198 



RAYMOND PEARL AND MAUD DEWITT PEARL. 



frequency. It has been shown ^ that if j'^ be any class frequency 

 within a sample containing 7n individuals altogether, then 



^- ^■ys= -^7449 aJj'.s{' -'jf^ 



From the "absolute" probable error so obtained the probable 

 error of the sex-ratio is easily deduced. 



Table II. 

 Males to 100 Females from Totals of Table I. 



Mating. 



Argentine^ Argentine? 

 Italian ^ Italian 9 

 Spanish ^ Spanish j 



Sex Ratio. 



103.26 ±.34 

 100.77 =b .20 

 105.55^-36 



Mating. 



Italian ^ Argentine 9 

 Spanish ^ Argentine 9 



Sex Ratio- 



105.72 

 106.69 



.46 



•74 



From this table the following points are to be noted : 



1. The number of males to 100 females varies between ap- 

 proximately loi and 107 in the different matings. There is an 

 excess of males in every case. Further, this excess is significant 

 in amount as is indicated by the probable errors. The present 

 statistics agree with other large collections of data regarding the 

 sex-ratio of human births. There appears to be no doubt that a 

 tendency towards the production of a greater number of males 

 than of females is normal for Caucasian races at least. 



2. The sex-ratio is in each case higher for the cross matings 

 than for the pure. That is, there are more males per hundred 

 females produced when the parents are of different racial stocks 

 than when they are of the same. 



The answer to the first question propounded above (p. 196) 

 then is that there is a definite tendency towards an excessive pro- 

 duction of male offspring in cross as compared with pure matings 

 in the data here considered. Further, it appears that within the 

 limits of the present material this tendency is uniformly exhibited 

 in all the matings. 



Attention may next be turned to the second part of the 

 second question, which may now be put as follows : 



' Editorial 

 II., p. 274. 



The Probable Errors of Frequency Constants," Bto?netrika, Vol. 



