62 ANALYSIS AND ADULTERATIONS OF BUTTER. 



either be a dried and weighed one, or which, afterwards must 

 be freed by means of ether from the small quantity of fat 

 adhering to it. To strain off through a hit of muslin without 

 filtration, as recommended, is quite inadmissible, as in all 

 cases a milky liquid is obtained, which contains in suspension 

 a notable quantity of fatty acids ; as Dr. Muter employs, how- 

 ever, a comparatively large quantity of fat (ten grammes), 

 this loss is somewhat diminished in importance, but stDI it is 

 avoidable, and ought not to be admitted. Secondly, as it is 

 necessary every time to cool down the water, in order to 

 solidify the fatty acids, it is evident that the washing is really 

 done with cold water, and it is by no means impossible 

 that part of those acids which are but slightly and with 

 difficulty dissolved by even hot water, separate again on 

 cooling. In fact it may easily be observed that the Kquid is 

 more strongly acid before it is quite cool than after the 

 solidification of the fatty acids. Hence a very much larger 

 quantity of water ought to be used for washing, than when 

 boiling water is used, and filtered off as such. As for about 

 three grammes of butter 700 c.c, or even more, of boiling water 

 are required to remove all soluble acid, ten grammes of butter- 

 fat ought to be washed with at least 2^ litres of boiling, and a 

 much larger quantity of cold, water. But Dr. Muter takes no 

 more, and even less than one litre of cold water ! Hence it is 

 obvious, that his results are too low from loss, and too high 

 from incomplete washijig. These errors may counterbalance 

 each other, but not necessarily so, and in general there will 

 be a tendency to get results higher than by the other method, 

 because a large proportion of really soluble acids has not 

 been removed, and does not volatilise on heating to 100° C, 

 from so considerable a quantity of fatty acids as is yielded 

 by ten grammes of fat. 



But we know from experience that the results obtainable 

 by either process do as nearly as possible correspond, if equal 



