Of 220 children referred for suspected 'hyperactivity', 55 were subjected to a 6 week trial of the Feingold diet. Forty (72.7%) demonstrated improved behaviour and 26 (47.3%) remained improved following liberalization of the diet over a period of 3-6 months. The parents of 14 children claimed that a particular cluster of behaviours was associated with the ingestion of foods containing synthetic colourings. A double-blind crossover study, employing a single-subject repeated measures design was conducted, using eight of these children. Subjects were maintained on a diet free from synthetic additives and were challenged daily for 18 weeks with either placebo (during lead-in and washout periods) or 50 mg of either tartrazine or carmoisine, each for 2 separate weeks. Two significant reactors were identified whose behavioural pattern featured extreme irritability, restlessness and sleep disturbance. One of the reactors did not have inattention as a feature. The findings raise the issue of whether the strict criteria for inclusion in studies concerned with 'hyperactivity' based on 'attention deficit disorder' may miss children who indicate behavioural changes associated with the ingestion of food colourings. Moreover, for further studies, the need to construct a behavioural rating instrument specifically validated for dye challenge is suggested.