This paper considers the impact of exposure to civil conflict on health inputs and outcomes from conception to age five, using the recent Maoist insurgency in Nepal as a case study. Conflict intensity is measured by the number of conflict deaths by district and month and merged with pregnancy histories from the 2001 and 2006 Demographic and Health Surveys. Within-mother estimates show that civil conflict increases the likelihood of miscarriage, so that exposure to conflict in utero has not only a scarring effect but also a selection effect on survivors, most likely due to a combination of maternal stress and malnutrition.