344 BUTTER. 



boiling water is liable to greater experimental error than deter- 

 minations at lower temperatures, and, as the experiments of 

 Skalweit tave shown, that the effect of experimental error is 

 magnified at 100° C, owing to there being a smaller difference 

 between the densities of butter and margarine at this tempera- 

 ture than at lower ones. It is desirable not to adopt this method 

 where accuracy is, as it always should be, a desideratum. 



On the whole, it seems desirable to adopt 100° F. (37 "8° C.) as 

 the standard temperature at which determinations should be 

 made, because it is sufficiently near Skalweit's minimum to give 

 a large difference between butter and margarine, and because a 

 large number of experiments on genuine butters have already 

 been made at this temperature. 



Determination. — The density of butter is best determined by 



the pycnometer. This is filled with distilled water, and the 



weight of the water which it holds at 37 '8° determined. After 



drying, by placing in the water oven and drawing a current of 



air through it, it is filled with the fat and placed in water at 



37 '8° C. till the volume is constant ; the temperature must be 



accurate to 0'1° C. if the result is required to be exact to the 



fourth place of decimals. The weight of fat divided by the 



S7 '8° 

 weight of water will give the density at „ „„ . 



o7 o 



The Westphal balance may be employed, the apparent density 



of water at 37 '8° must be determined, and the density of fat 



indicated by the instrument divided by this to obtain the density 



37-8° 

 at 3^ .go. 



The density is also sometimes determined by a hydrometer. 

 If this instrument be used, it should be tested in fats of known 

 density, and its indications thus controlled. A. Meyer states 

 that the height of the meniscus depends somewhat on the baro- 

 metric pressure, but the error due to this cause is not likely to 

 exceed the experimental error of reading. Should the tempera- 

 ture not be exactly 37 '8° C, a correction of 0'0007 for each 

 degree may be added for temperatures above and subtracted for 

 temperatures below, 37 '8° C. 



100° 

 If it be desired to take apparent densities at , ,, in glass, 



the instruments should be standardised at 15 '5°, and the density 

 determined as above. 



The author has used a bulb of specific gravity 0'865 at 15 '5° 

 for the purpose of determining rapidly an approximate density. 

 A test tube is filled with the fat, the bulb dropped in, and the 

 tube placed in boiling water. If the bulb floats at the top. the 



