Analysis of outcome data from 305 of the 414 offspring from the Maternal Phenylketonuria Collaborative Study (MPKUCS), plus 70 control offspring, revealed significant deficits in the IQ (intelligence quotient), as measured by the Wechsler Intelligence Scale for Children--Revised (WISC-R), when maternal metabolic control during pregnancy was delayed and/or inadequate. There were, however, 23 'outliers' (7.5% of the 305) in which the offspring's intellectual IQ was worse (n =10) or better (n =13) than expected. The aim of this study was to determine whether collection parameters were incomplete or whether these subjects were true biological variants influenced by other undetected factors or, perhaps, by modifier genes. Among the 10 offspring whose intellectual functioning was worse than expected, additional complications were uncovered that could explain the poor outcome. Four of the 13 offspring with higher than expected IQ had mothers with mild variants of PKU in which the insult to the fetus would not be expected to be as profound. For the other nine offspring whose intellectual performance was better than expected, there was no explanation, based on the parameters studied. We hypothesize that modifier genes will, at times, protect the fetus despite high maternal concentrations of phenylalanine. Not all offspring from the same (untreated) PKU mother may be similarly affected. Finding the source of these modifiers might effect the treatment of MPKU.