571 



THE AMERICAN NATURALIST [Vol. XLIX 



When mongrels and the more fertile hybrids are propagated for 

 several generations, an c\l reme anioiinl of variability in the offspring 

 in both cases is notorious; but some few instances of both hybrids and 

 mongrels long retaining a uniform character could be given. The vari- 

 ability. however, in the successive generations of mongrels is, perhaps, 

 greater than in hybrids. 



One of the causes of ordinary variability . . . is . . . that the repro- 

 ductive system from being eminently sensitive to changed conditions of 

 life, fails under these circumstances to perform its proper function of 

 producing offspring closely similar in all respects to the parent form. 



From "Plants and Animals under Domestication" we find 

 the following. 



an element, probably a potent one, in causing variability. 



A variation to be effective in species formation must reappear 

 in some of the descendants. That a variation could, through 

 selection within a pure strain be increased or decreased in the 

 direction of selection to form a stable species has been seriously 

 questioned since Johannsen's classic experiments. It is well 

 understood, on the other hand, how selection in a mixed popula- 

 tion could cause the variation to move in the direction of selec- 

 tion up to a certain point. 



The first mutant stock selected for the experiment was cherry 

 club vermilion. The factors for these three characters are 

 linked together and are also linked with sex; the second stock 

 was black pink bout, which has the three factors independent of 

 each other and none is linked with sex. These factors are sup- 

 posed to lie in the second, third and fourth chromosomes, re- 

 spectively. The third stock was black purple vestigial arc speck, 

 which has the five factors linked together. They lie in the sec- 

 ond chromosome. A stock from France was crossed to the mu- 

 tant stock several months after the other crosses were made, and 

 eosin tan vermilion was substituted for the cherry club ver- 

 milion, and pink kidney sooty rough for the black purple ves- 

 tigial arc speck stock because flies of these particular stocks 

 were not to be had at the time desired. 



These forms were chosen because it was thought that if muta- 

 tions do arise from hybrid forms there would be more probability 

 of their origin from a mutant varying in several characters when 

 crossed to wild than if it varied in only one character. Also by 

 using stock containing several recessive characters a check could 



