Testing the roles of competition, facilitation and stochasticity on community structure in a species-rich assemblage

1. There is an ongoing debate about whether communities are closely integrated and bound together via interactions such as competition or facilitation, or are disintegrated and dominated by chance. We still lack community-wide data on the intensities of interactions and randomness, and measurements of their impacts on community structure. 2. Using a long-term data set, we sought to measure the effects of interactions and stochasticity in structuring a highly diverse (>100 species) semi-arid grassland plant community, testing for positive and negative interactions at different stages of population growth of all species. 3. During the colonization of new patches, most species were facilitated or inhibited by several others. These opposite effects can potentially have a large effect on species abundances, but they were correlated and cancelled out at the community level. Nevertheless, competition during colonization was strong enough to cause poor competitors to have small population sizes. 4. The subsequent phase of population growth (increase and subsequent change in numbers within occupied patches) was mainly driven by intraspecific density dependence, and we found little evidence for interspecific interactions. 5. Model results showed that stochasticity and recurrent colonization of transient, favourable patches maintained diversity, keeping poor competitors from becoming extinct. 6. Synthesis: Our results, taken together with recent studies on tropical forests, suggest that weak interactions among established plants may be a general phenomenon, but that local interactions during colonization are important drivers of community composition. Most of the variance in species abundance in our community was explained by intraspecific competition and stochasticity, with interspecific interactions playing a minor role due to their overall weakness, interaction changes over ontogeny, and the cancellation of opposite-sign interactions when all the species in the community are considered. Despite this, some species were rare seemingly because they cannot withstand interspecific competition. Thus, to untangle the effects of interactions on community structure, future research should focus on interactions occurring at different phases of population growth and on whole communities.