Spatial patterns of bee captures in North American bowl trapping surveys

Bowl and pan traps are now commonly used to capture bees (Hymenoptera: Apiformes) for research and surveys. Studies of how arrangement and spacing of bowl traps affect captures of bees are needed to increase the efficiency of this capture technique. We present results from seven studies of bowl traps placed in trapping webs, grids, and transects in four North American ecoregions (Mid-Atlantic, Coastal California, Chihuahuan Desert, and Columbia Plateau). Over 6000 specimens from 31 bee genera were captured and analysed across the studies. Based on the results from trapping webs and distance tests, the per bowl capture rate of bees does not plateau until bowls are spaced 3-5 m apart. Minor clumping of bee captures within transects was detected, with 26 of 56 transects having index of dispersion values that conform to a clumped distribution and 39 transects having positive Green's index values, 13 with zero, and only four negative. However, degree of clumping was slight with an average value of only 0.06 (the index ranges from -1 to 1) with only five values > 0.15. Similarly, runs tests were significant for only 5.9% of the transects. Results indicate that (i) capture rates are unaffected by short distances between bowls within transects and (ii) that bowls and transects should be dispersed throughout a study site.