No. 613] EXPERIMENTS CONCERNING EVOLUTION 3 L 



In crosses of these different types Fj is about inter 

 mediate, as some curves in Table II show. 1 



And F 2 gives a 1 : 2 : 1 ratio, or 3 light + medium : 1 dark, 

 whatever races are involved. (This statement should be 

 taken only on its face value. As a matter of fact, we find 

 here, within the invariably present ratio of 3:1, very 

 strange details of the kind described as ''gametic con- 

 tamination," and, furthermore, an obscuring of the ratio 

 in earlier stages followed later by the right ratio, ap- 

 parent lack of segregation, etc. From a purely genetic 

 point of view, the analysis of these phenomena consti- 

 tutes the most interesting part of this work, but it has no 

 special relation to the problems here under discussion.) 

 Back crosses, however, give a 1 : 1 ratio. The following 

 table gives a few data of this kind. 



TABLE III 



