The number of ultrafine particles in urban air may be more health relevant than the usually measured mass of particles smaller than 2.5 or 10 μm. Epidemiological studies typically assess exposure by measurements at a central site. Limited information is available about how well measurements at a central site reflect exposure to ultrafine particles. The goals of this paper are to assess the relationships between particle number (PN) and mass concentrations measured outdoors at a central site, right outside and inside the study homes. The study was conducted in four European cities: Amsterdam, Athens, Birmingham and Helsinki. Particle mass (PM10 and PM2.5), PN, soot and sulfate concentrations were measured at these sites. Measurements of indoors and outdoors near the home were made during 1 week in 152, mostly non-smoking, homes. In each city continuous measurements were also performed at a central site during the entire study period. The correlation between 24-h average central site outdoor and indoor concentrations was lower for PN (correlation among cities ranged from 0.18 to 0.45) than for PM2.5 (0.40-0.80), soot (0.64-0.92) and sulfate (0.91-0.99). In Athens, the indoor-central site correlation was similar for PN and PM2.5. Infiltration factors for PN and PM2.5 were lower than for sulfate and soot. Night-time hourly average PN concentrations showed higher correlations between indoor and central site, implying that indoor sources explained part of the low correlation found for 24-h average concentrations. Measurements at a central site may characterize indoor exposure to ambient particles less well for ultrafine particles than for fine particle mass, soot and sulfate. © 2007 Elsevier Ltd. All rights reserved.