394 THE AMERICAN NATURALIST [Vol. LVI 



bred for 21 generations. The individuals for the experi- 

 ment came from 20 mass matings of 3 pairs of parents 

 each from this line. The bottles were started December 

 16, 1921, and the flies used in the experiment emerged 

 December 28, 1921, to January 9, 1922. 



The flies were counted through the counting tube into 

 our standard shell-vials in groups of 50 each. Each 

 day's bottles were divided into three groups at the be- 

 ginning of the experiment, but all had the same regular 

 treatment until the flies in them were 30 days old. This 

 is a' point where the L line is beginning distinctly to turn 

 downward. From that time on until the end of their life 

 one series of flies was given chicken juice in their food, 

 and one series the juice and pulp of crushed Drosophila 

 larvae. The chick embryos used were 14 days old. The 

 juice was extracted in a beef -juice extractor, and added 

 to the regular food at the rate of approximately 2 c.c. 

 to 100 c.c. of food. With the Drosophila larvae, the whole 

 pulp was used, and that too was added to the regular 

 food at the rate of approximately 2 c.c. to each 100 c.c. 

 of food. All the flies, experimental and control, were 

 transferred every day to fresh food, made up that day, 

 except on Sundays. 



On Feb. 8 an accident happened to the incubator at the 

 source of our chicken supply,^ so that for 12 days no 

 chickens were obtainable. 



In all particulars except those specified above, the 

 procedure in these experiments was the standard tech- 

 nique of this laboratory in duration of life work de- 

 scribed in (27). 



Results 



The survivorship distributions are given in Table V, 

 on the basis, A, of 1,000 starting at emergence, and, B, 

 of 1,000 starting at age 31 days, that is at the time when 

 the experimental feeding began. 



2 We are greatly indebted to^ our colleagues, Dr. and Mrs. Warren H. 



