The rapid increase in adult mortality due to the AIDS epidemic in sub-Saharan Africa raises great concern about potential intergenerational effects on children. This article estimates the impact of AIDS-related adult mortality on primary school attendance in rural Kenya using a panel of 1,266 households surveyed in 1997, 2000, and 2002. The paper distinguishes between effects on boys' and girls' education to understand potential gender differences resulting from adult mortality. We also estimate how adult mortality affects child schooling before as well as after the death occurs. The paper also estimates the importance of households' initial asset levels in influencing the relationship between adult mortality and child school attendance. We find that all of these distinctions are important when estimating the magnitude of the effects of adult mortality on child school attendance. The probability that girls in initially poor households will remain in school prior to the death of a working age adult in the household drops from roughly 88% to 55%. Boys in relatively poor households are less likely than girls to be in school after an adult death. By contrast, we find no clear effects on girls' or boys' education among relatively non-poor households, either before or after the timing of adult mortality in the household. We find a strong correlation between working-age adult mortality in our data and lagged HIV-prevalence rates at nearby sentinel survey sites. The evidence indicates that rising AIDS-related adult mortality in rural Kenya is adversely affecting primary school attendance among the poor. However, these results measure only short-term impacts. Over the longer run, whether school attendance in afflicted household rebounds or deteriorates further is unknown.