The aim of this article is to describe PM(2.5) personal exposures within the Grenoble population and to identify the implications of such measurements on epidemiological or risk assessment studies of air quality. Non smoking adult volunteers, selected in summer 1996 (n=40), in winter 1997 (n=40) and in summer 1998 (n=20), carried a case containing 2 PM(2.5) personal monitors. One of the monitors was running continuously for 48h. (cumulative exposures), the other one was running only while indoors (indoor exposure). PM(2.5) masses were determined by reflectometry (black smoke method) and by deionised weighting (Mettler MT5 micro-balance; gravimetric method). Cumulative PM(2.5) personal exposures ranged on average from 21.9 in summer to 36.7 microgram/m(3) in winter (arithmetic mean), using the gravimetric results; the dispersion of these personal exposures was greatest in winter (s.d.=23.1 microgram/m(3)) than in summer (s.d.=10.4 microgram/m(3)). There was a good correlation (R=0.7) between the reflectometry and weighting results in winter, but not in summer. Outdoor personal exposures, determined by difference between the cumulated and indoor masses, were slightly higher than the cumulative personal exposures: the geometric means was 49.6 microgram/m(3) (geometric standard deviation=2.7 microgram/m(3)) in summer and 55.1 microgram/m(3) (3.7 microgram/m(3)) in winter (gravimetric results). Due to these greater outdoor concentrations, the fraction of outdoor exposure was high (25%) relative to the small amount of time spent outdoors (less than 10%). These descriptive data, consistent with the literature, show the importance of "expology" studies aiming at characterizing PM(2.5) personal measurements across the year. This would lead, in the future, to optimizing the use of "ecological" estimates of exposures from ambient air concentrations provided by the ambient air quality networks, for the characterization of exposure in epidemiological or risk assessment studies.