Seedling recruitment limitation could be an important influence on forest community structure. Typically tree seed or seedling limitation has been assessed at spatial scales ≤ 1 m², which is much smaller than the spatial scale of mature trees that the seedlings would replace. We tested effects of spatial scale on seedling prevalence, based on >19k newly germinated seedlings of 106 more common woody species (N³10 seedlings in at least one transect) that have been censused at 6 week intervals over 5.5 years in five 200 m belt transects in a Costa Rican wet tropical forest. Species prevalence was defined as the proportion of sampling units having ³1 seedling, which is equivalent to (1- fundamental seedling limitation). In particular, we compared species prevalence based on occupancy of 1 m² sampling units versus occupancy of sampling units that combined adjacent quadrats to approximate crown diameters of woody plants per five lifeforms: understory plants (2 m crown diameter), treelets (8 m), subcanopy (15 m) , canopy trees (20 m) and lianas (20 m).

Results/Conclusions

At the 1 m scale, most species were present in <15% of sampling units, which is consistent with previous findings for more common species in tropical forests. However, at spatial scales commensurate with the crown diameter of mature individuals, most species were represented in >60% of sampling units. Over the 5.5 year period, large gaps in species occupancy along the belt transect were rare, which mitigated fundamental seedling limitation at spatial scales commensurate with organism size. For example, >85% of gaps in species occupancy (i.e., where a seedling of a species was not present over the 5.5 year period), were ≤ 10 m in length. These findings support that extrapolating from fine spatial scales exaggerates the degree of seedling limitation in more common species. However, <10% of the 106 species had at least 10 seedlings at all 5 sites, suggesting seedling limitation for most species at the landscape level.