When isolated proteins are fed singly, striking differences in nutritive value appear, as has been shown in Chapter III (Proteins). In view of this fact it may seem strange that in the experiments hitherto conducted to determine the protein requirement of man the kind of protein fed has not exerted a more striking influence upon the results obtained. There is, however, no real discrepancy between the two sets of findings. The experiments described in Chapter III were for the purpose of comparing individual proteins isolated even from the other proteins which always accompany them in natural or commercial food materials, and were conducted largely upon rapidly growing young animals, in which there is an active synthesis and retention of protein, so that a deficiency in the supply of any amino acid which is required in the construction of body protein is apt to be quickly and plainly reflected in a diminution or cessation of growth. On the other hand, in experiments like those described in the preceding section, where the purpose is not to compare proteins but to measure the normal protein requirement, the diet is naturally made up, not of isolated proteins or even of single or unusual foods, but (ordinarily at least) of such combinations of staple foods as is believed to represent a normal diet, so that even a relatively simple ration arranged for the purposes of such an experiment would probably contain a number of different proteins among which any peculiarities of amino acid make-up would be apt to offset each other. Moreover the experiments of the latter group have been made entirely upon adults whose protein requirement was limited to that of maintenance. In such cases there is no longer a demand for amino acids to be built into new tissue, but only to maintain the equilibrium which now exists between amino acids and proteins in the tissues already full grown. Any of the amino acids whose radicles are contained in tissue proteins may be expected to contribute something to the maintenance of such an equilibrium, whereas there can be no growth unless all the necessary amino acids are present. In a corresponding series of experiments upon growing children or nursing mothers the influence of food selection would probably be much more pronounced.

Even for the maintenance of adults protein requirement may be found to be considerably influenced by food selection when experiments suitably planned to test the question are carried out. The inadequacy of gelatin as a sole protein food and its inferiority to meat or milk protein when substituted beyond a certain proportion are well known. A series of experiments, designed to demonstrate differences in nutritive efficiency for man of the protein supplied by different staple articles of food, was carried out by Karl Thomas in Rubner's laboratory and the striking results obtained have been widely quoted, often on Rubner's authority. These results, however, have as yet failed of confirmation, and on some important points have been so directly refuted by later workers using longer experimental periods, as to make it appear that Thomas's plan of experimenting and mode of interpretation were not entirely suited to the solution of the question at issue.

Thomas1 thought he had demonstrated that meat protein was greatly superior to bread or potato protein for the maintenance of body tissue; but Hindhede finds no such difference, being able to maintain normal nutrition with either bread or potato nitrogen in relatively small amounts.

1 Thomas, Archiv fur Anatomic und Physiologie, 1909, pages 219-302.

Rose and Cooper * have also demonstrated the high value of potato nitrogen in the maintenance of nitrogen equilibrium, and preliminary experiments in the writer's laboratory†have tended to confirm Hindhede's finding that nitrogen equilibrium may be maintained on a relatively low intake of protein in the form of bread.

Of greater practical importance than the experiments with bread alone are those † which show the maintenance of nitrogen equilibrium over long periods on low protein diets in which bread is the chief source of protein, but is supplemented by small amounts of milk.

Since estimates of protein requirement, in order to be of general application, should provide for the needs of growth, reproduction, and lactation, as well as for maintenance, it will be well to consider more fully the results obtained in feeding experimental animals upon known rations throughout the period of growth or the entire life cycle.

It will be remembered that Osborne and Mendel, feeding isolated proteins in liberal proportion (18 per cent) in diets adequate and well balanced as regards all other factors, found that edestin, a typical vegetable globulin, was able to supply all the protein requirements of maintenance, reproduction, and growth, even through three generations of rats. With gliadin as the sole protein, maintenance was satisfactory but growth was inhibited; but an addition of lysine to this diet caused an immediate resumption of growth. When the supply of lysine was cut off, growth again ceased. A ration containing zein as the sole protein did not suffice even for maintenance; but when tryptophane was added to it, or gliadin, which contains tryptophane, it served to maintain body weight, and on further addition of lysine, growth ensued.

In order to emphasize such differences as these it is sometimes thought advantageous to classify proteins as:

A. Complete: Capable of maintaining adults and providing for normal growth of the young when used as a sole protein food. Casein and lactalbumin of milk; ovalbumin and ovovitellin of egg; glycinin of soy bean; excelsin of Brazil nut; edestin, glutenin, and maize-glutelin of the cereal grains.

* Rose and Cooper, Journal of Biological Chemistry, Vol. 30, pages 201-204. † Not yet published.

B. Partially Incomplete: Capable of maintaining life but not of supporting normal growth. Gliadin of wheat is the well-demonstrated example of this class.

C. Incomplete: Incapable either of maintaining life or of supporting growth, when fed as the sole protein. Zein of corn (maize), and gelatin are the conspicuous examples.

Any such grouping of the proteins, however, must be used with much discrimination, and with great care to insure an understanding of the quantitative aspects of the experimental data, if misconceptions are to be avoided. Edestin is a conspicuous example of a "complete" protein, having served as above noted as the sole protein food of a family of rats for three generations; but when the percentage of edestin in the food mixture was considerably reduced, results like those above described for gliadin were obtained - the diet did not support a normal rate of growth, but this could be secured by adding lysine to the food mixture. Similarly casein when fed in reduced proportion to the total food mixture did not support normal growth; but growth became normal when cystine was added. Thus "complete" proteins may behave as "partially incomplete" when fed in reduced proportion. It is also to be remembered that varying rates of growth in different species (not to mention other differences) make inadmissible any broad generalizations as to the proportion in which any protein should be fed to species other than that with which its "completeness" or "incompleteness" has been demonstrated.

In some of their most recently published experiments (1916) Osborne and Mendel give quantitative measurements of the relative efficiency (for support of growth in young rats) of some of the "complete" proteins. The rate of gain obtained with 8 per cent of lactalbumin required 12 per cent of casein or 15 Q per cent of edestin; or, as they also state the result, "to produce the same gain in body weight 50 per cent more casein than lactalbumin was required, and of edestin nearly 90 per cent more." In maintenance experiments, 2.4 to 3 per cent of lactalbumin was as effective as 3.5 to 4 per cent of casein or edestin.

On extending their experiments from rats to chicks, Osborne and Mendel again found that proteins rich in lysine are much more effective for growth than those in which the proportion of lysine is much smaller.

McCollum found milk protein much more efficient than wheat or maize protein in supporting the growth of young pigs.

As in growth, so in lactation, the demand for material for the construction of new protein creates a condition in which differences of value in the protein fed may readily become more apparent than when only maintenance is involved. Hart and Humphrey find that in meeting the protein requirements of milch cows, milk protein and the protein of flaxseed, "oil meal," are about 50 per cent more efficient than the proteins of the corn (maize) or of the wheat kernel; and Hoobler has shown that milk is the best form of food protein for the production of human milk and the protection of the body protein of the nursing mother.