Two regimens (A and B) for TPN were designed to meet the requirements of newborn infants for calories, amino acids, fatty acids, electrolytes, trace elements and vitamins. Both "A" and "B" included fat emulsion (Intralipid). "A" contained fructose and glucose, "B" glucose only. "A" provided amino acids (Vamin) in proportions similar to those of whole egg, "B" similar to those of human milk. All nutrients were given simultaneously into peripheral veins by constant infusion. Nineteen patients (11 newborns, 8 infants) were studied for 1-28 days. Twelve infants recovered, 7 died. In none could TPN be regarded as the cause of death. Treatment was complicated by sepsis in 5 infants. During the course of treatment, blood levels of substrates and insulin were measured before, during and 30 min after discontinuation of TPN. Highly raised concentrations of circulating substrates seen in 3 infants seemed to be related to a poor clinical condition rather than to the regimen used. Infants in good condition tolerated TPN well. Low levels of branch-chained amino acids and tendency to ketonemia, when infusion was stopped, suggested that minimal rather than optimal supply of energy and of amino acids in relation to energy was provided with both regimens. Low insulin levels associated with elevated blood levels of substrates suggested that insulin administration to selected cases might be indicated. Fructose (0.30 g/kg X hour-1) given with regimen A increased blood lactate concentrations. Homocystinaemia appeared in 2 cases; disappearance after excess vitamin B6 administration indicated increased B6 requirement.