Eight lower extremities from cadavera were tested for anterior-posterior laxity in two positions before and after transection of the anterior cruciate ligament. At critical points in the tests, electrogoniometric and radiographic measurements of tibiofemoral translation were compared. By direct measurement, we determined the accuracy of the radiographic method to +/- 0.4 millimeter (95 per cent) in measuring anterior-posterior translations of the tibia with respect to the femur. The electrogoniometer estimated displacement of the tibia with respect to the femur during the anterior drawer test to be 3.5 +/- 8.2 millimeters at 90 degrees of flexion of the knee and 11.1 +/- 16.1 millimeters at 30 degrees of flexion. Direct comparison of these measurements with those obtained by means of the radiographic technique showed that the electrogoniometer tended, on average, to overestimate the tibial translation. The amount of overestimation was 0.7 millimeter for intact knees and 1.9 millimeters after sacrifice of the anterior cruciate ligament. Despite this small average error in measurement of tibial translation, the difference between individual electrogoniometric and radiographic measurements varied greatly, with a 95 per cent confidence limit of +/- 5.5 millimeters. The error of the electrogoniometric measurements varied with the angle of flexion of the knee during testing, both the accuracy and the reliability of the electrogoniometric measurements being greatly diminished at 30 degrees of flexion. The electrogoniometric method also tended to overestimate tibial internal rotation (by an average of 10.5 degrees) and external rotation (by an average of 9.3 degrees); the reliability of these measurements was +/- 6.9 degrees.(ABSTRACT TRUNCATED AT 250 WORDS)