With the widespread availability of capnography, many anaesthetists have swung away from formally verifying hypocapnia by intraoperative arterial blood gas analysis and, instead, have come to rely upon capnography as an acceptable and constant predictor of arterial CO2 tension (PaCO2) during neurosurgery. However, the nature of the arterial-end-tidal CO2 gradient is complex, and is frequently unexpectedly large, or even negative. The importance of close intraoperative CO2 control during neurosurgery--more specifically, routine hyperventilation, and our reliance upon capnography to guide intraoperative management--is reappraised. There is a growing appreciation of the adverse effects of hyperventilation and hypocarbia, especially upon abnormal or ischaemic brain, and it is clear that capnography alone cannot be used to confidently predict the true PaCO2 during neuroanaesthesia.