In a prospective study of 100 patients with internal derangement of the knee, clinical diagnosis, arthrography, arthroscopy, and arthrotomy findings were compared. The clinical diagnosis was correct in seventy-two patients, correct but incomplete in ten, and incorrect in eighteen. Arthroscopy was accurate in ninety-four, influenced surgical therapy in fifty-five, and revealed unexpected disease in twenty-five patients. It was found to be critical for diagnosis in sixteen patients. Twenty-three additional patients were analyzed who underwent arthroscopy but not arthrotomy. Unnecessary operations were avoided in twenty-one of these twenty-three patients.