Measurements of particle number concentration were made in pulse count mode of a condensation nucleus counter (CNC) in order to evaluate particle coincidence error at high concentrations. A screen-type diffusion battery was used to change particle concentration by a constant factor. Existing correction theory was tested against the experimental data. The results showed that two parameters related to the viewing volume of the CNC was very important for the correction. One of them is sampling flow rate, and the other is effective time particles spend m the viewing volume. When the coincidence is appropriately corrected, concentration measurable in pulse count mode of the CNC (TSI model 3760) can be extended to 105 particles/cm3.