﻿Manchester Memoirs, Vol. xlix. (1905), No. <>. 15 



and cur application of atomic and molecular hypotheses 

 to problems in heat, elasticity, and cohesion is essentially 

 based on statistics of average conduct. Corpuscles in each 

 other's presence are supposed to obey certain laws of 

 motion, but no explanation has hitherto been given of 

 these laws. So it is with vital units ; they vary, why they 

 vary we know not, and we explain nothing by attributing 

 it to bathmic influences. As we can predict little or noth- 

 ing of the individual atom, so we can predict little or 

 nothing of the individual vital unit. We can deal only 

 with statistics of average conduct. We have laws of varia- 

 tion and laws of heredity, in themselves quite as general 

 and as definite as the majority of those we meet with in 

 physics." 



I may perhaps take this opportunity to explain that 

 I have used the term biometric theory advisedly, and that 

 the definition of it that I have had before my eyes is not 

 merely "the application of exact statistical methods to the 

 problems of biology,"* but the aspect of vital phenomena, 

 just quoted from the " Grammar of Science," which prompts 

 that application. 



And I believe that I am justified in including under 

 the term "biometric" both Pearson's and Galton's theories 

 which, though in one respect they are radically different/)* 

 resemble each other in regarding heredity as a mass- 

 phenomenon and in treating it by the statistical method. 



The Mendelian, on the other hand, with a new appli- 

 cation of experiment, is a biological demon who, " per- 

 ceiving " and " handling the separate " units themselves, 

 tries to find out their properties by mating them with 

 other units. But, here again, I need not expatiate on 

 the closeness of the parallel I have suggested when we 



* Nature. Oct. 27, 1904. P. 626. 



t Karl Pearson. Biometrika. Vol. III., pp. no and 11 1. 



