Vehicle exhausts are a well known source of aromatic hydrocarbon pollution in urban environments. The paper reports the results of environmental and biological monitoring of benzene exposure in traffic wardens carried out over a 5-hour workshift. Subjects (n = 131) were grouped according to smoking habits and job task as follows: group (A) 52 nonsmoking office workers, (B) 43 nonsmoking outdoor workers, subdivided into (B1) 36 working on foot and (B2) 7 cyclists; (C) 20 smokers office workers, (D) 16 smokers outdoor workers, subdivided into (D1) 11 working on foot and (D1) 5 cyclists. The median indoor environmental benzene concentration (26 micrograms/m3, n = 50) was significantly lower than the outdoor concentration (45 micrograms/m3, n = 43) (p < 0.01); median exposure value of cyclists was 78 micrograms/m3 (n = 12). For biological monitoring, urinary excretion of trans,transmuconic acid was determined in spot samples collected at 7:30 h (MAit) and 12:30 h (MAft). The MAftA median value (63 micrograms/l, range 2-242 micrograms/l) was not statistically different from MAftB (74 micrograms/l, range 15-216 micrograms/l), while the MAftB2 value of 96 micrograms/l was higher than both MAftB1 (71 micrograms/l) and MAftA. In group (B) there was a relationship between airborne benzene levels and MAftB excretion (y = 17.2 + 1.1x, r = 0.62, n = 35, p < 0.01). The influence of smoking on urinary MA excretion was studied by comparing the results obtained in all nonsmokers (AB) with smokers (CD). MAftCD (192 micrograms/l) was significantly higher than MAftAB (69 micrograms/l) (p < 0.01). In smokers, statistically significant relationships were observed between urinary excretion of MAft (y, microgram/l) and cotinine (x, microgram/l) (y = 83 + 0.08x, r = 0.73, n = 23, p < 0.01), and smoking (x, number cigarettes/day) (y = 87.4 + 4.4x, r = 0.53, n = 29, p < 0.01). Comparison between MAft and MAit median excretion values, calculated for each of the 6 exposure groups, showed that MAft was always higher than the corresponding MAit value. A rough estimate of the total dose of benzene ("index of exposure", EI) inhaled by each subject during the 5-hour working shift as a consequence of air pollution and smoking was also made. Considering the entire group of subjects, a significant association was observed between EI and MAft values (y = 43.4 + 0.39x, r = 0.65, n = 104, p < 0.01). Individual values of MA it were correlated with MAft according to the equation y = 43.6 + 0.82x (r = 0.62, n = 105; p < 0.01) and were also positively associated with EI values (y = 42.3 + 0.20x; r = 0.55; n = 74; p < 0.01). In conclusion, the results suggest that the measurement of urinary MA excretion is a poor indicator for assessing environmental benzene exposure at levels below 100 micrograms/m3, such as those seen in this study; MA can however be reliably used as a biomarker for higher exposures such as those, for example, due to smoking.