Conservationists have often assumed that the least disturbed habitats harbor the greatest biodiversity, and that human presence is therefore uniformly a subtraction from nature's richness. Studies of certain long-inhabited landscapes trouble this assumption. In some Amazonian forests, the concentration of useful tree species is highest precisely where pre-Columbian peoples once lived and cultivated, suggesting that centuries of human management enriched rather than depleted the local flora. The point is not that all human activity benefits ecosystems — much of it plainly does not — but that the equation of pristineness with abundance rests on a picture of humans as necessarily external to the systems they inhabit.