306 



THE AMERICAN NATURALIST [Vol. XLI 



of male and female births in London, in order to determine whether 

 or not the sexual ratio is affected by the nutrition of the parents. 

 For this purpose he divides London society into three groups, 

 following Rowntree's (:02) division of the society of York. These 

 groups are (1) the servant keeping class, (2) the artisan class in 

 which the family earnings are in excess of 26 shillings a week, 

 and (3) the laboring class in which the family earnings fall below 

 26 shillings a week. Rowntree found that, compared with a 

 standard dietary containing 125 grams of proteid and possessing 

 a total energy content of 3500 calories, the first group has a dietary 

 containing more food than is necessary for the maintenance of 

 health; that the second class has, in general, a sufficient diet, 

 although the family must practice strict economy in order to pro- 

 cure it; the third class is, as a rule, seriously underfed, the average 

 deficiency in proteids amounting to as much as 29 per cent. Assum- 

 ing that these considerations apply to I^ondon as well as to York, 

 Punnett finds that there is either no effect upon the sexual ratio 

 which can be attributed to parental nutrition, or, at most only a 

 very small effect. He finds also that the statement of Diising as 

 to the greater proportion of males among the first born children 

 is supported by the statistics of the English lying-in hospitals. 

 Furthermore, mothers whose first birth occurs between the ages 

 of nineteen and twenty-three years bear a larger proportion of 

 males at this birth than mothers whose first birth occurs either 

 earlier or later in life. 



IIL Statistical Data. 



On the fundamental errors in the statistics. — Rauber (:00) has 

 considered the errors in even the best statistics, and only a 

 brief discussion of them will be given here. In order to compute 

 the exact sexual ratio, it is ntH-essarv to obtain statistics of all 

 births, both premature and full term, livinu' (»r still born. The 



homologous or duplicate twins, (lc\ eloped tVoin a sinulc ovum, 

 and invariably of the same sex, should be coiintcil as a single 

 birth. The author has at hand no snfhcicnt data n|)on which to 

 base an idea of the ina^-nitndc of the error which niiuht be uitro- 



