During 1995–1998, we tested 134 geographically discrete populations of Colorado potato beetle, Leptinotarsa decemlineata (Say), from the United States, Canada, Germany, France, and Poland for susceptibility to imidacloprid. Neonates were assayed on potato-based agar diet incorporated with imidacloprid and exposed on filter paper to esfenvalerate, azinphosmethyl, and carbofuran to characterize cross-resistance. In all 4 yr, Long Island populations were the most tolerant to imidacloprid, with LC₅₀s ranging up to 29 times higher than the most susceptible populations. Responses to imidacloprid did not change significantly on farms where populations were assayed over time, except for those from Long Island, which doubled in overall tolerance to imidacloprid since 1995. Much of this tolerance was already present before imidacloprid was used on Long Island. Correlative analysis of the populations tested over the 4 yr indicated positive cross-resistance patterns with esfenvalerate and azinphosmethyl. This response was probably caused by preexisting metabolic and excretion mechanisms selected by previous exposure. There was no significant pattern of cross-resistance with carbofuran or bensultap. Regression slopes were also significantly negatively correlated with LC₅₀ values for imidacloprid, indicating higher heterogeneity, which could lead in further resistance development. We discuss the relative sensitivity of diet-incorporated assays with neonates compared with other bioassay studies. Based on a pooled group of susceptible populations tested in 1995, a baseline LC₅₀ of 0.39 ppm and a discriminating concentration of 8 ppm were suggested to detect early stages of resistance in “suspect” populations. We also suggest application strategies for imidacloprid that reduce selection pressure.