ESTI.MATIOX OF FAT. 121 



The results agree well with the Adams method, except in the 

 case of very highly skimmed milks, when there is a tendency to 

 be low. 



This method has the advantage that the fat can be determined 

 on the same portion of milk used for the estimation of proteins 

 and milk-sugar. 



The following comparative figures will show the results which 

 may be expected : — 



Harrison and Goodson, working in the author's laboratory, 

 have shown that this method cannot be used for the estimation 

 of fat in sterilised or condensed milks, nor is it available for 

 homogenised milk. 



Other Methods of Gravimetric Fat Determination. — The 

 other methods for the estimation of fat in milk are vor\' numerous ; 

 a few of those which have been proposed may be brieHy noticed. 



\\'anklyn's method consists in extracting the fat from the 

 solids of milk dried -per .sr. The totality of the fat is never 

 obtained, as, owing to the hard, horny character of the residue, 

 a considerable proportion of the fat is protected from the ether. 

 This method attained considerable notoriety, owin^ to its semi- 

 official adoption by the Society of Public Analysts more than 

 twenty years ago, but has now fallen into almost complete dis- 

 use. It has been modified by stirring the residue during evapo- 

 ration to obtain a more granular residue, and by evaporating 

 in a conical flask to expose a large surface to the ether, but the 

 results, even with these modifications, have been unsatisfactory. 



Hoppe-Seyler and, later, Liebermann have proposed to shake 

 the milk with potash (to dissolve the casein), and then to extract 

 with ether, but the separation of the ether is so slow as to render 

 this method impracticable. 



Morse, Piggott, and Burton add the milk to anhydrous copper 

 sulphate, which combines with the water, giving a dry residue 

 at once ; they then extract with petroleum ether. 



Baynes proposes to dry the milk on powdered glass, a method 

 essentially the same as that of Storch. Sand has also been 

 largely used in Germany ; but as it is very difficult to grind this 

 up, its use is not to be recommended. 



Marpmann has proposed the use of cotton-wool, Ganntter of 

 wood-fibre, and Duclaux of sponge ; the principle of these 

 methods is the same as that of Adams. 



Fernandez-Krug and Hampe mix a measured volume of milk 

 with a finely-divided mineral substance — usually 5 c.c. of milk 

 with 7 J grammes of washed and dried kaolin — and add 5 gi'ammes 



