PURPOSE: To determine the MR imaging criteria by which bland and proliferative effusion of the knee may be differentiated. MATERIALS AND METHODS: We retrospectively reviewed the MR images of 64 patients (65cases), in whom T2-weighted sagittal scans revealed anteroposterior distension of the suprapatellar bursa of at least 0.5cm. The patients were divided into two groups: bland effusion (n=36) , and proliferative effusion [(n=29); pigmented villonodular synovitis (n=5), rheumatoid arthritis (n=6), septic arthritis (n=6), chronic synovitis (n=5), gouty arthritis (n=3), tuberculous arthritis (n=2), and lipoma arborescens (n=2)]. All conditions were diagnosed on the basis of operative data or clinical criteria. The knee joint space was divided into four compartments: the suprapatellar pouch, central zone, posterior femoral recess, and subpopliteal recess, and the amount and distribution of effusion was then compared between the two groups. The ratios of the width and the length of the lateral recess of the suprapatellar bursa to those of its medial recess were deter mined, and the findings for the two groups were compared. Abnormality of the intracapular fat pads (prefemoral fat, Hoffa 's fat, and quadriceps fat sign) as seen on sagittal scans, is a predictor proliferative effusion, and any such abnormality was evaluated. The synovium was classified as either thin or thick, and as having either a smooth or an irregular margin, as seen on Gadolinium-enhanced T1W1 images. RESULTS: As compared with bland effusion, proliferative effusion involved more prominent joint effusion in the suprapatellar pouch and posterior femoral recess, and in the suprapatellar bursa, the ratio of the width of the lateral recess to that of the medial recess was greater. When comparing the ratio of the length of the lateral recess to that of the medial recess, however, no significant statistical difference was noted. Sensitivity: specificity for proliferative effusion was 58%: 86% on coronal scan and 64%: 93% on axial scan at a threshold value of 0.7 (the ratio of the width). The prefemoral fat pad sign was 41% sensitive and 100% specific, while Hoffa 's fat pad sign had a sensitivity of 32% and a specificity of 95%. The corresponding figures for the quadriceps fat pad sign were 14% and 100%. The pattern of the synovium in bland effusion was thin and smooth in two, thick and smooth in one, and thin and irregular in one. In proliferative effusion, the pattern was thick and smooth in 11 cases, thin and irregular in four, and thick and irregular in 14. CONCLUSION: In proliferative effusion, the synovium tended to be thick and irregular. Proliferative effusion demonstrated greater predilection for the suprapatellar pouch, especially the lateral recess, and posterior femoral recess, than did bland effusion. Difference in the distribution of joint effusion effectively predict both proliferative effusion as well as intracapsular fat pad signs.