Biased. Low-ranking Rwandan mothers tend to have more girls than boys.

Dave Bartruff/Corbis

Fewer Resources Means Fewer Sons

Desperate to have a baby girl? It helps to be poor. That's the conclusion of a study in Rwanda, which shows that, when men marry multiple women, low-ranking wives are more likely to have daughters than sons. The findings indicate that the social ranking of human mothers can influence the sex of their offspring.

In 1973, evolutionary biologists Robert Trivers and Dan Willard predicted that in many species, social and environmental factors may influence whether a female has more sons or daughters. For instance, in a polygynous species, where males mate with more than one female, males in good condition are at a reproductive advantage over less-fit males because they have more mating partners. According to the Trivers-Willard hypothesis, well-off females in these societies are better off having sons, because sons will have more chances to pass on their parents' genes. However, if moms in polygynous unions don't have many resources to invest, they're better off producing daughters, because only affluent males have multiple wives; daughters will be mated with regardless of status.

Over the years, several studies have supported the Trivers-Willard hypothesis, including work in red deer, mice, and a variety of nonhuman primates. And in humans, studies of Hungarian Roma and mothers in rural Ethiopia have shown evidence of a Trivers-Willard pattern, but others looking into modern Venezuelan society and the Sudanese Bari ethnic group haven't returned the same results.

Thomas Pollet, an evolutionary psychologist at the University of Groningen in the Netherlands, thinks the confusion is partly a result of researchers struggling to determine what constitutes better or worse conditions for mothers. So Pollet and his colleagues looked at a novel maternal situation: the ranking of wives in polygynous Rwandan societies, within which the links between social standing, resource availability, and offspring are easy to tease out.

The team examined census data for 843,000 women in 12 regions in Rwanda and compared the sex ratio of offspring among three groups of women: those in monogamous unions, the first or second wives in polygynous unions, and third wives. The researchers found that, in addition to having fewer children in general, the third wives were 9% to 12% more likely than higher-ranking wives to have surviving infant daughters than sons. The monogamous, first and second wives had more sons than daughters.

Pollet and colleagues believe these results fit with the Trivers-Willard hypothesis. One possible explanation, the authors report today in Biology Letters, is that women in dominant positions, such as first or only wives, have higher levels of testosterone, making them more likely to have sons.

Lena Edlund, an economist at Columbia University who has previously studied how the Trivers-Willard hypothesis fits with social status, says Pollet's methods are sound and that wife rank seems to be a good measure of women's status. One possible mechanism not mentioned in the study, she says, is that lower-ranked wives might suffer nutritionally, which has been shown to influence a child's sex. Low glucose levels, she says, could lead to more daughters, for example.