THE FACTORS WHICH DETERMINE SEX 687 



tention there is other evidence, as will be shown later, that although 

 a certain chromosome constitution, like a secondary sexual character 

 in an adult, is usually correlated with one sex, this constitution may, 

 imder certain conditions, be overridden and the other sex develop. 



Statistical Investigations on Man. — Statistics of human births 

 have been brought forward in support of the view that the 

 proportion of the sexes varies with the conditions of nutrition. It 

 has been pointed out that in Prance the proportion of births of boys 

 and girls is 104-5 to 100 for the upper classes (who are presumably 

 best nourished), and 115 to 100 for the lower classes (who are more 

 poorly fed). In the " Almanack of Gotha " the proportion recorded is 

 105 boys to 100 girls, while for Russian peasants this proportion 

 is 114 to 100. Among the nobility of Sweden statistics show a 

 proportion of 98 male to 100 female births, but that given for the 

 Swedish clergy is 108'6 boys to 100 girls.^ There is therefore some 

 slight evidence that the percentage of female births is a little higher 

 among those classes which are best nourished or subject to more 

 favourable circumstances, but the differences are very small. 



Punnett^ has examined the statistics collected in the official 

 census of the county of London for the year 1901, with a view to 

 determining the relative proportions of the sexes amongst different 

 classes of society. The following is his summary and conclusion : — 



" If the population of London be divided into three portions 

 exhibiting graduated poverty, it is found that the proportion of male 

 to female infants produced [or rather which have survived] is lowest 

 in the poorest portion, highest in the wealthiest portion, and inter- 

 mediate in the . intermediate portion. The proportion of males is 

 highest of all in a number of births taken from ' Burke's Peerage,' 

 where the nutrition may be supposed to be of the best. From this 

 alternative conclusions may be drawn : that either more favourable 

 conditions of nutrition (1) may result in a large proportion of male 

 births [a conclusion which is contrary to that indicated in the returns 

 mentioned above, but which nevertheless appears to be warranted at 

 first sight], or (2), may have no effect on the proportion of the sexes, 

 or (3) may even result in a relative preponderance of female births, 

 but that in the last two cases the effect is masked by other factors 

 which affect unequally the different strata of society. Such factors 

 are shown to exist in a differential infant mortality, a differential 

 birth-rate, and probably also in a differential marriage-age. These 

 factors all tend to diminish the proportion of males in the poorer 

 portions of the population, and consequently render the first of the 



1 See Morgan, loc. ait. 



2 Punnett, " On Nutrition and Sex-Determination in Man," Proc. Canib. Phil. 

 Soc, vol. xii., 1903. 



