1. Classical and more recent models of secondary succession have been predominantly formulated and tested in temperate zones, and remain rarely evaluated for tropical systems.
2. To test predictions of the relay floristics, initial floristic composition, and nucleation models, and Connell & Slatyer's interaction categories of facilitation, inhibition and tolerance, we monitored demography and mapped coordinates of woody colonists across eleven years from the initiation of succession in five abandoned pastures of southern Costa Rica, which had similar land-use history, climate, elevation and topography.
3. The sites differed five-fold or more in stem density and basal area throughout the study. Floristic similarity among sites was modest at 1.75 years age, and increased slightly by year eleven, due mostly to one species becoming abundant in several sites. In some sites, woody species richness increased after the first inventory, while in others there was little change in the same time period.
4. Temporal dynamics of species abundances were not synchronized, showing no successional stages and offering little support for the concept of relay floristics. To test IFC, in both the ninth and eleventh years we sampled pasture quadrats for small seedlings of canopy tree species; we found very few (0.2 - 0.6 stems m-2), and most were early-successional species. Thus IFC has at most a minor influence on forest succession in these sites. To test nucleation, we characterized colonist densities around remnant trees and compared those densities to colonist densities away from remnant trees. The density of woody colonists around remnant trees was in most cases less than one standard deviation above the mean density around random points, suggesting at most a modest role for nucleation.
5. New seedling recruitment in recent years has been highest in the most densely-colonized sites, suggesting facilitation among cohorts; there was little evidence of establishment inhibition among woody colonists.
Synthesis: Overall, we find little evidence for the relay floristics, initial floristic composition, or nucleation hypotheses, and interactions among the woody colonists are predominantly facilitative. Idiosyncratic successional trajectories in these sites supports an individualistic hypothesis, and explanation should be sought in propagule availability and influences on germination.