Natural cross‐pollination rates of blue lupine (Lupinus angustifolius L.) in Georgia and Florida ranged from 0 to 11.95% for different genotypes, test sites, test years, and honey bee population densities. The larger and fleshier‐flowered, earlier‐flowering biotype (common bitter blue lupine and ‘Rancher’ sweet blue lupine) had much lower cross‐pollination rates (0 to 3.45%) than the smaller and more fragile‐flowered, late‐flowering, elite winter‐hardy line 65G‐251. Nine other elite winter‐hardy lines were similar to 65G‐251 in floral, maturity, and cross‐pollination‐rate characteristics. We attribute the difference in natural cross‐pollination rates between the two major biotypes to their relative resistance to tripping by honey bees (Apis mellifera L.) or the flowering of the late type at temperatures more favorable for honey bee activity, or both.Results from tests including treatments designed to prevent flower‐visitation by one or more of the possible insect vectors of cross‐pollination indicate that: (i) Bumble bees (Pyrobombus spp.) easily tripped all types of blue lupine flowers and are cross‐pollination vectors, but their numbers were too few to be the principal vector; (ii) Honey bees were the principal cross‐pollination vectors; (iii) Thrips (Frankliniella spp.) were not cross‐pollination vectors, although they may render some flowers male sterile hy eating all the pollen; (iv) Wild bees were not vectors; and (v) Blue lupine is highly self‐compatible, highly self‐pollinated, and fully capable of automatic self‐pollination, i.e., self‐pollination within the closed flower is independent of insect visitation.In all tests at all locations cross‐pollination rates were positively correlated with honey bee population densities. Wind protection of the test site favored honey bee activity and higher cross‐pollination rates.We made suggestions based on these findings for use in breeding, genetic, and variety seed production operations.