Search for a command to run...
The spatial distribution of seeds can influence several parameters of the natural regeneration of plant populations. Factors shaping seedfall patterns have been typically explored from the tree perspective (seed shadow) or from the population perspective (seed rain). However, the seed rain is actually composed of multiple seed shadows. In this paper, we use this framework to explore the variables shaping the seedfall pattern of the wild olive tree (Olea europaea var. sylvestris), a fleshy-fruited tree of the Mediterranean scrublands. We monitored the movement patterns of avian seed dispersers and the seedfall around each fruit-bearing O. europaea tree in two contrasting sites, differing in the degree of human management and abundance of scrub cover. None of the seed dispersers moved between microhabitats (different shrub species and open interspaces) as a function of microhabitat relative abundance. All dispersers foraged preferentially at O. europaea; only the smaller species visited open interspaces, and these only sporadically. Avoidance of open sites by frugivores, especially larger species, and their attraction to source trees were the major determinants of the seedfall pattern. Regarding seed size distribution among microhabitats, we found that only small seeds were significantly overrepresented in open interspaces. From the seed shadow perspective, seed density under trees was similar in the two study sites, but it decreased with distance at a higher rate in the disturbed scrubland. From the seed rain perspective, seed density was significantly different among microhabitats, with open interspaces collecting few, if any, seeds, and places under source trees receiving the highest densities. Results from seed shadow and seed rain analyses were integrated to construct a path model to explore the relative contribution of spatial, microhabitat, and individual tree features to the seedfall pattern. Factors related to microhabitat were identified as having a major role in shaping seedfall pattern. Open interspaces collected few seeds, points under shrub species with sparse foliage collected moderate seed densities, and points under shrubs with dense foliage collected the highest seed densities. The comparison between sites suggests that habitat alteration (a reduction of the scrub layer) can lead to contrasting seedfall patterns. The occurrence of an abundant scrub layer in the well-preserved scrubland allowed a complex pattern of seed rain. In contrast, the sparse scrub cover in the disturbed scrubland yielded a seedfall pattern composed of multiple seed shadows constrained to the vicinity of tree crowns. Thus, our sampling design and analytical approach have proved useful in describing the relative importance of the set of variables that shapes the complex seedfall pattern of a bird-dispersed plant. This seedfall pattern, in turn, is central to the understanding of the spatial patterns of plant recruitment and the efficiency of the dispersal mutualism.