38 Zoologia: N. Y. Zoological Society. [Vol. I 



taking place under conditions approximating those of the forma- 

 tion of the Salton Sea, in the Colorado desert in southern 

 California, where the sudden appearance of a large body 

 of water in an arid region might be accompanied by a con- 

 siderable increase in the relative humidity. If any change in 

 color in the plumage of the local species of dove should prove of 

 adaptive value, in influencing even to a slight degree the vitality 

 of the birds, the species would at once benefit. Natural selection 

 would exert an important influence in the elimination of those 

 with less plastic pigment-forming enzymes, or it might operate 

 in an antithetical way by raptorial elimination of the darkest, 

 most conspicuous, individuals. Be this as it may, we can be 

 reasonably certain that this ontogenetic modification is only so- 

 matic, and that the offspring would acquire the dark color anew 

 only with the first few moults. Until some of the problems 

 offered by Scardafella, mentioned on page 33, have been solved, 

 we have no means of telling how soon or in what manner this 

 modification would become congenital ; or, in other words, would 

 be transferred from an acquired to an inherited character. 



But this is not necessary to the present argument, that, in 

 the Scardafella color modification, organic selection finds strong 

 support. If the new character, ontogenetically acquired, is in 

 any way adaptive, it might easily be the means of preserving 

 the species until phylogenetic variation had impressed it upon 

 the race. 



That we have much to learn concerning these greatly dis- 

 puted classes of variation is shown by another phase of this very 

 experiment. We have seen that the specific and apparently per- 

 manent color distinction between S. inca and S. ridgwayi is, in 

 reality, so plastic that a year or two of superhumidity will annul 

 it. Thus two species widely separated by a geographical hiatus 

 may actually be differentiated by the most evanescent of acquired 

 characters. So, while theoretically explainable by organic selec- 

 tion, yet we see that a character, which few evolutionists would 

 have considered apart from regular inherited, specific characters, 

 might, after we know not how many years and generations, still 

 prove to be of the most superficial and transient nature. 



Bearing in mind the Porto Santo rabbits and the probably 

 considerable time during which Scardafella has inhabited Mex- 

 ico, we can easily see that an ephemeral, acquired, yet adaptive 

 character, might render life possible in a region otherwise un- 

 tenable, through a sufficiently long period of time for congenital 

 variation plus natural selection to bring about skeletal and other 

 modifications of generic, if not indeed of family, importance. 



