LAUSR

Theory predicts that species diversity of isolated habitat patches depends on patch size and isolation. However, there are few previous studies of how patch size and isolation influence community composition. This study tested theory using small and isolated farmland patches in a forested matrix. I tested whether bird species diversity and community composition of isolated farmland patches depended on patch size, measures of isolation (e.g. distance from the main farmland areas; up to 15 km), and other variables such as presence of human habitation. Farmland bird species diversity and community composition (relative proportions of different ecological guilds) of 130 farmland patches within forest landscapes were recorded in Norway. Species diversity and proportion of the species that nested on the ground or low in bushes increased with patch size. Distance from the main farmland region was not related to species diversity or community composition, but species diversity increased when there were stepping stones present. Sites inhabited by humans had higher species diversity, a larger proportion of species nesting high in trees or in cavities (such as in buildings), and a larger proportion of resident species. Other variables were less important. There was also a strong relationship between the occupancy of individual species in the main farmland region and their occupancy in the forest region, suggesting that colonisation pressure also influenced species diversity of these habitat islands. I found weak evidence of isolation effects on bird species diversity and community composition of farmland patches within forest, whereas most of the variation was explained by patch size and presence of humans.