III.] SUB-REGIONS. 135 



necessary to look elsewhere for the origin of the Neogaeic fauna. 

 That Africa has been the great feeder appears the most probable 

 explanation ; although in the case of the marsupials it seems 

 necessary to look to Notogaea as the point of origin. Clearly, 

 however, the presumed connections between Neogaea, Notogsea, 

 and Africa have not been very continuous or- very extensive 

 in Tertiary times, or the faunas of these areas would have been 

 more alike than they are ; and this suggests that the northern 

 extension of Antarctica has not been so great as has been 

 supposed. Whether the presumed connection between Notogaea 

 and Neogaea has taken place by way of Antarctica or Polynesia 

 may be left an open question. With regard to Africa, the recent 

 researches of Dr Gregory 1 on the West Indian corals, in the 

 course of which it is urged that a shallow-water connection "ex- 

 tended across the Central Atlantic in at latest Miocene times," 

 while the southward extension of the Atlantic is a comparatively 

 recent feature, indicate the possibility that the land-connection 

 which existed in Jurassic times between Brazil and Western 

 Africa may have persisted till the Tertiary era. 



As already mentioned, the Neogaeic realm includes but a 

 single region the Neotropical : and in this four 



,..,,..,, , Sub-regions. 



sub-regions have been defined, and are named as 

 follows. Firstly we have the Brazilian sub-region, which includes 

 not only Brazil, but likewise the Guianas, Venezuela, Colombia, 

 Ecuador, Paraguay, and those portions of Peru and Bolivia lying 

 on the Brazilian side of the Andes, together with the eastern 

 slopes of that portion of the great mountain-chain itself. This is 

 essentially an area of dense tropical forests, locally interspersed 

 with open pastures, or "campos." The second is the Chilian 

 sub-region, comprising Chili, Argentina proper, Uruguay, Pata- 

 gonia, and such portions of Peru and Bolivia as are not included 

 in the preceding. It is chiefly an area of open plains and 

 pampas, although including the high Andes. Thirdly, there is 

 the Mexican sub-region, which embraces the isthmus of Panamk, 

 Central America, and Southern Mexico, and may be regarded to 

 a great extent as a transitional tract between the typical Neo- 



1 Quart. Journ. GcoL Soc. Vol. LI. pp. 306 307 (1895). 



