SEX-RATIO OF THE DEER-MOUSE. 157 



same year-to-year fluctuations were undergone by both pure and 

 hybrid stock, with the exception of the last (from 1920 to 1921). 



When each of these years is divided into 3-month periods 

 (February to April, etc.), we find that, on the whole, the years 

 of low and high ratio show this tendency for each of the four 

 " seasons." Graphs (not here reproduced) show that the " curve " 

 for 1917, while undergoing closely similar fluctuations to that for 

 the entire series, remains throughout far below the latter. Of 

 the other two "low" years, 1921 remains throughout below the 

 mean, while 1920 gives a lower ratio in three out of four of these 

 periods. Of the two years which show ratios appreciably higher 

 than the mean, each gives a " curve " which keeps far above the 

 mean during three of the four seasons. The exceptions fall at 

 different times in the two cases. 



These considerations, while they add little to the evidence for 

 the statistical reality of these yearly differences, show that the 

 latter do not depend upon conditions which act primarily during 

 any particular period of the year. To be sure, computations 

 have been made which show that the departures of these yearly 

 ratios from the mean condition occur preponderantly in the fall 

 and winter months. This relation is probably due to chance, 

 however, so that the figures need not be given here. 



Similar differences have been recorded by a number of in- 

 vestigators between the sex ratios of different years for mankind, 

 but so far as we know, these differences have not been discussed 

 by them. They have been regarded as accidental, or at least as 

 irrelevant and negligible. Thus in the study by the Pearls, al- 

 ready referred to, of the sex ratios of " pure " and " hybrid " 

 races in the city of Buenos Aires, the authors confine their dis- 

 cussion to the apparently higher ratios which are found among 

 offspring of mixed parentage. They quite overlook the fact, 

 however, that the sex ratio of the entire population varies mate- 

 rially from one year to another, and that these yearly differences 

 are in some cases both absolutely greater and statistically more 

 certain than are those between the pure and hybrid elements of 

 the population. Thus the ratio for the year 1900 is 97.0 .go, 

 while that for 1904 is 105.4 .95. Each of these figures is based 



