"The necessity of long continued experiments calls for particular emphasis. Physiological alterations dependent upon the gradual depletion of a small store of essential tissue material may manifest themselves with extreme slowness; and the fact that a satisfactory nutritive balance can be maintained for a week or two or even a month in some cases is no guarantee of either the ultimate success of the dietary or of the impossibility of a decline owing to the in-appropriate exhibition of an essential ingredient - a deficiency in some essential ingredient may not make itself manifest for a long time, and even then be difficult to detect; for many other causes than defects in the food may lead to the decline or death of animals during experiments lasting over many months. It is only when the failure to be maintained is the invariable outcome of the prolonged feeding on a given diet that we are justified in assuming that the diet is in some way inadequate, and then only when a prompt and complete recovery ensues when the diet is changed to one that is known to be in all respects sufficient.

... If we had been content to discontinue the experiments after a reasonable period many of the declines evidently associated with imperfections in the dietary and readily checked by a change in feeding would have escaped attention".

A difficulty which is met with at the outset of all experimental work is the great variability in animals of the same species with regard to growth, impulse and vitality. This has been repeatedly emphasised by the various experimenters. Osborne and Mendel 26 noted that when rats were kept on a diet of purified protein, lard, starch, and protein-free milk, growth ceased after sixty days in some cases while others continued to grow for a hundred days or more. They report the case of one rat among hundreds tested on inadequate lard diets, continuing to thrive for the exceptional period of 252 days before the inevitable failure ensued.27

25 Osborne and Mendel, J. Biol. Chem. 13, 233, 1912-13. 26 Osborne and Mendel, J. Biol. Chem. 15, 313, 1013. 27 Osborne and Mendel, J. Biol. Chem. 20, 383, 1915.

If success in growth or reproduction is to turn on the single factor of the lack of a certain complex or complexes in the ration, McCollum and Davis assert that we must know the animals are so vigorous that they are accomplishing everything which any individual of the species is prepared to accomplish, as far as the chemical character of the ration will admit.

McCollum, Simmonds and Parsons29 call attention to the same point, observing that when an animal in infancy is fed a diet unsatisfactory in any respect it does the best it can under the circumstances. There is a variable factor in the vitality with which the individual is born, which in some measure will determine its ability to utilize food faulty in any respect. Only in this way can an explanation be provided for the great variation in the ability of several individuals in the same experimental group to grow on the same diet. Since it is not possible to determine the extent to which any individual possesses natural vigor, an experiment should be conducted with several animals so that one or more will be of great vitality. The elimination of the unfit in the breeding stock is important for this reason.

Funk and Macallum30 similarly note:

Experience demonstrates that there are wide variations depending on the constitution of the individual rats. Every rat taken individually is not suitable for this class of work. As a matter of fact in one case, 80 per cent of the rats purchased from dealers were rejected on account of physical defects not apparent before the initiation of the experiments. A second complication is a diminished resistance to infection which follows the use of all artificial diets. The meager knowledge we possess of the pathological conditions in rats may lead to a condemnation of the diet; whereas actually the condition could be remedied without change of diet, if we were able to recognise its nature. As an example of this, rats on artificial diets frequently contract an eye infection which can be treated with a certain degree of success by an application of a few drops of sine sulphate solution. If untreated this condition is accompanied by loss in weight, becomes acute, and terminates fatally.

28 McCollum and Davis, J. Biol. Chem. 20, 641, 1915.

29 McCollum, Simmonds and Parsons, J. Biol. Chem. 37, 159, 1919.

30 Funk and Macallum, J. Biol. Chem. 27, 51, 1916.

Funk recommends the use of rats specially bred for the purpose in the laboratory. He31 laments that the animals furnished by dealers seemed to lack that power of resistance which would make them suitable for feeding experiments. This was manifested by a loss of hair on being placed in the metabolism cages, and early death before the experiments were well under way. This difficulty was promptly eliminated when rats bred in Funk's laboratory under proper dietary conditions were substituted. The animals which were used for breeding purposes thrive best when kept on a diet of oats, bread, condensed milk, and yeast. The success of this diet was evidenced by the production of better animals from the physical as well as from the experimental standpoint.

In order to determine the degree of variability among young rats on a normal diet, Hopkins32 transferred fifty rats from the regular diet of the laboratory - dog biscuit and oats - to a diet of bread and milk, and noted the time required for them to double their weight. The change of diet was made in order to reproduce more exactly the conditions under which the usual laboratory experiment is carried on, a sudden change of diet at the beginning of the experiment being inevitable in these cases.33

Half the animals used were bucks, and half does. The most striking result of the test was its demonstration of a much greater regularity in the variation of the bucks. The initial weights of the animals fed, while mostly near to the average of 50 grams, showed in certain cases a considerable departure from this (46 to 63 grams). Nevertheless, the periods of weight doubling in the case of bucks departed comparatively little from the mean value. ... In the case of 19 rats out of 25, or 76 per cent, the completion of the weight-doubling period occurred within three consecutive days (thirteenth to the fifteenth). Two other rats may be grouped with these, one completing its period on the 11th, and one on the 16th day. Of the remainder two doubled their weight on the 20th, one on the 21st, and one on the 26th day. Over 80 per cent therefore showed periods near to and normally grouped round the mean.

31 Funk, J. Biol. Chem. 27, 1, 1916.

32 Hopkins, J. Physiol. 44, 430.

33 MacArthur and Luckett, J. Biol. Chem. 20, 164, point out that when* ever any food mixture was given to a mouse accustomed to an ordinary diet of corn, dog biscuit, sunflower seeds, carrots, and meat, there was a short period of slower growth, or even no growth, followed by nearly normal growth if the food mixture was complete.