258 



SCIENCE 



[N. S. Vol. XXXI. Xo. 790 



or shallow, to a chilling or warming or 

 other physical change, often as fatal to 

 one fauna as it was favoring to another. 

 The controlling factor was geographic; it 

 set the hour of immigration; and through 

 knowledge of it alone can we estimate the 

 time elapsed during the wandering. 



The case stated is that of passage from 

 diversified to unified environments. It 

 has repeatedly characterized geologic peri- 

 ods; and it has repeatedly culminated in 

 the evolution and distribution of cosmo- 

 politan faunas, which simultaneously peo- 

 pled remote realms with like species. The 

 coefficient of uncertainty, with which we 

 must qualify any correlation that depends 

 solely on identity of species, is reduced to 

 a minimum at the time of culmination of 

 uniformity. 



Uniformity has in turn repeatedly 

 yielded to diversity. Large marine f aunal 

 realms have been divided into provinces 

 by emergence of continents, by diversion 

 of ocean currents, by diiferentiation of 

 ■climates, by local dilution or concentration 

 of ocean waters and by the other changes 

 Avhich establish physical barriers. Lands 

 have been diversified in like manner. In 

 each such province evolution re-began by 

 extinction of the unadapted and survival 

 of the fitter forms. Originating in a com- 

 mon ancestry, the faunas of two neighbor- 

 ing provinces may for a time have had 

 much in common. As they developed dif- 

 ferences, Jordan's law of geographic iso- 

 lation came into play. The resulting 

 geminate species may have been closely 

 contemporaneous; but continued contem- 

 poraneity depends upon uniform rates 

 of differentiation, which the changing 

 environments do not favor. Several rela- 

 tions, other than uniformity, are conceiv- 

 able; isolate a fauna under static condi- 

 tions and contrast it with the same fauna 

 under changing conditions. Assume the 



changes to be favorable or unfavorable to 

 the indigenous fauna. Contrast isolation 

 with more or less free emigration and im- 

 migration. Consider the many factors 

 of environment and the many possibilities 

 of variation in sensitive, highly developed 

 organisms. 



Must we not conclude that diversity of 

 species rather than likeness will be the 

 rule at such a time among contemporane- 

 ous faunas? 



But it is upon likenesses that we rest our 

 faunal correlations. "What do they sig- 

 nify? Simply that the surviving species 

 of a fauna have remained unchanged dur- 

 ing a longer or a shorter period, either at 

 home or during migration, or that varia- 

 tions have developed similarly in two 

 provinces from an ancestral stock similarly 

 conditioned. 



When we find a German fauna in New 

 York or a Russian fauna in western North 

 America, the occurrence means that the 

 particular fauna persisted in an environ- 

 ment which offered it no stimulus to varia- 

 tion during the period of migration from 

 the ancestral home to the new domain. 

 But that period remains indeterminate. 

 It was the shifting of the habitat that made 

 the migration possible and that set the rate 

 of progress. It was a geographic move- 

 ment first and the faunal journey was a 

 consequence. 



Or in case of similar variations from an 

 ancestral stock, can we assume that the 

 stimuli acted in different provinces at the 

 same times and at the same rates 1 "When 

 they did, but only when they did, were the 

 similar variations contemporaneous; and 

 the cases of such coincidence may reason- 

 ably be regarded as exceptional. 



Hence I conclude that : 



Correlation hy identical or closely re- 

 lated species, faxmules or floras is subject 

 to a coefficient of error, which is a function 



