The ideal technique for identifying the epidural space remains unclear. Five-hundred-forty-seven women in labor who requested epidural analgesia were randomly allocated to three groups according to the technique by which the epidural space was identified: 1) loss-of-resistance with air (air; n = 180), 2) loss-of-resistance with lidocaine (lidocaine; n = 185), and 3) loss-of-resistance with both air and lidocaine (air-plus-lidocaine; n = 182). We assessed ease of epidural catheter insertion, characteristics of the blockade, quality of analgesia, and complications. The inability to thread the epidural catheter occurred in 16% of the air, 4% of the lidocaine, and 3% of the air-plus-lidocaine patients (P < 0.001). More patients from the air group had unblocked segments (6.6% versus 3.2% and 2.2%, respectively; P < 0.02). The incidence of accidental dural puncture was greater in the air group (1.7% versus 0% in the other two groups; P < 0.02). Pain scores, time to onset of analgesia, upper sensory level, motor blockade, and the incidence of hypotension, transient neurological deficits, postpartum urinary retention, and postdural puncture headache were comparable. Identification of the epidural space with air was more difficult and caused more dural punctures than with lidocaine or air plus lidocaine. Additionally, sequential use of air and lidocaine had no advantage over lidocaine alone.