If both a null model and a deterministic model can predict the shape of the same empirical distribution of dispersal distances, the underlying processes of dispersal cannot be safely attributed to the process described by the deterministic model. The author provides a null model against which these geometric models can be tested. -from Author