NO. 519] THE LIFE CYCLE OF HYDATINA SENT A 149 



dence on the question of starvation. It is not practicable 

 to diminish the quantity of food placed in the dishes, 

 without at the same time diminishing the quantity of 

 other substances introduced. Several experiments were 

 made to test the apparent effect of starvation. When 

 the young were isolated, they were given only a> much 

 food as was judged necessary to support life and enable 

 them to produce a moderate family. A control line, 

 which was fed abundantly from the same cultures, was 

 maintained in each case. One experiment, Table V, 

 included 55 generations and covered a period of three 

 months. In this time, the well-fed control varied so 

 greatly in the proportion of sexual to parthenogenetic 

 females produced, that it is important, not merely to 

 note the totals, but to divide the experiment into periods,, 

 and compare the results in each. 



Table V 



While the experiment as a whole shows a decidedly 

 higher percentage of sexual forms in the starved families, 

 there are parts of it that show a slightly lower proportion 

 in the starved line. The variation in the percentage of 

 sexual females from one period to the next in the starved 

 line is of the same sign as in the corresponding periods 

 of the well-fed line, but is smaller in every case. While 

 the major fluctuations of the starved line occur simul- 

 taneously with those of the well-fed line, they are less 

 in degree ; the extremes of the starved series are always 

 well within those of the well-fed. I can find only one 



