Incorporating patterns of disperser behaviour into models of seed dispersal and its effects on estimated dispersal curves.

The processes determining where seeds fall relative to their parent plant influence the spatial structure and dynamics of plant populations and communities. For animal dispersed species the factors influencing seed shadows are poorly understood. In this paper we test the hypothesis that the daily temporal distribution of disperser behaviours, for example, foraging and movement, influences dispersal outcomes, in particular the shape and scale of dispersal curves. To do this, we describe frugivory and the dispersal curves produced by the southern cassowary, Casuarius casuarius, the only large-bodied disperser in Australia's rainforests. We found C. casuarius consumed fruits of 238 species and of all fleshy-fruit types. In feeding trials, seeds of 11 species were retained on average for 309 min (+/-256 SD). Sampling radio-telemetry data randomly, that is, assuming foraging occurs at random times during the day, gives an estimated average dispersal distance of 239 m (+/-207 SD) for seeds consumed by C. casuarius. Approximately 4% of seeds were dispersed further than 1,000 m. However, observation of wild birds indicated that foraging and movement occur more frequently early and late in the day. Seeds consumed early in the day were estimated to receive dispersal distances 1.4 times the 'random' average estimate, while afternoon consumed seeds received estimated mean dispersal distances of 0.46 times the 'random' estimate. Sampling movement data according to the daily distribution of C. casuarius foraging gives an estimated mean dispersal distance of 337 m (+/-194 SD). Most animals' behaviour has a non-random temporal distribution. Consequently such effects should be common and need to be incorporated into seed shadow estimation. Our results point to dispersal curves being an emergent property of the plant-disperser interaction rather than being a property of a plant or species.