228 Experimental Zoology 
extent sterile with the parent type, in the same sense that even 
closely related species are often sterile. De Vries’s statement 
gives the impression that he regards this property of sterility as 
a general property of mutations. But there are other cases of 
mutations where nothing of the sort is evident. Such types as 
the ancon ram, the merino sheep, the turnspit dogs, the Japa- 
nese peacock, the guinea pig with whorled hair, certain breeds of 
fowls with different combs, show no evidence of infertility; when 
crossed with the parent type. Furthermore if many of our do- 
mesticated races owe their origin to the appearance of mutations, 
and to the introduction of wild elementary species (which in 
principle may amount to the same thing), it is surprising to find 
no trace of sterility between the different breeds, as we should ex- 
pect if sterility is one of the general peculiarities of mutations. 
If many Linnean varieties are elementary species’ that have 
arisen as mutations, it is well known that they are, as a rule, per- 
fectly fertile when crossed. 
It seems likely, therefore, that the sterility of some of the 
cenotheras may be the exception and not the rule for elementary 
species, and that the same principle holds here that Darwin 
has so ably developed, viz. that fertility is a graded process, 
not strictly following the lines of species and varieties, but 
dependent on the extent of physiological differences between 
two forms. 
Our examination of the possibility of new mutations becoming 
swamped by intercrossing shows that there is no real difficulty 
to be met in this direction, although it is also apparent that some 
kinds of mutations may, of course, become lost by crossing.’ 
The opportunity exists, however, for some of them, especially 
those that dominate, to find a place in the economy of nature. 
No one will pretend that all new forms that appear will survive. 
1 Some of them may be ‘elementary varieties,’ others local varieties due to 
external conditions, and others hybrids in which there is a blending of the new 
and old characters. 
? Those, for example, in which blending occurs, for this, too, is one of the re- 
sults of hybridizing even where we have reason to think that the new character 
arose as a discontinuous variation. 
