STATISTICAL AND PHYSIOLOGICAL LAWS 335 



In a certain number of cases a knowledge gained by experiment 

 enables us to say that from two parents D and R a certain kind 

 of offspring will result — D(R) — and that the progeny of D (R) 

 X D(R) will be in the proportions iDD + 2D(R) + iRR ; and 

 we can interpret the result by a theory of the segregation of the 

 gametes of D(R) into two sets of pure gametes which bear in 

 potentia the contrasted characters embodied in the original 

 parents D and R. 



In other cases, however, a knowledge of the characters of the 

 parents does not enable us to predict the results of an individual 

 pairing, and we fall back on the law of ancestral inheritance 

 which states the average result — the most probable result. 



As it appears to us, Mendelian phenomena are illustrated 

 when the parents differ in sharply defined contrasted characters 

 which cannot blend or compromise, and the extension of ex- 

 periment will doubtless go on increasing our knowledge of 

 these unit characters and their behaviour. The formulation 

 will remain whether the theory of the segregation of pure gametes 

 be confirmed or not. In other cases, however, the Galtonian for- 

 mulation seems the only one applicable, and here the need is to 

 work out — perhaps along the lines of Weismann's germinal selec- 

 tion of determinants — a conceivable physiological interpretation. 



Unless we misunderstand the situation, there is this clear 

 difference between Mendelian and Galtonian formulas — that 

 Mendelian formulae apply to the progeny of known crosses or 

 hybrids, while Galtonian formulae apply to intra-racial heredity. 



We must refer the reader to Mr. Yule's discussion (1902) 

 of the supposed antagonism between Mendelian and Galtonian 

 conceptions — a discussion which leads this expert to conclude 

 " that Mendel's Laws and the Law of Ancestral Heredity are 

 not necessarily contradictory statements, one or other of which 

 must be mythical in character, but are perfectly consistent 

 the one with the other, and may quite well form parts of one 

 homogeneous theory of heredity." 



