174 ELMER ROBERTS 



There are other cases in Drosophila of factors which find ex- 

 pression only under certain environmental conditions. Miss 

 Hoge ('15) found that in a certain stock of Drosophila super- 

 numerary legs were produced in a Mendelian ratio at about 10°C., 

 but at normal temperature this character was very rare. Mor- 

 gan ('15) described the case of abnormal abdomen in which the 

 black bands, normally present, are broken and irregular or en- 

 tirely absent. This abnormal condition is induced by the pres- 

 ence of moist food, but when the food becomes more dry, the 

 individuals that emerge are entirely normal in appearance. 

 Dexter ('14) also found that more beaded wings were produced 

 in a wet culture than in a dry, and more in alkaline than in acid 

 cultures. Extra number of bristles in Drosophila was decreased 

 by a scanty food supply (MacDowell, '15). 



Further investigation under a constant environment will be 

 necessary before one can come to a conclusion concerning the 

 efficacy of selection on the size of vestigial wing. On the basis 

 of this explanation, selection failed because the somatic expres- 

 sion of this character was not indicative of the full germinal 

 potency. 



The second plausible explanation of the increased size and 

 variability of the vestigial wing in the 'crossed-in' series is fac- 

 torial contamination. Castle, particularly, held that such a con- 

 tamination might occur — that gametes are subject to change and 

 inconstant. Castle and Forbes ('06) concluded that there was 

 contamination of the germ-cells produced by the cross-breds of 

 long- and short-haired guinea pigs, due to the association of the 

 factors for long and short hair. Likewise, MacCurdy and 

 Castle ('07) were of the opinion that the gametes of cross-breds 

 from a cross of hooded and Irish rats had changed because the 

 extracted patterns were somewhat changed. Contamination of 

 the gametes would mean variation of the unit-character in ques- 

 tion. It was the variabihty in this hooded pattern of rats which 

 furnished the foundation for the extensive selection experiment 

 by Castle and Philhps ('14). The results of their experiment 

 showed that the amount of pigment could be diminished or in- 

 creased at will by selection. They adopted the hypothesis that 



