Forests can provide refuge, funds, and food for combatants in civil war. Insurgents may also use forested regions to hide from government troops and the government may choose to ignore them if they remain in remote forested regions. Case studies of Burma, Cambodia, and other countries show that forest resources have served to fuel armed conflict. Several scholars and NGOs have portrayed this as a world-wide problem, but there are hardly any relevant statistical studies. This article first develops theoretically the different mechanisms relating forest resources to conflict. It then looks at the empirical relationship for the onset and duration of conflict at the country level and finds very little support for a general and direct relationship. At this level, it would appear that many have generalized too hastily from the case studies. The analysis then moves to the disaggregated level and looks at the duration of conflict when forest resources are available in the conflict zone. Again, no general relationship is found. We find that a shorter distance to the coast tends to make the conflicts in forested conflict zones longer. On the other hand many of the significant results we find are driven by one or two cases. While several generalizations from case studies seem overly ambitious, a more careful modeling of the mechanisms revealed by the cases finds some support in the statistical study.