ANALYSIS OF BUTTER-FAT. 61 



acid aqueous solution, some of the fat being alleged to be 

 lost by creeping through the pores of the filter. But both 

 these conclusions are, as we have shown, fallacious. Ether does 

 not extract more than at most two milligrammes of fat from 

 the glass rod and basin, provided they are properly washed ; 

 the filtrate is clear and bright, and remains so on standing, 

 no oily drops being visible even when examined under the 

 high powers of the microscope, provided the filter-paper is 

 well selected, close and firm, and of the best possible quality. 

 Dr. Muter, moreover, compares his results with figures which 

 we have long acknowledged to be defective, because directly 

 after the introduction of alcohol as a solvent for the fats, it 

 became evident that the loss by spurting during the long- 

 continued boiling amounted to 1 or even 2 per cent. It 

 was only this spurting, which was unavoidable, which caused 

 the low results we at first obtained and published, and which 

 we subsequently, whenever the occasion arose, acknowledged. 

 As soon as this source of error was avoidable and done away 

 with we got results by the filter wash method, which are 

 practically identical with those yielded by the flask-washing. 

 As an instance we niay here again refer to the analysis of 

 the sample of oil-cake butter, quoted at page 59 ; the figures 

 marked (a) were obtained by filter washing, those marked (6) 

 by flask washing, by Dr. Muter himself.' They approach 

 so closely to our own results as to be within the limits of 

 experimental error. 



While, therefore, we readily concede to the flask washing 

 if properly done an accuracy equal to filter washing, we yet 

 must point out some sources of danger which must carefully 

 be guarded against when the former process is employed, 

 and into which Dr. Muter himself has fallen. Krst, the 

 solidified fatty acids cannot simply be separated from the 

 aqueous solution by straining off the latter; the liquid 

 must necessarily be passed through a filter, which must 



