Abstract
In a “terra firme” rain forest (Amazonian Ecuador) we quantified the species abundance distribution, gap size, gap isolation, sapling diversity, and gap community similarity for 24 initial gap communities. The pooled community distributions showed a log-series distribution, mainly because 2-D space was partitioned during the forest cycle's first stages, and randomness is accepted as influencing both species arrival and gap formation. Gap size interacted with species richness and evenness due to the low species abundances. The species composition in each gap was an almost unique set: on average, a gap shared <20% of its species with any other gap. A cluster analysis confirmed this observation: the first fusion occurred at a similarity level of 0.42, which related to <27% of common species. Despite this high intergap dissimilarity, association analysis showed that 33.7% of all pairwise gap comparisons represented a significant association, indicating that more species were in common than expected with random assumptions. A principal components analysis revealed three dimensions in the data; sapling diversity, community similarity, and gap size were found to be independent of gap isolation. The uniqueness of the communities in terms of sapling composition highlights the value of gaps for conservation.