The probability that any given host will be exposed to a vector-borne pathogen increases with increasing population density of the vector and with increasing proportion of vectors that are infected.  Vector density and infection prevalence, however, might be affected by different ecological factors.  For vectors with inefficient vertical (transovarial) transmission, infection prevalence is determined by the probability that previous blood meals were taken from an infected reservoir host.  Determinants of vector density are more complicated, and might include climate, availability of breeding sites, and availability of hosts.  Quantitative and conceptual models generally do not consider host species identity important in affecting vector density.  We tested the ability of larval blacklegged ticks, the main North American vector of Lyme disease (LD), anaplasmosis, and babesiosis, to survive while attempting to feed on six species of commonly parasitized vertebrate hosts.  We then used a simple model to project tick abundance and infection prevalence when the diversity and species composition of the host community varies.Results/Conclusions

Species varied significantly in their quality as hosts for ticks.  Almost half of the larval ticks that were placed on white-footed mice fed to repletion, while only 3.5% of ticks that fed on opossums did.  Eastern chipmunks, gray squirrels, veeries, and gray catbirds were of intermediate quality.  The remaining ticks were not recovered, indicating that they had been consumed during host self-grooming.  Among these six species host quality was positively correlated with reservoir competence (r = 0.85, P = 0.03); hosts that better support tick survival also more efficiently infect the ticks with the LD bacterium.  We calculated the density of infected nymphs (DIN, the main human risk factor) for a host community consisting of all six species, and then used a simple model to remove species in approximately the order they disappear as forests are fragmented.  Reduction in host diversity dramatically increased DIN.  We conclude that: (1) some naturally-infested hosts act as ecological traps that attract vectors but kill most of those that attempt to feed; (2) trap species also tend to be dead-ends for the pathogen; (3) trap species serve a potent protective role and their loss exacerbates disease risk.  We suspect that common life-history traits influence host suitability for both vectors and vector-borne pathogens; if so, results from the LD system should be generalizable to other vector-borne zoonoses.