In this connection the experiments of Schweizer70 are interesting. The action of various fractions of autolysed yeast upon living yeast suspended in pure sucrose solution was measured by the amount of carbon dioxide evolved. The fractions used in the first series of experiments were the alcohol-soluble and alcohol-insoluble portions, and in the second series the cell walls and the cell contents. In both series the separate action of the fractions was small compared with that of autolysed yeast itself. In each case the two fractions when reunited appeared to have the same effect as the original yeast.

Emmett71 calls attention to the variation in methods employed by different observers in the study of vitamins and recommends that standardized methods be adopted; his suggestions are:

1. There should be a strict uniformity in the basal diets. Thus, in the case of the growing rat, we should agree upon a basal diet that is complete for normal growth and reproduction; upon one deficient only in the water-soluble B vitamin, and upon one deficient only in the fatrsoluble vitamin or vitamins. And in the case of the guinea pig, we should have a complete basal diet and one deficient in the water-soluble C. In these diets, it should be understood that, the percentage of protein, carbohydrate, fat and mineral salts, and the kind and quality of these nutrients should be the same throughout, for the particular purpose intended. That is, all laboratories should use exactly the same basal diets. It would then be possible always to employ these rations for the normal and pathological control groups and then to compare with these control findings, the results obtained from any other combination of nutrients that seemed desirable in connection with a particular phase of a vitamin project that one might have under way.

That is, if the control data obtained from these basal diets were always reported along with the experimental results, it would be easier to interpret and compare the conclusions from different workers.

2. In the preparation of these basal diets, definite methods should be outlined, with respect to detailing just how to purify the various constituents used - the proteins, carbohydrates, fats, and mineral salts. Besides, exact methods should be given for preparing the vitamin extracts that are to be incorporated. Then again, it should be stated clearly how and in what order these ingredients are to be combined; for unless this is done the texture of the ration will often vary considerably, and thereby an unintentional error will be introduced which might affect the food consumption, and this in turn might influence the trend of the weight curves, etc.

3. It is equally important that there be some definite statements made with respect to the experimental animals. In the case of rats and guinea pigs, the age, weight, sex, and exact physical conditions should be carefully observed and stated. It is known, for example, that, for the best comparisons, the animals should be of about the same age and weight at the start, and that younger rats are preferable to older ones. In Emmett's work where he used some 2,000 pigeons, he found that it was necessary to consider breed, previous feeding, age, body conditions, weight, season of the year, etc., in order to obtain the beet results in testing for the antineuritic vitamin. Also, in the case of handling and taking care of the animals, it is essential to bear in mind that one should avoid exposing them to sudden changes in temperature and to undue excitement, and to too much handling. Systematic attention should be given to cleaning and disinfecting the cages, bedding, food cups, and water receptacles. Vermin or lice should be scrupulously excluded. Special care should be given an animal when it becomes appreciably affected by the deficiency disease, and so on.

In fact, it is only after considerable experience that one can always be sure the effect of the experimental diet is a bigger factor than some of the other variables that may have been introduced.

70 Schweizer, Bull. Assoc. Chim. Sucr. 38, 304,1021; J. 8. C. I. July 30,1921, 524A.

71 J. Ind. Eng. Chem. 1021, 1104.

4. It is very important to state whether the animals are to be force-fed or not; whether the food in-take is to be accurately recorded or not; whether the rations are to be made frequently or in large enough quantity to last for a considerable time; whether the consistency of the food or ration is appreciably altered when the changes in the diet are made, and, if vitamin extracts are to be used, whether the animals are to be treated during the test period, or fed at this juncture a mixture of the basal diet and the extract.

5. In the feeding, the question arises whether the rations should be tested from the standpoint of a prophylactic or a corrective. That is, should the test be one of preventing the usual decline in weight, etc., that comes about from the absence of the particular vitamin, or should it be one endeavoring to bring about a normal response after the animal has been affected? In many ways, it is often preferable to follow the latter method.

6. It is very important that the tests should be run long enough to be conclusive. Emmett reports instances in the case of the rat, where the pathological animals did not respond quickly when put upon a curative diet. Sometimes they did not manifest any signs of improvement for almost a week, when they began to show the effects of the new diet. Here it is not a case of the vitamin being absent but simply a question of the animal not eating the new ration at first When they do begin to take the food, the effect of the treatment will become very evident. However, the precaution should be taken to make sure that any gains in weight are due simply to a change in the physical nature of the food, rendering it possibly more palatable. Noting the physical condition of the animal and continuing the test long enough will settle this point. Where it is possible, as in using extracts, it is generally much better to keep the basal diet the same throughout and give the animals definite doses of the extracts. In this way, one can make sure that the material has been consumed.

This, of course, cannot be done where one wishes to test a food in its entirety.