246 THE ANALYSIS OF BUTTER FAT. 



so far as the saponification is concerned the method is that of 

 Leffmann and Beam. 



Five grammes of butter fat are weighed into a 300 c.c. Jena 

 glass flask (Fig. 33), and 20 grammes of glycerol are added, the 

 weight of the glycerol being exact to 0-1 gramme ; 2 c.c. of 50 per 

 cent, caustic soda solution are added, and the saponification 

 carried out by heating over a naked flame ; at first a considerable 

 amount of frothing takes place, and on continued heating the 

 solution suddenly becomes clear, after which the flask should 

 be set aside to cool slightly, and 100 c.c. of well-boiled distilled 

 water added. A little ignited pumice, which has been powdered 

 and sifted through muslin (Harris recommends that 0-1 gramme 

 should be used), and 40 c.c. of sulphuric acid (25 c.c. per litre) 

 added, and the flask immediately connected by a bulb tube 

 to a condenser. The sulphuric acid solution should be of such 

 strength that 2 c.c. of caustic soda solution should neutralise 

 about 35 c.c. The apparatus should have exactly the dimensions 

 given in the figure, and the temperature of the cooling water 

 should be such that the distillate enters the flask at about 20 C. 



The flask is heated by a small flame till the fatty acids are 

 just melted, and the flame then turned up to such a height that 

 110 c.c. of distillate are collected in from 19 to 21 minutes, when 

 the flame is immediately removed, and the flask replaced by 

 a cylinder. 



The flask is placed for ten minutes in water at 10 C.,* and after 

 the physical condition of the insoluble fatty acids has been 

 noted, the contents are mixed well, filtered, and 100 c.c. are 

 titrated as in the Reichert-Wollny process. 



The whole of the distillate is passed through the filter, and 

 the condenser is washed out with 18 c.c. of water, this being 

 collected in the cylinder, poured into the flask, and used to 

 wash the filter ; the last 10 c.c. of filtrate should be neutralised 



N 

 by one drop of alkali solution. The funnel is removed to the 



flask in which the distillate was collected, and four successive 

 portions of 10 c.c. of neutral 90 per cent, alcohol (methylated 

 spirit will serve) are poured through the condenser, cylinder, and 



"NT 



filter, and the combined alcoholic filtrates titrated with ^ baryta 



solution ; the number of cubic centimetres used gives the Polenske 

 figure. 



This method has been investigated by many observers, and it 

 is found that the specification of the method is not quite sufiicient 

 to allow of absolutely concordant results being obtained. The 



* The author and Hall have shown that variations of this temperature 

 from 5 to 20 do not affect the results. 



