AN ANALYSIS OF GEISSLER'S DATA ON THE HUMAN SEX RATIO

Abstract

Summary Skellam's modified binomial distribution is described, and fitted to Geissler's data on the sex ratio in human families of from 2 to 12 children. It is concluded that p , the probability of a birth being male, varies between families of the same size, having a variance of about 0–0025 amongst a complete cross‐section of a nineteenth‐century German population, and that there is no evidence for the existence of parents only capable of producing unisexual families. Assuming that p varies within a family, the apparent anomalies would appear to be explicable. The critical conclusions of Lancaster regarding the reliability of Geissler's data are briefly examined. In an appendix, the derivation of the posterior probability distribution of p for a couple with a known number of boys and girls is given.