80 DISTRIBUTION OF ANIMALS. [paut i. 



Encyclopaedia Britannica, thinks that as regards tliat class it 

 can hardly claim to be more than a sub-region of the Neotropical. 

 These views are mutually destructive, but it will be shown in 

 the proper place, that on independent grounds the ISTearctic 

 region can very properly be maintained. 



Suhdivisioiis of the Neardic Region. — The sub-regions here 

 depend on the great physical features of the country, and have 

 been in some cases accurately defined by American naturalists. 

 First we have the Californian sub-region, consisting of Cali- 

 fornia and Oregon — a narrow tract between the Sierra Nevada 

 and the Pacific, but characterized by a number of peculiar 

 species and by several genera found nowhere else in the region. 



The second, or Eocky Mountain sub-region, consists of this 

 great mountain range with its plateaus, and the central plains 

 and prairies to about 100" west longitude, but including New 

 Mexico and Texas in the South. 



The third and most important sub-region, which may be 

 termed the Alleghanian, extends eastward to the Atlantic, in- 

 cluding the Mississippi Valley, the Alleghany Mountains, and 

 the Eastern United States. This is an old forest district, and 

 contains most of the characteristic animal types of the region. 



The fourth, or Canadian sub-region, comprises all the northern 

 part of the continent from the great lakes to the Arctic ocean ; 

 a land of pine-forests and barren wastes, characterized by Arctic 

 types and the absence of many of the genera which distin- 

 guish the more southern portions of the region. 



Ohscrvations on the series of Suh-regions. — The twenty-four sub- 

 regions here adopted were arrived at by a careful consideration 

 of the distribution of the more important genera, and of the 

 materials, both zoological and geographical, available for their 

 determination ; and it was not till they were almost finally 

 decided on, that they w^ere found to be equal in number through- 

 out all the regions — four in each. As this uniformity is of great 

 advantage in tabular and diagrammatic presentations of the 

 distribution of the several families, I decided not to disturb 

 it unless very strong reasons should appear for adopting a greater 

 or less number in any particular case. Such however have not 



