No. 6. DEPARTMENT OF AGRICULTURE. 649 



the fat meets the side of the tube and not from the surface of fat 

 in the center of the tube nor from the bottom of the dark line caused 

 by the refraction of the curved surface. 



The reading- may be made \Yith less liability of error by meas- 

 uring the length of the column of fat with a pair of dividers one 

 point of which is placed at the bottom and the other at the upper 

 limit of the fat. The dividers are then removed and one point being 

 placed at the o mark of the scale on the bottle used, the other 

 will be at the per cent, of fat in the milk examined. 



Sometimes bubbles of air collect at the upper surface of the 

 column of fat and prevent a close reading; in such cases a few 

 droijs of strong alcohol (over 00 per cent.) put into the tube on top 

 of the column of fat, will cause the bubbles to disappear and give 

 a sharp line between the fat and alcohol for the reading. When- 

 ever alcohol is used for this purpose ,the reading should be taken 

 directly after the alcohol is added, as after it has stood for a time 

 the alcohol partially unites with the fat and increases its volume. 



Whenever the fat is not quite clear, more satisfactory results 

 may be obtained by allowing the bottles to stand until the fat has 

 crystallized, and then warm them by placing the bottles in hot 

 water, before taking the reading. 



If the column of fat is less than about one division, as will often 

 happen with skim-milk and buttermilk, it may assume a globular 

 form instead of a uniform layer across the tube; v/hen this occurs 

 the fat can usually be estimated with sufficient accuracy by simple 

 inspection, but in such cases it is better to use specially constructed 

 bottles, like those illustrated in Figs. 20 and 21. 



85. Testing Cream by Babcock Test. 



Accurate results can be obtained by the Babcock test in ascer- 

 taining the amount of fat in cream, but much greater care has to 

 be taken in sampling cream. Cream that is sour, or that has been 

 exposed to air until the surface has dried, cannot be accurately 

 sampled. The same is true of centrifugal cream that is badly 

 frothed. Sweet cream, from Cooley cans, that is not too thick to 

 flow readily from the pipette may be tested with satisfactory results. 

 The process, however, must be modified slightly from that used 

 with milk, as the amount of fat in cream is so large that it can- 

 not be measured in the ordinary test bottle, if the usual quantity 

 is taken for the test, besides a much greater error results from the 

 creaLn which adheres to the pipette than with milk. Both of these 

 difficulties may be overcome by takiag two or three test bottles 

 and dividing the test sample between them into as nearly equal por- 

 tions as can be judged by the eye. The pipette is tben filled with 

 39 



