THE MEASUREMENT OF VARIATION. 3 



been regarded as quite outside all law was in reality 

 amenable to treatment according to the well-known 

 Laws of Chance. Again, investigations from the ex- 

 perimental side have suggested much concerning the 

 causes of variations, both genetic and somatic. Still 

 again, a fresh burst of activity in the collection of data 

 regarding the actual facts of variation, more especially 

 in respect of organisms found in a state of nature, has 

 shown us how much in this branch of the subject there 

 remains for us yet to learn. 



Perhaps the keynote of most of the recent work on 

 variation lies in the recognition of the fact that almost 

 all the problems to be solved must be attacked from a 

 numerical standpoint. It is no longer sufficient to 

 say that such and such a kind of variation is frequently 

 or occasionally found. It is necessary to know the 

 exact amount of the variation, so far as it is measur- 

 able, and the exact proportion of cases in which 

 it occurs. Only by obtaining data of this kind 

 can we hope to ascertain with any certainty the 

 probable degree of importance of any particular vari- 

 ation in the evolution of a species, and whether such 

 evolution is actually taking place at the present 

 day. ISTo apology is therefore needed for the fre- 

 quent introduction of figures into the study of ques- 

 tions of variation. Rather is this necessary if one 

 should attempt to found theories and deduce conclusions 

 from generalised statements and opinions, unsupported 

 by such evidence. To say that any particular organ is 

 very variable means but little, for so much depends 

 upon the personal opinion of the observer as to what 

 constitutes a great and what a slight variation. But 



