Seed bank versus seed rain in the regeneration of a tropical pioneer tree

Publisher Website
Google Scholar
Abstract
We used the tropical pioneer tree, Cecropia obtusifolia to evaluate the relative importance of different sources of seeds in the regeneration of species that depend on ephemeral sites. We studied seed production in a population established in a 5 ha plot, and dispersal, dormancy and seed predation in two recent treefall gaps (35 years since disturbed) for a one year period at Los Tuxtlas (Mexico). Flowers and fruits were counted at monthly intervals. Annual fecundity per tree ranged from 1.4×104 to 1.4×107 seeds. Seeds were continuously available on the trees and on the ground. Average annual seed rain per m2 (as measured by 0.5×0.5 m seed traps) varied from 184 to 1925 among the six sites. Distance to nearest seed source and patch type explained more than 60% of the seed rain variation among sites. Soil seed density, estimated by counting seeds from ten samples (78.5 cm2×10 cm deep) collected from each site in October and January, ranged among the six sites from 269 to 4485 seeds per m2 in January and from 204 to 5073 in October. Soil seed viabilities were much lower (17.1% in October and 5.1% in January) than those of rain seeds (48.26%). Annual survivorships of 2.2% were estimated for seeds artificially sown on the soil surface of a gap and a mature patch, and 3.75% in a building patch. In two other experiments seed removal rates ranged from 27% to 98% in 4 days. Removal rates were significantly higher in gap and mature patches than in building patches. Ants (Paratrechina vividula) and grasshopper nymphs (Hygronemobius. sp.) were the main predators. We draw three main conclusions from our data: (1) Pathogens and predators determine low survivorship of C. obtusifolia's seeds in the soil and a rapid turnover rate (1.07 to 1.02 years) of its seed bank; (2) a continuous and copious seed production and an abundant and extensive seed rain replenish the soil seed pool in patches with different disturbance ages at least up to 86 m from nearest source; (3) more than 90% of the seeds contributing to C. obtusifolia seedling recruitment in gaps are less than one year-old. We discuss our results in the context of previous similar studies for tropical forests.