A double-blind crossover trial involving a control diet and a diet eliminating artificial flavors, colors and natural salicylates as recommended by Feingold was conducted on 15 hyperkinetic children. Teachers and parents observed the children for 1 mo. prior to treatment, using standardized rating scales. Both parents and teachers reported fewer hyperkinetic symptoms on the K-P diet as compared to the pretreatment baseline. The teachers noted a highly significant reduction of symptoms on the K-P diet compared to the control diet but the parents did not. The control diet ratings did not differ from the baseline period ratings for either parents or teachers. The K-P diet may reduce hyperkinetic symptoms, but this is put forth with caution in view of several features which need further evaluation, including objective measures of change, manipulation of the independent variable and reducing the independent variable to more specific components.