Abstract Maternal phenylketonuria results in a high incidence of children born mentally retarded. We showed that the large neutral amino acids valine, isoleucine, and leucine (VIL) ameliorate the effects of intrauterine hyperphenylalaninemia in rats on a test of complex maze learning. To further test the ameliorative effects of VIL on intrauterine CNS development during hyperphenylalaninemia, gravid rats were administered a phenylalanine/p-chlorophenylalanine (index group) supplemented diet with or without VIL added. Controls were given standard diet with or without VIL. All groups were pair-fed to the index group. As adults, the progeny exposed in utero to hyperphenylalaninemia showed characteristic learning impairments in a complex water (Cincinnati) maze on forced and elective-choice phase of the task and deficits in radial-arm maze and Morris maze acquisition, whereas those exposed to hyperphenylalaninemia combined with VIL showed no deficits in the forced-choice phase of Cincinnati maze learning and no evidence of radial-arm maze deficits. However, the improvement was not complete, with no ameliorative effects obtained on the elective-choice phase of the Cincinnati maze or on the Morris hidden platform test. No deficits were seen on phases containing test trials for memory function (Olton and Morris mazes). The acquisition differences occurred in the absence of any effects of VIL on maternal weight gain during gestation, maternal serum amino acid concentrations of phenylalanine or tyrosine, or effects on offspring growth. VIL alone produced no adverse or enhancing effects on learning or memory. Based on these data it was concluded that the VIL supplement continues to show promise as a potential treatment for intrauterinely acquired mental deficiency associated with maternal phenylketonuria.