There is evidence that matrix features similar in structure to breeding habitat patches increase species movement. The evidence base is broad and heterogeneous. The overall strength of the evidence is fairly low, which may be at least partly accounted for by the variability in individual species’ responses to landscape features, and also by the relative infrequency of dispersal events. Some of the most abundant evidence was for the relative importance of linear features to movement. The evidence supports assumptions made in a range of modelling studies (described in Watts et al., 2005).