356 CARNEGIE INSTITUTION OF WASHINGTON. 



In order to test any given substance for the presence or absence of 

 either the fat-soluble or the water-soluble vitamine, or to make quan- 

 titative comparisons of different substances with regard to their value 

 as sources of either of these vitamines, one must be certain the basal 

 ration is as free as possible from whichever one of these food essentials 

 is being tested. Inasmuch as neither of these vitamines has been iso- 

 lated nor chemical tests for its presence devised, the only method of 

 assuring oneself of the suitability of any given diet as a basal ration is to 

 feed a number of rations containing various ingredients and compare 

 the behavior of the animals on the different diets. The starch, salts, 

 and fats used in the foods have been proved to contain little if any of 

 the water-soluble vitamine, hence the protein is the only constituent 

 which might be contaminated. To test this point, and to determine the 

 most satisfactory method of preparing protein for use in basal rations 

 for studies on the distribution of the water-soluble vitamine, a series 

 of rats was fed on diets essentially alike, except that the protein was 

 derived from widely differing sources and prepared by widely differing 

 methods. When the protein was casein (either very crude or very 

 carefully purified), washed meat residue, lactalbumin, gliadin, ovo- 

 vitellin, or very carefully purified edestin, the animals began to decline 

 within two or three weeks and with very few exceptions were either 

 dead or moribund in 40 to 60 days. On the other hand, when the pro- 

 tein was either cottonseed globulin or less carefully purified edestin, 

 the animals were maintained without appreciable loss of weight for 100 

 days, and when the amount of protein in the food was doubled the 

 animals gained slightly in weight, demonstrating that the water-soluble 

 vitamine had not been entirely removed from the preparations of these 

 two proteins. The uniform results obtained with the other isolated 

 proteins indicates that they are nearly, if not wholly, free from the 

 water-soluble vitamine and also that any one of them is suitable to use 

 in a basal ration for experiments concerning the water-soluble vitamine. 



A similar series of experiments has been started to test a number of 

 different isolated proteins for the presence of adherent fat-soluble 

 vitamine. These experiments have not progressed sufficiently to justify 

 final conclusions, but the fact that almost without exception the rats 

 on such diets have grown at least some, would indicate that either these 

 diets contain traces of fat-soluble vitamine or else animals can grow 

 for a short time without being furnished more than very slight traces 

 of this vitamine in the diet. It is probable that the lard used did not 

 contribute an appreciable amount of this food substance to the ration^ 

 because "lard oil," prepared by extracting lard with absolute alcohol 

 at about 70°, according to the method used in making our very effective 

 "butter oil," did not prevent the decline of rats fed on a diet containing 

 no known source of fat-soluble vitamine. 



