Ambient aerosol has been identified as a major pollutant affecting human health. Standards to reduce particles mass concentrations have therefore been established in many countries. Recent studies suggest that the number concentration of aerosol particles, which is dominated by the ultrafine size range smaller than 100 nm in diameter, may be independently associated with health effects. Currently, epidemiological evidence for such effects is conflicting. We have measured aerosol size distributions at two stations (urban background, street canyon) located at a distance of 1.5 km for a time period of 1 year. Number concentrations and particle size distributions at both sites were significantly different. Short-term correlation between the two sites was weak for individual measurements of number concentrations and size bins of ultrafine particles (0.19-0.46). Correlation coefficients for hourly and daily averages in selected size ranges ranged from 0.35 to 0.46. On the other hand, the correlation coefficient for daily average particle volume concentrations was found to be 0.67. About 10% to 20% of the population of European cities lives close to roads with traffic densities comparable to our site. The underestimation of the exposure of a considerable part of a study population may therefore severely influence the outcome of epidemiological studies focused on health effects associated with ultrafine particles. A single background measurement site may not be sufficient for exposure assessment in these studies without taking spatial and temporal variability into account.