298 LOUISE H. GREGORY. 



The first experiments were carried on during the winter 

 of 1913-1914 in the greenhouse at Columbia University 

 (Table I.). Five young of a wingless mother were deprived of 

 food 7! hours daily for four days. These (Ai series) produced a 

 total of 352 young, 254 or 71 per cent, of which were winged. 

 In the second series (A2) younger members of the same family, 

 8 aphids were starved 8 hours daily for four days. The total 

 number of offspring in the eight families was 463, of which 249 

 or 53 P er cent, were winged. In the control series of a total of 

 228 (young of five families) 23 per cent, were winged showing a 

 difference of 48 per cent, if compared with the Ai series and of 20 

 per cent, with the A2 series. If the entire A series is considered 

 as a whole, among the young of 13 families 61 per cent, were 

 winged. Apparently the lack of food has had its effect on the 

 second generation, none of the starved individuals developing 

 wings. 



The experiments were repeated at Woods Hole in the summer 

 of 1915 (Table II.). Again the young of a wingless mother 

 were used for the experiment and control. The families of seven 

 wingless females starved 45 hours, were observed and out of a 

 total of 1 88 young, one winged form only appeared. The seven 

 families of the control series showed 3 winged individuals. 

 Apparently the effect of fasting was overcome in some way. 



Considering that the atmospheric conditions had been unusu- 

 ally favorable for the development of the peas, the weather 

 having been very hot with much humidity, I concluded the 

 food had been sufficiently rich to overcome any adverse effects 

 of starvation, and that in order to induce the effects of starvation, 

 it would be necessary to increase the number of starvation 

 periods. Consequently the experiments were repeated the 

 following summer in Princeton, Mass. (Table III.). 



Five young starved 60 hours produced a total of 254 offspring, 

 30 per cent, of which were winged. Ten individuals forming the 

 control series produced 514 young, of which 56, or 9.7 percent., 

 were winged. Here again starvation has had its effect. 



Summarizing all the experiments on wingless parents, it is 

 seen in Table III., that 25 starved individuals produced 1,257 

 young, 46 per cent, of w r hich had wings: that 22 normal indi- 



