Historians have frequently suggested that droughts helped facilitate the African slave trade. By introducing a previously unused dataset on 19th century rainfall levels in Africa, I provide the first empirical answer to this hypothesis. I show that negative rainfall shocks and long-run shifts in the mean level of rainfall increased the number of slaves exported from a given region and may have had a persistent impact on the level of development today. Using geocoded data on 19th century African conflicts, I show that these drought conditions also increased the likelihood of conflict, but only in the slave exporting regions of Africa. I also explore the role of household desperation, the internal African slave market, and disease outbreaks in explaining the negative relationship between droughts and slave exports. I find limited evidence for for these alternative mechanisms, with household desperation having the most empirical support. These results contribute to our understanding of the process of selection into the African slave trade.JEL Classification: N37, N57, O15, Q54