BACKGROUND: The appropriate landmark of spinal puncture is important for preventing spinal cord injury. L4 spinous process (SP) is the largest in size and L5 is the smallest. In this study 'height of SP' is the longitudinal length of SP on lumbar AP view. The purpose of this study was to identify the L4-5 interspinous space (ISS) using difference between L4, 5 SP heights. METHODS: Sixty-six patients scheduled for spine surgery were enrolled. After induction of general anesthesia, patients were changed to a prone position. The lumbar vertebrae were palpated from the lowest point of the lumbar spine and cranially. We palpated the difference in spinous process heights and marked the point of step-off from L4 SP to L5 SP. The level was radiologically confirmed. Direction of estimation error and the effects of spondylolisthesis, sex, and obesity were also analyzed. RESULTS: The number of accurate identification of the L4-5 ISS in males was 36 (85.7%), in females was 17 (70.8%), and in total 53 (80.3%). The difference between L4 and L5 SP heights (DL4-5SPHs) and sex affect the results. DL4-5SPHs were larger in concordant patients than in discordant patients (7.2 +/- 4.9 mm: 5.0 +/- 1.6 mm, P < 0.05). Among errors, there were more cephalad identified cases compared to caudad identified ones (12.1%: 7.6%). CONCLUSIONS: The identification of L4-5 ISS using SP height difference was considerably accurate. So, we considered this might be a useful method for spinal anesthesia.