ship and later fertility, as well as ameliorated pathophysiology. In effect, the Virginia opossums have been a reasonable "natural experiment," like those in the laboratory with Drosophila.
But the force of natural selection does not entirely determine the pattern of demographic evolution. An additional determinant is the nature of the genetic mechanisms that affect particular life-history characters. These genetic mechanisms can be grouped under two broad headings: nonpleiotropic genetic effects and pleiotropic genetic effects. I discuss each further in turn.
The most elegant possibility for the genetic mechanisms of demographic evolution is for each age-specific life-history character to evolve on its own, with the alleles that affect it lacking in effects on any other such character. It should be noted that this is a severe assumption, which is being offered here primarily for its simplicity. In particular, I do not wish to convey the impression that I think that this is the normal evolutionary situation where aging is concerned. At the same time, this severe constraint on allele action makes the mathematical specification of the evolutionary dynamics of life history, demography, etc. extremely elegant. Essentially, the entire evolutionary process reduces to two components: age-specific mutation and age-specific selection. In particular, because few beneficial alleles will be segregating at high frequencies at any one time, the evolutionary dynamics become mutation-selection balance, where there would typically be one favored allele and one or more deleterious alleles. For simplicity, the favored and deleterious alleles can be grouped into two allelic classes, and the frequency of the deleterious allele, when it is of minor effect and not fully recessive, can be approximated as
where u is mutation rate from beneficial to deleterious alleles and the ij subscripts refer to genotypes having alleles i and j at the locus in question (Charlesworth, 1994). The beneficial allele is indicated by the numeral 1, and the deleterious allele is indicated by 2. Note that since the demography of the population will be determined largely by the homozygote of the fitter allele, the scaling function is specified by that demography. As discussed above, the scaling function is going to fall with age after the onset of reproduction. And because the scaling function is in the denominator of the allele-frequency formula, the frequency of deleterious alleles will theoretically increase enormously at later ages, producing a deterioration in survival probability. At early ages, of course, selection mutation should keep the frequency of deleterious alleles low, and thus age-specific survival would be high. An additional expectation, with this population genetic mechanism, is that the additive genetic variance for life-history characters should