194 



A. FRANKLIN SHULL. 



Notwithstanding the great fluctuation of the proportion of the 

 male-producing females in famihes of different sizes, it might be 

 possible to see in the percentages given in the last columns of 

 these two tables (or perhaps only Table IV.), a slight increase 

 from top to bottom, and hence a correlation between size of 

 family and male production; though the degree of correlation 



Table IV. 



Showing Size of Family and Proportion of Male-producing (cf 9 ) and 



Female-producing (99) Females in a Single Parthenogenetic Line 



of Hydatina senta. 



certainly can not be high. However, even if such correlation 

 exist, it does not follow that the percentage of male-producing 

 females is dependent upon the degree of nutrition, which deter- 

 mines size of family. The one way to test the effect of nutrition 

 is to alter it artificially, and note the results. That has been 

 done in my starvation experiments (Shull, '10, pp. 320 ff.), with 

 results that were positive but of such a character that they could 

 be explained as due to changes in the chemical composition of the 

 medium, rather than changes of nutrition. Mitchell ('13) ob- 

 jects that in drawing conclusions from these starved families, 

 I have regarded only the totals; had I observed individual 

 families, he believes, I would have reached a different result. 

 Mitchell states that the male-producing females in my starvation 

 experiments did not appear in the smallest families, which were 

 presumably the offspring of the most starved parents, but in the 

 larger families, produced by the better nourished females; and 

 from this supposed fact concludes that abundant male production 

 is due in Hydatina, as he believes it to be due in Asplanchna, to 



