122 



THE AMERICAN NATURALIST [Vol. LII 



fers that within this group of species there has been a 

 definite evolution of chromosome groups. 



The known methods of species formation, therefore, 

 may be described as follows: (1) Factor mutations, caus- 

 ing more or less extensive heritable somatic changes, some 

 of which are adapted to the environment and persist. 

 These, under the influence of natural selection, provide 

 the means for gradual differentiation of groups having 

 the same chromosome numbers. Presumably these groups 

 would be recognized at successive stages in the process 

 as geographical or ecological forms or races, distinct 

 varieties and, ultimately, related species. (2) Chromo- 

 some group alterations, which produce new and sometimes 

 inconstant forms, but which may also produce true species. 

 (3) Species crosses, which are known to give rise to new 

 types, some of them constant, but mostly inconstant 

 forms, all of which are cryptomeres, potentially capable 

 of throwing new combinations of parental characters in- 

 definitely. The possibility should also be noted here that 

 new constant types, having different numbers of chromo- 

 somes from the parent species, might originate as species 

 hybrids. 



Factor mutations occasionally produce new dominant 

 characters. This fact, now fully established by various 

 investigations, is of considerable theoretical significance. 

 It has been a common practise during recent years to ex- 

 plain the origin of recessive characters as due to verlust 

 mutations, i. e., mutations due to the "loss" of factors. 

 This conception has been associated with the much used 

 though inadequate "presence and absence" hypothesis, 

 according to which the only relations which can exist 

 with respect to a certain factor depend on its presence or 

 absence from the hereditary material. Difficulties are 

 met when attempts are made to explain the origin of domi- 

 nant mutations in terms of this hypothesis, for in such 

 cases it is necessary to assume that a factor has been 

 added to the hereditary material. As a result of employ- 

 ing the presence and absence hypothesis in genetic nomen- 



