Dispersal-dormancy relationships in annual plants: putting model predictions to the test.

Bet hedging is a means to increase fitness in environments that vary unpredictably in space and time. In such environments, models predict a trade-off between the bet-hedging strategies dispersal and dormancy, while the increasing importance of risk reduction with decreasing predictability should lead to an increase in dispersal and dormancy along gradients of environmental predictability. However, so far there has been no experimental study to test these predictions in the field. Here, we used a set of novel field experiments that enabled us to quantify and separate seedling recruitment from three sources: local reproduction, dormancy, and dispersal. The study included the entire plant community from five environments differing considerably in predictability. Evidence for both the existence of a trade-off between dispersal and dormancy within environments and their increased use in unpredictable environments was very weak. The importance of dispersal for population and community dynamics in our system was extremely low relative to dormancy and local reproduction. This indicates that the role of dispersal for buffering environmental variation may be negligible compared with other risk-reducing strategies. Our findings highlight the urgent need for multispecies and multisite experiments in empirical tests of theoretical predictions.