SEX-RATIO OF THE DEER-MOUSE. 143 



King and Stotsenburg obtained conflicting results from two 

 large subdivisions of their material, when arranged to test the 

 possible correlation between sex ratio and size of litter. The 

 authors cautiously conclude that " the lack of uniformity in the 

 results of this arrangement of data indicate that apparently there 

 is no well-defined relation between litter size and sex in the 

 albino rat." 



Weldon (1907), on the other hand, believed that he had ob- 

 tained some evidence that a positive correlation existed between 

 the size of the litter and the sex ratio in mice. This relation was 

 discernable only when the mean size of litter in different genera- 

 tions was considered, not being evident in respect to individual 

 litters within a generation. The figures presented are not, how- 

 ever, very impressive. 



Combinations of the Sexes in Individual Broods. Thus far, the 

 relative numbers of the sexes in broods of different size have 

 been dealt with by methods of mass statistics. The total for each 

 sex has been computed, and the ratio between these totals ob- 

 tained. Such treatment would entirely obscure one possible 

 phenomenon of high interest, namely the tendency of members of 

 a litter to agree with one another in respect to sex. Do we, for 

 example, encounter broods consisting of four or five of the same 

 sex more frequently than would result from chance? 



This question we have endeavored to answer by arranging 

 broods of each size in groups according to the number of each 

 sex present. For example, broods of three present four possible 

 combinations : 3 J\ 2 <$ + i ?, I J 1 + 2 ?, 3 ? In table A are 

 included only broods in which no deaths are known to have oc- 

 curred. There are given the actual number of broods containing 

 a given combination of males and females, likewise the "ex- 

 pected " number, to the nearest integer. In computing these last 

 figures, it has been assumed that males and females are equally 

 likely to be produced. The close approximation to equality in 

 the material as a whole seems to warrant this procedure. 



When we consider the comparatively small number of broods 

 present in most of these groups, the agreement between the actual 

 and the expected figures is remarkably close. This agreement is 



