Based on observations from all three tropical continents, there is good reason to believe that poor service providers can broadly gain access to payment for environmental services (PES) schemes, and generally become better off from that participation, in both income and non-income terms. However, poverty effects need to be analysed in a conceptual framework looking not only at poor service providers, but also at poor service users and non-participants. Effects on service users are positive if environmental goals are achieved, while those on non-participants can be positive or negative. The various participation filters of a PES scheme contain both pro-poor and anti-poor selection biases. Quantitative welfare effects are bound to remain small-scale, compared to national poverty-alleviation goals. Some pro-poor interventions are possible, but increasing regulations excessively could curb PES efficiency and implementation scale, which could eventually harm the poor. Prime focus of PES should thus remain on the environment, not on poverty. © 2008 Cambridge University Press 2008.