The subacute oral and inhalation toxicity of furfural vapour was studied in Fischer 344 rats to investigate whether route-to-route extrapolation could be employed to derive the limit value for inhalation exposure from oral toxicity data. Groups of 5 rats per sex were treated by gavage daily for 28 days at dose levels of 6-192 mg/kg bw/day, or exposed by inhalation to concentrations of 20-1280 mg/m3 (6 h/day, 5 days/week) or 160-1280 mg/m3 (3 h/day, 5 days/week) for 28 days. Controls received vehicle (corn oil) or were exposed to clean air. Daily oral treatment with the highest dose of furfural (initially 192 mg/kg bw/day, later reduced to 144 mg/kg bw/day and finally to 120 mg/kg bw/day) resulted in mortality, and in increases in absolute and relative kidney and liver weight in surviving females of this group. Exposure of rats by inhalation for 6 h/day, 5 days/week for 28 days induced mortality at concentrations of 640 mg/m3 and above within 1-8 days. At 640 mg/m3 (3 h/day) and at 320 mg/m3 (3 and 6 h/day) and below, however, exposure was tolerated without serious clinical effects. In contrast, histopathological nasal changes were seen even at the lowest concentration of 20 mg/m3. With increasing exposure concentration, the nasal effects increased in incidence and severity and also expanded from the anterior part to the posterior part, including the olfactory epithelium. It was concluded that the no-observed-adverse-effect level (NOAEL) for oral toxicity was 96 mg/kg bw/day. The NOAEL for systemic inhalation toxicity was comparable, i.e. 92 mg/kg bw/day (corresponding to 320 mg/m3 (6 h/day) or 640 mg/m3 (3 h/day)) assuming 100% absorption. The presence of the histopathological nasal changes at the lowest tested concentration of 20 mg/m3 (corresponding to 6 mg/kg bw/day) proves that for locally acting substances like furfural extrapolation from the oral to the inhalation route is not valid.