Summary, in English
To investigate recent declines in bumblebee populations, I carried out landscape analyses, made field studies of bumblebees, flowering plants and pollination in an agricultural region in southernmost Sweden. Studies were carried out in landscapes of contrasting agricultural intensity and land-cover configuration; "simple" landscapes highly dominated by large crop fields, and "complex" with mixed farming, grasslands and smaller fields. Factor Analysis of land-cover and measures of agricultural practice showed that intensity is not to be equated with landscape complexity. A relatively complex landscape could thus provide high yields from farming, while retaining small non-crop elements. This is promising for persistence of farmland biodiversity. Total number of bumblebees was ca. 30 times lower in simple compared to complex landscapes in July. However, numbers did not differ significantly in June, despite ca. 30 times more of flowering plants in complex landscapes. This may be explained by the larger area of oilseed rape grown in simple landscapes, since oilseed rape may aid early colony growth. Effects of landscape context on bumblebees was modified by colony-based, ecological and life-history traits. Species with early emerging queens, large colonies, under-ground nests or short colony reproductive cycle, produced equal numbers of male offspring in both simple and complex landscapes. However, all species produced smaller workers in simple landscapes. This may be caused by a shortage of food, which hampers larval growth and adult size. It may also be an adaptive strategy to produce more but smaller workers under food stress. Individual-based traits of workers: thorax width and proboscis length modified foraging habitat preferences. Likely because composition of flowers differed among habitats and therefore also differed in value depending on bumblebee species. Higher abundance and richness of wild bees were found near domestic gardens in simple landscapes, compared to 140m further away. Also, seed set of Campanula persicifolia was higher nearer to a garden. Gardens may thus act as refuges for wild bees, constitute sources of pollinators to surrounding habitats and benefit pollination in these landscapes. However, pollination of wild plants may already be hampered in simple landscapes, since seed set of C. persicifolia was lower already at 140m from a garden. Measures to enhance bumblebee populations and pollination in farmland regions should focus on increasing landscape complexity by restoring and recreating flower-rich non-crop habitats. Importantly, this must be implemented over whole regions and match the spatial scales and phenology at which declining species operate and flowering plants must also meet the preferences of these species.