DETERMINATION OF CONSTITUENTS OF BUTTER. 13 



scales. This method is utterly tmreliable, unless the filter 

 be specially clipped ■with the scissors to the correct weight. 

 Some filters stamped out from the finest Swedish paper 

 differ very considerably, sometimes as much as 0'05 gramme. 

 The residue is now burned in the weighed basia to a white 

 ash, which consists, ia genuine butters, practically of salt 

 only, and is accordingly calculated as such. The difference 

 T)etween the weight of the ash and that of the dry solids 

 not fat, gives the amount of casein. If an accurate deter- 

 mination of the salt in the mineral residue obtained be 

 deemed desirable, the chlorine must be determined by the 

 usual nitrate of silver method, either volumetricaUy or gravi- 

 metricaUy, as circumstances may require. 



The fat may be estimated either directly or by difference. 

 The ethereal solution obtained may be evaporated in a small 

 weighed flask or beaker ; the ether may easily be saved by 

 connecting the flask with a condenser. We can recommend 

 a simple and effectual one, easily made by bending a narrow 

 glass-tube into a Y shape, each arm being a foot long. The 

 bent tube is immersed in cold water, and the ethereal solu- 

 tion heated to boiling by means of a water-bath. When 

 nearly dry, the flask containing the butter fat is placed into 

 the steam-bath, and dried till a constant weight is obtained. 

 The indirect determination is much the quickest, and capable 

 of yielding very accurate results. 



In lieu of ether, petroleum spirit or benzoline, recently 

 distilled, may with advantage and considerable saving be 

 employed. 



According to the plan of analysis described in this chapter, 

 a great number of determinations were made by us, some 

 of which are appended to illustrate the composition of the 

 butters of commerce : — 



