Hart, Halpin and McCollum10 found a striking difference in the mineral requirements at least and possibly in other nutritive needs of chickens as compared with mammals. In this connection they remark that in investigations involving rats, swine, and cattle, there has been constant agreement among these three species in the effects produced by the several rations studied. There may have been some slight differences between the resistances of swine and rats to certain rations, but on the whole substantial agreement has resulted. It does not follow, however, that this would be true with every ration that may arise for investigation nor that the requirements for the normal factors of nutrition be exactly identical among all species of animals. For example, chickens behave entirely differently from either rats or swine on rations limited to corn or wheat grains and their products. Chickens started at half normal weight can maintain themselves and even make slow growth and produce fertile eggs on rations restricted to gluten feed, calcium carbonate, and either corn or wheat meal, while swine and rats on this same food suffer serious loss of weight and on the wheat ration develop symptoms resembling those of polyneuritis.

Later work showed 11 that younger birds tolerate wheat much less effectively than do more mature fowls, requiring addition of casein and A vitamin, as well as adjustment of the mineral content of the ration, whereas in the case of corn, additions of casein and common salt were sufficient to make it a satisfactory food for baby chicks.

The experimental methods adopted by Osborne and Mendel in their early work12 will serve as typical of most of this class of work.

The experimental animals were white rats. These were kept in metabolism cages modelled after those described by Henriques and Hansen.13 Where it was thought desirable to collect the feces and urine for analysis this was arranged for by resting the cage on a large glass funnel from which the liquid trickled down into a bottle containing chloroform and boric acid, while the solids were retained by a grating in the funnel. In later experiments the funnels were dispensed with, the cages being placed over a frequently changed sheet of absorbent paper upon an enamelled tray or pan. For the most part, however, the analysis of the excreta is unnecessary in work on vitamins.

10 Hart, Halpin and McCollum, J. Biol. Chem. 29, 57.

11 Hart, Halpin and Steenbock, J. Biol. Chem. 31, 415. 12 Osborne and Mendel, Carnegie Publication 156, 1912.

13 Henriques and Hansen, Zeitsdhr. f. phynol. Chem. 43, 418, 1904.

The food was mixed into a paste with melted lard to prevent the animals from scattering it as they ate. Osborne and Mendel say:14 "The inclusion of 20 to 45 per cent of fat in the diet - a condition necessitated by the requirements of the experiment as outlined - seems like an excessive amount, nevertheless the utilisation appears to be satisfactory and attempts to devise less objectionable modes of feeding have been unsuccessful in our hands".

In order to avoid frequent weighings of food, this was introduced into a glass cylinder about 25 cm. in length and 3 cm. in diameter. A rubber stopper was inserted into one end which could be moved forward like a piston head and the food expelled from the other end of the cylinder into the food receptacle. The exit end of the cylinder was kept stopped when the food was not being expelled and the entire apparatus with its food content was kept in an ice-box which made possible its preservation, for long periods without deterioration of the diet. The food eaten can thus be renewed at intervals and the quantity fed determined when desired by ascertaining the loss of weight of each food tube.

Great care was taken to secure optimum conditions of temperature, etc., since rats are sensitive to marked changes. To ascertain whether these conditions were suitable a control series of animals on what was supposed to be an entirely satisfactory diet of ground dog biscuit and lard were kept under observation and their average growth and development compared with that of the group on the experimental diets. Donaldson's study of growth in the rat was also used as a standard of normality. From their observations they concluded that neither caging nor monotonous diet reacted unfavorably upon the health and vigor of the animals.

While they were accustomed as a general procedure to add a small quantity of agar-agar to the food mixtures in order to provide roughage and facilitate evacuation, they succeeded in maintaining rats for over a year without this or any other indigestible constituent in the food.

In McCollum's work the rats were kept in cages two feet square and twenty inches high, made of wire netting nailed to a wooden frame. The entire floor space was covered by a square pan of galvanized iron which fitted snugly. In the pan clean wood shavings covered the bottom. The food and water supply was suspended so as not to become mixed with the shavings. In this way a good opportunity was provided for ventilation and exercise. Several rats were always kept together.15

14 Osborne and Mendel, 1. c. Pt. 2, p. 61.

Mitchell and Nelson criticize the use of wood shavings, which they replaced by paper excelsior. They say:16

"We cannot agree with the statement of McCollum and Davis that they do not look upon the consumption of a small amount of wood fibre as objectionable to any greater degree in this type of experiment than is the feeding of agar-agar. Also, the availability and nutritive value of the nitrogenous and mineral substances of the wood cannot be as lightly disregarded, we believe, as they have been by these investigators".

The conditions and technique made use of by Hopkins may be taken as representative of the methods of the English investigators.17