This study characterizes the personal, indoor, and outdoor PM2.5, PM10, and PM2.5-10 exposures of 18 individuals with chronic obstructive pulmonary disease (COPD) living in Boston, MA. Monitoring was performed for each participant for six consecutive days in the winters of 1996 or 1997 and for six to twelve days in the summer of 1996. On each day, 12-h personal, indoor, and outdoor samples of PM2.5 and PM10 were collected simultaneously. Home characteristic information and time-activity patterns were also obtained. Personal exposures were higher than corresponding indoor and outdoor concentrations for all particle measures and for all seasons, except for winter indoor PM2.5-10 levels, which were higher than personal and outdoor levels. Higher personal exposures may be due to the proximity of the individuals to particle sources, such as cooking and cleaning. Indoor concentrations were associated with both outdoor concentrations and personal exposures (as determined by individual least square regression analyses), with associations strongest for PM2.5. Indoor PM2.5 concentrations were significantly associated with outdoor and personal levels for 12 and 15 of the 17 individuals, respectively. Both the strength and magnitude of the associations varied by individual. Also, personal PM2.5, but not PM2.5-10, exposures were associated with outdoor levels, with 10 of the 17 subjects having significant associations. The strength of the personal-outdoor association for PM2.5 was strongly related to that for indoor and outdoor levels, suggesting that home characteristics and indoor particulate sources were key determinants of the personal-outdoor association for PM2.5. Air exchange rates were found to be important determinants of both indoor and personal levels. Again, substantial interpersonal variability in the personal-outdoor relationship was found, as personal exposures varied by as much as 200% for a given outdoor level.