: Sierra de Anafe , 27 December 1911, Wilson & León 11551 ♀ (K!, NY!, US!). The Cuban material is D. psilostachya . PUERTO RICO. Mayaguez: Monte Mesa, 24 Oct. 1884, Sintenis 109 ♀ [GH!, K! (has both ♀ and ♂ elements: only the ♀ element was cited by Knuth, but they are the same taxon), S!, US!]; Sierra de Luquillo : in Monte Yunque, 14 July, 1885. Sintenis 1384 ♂ (GH!, US! (mixed sheet, part D. pilosiuscula , part D. cordata ); U.S. VIRGIN ISLANDS. St. Thomas : without locality, no date, Eggers 279 ♂ (presumably at B, n.v.).

Rajania ovata var. ehrenbergii (Uline ex Knuth) Knuth (1917: 219) . Rajania cordata var. ehrenbergii Uline ex Knuth (1917: 219) . pro syn. Type:— HAITI. Ehrenberg s.n. ♀ (holotype B). I haven’t seen the specimen, however Kunth (1850) provided a richly detailed description (as R. cordata L.), making a determination possible. Knuth (1924) ultimately chose not to recognize infraspecific taxa in R. ovata and sunk his own variety.

Rajania venosa Knuth (1917: 219) . Type:— PUERTO RICO. Luquillo Mountains, July 1902, P. Wilson 163 ♂ (holotype B!, isotypes K!, NY!, US!).

Rajania cordata var. eucordata Knuth (1924: 334) . Illeg. The autonym R. cordata L. var. cordata , here becomes Dioscorea cordata (L.) Raz var. cordata .

Dioscorea spiculoides Matuda (1953: 57) . Type:— MEXICO. Oaxaca, 1841–1843, Liebmann 14561 ♂ (holotype F!, isotype C!).

Notes: Dioscorea cordata is the most widespread and variable of the Dioscorea sect. Rajania species, occurring in both the Lesser and Greater Antilles (with the notable exception of Cuba). It exhibits complex patterns of variation across its range and further studies are needed to interpret this complexity; genetic data in particular are desirable. There is archeological evidence to suggest that this species may have been an important food plant in Puerto Rico as early as 6000 B.C. ( Pagán Jiménez 2011), and it is likely that human activities (migration, trade) have influenced its distribution. The role, if any, of human selection on observed diversity, is unknown.

Across the Antilles, leaf outline in D. cordata varies from rounded to narrowly elliptic to deltate forms, with cordate to hastate basal lobes, however the recognition by Knuth of distinct varieties of R. cordata based on leaf outline seems not to be justified. Field observations of populations of D. cordata in Puerto Rico and Jamaica show plasticity to be the rule, not the exception, and variation is consistent with morphogenetic studies by Burkill (1960), who illustrated how differential expansion of the apex and basal lobes can produce a range of morphologies within an individual species (using D. alata L. and Tamus communis L. as case studies).

Robustness of the flowers and inflorescences also varies within D. cordata , particularly in Puerto Rico, where smaller flowered forms (generally lighter in flower color) are associated with serpentine and karst substrates (see below), and more robust forms occur on volcanic soils in the Luquillo Mountains at the extreme eastern end of the island.

A general description of D. cordata follows: it is intended to complement the information provided in regional treatments ( Adams, 1972; Howard, 1979; Acevedo-Rodríguez & Strong, 2005), and for comparison with D. cordata var. cymulifera , the one variety I am currently recognizing (diagnostic characters given in 6a, below).

Tuber annual, elongate (occasionally branched) to 40 cm or more, generally wider at the distal end (to ca. 8 cm), the epidermis tan, smooth but with wiry roots, evenly distributed, these leaving circular scars <1 mm diameter, starchy parenchyma white, mucilaginous, deeply buried at the end of a filipendulous cylindrical stalk up to 65 cm, the latter arising from the hypocotyl.Stems terete, without exudate.Leaves +/-coriaceous; primary and secondary veins frequently prominent abaxially. Inflorescences of staminate and pistillate plants are often robust and frequently tanniniferous, occassionally pubescent. In the staminate inflorescences, the internodes between the cymules are frequently shortened, but may become elongated; flowers typically 3–8, in subsessile glomerules, or with floral pediments elongated (length varies continuously), sometimes becoming scorpioid; tepals (staminate and pistillate flowers) commonly light brown or yellowish (-greenish or cream), elliptic to ovate, and both whorls spreading (rotate fl.) or sometimes just the outer whorl spreading. The torus is frequently conspicuously swollen; stamens typically inserted at the perifery of the torus, the anthers stout and subsessile; pistillode inconspicuous (three minute slits in center of torus). Pollen ornamentation perforate-reticulate, the murae completely interconnected, no free ends. Samaras frequently tanniniferous, the wing 2–3 cm long. Plants of serpentine and karst substrates have smaller samaras with much lower tannin content.

Plants from Hispañola called Rajania ovata var. ehrenbergii by Knuth have attributes of both D. cordata and D. haitensis . At the base of the stem, cordate leaves prevail, and lance-ovate leaves subtend the inflorescences. Knuth’s (1917, 1924) placement of these plants within R. ovata was based exclusely on leaf outline, but the floral morphology is consistent with that of D. cordata . Pending further study of the Hispaniolan plants, I am not currently recognizing this variety.

The appearance of Dioscorea spiculoides in this list of synonyms is novel. It was described as an endemic species from Oaxaca, Mexico, based on the Liebmann type, but it is likely that the locality was recorded incorrectly on the label (see discussion in Téllez and Geeta, 2007). Liebmann is known to have traveled to Puerto Rico on his way to Mexico, and in fact the type of D. spiculoides is a D. cordata form that occurs on that island. The F sheet of D. spiculoides is the only duplicate cited by Matuda and should be considered the Holotype, not the C sheet, as cited by Téllez and Geeta (2007).

See Dioscorea quinquefolia , below, for discussion of lectotypification of Linnaean names.

Figures

FIGURE 1. Lectotype of Rajania cordata L.—Plate 491 of the “Codex Boerhaavianus”. It is a copy, made for Boerhaave, of a Plumier drawing representing Janraja scandens folius tamni (Plumier 1703: 33), a verion of which appeared subsequently in Plumier (1758: Tab. 155 Fig. 1) as “Raja foliis simplicibis, cordatus”. Courtesy of the library of Rijksuniversiteit, Groningen.