Abstract Density dependence is of fundamental importance for population and range dynamics. Density‐dependent reproduction of plants arises from competitive and facilitative plant–plant interactions that can be pollination independent or pollination mediated. In small and sparse populations, conspecific density dependence often turns from negative to positive and causes Allee effects. Reproduction may also increase with heterospecific density (community‐level Allee effect), but the underlying mechanisms are poorly understood and the consequences for community dynamics can be complex. Allee effects have crucial consequences for the conservation of declining species, but also the dynamics of range edge populations. In invasive species, Allee effects may slow or stop range expansion. Observational studies in natural plant communities cannot distinguish whether reproduction is limited by pollination‐mediated interactions among plants or by other neighbourhood effects (e.g. competition for abiotic resources). Even experimental pollen supply cannot distinguish whether variation in reproduction is caused by direct density effects or by plant traits correlated with density. Finally, it is unknown over which spatial scales pollination‐mediated interactions occur. To circumvent these problems, we introduce a comprehensive experimental and analytical framework which simultaneously (1) manipulates pollen availability and quality by hand pollination and pollinator exclusion, (2) manipulates neighbourhoods by transplanting target plants, and (3) analyses the effects of con‐ and heterospecific neighbourhoods on reproduction with spatially explicit trait‐based neighbourhood models. Applying this framework to Senecio inaequidens, one of Europe's fastest plant invaders, we found that the seed set was strongly pollen limited. Reproduction had increased by pollinator‐mediated facilitation by both con‐ and heterospecific neighbours which may lead to (community‐level) Allee effects. Pollination‐independent interactions, such as amelioration of abiotic conditions through neighbours, contributed to additional positive neighbour effects. However, these pollination‐independent interactions were weaker than the pollination‐mediated interactions and they occurred over smaller spatial scales. Finally, the strength and direction of neighbourhood effects depended on neighbour traits and thus changed with the trait composition of the neighbourhood. Synthesis. By manipulating both pollen availability and target plant locations within neighbourhoods, we can comprehensively analyse spatially explicit density dependence of plant reproduction. This experimental approach enhances our ability to understand the dynamics of sparse populations and of species geographical ranges.
Read full abstract