The reliability and accuracy of two-compartment temporomandibular joint (TMJ) arthrography was compared with MR imaging on the basis of an analysis of surgical findings obtained from joints that had been studied preoperatively with arthrography or MR or, in some cases, both procedures. Seven hundred forty-three consecutive TMJ arthrograms were successfully obtained in a total of 443 patients by using a Single 27gauge needle and a two-compartment technique in each joint. There was a 100% correlation with surgical findings in 218 radiologically abnormal joints operated on within 90 days of arthrography with respect to the presence or degree of meniscus displacement and normal or abnormal disk morphology and function. In 604 patients 1052 TMJs were studied with high-field-strength surface-coil MR. Surgical findings were available for correlation in 170 of the joints studied. Forty-three joints were studied with both twocompartment arthrography and MR. Eight operated joints had been imaged successfully with both two-compartment arthrography and MR. Both methods of evaluation provided highly reliable and accurate information regarding meniscus position and shape. Arthrography was superior to MR in detecting capsular adhesions and the presence or absence of perforation of the disk or meniscus attachments. Simple meniscectomy (with or without insertion of a temporary Silastic TMJ implant) was the most frequently performed surgical procedure in the series, followed by meniscus repositioning procedures. Joint effusions, failed TMJ implants, and avascular necrosis were demonstrated best with MR. Soft-tissue lesions, including intrinsic degeneration of the meniSCUS, anomalous muscle development, muscle atrophy, tendinitis, and injuries such as contusions and hematomas, were demonstrated only with MR. Partial-flip-angle GRASS (gradient-recalled acquisition in the steady state) techniques permit both fast scanning and study of functional joint dynamics. Joint fluid may appear as high signal intensity on GRASS images because of T2*-weighting. We recommend MR as the procedure of choice for diagnosis of uncomplicated internal derangements of the TMJ. Two-compartment arthrography with videofluoroscopy is an important ancillary procedure that should be performed whenever capsular adhesions or perforations are suspected and not demonstrated with MR and whenever MR is inconclusive.