CORRELATION 179 



gene acts as a differential, turning the balance in a given 

 direction, affecting certain characters more conspicuously 

 than others.' This view certainly harmonises better with the 

 data of genetics, but it does not enable us to envisage the 

 process by which complex structures develop harmoniously. 



This is the question which has been raised by Russell 

 (1930). He points out that there is no evidence for a qualita- 

 tive division of the chromosomes at any stage of development. 

 Each cell (in typical cases) has the same equipment of hereditary 

 material. The fact that different cells give rise to such varied 

 structures can only be explained by considering the spatial 

 relations of the cell to the whole. Russell is so impressed by 

 this antinomy that he is prepared to discard the whole unit- 

 character hypothesis of heredity. But this extreme attitude 

 appears perverse. Somehow or other the quantitative pre- 

 dictions which can be based on the chromosome theory must 

 be accounted for. The difficulty here raised has also been 

 considered by Woodger (1929, chapter ix, especially sect. 9). 

 He attempts to visualise development as a process of gradual 

 realisation of spatio-temporal parts, while genes are concerned 

 only with the characterisation of the parts. In order to include 

 those cases in which whole parts may be inherited on Mendelian 

 lines (e.g. vertebrae) he suggests that, for the purposes of 

 genetics, the part should be defined rather by its dimensions, 

 so that ' absence ' is merely the end term in a gradual process, 

 rather than something sharply different from ' presence.' 

 This idea of the relation between heredity and development 

 seems helpful in trying to orientate our fragmentary knowledge, 

 but scarcely helps us as yet in the matter of character correla- 

 tions. The characters do not act as separate units in develop- 

 ment, and we cannot help suspecting that whatever controls 

 the orderly unfolding of the inherited organisation must be 

 deeply concerned with the correlation of the characters on 

 which the end result largely depends. 



We feel that there is a very real difficulty here. On the 

 one hand we have the obvious and incontestable fact that 

 (p. 163) the characters defining species are rather loosely 

 correlated, we have produced certain reasons (p. 172) for not 

 considering their association as of a ' physiological ' (i.e. 

 intimate and causal) nature, and we have definitely suggested 

 that it is in the bulk of cases due to the members of species being 



