D. Heron 
85 
Equating these respectively to '589 and "456, the sex-ratios for paternal and 
maternal sibships, we fiud : 
11 = 1-62 and s = -526. 
The latter value is precisely the value found for all sibships of the parental 
generation. The former should represent the average number in a sibship of 
the parental generation. It cannot be very far from its true value, because all 
sibships without at least one male (or it may be one female) have been by the 
nature of the case excluded, and further, no sibship has been used with fewer 
tlian four members. It will thus be seen that (jur human data ai"e in good 
accordance with each other. 
So far as we can judge, in the second generation of thoroughbred horses under 
consideration there was a preponderance of mares born, the sex-ratio being '478, 
and differing from '500 by at least seven times the probable error. In the first 
generation, since there must be one filly in the produce at least, we have 
(?i-l)s/u = '463, 
and if n lie between 10 and 15 as it does, this gives s = •5 within the probable 
error. In other words, the sex-ratio between the two generations appears t(j have 
fallen from equality to about "48, a substantial alteration. 
Turning now to the main portion of the present enquiry, we have : 
TABLE VII. 
Con-elation between Sex-Ratios in Successive Generations. 
Nature of Sibships 
Correlation 
Sibships of Father and Offspring, Man 
Sibships of Mother and Offspring, Man 
Sibships of Parent and Offspring, Man... 
Joint Parental Sibship and Offspring, Man 
Produce of Mother and Daughter, Thoroughbred Horse 
•053 + -020 
•001 + ^021 
•021 + •OU 
•043 + -036 
•034+ -021 
It is true that all these correlations are positive, but not one of them is 
definitely significant, having regard to its probable error. Thus on rather wider 
data — in horse as well as in man — Dr Woods' position is confirmed ; there is 
no inheritance, or at least no sensible inheritance, of sex. The persistent and 
sensible differences from •o which occur in various races for the sex-ratio are 
therefore not racial in tlie sense that they are an inherited characteristic of 
the race; they must be in some manner associated with environment, nutrition, or 
habit. They appear to be a more universal, if less marked, result of such 
causes as lead certain species which usually reproduce parthenogenitively to 
occasionally reproduce sexually. It is conceivable that the sex-ratio of produce 
may not exhaust all the characters associated with an individual which are not 
subject to the general rule of inheritance. 
