Ohio's oral rabies vaccination (ORV) program was established to prevent the westward spread of the raccoon (Procyon lotor) rabies virus (Lyssavirus, Rhabdoviridae) in Ohio, USA. The program, which targets raccoons, distributes vaccine-bait units (VBU) at a target density of 75 units/km². Few studies have examined the relationship of VBU density and target population density to the prevalence of rabies virus-neutralizing antibodies (RVNA). We conducted experimental VBU distributions in August 2003 and August 2004, 150 km west of the ORV zone where there was no history of raccoon rabies. We measured change in RVNA titers in blood collected from live-trapped raccoons before and after VBU distributions. A closed population mark–recapture estimate of the size of the target population was 91 raccoons/km², compared to the realized VBU distribution density of 70 units/km². Surprisingly, 41% of 37 serum samples were RVNA-positive (≥0.05 IU/ml) before VBU distribution in 2003, but all titers were <0.25 IU/ml. Although viable VBUs were distributed in August 2003, only 21% of 315 samples were RVNA-positive before VBU distribution in 2004, but 9% had titers ≥0.25 IU/ml. Tetracycline (biomarker in bait) prevalence in teeth indicated that 57% of raccoons ingested VBUs after distribution in 2003, and 54% ingested VBUs after distribution in 2004. However, only 8% and 11% of sera were positive for RVNA (≥0.05 IU/ml) after VBU distribution in 2003 and 2004, respectively. Only 4–5% of sera collected after bait distribution had titers ≥0.25 IU/ml each year. The standard distribution density of 75 VBUs/km² was insufficient to produce a population-wide immunoprotective response against rabies infection in our high-density target population. Presence of RVNA in a presumed naïve population before baiting demonstrates that estimating prevalence of RVNA after oral rabies vaccination can be problematic without knowledge of background titers and seasonal changes in prevalence of RVNA before and after baiting.