Technical Abstract: We used pan traps in three colors (blue, yellow, white) to sample the bee fauna on five sand dune sites on military lands in the west desert of Utah. Our primary objectives were to 1) compare bee species richness and similarity among the three colors; 2) estimate the incidence of color preference among bee species; 3) examine the effect of background bloom color on pan trap color preference. We found the colors to be equivalent in the numbers of bee species captured: each color captured about two-thirds of the 84 total species recorded with an average similarity between colors of about 0.70 (Sorenson Index). Average similarity among colors was much higher (0.93) when only abundant species were considered. All bee species represented by sufficient numbers displayed significant color preferences. Each color was preferred by at least one species. We found no indication that the prevalent color of flowers in bloom during each site-week had an effect on occurrence in pan traps. Those seeking to exhaustively enumerate the bee fauna of an area using pan traps should employ a diversity of colors to ensure capturing uncommon species with color preferences.