The type of experiment which has thrown most light on the comparative values of the proteins of different foods is that first employed by McCollum and Davis (57) for showing the nature of the dietary deficiencies of the cereal grains. Their method has since been applied to other important seeds and to leaves, tubers and roots, as well as foods of animal origin (58). The details of this method have already been described (p. 27) as an illustration of the method for the biological analysis of a food-stuff. It consisted in feeding to one group of animals a natural food as the sole source of nutriment, and to other groups of animals the same food supplemented with single or multiple additions of purified food-stuffs. The interpretation of the results was based on the number of food factors which had to be improved before growth could take place, and the number which, when improved in quality, could be shown to make the diet of better quality, as revealed by fertility, success in rearing young, span of life, and general well-being.

The results showed that the quality of the proteins of the seeds of wheat, maize, oats, rice, pea, bean, flax, millet, and kaffir corn is such that, when fed at the planes of intake that can be secured by including in the diet as much of the seed as possible, the optimum condition of nutrition is not attained, even though all other factors in the diet are so adjusted as to be fairly satisfactory (59). When the latter conditions are fulfilled, the nutrition of the animals could in all cases be improved by the inclusion in the diet of a pure protein such as casein.