It became very early the principal activity of investigators in nutrition to discover and perfect methods for the analysis of foodstuffs, in order to determine the amounts of protein, carbohydrate, fat and ash contained in them. There were observed pronounced differences in the composition of the many substances which serve as food for man and animal. Meats, milk, eggs and a few seeds, such as the pea and bean, are very rich in protein. The cereal grains contain much less of this foodstuff, and the fruits, tubers and vegetables, especially in the fresh condition, contain very little. Equally great variations are observable in the water content of foods, and in their yields of fats and carbohydrates. One of the great epochs in the development of the science of nutrition is that in which Atwater and his associates in America, and Voit in Germany, examined and tabulated in classified form the chemical composition of an extensive list of human foods (1). Following this, Agricultural Experiment Stations accumulated similar data concerning substances used for feeding animals. Previous to 1900 the idea that there was any marked variation in the quality of proteins from different sources did not become generally appreciated.

It seems remarkable, in the light of the revelations in the field of nutrition during the last few years, that close students of nutrition accepted for so long, and without experimental proof, the belief that the results of a chemical analysis revealed the dietary values of foodstuffs.