Seed trapping or a nurse effect? Disentangling the drivers of fine-scale plant species association patterns in a windy environment

Plant co-occurrence patterns can be driven by both abiotic and biotic conditions. Biotic interactions can affect the co-occurrence of seedlings and adult plants through the nurse effect, a form of facilitation where adult plants create favourable conditions for seedlings. However, other mechanisms, including seed trapping, can also cause the aggregation of seedlings around adult plants, resulting in similar co-occurrence patterns as the nurse effect, but this is yet to be examined. This study, therefore, tests if seed trapping drives co-occurrence patterns of two dominant plant species on sub-Antarctic Marion Island; the grass Agrostis magellanica and the cushion plant Azorella selago . We compared A. magellanica seed, seedling and reproductive plant densities around A. selago plants, around similarly sized rocks and on the adjacent substrate. The abundance of A. magellanica seeds and seedlings did not differ significantly between different substrate types or direction (reflecting differing wind exposure). The abundance of mature A . magellanica individuals was significantly related to both substrate and wind exposure, although only weakly so. This suggests that neither cushion plants nor rocks trap seeds and that the high densities of A . magellanica observed with A . selago reflect stage-specific facilitation (i.e. improving the growth and/or survival of established plants). However, this is not consistently the case since complementary seed density datasets demonstrated that under some conditions cushion plants do act as seed traps. This study, therefore, highlights the need to explicitly consider the role of both seed dispersal and biotic interactions in plant communities, particularly before inferring nurse effects based on spatial co-occurrence patterns.