On Nutrition and Sex-determination in Man. 



265 



in group C where the nutrition hypothesis of sex would lead us 

 to expect the largest proportion of $ s we find, on the contrary, 

 the greatest proportion of Js. In other words we arrive at the 

 conclusion that the more favourable the conditions of nutrition, 

 the greater the proportion of </s produced. This conclusion is 

 accentuated by the following piece of evidence. Records of 5225 

 births were made (for purposes referred to below) from Burke's 

 Peerage. Of these 2708 were Js and 2517 $s, thus giving a 

 proportion of 107*6 <fs per 100 $s. And here one is justified in 

 assuming the very best conditions as regards nourishment: — i.e. 

 conditions considerably more favourable than in group C above. 

 This evidence, therefore, considerably strengthens the conclusion 

 arrived at just above. Before, however, such a conclusion can be 

 regarded as having any value several points must be taken into 

 consideration. 



(1) Comparative mortality in £ and % infants. 



It is well known that during the first year or two of their 

 existence </ infants are subject to a higher death-rate than $s. 

 Now in the statistics dealt with here, there is given, not the birth- 

 rate, nor the total number of infants born in a given year, but 

 the number of infants born in a given year, and which remain alive 

 at the end of it. The average age, therefore, of these infants will 

 be about six months ; and in order to obtain the relative numbers 

 of the sexes at birth, the greater mortality among the <f s during 

 this period of six months must be taken into account. Reference 

 to Tatham's English Life-Table (1897), shews that of 1,000,000 

 infants of each sex, 161,036 cfs and 131,126 $s die during the 

 first year, i.e. 29,910 more £ infants die than $s during this 

 period. Hence for six months 14,955 more Js than $s will die, 

 which is approximately 1*5 °/o- Consequently we may roughly 

 emend the figures previously arrived at by adding 1*5 °/ to the 

 percentage of males. In the case of the data from Burke's Peerage 

 no such emendation is required. 



Table II. 



