There are two theories regarding the origin of the lumbar cerebrospinal¯uid pulse wave (L-CSFPW): that it arises from the arteries supplying the spinal cord, and that it is due to the pulsations of the brain transmitted through the subarachnoid space of the spine.We investigated L-CSFPW of 11 myelopathic patients with a complete (®ve patients, CBgroup) or an incomplete spinal block (six, ICB-group) on myelography to determine the origin of L-CSFPW. Since arterial pressure amplitude (APA), the energy source of L-CSFPW, is not the same between individuals or between before and after operation, not only L-CSFPW itself but also the transfer function between the arterial pressure wave and the L-CSFPW calculated by the system analysis method was analyzed to eliminate the in¯uence of hemodynamic uctuations. In the system analysis, the arterial pressure wave, L-CSFPW and transfer function were decomposed into ®ve harmonic waves (HW).In the CB group, L-CSFPW was observed to be 0.72 mmHg on average (range, 0.25 ± 1.00) in spite of blocking pulsations of the brain, showing that there was a contribution to L-CSFPW unrelated to the brain, that is, the spinal cord. In the CB group, however, the preoperative transfer function value of HW1 (mean, 0.056; range, 0.012 ± 0.170) was lower than that in the ICB group (mean, 0.137; range, 0.061 ± 0.236) (P50.05), indicating that the brain pulsation also contributed to L-CSFPW.In the ICB group, there was signi®cant reduction of HW1 (P50.01) and HW2 (P50.05) transfer function after posterior decompression surgery in spite of improvement in the subarachnoid space narrowing: preoperative HW1, mean, 0.137, range, 0.061 ± 0.236; postoperative HW1, mean, 0.065, range, 0.021 ± 0.153; preoperative HW2, mean, 0.092, range, 0.011 ± 0.148; postoperative HW2, mean, 0.044, range, 0.030 ± 0.066. It has been reported that the spinal cord blood¯ow is decreased 20% or more by laminectomy, therefore, L-CSFPW measurement may be sensitive enough to detect a 20% or higher decrease in this ow. This suggests that L-CSFPW could possibly be used clinically as a non-invasive method of monitoring the spinal cord blood¯ow. For broad clinical application of CSFPW, however, further studies are needed, especially on the direct relationship between CSFPW and spinal cord blood¯ow itself.