It is a self-evident fact that conclusions concerning any one factor in a diet will be reliable only if the ration fed is entirely satisfactory in every respect save the particular factor under investigation. Much of the early work on vitamins was unsatisfactory because of the faulty character of the basal diet, the fault being due to lack of adequate knowledge rather than to any carelessness on the part of the investigators. In recent years great advances have been made in our knowledge of the quantitative composition of the natural foodstuffs and of the requirements of young and adult animals, especially with regard to amino acids and mineral salts.

McCollum is responsible for the introduction of what he has designated "biological analysis" of foods, that is, a systematic series of feeding experiments in which first one dietary factor at a time, and then if necessary combinations of two or more factors, are altered in order to ascertain the true function of each. This method is very well illustrated by the experiments of McCollum and Davis on the dietary value of the wheat grain.1

It was reasoned that, since all the dietary essentials, except possibly the one which is not present in vegetable fats, are certainly present in the wheat kernel, the faults in the latter must depend upon a lack of the unknown substance contained in butter-fat, or on the quantity of some one or more of the well-recognised constituents of the diet. It seemed possible to discover by means of a systematic series of feeding experiments in which the quality of the seed should be improved with respect to one dietary factor at a time, which factor was interfering with growth. Accordingly wheat was fed in the following combinations, and with the results noted:

(1) Wheat alone ........................................no growth, short life.

(2) Wheat plus purified protein..........................no growth, short life.

(3) Wheat plus a salt mixture which gave it a mineral content similar to that of milk...............................................very little growth.

(4) Wheat plus a growth promoting fat (butter-fat) ...............no growth.

From these results it seemed apparent that either their working hypothesis regarding the factors which are necessary in an adequate diet, must be wrong, or there must be more than a single dietary factor of poor quality, and jointly responsible for the poor nutrition of the animals. In order to test this theory another series of feeding experiments was carried out in which wheat was fed, supplemented with two purified food additions.

1 McCollum, "Newer Knowledge of Nutrition," 1919, p. 20.

(5) Wheat plus protein, plus the salt mixture..........Good growth for a time.

Few or no young. Short life.

(6) Wheat plus protein, plus a growth-promoting fat (butter-fat)....No growth.

Short life.

(7) Wheat plus the salt mixture, plus the growth-promoting fat (butter-fat)

...................Fair growth for a time. Few or no young. Short life.

Following this method McCollum and his co-workers have made an exhaustive study of a great many typical foodstuffs and have arrived at results of immense practical value, which must be taken into consideration in all feeding experiments in which natural foodstuffs form part of the diet. In so far as these deal with factors other than vitamins they are briefly summarized here for reference.

Experiments with swine 2 showed that there was very little difference in the nutritive value of the proteins contained in wheat, oat and corn kernels, and at a single experiment in which the three grains were fed together in equivalent amounts there appeared to be little advantage through the supplementary action of the proteins from one source on those from another.

Hart and McCollum3 showed that when swine are restricted to corn meal and gluten feed little or no growth can be secured, but with an addition of salts, making the entire ash content of the ration very similar in quality to that of milk, growth approximating normal was secured to at least 275 pounds. On an exclusive diet of wheat kernel, growth was again limited, both with swine and rats. Correcting the mineral content alone induced a certain amount of growth, but the benefit was only temporary. Supplementing the wheat kernel with salts and butter-fat caused the animals to remain vigorous and strong for a very much longer period, and induced improved growth in both species, although not up to normal. When a further addition of casein to the extent of 2.5 per cent of the ration was made a normal curve of growth was secured with both swine and rats. A diet of mixed grains and distilled water did not allow normal growth in swine.

In a study of the influence of protein intake on growth, McCollum and Davis4 found that there is a progressive increase in the rate of growth on rations derived from milk as the plane of protein intake is raised between three and eight per cent of the diet, and that, during six weeks at least, a ration carrying but four per cent of protein from wheat embryo compares favorably with a similar plane of protein intake derived from milk powder, and is somewhat better than six per cent of protein from the entire kernel, on which rats may grow at about half normal rate.

2 McCollum, J. Biol. Chem. 19, 823, 1914.

3 Hart and McCollum, J. Biol. Chem. 19, 873.

4 McCollum and Davit, J. Biol. Chem. 30, 415, 1915.

An investigation on the influence of the mineral content of the ration on growth and reproduction 5 led to the conclusions that provided the other factors in the ration are adequate, young rats can grow normally and remain in apparent good health on rations whose base content varies widely in amount (from 0.2 to 8 per cent); and that high acidity is not necessarily injurious.

McCollum and Davis6 studied the dietary deficiencies of rice and showed that polished rice must be supplemented with protein and salts, as well as with vitamins A and B, in order to make it a complete ration.