84 On the Inheritance of the Sex-ratio 
her offspring, and is, perhaps, more satisfactory than the human determination. 
Reducing this material, we obtained the following results : 
TABLE VI. 
Mean and Standard Deviation of Sex-Ratios. 
Group 
Mean 
Standard 
Deviation 
1 
Father's Sibship, Man ... 
•589 + -004 
•178 + ^002 
2 
Mother's Sibship, Man ... 
■4.56 + -00.3 
•167 + ^002 
3 
All Parental Sibships, Man 
•526 + -003 
•185 + ^002 
h 
Filial Sibship, Table I., Man 
•522 + -004 
•208 + ^003 
5 
Filial Sibship, Table II., Man 
•520 +^005 
•218 + •003 
6 
Filial Sibsliips, Table III., Man 
•521 + -003 
•210 ±-002 
7 
Filial Sibships, Table IV., Man 
•504 + -007 
•193+ 005 
8 
Joint Parental Sibships, Table IV., Man 
•521 + -005 
•130 + ^003 
9 
Mare's Prodnce, Mother ... 
•463 + •OOS 
•148 + ^002 
10 
Mare's Produce, Daughter 
•478 +^003 
•151 + ^002 
Now, if we examine this table, we cannot in the case of thoroughbred honses 
assert that any difterence exists in the variability of the sex-ratio for the two 
genei'ations. But in the case of man there certainly is a significant difiference 
in the variability. While there is no significance in the difference of the varia- 
bilities denoted by the row numbers 4, 5, and 6, and pos.sibly not in 7, there 
is a difference more than six times the probable error of the difference between 
these variabilities and that of 3. There is, however, no difference in type between 
3, 4, 5, 6, 8, and possibly, but not certainly, 7 *. These figures demonstrate the 
point referred to above, that in the free mating of man, families with a preponder- 
ance of female or male elements are not drawn upon equally with families in 
which the sexes are more equally balanced. In the controlled mating of horses 
this result is not apparent. 
We have already noted that in the sibships which are not selected so as to 
have at least one male or one female, the type is fairly constant and gives a sex- 
ratio of about '522, which coi'responds to 109 male births as compared with 100 
female births, a quite good result. We next ask how does this agree with the 
values found for sibships which must have at least one male or female ? Let n be 
the average number in a sibship, and s be the sex-ratio. Then if we choose 
sibships in which there is at least one male, we might expect the sex-ratio to be 
and that for sibships with at least one female to be 
{n — 1) sjn. 
* The data for 7 include a Cornish fishing village where the sex-ratio is far more nearly one of 
equality than elsewhere in this country; owing to the persistence of large families in this district, it 
therefore figures disproportionately in the results. 
