156 F. B. SUMNER, M. E. MCDANIEL AND R. R. HUESTIS. 



sampling. We must, of course, consider the possibility that these 

 differences among the sex ratios of the various years are due to 

 some of the other factors already considered. May they not, in 

 part at least, be due to the unequal representation in these dif- 

 ferent years of broods born in months of high or of low sex ratio? 

 Aside from the fact that these yearly differences are even more 

 extreme than any of the seasonal ones, we have excluded this 

 possibility by a procedure previously adopted in similar cases. 

 The number born in each month of each year was multiplied by 



,60 



A 



140 



120 



_ 6.82. 



80 



60 



1915 1916 1917 1916 1919 1920 I9ZI 



FIG. 7. Differences in the sex ratio during the different years comprised 

 in our records. Continuous line pure. Dotted line = hybrids. 



the sex ratio for that month (based on the entire material). The 

 weighted mean for the year in question was then computed. None 

 of the yearly ratios thus obtained differed by more than 4.0 from 

 the general ratio (97.37), showing that seasonal distribution 

 cannot account for these yearly differences. 



Nor are the latter to be accounted for by the possibly unequal 

 distribution of pure and hybrid births. Figure 7 shows that the 



