1. Adult rats were fed on diets free of either lysine, methionine, threonine or protein. The threonine- and protein-deficient animals lost weight at approximately the same rate, about 100 g in 14 weeks, at which time several were moribund. In contrast, lysine-deficient animals lost only about 30 g in 14 weeks and had lost only 46 g after 22 weeks, when they were killed. Methionine-deficient animals showed an intermediate response. Losses in weight of several tissues – kidney, heart and two muscles – were related to, but not necessarily proportional to, the loss of body-weight. Liver weights relative to body-weights were large in lysine- and threonine-deficient animals and smallest in methionine-deficient animals.

2. Adult rats were fed on diets containing zero, a moderate amount (about twice the estimated minimal requirement) or an excess (about four times the estimated requirement) of lysine or threonine in all combinations (3 × 3 design). Analysis of variance of the body-weights, tissue weights and tissue nitrogen contents indicated, in general, a significant effect of each amino acid, as expected, but also, in most instances, a significant interaction. Plasma concentrations of lysine and threonine were affected by the intakes of the respective amino acids, but plasma lysine concentrations were also affected by the threonine intake.

3. Liver histology also suggested significant interactions between the two amino acids. Animals given no lysine but moderate amounts of threonine developed severely fatty livers; next most severely affected were animals receiving excess of both amino acids. Threonine deficiency, in the presence or absence of lysine, produced moderately fatty livers similar to those seen in protein-deficient animals.

4. Since animals have varying ability to conserve body nitrogen when they are fed on diets limiting in different essential amino acids, measurements of biological value (BV) and net protein utilization by conventional methods, over a short period of time, over-estimate nutritive value relative to amino acid score and probably over-estimate the true nutritive value of poor-quality proteins, particularly those limiting in lysine. If so, this is a serious error, since it leads to underestimates of the protein requirements if BV is used. The fact that certain tissues, particularly the liver, do not necessarily lose nitrogen in proportion to total body nitrogen and may show specific pathological effects depending on the limiting amino acid or the proportions of amino acids in the diet also indicates that general measures of nitrogen economy may not be sufficiently discriminating tests of the nutritive value of proteins.