Diesel exhaust is a known mutagen and a potential human carcinogen. Recent epidemiological studies have demonstrated a small increase in the risk of lung cancer from diesel exhaust exposure. However, many epidemiological studies have used crude estimates of exposure, and even accurate measures of exposure may not be accurate estimates of the internal dose received. Measurement of diesel exhaust exposure also has been limited by the absence of a standard marker. This study was undertaken to evaluate the usefulness of urinary mutagenicity as a biological marker of diesel exhaust exposure in the workplace. We measured the exposure of individual railroad workers to diesel exhaust by using personal air samples taken over two consecutive work shifts. Nicotine in the samples was measured to adjust the respirable particle concentrations for active and passive cigarette smoking. Urine samples were collected at the end of the study work shifts and were analyzed for markers of cigarette smoking (nicotine, cotinine) and for mutagenicity, using a sensitive microsuspension assay (micro preincubation assay; Salmonella strain TA98 with or without S9 enzyme). The number of cigarettes smoked on the study shift was recorded, and subjects completed a questionnaire at the end of the second day on personal habits and exposures at home and work. Multiple regression analyses were used to analyze independent determinants of urinary mutagenicity, including a generalized least-squares analysis that divided residual variation into between- and within-person components. Eighty-seven subjects completed 151 two-day protocols; an additional four subjects provided usable data for a single day (n = 306 samples). Respirable particle concentration was not a good marker of diesel exhaust exposure when contamination by environmental tobacco smoke existed in the work location, but respirable particle concentration that was adjusted for environmental tobacco smoke correlated with a priori assessments of diesel exhaust exposure by job grouping. Phenanthrene concentration, as a potential marker, was measured in a subset of personal samples, and correlated with known diesel exhaust exposure by job grouping. A constant ratio of phenanthrene to respirable particles in area samples from diesel exhaust-exposed work locations suggested that phenanthrene is promising as a marker for diesel exhaust. Mutagenic activity was also measured from extracts of respirable particles in a few personal filter samples, and this technique may be useful for further investigation in epidemiological studies.(ABSTRACT TRUNCATED AT 400 WORDS)