To evaluate the validity of methanol (MeOH) and formic acid (FA) in urine as biological indicators of methyl formate (MF) exposure in experimental and field situations. The subjects were 28 foundrymen and two groups of volunteers (20 control and 20 exposed). Exposure assessment of the workers was performed by personal air and biological monitoring. Methyl formate vapour collected on charcoal tube was analysed by gas chromatography. The concentration of MF in the exposure chamber (volunteer-study) was monitored by two independent methods [flame ionisation detection (FID) and Fourier transformation infra-red detection (FTIR)]. Urinary metabolites (MeOH and FA) were analysed separately by headspace gas chromatography. The volunteers exposed to 100 ppm MF vapour at rest for 8 h excreted 3.62 +/- 1.13 mg MeOH/l (mean +/- SD) at the end of the exposure. This was statistically different (P < 0.001) from pre-exposure MeOH excretion (2.15 +/- 0.80 mg/1), or from that of controls (1.69 +/- 0.48 mg/l). The urinary FA excretion was 32.2 +/- 11.3 mg/g creatinine after the exposure, which was statistically different (P < 0.001) from pre-exposure excretion (18.0 +/- 9.3 mg/g creatinine) or that of controls (13.8 +/- 7.9 mg/g creatinine). In foundrymen, the urinary FA excretion after the 8 h workshift exposure to a time weighted average (TWA) concentration of 2 to 156 ppm MF showed a dose-dependent increase best modelled by a polynomial function. The highest urinary FA concentration was 129 mg/g creatinine. The pre-shift urinary FA of the foundrymen (18.3 +/- 5.6 mg/g creatinine) did not differ from that of controls (13.8 +/- 7.9 mg/g creatinine). The urinary MeOH excretion of the foundrymen after the shift, varied from < 1 to 15.4 mg/l, while the correlation with the preceding MF exposure was poor. The foundrymen excreted more (P = 0.01) FA (2.12 +/- 3.56 mg/g creatinine) after the workshift than experimentally, once-exposed volunteers (0.32 +/- 0.11 mg/g creatinine) at a similar inhaled MF level of 1 ppm). In spite of its high background level in non-exposed subjects, urinary FA seems to be a useful biomarker of methyl formate exposure. The question remains as to what is the reason for the differences in chronic and acute exposure respectively.