22 



MILK-ANALYSIS. 



When I first worked this process, I employed a pipette 

 •which discharged, not 5 c. c, but 5 grammes of milk, of 

 average density ; and in that way obtained results which, 

 multiplied by 20, expressed percentage. I have, however, 

 come to the conclusion, that it is better to express the result, 

 not exactly in percentage, but in grammes yielded by 100 

 c. c. of milk, and that mode of statement I am now in the 

 habit of adopting. 



As before said, if the milk be curdled, it is not well to use 

 the pipette, and to take the 5 c. c, but to weigh out an 

 irregular quantity of the milk (about 5 grammes), and dry 

 it up. 



The following examples wiU serve to illustrate the degree 

 of accuracy easily attainable by this process. 



A sample of good country milk was submitted to the 

 process four times, with the following results: — 



Expressed in percentage, this is equivalent to- 



