Ambient and biological monitoring of hexane exposure were repeatedly carried out in 14 female shoe makers. Airborne hexane (Ci-H) was measured in 4-h samples collected by a diffusive method. Urinary spot samples were collected before, during (at noon), and at the end of a work shift. 2,5-Hexanedione (2,5HD) in urine collected at noon was poorly related to morning Ci-H. End-of-shift 2,5HD were also poorly related to afternoon air samples. The correlation was still relatively low when end-of-shift 2,5HD was related to 8-h TWA Ci-H (r = 0.44; P < 0.01 on a linear scale, and r = 0.58, P < 0.01 on a log-log scale). End-of-shift 2,5HD levels estimated on the basis of pre-shift values using a mathematical model were much higher (2.3 times on average) than those experimentally measured during the study period. Owing to its relatively long half-time, 2,5HD seems to be influenced not only by current exposure, but also by hexane absorbed during the day(s) preceding sampling. The lack of a sampling strategy may account not only for inconsistencies between environmental and biological data, but also for a possible misuse of biological monitoring when utilized for risk assessment. Despite sometimes poor correlations with Ci-H, 2,5HD may still be preferred to other indicators as a marker of effective internal dose. A sampling strategy should ensure that measured values are representative of the individual risk for adverse effects.