Estimating adhesive seed‐dispersal distances: field experiments and correlated random walks

Abstract

Summary In this study we aimed to estimate distance distributions of adhesively dispersed seeds and the factors that determine them. Seed attachment and detachment were studied using field experiments with a real sheep, a sheep dummy and a cattle dummy. Seed‐retention data were used in correlated random walk models to simulate adhesive seed dispersal. Seed attachment to the sheep dummy was larger in quantity and in number of species, and stronger in relation to seed density in the vegetation, than was seed attachment to the cattle dummy. Species found on the real sheep were also found on the sheep dummy. Detachment from sheep wool differed little between smooth, bristly, small or large seeds, but smooth seeds detached from cattle fur within a few metres. Seeds applied within reach of vegetation detached sooner than seeds applied higher on the dummy. The simulations showed that sheep are long‐distance seed‐dispersal vectors for seeds of any morphology (99 percentile distance, 2·9 km). The virtual cattle and Fallow Deer dispersed bristly and hooked seeds over long distances (99 percentile distance, 435–840 m), but not smooth seeds. Wood Mouse simulations generated only short‐distance dispersal (99 percentile distance, 12 m).