The past year brought no epochal advances in anaesthesia. Developments in this specialty are not as striking as in other areas of research, although it can claim some records for longevity: morphine 1806, nitrous oxide (N2O) 1844, spinal anaesthesia 1898. Among the articles appearing in specialist journals this year, a handful stand out. Glenn Russell has shown that nasogastric tubes are not routinely necessary for cardiac surgery (in patients without risk factors for gastro-oesophageal reflux) and that these tubes may do more harm than good, if managed without continuous suction, by promoting aspiration of gastric contents (detected by tracheal pH monitoring). Will this finding be confirmed in other settings? Will it inspire critical reappraisal of other procedures, ubiquitous in medical practice, of unproven benefit yet blindly continued? Will it mark the start of a movement to abandon routine procedures that cause needless discomfort, increase workloads, and usurp limited resources that could be put to better use? Current anaesthetic practice has as much to gain as other disciplines from evidence-based scrutiny. Pioneering this task, Tramèr, Moore, and McQuay applied the technique of systematic review of randomised clinical trials to unresolved questions about the emetic effect of anaesthetic regimens that include N2O. They concluded that leaving out N2O does not decrease postoperative nausea, unless baseline risk of vomiting is high. N2O-free regimens do, however, entail the risk of intraoperative awareness. Other investigators have shown, disproving ancestral belief, that the prone position (allowing unrestrained abdominal movements) improves rather than worsens, pulmonary function in obese patients under general anaesthesia (Anesth Analg 1996; 83: 578). Meanwhile, further attention has been given to ways of safely minimising perioperative suffering and discomfort. Rigid traditional preoperative fasting is being increasingly relaxed: a survey reveals that clear fluids, for instance, are safely and widely allowed 3 h before elective surgery (Anesth Analg 1996; 83: 123). The neglected issue of postoperative sleep disturbance was addressed by Rosenberg-Adamsen who urged us not to forget such preventable causes of sleep disturbance as noise (on intensive-therapy units), nocturnal nursing procedures, starvation, and heat or cold (as well as, of course, pain). Beyond the subjective wellbeing afforded, good postoperative pain relief has objective benefits. Hirose reports that newborn infants of mothers given satisfactory analgesia for 3 days (continuous epidural anaesthetic infusions) took more breast milk and gained more weight 11 days after their mother's caesarean delivery. What would be welcome next year? Elegant physiological investigations such as those of Selldèn (interesting in our era of technological infatuation because they could have been obtained at any time since anaesthetics were first given, thermometers were used, and nutrient-induced thermogenesis was known). Preoperative infusions of aminoacids (requiring access to a central vein, alas) prevent postoperative hypothermia—a condition neither comfortable to patients nor devoid of untoward effects (Lancet 1996; 347: 1199). Also, now that renewed hints suggest that use of pulmonary artery catheters may do more harm than good (Lancet 1996: 348, 1324), results of trials addressing this issue (Swan-Ganz catheter's swan song?) will be eagerly awaited.