The elderly are one of the key groups of survivors who may be adversely affected by the death of prime-age adults from AIDS. We use a longitudinal survey of households from Northwestern Tanzania in 1991–94 to compare the activities and welfare of the elderly in households before and after the death of a prime-age adult with that of the elderly in households that did not have an adult death. The elderly in households that had an adult death were more educated, more likely to engage in wage employment and spent less time farming than the elderly in households that did not have a death during the survey. Time spent by the elderly in household chores rose following an adult death and participation in wage employment fell; there was no evidence of increased participation in farm work. The physical well-being of the elderly, as measured by body mass index (BMI), was lower prior to an adult death and higher afterward. Finally, the elderly with the lowest BMI are those in poor households that did not have an adult death during the survey. Thus, policymakers should be concerned about the adverse impacts of adult deaths on the physical well-being of the elderly—primarily during the period of illness prior to a prime-age adult death—but they should also focus on the larger groups of poor elderly with much lower physical health status.