Nunn asks whether the four slave trades that drained Africa between 1400 and 1900 help account for the continent's later poverty, and he assembles the data needed to test the claim. Combining shipping records with historical documents that report the ethnic identity of captives, he estimates how many people were taken from each modern country in each century. The countries that lost the most are the poorest now. The obvious objection is selection, that already weak societies supplied the captives, and he answers it twice over. African historians and pre-trade population densities both indicate that the more developed and densely settled regions were drawn in most heavily, which would push the estimate toward zero. He then instruments slave exports with sailing distance to each trade's markets and recovers larger negative coefficients. Turning to mechanisms, he links heavy exports to present ethnic fractionalization and to weak nineteenth-century states, since raiding and kidnapping broke villages apart.