Spinal anaesthesia has the definitive advantage that profound nerve block can be produced in a large part of the body by the relatively simple injection of a small amount of local anaesthetic. However, the greatest challenge of the technique is to control the spread of that local anaesthetic through the cerebrospinal fluid (CSF), to provide block that is adequate (in both extent and degree) for the proposed surgery but without producing unnecessarily extensive spread and so increasing the risk of complications. The great interpatient variability in spread was observed and described as ‘lauenhaft’(waywardness) by August Bier, 10 the first person to use the technique clinically, and has challenged many subsequent workers. In fact, the definitive studies were performed nearly 100 years ago by Arthur Barker, a London surgeon who was the first to use solutions made hyperbaric by the addition of glucose, 7 but his principles have had to be re-learned virtually each time a new drug has been introduced for the technique. This review focuses in particular on work published in the last decade (during which time ropivacaine and levobupivacine were introduced) in trying to provide guidance on making spinal anaesthesia as predictable as possible.