A large number of animal and vegetable fats and oils were tested, and it was found that all were preventive except linseed.30a When milk fat is eliminated from the diet by using separator milk, and linseed oil was the only fat in the basal diet, it was found that the value of the oils is graded, cod-liver oil being the best and linseed oil worst. Animal fats are more antirachitic than vegetable fats, and the latter differ among themselves. The best vegetable fats are peanut and olive, the least effective of those examined include linseed, cottonseed, ba-bassu oil, a hydrogenated fat, and coconut oil. From this it would seem that if A and the antirachitic factor is identical we must assume that development or lack of development of rickets in puppies is a much more delicate test for the presence of A than are growth experiments with rats.

30 Mellanby, Lane. 1919, i. 407; 1920, i. 856.

30a A diet given to kittens to determine the effect of a deficiency of animal fat (Mackay, Biochem. J. 15, 19-27, 1921) consisted of white bread ad lib., orange juice 1 c. c, alternate days; marmite (a variety of yeast) 2 g. given in milk, whole not consumed. In addition, control animals received whole milk ad. lib., and the experimental animals machme-ekimmed milk with 3% olive oil ad. lib. Both the whole and skimmed milk were brought to a boil and allowed to cool before being given to the kittens. The fat value of the skimmed milk was 0.15%. The olive oil had previously been autoclaved, and had been shown to be free of A. The absence of A in the diet of the experimental animals gave rise to emaciation, arrest of growth, abdominal distension with atrophy of the walls of the stomach and intestines in some cases and changes in the oosto-chondral junctions similar to those seen in guinea pigs and rats suffering from deficiency of the same accessory food factor. The appearance presented by these animals was similar to the clinical picture of celiac disease in children.

No evidence of rickets was found in any of the kittens. (See also Tozer, Biochem. J. 15, 28, 1021).

It was found that the addition of orange juice (one-fourth of an orange per day) did not prevent rickets, also that adding 5 g. of calcium phosphate or increasing the calcium intake by doubling the amount of separator milk, had no preventive effect, but that under these conditions the growth and general health of the puppies was better, and in general the better an animal grows on a rachitic diet the more easily is rickets produced.

On diets III and IV it was found that small quantities of meat and meat extracts did not prevent rickets from developing as had previously been observed when used in addition to diet II. Although meat did not prevent rickets, however, it was found to have some inhibitory effect, which was best seen when small quantities of meat were given with quantities or types of fat otherwise ineffective in preventing rickets. This probably accounts for its action on diet II, where a small amount of butter-fat in the milk had its antirachitic effect enhanced by meat. In this connection Mellanby remarks:

Meat evidently has a stimulating action on growth far beyond its A content, and the antirachitic action is also greater than would be expected from the amount of A supposed to be present. It is known that meat, more than any other foodstuff, has a stimulating effect on the total chemical changes taking place in the body. In thus serving as a stimulant, it will increase the effectiveness of any A in the diet, and will tend to prevent storing up and deposition of this substance in subcutaneous and other tissues. Also, any A in the tissues will be more readily metabolised under the stimulating influence of the metabolising meat. It is probable that the antirachitic action of meat may therefore be due more to making the fires burn more brightly and thereby increasing the effectiveness of any A present in the body rather than to the amount of this factor which it possesses in itself.

Substances found to have no preventive action against rickets are: Separator milk, bread, protein of meat, yeast, linseed and babassu oils, and certain hydrogenated fats.

Substances with well marked preventive action: Cod-liver oil, butter, suet.

Substances with definite, but slighter, preventive action: Meat, meat extract, malt extract, lard, peanut, and olive oil.

In his later paper Mellanby emphasises the importance of the mutual relation of all the dietetic elements.

Any condition which leads to laying on of tissue seems to necessitate a greater intake of antirachitic accessory factor to prevent rickets, while conditions which stimulate metabolism and increase the loss of heat relatively to the energy of the stored food, work together with the antirachitic accessory factor, and reduce the amount necessary for normal growth. Excessive carbohydrate in the diet often leads to obesity and lethargy; confinement works in the same direction, and the diet' under these conditions must have more than the normal amount of antirachitic accessory factor to prevent the development of rickets. Proteins and exercise, on the other hand, are stimulants to metabolism, and when the diet has a relatively high protein content and the animals are active less antirachitic accessory factor is necessary.

Hess 31 reports a case of five infants varying in age from five to twelve months who were given a diet which was complete except for a very small content of fat-soluble vitamin. They thrived on this diet for eight to nine months and showed none of the well-established signs of rickets. From this, Hess concludes that either a very small amount of this vitamin suffices to supply the needs of human nutrition or this deficiency has to be maintained for a period of years in order to bring about any harmful result. Hess does not believe that rickets is brought about merely by a deficiency of A, as this disorder developed in infants receiving large quantities of milk containing an ample supply of this vitamin. Moreover, neither cream nor the leafy vegetables, both of which are rich in A, are regarded as comparable to cod-liver oil as antirachitics.31a.

31 Hess, Proc. Am. Soc. Biol. Chem., J. Biol. Chem. 41, xxxii. 1920; J. Am. Med. An. 76, 603, 1021.

31a Hess and Unger (Proc. Soc. Exptl. Biol. Med. 27, 220, 1020) also finds that many diets supposed to be conducive to rickets result in normal nutrition of infants. Condensed milk, only exceptionally induced rickets. A " protein milk " made by precipitating buttermilk with heat regularly led to rickets. It contained 0.44 per cent ash, in which Ca and P stood midway between that of human and cow milk; its Na content was even higher than in oow milk. Its fat-eoluble vitamin content was high, the water-soluble vitamin low.