McCollum, Sim-monds and Parsons (42) restricted young rats to a diet which contained lactalbumin as the sole source of nitrogen and could not secure any growth, even with 18 per cent of protein in the diet. They drew the obvious conclusion that since the same diet was complete when casein replaced the lactalbumin, the latter must be a poorly constituted protein. This has been confirmed by Emmett and Luros (43). A later experience of Osborne and Mendel (32) convinced them also that lactalbumin was lacking in some constituent which is essential in proteins of the "complete" type. They found a diet consisting of purified food substances, including 18 per cent of lactalbumin, inadequate for the support of growth even when small additions of yeast proteins were present in the diet. Yeast proteins do not appear to supplement lactalbumin, whereas the small amount of nitrogenous substances in "protein-free milk" are very effective for this purpose. In this connection Osborne and Mendel state "For some as yet unknown reason the majority of the rats grew normally (on the diets containing "artificial protein-free milk," butter fat, starch, lard and yeast)* when the protein used was casein, whereas they have usually failed when it was edestin, and almost invariably when lactalbumin, cottonseed globulin, cottonseed proteins, or squash seed globulin was fed. This result surprised us because all of these proteins had earlier led to normal growth when used in rations containing natural "protein-free milk." The failure to grow on the "artificial protein-free milk"-yeast was especially unexpected in the case of lactalbumin for our former experience had demonstrated that even exceptionally small amounts of this protein promoted normal growth." It is certain that any experiments directed toward making a quantitative comparison of the nutritive values of isolated proteins, have in them a disturbing factor wherever "protein-free milk" was employed.