VII.] SUB-REGIONS. 229 



certain amount of faunal peculiarity superadded as well, but this is 

 not sufficiently pronounced to permit of the separation of this 

 tract from the tract lying immediately to the north. We have thus 

 on the continent three strictly defined faunal sub-regions : (i) the 

 pasture-lands already described, constituting the East Central 

 African sub-region, through whose vast expanse there is manifest 

 a strong identity in the character of the animal products, the same 

 or very closely related forms being in many instances found at the 

 extreme points of this sub-region ; (2) the forest-tract, constituting 

 the West African sub-region, whose animal products naturally 

 differ very essentially from those of the last ; and (3) the desert or 

 Saharan sub-region, containing a comparatively limited fauna, 

 which, with almost insensible gradations, merges into the fauna of 

 the Mediterranean tract. To the same division belong in great 

 measure the desert tracts of Arabia, or that portion of the 

 peninsula lying to the south of the tropic of Cancer." 



Although Dr Wallace had previously divided continental 

 Ethiopia into an East African, West African, and South African 

 sub -region, the foregoing arrangement seems, on the whole, 

 preferable. There are, however, considerable reasons for regard- 

 ing Somaliland as a sub-region by itself, and South Arabia should 

 perhaps constitute another. Although the precise determination 

 of such areas does not come within the province of this work, it is 

 most important to notice that the West African or Equatorial 

 forest tract continues right across the continent as far eastwards as 

 the Congo- Nile watershed, that is to say, close up to Wadelai, 

 where all traces of the West African fauna are suddenly lost. On 

 this point Mr O. Thomas 1 writes that "the abruptness with which 

 the change of fauna occurs on the watershed is, considering the 

 insignificant nature of the physical barriers, very remarkable, and 

 almost unequalled in the distribution of the mammals of any part 

 of the world. The reason of the change is, however, clear enough, 

 being not the occurrence of such barriers to migration as moun- 

 tains or rivers, but the abrupt ending of the great West African 

 forest, which, as we know from the travels of Schweinfurth and 

 ethers, extends quite into this region, but abruptly ceases before 

 the slopes of the upper Nile basin are reached." 

 1 Proc. Zool. Sec. 1888, p. 17. 



