No. 6. DEPARTMENT OF AGRICULTURE. 651 



use of a special test bottle like that iii Fig. 20 or 21. Less acid is re- 

 quired for whe}' than milk. 



If o«ily traces of fat appear iu the ueck of the bottle, the fat in 

 the milk examined may be nearly 0.1 per cent, and this reading 

 will be more nearly correct than estimates of from .01 to .05 per 

 cent., which often appear in (he agricultural papers. The reason 

 for this is that minute quantities of fat are either dissolyed or not 

 separated by the method. The amount of fat lost in this way is 

 about the same for all milks; it is compensated for when suffi- 

 cient fat is i)resei!t to form a complete layer across the neck of the 

 bottle by readiog to the point where the fat meets the glass instead 

 of at the concave surface. 



87. Testing Condensed Milk. 



The estimation of fat in condensed milk is accomplished in ex- 

 actly the same way as with cream. As a rule, condensed milks 

 are so thick that it is impractical to measure the test sample di- 

 rectly with a pipette. This difficulty may be overcome by care- 

 fully diluting the milk with a known volume of water, making 

 the analysis of this and correcting the result for the quantity of 

 water added. The best method is to weigh the sample into a test 

 bottle, taking about 8 grams, and after adding about 10 c. c. of 

 water, completing the test in the same manner as with milk, the 

 per cent, of fat being obtained by multiplying the reading by 18 

 and dividing the product by the weight, in grams, of the substance 

 taken. The results are satisfactory. 



88. Testing Cheese. 



The examination of cheese is not as satisfactory as that of other 

 dairy products. The chief reason for this is the unequal distribu- 

 tion of moisture and fat in the cheese, making it very difficult to ob- 

 tain representative samples. On account of this, tests made from 

 different parts of the same cheese, especially if it be very rich, often 

 vary as much as two or three per cent, in the amount of fat found. 

 To avoid this as much as possible, samples should be taken in a uni- 

 form manner. 



Where the cheese can be cut, a narrow wedge reaching from the 

 edge to the center of the cheese, will more nearly represent the 

 average composition of the cheese than any other sample. This 

 may be chopped quite fine, with care to avoid evaporation of water, 

 and the portion for analysis taken from the mixed mass. When 

 the sample is taken with a cheese trier, a plug taken perpendicular 

 to the surface, one-third of the distance from the edge to the center 

 of the cheese should more nearly represent the average composition 

 than any other. The plug should either reach entirely through 



