Samples of cerebrospinal fluid obtained from 130 patients undergoing spinal anaesthesia were examined microscopically. Subarachnoid puncture was performed using either a 25G Whitacre or 25G Quincke spinal needle. Two samples were collected from each patient and the red blood cell count of the second sample collected was taken as a measure of the vascular trauma associated with the procedure. Red blood cells were seen in 50 (38%) of these samples, of which 18 (14%) contained >β100 red blood cells.mm^β3^. Paraesthesia was felt by 11 (8.5%) patients and the occurrence of paraesthesia was associated with significantly raised red blood cell counts (pβ<β0.0001). There was also a correlation between the number of needle passes made at lumbar puncture and the red blood cell count in the sample (pβ<β0.0001). Neither spinal needle type nor antiplatelet drug therapy influenced red blood cell counts (pβ=β0.66 and 0.37, respectively). These findings suggest that routine spinal anaesthesia is often complicated by minor degrees of vascular trauma, especially when paraesthesiae or technical difficulty occur at subarachnoid puncture.