146 MILK AND DAIRY PRODUCTS PART 



moistened with a little alcohol and ether, is recommended. A 

 few drops of clear filtered butter fat are brought on to the 

 upper face of the lower movable prism. It is best to do this 

 whilst holding the refractometer in the left hand, and inclining 

 it until the face of the prism in question is horizontal. The 

 butter need not necessarily be filtered, and the requisite 

 quantity can be got by dipping a well-rounded glass rod into 

 the melted fat, taking care that none of the impurities floating 

 on the surface are withdrawn. 



When the prism box has been closed again the refractometer 

 is replaced in its original position, and the refractive index can 

 be read off as with the milk refractometer. Daylight or lamp- 

 light is used for illumination. 



Before taking a reading it must be seen whether the space 

 between the faces of the prisms is evenly filled with butter fat. 

 To do this the image of the faces of the prism which lies about 

 1 cm. before the ocular is examined with a magnifying glass or 

 with the naked eye, the latter being held at a certain distance 

 from the ocular. In this way air bubbles in the layer of fat are 

 easily detected ; they exercise a disturbing effect upon the 

 sharpness of the critical line. 



Before taking the reading of the scale a minute or two should 

 be given for the butter fat and the prism to take the same 

 temperature, for it is only when this has been reached that the 

 critical line becomes clear and definite. 



For every whole degree of temperature above or below that 

 for which the refractometer is graduated a correction of 0'55 

 must be made, and this number added to or deducted from the 

 refractometer reading. 



According to Wollny, the limits for natural butter are 

 49'5-54'0 divisions of the scale at 25 C. The corresponding 

 indices of refraction are 1'4590-1'4620. At 40 C. the limits 

 are from 40'5 to 44*4. 



Along with each refractometer there is supplied a standard 

 liquid for testing the accuracy of the instrument, and this must 

 be done from time to time and the results compared with the 

 figures given on the opposite page. 



A fraction of a degree is easy to calculate (0'1 = 0'06 division 

 of scale). Differences of one- or two-tenths of a division are 

 of no importance, and they are generally the result of faulty 



