Snags are standing dead trees that are an important component in the nesting habitat of birds and other species. Although snag availability is believed to limit populations in managed and non-managed forests, little data are available to evaluate the relative effect of stand conditions and management on snag occurrence. We analyzed point sample data from an intensive forest inventory within an 80,000 ha landscape for four major forest types to support the hypotheses that routine low-intensity prescribed fire would increase, and thinning would decrease, snag occurrence. We employed path analysis to define a priori causal relationships to determine the direct and indirect effects of site quality, age, relative stand density index and fire for all forest types and thinning effects for loblolly pine and longleaf pine. Stand age was an important direct effect for loblolly pine, mixed pine-hardwoods and hardwoods, but not for longleaf pine. Snag occurrence in loblolly pine was increased by prescribed fire and decreased by thinning which confirmed our initial hypotheses. Although fire was not important in mixed pine-hardwoods, it was for hardwoods but the relationship depended on site quality. For longleaf pine the relative stand density index was the dominant variable affecting snag occurrence, which increased as the density index decreased. Site quality, age and thinning had significant indirect effects on snag occurrence in longleaf pine through their effects on the density index. Although age is an important condition affecting snag occurrence for most forest types, path analysis revealed that fire and density management practices within certain forest types can also have major beneficial effects, particularly in stands less than 60 years old.