It is hypothesised that hormone levels of both parents at the time of conception affect the probability of a male birth, high levels of estrogen and testosterone increasing this probability, and high levels of gonadotropin decreasing it. This hypothesis satisfactorily accommodates the observed variation of sex ratio with: 1. hormonal induction of ovulation, 2. time of insemination within the cycle, 3. wartime and 4. the distribution of the combinations of the sexes in human dizygotic twin pairs. It has been argued that the hypothesis may also offer a basis for explanations of sex ratio with season, race, maternal smoking, social and psychological factors and with artificial insemination.The most promising approach to an understanding of the determinants of sex ratio is via those variables with which sex ratio varies substantially rather than those with which the variation is slight (even though statistically significant). So there are two powerful methods of testing the hypothesis. The first consists in contrasting observed and expected hormone profiles of those classes of patient who reportedly produce children with substantially biassed sex ratios, e.g., men treated hormonally for subfertility, men with prostatic cancer, women with multiple sclerosis and patients with non-Hodgkin’s lymphoma. If the hormone profiles of these patients are not as predicted by the hypothesis, it will at least need amelioration. I should be glad to hear from workers with data of this sort.The second method of testing the hypothesis is to examine the sex ratios of offspring of those categories of patients who are known to suffer imbalance of the relevant hormones. Patients with many different sorts of disease suffer such imbalance either as a consequence of the disease or its treatment.
James, William H.
"The Human Sex Ratio. Part 2: A Hypothesis and a Program of Research,"
6, Article 3.
Available at: https://digitalcommons.wayne.edu/humbiol/vol59/iss6/3