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



Since, as will be shown presently, there are rather wide seasonal 

 differences in the sex ratios found by us, and since the different 

 months are represented very unequally in our records, it is of 

 interest to present the mean of these separate monthly ratios. 

 This figure is 95.65. 



Both of the foregoing figures are distinctly smaller than have 

 been given by various writers for rats and mice, 1 as well as for 

 man and some other animals. 2 In most cases a decided excess of 

 males has been reported. Miss King (1918), for example, from 

 the records of 2,818 white rats born in her "stock" (i.e., unse- 

 lected) series, obtained a sex ratio of 104.6. 



Some attention should here be devoted to the possibility, al- 

 ready referred to, that the sex ratio which we have obtained for 

 Peromyscns has been influenced by differential mortality. As is 

 well known, the sex ratio among stillborn infants is very high, 

 being frequently given as 130 or more; 3 and for cattle a similarly 

 higher prenatal mortality among the males has been reported. 4 

 For Peromyscns we have no data on this subject since the sex of 

 stillborn young -was in no case determined. As regards early 

 post-natal mortality, also, our direct evidence is very meager, so 

 much so as to be almost worthless. 424 deaths occurred be- 

 tween the date of counting and the date of marking and registra- 

 tion. Owing to cannibalism and other causes, it was frequently 

 impossible to determine the sex of these dead individuals, and in 

 many other cases we neglected to do so; but this was done in 40 

 cases. Restricting our consideration to those mice which died 

 during the first two months of life, we have 31 individuals, of 

 which 20 were males and n females. From these figures one 



twice as great as those which would be obtained by another formula which has 

 been widely followed (see Pearl and Pearl, 1908), and are therefore much 

 safer as a basis for estimates of the significance of results. Indeed it turned 

 out that certain highly interesting conclusions which we had drawn at the 

 outset had to be relinquished on this account. It may be added that the con- 

 clusions of certain other writers are greatly weakened if the formula here 

 employed is substituted. 



1 King, IQII, 1918. 



2 Morgan, 1907, pp. 365-366; 1913, pp. 230-231. 



3 Morgan, 1907, p. 368; Schultz, 1918, p. 264. 



4 Jewell, 1921. 



