Seed size and number make contrasting predictions on seed survival and dispersal dynamics: A case study from oil tea Camellia oleifera

Abstract

Seed size and number vary greatly both within and among plant individuals, populations or species. However, we know little about how individual variation in seed size and number relates to seed survival and dispersal dynamics in animal-dispersed plants. Based on the large-seed hypothesis and the predator satiation hypothesis, we developed a synthetic framework and testable predictions for how seed size and number interact to predict seed survival-dispersal patterns at the seed and tree scales, and tested for these effects using rodent-dispersed oil tea, Camellia oleifera. Our results showed that seed size and number varied considerably at both fruit and tree scales, and mean seed size for each fruit was negatively correlated with the number of seeds per fruit. We placed groups of individual-weighed seeds in a way that mimicked seed size and number from individual trees and tracked the fate of these seeds. This showed that seed survival in situ at source trees decreased with increasing seed size, but seed dispersal and the subsequent survival of dispersed seeds increased with larger seed size. In addition, seed survival in situ at source trees increased with larger seed number per tree, but the proportion of dispersed seeds decreased as seed number per tree increased. We demonstrate that individual variation in seed size and number have strong but contrasting effects on seed survival and seed dispersal in oil tea. Moreover, these two factors are not independent in their effects on seed survival at the source tree, but after dispersal, the size of individual seeds may be more important in determining seed survival or further handling by scatterhoarding animals.