The demographic impacts of harvesting nontimber forest products (NTFP) have been increasingly studied because of reports of potentially unsustainable harvest. Nevertheless, our understanding of how plant demographic response to harvest is altered by variation in ecological conditions, which is critical for developing realistic sustainable-use plans, is limited. We built matrix population models to test whether and how variation in ecological conditions affects population responses to harvest. In particular, we examined the effect of bark and foliage harvest on the demography of populations of African mahogany (Khaya senegalensis) in two contrasting ecological regions of Benin, West Africa. K. senegalensis bark and foliage harvest significantly reduced its stochastic population growth rates, but ecological differences between regions had a greater effect on population growth rates than did harvest. The effect of harvest on population growth rates (Deltalambda) was slightly stronger in the moist than in the drier region. Life-table response experiments revealed that the mechanism by which harvesting reduced lambda differed between ecological regions. Lowered stasis (persistence) of larger life stages lead to a reduction in lambda in the drier region, whereas lowered growth of all life stages lowered lambda in moist region. Potential strategies to increase population growth rates should include decreasing the proportion of individuals harvested, promoting harvester-owned plantations of African mahogany, and increasing survival and growth by promoting no-fire zones in gallery forests. Our results show how population responses to harvest of NTFP may be altered by ecological differences across sites and emphasize the importance of monitoring populations over the climatic range in which they occur to develop more realistic recommendations for conservation.