Age-estimation methods tend to under-estimate advanced age and consequently life expectancy of populations. Many methods use age-scales ending in categories such as “50 years and over” without explaining how the upper limits are decided; the general belief that in early times life was short may play a part. Instances now exist of contemporary written records showing that, even when expectation of life is low, some people become septuagenarians or even older. Most methods of age-estimation, commencing with Gustafson's using dental criteria, predominantly root translucency, have recently been scrutinized by several workers but tooth wear as a method has escaped such scrutiny. A wide range of published findings based on tooth wear is therefore reviewed here. Most age-estimating criteria can only be assessed in terms of ordinal stages, unlike root translucency which can be measured on a continuous scale. Tooth wear ultimately ends in total loss of the dentition but, until that stage, has the potential, not yet fully developed, of also being so measured. The revision is reported of the 1989 estimated age-at-death distribution of a skeletal population. Based on tooth wear and tooth loss, the ages of most over 50 years of age were advanced by about 15 years and reasons are given for regarding four as likely to have been over 75 years of age. Palaeodemography today needs such survivors to be identified, however tentatively.