VARIATION 117 



The fact of multiple allelomorphs throws new 

 light on this matter, and we must doubtless say 

 that the recessive determiner is not simply a 

 vacant spot, but is a real factor which does not 

 function as does the dominant. Hence this 

 class of cases falls under our group <z, rather 

 than under the present group. Still, we must 

 admit that sometimes there is actual loss of sub- 

 stance instead of modification, and there is 

 reason for thinking that this may very rarely 

 be brought about by environmental factors. 

 There may be, in some cases, a selective de- 

 struction of the items of inheritance. 

 The above two types of variation are the most diffi- 

 cult to study and understand, but also the most im- 

 portant, since they will permanently modify the 

 material of inheritance. Could we bring them about 

 experimentally, we could practically produce new 

 species. Even then, we could work along only certain 

 lines which the character of the germinal substance 

 permits, just as the chemist can make only certain 

 compounds. It is probably fortunate that man has 

 not been able in this manner to play the parfe of a 

 creator ; he would doubtless have made a mess of 

 things. Nature may be "blind," but working in the 

 long run and the fullness of time, she does her work 

 better than we could hope to imitate. 



3. Two other classes of variations have nothing to New corn- 

 do with any change in the germ plasm itself. ShS 

 c. Variations due to new combinations. These have qualities 

 been discussed under Mendelism and the Red 

 Sunflower. It is evident that they will break 

 up again, forming still other combinations, ex- 

 cept when they become homozygous. In the 



