CORRELATED VARIATIONS. 97 



is perfectly constant in one species of a group, it will 

 tend to be constant in allied species." The general 

 truth of this law seems highly probable on the face of 

 it, because most allied species have presumably split off 

 from their common ancestor at no very remote period, 

 and so would still resemble each other more or less 

 closely in respect of variability, just as they do in re- 

 spect of morphological structure. The results, already 

 quoted, of Weldon and Warren for the correlation of 

 various organs in Carcinus and Portunus, afford direct 

 experimental evidence in support of this law. 



In conclusion, it may not be out of place to make one 

 or two brief remarks as to the general bearing of the 

 evidence which has been adduced in these three chap- 

 ters concerning the " facts of variation." I believe 

 that they include most of the more important and more 

 recent contributions to our knowledge of the subject, 

 especially those in which the results have been ex- 

 pressed in exact numerical terms. It may very likely 

 be objected that insufficient importance has been at- 

 tached to the kind of information which Darwin col- 

 lected so thoroughly and in such profusion in his work 

 on " Variation in Animals and Plants under Domestica- 

 tion." 



The reason of this is twofold. In the first place 

 it seemed unnecessary to refer at great length and 

 with much frequency to data with which most seri- 

 ous students of Biology must be already conversant: 

 whilst in the second place, comparatively little infor- 

 mation of this kind has, as far as I am aware, been 

 recorded in scientific journals since Darwin's time. It 

 has, indeed, been recognised that the facts of variation 



