The Trivers-Willard hypothesis joins the ideas of R.A. Fisher and A.J. Bateman to model parental investment strategies. Trivers and Willard argue that any overall investment bias favoring either daughters or sons would be maladaptive. Nevertheless, they suggest that, in effectively polygynous species, more complex, conditional sex biases could be adaptive. In particular, they predict that parents in good condition will bias their investment toward sons and that parents in poor condition will bias their investment toward daughters. Among a sample of approximately 900 U S . mothers we examined several measures of maternal investment including birth weight, interbirth interval and lactational commitment. Maternal condition was assessed by income and by the presence or absence of a coresident adult male. Some measures of investment (five of 14 statistical tests) showed marked and significant sex-by-condition interactions of the type and in the direction predicted by Trivers and Willard; none showed significant effects in the opposite direction. No conscious mediation is required to produce the observed investment patterns.
'The precise equilibrium value will depend not only on the benefits of producing sons versus daughters but also on the relative costs of the two sexes; this complication is unnecessary for our present purposes.