This report describes two studies which compared the results of the analyses of four heterocyclic aromatic amines (HAAs): 2-amino-3-methylinidazo[4,5-f] quinoline (IQ); 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx); 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (4,8-DiMeIQx) and 2-amino-1-methyl-6-phenylimidazo [4,5-b]pyridine (PhIP), first as pure methanolic solutions and, in a second step, in a food matrice (beef extract) spiked with known amounts of these four HAAs. Details are given for the preparation of the methanolic solutions of the four HAAs and for the homogeneity and stability studies. The results of different statistical treatments revealed no significant heterogeneity within or between ampoules. The results of the stability studies clearly indicated that, except for PhIP, no effect of storage period (up to 6 months) or storage temperature (up to 25 degrees C), existed for the five HAAs methanolic solutions. Each of the eight European participating laboratories, which has leading experience in the analysis of HAAs, received sealed ampoules containing the pure reference solutions of the four HAAs together with a mixture of unknown identity and concentration. For the analysis of the unknwon sample, participants followed, a common work programme, but used different columns, solvent gradients and detection systems (UV, fluorescence, mass spectrometry and electrochemical detection. The analysis of the results of this first comparison revealed a good correlation between the results provided by the participants and high precision regarding the target values, independent of the experimental conditions used. For the second comparison, a common batch of commercial beef extract was prepared and spiked with known amounts of the four HAAs. The long-term stability study at -18 degrees C, 4 degrees C, 25 degrees C, 40 degrees C and 60 degrees C revealed high stability of these four HAAs, during up to 6 months of storage. At 40 degrees C and 60 degrees C, however, a significant loss was observed, in particular for PhIP. On the other hand, the 1-year stability study revealed that the HAAs content did not change when beef extract was stored at -18 degrees C. Details of these homogeneity and stability studies are provided. The sealed ampoules containing beef extract, together with the reference methanolic solutions were sent to the participants in refrigerated container. The eight European laboratories, which participated in the first comparision, adopted the work programme of this exercise. They generally followed a previously agreed upon solid-phase extraction procedure, very similar to that described by Gross [8], with analysis by HPLC. Column conditions, solvent elution and detection by UV, fluorescence, mass spectrometry and electrochemical detection varied between laboratories. The objectives of this second phase of the project were to compare and improve usual routine laboratory methods for the determination of IQ, MeIQx, 4,8-DiMeIQx and PhIP in the range of 1-30 ng/g, in a commercial beef extract. The comparision of the results revealed, however, large variations, not only beween but also within laboratories. Major difficulties were encountered by the participants, mainly for the determination of PhIP. Acceptable recovery levels were agreed between participants and different sources of variability in the extraction procedures were identified. In conclusion, whereas the analytical determination of HAAs in beef extract appeared to be satisfactory, the procedures of isolation and purification require further improvement.