On the probable age structure of the Roman population.

Abstract The average expectation of life has often been calculated from ages given on the many thousands of surviving Roman tombstones. But the distribution of these ages at death is demographically most improbable. However, this can be easily explained once attention is paid to the patterns of commemoration between relatives, for in some inscriptions the age at death is given, in others a relationship (e.g. marriage) is commemorated, in yet others both are recorded. But the distortions cannot be corrected; these ages at death must be discarded as useful evidence for estimating life expectation.