Chemical Analysis of Milk and Its Products. 237 



method: First, care must" be taken not to heat the 

 beaker too fast so that spattering occurs ; there is not so 

 much danger from this source when an alcohol lamp is 

 used as with a gas burner, which easily raises the tem- 

 perature too high, causing a fine spray of material to 

 be thrown about, and thus giving too high results for 

 water content. Second, it is important to discontinue 

 the heating at the exact point when all the water has 

 been driven off and before burning of the non-fatty 

 solids (casein, milk sugar, and organic acids) occurs, 

 as indicated by a slight darkening in color. It is not 

 necessary to cool the beakers in water, but they can be 

 left to cool in the air. The determination of water in 

 butter by this method can be finished in ten minutes 

 or less by an experienced operator. 



The Irish test is similar to the method described in 

 the preceding, differing from the same mainly in the 

 shape of the aluminum dishes used. Modifications of 

 this test have also been worked out by the Iowa and 

 Cornell experiment stations, which are designed to pre- 

 vent losses by spattering when the dish is heated. In 

 the Ames method the aluminum dish containing the 

 sample is heated with a paraffme bath, while in the Cor- 

 nell test a thin sheet of asbestos is placed between the 

 flame and the dish holding the samples. 



276. Dean's method^ Three cc. of a melted sample 

 of butter are placed in an ordinary " patty-pan " tin 

 dish (about 2y 2 inches in diameter and y 2 inch deep) 

 ind accurately weighed; the dish is then placed in a 

 steam oven provided with a pop safety valve, a steam 



