1 .-,( ) 



THE AMERICAN NATURALIST 



[Vol. XLIV 



factor which operated on both lines in the same way, 

 but to less degree in one line than the other, namely, the 

 food culture. 



Of especial interest are the second and sixth periods 

 Table V. In these, not only the parentage, but also the 

 absolute number, of sexual females is higher among the 

 starved individuals. This shows that the percentage is 

 not raised by the mere elimination of some of the parthen- 

 ogenetic females. 



The experiment with starvation was repeated three 

 times, each repetition including eight to sixteen genera- 

 tions. Each time there was a considerably higher pro- 

 portion of sexual females in the starved families. 



While it is conceivable that several factors may be at 

 work producing these different proportions of sexual to 

 parthenogenetic forms in the well-fed and starved fam- 

 ilies, the experiments with the filtrate from old food 

 cultures show that the different quantity of dissolved 

 substances incidentally given with the food is sufficient 

 to explain the results. If this is the correct interpreta- 

 tion, the larger proportion of sexual forms in the starved 

 families is not due to lack of food, but to the absence 

 of chemicals which, in the well-fed families, prevent the 

 appearance of the sexual forms. 



