114 



THE JOURNAL OF BOTANY 



The SjncatcB would then emerge as the result of suppressing the 

 lateral branches in LaxcB; this transition is reflected in B. bracteo- 

 lata and B. macrocarpa (q. v.). Finally, the Capitatce may have 

 originated from SpicatcB by shortening and ultimate suppression 

 of the rachis ; or, on the other hand, it might be conceived not 



Fig. 1. — B. hrevifiora 



B. laxissima. 



unreasonably that the DivaricatcB may have given rise to the 

 CapitatcB more directly by suppression of all the axes. B. Laur- 

 entii, however, seems to leave little doubt as to which supposition 

 we should incline to adopt. In this species a single abbreviated 

 but well-pronounced main axis is present, and this is significant 

 as indicative of the probable origin of Capitated directly from 

 SpicatcB. 



Thirty-three species are recognized in the present paper — 

 twenty-one from the African continent, seven from the New 

 World, one from Madagascar, and four from the other Mascarene 

 Islands. The five latter and the American species, with one 



