It is often argued that leakage of tourism revenue results in minimal economic benefits for host communities, particularly in rural areas of developing countries. However, leakage studies often employ flawed methods, and rarely compare retained revenue with other sources of income to such areas, which can be limited. This paper estimates these values for gorilla tracking tourism at Bwindi Impenetrable National Park, Uganda. Leakage was over 75%, but retained revenue was nevertheless greater than all other sources of revenue to the area combined. It is therefore argued that tourism can be highly significant in the local context despite considerable leakage. Copyright © 2008 John Wiley & Sons, Ltd.