20 Darbishire, Laws of Heredity. 



that Mendelism dealt with the units while biometry 

 was concerned with masses. Now I see that my 

 inability to do so 'was due to the fact that biometry 

 meant to my friend that the only way to measure the 

 resemblance between parents and children is the method 

 of the correlation table ; in which he was quite right : 

 while biometry called up in my mind the Law of Ancestral 

 Inheritance, and especially the manner in which data 

 designed to establish that Law are collected and dealt 

 with ; such, for example, as in the case of one of the last 

 series of data from which such correlation coefficients 

 have been worked out — that of greyhounds. I had in my 

 mind the collective treatment in one correlation table of 

 such different characters as Black, Brindled, Fawn, White, 

 and Red ; * while he was thinking of the only method of 

 measuring the intensity of inheritance within a single such 

 character — say, Red. Now this I believe brings us to the 

 heart of the matter. When I say that the Mendelian deals 

 with units and the biometrician with masses, I mean, not 

 that the former deals with a few and the latter with many, 

 but that the former first settles what character he is 

 going to treat as a unit and then only deals with it 

 in large numbers when he is sure that the component 

 units of this number are identical ; their sameness having 

 reference to such properties as can be discovered by 

 mating them with their like and with their unlike. It 

 is just as necessary for the Mendelian to have a large 

 record of matings, as the biometrician, to establish his 

 generalizations. But though the Mendelian might allow 

 that the only method of measuring the similarity between 

 parents and offspring within a group (such as the unit 

 character Red) was that of the correlation table, he would 

 vehemently maintain that the biometric method over- 



* Rartington, etc. :05, p. 264. 



