274 Experimental Transformation of Species 



entiating them from the parental type. These '* mutations " 

 are found to transmit their distinctive traits either as re- 

 cessives or dominants, and thus indicate some change in the 

 chromosome mechanism. In addition, therefore, to the 

 many recombinations of traits that hybridization may bring 

 about in accord with Mendel's law of segregation, there ap- 

 pear new traits making possible more new combinations. 



Most mutations involve very slight although distinct 

 changes, and therefore do not raise the question of a new 

 " kind." In other cases the entire appearance of a plant or 

 animal is altered, and yet a freak sheep is still sheep enough 

 to be recognized as such and at the same time mutation 

 enough to be considered a freak (see page 182). One sig- 

 nificant fact about mutations is that when such a character 

 appears it comes in a single step and not by a succession of 

 small steps. In many cases the mutations appear in one sex 

 and remain sex-linked in succeeding generations. In such 

 cases we would assume that the germinal change takes place 

 in one of the sex chromosomes. In other cases a mutation, 

 although appearing in a single individual in one or the other 

 sex, becomes in a few generations characteristic of the whole 

 stock, that is, of both sexes. 



The frequent appearance of sex-linked mutations 

 among the experimental insects has led to a reexamination 

 of certain well known species, especially butterflies, in which 

 the females differ from the male, and in some of which there 

 are two or more distinct female forms. Darwin had at- 

 tempted to explain these sex differences as due to the result 

 of sexual selection, assuming that the males preferred, in the 

 female, one type of pattern to the other, with the result 

 that in the course of generations the preferred type came 

 to prevail. In other species the appearance of the distinct 

 patterns was explained as having a protective value to the 

 species, since some patterns resembled forms that were not 

 destroyed by the enemies (birds, lizards) to the same ex- 

 tent (see page 346 and Fig. 88). Breeding experiments 

 among the insects showing these many forms would indicate 



