;326 FORMER LAND COXNECTIOXS. 



Tlius, in this group of animals, which is widely distributed 

 ,in the warmer parts of the world, we have two striking instances 

 of Afro-American affinity. In both cases, this is greater than the 

 affinity between the African and their nearest Old World allies ; 

 further, the genera of Peripatidae seem to constitute a regular 

 succession of forms, just as we should expect to find on the 

 assumption that the Malay region, Central Africa and Tropical 

 America were once connected up into one land mass. 



Here I should add that, according to A. H. Clark, the 

 Peripatopsidae most probably were distributed through Antarc- 

 tica, which indeed seems very reasonable. 



(3) The limbless amphibians composing the order Gj'mno- 

 phiona include 19 genera distributed through the tropics as 

 follows: — Dermophis, a primitive genus, occurs in South and 

 Central America (5), in San Thom^ Island (1), in British East 

 Africa (1), and in the Seychelles (1). Herpele, a somewhat 

 specialised genus, occurs in Panama (1), in West Africa (3) and 

 1 in India (Assam): its nearest generic ally appears to 'be the 

 South American Gymnopis. Uraeotyphlus, a primitive genus, has 

 one species in W'est Africa, and two in India (Malabar). 



Hypogeophis, a primitive genus, has one species in Zanzibar, 

 and three in ihe Seychelles. 



In addition, South and Central America have seven peculiar 

 genera, (two primitive, two somewhat specialised, and three scale- 

 less and thus degraded) : East Africa has three peculiar genera 

 (two the most degraded of all, and the third also very degraded) : 

 the Seychelles have two peculiar genera (one primitive, and one 

 somewhat specialised) : West Africa has one peculiar but primi- 

 tive genus : the Oriental region, including Malayan Islands and 

 Indo-China, has two peculiar genera (one primitive, and one very 

 degraded). Now all the scaleless genera (7) which are presumably 

 degraded and recent, are strictly localised (three in America, three 

 in Africa, and one in Travancore) : from this it seems a reason- 

 able inference that they at least have actually originated in sihi 

 from the local primitive stock. 



Again, besides the three more or less widely distributed 

 primitive genera above-mentioned, there are five others, appar- 

 ently equally primitive but more localised, one ranging between 

 Ceylon, Malay region and Indo-China, two in South America, 

 one in West Africa and one in the Seychelles. 



I have mentioned these details in order to emphasize the 

 pronounced tendency to differentiate into local genera, not only 

 in the most degraded groups, but also in the most primitive ones. 

 This reduces the probability of such extensive migration as is 

 implied in the theory that Herpele and Dermophis have migrated 

 unchanged from Eurasia to the various localities they now 

 occupy, as Wallace seemed to hold (see " Island Life," p. 404) 

 from the scanty data that were then available. 



Or, admitting the possibility, it still remains unexplained 

 why the primitive Dermophis is absent from that region we 

 should expect it to occupy on the migration hypothesis, namely, 

 S.E. Asia. 



