Annual early fire, selective tree cutting and exclusion of grazing are currently used as management tools in the Sudanian savanna of Burkina Faso although their longterm effects on seedling recruitment are poorly documented. A factorial experiment involving fire, grazing and cutting, each with two levels, was established in 1992 to study the effects of these management regimes and their interactions on the regeneration of woody species, and examine whether their effects varied temporally. Species richness, density and the morphology of seedlings were assessed in 1997 and 2002, and their relative changes were determined. The change in species richness of multistemmed individuals was significantly higher (P = 0.018) on plots that received fire · cutting treatment than the control plots. Significantly more species with single-stem were found on unburnt than on burnt plots (P < 0.001). Grazing tended to reduce the change in total density while fire (P < 0.001) and grazing (P = 0.029) significantly reduced the change in density of single-stemmed individuals. Selective cutting did not affect the total seedling density, but tended to reduce the change in single-stemmed seedling density. Principal component analysis revealed species-specific responses to treatments, particularly the relatively high abundance of lianas compared with other species.