Following their failure to nourish animals satisfactorily on a diet of isolated and purified foodstuffs, Osborne and Mendel (2) adopted what they apparently regarded as essentially the equivalent of a mixture of purified food-stuffs. The basis of their new diet was 28 per cent of "protein-free milk," a product made by removing as completely as possible the fat, casein and lactal-bumen from milk, and evaporating the resulting whey to dryness. The residue was a yellow solid, easily ground to a powder, consisting for the most part of milk sugar and the mineral salts originally present in the milk. In fact, no other components of the product could be named. "Protein-free milk" was found to contain about 0.7 per cent of nitrogen, but the form in which it was present could not be ascertained.

They prepared diets, using 28 per cent of this "protein-free milk," the remainder of which consisted of starch, lard, agar-agar, and the protein which they desired to study. This combination had certain properties which could not be accurately defined, but which made it far superior to any of the simple mixtures used by others. With these diets they were able with certain proteins to secure in rats normal growth to the full adult size. The results, judging from the widespread comments on these experiments, were regarded by many as the final triumph in this type of experimental work.

Since Osborne and Mendel considered their diet containing "protein-free milk" as suitable for the purpose which they had in view, the comparison of the relative values of various isolated proteins, they apparently abandoned further efforts to solve the problem of what was lacking in their former diets of purified food-stuffs. They did, however, prepare an "artificial protein-free milk" from what they regarded as pure milk sugar and a mixture of mineral salts made up from reagent bottles, -an attempt being made to imitate exactly the composition of the "natural protein-free milk." With this mixture in place of the natural product they secured excellent growth records over periods covering two months or a little longer (3). They made a special effort to secure lactose and salts of a high degree of purity for certain of their experiments in which "artificial protein-free milk" was used. This preparation gave negative results in most cases, and in all the amount of growth was decidedly less than when materials of ordinary purity were used. They carried their investigations in this direction no further than to make the addition of traces of iodin, manganese, fluorin and aluminum. These additions, the "impurities" suspected in their reagents of poorer quality, actually improved the diet, making it capable of inducing better and more continuous growth, which, however, ceased before growth was completed. Osborne and Mendel were for a time inclined to accept the view that the peculiar virtue of their natural "protein-free milk" lay in the nice adjustment of the inorganic elements and radicals it contained.