Personal exposures to ozone were measured in the summers of 1992 and 1993 in the Fraser Valley, a suburban and rural river valley bordering the metropolitan region of Vancouver, British Columbia (population approximately 1.8 million). In 1992, two groups of 25 healthy individuals were selected for 14 consecutive days of 24-hour personal monitoring, based on prior expectations of their activity patterns. The first group, composed of adult healthcare workers, was expected to spend most of the work day indoors or commuting. The second group, teenage camp counselors, was expected to spend more time outdoors. Time-activity data were collected to investigate the association between activity patterns and ozone exposures. In 1993, 15 adult farmworkers wore personal ozone samplers during their workday, which was spent entirely outdoors. Personal monitoring was conducted with a nitrite-coated filter passive ozone sampler described by Koutrakis and colleagues. Ozone collection rates were determined empirically by collocation of the passive samplers with continuous ozone analyzers at fixed sampling sites. Sampler performance was assessed by collecting several duplicate personal samples. Although ozone levels were low (< 35 ppb 24-hour average) during both summer sampling periods, the passive ozone samplers agreed well with collocated UV absorption ozone measurements. Based on replicate personal samples, we estimated that personal measurements differing by more than 35% were associated with true differences in exposure. Within the two groups with varying time activities, differences in ozone exposures were associated with the fraction of time spent outdoors. Similarly, differences in ozone exposures were observed between the three sampling populations, with higher exposures recorded in the group that spent more time outdoors.(ABSTRACT TRUNCATED AT 250 WORDS)