The basic finding is that mortality decelerates at older ages. For some species, such as humans, death rates keep on going up with age up to advanced ages but the rate of increase slows down. For other species, such as medflies, death rates reach a plateau and then fall dramatically. Why mortality decelerates at older ages, in all the species for which large, careful studies have been conducted, is a puzzle. The chapter concludes with an explanation of why this is a puzzle and what might be the key to an answer.
The research reported here is based on collaborative efforts by scientists from different backgrounds. Hence, the chapter constitutes a case study of the topic I was asked to address—namely, "demographers, ecologists, and evolutionary biologists: what can we learn from each other?"
Medvedev (1990) reviews more than 300 theories of aging; most could be used to derive testable predictions about the trajectory of mortality at advanced ages. These theories, however, are, for the most part, simply sets of ad hoc assertions derived from or consistent with a few observations and largely based on strongly held prior beliefs. For example, Buffon, the eighteenth century French naturalist, hypothesized that each species has a characteristic maximum life span and that this maximum life span is six or seven times the duration of the period of growth. A long stream of variants on this theme culminated with Fries (1980), who popularized the conjecture that all individuals are born with a genetically determined maximum life span. Individuals who do not die prematurely die of senescence shortly before their fixed allotment of life.
In the 1950s and 1960s, Medawar (1952), Williams (1957), and Hamilton (1966) developed a more cogent theory, based on evolutionary arguments, of why mortality increases at adult ages. Partridge (in this volume) and Rose (1991) review the development of this theory. Charlesworth (1994) summarized it as follows: at older ages "mutation pressure overpowers selection," leading to "mutational collapse." The basic idea is that there will be little or no pressure of evolutionary selection against mutations that have detrimental effects at older ages but neutral or positive effects at younger ages. The elderly have little impact on evolution because only a fraction of a cohort live to old age and few, if any, of them produce offspring. For humans and other species that care for their children and grandchildren, the elderly may have some evolutionary role, but this role diminishes with advancing age. The prediction of this line of thinking is that the age-trajectory of mortality should shoot up at postreproductive ages, especially for species in which parents do not help their progeny (Curtsinger, 1995b).
Until recently, the evolutionary theory of aging largely rested on theoretically plausible arguments that had not yet been empirically tested; Rose (1991) reviews the limited empirical results. Although there have been some important recent experiments (such as Hughes and Charlesworth, 1994: Charlesworth and