Concern over insecticide resistance has led to the suggestion of spatially variable within-field management of Colorado potato beetle, Leptinotarsa decemlineata (Say). Here we compare L. decemlineata spatial and temporal dynamics, and potato yield, in fields treated with a narrow perimeter (5.5 m) of systemic imidacloprid supplemented with spatially targeted sprays to untreated fields and to fields where all rows received the systemic. The systemic targeted immigrating individuals which, having acquired the field through either flight or walking, first established themselves in the outer 5.5 m of the field. The perimeter treatment (≈25% of field area) reduced mean densities with no effect on timing of peak densities. Immigrating adults established similar spatial trends in both perimeter and untreated fields. Although trends in F₁ larval densities have been shown to follow the patterns established by immigrating adults, trends in the F₁ larval densities of the perimeter treatments diverged from adult patterns and developed highest densities in field centers. Immigrating adults had little to no spatial dependence in the covariance structure in any treatments. Spatial dependence in the covariance structure of F₁ larval and F₁ adult populations developed as density increased in both perimeter and untreated fields, with a tendency for increasing spatial dependence in perimeter fields, though this was not statistically significant. Comparing the perimeter to untreated fields, yields increased at a proportion that was higher than the proportion of land area treated, but remained significantly lower than the whole-field treatment. These results suggest that the perimeter tactic has promise as a site-specific resistance management program, but that refinement of border width is needed to optimize trade-offs among yield, quality, and long-term maintenance of susceptibility.