Background. We tested the hypothesis that ultrasound guidance may reduce the minimum effective anaesthetic volume (MEAV(50)) of ropivacaine 0.5% required to block the femoral nerve compared with nerve stimulation guidance. Methods. After standard premedication and sciatic nerve block were given, 60 patients undergoing knee arthroscopy were randomly allocated to receive a femoral nerve block with ropivacaine 0.5% using either nerve stimulation (group NS, n=30) or ultrasound (group US, n=30) guidance. The volume of the injected solution was varied for consecutive patients based on an up-and-down staircase method according to the response of the previous patient. The initial volume was 12 ml. A double-blinded observer evaluated the occurrence of complete loss of pinprick sensation in the femoral nerve distribution, with concomitant block of the quadriceps muscle: positive or negative responses within 30 min after the injection determined a 3 ml decrease or increase for the next patient, respectively. Results. The mean (SD) MEAV(50) for femoral nerve block was 15 (4) ml (95% CI, 7-23 ml) in group US and 26 (4) ml (95% CI, 19-33 ml) in group NS (P=0.002). The effective dose in 95% of cases (ED95) calculated with probit transformation and logistic regression analysis was 22 ml (95% CI, 13-36 ml) in group US, and 41 ml (95% CI, fs 24-66 ml) in group NS. Conclusions. Ultrasound guidance provided a 42% reduction in the MEAV of ropivacaine 0.5% required to block the femoral nerve as compared with the nerve stimulation guidance.