THE EXPERIMENTAL STUDY OF GENETIC RELATIONSHIPS 1 35 
his contemporaries, although they were stated with unsurpassed clear- 
ness, and would have changed the current of biological thought if 
anyone had realized their bearing on the theory of evolution. Another 
was Focke, who laid emphasis upon the importance of hybridization 
in species formation. Even then, as now, he believed that, in such 
large and polymorphic genera as Rubus and Rosa, many of the species 
have originated by hybridization, followed by the sorting out of stable 
forms. 
From Darwin's time until very recently, however, it has been the 
prevailing view that the selection of extremes from a continuously 
varying population would result in a continuous, gradual modification 
of the entire population, and that such selection had brought about 
the formation of varieties in cultivation and of species in nature by the 
simultaneous transformation of masses of individuals. It need only 
be said that carefully planned and executed experiments lend no 
support to this view. All the evidence, on the contrary, seems to 
show that no amount of selection will suffice to modify the range of 
fluctuating variation of an organism. On the botanical side there is 
little evidence of the efficacy of selection as a factor in evolution; on 
the zoological side there are certain selection experiments of Castle's, 
carried out with characteristic care and accuracy, but surely capable 
of a different explanation from that which he gives. Even if correctly 
interpreted they have at best a dubious bearing on the problem of 
species formation. His experiments deal with the inheritance of a 
certain color pattern in rats, which, in a presumably homozygous race, 
may be modified in either direction by selection. The changes from gen- 
eration to generation are very slight, however, and we cannot conceive of 
any agency in nature which would bring about assortive mating among 
such slightly dissimilar individuals. While it is becoming increasingly 
clear that the old selection theory is untenable, we are becoming more 
and more convinced that evolution does take place with measurable 
rapidity, and that the factors concerned with it are mutation and 
hybridization. 
The new point of view we owe primarily to de Vries, who has deter- 
mined the distinction between non-heritable fluctuating variations and 
inheritable germinal variations, or mutations, and has developed the 
mutation theory. It should appeal especially strongly to systematists, 
most of whom have really never been convinced of the adequacy of the 
discarded selection hypothesis. Why, if species had come about by 
