Although PCB in caulking materials has been forbidden for many years in most of Europe, including Denmark, there has been continued interest to measure PCB levels in the air of contaminated buildings and blood of the occupants (Mengon and Schlatter 1993, Fromme et al. 1996, Ewers et al. 1998, Currado and Harrad 1998, Gabrio et al. 2000). The relatively low priority for investigations of this contamination is probably due to the small quantities inhaled compared to exposure via food, and the rapid metabolism of the most volatile congeners demonstrated by low concentrations of all congeners in the blood of exposed persons (Ewers et al. 1998, Gabrio et al. 2000). There is, however, evidence that PCB containing caulking materials have been used even during the '90s (Fromme et al. 1996). In Denmark, it is estimated that 75 t PCB is still in buildings (Organization of Sealant Branch's Manufacturers and Distributors 2000). During an investigation of dust from buildings with excessive microbial growth (including 35 rooms from 9 buildings), the analysis of semivolatile compounds by thermal desorption-GC/MS of samples from a single building surprisingly revealed large amounts of PCBs containing 3, 4 and 5 chlorine atoms, 10-20 times the amounts found in samples from other buildings. Extraction of the dust by SFE followed by GC/ECD analysis for 12 PCB congeners showed that there was approximately 20 times the total PCB concentrations in dust from the polluted building compared to the levels in the other buildings. Subsequent headspace analysis of caulking material from the polluted building revealed this to be the source. Shelf dust functions as a passive sampling medium and, thus, can be used as a screening method to detect PCB and other semivolatile pollution indoors.