124 THE AMERICAN NATURALIST [Vol. LII 



strengthening evidence that factor mutations furnish the 

 means for evolutionary change within existing species, it 

 must be admitted that there is a wide gap which needs to 

 be rilled. May we not look to future studies on the 

 phylogeny of the chromosomes to supply this need in some 

 measure? I venture to suggest that more work like that 

 of Metz on the chromosome groups of related species will 

 prove to be an important source of further light on this 

 problem. 



We may now consider the role of factor mutations in 

 more detail. It is certain that even those species having 

 the same chromosome number differ as a rule in many 

 unit factors. Hence in order to explain the origin of one 

 from the other it is necessary to assume one of three pos- 

 sible methods of procedure. 



1. There may have been one or more factor mutations 

 having manifold somatic effects. That profound somatic 

 differences such as would distinguish species from one an- 

 other are sometimes produced by single factor mutations 

 is proved by the results of the crosses between the oak- 

 like walnut and its parent, the California black walnut, 

 which I have described in earlier papers. The mutant 

 form, unlike most so-called monophyllous varieties, differs 

 from its parent in every gross external feature, yet it 

 behaves as a simple recessive in Fj and F 2 . However, 

 factor mutations wliuli induce such extensive somatic 

 changes seem to be exceedingly rare. 



2. There may have been simidtaneou^s mutations in sev- 

 eral factors, thus producing full-fledged an individual of a 

 new species. Objection to this hypothesis is found in 

 the observation that factor mutations always or nearly 

 always occur singly, i. e., a single factor mutation in a 

 given individual at a given time. This observation is 

 mathematically predictable since the probability of the oc- 

 currence of two factor mutations in the same individual 

 at the same time, according to the principle of least 

 squares, would be the inverse ratio of the square of the 

 number of typical (unchanged) individuals in the popu- 



