r 



76 ME. BENTHAM ON LOGAKIACE-S:. .J 



occurs 



the genera with more than two cells in the ovary, usually classed 

 among Isertiece and SameliecB. In Apocynece we have a somewhat 

 similar fruit in Melodinus. In external appearance those of the 

 common species of both genera are described as exactly like 



occur 



cyneous 



corolla forms a decided separation, confirmed by the peculiar 

 foliage, and StrycTinos (including Brehmid) stands more isolated 

 than almost any other Loganiaceous genus. 



The species more or less known, both in the new and the old 

 world, are now numerous. Unfortunately the great similarity of 

 their foliage, the impossibility of preserving their ripe fruits in 

 herbaria, and the difficultv of obtaining even ffood flowerinef spe- 



cmiens 



identification of many of them a difficult task. The arboreous or 

 climbing habit, which has been adopted as one of the first prin- 

 ciples of division, is very difficult to judge of from dried specimens. 

 The presence or absence of the peculiar hooked tendrils of the 

 genus is by no means a safe criterion. They are sometimes so 

 few in some of the most climbing species that herbarium speci- 

 mens are mostly without them, whilst, on the other hand, some 

 of the smaller American kinds are described by all collectors 

 aa erect shrubs, although the specimens bear several cirrhi. 

 There is reason indeed to believe, that some species which 

 are erect and bushy in open situations become more or less 



scandent when growing 



A much 



more marked distinction lies in the shape of the corolla. The 



some tunes 



four or five times the length 



gradu 



The hairs of the lacinise occupy either neai.^ 



surface, or the base only, or form a ring round the throat or a 

 curved line across the middle of the lobe, or again are wanting 

 altogether. But all these distinctions pass so gradually one into 

 the other, that, however different for instance may be the flowers 

 of S. ntiX'Vomica from those of S, rubiyinosa, no good sectional 

 line can be drawn between them. TPhe number of parts of the 

 flower, quaternary or quinary, is barely of specific value ; it has 

 no relation to habit, and both numbers are occasionally to be 

 found on the same species, although it may often occur in two 

 species otherwise very nearly allied that the one or the other may 



I 



: c 



^ - 



-^■ 



-^-^ 



I L- 



'^■'^ 



■^ ^ . ^ 



4 ■ 



■ -:" ■,...: 



