200 G. Udny Yule. 
“ Not to mention moderns, these high hopes” (of the generality 
of the law) “ had been finally disposed of by the work of the 
experimental breeders such as Kolreuter, Knight, Herbert, Gartner 
Wichura, Godron, Naudin, and many more. To have treated as 
non-existent the work of this group of naturalists, who alone have 
attempted to solve the problems of heredity and species—Evolution 
as we should now say—by the only sound method —experimental 
breeding —to leave out of consideration almost the whole block of 
evidence collected in Animals and Plants — Darwin’s finest legacy as 
I venture to declare—was unfortunate on the part of any exponent 
of Heredity, and in the writings of a professed naturalist would 
have been unpardonable.” (Mendel’s Principles, pp. 112-13). 
The “ experimental breeders ” referred to never touched the 
questions of intra-racial individual heredity at all. The work 
of Darwin bears chiefly on racial heredity, and on hybridisation ; the 
relatively small portions on intra-racial heredity do not give inform¬ 
ation in any form which enables one to apply it to the criticism of 
statistical laws. In the present state of our knowledge it is impossible 
to confuse the subjects in so loose a fashion. Laws of hybridisation 
cannot be admitted as general principles of heredity until they have 
been proved to hold as such—the title of Mr. Bateson’s volume 
begs the whole question—nor can laws of heredity be in general, 
or necessarily, expected to hold good in cases of hybridi¬ 
sation. “ Experimental breeding” is certainly a sound method for 
the study of heredity: it is not the only sound method, for equally 
good material may be obtained by simple observation, as in the case 
of man. But if by “experimental breeding” is meant solely hybridi¬ 
sation (i.e. crossing of different races, varieties, species or 
genera in general), then I join issue with Mr. Bateson altogether. 
Experiments on crossing can give nothing but laws of crossing; it 
may be possible that some of these laws are applicable to the 
breeding of pure races, but this cannot be decided without definite 
trial. The work of the whole of the “ group of naturalists” he 
mentions is valueless for the branch of work on which the 
biometrical school has been engaged. 
So far we have dealt solely with the direct heredity between 
parent and offspring ; now let us consider the question of inheritance 
from the remoter ancestry. Supposing a series of grandparents and 
their grandchildren to be measured as before, it is evident that we 
could construct an “ estimating equation” just like (i), but giving 
the mean character of the gra/zcfchildren in terms of the character 
