Relative pollinator quality can be measured in terms of the amount of pollen removed from anthers and subsequently deposited on stigmas, combined with the rate of floral visitation. Most studies compare crop pollinators on the basis of seed yield or pollen deposition. Considering pollen removal may give additional insight into pollinator quality, particularly in systems where pollination is limited by insufficient pollen transfer. Using a pollen removal-deposition measure, we compared three native bees and honey bees (Apis mellifera L.) visiting the native annual Phacelia tanacetifolia Benthem. Honey bees removed significantly less pollen per visit than any native pollinator. Bumble bees (Bombus spp. Latreille) and Anthophore urbana Cresson deposited relatively more pollen on stigmas than either honey bees or Osmia lignaria Cresson. Per-flower visitation rates also differed among bees. A. urbana were at least three times faster than any other species. According to a pollen removal and deposition measure of quality, Bombus spp. and A. urbana are relatively good P. tanacetifolia pollinators. Honey bees are poorer because they move little pollen into circulation. O. lignaria are the worst because they deposit little on stigmas and may remove pollen from the system that could be deposited by other species. A removal and deposition perspective cannot be substituted for comprehensive study of seed yield by different pollinators. However, the removal and deposition method, combined with pollen transfer models, may provide a powerful complement to explore systems with multiple visitors and screen candidate pollinators in conjunction with studies of pollinator biology.