Australia's National Health Policy on Alcohol has recommended that beverage containers be labelled so that alcohol content is 'readily understandable by the public'. Health promotion to increase the responsible use of alcohol now relies extensively on the concept of a standard drink--usually defined as 10 g of ethyl alcohol. Numerous difficulties confront a drinker who wishes to apply the standard drink system to monitor alcohol intake. This report describes a series of experimental tests of the proposal that these difficulties are minimised if alcohol containers have their alcohol content indicated in terms of standard drinks in addition to the usual percentage alcohol by volume method. Subjects were drinkers recruited from a Perth shopping mall and were tested only on beverage types they had consumed within the previous week. They were required to pour what they judged to be a single standard drink from a 750 ml bottle of either wine or beer. Beer drinkers achieved greater accuracy in this task when the bottles had standard drink labels, even when the glass size and beverage strength were varied. Wine drinkers had equal difficulty with this task whether standard drink or percentage labels were used. The addition of a 'ladder' up the side of a wine bottle with graduations in standard drinks would be necessary for wine drinkers to achieve a high level of accuracy. We conclude that labelling drink containers with their alcohol content in terms of standard drinks would better equip all drinkers to follow the advice of health educators.