Large‐scale patterns of distribution and persistence at the range margins of a butterfly

One major goal of conservation biology in fragmented landscapes is to identify the attributes of habitat networks that influence metapopulation persistence, and thereby to determine whether particular areas hold the key to regional persistence. In Britain, the genus Aricia is represented by A. agestis in the south and A. artaxerxes in the north. We mapped the distributions of A. agestis and A. artaxerxes at their range margins, where both species are restricted to isolated fragments of calcareous grassland containing the larval host plant Helianthemum chamaecistus. We present evidence that Aricia is likely to persist as metapopulations at its range margins. First, both individual habitat patches and semi‐independent networks of patches showed colonizations and extinctions of Aricia. Second, the smallest and most isolated habitat patches were least likely to be occupied. Third, persistence was more closely related to the total area and configuration of habitat in each network (consistent with a metapopulation interpretation) than to the area of the largest patch (predicted for mainland–island systems). In addition, the occupancy of entire networks depended on their isolation from other networks, well beyond the empirically measured dispersal range of the species. We suggest that species may persist as “metapopulations of metapopulations” at coarse spatial scales. In some systems, all networks may be prone to extinction, but we suggest that regional persistence more often depends on the continued occupancy of extinction‐resistant core areas (“core/marginal metapopulations”). Extinctions of small metapopulations and very long‐distance colonizations may profoundly influence the distributions of species in fragmented landscapes, and set the dynamic margins of species ranges. As a result, identification and protection of extinction‐resistant core areas should be priorities in conservation programs.