During the Belgian PCB crisis (1999/2000) the quality of the data from the PCB monitoring were studied with a proficiency testing experiment. Pork fat that was spiked at the Institute for Reference Materials and Measurements (EC-JRC-IRMM) was sent out as unknown to all laboratories that participated in the monitoring. In parallel, the material was certified on the basis of the results of several leading PCB laboratories throughout Europe that did not participate in the monitoring. During the first round 15 to 25% deviation was experienced, while in the end of this multistep intercalibration procedure the average deviation for the respective PCBs was 10.9% (PCB 28), 13.1% (PCB 52), 10.1% (PCB 101), 10.7% (PCB 118), 10.7% (PCB 138), 9.1% (PCB 153), 8.1% (PCB 180) and 8.2% for the sum of the 7 PCBs. The concentrations measured for the higher volatile PCB 28 had a tendency to show lower levels, while for PCB 180 this was less pronounced. On the other hand, PCB 153 showed results rather to the higher side. During this exercise the material was also tested for stability by one of the certifying laboratories.