Analysis of Heteeogeneity in Some Simplest Chaeacters 237 



ment, evaporation, and soil-moisture. The temperature was set at 21° C, the 

 relative humidity at 75 per cent, and the air was changed once each minute, 

 giving uniform evaporation rates. 



In September 1907, I mated, under the controlled conditions, 3 pairs from 

 pair 1 of stock 6, 2 from pair 8, 4 from pair 10, all of these taken at random in 

 the population of each parent fraternity. The method of doing this was to 

 separate the sexes, place each in a dish under a dark or opaque cloth, and then 

 draw them out singly as needed for the matings, thus preventing any possible 

 selection in the matings of the pairs within the population. These 9 pairs of 

 random matings in different lines were placed in the tanks, and there 8 of them 

 gave progeny. At the same time 3 mass cultures were mated at random, and 

 consisted of 3 of each sex taken from the progeny of pairs 2, 14, and 16, and 

 these were allowed to breed under the conditions of the experiment as were the 



Fig. 32. — Diagrams showing different measurements taken to determine the body 

 index in these animals. (See text.) 



pairs, all giving progeny. The small size of the breeding-chambers made it 

 necessary to grow only small populations, but each pair was allowed to breed 

 until death; the elimination was made in the eggs and was, of course, purely a 

 random one, and consisted of destroying about half of every batch of eggs laid. 

 Mated pairs and mass cultures gave normal, strong progeny in the middle of 

 October, and presented no difference in the indices or in the range thereof in 

 any of the populations resulting. At the end of October each line was continued 

 by another random mating, and in the same tanks as before, and these all gave 

 progeny, reduced in number as before, in the first half of December 1907. 

 The resulting progeny, however, showed exactly the same ranges in the index 

 as did the previous generation and as did the other portions of the series, namely, 

 the progeny of the other stocks that were reared at the same time, but not under 

 accurately controlled conditions. I was greatly surprised at this result, having 

 expected that there would be an alteration of the range in the index in at least 



