COMPARISON OF 1 MILK AND MIXED DIET. 8 1 



water which contained 1.64 c.e. of concentrated hydrochloric acid. 

 The flocculent precipitate of casein was strained out on cheesecloth 

 and the very nearly clear solution was filtered through a pulp filter. 

 The filtrate, which at the most was very slightly turbid from sus- 

 pended fat, was tested carefully by the alternate addition of dilute 

 alkali and acid to determine whether any more casein could be sepa- 

 rated from it. The addition of alkali caused a slight precipitate 

 which did not increase on adding more alkali or dissolve on the addi- 

 tion of even relatively large amounts of alkali. This was presumably 

 chiefly calcium phosphate. The addition of acid in no case caused 

 any further precipitation. The filtered milk serum was then heated 

 to boiling for a few minutes and filtered. The filtrate, which was in 

 all cases water clear, was then neutralized to litmus with a dilute 

 solution of sodium hydroxide and evaporated to dryness on a steam 

 bath at a temperature of about 70 . The product thus obtained 

 formed a friable, pale yellow mass which was easily reduced to a fine 

 powder by grinding in a mill. Several grams of this powder were 

 tested for protein by dissolving in about 30 c.c. of water containing 

 a little hydrochloric acid and warming gently. The solution was then 

 saturated with ammonium sulphate. The precipitate, which appeared 

 to consist chiefly of calcium sulphate, was separated by centrifugation, 

 dissolved in a little water, and potassium hydrate solution and copper 

 sulphate added. The solution showed no evidence of the biuret 

 reaction until it was saturated with potassium hydroxide and shaken 

 with alcohol. It then separated into two layers, the upper alcoholic 

 layer showing a slight but positive biuret reaction. Millon's reaction 

 tried on portions of 2 or 3 grams of the substance did not give a posi- 

 tive reaction. Nitrogen determinations in several lots of the protein- 

 free milk powder thus made showed them to contain 0.66, 0.59, 0.60, 

 0.72, 0.71, 0.67, 0.75 per cent of nitrogen. Munk* states that if the 

 proteins of milk are precipitated by alcohol, or separated according to 

 Hoppe-Seyler, from one-thirtieth to one-fifteenth of the protein 

 remains dissolved. All the proteins can be precipitated only by 

 tannin in the cold or by copper hydroxide on heating. He further 

 states that cow's milk contains about one-sixteenth of its nitrogen in 

 non-protein form. Since our protein-free milk powder was equal to 

 50 per cent of the total solids of the milk, it should, if Munk's state- 

 ments are correct, contain 0.48 per cent of non-protein nitrogen, thus 

 leaving at the most only 0.28 per cent of protein nitrogen, equal to 

 1.69 per cent of protein. Since 100 grams of the food mixture 

 employed in our experiments contained 28.2 grams of protein-free 

 milk powder, we can assume that at the most the food pastes thus 

 made contained only 0.48 per cent of milk protein. The protein-free 



*Munk: Virchow's Archiv fur pathologische Anatomic, 1S93, 134, p. 501. 



