ME. BENTHAM ON LOGAJflACE-aE. 



77 



F^ 



be the prevailing number. I have therefore been unable to retain 

 even as a section the genus Rouhamonj adopted with some hesi- 

 tation by Alph. DeCandoUe ; and if Brehmia may yet be main- 

 tamed as distinct, it is not on accoxmt of its supposed tetramerous 

 flowers, for I find almost as often five as four parts ; but because 

 a combination of minor characters gives to the flower so different 

 an appearance, that I have been imwilling to suppress it so long 

 as no second species is discovered to connect it more closely with 

 other Strychni, There are also among Strychni considerable diP- 

 lerences in inflorescence, yet seldom in sufficient accord with 

 other characters to make good sections, I have therefore, in the 

 subjoined enumeration of species, thought it most convenient to 

 commence by separating those of the old world from the Ame- 

 rican ones ; in the former case to adopt DeCandoUe's division into 

 arborescent and climbing species, and to arrange these as nearly . 

 as possible according to the form of the coroUa. Among the 

 American species, inflorescence combined with the form of the 

 corolla appears to afford the best primary characters. 



The Asiatic species present some difficulty in the identification 

 of those already published. The S. nux-vomica, a common tree 

 on the Indian coasts, is iadeed easily recognized, not only by its 

 arborescent stem and corymbose inflorescence, but essentially by 

 the long tube and naked throat of the corolla. It has, however, 

 frequently been confoimded with 8. colubrina, a scandent short- 

 flowered species, and several of the figures usually quoted do not 

 nelp to clear up the confusion. Eheede's Caniram, vol. i. t. 37, 

 represents the leaves as alternate, although he describes them as 

 opposite, and the flowers are very rudely drawn. Wight's plate 

 434 of his ' Icones ' is a very good representation of the foliage 

 and flowers of a luxuriant specimen j but there is a tendril repre- 

 sented on the stem, and the figure is therefore referred to S. 

 colubrina, though quoted by DeCandolle under /S. nux-vomica. 

 The drawing was one of Eoxburgh's, and it is impossible now to 

 say whether the tendril was an error of the artist having mixed 

 ^P specimens of two species, or whether in the individual repre- 

 sented a tendril had been accidentally formed, aa occurs occasion- 

 ^7 in some of the erect American species, or whether again it 

 WM really a somewhat anomafous specimen of one of the long- 

 flowered scandent species, such as 8. avalifolia. Eoxburgh's own 

 plate 4 of his Coromandel plants is an excellent representation of 



-Sf, 



nux'Vomica, 



^_"■^■ 



Blume's figure and description of his S. liguatrina (Rumphia, 



■-Tl 



*^^- 



