The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
cases where velocities approach the speed of light or very high masses are involved. Normally, Newtonian theory "works." But we know now that this is because it is only a first-order approximation to the truth, which is better approached by relativistic mechanics.
The strategy that Laurence D. Mueller and I have adopted is to look for a theory that can naturally, without forcing, generate phenomena like plateaus in mortality rates among the oldest old. This theory is thus the analog of relativistic mechanics in physics. We feel that we have taken the first significant steps toward developing such theory, but I do not have the opportunity to introduce this theory for evolutionary demography here.
Abrams and Ludwig (1995) have recently published an optimality approach to the kind of the theory that Mueller and I have tried to develop using population genetics. As a population geneticist, I have long had grave reservations about the use of optimality models for biological situations that are not defined by "free choices," such as those of sex allocation (Charnov, 1982), for example, in which there seems little reason to doubt that an optimal phenotype can evolve. Life-history evolution, instead, involves material resources and genetic constraints that may entirely prevent the attainment of any optimal phenotype (cf. Gould and Lewontin, 1979), and it is at least arguable that the evolution of actual life histories does not proceed toward optimal outcomes (Rose et al., 1987). For example, is it likely that the later life history will evolve toward optimal values if the allelic variation that shapes it is subject to very weak natural selection or none at all, given recurrent mutation pressure? For these reasons, we have proceeded instead to develop a population genetics theory for demography.
Recent work on age-specific mortality rates strongly suggests that conventional demographic models are in need of repair. Rather than resorting to yet another ad hoc tuning up of the same mathematical tools, consider the alternative approach provided by the population genetics theory of age-structured populations. This theory has been well-characterized mathematically and has been extensively supported in experimental genetic systems, although neither theory nor experiment are complete or perfect. Results obtained from evolutionary theory offer the best prospects for the development of demographic models, in general. In this way, demography could leave its ad hoc traditions behind and join together with evolutionary biology to forge a much stronger theoretical foundation. While the Gompertz model is an excellent approximation to mortality rates for most individuals from some iteroparous populations, it appears inadequate for the description of even the qualitative pattern of the full life history of most organisms, granting that such full life histories are not usually adequately observed. Biodemography should now move on to the task of developing proper evolutionary foundations.