Critique of the demographic evidence for 'late-life non-senescence'.

Although the Gompertz formula accurately describes observed mortality distributions over most of their extent, their 'tail' is much longer than that of a Gompertz curve fitted to the whole data set. A simple candidate explanation is that the longest-lived subset of any population will necessarily be enriched in individuals that age more slowly than the average of that population. However, some investigators have suggested that, instead, individuals actually cease to senesce after a certain age. Here, using a new approach to determining the best-fit degree of heterogeneity in the Gompertz slope parameter, it is shown that observed distributions can in fact be fit quite accurately by purely 'heterogeneous Gompertz' curves. Either explanation may therefore be correct.