An experimental test of corridor effects on butterfly densities

In a large-scale experiment, we found that three habitat-restricted butterf ly species reached higher densities in patches connected by corridors than in similar, isolated patches. We conducted our study in 27 equal-sized (1.6 4-ha) patches that varied in whether or not they were connected to another patch by a corridor. Patches and corridors were open, early-successional ha bitat that contrasted strongly with the surrounding pine forest. Of four bu tterfly species studied, three open-habitat specialists (Junonia coenia, Eu ptoieta claudia, and Phoebis sennae) reached higher densities in patches co nnected by corridors than in isolated patches. A fourth species, Papilio tr oilus, showed no preference for open habitat or pine forest, and its densit y did not differ in connected or isolated patches. Although butterfly densi ties were often significantly, positively associated with densities of host plant or flower resources, plant densities did not confound corridor effec ts on butterfly densities. Higher densities in patches connected by corridors may have been caused by three different factors. First, species with higher densities in connected patches also moved more frequently between connected patches, and higher mo vement rates may have helped to sustain higher population sizes. Second, sp ecies with higher densities in connected patches also had higher densities farther from the forest edge. Corridors increased the area of a patch that was farther from the forest edge, which increased the "effective area" of c onnected patches and may have increased butterfly densities. Third, corrido rs may have acted as "drift fences," intercepting dispersers from the surro unding forest and directing them to connected patches. We could not determi ne the relative contribution of each factor, and it is possible that all th ree contributed to higher densities of habitat-restricted butterflies in pa tches connected by corridors.