Life Cycle of Hydattna Seuta 351 



was altered by the change of medium, the numbers used were too 

 small to decide; but the influence felt in the first generation is 

 at most only a fraction (probably a small one) ot that apparent 

 in the second generation. Since this is the case, it is readily seen 

 how a reduction in the quantity of substance in the water at a 

 comparatively late stage of the rotifer's life, when the ongonia or 

 eggs are well developed, might result in more male-producers in the 

 next generation, whereas a similar reduction just after hatching 

 might have no efi-ect. This view harmonizes with the result that, 

 while starvation throughout life is followed by an increase in the 

 proportion of male-producers, starvation for only a short period 

 after hatching has no effect. 



There remains the further possibility that starvation just after 

 hatching may defeat its own purpose. Doubtless some unassimi- 

 lated food comes over from the egg to the young female. How 

 long this lasts has not been determined. Since a young female 

 eats very soon after hatching, to deprive her of food must dis- 

 turb her normal processes; but it seems to me doubtful whether 

 as great a degree of starvation has thereby been obtained as has 

 been supposed. 



Not less important than the relation of my experiments to 

 those of previous workers, is the examination of the experiments 

 for defective points. The assumption that male-producers 

 appear chiefly in a given part of the family, and that reducing the 

 size of the family may be accomplished by omitting, unaltered, 

 some specific portion of the family, have been shown to be with- 

 out foundation. As a corollary of this, the similarity of large and 

 small families with respect to the relative position of the male- 

 producers in them shows that in general the value of experi- 

 ments with Hydatina is not diminished because the families are 

 small, provided the requisite aggregate number of individuals is 

 obtained. 



Death losses sometimes invalidate experiments, and must 

 always be taken into account. If any safe conclusion is to be 

 reached, the differences caused by the conditions of the experi- 

 ment must be so great as to make the death losses insignificant, 

 or it must be shown that these losses are not selective. In the 

 experiments with the filtrate from old food cultures, it has been 



