Aims: Firm bandaging of the knee following knee replacement may prevent bleeding into the joint by a tamponade effect, but the presence of a tight bandage around the knee may obstruct venous return and lead to thromboembolic complications. We have studied the pressure required, and then compared clinically the use of a compression bandaging with the use of either a drain, or a standard crepe bandage. Method: Transducers were used to measure the pressure achieved on the surface of the knee under different bandages, and within the knee following release of tourniquet. Three series of 50 patients have been compared: with compression bandaging from toes to mid-thigh, with crepe alone, or with suction drain and crepe. Results: The pressure within the joint at which tamponade occurs is 52–57 mm Hg. The pressure on the skin under a properly applied compression bandage is between 28 and 32 mmHg and this controls bleeding within the joint. Patients treated with compression bandaging recovered quicker from the operation had a shorter hospital stay and a greater range of ßexion on discharge. They had no swelling of the limb, rarely suffered a tense haemarthrosis and had fewer complications. Conclusions: The use of compression bandage incorporating the foot and calf following knee replacement surgery confers speciþc advantages over the use of crepe bandage alone.