In October 1976 a randomized, paired, prospective study comparing the compression sliding hip screw with Ender's pins in the treatment of extracapsular hip fractures was begun on the orthopaedic service at San Francisco General Hospital. The results in 100 patients with a minimum follow-up of six months showed that Ender's procedure required less operative time, had less blood loss, and had a lower postoperative incidence of medical complications. A 26 per cent incidence of technical problems with Ender's pins (such as postoperative backing-out of the pin) required reoperation in 16 per cent of the patients compared with an 8 per cent rate of reoperation in patients treated with the compression screw. The majority of these complications occurred in unstable fractures. There were no deep infections in the patients with Ender's pins but a 6 per cent incidence of deep infection was seen in those with the hip screws. There were no non-unions in either group. Although hip function at follow-up was equal in both groups, there was a 41 per cent incidence of postoperative pain and stiffness of the knee in the group with Ender's pins. It is our opinion that Ender's pins are a valuable addition to the armamentarium of surgeons treating extracapsular fractures of the hip and are ideally suited for elderly patients with stable fractures, particularly if the surgical risk is high. They must be used with caution in unstable fractures and postoperative protection in traction may be necessary. However, the occasional external rotation deformity and the high incidence of problems with the knee makes their use in younger, more active patients less desirable than the compression hip screw unless their unique advantages justify their use.