Nymphal and adult Ixodes scapularis Say were sampled by flagging at 2 sites on a barrier island, Fire Island, NY, and at 2 sites on the nearby mainland. Nymphal densities did not differ consistently between island and mainland sites, but adult densities were consistently lower on the island. We tested whether lower adult densities on the island resulted from greater nymphal mortality on the island than the mainland, or whether adult ticks on the island were poorly sampled by flagging because they had attached abundantly to deer, which were common on Fire Island. Differential nymphal mortality on islands vs. mainland did not explain this difference in adult densities because survival of flat and engorged nymphs in enclosures was the same at island and mainland sites. Ticks were infected by parasitic wasps on the island and not the mainland, but the infection rate (4.3%) was too low to explain the difference in adult tick densities. In contrast, exclusion of deer by game fencing on Fire Island resulted in markedly increased numbers of adult ticks in flagging samples inside compared to samples taken outside the exclosures. Therefore, the scarcity of adult ticks in flagging samples on Fire Island resulted, at least in part, from the ticks being unavailable to flagging samples because they were on deer hosts. Differences in the densities of flagged ticks inside and outside the exclosures were used to estimate the percentage of questing adults on Fire Island that found deer hosts, excluding those that attached to other host species. Approximately 56% of these questing adult ticks found deer hosts in 1995 and 50% found deer hosts in 1996. Therefore, in areas where vertebrate hosts are highly abundant, large proportions of the questing tick population can find hosts. Moreover, comparisons of tick densities at different sites by flagging can potentially be biased by differences in host densities among sites.