In groups of six, 30 female subjects were exposed for 4.8 h in a low-polluting office to each of two conditions the presence or absence of 3-month-old personal computers (PCs). These PCs were placed behind a screen so that they were not visible to the subjects. Throughout the exposure the outdoor air supply was maintained at 10 l/s per person. Under each of the two conditions the subjects performed simulated office work using old low-polluting PCs. They also evaluated the air quality and reported Sick Building Syndrome (SBS) symptoms. The PCs were found to be strong indoor pollution sources, even after they had been in service for 3 months. The sensory pollution load of each PC was 3.4 olf, more than three times the pollution of a standard person. The presence of PCs increased the percentage of people dissatisfied with the perceived air quality from 13 to 41% and increased by 9% the time required for text processing. Chemical analyses were performed to determine the pollutants emitted by the PCs. The most significant chemicals detected included phenol, toluene, 2-ethylhexanol, formaldehyde, and styrene. The identified compounds were, however, insufficient in concentration and kind to explain the observed adverse effects. This suggests that chemicals other than those detected, so-called 'stealth chemicals', may contribute to the negative effects.