564 



SCIENCE, 



[N. S. Vol. XI. No. 276. 



ited ia number and kind by hereditary, 

 stirp or germinal influences, and thus re- 

 sult in the independent evolution of sim- 

 ilar types in widely-separated regions under 

 the law of parallelism or homoplasy. 



This law^causes the independent origin 

 not only of similar genera but of similar 

 families and even of our similar orders. 

 K'ature thus repeats herself upon a vast 



above orders, and the Hy stricomorph rodents 

 enjoyed their chief radiation. In Notogaa 

 two orders were cut off by the sea, one of 

 them a rapidly declining type, the Mono- 

 tremes, the other the Marsupials enjoying 

 a very highly diversified radiation. This 

 hypothesis is expressed in Chart IV. Two 

 other orders of mammals, the Sirenia (prob- 

 ably a branch of the hoofed tribe), took the 



Chart II. — Division of the World into three Eealms and nine main Geographical Regions. 

 nental platform is raised to the 200 metre line showing the main Tertiary land connections. 



scale, but the similarity is never complete 

 and exact. When migrations are favored 

 by over- population or geographical changes, 

 a new and severe test of fitness arises by 

 the mingling and competition of the parallel 

 types. 



Now under the operation of these laws a 

 most interesting generalization or hypothe- 

 sis can be made as to the three realms, 

 geographical isolation has been so continu- 

 ous and prolonged that great orders of 

 mamnials have been evolved (Chart IV.) in 

 each. Thus Ardogcea containing the broad- 

 est and most highly diversified land area, 

 appears hj'pothetically as the center in 

 which fourteen primitive and specialized 

 orders radiated from each other. In the 

 southern portion of Neogxa, at least four 

 orders sprang from primitive members of the 



rivers and coasts of America, Europe and 

 probably Africa as tlieir radiating center, 

 w^hile the Cetacea occupied the fourth or 

 oceanic realm. 



Now, we mean to express by this hy- 

 pothesis that Realms were the main centers 

 of adaptive radiation of orders, by no means 

 the exclusive areas of distribution, for dur- 

 ing the periods of land contact certain 

 members of these orders found their way 

 into adjacent realms. Each realm, there- 

 fore, contains its pure autocthonous types 

 and its migrant or derived types. Regions, 

 on the other hand, may be distinguished 

 from realms as geographical and zoological 

 areas, which have been isolated from each 

 other for shorter periods, either by climatic 

 barriers, as in the case of the Arctic or cir- 

 cumpolar region, or bj' great physical bar- 



