Bait stations with toxic baits are an emerging technology for eradication of the invasive brown tree snake (Boiga irregularis) on Guam, yet potential interferences by nontarget species are largely unknown. We tested the efficacies of three bait station designs together with three commonly used station support structures to exclude nonnative rats (roof rat, Rattus rattus; Norway rat, Rattus norvegicus; Polynesian rat, Rattus exulans) and native coconut crabs (Birgus latro). When directly presented, all species readily consumed the dead neonatal mouse baits (nontoxic) including those replicating decomposing baits in the field. When bait stations were made easily accessible by placement near ground level, all rat species were able to enter all station types, but some individual roof rats and Norway rats exhibited apparent neophobia. When stations were placed up on support structures, simulating those in the field (∼1 m above ground level), numbers of station accessions by roof rats and Norway rats remained essentially unchanged, but Polynesian rats then showed almost no inclination to enter stations. However, ability to access entrances of stations (but not interiors) when on support structures was extremely high for roof rats and appreciable for the other rat species, including Polynesians. The station type currently in widest use, when placed on chain-link cyclone fence, had the highest probability of accession. Crabs readily accessed station entrances but never interiors. The two downward-angled station designs, when placed in simulated vegetation, had the lowest probabilities of accession. In areas where nontarget species are a concern, we recommend use of either of the downward-angled station designs and suspension from vegetation wherever possible.