!l!HE ANALYST. 125 DBTERMINATIONATION 02' FAT AUIDB IN OILEL BY CE. E. 8-. NE~~LY dl vegehble oils aro aubject, more or lem, to fermentatioii, and the fementa- tive action caws fat acids to separate from glycerine with the formation of free midie. Whon the oil is wed for soap-making or wool-cleasing the presence of the €at aaids has little or no deleterious effect ; but when wed for machinery the cam is different, as they sot on the metal bearings in a similax manner to minerd m i , although beas violently,The process used for the estimation of fat acids is that of Burstyn, and is based on the property possessed by strong alcohol of dissolving the fat acid8, while neutral €ate are not perceptibly soluble. The prooess is carried out by shaking up 100 gram me^ of the oil with 100 grazlunes of 90 per cent.alcohol. The alcohol separates from the oil, carrying with it the fa6 aoids. By means of a soparating funnel the alcohol layer can easily be removed and 20 c.c., titrated with nomal alkali. The aaid obtained corresponds to sulphurio acid ; this, multiplied by 6, w i l l give the total gumtity of mid as oleio mid. A dispute having arisen about some oil purchased by a house in Lille, the author was led to examine B~nstp's process. A portion of the dcoholic solution, equal to about 20 o.c., was evaporated, and dried at a temperatme of 100" U. to 1 0 5 O C., until the weight beoame constant. The followkg 02s were tested :- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bur~tp's Pivceas. '37 0614 6*83 9.23 12-10 -85 to *90 9 76 *677 .. .. .. .. .. .. .. .. By Weight. -28 -600 6- 10-16 13- 065 CO '90 -26 '422 The prooess of Burstyn may, therefore, be considered to give satiafactory results, although it is dear that alcohol dissolveB volatile acids, WE& me lost by evaporation, and also colouring mattem, which have no action upon an alkaline solution. Volatile substances tend to give graUimetrc results lower than those by Burtdyn'a process, while colouring and odorous substances give higher redb, as they have no action on standard alW. In titrating, the author haa found that tmeric gives more aat;sfact;org resdh than either litmus or phenolphthdein.-Xitii &imt@p, 3 xi?., 205.