A previously established method for the analysis of chlortetracycline by liquid chromatography using polystyrene-divinylbenzene stationary phases was examined in a multicentre study involving four laboratories and a total of 12 columns. Three chlortetracycline hydrochloride samples were analysed. The main component and the impurities were determined. An analysis of variance, treating each column as a different laboratory, showed absence of consistent laboratory bias and presence of significant laboratory-sample interaction. Estimates for the repeatability and reproducibility of the method, expressed as relative standard deviations of the result of the determination of chlortetracycline hydrochloride, were calculated to be 0.7 and 1.2%, respectively. When the analysis of variance was performed using only the results obtained on the wide pore (1000 A) stationary phases, the laboratory-sample interaction strongly decreased. It is therefore proposed to use such materials for the analysis of chlortetracycline