DAIRY PRODUCTS, 95 



The separation took place perfectly in ten minutes, with a rate of rev- 

 olution of about 300 per minute. 



The second set of four samples was treated in the same way and sep- 

 arated completely in eight minutes. The following readings were ob- 

 tained: No. 1 gave 2.36 per cent, fat; No. 2, 2.34 per cent, fat; No. 3, 

 2.31 per cent, fat; No. 4, 2.30 per cent. fat. 



The third set of samples separated by the centrifugal showed the 

 following percentages: No. 1 gave 2.23 per cent, fat; No. 2, 2.30 per 

 cent. fat. 



The volume of the clear ether-fat solution in each case was about 40cc. 



The next trial was with milk also purchased in the Department res- 

 taurant. I fc proved to be one of the rare cases in which a reasonably 

 prompt separation was secured by the old method. After thirty min- 

 utes about 25ec. of the ether solution had separated, which was enough 

 to get a reading. Duplicate determinations were made : No. 1 gave 

 2.08 per cent, fat; No. 2, 2.04 per cent. fat. 



Four separations of the same milk were also made with the centrif- 

 ugal. Separation took place promptly in eight minutes at a speed of 

 about 200 revolutions per minute, and the volume of ether-fat in each 

 case was about 40cc. : No. 1 gave 2.01 per cent, fat; No. 2,2.01. per 

 cent, fat; No. 3, 2.00 per cent, fat; No. 4, 2.04 per cent, fat; which is an 

 agreement as close as any one could expect. 



Having thus shown that the centrifugal method was capable of making 

 the areometric method applicable to almost every sample of milk, I un- 

 dertook a new series of experiments. In all, 155 samples were subjected 

 to treatment. 



Of the 155 samples examined only 57 gave a good separation by the 

 Soxhlet method in thirty-minutes. Of the remaining 98, about half did 

 not separate at all so as to permit a reading, and the other half only 

 after several hours. Compare this with the centrifugal method, in 

 which only 6 samples out of the whole lot required over fifteen minutes 

 for separation and only one was abandoned as entirely inseparable, and 

 the more general application of the process is at once apparent. 



Of the 6 samples mentioned above, 3 were from the same cow, a 

 grade Shorthorn, four years old, weight about 800 pounds, in milk since 

 July 1, 1885. She gave G quarts of milk a day, was milked at 5 a. m. 

 and 5 p. m. The samples of milk sent were taken at 5 p. m., on April 13 r 

 17, and 22, respectively. The food received by this cow was the same 

 as for all the others (3C) from which samples were taken for analysis. 

 They received at 5 a. m. 3 pounds of wheat bran, and the same of hom- 

 iny chops, and then as much corn (maize) fodder as they could eat. The 

 bran and chops were fed dry. In pleasant weather the cows were out 

 until 3 p. m. They were then fed 10 pounds each of unthrashed oats. 

 At 5 p. m. they got a half peck of chopped turnips and a repetition of 

 the morning's feed of bran and chops. 



