Disentangling the mechanisms underlying the species–area relationship: A mesocosm experiment with annual plants

Abstract The increase in species richness with increased island area is one of the most widely documented patterns in ecology and biogeography but its mechanisms are still under debate. Disentangling these mechanisms in natural systems is challenging due to various kinds of confounding effects. Here we use a novel mesocosm experiment focusing on annual plants to disentangle the effects of two mechanisms that may lead to higher richness on large islands: increasing population sizes and increasing opportunities for within‐island dispersal. Theoretical studies show that both mechanisms may contribute to higher richness on large islands, but no previous study has attempted to separate their effects. We also test an alternative, ‘null’ hypothesis, according to which the only mechanism underlying the increase in richness with increased area is a pure sampling effect (the ‘passive sampling hypothesis’). As expected, increasing area had a strong positive effect on species richness. However, the main mechanism underlying the difference in richness between small and large islands was passive sampling. Moreover, our results indicate that dispersal had a negative rather than positive effect on island richness. Synthesis . Our results corroborate previous observational studies that failed to reject the passive sampling hypothesis. However, in contrast to previous studies, our findings are based on experimental manipulations of island area, under controlled uniform conditions, at small spatial scales that facilitate within‐island dispersal and in three independent systems that differed in resource availability and disturbance. These findings emphasize the need for a careful examination of sampling effects in future research of the species–area relationship.