Seed dispersal and seedling recruitment: The missing link(s)

Abstract

:Typically seed rain density decreases as a function of distance from the parent plant. Most models of recruitment assume a seed (and a seedling) mortality function monotonically decreasing with distance from the parent (hence, positively density-dependent). This results in either higher recruitment at a distance from the parent (Janzen-Connell model) or near it (Hubbell model), or else a spatially uniform recruitment (exact compensation model), depending upon the specific shape of the mortality function in relation to distance from the parent. However, mortality may increase with distance from the parent (negatively density-dependent) or be spatially uniform, both of those scenarios resulting in a recruitment similar to that predicted by the Hubbell model. Recruitment may also be unpredictable on the sole basis of the parent plant position if mortality is spatially variable and density-independent. For several tree species in a homogeneous cold temperate forest (Houle, 1992a), recruitment could not be predicted from seed rain spatial patterns. Seeds and seedlings were contagiously distributed and mortality, at both stages, was spatially variable and density-independent. In a similar forest type (present study), in which two species were segregated along a topographical gradient, recruitment of each species was higher where respective seed rain abundance was higher. At this study site, also mortality, for both seeds and first-year seedlings was not density-dependent, with the result that clumping at the seedling level and positive relationship between seed rain and recruitment were maintained through time. A chain of events from seed dispersal through seedling establishment and subsequent survival controls recruitment. Simple models tend to mask the complexity of the processes involved between dispersal and recruitment.