INSULAR FLORAS. 467 



branches ; the only instance known in the natural order to 

 which it belongs. As Dr. Balfour remarks, it is one of the 

 most remarkable discoveries of the expedition, not unlike 

 Colletia cruciata (Rhamnaceae), a Chilian shrub, occasionally 

 seen in gardens in this country. Lachnocapsa (Cruciferse) 

 is an endemic, monotypic shrub, having downy cordate 

 pods so like the thick leaves that it is difficult to distinguish 

 them. Dirachma is another extremely interesting endemic 

 monotypic genus, which Balfour refers to the Geraniaceae, 

 with the remark that it appears to be most nearly allied to 

 the South American genera Wendtia and Balbisia. This is 

 one more of the unexplainable isolated remnants of plants 

 of American rather than Asian or African affinities, cor- 

 responding to similiar anomalies in the present distribution 

 of animals. Thamnosma socotrana (Rutaceae) belongs to a 

 genus peculiar to the Mexican region, if we except a South 

 African plant of this affinity still imperfectly known. 

 Ccelocarpus (Verbenaceae) is another new genus, which Bal- 

 four finds most nearly allied to the American Citharexylum. 

 In this connection I may refer to the turneraceous 

 Mathurina pendulifiora (38), an arboreous member of the 

 order inhabiting Rodriguez, whose nearest allies are in- 

 habitants of Central America. Indeed, there is very little 

 to separate Mathurina from the American Erblichia, and 

 Hoffmann (13) describes an Erblichia madagascariensis, 

 which is also arboreous, and evidently congeneric with the 

 Rodriguez plant. 



Historically, the frankincense and myrrh trees of 

 Socotra are of more than ordinary interest. Most 

 of the species — three or four each of Boswetlia and 

 Bahamode?idron (Commiphora) — are apparently endemic, 

 though closely related to members of the same genera 

 inhabiting: Arabia Felix and Somaliland. 



The Leguminosae are numerically the strongest order 

 in the island ; but the only one calling for special remark 

 is Priotropis socotrana, which is closely allied to the only 

 other known species, P. cytisoides, a native of the Eastern 

 Himalaya and Khasia. In Pnnica protopunica we have 

 what Dr. Balfour, probably correctly, regards as the proto- 



33 



