The main reason for this, for which we refer the reader to the Discussion and Text S1, is the difficulty in discriminating between long-range dispersal, occurring within a census site, and primary infection incoming from outside.