The Glasgow Coma Scale, based upon eye opening, verbal and motor responses has proved a practical and consistent means of monitoring the state of head injured patients. Observations made in the early stages after injury define the depth and duration of coma and, when combined with clinical features such as a patient's age and brain stem function, have been used to predict outcome. Series of cases in comparable depths of coma in Glasgow and the Netherlands showed remarkably similar outcomes at 3 months. Based upon observations made in the first 24 hours of coma after injury, data from 255 previous cases reliably predicted outcome in the majority of 92 new patients. The exceptions were patients with potential to recover who later developed complications: no patient did significantly better than predicted.