GENE MUTATIONS AND EVOLUTION 89 



propagate; and in some of their descendants again useful 

 mutations occur, along with many that are not useful. Thus 

 in the long run, there appear in the line of descent only those 

 in which one useful mutation has followed another. Thus 

 genetic variation appears to have followed a definite straight 

 course; although as a matter of fact it has occurred in many 

 directions, most of which were not useful and hence are not 

 found in the pedigree of the final products. The fact that 

 genetic variations do occur in directions that are useless or 

 harmful is not in doubt, it may be pointed out, since useless 

 and harmful mutations actually occur in the genetics of living 

 organisms. 



But in view of what we have seen of gene mutations in 

 earlier pages, the question remains whether mutations indeed 

 do supply the constructive and useful changes required for 

 such evolutionary progress ; or whether progress results from 

 types of change not yet observed in living organisms. 



II. The second of the indirect methods of studying the 

 nature of evolutionary changes lies in the examination of the 

 diversities among living organisms, attempting to judge how 

 such differences must have arisen. On this matter there is of 

 course a vast literature, dealing with taxonomy, anatomy, 

 embryological development, geographical distribution and 

 the like. Much of this was written before the rise of modern 

 genetics, or by men not familiar with it, so that it is difficult 

 to interpret in genetic terms. Other parts of it do not suffer 

 from this difficulty. As an example of such work done in full 

 awareness of the methods and results of genetics, and indeed 

 making use of these, some features of the work of F. B. 

 Sumner^* may be sketched. 



Sumner has made extensive studies of varieties and subspe- 

 cies of certain wild mice — "deer mice" — of the genus Peromys- 

 cus; with attempts to judge of the nature of the differences 



