AMAItYLLIDAClliE. 
123 
it, to be only deeply cleft, and falsifies the generic character 
which calls it Sexpartitum, and it represents the filaments, as 
Salisbury described them, inserted in the bed,. I have had no 
opportunity of ascertaining from fresh specimens to whom the 
error is attributable. 
1. Pentaphylla. Willd. Rheede. Mai. 7. 35. — 2. Tri- 
phylla. Willd. Jacq. ic. t. 627. — 3. Quinqueloba. 
Willd. Ksempf. ic. t. 15. — 4. Braziliensis. Willd. — 
5. Heterophylla. Hort. Beng. — 6. Aculeata. Willd. 
Rheed. Mai. 7. t. 37. — 7. Rubella. Hort. Beng. — 
8. Purpurea. Hort. Beng. — 9. Angustifolia. Spreng. 
— 10. Nummularia. Willd. Rumph. Amb. 5. t. 162. 
— 1 1. Alata. Willd. Rheed. Mai. 7. t. 38. — 12. Bul- 
bifera. Par. Lond. 17. — 13. Crispata. Hort. Beng. — 
14. Altissima. Spreng. Plum. ic. 117. f. 2.- — 15. Co- 
riacea. Spreng. — 16. Angelina. Hort. Beng. — 17. 
Pulchella. ib. — 18. Atropurpurea. ib. — 19. Glo- 
bosa. ib. — 20. Fasciculata. ib. — 21. Glabra, ib. — 
22. Sativa. Willd. Rheed. Mai. 8. 51. — 23. Piperi- 
folia. Willd. Plum. ic. 117. f. 1. — 24. Nepalensis. 
Sweet H. Brit. — 25. Cinnamomifolia. Bot. Mag. 55. 
2825. — 26. Quaternata. Pursh. Walt. fl. c. 246. — 
27. Villosa. Jacq. ic. 626. Quinata. Walt. Panicu- 
lata. Mich. fl. A. 2. 239. — 28. Oppositifolia. Petiv. 
gaz. t. 31. f. 6. — 29. Lucida ? Brown. Prod. 
The tubers of Dioscorea are eaten and called yams. The 
flower is very insignificant. Jacquin’s figure of D. villosa 
makes the filaments bifid, each point bearing a detached cell 
of the anther which is bipartite. I cannot believe such an 
extraordinary feature, if correctly given, to be only a diffe- 
rence of species. A. Dioscorea, with large cordate leaves, at 
the Chelsea Garden, has 13 strong longitudinal veins, and 
the cross veins confluent, a little oblique, branching in vari- 
ous directions, some retroflex. Fifteen species, of which I 
have examined dry specimens, have the cross veins confluent, 
but those of lucida are most correctly parallel, very rarely 
bifid at the point, but in such case continuing their direction 
to the next rib or the margin, and never retroflex or running 
towards another parallel vein. Its fruit not having been 
seen, it is possible that it may belong to a separate, but 
closely allied, genus. 
22. Rajania. — Perianth deeply cleft, half-patent ; contracted 
