Historically, three assumptions have served as the foundation for understanding the dynamics of mortality and longevity in humans: (z) An increasing probability of death is an inexorable function of advancing age. (2) Female longevity advantage is a universal law of nature. And (3) species possess fixed lifespan limits. However, recent research on the mortality dynamics of several model species has cast doubt on these three canons (Vaupel et al.,1998). Whereas mortality patterns of virtually all species were previously thought to be fixed and intractable, it is now evident that the age trajectories of mortality are dynamic and diverse. These findings have encouraged gerontologists to rethink fundamental concepts concerning aging at the level of the whole organism, including the nature of actuarial senescence (referring to the statistical calculation of life expectancy and related measures based on studies of a population), malefemale differences in mortality, and lifespan limits. Our objectives in this paper are to provide a general overview of the importance, characteristics, and relevance of age-specific mortality to gerontology, to describe new insights derived from large-scale mortality studies in the Mediterranean fruit fly, and to relate the results to understanding longevity and aging in humans.
The most important measure in aging research at the level of the individual is age-specific mortality-the probability of death between two successive birthdays, starting at birth and extending through the last age of the contemporary human life course (greater than rao years). This actuarial parameter is useful in aging research because it is conceptually simple, easily measured, readily modeled, and applicable to all species. It is also useful because it serves as the analytical basis for computing summary measures of longevity such as expectation of life, cohort survival, and life endurancy (age at which go percent of a cohort has died).
The characteristics of the mortality patterns of contemporary humans living in developed societies very with life-cycle stage (Fig. 1). Age and sex (see Fig. 1 inset) account for most of the within-population differences in overall mortality. Newborns experience the highest death rates of youth. However, mortality drops quickly after the first few months of life and reaches the lowest point of the life course at age ii. An 11year-olds probability of dying by her twelfth birthday is less than r in 8,000. Mortality increases at a moderate rate through the early to mid aos, at which point a slight change in trajectory occurs, the details of which depend on sex. Male mortality exceeds female mortality, and the rate of increase is greater, resulting in a slight male "mortality bump." This sex difference can be linked to the high-risk behavior more common in males in the late teens and early 20s than in females of the same age. Between ages 50 and 90, the rate of change in mortality with age first accelerates, then decelerates, forming a bell-shaped curve, quite regular for women, though not for men (Horiuchi and Coale, 1990; Kannisto, 1991). Mortality continues to decelerate from ages 9o to iio, causing a moderate leveling-off and slight decrease in mortality at the most advanced ages (Thatcher, i99a). The cause-of death structure in developed countries like the United States progresses from congenital diseases, accidents, and homicides at younger ages to cancer, heart disease, stroke, and dementia at the older ages (Hoyert, Kochanek, and Murphy,1999).
USE OF MODEL SYSTEMS
IN ACTUARIAL AGING STuDIEs
A stumbling block to the serious use of model systems in studying actuarial aging has been the mistaken belief that, because causes of death in humans are unrelated to causes of death in invertebrates like nematodes (worms) and fruit flies, little can be learned from detailed knowledge of age-specific mortality in these model species. …