The optimal dietary management of children with phenylketonuria (PKU) has rarely been rigorously explored. The aim of this study was to assess longitudinally the effects of three factors thought to influence plasma phenylalanine concentrations in PKU: total energy intake; protein intake from natural foods allowed freely in addition to allocated phenylalanine exchanges; and the distribution of protein substitute throughout the day. Nineteen subjects, 15 girls and four boys aged 1-16 years, were enrolled. Food intake was weighed, and twice daily plasma phenylalanine concentrations measured during either 3-day or 4-day periods, for a total of 21 days throughout six months. There was a negative correlation between the percentage of protein substitute eaten by the time of the evening meal and the fall in plasma phenylalanine concentration during the day (r = -0.941; p < 0.0001). On average, 49% of pre-evening meal plasma phenylalanine concentrations were less than 100 mumol/l in children who had taken at least 65% of their protein substitute by the time of their evening meal. There was no correlation between excess natural protein intake from freely allowed foods and (a) pre-breakfast or pre-evening meal plasma phenylalanine concentrations or (b) the daily change between pre-breakfast and pre-evening meal concentrations. Nor was there any correlation between excess natural protein intake on the previous day and plasma phenylalanine concentration on the following morning. Energy intake was not correlated with plasma phenylalanine concentrations. It is therefore preferable to distribute the protein substitute evenly through the day in order to achieve stable phenylalanine concentrations, rather than to carry out further fine manipulation of the phenylalanine intake, which would make management of the diet even more difficult.