For the histopathological classification of the severity of osteoarthritic lesions of cartilage, the Mankin score is frequently used. A necessary constraint on the validity of this scoring system is the consistency with which cartilage lesions are classified. The intra- and interobserver agreement of the Mankin score was determined. The intra- and interobserver agreement of the 14-point Mankin score was adequate. Between observers 95% of differences were less than approximately 7 points. By a more strict definition of the elements of the Mankin score, the intraobserver differences were reduced only for some observers. The interobserver differences were only slightly reduced: between observers 95% of differences were less than approximately 6 points. We found the Mankin score to be an adequate histopathological tool.