252 CLAUDE W. MITCHELL 



In one instance only we wish to refer more minutely to the 

 results of Shull's experiments. In one of his earlier articles^i he 

 presents the results of the "Influence of food culture on the per- 

 centage of male producers." His immediate conclusion as to re- 

 sults is somewhat favorable to the positive action of starvation 

 as increasing male production. But the increase is not great 

 and is later ascribed to other causes. Without commenting upon 

 this general conclusion, we wish to call attention to the interesting 

 nature of the individual results in this starvation line. The first 

 striking fact shown by his table is an apparent rhythmic varia- 

 tion. Almost without exception we find one or two very 

 small families of, say, from three to twelve individuals followed 

 by a succession of from four to six families with from two to 

 three times this number. This variation in the size of family 

 obviously demonstrates that, however food was supplied, effec- 

 tive starvation was not uniform. The individuals producing the 

 larger families must have ingested more food than those produc- 

 ing the minimum number of offspring. In noticeable correla- 

 tion with this change in size of family, we find the fact of increase 

 and decrease in the number of male producers. The smaller 

 families scarcely produce males at all, whereas the larger fami- 

 lies frequently contain 50 per cent or more of male producers. 

 It seems very natural to interpret these results as due to the 

 fact that the degree of starvation which produced the smaller fami- 

 lies precluded, as we always found it to do in our experiments, 

 the possibility of male production; while the higher nutrition, 

 productive of the larger family made such production possible. 

 A more minute examination of Shull's table shows the results to 

 be in even closer accord with our own. Thus it is not the first 

 large family that usually shows the high rate of male production 

 spoken of; indeed in some instances the first large family follow- 

 ing upon one or more small ones shows not a single male producer. 

 This coincides strikingly with the results of our experiments 

 which show that it is not the first generation of high potential 

 individuals which produce males. They rarely do so, but them- 



'1 Jour. Exp. ZooL, vol. S, pp. 311-54. 



