On the other hand the does showed much greater irregularity. Eight animals, or 32 per cent, grew at about the same rate as the bucks; completing their weight doubling between the 13th and the 16th days; 6, or 24 per cent, completed it round about the 20th day; another five required about 30 days; and the remaining four grew still more slowly; one animal, though showing no signs of ill health, had not doubled its weight by the 56th day.

A longer comparison further brought out the relative slowness in the growth of the females. Thus on the 52nd day of the experiment the mean weight of the bucks was their mean initial weight multiplied by 3.58. In the case of the does the corresponding factor was 2.63.

Even where the most careful procedure is followed it is necessary to guard carefully against erroneous conclusions drawn from misleading evidence. Speaking from wide experience Davis 34 considers many confusing factors to exist in spite of the progress that has been made. Clear-cut results are the exception rather than the rule. An animal may be abnormal and recover without change in the ration, or it may be subnormal and not recover when taken off the experiment. Frequently the individual variations within a group are greater than the variations between groups. Some of these individual abnormalities are due to simple causes. It sometimes occurs that a rat will continually lose weight on a diet supposedly liberal in all respects. Such animals are not infrequently found to be suffering from pulmonary infections, hair balls in the stomach, tumor, etc., the discovery of which explains the unexpected results." 35 If however the complicating factor should happen to remain unknown very inaccurate conclusions as to the diet might be drawn.

Much depends on the condition of the animal at the beginning of the test. It is fairly generally accepted that the animal, especially at maturity, has some power of storing up a sufficient amount of the essential factors to tide over a certain period of deprivation, the length of this period depending on the natural vitality of the animal. Osborne and Mendel call attention to the possible importance of this factor from the records of the long periods of starvation - over 117 days in the case of the dog - which have been survived successfully.

In another connection, Wheeler36 reports that animals have a surprising power of getting along for a time on a qualitatively inadequate diet. Mice kept upon a ration in which two-thirds of the protein was gelatin and the other third casein, did not show a decline in body weight for 21 days, although the ultimate decline of every animal fed upon this food showed it to be insufficient; another group made material gains in weight for 17 days on a diet upon which all ultimately lost weight rapidly and died unless the food was changed. Similar observations have been made by other investigators.

34 Davis, J. Home Econ. 12, 207, 1920.

35 Osborne and Mendel, J. Biol. Chem. 22, 247, 1916.

36 Wheeler, Exper. Zool. 15, 210, 1913.

Moreover, the minimum requirement of vitamins may vary from time to time. Drummond37 believes that the requirements of the growing rat for A become less as the animal approaches maturity and that the amount of this factor necessary to induce normal growth in a rat which has ceased growing on a vitamin-poor diet is inversely proportional to the weight of the animal.

Besides the individual variability already referred to, Osborne and Mendel38 report a peculiar seasonal variation which affects their rats. During the late summer and fall over a period of three years, young rate on protein-free milk diets almost invariably fail to grow, while at other seasons on the same diet they make normal growth for about three months. "The failure to make early growth is accompanied by diarrhoea, diminished appetite, and, after three or four weeks, by inflamed eyes which soon develop into a purulent state".

Dietary factors other than the particular one under observation may be ill adjusted. The necessity for recognition and rigorous adjustment of all factors in the diet is strongly emphasized by McCollum 39 and by Osborne and Mendel.40 Only when we are assured that every factor but one is at its optimum can we successfully trace nutritional success or failure to its source, still less obtain a quantitative measure of the influence of the factor in question. The importance of the protein and mineral constituents apart from the vitamin content is referred to in a later chapter,41 but these must also be considered in relation to vitamins.

According to Funk42 less vitamin is required on a high protein diet and more when the carbohydrate content of the ration is large.

37 Drummond, Bioch. J. 13, 05; 14, 601, 1920.

38 Osborne and Mendel, J. Biol. Chem. 17, 401, 1914.

39 McCollum, Simmonds and Pitt, J. Biol. Chem. 26, 111; McCollum and Simmonds, Ibid. 32, 181; McCollum, Simmonds and Parsons, Ibid. 37, 159.

40 Osborne and Mendel, Ibid. 37, 159. See also Steenbock and Gross, Ibid, 40, 507,1919.

41 See Chap. III.

42Ftmk, Z. Physiol. 1914, 89, 378; Rank, Lyle and McCaskey, J. Biol. Chem. 27,173,1916; Funk and Dubin, Soi. 52,447,1920; Braddon and Cooper, Brit. Med. J. 1914, 1909; J. Hyg. 14, 351, 1914; Theilcr, Green and Viljoen, 3rd and 4th Rep. of Director of Vit. Res. Union of So. Africa, 1915, 9-68.

If this is so the percentage composition of the basal ration might have a pronounced effect upon the vitamin minimum.