that could have profound influences on characteristics of one or more ensuing generations.
More conventionally, the analysis of life history emphasizes the zygote and subsequent developmental stages through maturation and later postmaturational stages. At later postmaturational ages, mammals typically show declining reproduction and slowly manifest increasing physiological and functional losses that, in association with increasing mortality risk, define the phase of senescence.
One measure of the rate of senescence is the rate of increase of age-specific mortality, m(t). In many populations, m(t) increases exponentially after maturity according to the Gompertz formula: m( t) = exp[α(t)] (Finch and Pike, 1996; Finch et al., 1990; Mueller et al., 1995). A convenient basis for comparing mortality rate accelerations is the time required for mortality rates to double (MRDT = In 2/α), which is about 8 years in humans and about 0.3 year in laboratory rodents (Finch et al., 1990).
A huge body of work demonstrates many ways in which the phenomena of senescence are highly plastic and subject to modifications through environmental parameters, for example diet, as discussed below. This plasticity challenges traditional beliefs that the life spans of higher organisms are rigidly preprogrammed by their genes. I use the plural, life spans, to emphasize that there may be many statistically distinct life-history trajectories within a given human population, which are subject to myriad gene-environment interactions, including lifestyle choices (Finch and Tanzi, 1997). The plasticity of life histories is generally consistent with an evolutionary basis for the numeric life span as a life-history trait. The vast range of schedules shown by multicellular organisms, as described next, implies that the plasticity in life-history schedules and phenotypes is itself a general outcome of evolution by allowing multiple alternative adaptive schedules.
The life expectancy of an individual in a population is not constrained by any known intrinsic feature of aging at the molecular or cellular level that is not open to evolution. As discussed by Tuljapurkar and by Rose (in this volume), for a species to survive, at least one of its populations must, on the average, maintain non-negative growth. This statistical outcome is achieved by balancing cumulative survival across adult ages by cumulative fecundity. The Euler-Lotka ''equation of state" for population dynamics does not lead to any predictions about the duration of the developmental stages that precede reproduction. Although time is a necessary dimension in the parameters that are used to describe population growth, evolutionary biologists generally conclude that the magnitude of duration for any life-history stage is free to increase without constraint and that there is no biological limit to the maximum life span.
Nonetheless, phenomena of senescence leading to characteristic life spans in a population are expected to be the norm. Rose, Partridge, and Tuljapurkar (in