This study examines the role of marriage form in infant mortality and tests the relative effects and mechanisms through which polygyny affects infant survival. A sample of infants born in the 5 years preceding the 1988 Ghana Demographic and Health Survey was used. A causal model was constructed and polygyny was found to be an important mediating factor in understanding infant survival. The findings highlight the relevance of family structure (polygyny) as an intervening factor, but also reveal the complex role of ethnicity, dietary supplement and birth interval in accounting for infant survival.