n-Hexane is one of the solvents widely used in industry and well known to be neurotoxic. Recently it was clearly revealed that n-hexane is metabolized in vivo and its metabolites are excreted in the urine. However, the relationship between the exposed dose of n-hexane and the metabolites in the urine has not yet been substantially determined. Therefore, in this investigation we intended to clarify the above relationship in order to establish its usefulness for biological monitoring of n-hexane exposure. The exposed dose was measured by means of a personal monitoring badge worn by workers in seven factories manufacturing vinyl sandals. The time-weighted average (TWA) concentration of n-hexane was 0.2-47.4 ppm. The n-hexane metabolites in the urine of 22 workers were measured with modified Perbellini's method [12] in the early morning (6:00-7:00 hrs) and at 17:00 hrs. 2,5-Dimethylfuran, 2,5-hexanedione and gamma-valerolactone were identified by gas chromatography and mass spectrometory. At 17:00 hrs the means +/- SD of the metabolites were 0.21 +/- 0.11 mg/l for 2,5-dimethylfuran, 1.13 +/- 0.71 mg/l for 2,5-hexanedione, and 2.04 +/- 2.31 mg/l for gamma-valerolactone. The metabolites were also found in the urine in the early morning. 2-Hexanol was not detected in the urine of any worker examined. A strong correlation between TWA concentration of n-hexane and 2,5-hexanedione in the urine was found at 17:00 hrs (r = 0.895, P less than 0.001). The results suggest that the urinary metabolites of n-hexane, especially 2,5-hexanedione, could be useful indicators for biological monitoring of n-hexane exposure. Furthermore the present study offers the advantage of a better estimate of n-hexane TWA.