Clinical tests and examinations of knee joint have been expounded fully and frequently in medical textbooks; however, few researchers have attempted to evaluate their credit through statistical analysis. This prospective, cross – sectioual study has been carried out in 2001-2002 on 163 patients (overall 171 arthroscopies) to assess the accuracy and validity indices of clinical examinations. Data on age, sex, clinical symptoms, pre-and per anesthetic physical examinations and arthroscopic findings were collected Clinical presentations of the patients, in the order of prevalence, included pain (98.9%), locking knees (61.2%), swelling (33.9%), limited range of movement (11.7%) and clicks (3.5%). 75.5% were male and the rest female. The most prevalent group was 20 -39 age group. Arthroscopy done on the right and left knee were 50.9% and 49.1% respectively. ACL rupture was the most common in patients with locking plus meniscus tear and those with giving way. In patients with medial meniscus tear, joint line tenderness was primarily on the medial side of the knees; however, in patients with lateral meniscus tear, the most common side of joint line tenderness was on the opposite side (lateral side). The validity indices of knee stress tests (including anterior drawer test, Lachman test and McMurray test) were statistically analyzed and determined. McMurray test (for meniscus tear) and anterior drawer and Lachman tests (for ACL rupture) had a high specificity and positive predictive value Sensitivity of McMurray test decreased under general anaesthesia; while those of the other two tests increased significantly. Finally, a new, useful and simple clinical classification system is suggested for patients with internal derangement of knee.