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



ratios. The former, on the other hand, is 'Merely the lowest point 

 in a prolonged period of low sex ratios. 



When plotted according to 3-month periods, the records of 

 King and Stotsenburg show an annual minimum in the spring 

 ( March to May) and an annual maximum in the summer (June 

 to August), while an intermediate condition is indicated for both 

 fall and winter. 1 Our figures, on the other hand, show maxima 

 in the spring and fall, minima in the summer and winter, though 

 the " seasons " adopted by us commence a month earlier than 

 those adopted by the former authors. 



The marked differences found between the seasonal cycles of 

 Peromyscns and the white rat might plausibly be attributed either 

 to the difference of species or to differences in the environmental 

 conditions under which the two sets of experiments were con- 

 ducted. On the other hand, it appears to us that the reality of 

 the seasonal cycle described by King and Stotsenburg is subject 

 to exactly the same doubts as that described by us. The fact that 

 the two groups of years into which the rat experiments were 

 divided gave quite contradictory relations for the winter months 

 certainly gives us reason for such skepticism. 2 



Heape (1907) gives evidence for the existence of seasonal dif- 

 ferences in the sex ratios of dogs. The records for nearly 18,000 

 greyhounds show sex ratios which fluctuate irregularly between 

 in and 128 during the months of January to September inclusive. 

 In October, however, the curve rises to 145, in November to 180, 

 and in December to 195. As regards the dates of conception, 

 these last three months become August, September and October. 

 For collies, on the contrary, he says that there is " no evidence 

 that conception at any particular time of year affects the propor- 

 tion of the sexes born." It must be added, however, that his 

 figures for the collies indicate very considerable monthly differ- 



1 The foregoing statements refer to the entire data of King and Stotsenburg. 

 The material is divided by the authors into two groups which are partially 

 discordant with one another. One of these (that for 1911-1913) gives a graph 

 which is almost exactly the converse of ours. 



2 It is possible, however, that unknown influences (other than random sam- 

 pling) caused differences in the sex ratio from one year to another. That such 

 annual fluctuations actually occur in Peromyscns will be pointed out below. 



