Hopkins * found that the addition of very small amounts of milk to diets otherwise composed of purified foodstuffs sufficed to a fat-soluble substance carried by butter-fat and the fat of egg yolk and in much smaller quantities if at all by most vegetable and meat fats. This fat-soluble substance (or something showing the same growth-promoting property) has also been found by McCollum to occur in certain plant tissues not rich in fat, notably in alfalfa and cabbage leaves and presumably in leaves generally. Normal growth and full development, as shown by ability to produce and nourish healthy young, demands, therefore, in addition to adequate and appropriate supplies of proteins, fats, carbohydrates, and salts, at least two substances or kinds of substances which are distinguished by the solubility of one in water and of the other in fat. These substances, neither of which has yet been chemically identified, are variously designated by different writers. Hopkins used the term "accessory factors." Funk calls them "growth vitamines." McCollum criticizes the use of the term "vitamine" and proposes that until chemically identified the substances be known as "fat soluble A" and "water soluble B." The fats of milk, eggs, and certain organs, and also the leaves of certain plants, are particularly rich in "fat soluble A" whereas many staple foods are very poor in this constituent. "Water soluble B" is more widely distributed, being found in the foods which have anti-neuritic properties, and it probably is the same as the substance whose absence or insufficiency induces polyneuritis.

* As early as 1906, Hopkins had found experimentally and published in brief (The Analyst, Vol. 31, page 395) the fact that an animal cannot live "upon a mixture to induce growth in young rats (Fig. 12), and Osborne and Mendel demonstrated that a similar growth-promoting effect was obtained when they introduced into their rations of isolated foodstuffs a moderate amount of "protein-free milk" - a powder made by removing the fat, the casein, and the albumin from cow's milk and evaporating the clear filtrate to dryness. Since in both these investigations it was found that milk ash does not show this property, it follows that milk must contain some water-soluble organic substance which exerts a distinctly favorable effect upon growth. A little later it was found both by McCollum and Davis and by Osborne and Mendel that the fat of milk (butter fat) also exerts a growth-promoting influence, which, as it is shared by only certain other fats, is probably not due to the glycerides themselves, but rather of pure protein, fat, and carbohydrate, and even when the necessary inorganic material is carefully supplied the animal still cannot flourish." Seeking further light upon the chemical nature of the essential substance contained in milk and some other natural foods but not in the purified foodstuffs, he deferred publication of the details of the experiments until 1912 {Journal of Physiology, Vol. 44, page 425).

Fig. 12. - Growth curves of rats. Lower curve six rats on artificial diet alone. Upper curve six similar rats receiving in addition 2 cc. of milk each per day. Abscissae time in days: ordinates average weight in grams. Courtesy of Dr. F. Gowland Hopkins.

Thus the feeding experiments with isolated foodstuffs have resulted in establishing the fact (until recently unsuspected and doubtless responsible for many of the failures met in earlier experiments) that there are required for normal nutrition, and most conspicuously during growth and development, these two factors A and B in addition to the previously known factors of ample energy and adequate and appropriate supplies of protein and of inorganic foodstuffs. This has made it possible to proceed much more intelligently and effectively in the study of the relations of ordinary food materials to growth and development. In this connection it is important, as McCollum has emphasized, that growth and development be considered not only in terms of gain in weight at a normal rate, but also in reference to the capacity to produce and nourish healthy young at intervals normal for the species. A diet lacking in growth-promoting properties is apt to have an unfavorable effect upon reproduction and lactation. In some cases a deficiency may become manifest in connection with reproduction, even when it has not appreciably retarded growth.

In a recent summary,* McCollum points out that the deficiency of wheat as a sole food has been found to be associated with the nature of its proteins, of its ash constituents, its lack of the "fat soluble A," and possibly a toxic factor. He states that when wheat and a good salt mixture are fed there is improvement in the condition of the experimental animals for a limited time. A rat will grow for a month † on this combination and then stop, whereas he could not grow at all on wheat alone. On feeding wheat and casein only there is also a marked improvement for a time, and the same is true for a mixture of wheat and butter fat, "but in no case does the beneficial effect extend beyond the first month. These results we interpret to mean that there were two at least of the dietary factors involved, unless the trouble was all the result of toxicity in the wheat kernel. The next step was to feed wheat together with two purified additions as wheat, salts, and casein; wheat, salts, and butter fat . . . combinations (which) will make a young rat grow to practically the normal adult size and at. nearly the normal rate, but rats so fed will never produce young, and will never live much beyond a third of the usual length of life of a well-nourished animal. When we feed wheat with all three of the purified additions, salts, protein, and butter fat, the animals are perfectly nourished and not only grow up at the regular rate but they are able to reproduce at frequent intervals and to successfully rear their young, and these young can complete the life cycle with no other food than that on which their parents lived."

* McCollum. The Present Situation in Nutrition, Hoards Dairyman, Jujy-August, 1916.

† In a month a rat makes as large a fraction of his total growth as is made by a child in from one to two years.

Thus it now appears that the diet in order to be fully and permanently satisfactory must furnish (1) adequate energy value, (2) proteins sufficient in quantity and suitable in their amino acid make-up, (3) ash constituents each in sufficient quantity and all in well-balanced proportions, (4) "fat soluble A," and (5) "water soluble B." All of these factors are doubtless necessary in order to make the diet really adequate at any time, but it is through studies of growth that the last-mentioned factors were found, and all of the requirements are plainly more prominent in connection with growth, development, and reproduction than in the simple maintenance of healthy adults.

Recognition of some of the factors just mentioned is too recent to have influenced the arrangement of many of the feeding experiments which have been made for the purpose of studying the relation of diet to growth, so that it is not always possible to interpret the experimental data in terms of these five categories. This can, however, be done to some extent.