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



upon over 20,000 births, and the difference is 6 times its probable 

 error. Even greater and more " significant " differences are to 

 be found in the yearly ratios of some of the separate population 

 groups (in one case 8 X P. E.), and it is of interest here to note 

 that the years of high and low sex ratios do not tend to corre- 

 spond in these various groups, but that they sometimes show 

 exactly opposite conditions. 



It is, of course, inconceivable that a difference of calendar 

 year, as such, should have any more influence upon the sex ratios 

 of mice or men than the phases of the moon or the conjunctions 

 of the planets. It is likewise improbable that any mean differ- 

 ence in the weather from one year to the next, can be the re- 

 sponsible factor. For the meteorological differences between 

 the most widely unlike years are small in comparison with the 

 differences between the summer and winter seasons in any single 

 year. But we have seen that these yearly variations of sex ratio 

 may be even greater than the seasonal ones. Again, examination 

 of the composition of the stock during these widely divergent 

 years (say 1916 and 1917) gives no suggestion of a clue based 

 upon considerations of this sort. 



Huxley (1920) discusses a case in which a teleost fish pro- 

 duced, for nearly a year, three times as many females as males. 

 Later, this ratio among the young produced changed to 2 fe- 

 males : 3 males for a few weeks, after which the numbers of the 

 sexes became approximately equal and remained so for several 

 years. Huxley believes it probable that in the first stage a cer- 

 tain proportion of the individuals having the zygotic formula XY 

 became " somatic females " or " feminized males," and that such 

 individuals produced X and Y eggs, which by fertilization gave 

 rise in the next generation to an excess of males. A converse 

 hypothesis might, of course, be invoked to explain the change in 

 our stock, from a ratio of 125 in 1916 to one of 71 in 1917. But 

 this is suggested merely as a remote possibility, justified only by 

 the utter absence of any plausible scientific explanation to cover 

 the case. 



Finally, it should be urged that even highly improbable things 

 sometimes happen, and that it is not impossible that our most 



