ME. BENTHAM Olf JjOGA^NIACHM. 



69 



is therefore more exactly iatermediate between Bubiacece 



wcynecd than any other known genus. 



corolla is usually described as " subinfundibuliformis. fauce 



barbata. 



j> 



The tube is, however, so short that it is often almost 



rotate, and the hairs, when they exist, are rather on the upper 



surface 



A. 



Mauritius, but of these the G. parvifiorvm 



lanceolatum 



vanes 



shape of the leaf. On the other hand, some specimens transmitted 

 by Bojer under the name of Q. ohovatum belong to a really distinct 



with 



other species. 



twice 



natum. Wall 



5ve Polynesian species must be deducted G. acumi-' 

 , described from male specimens of a species of 

 U^rophyllum (Axanthes, BL) allied to U, gldbrum ; and Q. hcBmo- 

 ^permmn does not appear to me specifically to diiFer from Torster's 

 original G. rupestre, of which I have seen the specimen in the 

 -British Museum, and which I have also from the Teeiee Islands, 



Bliune 



Expedition 



It is well 



the branches often become glabrous. 



three 



with the leaves of G. liqus 



trifolmm, but differently shaped calyx 



and ft very largi 



'Um 



Tribe III. FAaEJEE^. 



This tribe corresponds with Gardenie^, characterized by an in- 



^etiiscent fleshy fruit with several ovules m each ceU ot the ova- 

 ri^im. The affinities, however, with Buhiace^, except in the case 

 of Fagrma itself, are not so close as in the case of the other 

 tribes. Strychnos is in the tribe the representative of Apocyn€<e, 

 ^d Nicodemia of ScropTiularinecd, but in both instances the con- 

 nexion is rather remote ; nor are all these genera very naturally 

 ^sociated with each other, but I have been unable to discover 

 ^^y better arrangement. Like the other tribes, it may be divided 

 ^<^ording to the aestivation of the corolla ; contorted in the Asiatic 

 ^^r^a, the American Desfontainea and Potalia, and the African 



■^^ 





^■. ..■-..■• --^ 





I , J 



. t 



