The effect of tropical deforestation on stingless bees of the genus Melipona (Insecta: Hymenoptera: Apidae: Meliponini) in central Rondonia, Brazil

Aims (1) 2 To determine whether deforestation has affected the incidence of species of Melipona in an area undergoing deforestation. (2) To highlight the potential use of GIS and remote-sensed environmental variables in studies of insects as bioindicators of landscape change. Location Central Rondonia, Brazil in the south-western Amazon Basin, an area under intense deforestation pressure for agriculture and cattle ranching since the 1970s. Methods Stingless bees of the genus Melipona were sampled outside of forest cover on iron-weed, Vernonia polynthes Less., at 69 locations along a deforestation gradient within a 3150-km study area. We related species richness and composition for each sample point to local and regional deforestation variables, including geographical position along the deforestation gradient, distance to the forest, and percentage of primary forest coverage within a 1 and 2 km radius of the sample points. Deforestation variables were generated using GIS and LANDSAT TM 3 imagery of the study area. Redundancy analysis was used to illustrate the relationship between species incidence, the deforestation variables, and other possible confounding environmental variables. Results Seven species were found within the study area. Two species (M. seminigra abunensis Cockerell, 1912 and M. grandis GueÂrin, 1844) appear not to be affected by deforestation yet, occurring evenly across the deforestation gradient. Two other common species (M. melanoventer Schwarz, 1932 and M. ru®ventris brachychaeta Moure, 1950), however, occurred mainly towards the end of the gradient where the forest was more intact, indicating relative susceptibility to deforestation. Melipona species richness, ranging from 1 to 5 species, was inversely related to distance to forest and directly related to percentage of forest cover. Main conclusions Adverse effects of deforestation on Melipona are detectable in the study area, despite the fact that signi®cant areas of tropical forest cover remain. The species that are most affected may be considered indicators of landscape change, and efforts to protect these species could involve their use in beekeeping programmes designed to raise rural incomes and maintain regional biodiversity.