In the past century we have learned that driving performance is impaired by alcohol even in low dosage, and that many other drugs are also linked to impairment. This paper is a summary of some of the more relevant studies in the past fifty years--an overview of our knowledge and unanswered questions. There is no evidence of a threshold blood alcohol (BAC) below which impairment does not occur, and there is no defined category of drivers who will not be impaired by alcohol. Alcohol increases not only the probability of collision, but also the probability of poor clinical outcome for injuries sustained when impaired by alcohol. This review samples the results of the myriad studies that have been performed during the last half century as experiments have moved from examination of simple sensory, perceptual and motor behaviours to more complex measures of cognitive functioning such as divided attention and mental workload. These more sophisticated studies show that significant impairment occurs at very low BACs (< 0.02 gm/100 ml). However, much remains to be determined regarding the more emotional aspects of behaviour, such as judgment, aggression and risk taking. Considering that the majority of alcohol related accidents occur at night, there is a need for increased examination on the role of fatigue, circadian cycles and sleep loss. The study of the effects of drugs other than alcohol is more complex because of the number of substances of potential interest, the difficulties estimating drug levels and the complexity of the drug/subject interactions. The drugs of current concern are marijuana, the benzodiazepines, other psychoactive medications, the stimulants and the narcotics. No one test or group of tests currently meets the need for detecting and documenting impairment, either in the laboratory or at the roadside.