In his extreme male brain theory of autism, Baron-Cohen postulates that having a typically male brain was adaptive for ancestral men and having a typically female brain was adaptive for ancestral women. He also suggests that brain types are substantially heritable. These postulates, combined with the insight from the Trivers-Willard hypothesis regarding parental ability to vary offspring sex ratio, lead to the prediction that people who have strong male brains should have more sons than daughters, and people who have strong female brains should have more daughters than sons. The analysis of the 1994 US General Social Survey data provides support for this prediction. Our results suggest potentially fruitful extensions of both Baron-Cohen's theory and the Trivers-Willard hypothesis.