Accurate and detailed pollen monitoring is useful for selection of medication and for allergen avoidance in patients with allergic rhinitis. Burkard and Durham pollen samplers are commonly used, but are labor and time intensive. In contrast, automatic pollen counters allow simple real-time pollen counting; however, these instruments have difficulty in distinguishing pollen from small nonpollen airborne particles. Misidentification and underestimation rates for an automatic pollen counter were examined to improve the accuracy of the pollen count. The characteristics of the automatic pollen counter were determined in a chamber study with exposure to cedar pollens or soil grains. The cedar pollen counts were monitored in 2006 and 2007, and compared with those from a Durham sampler. The pollen counts from the automatic counter showed a good correlation (r > 0.7) with those from the Durham sampler when pollen dispersal was high, but a poor correlation (r 0.7 during the pollen season. The accuracy of automatic pollen counting can be improved using a correction to include rates of underestimation and misidentification in a particular geographical area.