any case, is that the proper ground for building theories of demography, aging, life history, and the like is evolutionary theory.
The fundamental starting point for evolutionary analysis is an ecologically determined demography: particular age-specific survival probabilities and fecundities, varying also with time, organismal size, population density, and a range of other factors (Charlesworth, 1994). Frequently, however, factors other than age are neglected in the theory, and I will follow this practice for the sake of convenience and clarity. Some of the literature on the other characters (e.g., size) is discussed in Steams (1992) and Roff (1992). For the present purposes, however, evolutionary demography starts from a simplified evolutionary situation defined by a set of age-dependent survival probabilities, P(x) and fecundities, m(x).
When density-dependence can be neglected (cf. Charlesworth, 1994), population growth rate is given by the largest real-valued root, r, of
where l(x) = Πy=0 P(y) and x is age as well as the upper limit of multiplication. This ecological situation, in turn, determines the evolution of P(x) and m(x), where the equations for gene frequency change have terms weighted by
for selection coefficients involving In P(a) and m(a), with genetic effects at age a, respectively (Hamilton, 1966; Charlesworth, 1980; Rose, 1985). The s and s functions play a scaling role. When the scaling functions are large, selection is more powerful, and conversely, when these values are small, selection is weak.
The forms of the s and s' scaling functions reveal the most important features of evolutionary demography in its crudest form. The form of s(a) is such that it remains at 1.0 for ages below the start of reproduction. But for ages after the start of reproduction, s(a) falls with age until it approaches zero. In terms of selection equations, this makes allelic effects on early survival of large impact on the evolutionary outcome, but allelic effects on late survival have little or no impact. Except for populations declining rapidly to extinction, similar patterns apply for s'(a), which will usually be large at early ages and small at late ages. Put simply. this situation is summarized as "the force of natural selection tends to fall with adult age" (Medawar, 1952). This basic idea is hinted at in the classic writings of R.A. Fisher and J.B.S. Haldane and then broadly sketched by P.B. Medawar (e.g., 1952). However, it wasn't adequately developed mathematically until the work of Hamilton (1966) and Charlesworth (e.g., 1980), as outlined above. The most obvious corollary derivable from this situation is that components of fitness should deteriorate at later adult ages, due to the weakness of natural selection,