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



It will be seen (Fig. 6) that the sex ratio for broods containing 

 from 2 to 5 individuals, inclusive, are arranged in a regularly 

 ascending series, while the last group (7 and over) gives us a 

 ratio far in excess of any of the others. Exceptions to this gen- 

 eral trend are groups i and 6, though it will be noted that these 

 two (as well as the last) are based upon relatively small numbers. 



Despite the absolute magnitude of some of these differences, 

 however, their statistical significance is in every case very doubt- 

 ful. In most instances 'they are less than twice their own probable 

 errors. When we combine broods of i to 3 and 4 to 9, inclusive, 

 the sex ratios of these major size groups become 94.85 2.94 and 

 102.42 3.01, respectively. The difference between these values 

 is rather less than twice its probable error. Here, then, as in the 

 case of season, we cannot say with any confidence that the size 

 of the brood is a factor which influences the sex ratio. 



It might be inferred, moreover, that this correlation, even if 

 actual, results from the fact that the smaller broods are ones 

 which have been depleted by unrecorded deaths, and that the 

 death rate has here been higher among the males. This explana- 

 tion derives some support from the fact (p. 128) that among the 

 very small number of identified dead included in our records, the 

 males greatly preponderated. It is rendered improbable, how- 

 ever, by the fact (p. 129) that broods which are known to have 

 been depleted give a sex ratio which is not significantly lower 

 than that shown by the broods in which no deaths are known 

 to have occurred. 



Nor do our records afford any ground for the belief that these 

 differences in the sex ratios of broods of different size depend 

 upon the unequal distribution of such broods throughout the 

 year. And it is equally improbable that the seasonal cycle, dis- 

 cussed above, is dependent upon the seasonal distribution of 

 large and small broods. It happens that the seasonal cycle, in 

 respect to the mean size of the brood/ does correspond pretty 

 closely, on the whole, with the seasonal cycle for sex ratio (Fig. 

 i). But the differences, in the former case, are so slight that 

 they could not at most account for more than a fraction of the 

 differences in the latter. 



