347 



herbarium, both bearing the number 18422, the two statements 

 cannot be reconciled, though their joint testimony makes it 

 clear that 1). sativa, as understood by Willdenow, excluded 

 D. sativa, Linn. [2], [3], [5j, and [6] as well as D. sativa, 

 Linn. [4]. Even the explicit exclusion by AVilldenow of D. 

 sativa, Linn. [4], appears to have produced little benefit if we may 

 judge from the action of Alton in 1813. Aiton's D. sativa (Hort. 

 Kew. ed. 2. vol. v. p. 392), though by citation identical with 

 D. sativa, Willd., is in reality the West Indian species, intro- 

 duced by Houston, which was described as D. sativa by Miller in 

 1768. Now Houston's plant is the species figured by Plunder 

 which constitutes D. sativa, Linn. [4], the solitary component of 

 the Linnean species that "Willdenow has been at pains to transfer 



f 



Therefore D. sativa, Ait., 



and D. sativa, Willd., so far from being identical, are 

 mutually exclusive. The citation by Aiton of Rheede's Mu- 

 Kelengu, in place of mitigating, only accentuates this criticism, 

 for Aiton claims that his D. sativa is a native of the West Indies 

 whereas Rheede's plant is an Asiatic species.* The same con- 

 fusion was perpetuated by Sweet (Hort. Suburb, p. 216) in 1818. 

 Sweet's T). aculeata, introduced from India in 1803 and thought 

 by him to be Rheede's Katu-Kelengii (Hort. Malab. vol. vii. t. 37) 

 was almost certainly the Mu-kelengu. His D. sativa on the 

 other hand, which was introduced from the West Indies in 1733, 

 though Sweet imagined it to be Rheede's Mu-kelengu (Hort. 

 Malab. vol. viii. t. 51) was clearly the Houston-Miller plant. 



In 1827 Blume used the name D. sativa (Enurn. PI. Jav. p. 23) 

 for two plants from Java which he regarded as varieties of one 

 species and considered to be conspecific with the Mu-kelengu of 

 Rheede from Malabar. He cites in addition D. sativa, Willd., as 

 defined in 1806, so that nominallv his D. sativa should include 

 the Linnean elements [1], [2], [3], [51, and [6]. He also cites 

 Sprengel's reference of 1825 (Syst. vol. ii. p. 152) which is too 

 general to be a guide. The description of I), sativa which 

 Blume furnishes is incomplete; it gives no account of the tubers 

 or of the floral structure. The account of the disposition, shape 

 and especially the secondary venation of the leaves, renders it 

 clear, however, that the plant before Blume must either have been 

 the Mu-kelengu of Rheede, which he evidently believed it to be, or 

 have been that species included by Linnaeus under D. bulbifera in 



w *rd as D. bulbifera, Linn. [1]. The 

 statement that the. leaves of the Java species are glabrous renders 

 its identity with Mu-kelengu doubtful, since in Mu-kelengu the 

 leaves are almost always more or less pubescent ; the statement 

 that the stem is unarmed strengthens this doubt, since it is rare 

 for Mu-kelengu to be without prickles at least at the base of the 

 petiole. Finally the statement that in the male of the Java 

 species the spikes are fascicled and that in the female the spikes 



* We make this remark notwithstanding the circumstance that one 

 species from Central America, D. cymulosa, Hemsl. (Biol. Centr. Amer. 

 vol. iii. p. 355), was identified in 18P2 with the Mu-Kelengu of Rheede. 

 The particular plant with which D. cymulosa was. then associated (Flor. Brit. 

 Ind. vol. v. p. 291) is one from Burma (Wall. Cat n. 5103 C, F) which in the 

 first place is not the Mu-Kelengu (Wall Cat n. 5103 A, D, E) and further is 

 not D. cyritulosn. 



