GEOGRAPHIC DISTRIBUTION 323 



evidence; that is, they have not been invented merely to explain animal 

 distribution. This is particularly true of the changes in Central America 

 or the Isthmus of Panama. These changes could be safely assumed on 

 geological evidence alone. 



Zoogeographers have not hesitated, however, to assume former land 

 connections for which geology gives no support. Geologists have some- 

 times been the authors of such connections but have based them on the 

 facts of modern distribution. North America and Europe have been 

 assumed to be connected through a strip of land taking in Greenland, 

 arching north of the Atlantic, and joining Europe through the Scandi- 

 navian Peninsula and the British Isles. An antarctic land bridge con- 

 necting the tips of South America and Africa with Australia was proposed 

 by the British geologist Hutton to account for the large flightless birds in 

 those areas. This bridge has been adopted by many others since, but 

 it seems unnecessary, for the connection of the southern continents with 

 the northern land mass is adequate to account for the degree of similarity 

 of the animals. A land bridge has even been thro\\'n across the middle 

 of the Atlantic Ocean, from western Africa, say, to Brazil and the West 

 Indies. This bridge has been employed by many students of distribution 

 and is supported even now by reputable zoologists. The trend, how- 

 ever, has been away from extensive land bridges. They may have 

 existed, but some of them seem geologically so improbable that zoogeog- 

 raphers are seeking other explanations for similarities of faunas, or are 

 frankly leaving the facts unexplained rather than postulate the bridges. 



Major Realms. — From the beginnings of zoogeography, many 

 attempts have been made to divide the earth into half a dozen or so 

 major realms which would have significance for all kinds of animals. 

 Birds were first used for such a division, then mammals. For these 

 two vertebrate groups the boundaries of the realms were somewhat 

 similar, and the authors of the schemes believed that other animals 

 would fit into the same divisions. Much of the work of zoogeography 

 has consisted of fitting groups of animals into the realms and modifying 

 the boundaries when necessary. 



It has become increasingly clear, however, that different kinds of 

 animals do not observe the same distributional limits, and that theo- 

 retically they should not do so. Each group must be delimited by a 

 different scheme. New Guinea, with respect to its earthworms, belongs 

 with eastern Asia; but in its other animals it is Australian. The earth- 

 worms of Ceylon, on the contrary, are of Australian types, despite the 

 nearness of the island to Asia. Chile differs from the rest of South 

 America in its mollusks, fresh-water fishes and earthworms, but agrees 

 with other parts of the continent in its birds and mammals. It is true 

 that in highly isolated areas like the Hawaiian Islands, Madagascar, and 



