2IO 



RAYMOND PEARL AND H. M. PARSHLEY. 



Data. 



The sex distribution of 480 calves, in relation to the time of 

 coitus in the oestrous period, is exhibited in Table I. Besides 

 the absolute figure the sex-ratios calculated as number of c? d 71 

 per 100 9 9 and the percentages of cf cf and 9 9 and their 

 probable errors are given. 



TABLE I. 



SHOWING THE NUMBER OF CALVES OF EACH SEX, AND THE SEX RATIOS, 

 FOLLOWING COITUS AT SPECIFIED TIMES IN THE (ESTROUS PERIOD. 



From this table the following points are to be noted: 



1. The proportion of male births increases steadily with the 

 later occurrence of coitus in the cestrous period. 



2. Comparing the extremes there are more than 10 per cent, 

 more males produced when coitus is late in heat than when it is 

 early in the oestrous period. 



3. Taking the three groups together the total figures give a 

 sex-ratio intermediate between those- exhibited in the extreme 

 groups. 



Now to consider the important question: Are the differences 

 between the different groups shown in Table I., statistically 

 significant, or are they such as probably arise from the inevitable 

 "error" of random sampling? To get light on this point recourse 

 must be had to the probable errors of the sex-proportions. 



The probable errors given in the percentage column for the 

 male births are calculated from the usual formula 1 for the 

 probable error of a proportion or percentage, viz: 



P.E. = .67449 



"2 



n 



1 See G. Udny Yule's "An Introduction to the Theory of Statistics," London, 

 1911 (Griffin & Company), for a very clear discussion of this and other probable 

 error formulae. 



