1898,1 INSECTS AXD ARA-CHNIBS FROM SOCOTRA. 377 



of the Cape Colony ', it becomes very common at Durban, and 

 follows the coast-line northwards, occurring at the mouths of the 

 Zambesi, in Mozambique, and at Wasin, Though it seems to be 

 rarely if ever met with in the "bouth African " interior, it passes 

 inland up the Zambesi and is found in Matabelelaud and at Zomba 

 on the Shire; while the British Museum also contains specimens 

 from Nyasaland, Lake Mweru, Tanganyika, the Victoria and Albert 

 Nyanza, Wadelai, the Galla country, Abyssinia, Somahland, and 

 Aden. Like B. ilithyia, it shows some amount of local variation 

 which may perhaps justify the specific separation of certain 

 geographical forms ". 



It appears therefore that, so far as is known, the distribution of 

 the two species (or varietal groups) is fairly distinct, though their 

 respective ranges coincide for a small portion of the South African 

 subregion and to a larger extent in " East Africa," as at Wadelai, 

 in Somaliland, and at Aden. It is further evident that while the 

 distribution of the iUihyia form is continuous from India through- 

 out the "East African" subregion, that of the gotzius form is 

 almost if not quite discontinuous, its area being separated into a 

 western and an eastern division. In the light of these facts it is 

 remarkable that the Socotran form is most closely akin, not to the 

 ilithyia, but to the gotzius type, nearl}^ resembling in fact West 

 African specimens of B. gotzius, from wliich it is separated geo- 

 graphically by the u hole width of Wallace's East African subregion. 

 It is further of interest to note that the Madagascar and Comoro 

 Islands form (B. anvatara Boisd.), though no doubt distinct, is also 

 a modification, not of B. ilithi/ia, but of B. gotzius \ B. ilithyia, 

 being found at such distant points of the "East African" subregion 

 as Senegal, Angola, and Somaliland, as well as in Arabia, India, and 

 Ceylon, might well have been expected to be the form occurring in 

 Socotra ; and the fact that it is here replaced by a form of B. got- 

 zius suggests the possibility that this island, like the South African 

 subregion, and Madagascar with the Mascarene group, contains 

 relics of a more ancient African fauna that has been expelled or 

 excluded from the bulk of the mainland by the great irruption of 

 forms of life which is believed to have taken place from the north- 

 east ^. 



^ It is implied by Mr. Marshall (he. cif. p. 337) that this Southern race of 

 B. gotsius (to which be restricts Wallengren's name acheloia) occurs also in the 

 Western districts of South Africa, wbere, be states, the Cunene River (north 

 of Damaraland) appears to be its northern boundary. 



- Mr. Marshall (loc. cit. pp. 337, 338) recognizes three local races — the 

 Southern (acfieloia Wallgr., of which vulgaris Butl. is the wet-season form), 

 the Western and Central African (gofciits Ilerbst), and the North-eastern 

 (castanea Butl.). Prof. Bayley Balfour's Socotran examples of B. boydi, noted 

 by Butler as Hypanis com Feisth., are rather curiously ranked by Mr. Marshall 

 under var. acheloia Wallgr. 



3 Triinen, South Afr. Butt. vol. i. 1887, p. 2fi7. 



^ Wallace, ' Geographical Distribution, ' 1870, vol. i. p. 218, &e. Prof. 

 Bayley Balfour (Proc. Roy. Instit. vol. x. 1884, p. 2PB) discusses the affinities 

 of Socotran with West and South African plants, and observes that " Sokotra 

 indeed is, with Madagascar, to be regarded as the remains of a greatlv advancetl 



