Insect traps of different designs should sample some species more effectively than others. I used three different traps to survey longhorned beetles (Coleoptera: Cerambycidae) at 43 locations throughout the state of Indiana to test the prediction that these traps would sample different assemblages of cerambycids. Nonmetric multidimensional scaling revealed that, while there was little difference between the assemblages sampled by Lindgren multiple funnel traps and Intercept panel traps, window traps caught quite different assemblages. Several species of cerambycids that showed differences in trap efficiencies seem to be differentially caught by different trap designs according to the nature of the larval host. These results can be used to plan monitoring according to the guild of the target beetles.