458 EDWIN CARLETON MacDOWELL 



concludes that such racial differences are inherited. Mutations 

 were found that were believed to represent a different kind of 

 inheritance. 



I insist that the burden of the proof rests on those who contend 

 that these two types of variation and inheritance are reducible to a 

 single category, that of discontinuity. Anything like a proof of this 

 contention appears to be. lacking (p. 452). 



However carefully performed and interesting in themselves are 

 these experiments of Sumner, they professedly are not critical 

 and stand as opposition mainly on account of their general appear- 

 ance of blending, which is so interpreted with far less difficult 

 experimenting than is required by the current theory. It is 

 obvious that objections based on evidence of this sort do not 

 have much weight in combating the conclusions supported by the 

 array of analytical selection experiments. 



SUMMARY 



1. The early generations of a race of Drosophila that was 

 continuously selected for increased bristle numbers during forty- 

 nine generations show higher correlation coefficients than any 

 others. 



2. With the greatest uniformity of the environment, the last 

 thirteen generations show no correlation at all. 



3. A final test of the germ plasm of the selected race was made 

 by raising four more generations without selection. In gener- 

 ations 52 and 53, 31,000 bristle counts indicated that the higher- 

 grade parents did not produce higher-grade offspring; the cor- 

 relation was negative for parents by sons and by daughters as 

 well as for grandparents by grandsons and granddaughters. 



4. No evidence of sex linkage is found by comparing the co- 

 efficients of mothers and fathers correlated separately with their 

 sons and daughters. The difference between the sexes is inter- 

 preted as due to the general developmental conditions initiated 

 by the sex chromosomes, rather than to the linkage of a specific 

 gene. 



