38 Merriam — Geographic Distribution of Life. 



gress of si^ecies from the south. No mountain range or arm of 

 the sea or other tangil)]e obstacle marks the northern boundary 

 of the semi-troi)ical fauna of northeastern INIexico where it 

 ends abruptly near the Nueces River in Texas, or the semi-trop- 

 ical belt of Florida Avhere it ends near Tampa Bay on the west 

 and Cape Malaliar on the east. 



If the Troi)ical fauna and tlora stojDped at the narrow 

 Isthmus of Panama, or cA'en in southern Nicaragua, where 

 the last union of the North and South American continents prob- 

 al)ly took place, the case would Ije very ditferent ; but instead 

 of doing this it pushes northward 1,500-2,0(K) miles and ends al)- 

 ru})tly where the most painstaking search fails to reveal any 

 barrier to further extension except an uncongenial decrease in 

 temperature and humidity (see also remarks under change of 

 climate following Pleist(jceue times p. 44.) 



No more striking illustration could be desired of the potency 

 of climate compared with the inefficiency of physical barriers 

 than is presented l:)y the almost total dissimilarity of the North 

 American Tropical and Sonoran Regions, though in direct con- 

 tact, contrasted with the great similarity of the Boreal Regions 

 of North America and Eurasia — now separated by broad oceans, 

 though formerly united, doubtless, in the region of Bering Sea. 

 Of the thirty-one Boreal genera of North American mammals 

 all luit eight, or three-fourths, occur also in Eurasia, and but a 

 single family is restricted to cold-temperate America. This 

 family (the Aphidontidrr) is the sole representative of a group 

 approaching extinction, and the accident of its survival (in a 

 single genus ancl two closely related species) in a very limited 

 area along our west coast can hardly be construed as of much 

 faunal significance. Contrasted with this one family (which 

 ought not to be counted) and eight genera of Boreal North 

 American mannnals not occurring in Eurasia, Tropical North 

 America (Central America and part of Mexico, exclusive of 

 the West Indies) has no less than eight families and fifty-three 

 genera not l)elonging to the inmiediately adjoining Sonoran 

 Region of the southern United States and the plateau of Mexico. 



Tup: Sonoran not a Transition Region. 



Before leaving this part of the subject reference should Ijc made 

 to the view recentl}" advanced by some naturalists, notably by 



