Seed dispersal is a multi-step process that links successive generations of plants. The most immediate outcome of dispersal is a patchy distribution of seeds in the ground, which can later influence the overall magnitude of seed predation, germination, and seedling establishment and, ultimately, determine the spatial patterns of recruitment. Hence, a central question in plant population ecology is whether adult plant recruitment can be spatially predicted from the seed deposition patterns or whether post-dispersal processes alter, if not erase, the initial seed rain pattern. In the latter case, post-dispersal processes can lead to seedling establishment patterns that are discordant with those of seed rain. Here, we investigated the patterns of concordance between seed dispersal and seedling recruitment of a fleshy-fruited, bird-dispersed tree (G. viburnoides) in four different habitats in a Neotropical savanna: 1) vegetation patches with G. viburnoides, 2) vegetation patches without G. viburnoides, 3) forest islands, and 4) savanna sites. In each habitat, we quantified seed rain and set up experiments to determine i) seed predation, ii) seedling emergence and, iii) seedling recruitment (survival of first-year seedlings).
Results/Conclusions
The density of dispersed seeds differed among habitats, being greater in patches with G. viburnoides than in any other habitat. Survival analysis revealed that post-dispersal predation also depended on the habitat. The predation risk of seeds dispersed in the savanna or in vegetation patches without G. viburnoides was 101% and 81% lower than in patches with G. viburnoides, respectively. However, the predation risk in forest islands was comparable to patches with G. viburnoides. Seedling emergence was marginally affected by habitat, tending to be higher in patches with G. viburnoides than in any other habitat. The spatial dynamics of G. viburnoides recruitment was examined via habitat-specific transition probabilities. The results show that the most favorable places for seed survival (savanna sites) became the worst places for seedling emergence and survival, whereas the worst places for seeds (patches with G. viburnoides) became the best for seedlings. These results indicate that seed-seedling conflicts occur in this population, and, eventually lead to seedling establishment patterns of G. viburnoides that are discordant with those of seed rain. Ultimately, understanding the occurrence and outcome of these conflicts will increase our understanding of the role of seed dispersal in the development of the spatial pattern in plant populations, which is critical to explanations of plant community structure.
