We have studied in 56 patients the predictability of spread of repeated spinal anaesthesia in the same patient on the basis of a previous block. With plain 0.5% bupivacaine, prediction of the second block was accurate. A significantly higher or lower spread of analgesia than in the previous block was achieved when plain 0.5% bupivacaine was administered using a modified technique--sitting position or lower interspace, respectively. When hyperbaric 0.5% bupivacaine was injected instead of plain solution for the second block using a similar technique, no baricity-related correlation was found between the first and second blocks. Change in technique did not reduce interindividual variation in the spread of analgesia. We conclude that individual anatomical properties may play a more important role than, for example, baricity in subarachnoid distribution of a local anaesthetic solution.