Testing the Trivers-Willard hypothesis among groups of contrasting kinship systems

Abstract The Trivers-Willard hypothesis predicts that parents in better condition tend to invest in sons, whereas those in less favourable conditions favour investment in daughters. But in humans, it has been argued that this effect might be complicated by the kinship system, through the related norms of inheritance and dispersal. Using a valuable dataset of the matrilineal and patrilineal Pumi and the matrilineal and patrilineal Han of western China, we test whether the Trivers-Willard effects hold across different kinship systems. We collected data on 2,684 adults’ offspring information, annual household income, occupation status and education attainment. Using Generalized Linear Models (GLMs) with a binomial distribution and a logit link function, we examined whether the relationship between offspring sex ratio and parental condition varied between different groups. The results in the Han show partial support for the Trivers-Willard hypothesis, and there is also evidence for kinship systems complicating or obscuring Trivers-Willard effects (in the Pumi). Our findings also indicate that the importance of various parental condition measures might vary across different ecological settings.