SEX-RATIO OF THE DEER-MOUSE. 131 



and two low periods annually. The sex ratios for these four 

 periods are as follows : 



1 i ) February-April 1 04.23 +3.85 



(2) May-July 91.48 + 3.48 



(3) August-October .' 102.29 + 3-68 



(4) November-January 85.21 + 4.62 



The greatest difference between two of these ratios is that be- 

 tween the first and fourth periods. This is 19.02 6.01. Taken 

 by itself, such a difference is commonly regarded as having a 

 probable significance. 



When the same broods are grouped according to date of con- 

 ception, the figures become : 



(1) January to March 105.35 + 3-Ss 



(2) April to June 91.09 + 3.50 



(3) July to September 102.68 + 3.66 



(4) October to December 82.10 + 4.38 



Here, again, the greatest difference is b'etween the first and 

 fourth periods, being, in this case 23.25 5.83, or almost exactly 

 four times its probable error. 



Figure 2 is based upon the sex ratios for the four 3-month 

 periods of the year, both when the broods are grouped according 

 to date of birth and according to date of conception. The two 

 graphs are seen to be in very close agreement, though the divisions 

 between the 3-month periods have, in the second case, been ad- 

 vanced by one month. 



It is when we subdivide our material into groups of one sort or 

 another that the inconstancy of these seasonal differences becomes 

 evident. Dividing our broods according to years is not very in- 

 structive, owing to the relatively small number born in any one 

 year, taken singly, and particularly to the exceedingly meager 

 records for certain months. These graphs (not here reproduced) 

 show various extreme fluctuations which must be regarded as 

 purely accidental. Five 1 of the seven " curves " show, however, 

 what may be regarded as spring and fall maxima, though the 

 position of these varies somewhat from year to year. Two of 



1 Those for 1915 to 1919 inclusive, 1920 and 1921 being the exceptions. 

 No graph is possible for the first year ("1914) owing to the small number of 

 months represented in our records. 



