1920.] 



781 



(3) When milk samples have to be kept, the preservative 

 to be used must be formalin. It would appear that bichromate 

 of potassium is ill adapted for this purpose, but investigations 

 on this point are not yet complete. 



In discussing the results of their observations from the 

 experiments, the writers state: — 



(1) With fresh milk the agreement between the Hoy berg 

 and the Gerber process is fairly accurate, provided that the 

 fat determination process develops according to Mr. Hoy- 

 berg's desire, although differences up to 15 per cent., and in 

 one case 20 per cent., occurred, and in general the Hoyberg 

 method gives too low a figure. 



(2) The process is easy to carry out and the rules are 

 simple, although in spite of this simplicity some peculiarities 

 occur which are difficult of explanation. 



(3) In arranging the submersion in the water bath as a 

 means for doing the necessary shaking, the inventor has 

 made the shaking and turning portion of the process much 

 less troublesome, although this was not borne out with the 

 apparatus for six samples placed at the disposal of the 

 investigators. 



(4) In some experiments, in which three samples were 

 tested, one sample in two butyrometers showed a dark fat 

 layer or a flocculent serum, or both, w^hilst the other tw^o 

 samples remained perfect. In the course of the process of 

 fat determination, also, more or less serious deviation in the 

 readings was found. The investigators consider it quite 

 possible that anyone making constant use of the Hoyberg 

 method would find means of avoiding such undesirable 

 occurrences; at the same time, it must always be a disad- 

 vantage that the investigator can, unconsciously, make small 

 errors which could lead to a partial failure of the experiment 

 and. to erroneous conclusions. 



(5) Beckoning up the time spent in carrying out Hoy- 

 berg's process in accordance with his rules, the investigators 

 state that 25 minutes are required from the moment of 

 closing the butyrometer until the time of reading off. This 

 period is longer than with the Gerber process. 



(6) The great advantage of Hoyberg' s method over 

 Gerber' s is the omission of the shaking. The full advantage 

 of this is not felt when the milk experiments are being made 

 in a laboratory fitted up for this purpose, but it is a different 

 matter if the experimenter has to go from one place to 

 another to carry out his investigations. 



