A prospective investigation was performed to determine when to remove a suction drain following total knee arthroplasty (TKA). Forty-one TKAs were randomly allocated to closed suction drainage for either 24 or 48 hours. The drain was removed and the tip was cut off and processed by a method giving quantitative cultures. In the 48-hour group, 85% of the total volume was drained during the first 24 hours. During the following 24-hour period, a mean volume of only 50 ml was drained. No organism was isolated from cultures of drain tips sampled at 24 hours. However, at 48 hours, 25% of the drain tips yielded light growths of coagulase-negative staphylococci (four drain tips) and Staphylococcus aureus (one drain tip). Clinical evaluations of wound healing were comparable in the two groups. Clearly, nothing is to be gained by continuing drainage beyond 24 hours. If drainage is maintained for longer periods, there is an increased risk of contamination by bacteria.