35 2 DON C. WARREN. 



METHODS AND MATERIAL. 



The selection experiments were carried out upon three unre- 

 lated stocks. The stock for the first experiment was collected 

 in Bloomington, Ind.; for the second, in Saratoga, Ind.; and for 

 the third, in Warsaw, Ind. In the first two cases the experiments 

 were started as soon as the stocks were collected from nature but 

 the Warsaw stock was inbred in the laboratory for about six 

 months before the experiment was started. At all times the 

 greatest care was exercised to prevent contamination. The 

 flies were provided with an abundance of food and care was taken 

 to examine all the offspring of each pair. All matings were 

 single pair matings. 



Each experiment was conducted upon the offspring of a single 

 pair of flies, that is, for the first experiment, a virgin female was 

 selected from the Bloomington stock and mated to a brother. 

 From among the offspring of this pair of flies, the parents of the 

 two strains were obtained. The one strain in which there was 

 selection for a relatively high number of females in comparison 

 to the number of males, has been called the "high " strain and the 

 one in which the selection was in the opposite direction, has been 

 called the "low" strain. Before selection was started, a suffi- 

 cient number of flies were examined to obtain a fair estimate of 

 the sex-ratio of the stock under consideration. Moenkhaus's 

 high and low strains were not offspring of a single pair of flies but 

 were obtained by selecting from nature, pairs which had unusual 

 ratios in the desired direction. Furthermore, he knew nothing 

 of the original sex-ratio of his strains before selection began. 



In each generation an attempt was made to examine the off- 

 spring of ten pairs in each strain. Unsuccessful matings, usually 

 due to poor food conditions, sometimes prevented this. In each 

 generation, the pair which gave the most extreme ratio in the 

 desired direction was used as the parent of the succeeding genera- 

 tion. At times, technical difficulties prevented the most ex- 

 treme pair being used and in these cases the next best was used. 

 Pairs showing extreme ratios but producing a small number of 

 offspring were not used. Since it was not possible to know 

 which pair had the most desirable ratio till all of the offspring of 

 all the pairs in a generation were examined, the technique of the 



