Two economists ask why the map of African ethnic groups is so finely divided, and they answer that the slave trade helped divide it. Whatley and Gillezeau place two hundred evenly spaced points along the western coast, count the ethnic groups recorded in modern ethnolinguistic atlases inside circular zones drawn around each point, and compare those counts with the number of captives shipped from ports in the same zones. Ordinary least squares and instrumental variable estimates, the latter using distance to American and North African slaving destinations, both find more groups where more people were carried away. The mechanism they propose is economic. Once a captive was worth more than a taxable farmer, rulers had reason to raid instead of building wide states, to sharpen the boundary between insider and outsider, and to skip alliances with neighbouring villages. Ethnic fragmentation thus reads as a legacy of slaving, and controlling for it hides part of slavery's cost.