Recent experience with three children who had non-ketotic hyperglycinemia suggested that interpretation of blood glycine levels in children is complex. The elevations of blood glycine in these children were quite modest, and comparable to those of other children admitted to hospital with other diseases. Often we find elevations of blood glycine in children who have had some degree of starvation before blood was taken for amino acid analysis. In a "typical" patient, valine, leucine, isoleucine and occasionally threonine are depressed, while the glycine level is raised. As nutrition improves all of the amino acids return to normal. Our data suggest that blood glycine levels in children with an acute episode of a debilitating disease need to be interpreted with respect to the immediate state of nutrition of the children.