BRISTLE INHERITANCE IN DROSOPHILA 139 



DISCUSSION 



The role of environment 



When families raised at the same time are compared, the 

 means of the return, extracted and the main high race show a 

 parallehsm. Of these three races there is one with high parents 

 and high means, one with low parents and high means, and one 

 with low parent and low means, yet the means of all the off- 

 spring rise and fall with striking similarity. Differences in par- 

 entage, similarity in environment, similarity in variation; it is 

 an evident conclusion that these variations are due to some- 

 thing other than differences in the germ plasm. There seems 

 to be considerable justification for the belief that the greater 

 part of the variations found in the curves is due to differences 

 in the environments of the developing flies. An inverse rela- 

 tionship has been found to hold, in general, between the age 

 of a culture and the number of extra bristles on the flies 

 hatching; the older the bottle the smaller the flies are apt to be. 

 The number of extra bristles appearing can be controlled to a 

 certain extent. At any time small flies can be obtained by allow- 

 ing the larvae only a small amount of food; this may be done 

 either by allowing the bottle to dry, to mould, or to have too 

 small an amount of banana. Optimum conditions for the pro- 

 duction of small flies are the optimum conditions for the produc- 

 tion of low bristle grades. When a unique set of atmospheric 

 conditions are encountered, as in the constant temperature room, 

 there appears a unique modification of the means (generations 

 29 to 31, high race) It seems hardly possible to escape the 

 conclusion that the controlling factor in producing the varia- 

 tions in bristle numbers is environment and not heredity. 

 •This may incline some readers to immediately claim that, 

 since the influence of environment is so strong, it would be im- 

 possible to discover any changes in the germ plasm, and forth- 

 with condemn any conclusions that may be drawn from these 

 data. At this point it will be well to consider the basis for our 

 suppositions that we can know the character of the germ plasm 

 by the condition of its somatic bearer. 



