486 University of California Publications in Zoology [Vol. 12 



tion is cumulative, is indicated by a great many facts from zoo- 

 geography. Species known to have originated recently, as for 

 example, those in neighboring areas separated by a barrier not 

 absolute, are structurally much more alike than species known to 

 have been separated for longer times. Thus Castor subauratus 

 shastensis and Castor subauratus subauratus, occupying neighboring 

 areas, and only recently separated, are structurally much more 

 alike than are Castor fiber and Castor canadensis canadensis occupy- 

 ing areas in different continents, and separated for a much longer 

 time. Island forms of mainland stocks are often poorly differenti- 

 ated, the degree of divergence seeming in a measure to be asso- 

 ciated with the length of time of separation of the island from the 

 mainland. This cumulation is, of course, only within limits, for 

 there are notable examples of very long separation with compara- 

 tively little differentiation, and even a few cases wherein long sep- 

 aration has apparently failed to bring about any differentiation 

 whatever. 



To summarize the remarks in this section : It may be conceived 

 that isolation works with organisms in nature in some such way 

 as the experimentalist works with organisms in the laboratory. The 

 latter isolates a number of individuals and subjects them to a 

 diversity of surroundings known as the conditions of the experi- 

 ment. In nature the parent stock becomes separated, by some 

 means or other, natural barriers corresponding to the walls of the 

 container or of the laboratory, and the different environmental 

 conditions in new territory occupied, corresponding to the condi- 

 tions of the experiment. Differentiation may ensue in either case. 

 In the higher vertebrates, at least, it has ensued practically in all 

 cases. Geographical isolation probably conditions differentiation 

 by segregating organism and environment together and maintain- 

 ing their mutual interrelations over long periods of time. 



Evidence from geographic distribution of higher vertebrates 

 does not oppose, if it does not definitely support, the thesis that 

 speciation is dependent on modification in germinal constitution. 

 It appears to present unequivocal data that the response to the 

 different environments depends upon the nature of the organism. 

 It furnishes a mass of evidence against the theorj r of the possibility 

 of the appearance of a new species in territory occupied also by the 

 parent, so emphasizing the necessity of migration and geographic 

 isolation as conditions of polytypic evolution. It indicates that 



