Objectives-This study was undertaken to assess reliable biological indicators for monitoring the occupational exposure to inorganic arsenic (iAs), taking into account the possible confounding role of arsenicals present in food and of the element present in drinking water. Methods-51 Glass workers exposed to As trioxide were monitored by measuring dust in the breathing zone, with personal air samplers. Urine samples at the end of work shift were analysed for biological monitoring. A control group of 39 subjects not exposed to As, and eight volunteers who drank water containing about 45 µg/l iAs for a week were also considered. Plasma mass spectrometry (ICP-MS) was used for the analysis of total As in air and urine samples, whereas the urinary As species (trivalent, As Conclusions-To avoid the eVect of As from sources other than occupation on urinary species of the element, in particular on DMA, it is proposed that urinary As 3 +As 5 may an indicator for monitoring the exposure to iAs. For concentrations of 10 µg/m 3 the current environmental limit for iAs, the limit for urinary As 3 +As 5 was calculated to be around 5 µg/l, even if the wide variation of values needs critical evaluation and application of data. The choice of this indicator might be relevant also from a toxicological point of view. Trivalent arsenic is in fact the most active species and its measure in urine could be the best indicator of some critical eVects of the element, such as cancer. (Occup Environ Med 1999;56:825-832)