Personal exposures to particulate and gaseous pollutants and corresponding ambient concentrations were measured for 56 subjects living in Baltimore, Maryland, and 43 subjects living in Boston, Massachusetts. The 3 Baltimore cohorts consisted of 20 healthy older adults (seniors), 21 children, and 15 individuals with physician-diagnosed chronic obstructive pulmonary disease (COPD*). The 2 Boston cohorts were 20 healthy seniors and 23 children. All children were 9 to 13 years of age; seniors were 65 years of age or older; and the COPD participants had moderate to severe physician-diagnosed COPD. Personal exposures to particulate matter with aerodynamic diameters less than 2.5 microm (PM2.5), sulfate (SO(4)2-), elemental carbon (EC), ozone (03), nitrogen dioxide (NO2), and sulfur dioxide (SO2) were measured simultaneously for 24 hours/day. All subjects were monitored for 8 to 12 consecutive days. The primary objectives of this study were (1) to characterize the personal particulate and gaseous exposures for individuals sensitive to PM health effects and (2) to assess the appropriateness of exposure assessment strategies for use in PM epidemiologic studies. Personal exposures to multiple pollutants and ambient concentrations were measured for subjects from each cohort from each location. Pollutant data were analyzed using correlation and mixed-model regression analyses. In Baltimore, personal PM2.5 exposures tended to be comparable to (and frequently lower than) corresponding ambient concentrations; in Boston, the personal exposures were frequently higher. Overall, personal exposures to the gaseous pollutants, especially O3 and SO2, were considerably lower than corresponding ambient concentrations because of the lack of indoor sources for these gases and their high removal rate on indoor surfaces. Further, the impact of ambient particles on personal exposure (the infiltration factor) and differences in infiltration factor by city, season, and cohort were investigated. No difference in infiltration factor was found among the cohorts, which suggests that all subjects were exposed to the same fraction of ambient PM2.5 for a given ambient concentration. In addition, the results show significant correlations between ambient PM2.5 concentrations and corresponding personal exposures over time and provide further indication that ambient gaseous pollutant concentrations may be better surrogates for personal PM2.5 exposures, especially personal exposures to PM2.5 of ambient origin, than their respective personal exposures. These results have important implications for PM health effects studies that use regression models including both ambient PM2.5 and gaseous pollutant concentrations as independent variables, because both parameters may be serving as surrogates for PM2.5 exposures.