Dioscorea wrightii Uline ex Knuth (1917: 208)

Raz, Lauren, 2016, Untangling the West Indian Dioscoreaceae: New combinations, lectotypification and synonymy, Phytotaxa 258 (1), pp. 26-48 : 43-45

6. Dioscorea wrightii Uline ex Knuth (1917: 208) . Type:— CUBA. Pinar del Río, 29 Sept. 1860 –64, Wright 3255 ♂ ♀ (holotype B!, isotypes GH!, K!)

Rajania herradurensis Knuth (1917: 220) . Type:— CUBA. Pinar del Río: Vicinity of Herradura, 26–30 Aug. 1910, N. L. Britton, E. G. Britton, F.S. Earle, and C. S. Gager 6390 ♂ (holotype B!, isotype NY!). Dioscorea herradurensis (Knuth) Wilson ex Alain in León (1946: 321).

D. lutea auct non Mey in Grisebach (1866: 251); Knuth (1917: 208).

Notes: Endemic to western Cuba.Although Rajania herradurensis was described in the same publication as D. wrightii, Knuth did not include a description of the androecium of the former, suggesting he did not open the flowers. When Wilson transferred the Rajania name to Dioscorea , he too gave only a superficial description with no detail of the flowers.

Grisebach’s (1866) use of D. lutea Mey. was based on examination of Wright 3255, the type specimen of D. wrightii . He also treated “ Dioscorea lutea Mey ” in his Flora of the British West Indies (1864), citing primarily specimens from the Lesser Antilles, however he also included Cuba in the species’ range. Dioscorea polygonoides does not in fact occur in Cuba, but Grisebach did not distinguish the Cuban Wright material from the non-Cuban material (see also D. polygonoides ).

Although a holotype specimen is not explicitly cited in either the 1917 or 1924 works, the B sheet is the only one annotated “Typus” by Knuth.

Dioscorea sect. Chondrocarpa Uline (1897: 84) . Lectotype (designated here):— Dioscorea chondrocarpa Grisebach (1842:34)

Dioscorea chondrocarpa Grisebach ( 1842:34) . Lectotype (designated here):— BRAZIL. [near Rio de Janeiro], without date, Sellow 54 ♂ (lectotype K!). Remaining former syntypes: BRAZIL. Schott 4573 ♀ (specimens not found)

Dioscorea altissima Lamarck (1789: 231) . Dioscorea alata var. altissima (Lamarck) Grisebach (1864: 587) . Lectotype (designated in Pedralli 2004: 12):— MARTINIQUE. Surian 816 (P! 00307100), misapplied in Kunth, 1850; Knuth, 1924; León, 1946; Adams, 1972; Howard, 1979; Acevedo-Rodriguez, 2005. See below.

Rajania cordata Vellozo (1881: 124) . Lectotype (designated here):—(icon) Fl. Flum. Vellozo (1827: Pl. 116). There are no specimen citations in either work; the material upon which the illustration was based was apparently lost in transit from Rio de Janeiro to Paris.

Dioscorea cayennensis var. altissima pro syn in Knuth (1924: 81). Knuth erroneously ascribed this name to Grisebach (1864: 587), where it is not to be found.

Notes: This species, distributed in tropical South America and the Antilles, had traditionally been called Dioscorea altissima Lamarck. Doubts about the name arose when Acevedo- Rodriguez (2005) called into question the identity of the type, suspecting it to be D. cayennensis Lamarck (1789) . The two species were published simultaneously, based on different specimens. Lamarck described D. altissima as having opposite leaves, cylindrical stems and small, campanulate flowers, however none of these characters applies to Dioscorea altissima as it has been understood in modern treatments. In all references dating from Kunth (1850) to the present, “ D. altissima ” has been applied to a plant with prickly, quadrangular stems, aerial tubers (towards the base of the stem), mixed subopposite and alternate leaves, and large, rotate flowers, from the West Indies and lowland tropical South America. Lamarck’s description of D. altissima corresponds perfectly to Surian 816 in the Jussieu herbarium, but it does not represent the neotropical species in question.

The identity of the Surian specimen was a matter of debate some 75 years ago, although there is no published commentary on the issue. I. H. Burkill, the foremost authority on paleotropical Dioscorea in the first half of the 20th century, annotated the specimen as D. cayennensis (without date). Burkill worked extensively with cultivated yams and although he probably would not have been very familiar with D. altissima , he would have had ample experience with D. cayennensis . A second determination is attributed to Aug. Chevalier with the date October 18, 1936, and reads “ Dioscorea altissima Lamk. (non Dioscorea cayennensis Lamk )”, although in front of Dioscorea altissima the word “Not” appears, in an anonymous hand. Chevalier, a contemporary of Burkill, worked in tropical West Africa, where D. cayennensis is native, and he published several articles on cultivated yams from that region, but he is not likely to have been familiar with D. altissima (apart from the specimen citation in Lamarck). The final annotation (undated) reads “ Dioscorea berteroana ” written in one hand, and below it, in a different hand “(Sagot Impr.)”, a reference to Paul Antoine Sagot, who worked extensively in French Guiana (where the type of D. cayennensis was collected). Dioscorea berteroana Kunth is today treated as a synonym of Dioscorea cayennensis subsp. cayennensis ( Govaerts et al. 2015) . Pedralli (2004), though he did not annotate the specimen, did examine it in person, and cited it as the “ holotype ” of D. altissima , without questioning its identity. Acevedo-Rodriguez was the last person to see the specimen before bringing the issue to my attention. He corrected Pedralli’s “ holotype ” to “ lectotype ” (2005), because Lamarck cited multiple elements.

Such is the difficulty of Dioscorea that even specialists’ opinions are divided, but based on the characters cited above, as well as my own field observations of both species in the West Indies, it is clear that Surian 816 is in fact D. cayennensis . This specimen cannot therefore be used to represent Dioscorea altissima .

The other element cited by Lamarck is an illustration by Plumier (1757: Tab. 117 Fig 2 ). The figure depicts a single inflorescence axis with campanulate flowers, a cylindrical stem, and opposite leaves armed at the nodes with a pair of pseudostipular spines. While D. altissima is generally armed at the base of the petioles, D. cayennensis often has petiolar spines. Lamarck, however, did not mention spines at all. Regardless, the drawing is too schematic to resolve the ambiguity and would not be an appropriate element upon which to lectotypify D. altissima .

As the definition of Dioscorea altissima cannot be clearly established from Lamarck’s protologue, nor from the elements used to substantiate his description, selection of a synonym is preferable to conservation. Grisebach’s description of Dioscorea chondrocarpa is complete and accurate, and the type is consistent with modern collections from both the West Indies and South America.

Dioscorea sect. Dematostemon Grisebach (1842: 27) . Lectotype (designated here):— Dioscorea pilosiuscula Sprengel (1825: 152)

Dioscorea pilosiuscula Bertero ex Sprengel (1825: 152) . Helmia pilosiuscula (Bertero ex Sprengel) Kunth (1850: 434) . Type:—HISPANIOLA. Without locality, no date, Bertero s.n. (holotype TO!). Most likely from the Dominican Republic.

Dioscorea cuspidata Balbis ex Kunth (1850:434) . Pro syn.

Dioscorea duchassaingii Knuth (1924: 67) . Type:— GUADELOUPE (?), without date, Duchassaing s.n. (♂) (holotype B!).

Notes: This species is known from the Greater and Lesser Antilles, Central America and lowland tropical South America. Consult Govaerts et al. (2015) for complete synonymy. Staminate specimens from the WI are unknown and in all pistillate specimens the ovaries are immature. It flowers and sets fruit in other parts of its range, suggesting a recent introduction to the WI. Knuth described D. duchassaingii as a distinct species on the basis of differences in leaf outline alone, and in Acevedo & Strong (2012), it is listed as an accepted name, but this is an error that was carried over from Govaerts et al. (2015).

Dioscorea sect. Macrogynodium Uline (1897: 84). Lectotype (designated here):— Dioscorea trifida Linneaus (1782: 427)

Dioscorea trifida Linnaeus (1782: 427) . Type:— SURINAME. Napis, no date, Allamand s.n. (holotype LINN-1184.1!)

Dioscorea triloba Lamarck (1789: 234) . Type:— FRENCH GUIANA. Cayenne, no further information given (holotype: MPU 012349!). The protologue referes to a specimen in the Jussieu herbarium. No sheet has been discovered there, however the MPU specimen bears the annotations “D. triloba- Willdenov” and “ex herbaris musei Paris,” (musoai is my transcription, but it the writing is illegible) suggesting it is the original material consulted by Lamarck. It is thus interpreted here as the holotype.

“ Dioscorea tuberosa ” auct. non Vellozo in Duss (1897: 566).

Dioscorea trifida var. tuberosa (Vell.) Stehlé (1976: 447) . nom. illeg.

Notes: Widely distributed throughout the lowland neotropics. Specimens from the Greater and Lesser Antilles are probably all of cultivated origin (this information is not always explicit on herbarium labels). The species does not appear to be naturalized in the West Indian flora, although it does flower and set fruit in cultivation. It has been recorded in the West Indies from starch grains (Pagan-Jiménez 2011), and may have been brought by earlier migrants from S. America. Consult Govaerts et al. (2015) for complete synonymy.

Figures

FIGURE 2. Lectotype of Rajania quinquefolia L.—Plate 493 of the “Codex Boerhaavianus”. It is a copy, made for Boerhaave, of a Plumier drawing representing Janraja scandens quinquefolia (Plumier 1703: 33), a version of which appeared subsequently in Plumier (1758: Tab. 155 Fig. 2) as “Raja foliis quinis”. Courtesy of the library of Rijksuniversiteit, Groningen.