Résumé
In situ conservation often requires the designation of sites where land-use is restricted, such as protected areas and no-fishing zones. Such areas are designed to reduce human impacts on the ecosystem, but the overall benefits of this approach might be compromised if 'leakage' takes place - that is, if impacts that would take place inside the restricted area are displaced to a nearby, unrestricted area. Recently, Oliveira and colleagues became the first group to measure leakage from newly created forest concessions. They showed that restricting land-use reduced deforestation within the concession areas, but dramatically increased it in the surrounding areas. We discuss these findings in the wider context of growing global interest in quantifying the effectiveness of nature reserves.