In a prospective, randomized study, 34 patients (25 male, 9 female; mean age 27 +/- 8 years) with unilateral anterior cruciate ligament ruptures and arthroscopic reconstruction with patellar tendon grafts were allocated at random to either early active motion only (AM; n = 17) or active motion in combination with continuous passive motion (CPM; n = 17). Range of motion was measured with a goniometer and joint swelling with a tape measure, preoperatively and at 6 weeks postoperatively. Neither associated injuries nor the age of the patients differed in the two groups. There was no difference in the range of motion between the two groups at 6 weeks' follow-up. Joint swelling was more pronounced in the AM group both preoperatively and at 6 weeks' follow-up. In this investigation the range of motion was not improved by CPM. The difference in joint swelling between the two groups may be explained by a persistent preoperative variation in joint effusion due to an imbalanced distribution of acute and chronic cases.