parental investment decisions into fertility. The key to this system is that maximizing lifetime-expected resource production through optimal allocation of activities and wealth flows will tend also to maximize fitness when all wealth is in the form of food and extra food translates into higher fertility.
There have been several empirical applications of optimality models, designed to determine whether the onset and termination of reproduction and the size of interbirth intervals actually maximize fitness. The results of those analyses have been mixed. Blurton Jones et al. (1989) analyzed the relationship of the birth interval following a child who survived until the birth of its next sibling to the survival of either member of the sibling pair among the !Kung (Jo’hansi) in order to determine whether the observed mean interval of 48 months maximized the total number of surviving offspring. This analysis showed that 48-month birth intervals were, in fact, optimal (see Harpending, 1994; Pennington, 2001, for a critique).
In contrast, similar analyses among the Ache showed that observed birth intervals were longer than optimal (Hill and Hurtado, 1996). Phenotypic covariation, however, poses a major problem in those analyses. If healthier women have shorter birth intervals than less healthy women because they have larger effective energy resources, the estimated effect of birth intervals on offspring survival would be downwardly biased. In fact, studies examining natural variation among nonhuman organisms often show a positive, rather than a negative, relationship between fertility rate and offspring survival for the same reason (Partridge and Harvey, 1985). When fertility rate is experimentally manipulated among those same organisms, the relationship is reversed, as would be expected by a quantity-quality trade-off. Consistent with the self-selection hypothesis, Gambian women who had higher hemoglobin levels following the birth of a child exhibited both shorter interbirth intervals and higher child survival (Sear et al., 2001). Thus, it may be that women’s physiology tracks its own condition in such a way as to maximize their individual fitness.
Hill and Hurtado (1996) also examined age of first reproduction and menopause to test optimality models. They found that women began reproducing at the optimal weight to maximize their lifetime fitness. However, they did not find support for the proposition that menopause maximizes fitness, as expected by the “grandmother hypothesis” (Williams, 1957). The effects of grandmothers on the fertility of their children and on the survival of their grandchildren were not large enough to overcome the fitness costs of reproductive cessation. In a parallel “contrary-to-prediction” analysis, Rogers (1993) found that the expected reproductive success of older women would have to be implausibly low to favor menopause. However, as Hill and Hurtado (1991, 1996) point out, self-selection also complicates this analysis. Grandmother’s death is used to assess the impacts