296 
GENETICS: E. C. MacDOWELL 
fication. By plus selection from hooded rats that came from a cross 
between the minus selected race and wild, the minus race was immediate- 
ly transformed into the plus race, whereas similar plus selection from the 
minus race without the cross with wild, required 6 generations to move 
the means even up to the "0" grade. 
There are then, four lines of evidence that support the conclusion that 
selection, in the case of extra bristles has separated genetic differences, 
or units, that existed at the beginning of the experiment, and that did 
not reappear de novo: (1) at first correlation coefficients are constantly 
positive and significant, later they are mainly not significant and fluctu- 
ate between positive and negative; (2) selection advances the means at 
first, later it does not; (3) before selection makes its advance, return, or 
low, selection separates a distinct race, while afterwards, the same pro- 
cedure fails to accomplish the same result, however, (4) if a cross with 
normal precedes, the low selection becomes as effective as it was at the 
beginning. 
As the material under discussion does not constitute a pure line in 
any sense, the conclusions drawn do not have any immediate bearing 
on the pure line theory. Moreover, environment plays such a con- 
spicuous role that nothing positive can be said about any hypothetica 
variations in the germ plasm that may be hidden by variations in the 
environment, and accordingly be too small to be determinable. It may 
be suggested that if the environment were not such a controlling factor, 
or if it were possible to reduc'e the variability of the environment, smaller 
changes in the germ plasm might be discovered. This claim may be 
made for any work in selection that could be imagined; no failure of 
selection, however stationary the environment may seem to be, can 
escape this claim, namely that variations in the germ plasm may be 
taking place, although they are not distinguishable in the soma. The 
claim may be made, but it will give neither the breeder, nor natural selec- 
tion any opportunity to make progress. It appears that instead of 
offering a fatal objection to the work here reported, the evident environ- 
mental factor serves well to emphasize the utter futility of attempting 
to deal in theory or fact with supposed germinal phenomena that can 
never be demonstrated or utilized. The variations in environment 
commensurate with the viability of the fly are great, but they do not 
hide the fact that there are germinal differences in regard to the numbers 
of extra bristles ; they do not hide the fact that two races may be raised 
at the same time under the same conditions and maintain their individ- 
uality ; nor do they explain the fact that after selection has made a cer- 
tain amount of progress it is no longer possible to raise the means of 
