80 HEREDITY 



merely not opposed to a Mendelian interpretation oi 

 the same facts. For Mendel's law and the pure line 

 conception are far simpler generalisations than the 

 statistical laws of heredity. Moreover, they give us 

 an insight into the actual mechanism of heredity which 

 we could never gain by regarding the subject statisti- 

 cally. In fact, it is becoming clear that statistical 

 results have no bearing whatever on the physiological 

 process and mechanism of heredity. Where a Mende- 

 lian interpretation is possible, then, it is more valuable, 

 both theoretically and practically, than a statement 

 of a statistical result. 



Yet it seems to be the case that, no matter how 

 universally Mendel's principles may be found to apply, 

 there will remain a wide field for valuable statistical 

 investigation ; namely, in those cases which are too 

 complex for Mendelian analysis. It is tolerably certain 

 that such cases will arise in the future, due either to 

 the large number of factors having an influence on a 

 particular character, or else to the obscuring of Mendelian 

 inheritance by the occurrence of relatively large modi- 

 fications. We have seen that there are three inde- 

 pendent factors for such a comparative detail as red 

 colour in wheat, and in the common garden snapdragon 

 the various colour varieties, numbering over a hundred, 

 are produced by the presence or absence of some fifteen 

 independent factors, which have somewhat complex 

 inter-relations. The researches by which these factors 

 have been analysed have necessarily been difficult 

 and laborious, and with a slow-breeding organism, or 

 one which could not be bred in great numbers, the work 

 would have been altogether impracticable. With 

 flower colour, moreover, we are dealing with fairly 

 definite and easily distinguishable characters. If, on 

 the other hand, we were to attempt to unravel the 

 heredity of such a character as size in beans, where 

 many factors are certainly involved, and where the 

 various types are indistinguishable without extensive 

 breeding tests, we should have a problem of much 

 greater complexity. 



