This study was undertaken to measure the passive smoking exposure of prisoners at three correctional facilities in the US and to evaluate the effectiveness of a ban on smoking in reducing these exposures at two of these facilities. The average weekly concentration of nicotine was measured in fixed locations within the correctional facilities using passive samplers. Samples were collected before and after a smoking ban was instituted, and after the policy was modified to allow smoking outdoors. Samples were collected in the living areas, near where inmates slept and watched TV, and in selected central facilities, including dining halls, visiting rooms, booking areas, and learning centers. Average weekly concentrations of nicotine were measured in 84 locations while smoking was allowed; changes in these concentrations were measured with 112 weekly samples 4 and 9 months after the policy restricting smoking was implemented The average concentrations of nicotine were high while smoking was allowed: most living and sleeping areas averaged 3-11 microg/m(3), but the gym that was used as a bunkroom averaged 25 microg/m(3); these values compare to an average of 2 microg/m(3) in the homes of smokers. The smoking ban significantly reduced nicotine concentrations in the living areas (P<0.01 at facility A and P<0.05 at facility B) to averages of 1.5-2.2 microg/m(3); all postban samples were less than 5 microg/m(3). In conclusion, secondhand smoke concentrations in correctional facilities can be quite high; however, policies banning smoking are effective in reducing, but not eliminating, these exposures.