Cattle influences gap dynamics in pastures in two ways: (1) by creating gaps and (2) by affecting the colonization process. This effect of cattle activity on gap revegetation can be subdivided in three main factors: herbage removal, trampling and dung and urine deposition. The objective of this study was to assess how these three effects moderate the plant succession following gap creation. In an exclosure, four controlled treatments simulating cattle activity (repeated mowing, trampling, manuring and untreated control) were applied on plots of 2 X 2 in. In the centre of each plot, one artificial gap of 60 X 60 cm was created. During three years, vegetation changes were monitored in spring and in autumn, with 2 a square grid of 100 cells of 0.01 m(2) centred on the gap. Our experiment confirmed that fine-scale gap creation may have a high impact on relative abundances of species in the community. The gap environment acts on species as a filter and this filtering was described in terms of regenerative attributes. Colonizers were species with small seeds, unspecialized seed dispersal, persistent seed bank and high vegetation spread. However, the role of dung deposition, herbage removal or trampling by cattle did not seem to be of primary importance in the revegetation process, but could moderate vegetation response. Therefore, the different cattle effects act as secondary filters that selectively favoured or disadvantaged different species from the gap-regenerating community. These complex interactions are probably keys to understand plant coexistence in perennial grasslands.