SEX-RATIO OF THE DEER-MOUSE. 137 



out the twelve months, or to the unequal part which the mice of 

 different years may have had in determining the various monthly 

 ratios ? 



The first of these possibilities is set aside by reference to Figure 

 4, showing the seasonal cycles of the pure and hybrid stocks 

 plotted separately. The second we have tested by ascertaining 

 the actual effect of each year's quota in determining the sex ratio 

 for each of the twelve months. 1 But an examination of these 

 figures (which it is not worth while to present here) does not 

 support such an explanation of our seasonal cycle. Unless, 

 therefore, all of these monthly differences are due to errors of 

 random sampling, they are probably caused in some more or less 

 direct way, by seasonal changes of temperature or some other 

 physical agent. 



The findings of certain other investigators regarding seasonal 

 differences in the sex ratio may be appropriately considered here. 

 King and Stotsenburg (1915) have presented data from 7,619 

 white rats reared at the Wistar Institute. The exact temperature 

 conditions existing in the animal quarters are not discussed by 

 the authors, though it is stated that the provisions for heat regula- 

 tion were inadequate, and that the rats suffered greatly from heat 

 during the summer months. It may be presumed that the rooms 

 were heated to some extent during the winter. 



When grouped by months, the entire data of King and Stotsen- 

 burg show what might be construed as a biennial rhythm. As 

 regards the position of these annual maxima and minima, their 

 results are in some respects in direct contradiction to ours. One 

 well-defined maximum covers the period from June to August, 

 another, the period from October to December. Minima occur 

 in March and September. The second of these is of brief dura- 

 tion, being bounded on either side by months having high sex 



i A new set of monthly ratios was computed as follows : For each month, 

 the number born in each of our years was multiplied by the general sex 

 ratio for that year, and the sum of these products was divided by the total 

 number born in that month. If the monthly differences which we have dis- 

 cussed above were due to the different seasonal distribution of the mice born 

 in different years, this new set of monthly ratios ought to show much the 

 same relations as the set computed earlier. 



