| The objectives of these studies were: (1) to develop a method for pan trapping bees in highbush blueberry (Ericaceae: Vaccinium corymbosum L.); (2) to describe the structure of the endemic bee community associated with blueberry and determine to what extent wild bees contribute to its pollination; (3) to examine the relationship between wild bee community structure and both local habitat features and pest management practices; (4) to evaluate native perennial plants for attracting bees; and (5) to examine blueberry bee community structure in the context of landscape spatial scales. Pan traps elevated in the canopy of flowering blueberry collected more bees than traps placed on the ground or above the canopy. This method was used at 15 blueberry farms in southwest Michigan to characterize the bee community present prior, during, and after blueberry bloom for three years (2004-6). Honey bees, primarily from rented hives, comprised 50-66% of all the bees captured in pan traps each year during bloom. The remainder were wild bees, mainly soil nesting bees in the families Andrenidae and Halictidae. Andrena carolina, a solitary bee that is oligolectic on Ericaceae, was the most abundant species captured in pan traps during bloom (14%). Bees collected while foraging (n = 22 species) and pollen carried by bees in pan traps (n = 126 specimens) were used to confirm which bees were foraging on blueberry. In pan trap samples during bloom, there was high species turnover from year to year with 79 species recorded on average each year out of a total of 120 species over the three years. Wild bees were more often captured in traps at the field edge than in the interior. In simple linear regressions (SLR) of bees and field characteristics, bee species richness increased with flowering plant species in 2005, and declined with the local frequency of adjacent blueberry fields. Bee diversity (H') was also lower in fields with more nearby blueberry fields. Native bee abundance and richness in 2004 along with bee diversity in 2005 declined with increasing insecticide program toxicity (IPT). In a redundancy analysis of the same characters, IPT explained the greatest amount of variation in the bee species data in 2004, and vegetation attributes explained variation in the species in 2005 and 2006. Of the 43 native perennial plants that were evaluated, all but 2 were visited at least once by no-Apis bees (n = 1393), but there were 9 that were visited most frequently. Floral area explained 33% of the bee diversity at particular plant species. The response of wild bees to the proportion of different landscape types around blueberry fields was evaluated at 5 nested scales using 250, 500, 750, 1000, and 1500 m radius circles. Forest margins (<10 m from the forest edge), human settlement, annual cropland, and blueberry plantations were the most abundant landscape types at each of the spatial scales. Total bee abundance, richness, and diversity did not vary significantly with any of 6 categories of land use at any of the spatial scales, however, A. carolina responded to the proportion of human settlement at the 1500 m scale, and Ceratina calcarata/dupla responded to the proportion of blueberry plantations and semi-natural habitat at the 250 m scale. Three of 4 blueberry fruit yield attributes increased with the proportion of blueberry habitat and decreased with the proportion of semi-natural habitat within 500 m of the focal field. Implications for the conservation of native bees and their importance in blueberry pollination are discussed. |
