Experience with a quality control programme for a multicentric collaborative study of antibiotic susceptibility of Neisseria gonorrhoeae is described. The Australian Gonococcal Surveillance Programme (AGSP) has monitored the antibiotic sensitivity of gonococci by surveying results of quantitative sensitivity tests performed in a number of centres throughout the country since 1981. A standardised system of testing by means of agar plate dilution was adopted to ensure uniformity of methods and results and to permit temporal and regional differences in antibiotic sensitivity to be identified and compared. A control programme was developed to provide quality assurance regarding each laboratory's performance. Over a period of nine years more than 3600 MIC determinations on control strains, examined as unknowns, were performed. An overall error rate of 3.1% was observed. Higher error rates were noted in the beginning of the programme but these declined as the study progressed. In addition repeat and duplicate testing of strains was performed. The results of these tests also indicated an improvement in performance over the study period. Close attention to all aspects of the testing procedure reduced the number of errors recorded. At the beginning of the programme, various factors were investigated as possible causes of error. The formulation and preparation of media were found to be potential sources of result variation. Once these variables were controlled, lack of experience with end-point determination and errors in preparation of dilutions of antibiotic solutions emerged as the commoner causes of discrepancy. In the experience of the AGSP, a long-term multicentric evaluation of gonococcal sensitivity can achieve accurate and reproducible results through standardisation of methods and application of a continuing quality assurance scheme.