146 THE ARCTOG^EIC REALM. [CHAP. 



already stated, takes Arctogaea as one of the three primary 

 divisions of the globe, proposed to subdivide it as follows, viz.: 



1 . Madagascar. 



2. Africa, south of the tropic of Cancer. 



3. Oriental, South-eastern Asia, and Malayan islands to 



Wallace's line. 



4. Aquilonian, Europe, Asia north of the Himalaya, Africa 



north of the tropic of Cancer, and America north of 

 about 45. 



5. Media-Columbian, America, between about 25 and 45 



north latitude. 



The importance of this division was, firstly, the recognition of 

 the right of Madagascar and the adjacent islands to form a region by 

 themselves ; and, secondly, the separation of the Medio-Columbian 

 region from the rest of North America. And it will be noted that, 

 if we take away that area, the Aquilonian region corresponds to 

 the Holarctic of Newton and Heilprin. A further modification was 

 proposed in 1892 by Dr C. H. Merriam 1 , who gave the name of 

 Sonoran region to the area corresponding approximately with the 

 Medio-Columbian of Dr Blanford, and suggested that the southern 

 portion of the Eastern Holarctic region should form a region by 

 itself; the name of Boreal region being adopted for what remained 

 of the Holarctic after the subtraction of the Sonoran region and 

 a corresponding area in the Eastern Hemisphere. 



Although from many points of view the retention of such 

 well-known terms as Palsearctic and Nearctic would be a great 

 convenience, the close resemblance of the existing mammalian 

 fauna of the whole of northern Arctogaea compels us to adopt 

 the view that the area forms but a single zoological region. For 

 this region the name Holarctic may be retained ; while for the 

 southern portion of the old Nearctic region, the term Sonoran is 

 the most appropriate. In the Eastern hemisphere the whole of 

 that portion of Arctogsea not included in either the Malagasy, 

 Ethiopian, or Oriental regions is provisionally included in the 

 Holarctic, although when our knowledge of distribution is less 



1 Appendix, No. 19. 



