922 PRICE, ON ESTER-SAPONIFICATION. 



The results shewed that the constant depended chiefly on the 

 nature of the acid radicle, Thus the constants of the methyl, 

 ethyl and propyl esters of acetic acid were fouiid to be the same 

 within the limits of experimental error. The constants of the 

 methyl esters of formic, acetic, propionic and butyric acids how- 

 ever differed greatly from each other. 



Since the constants of the various esters of a given acid 

 were found to be practically the same at the arbitrary tempe- 

 rature of 25% it was probable that the temperature coefficient 

 would also be the same in each case. It was not certain how- 

 ever that this would be true of the difFerent esters of the saiiie 

 acid and the following work was undertaken at the suggestion 

 of Prof. Arrheniüs, to whom I raust express my indebtedness, 

 to clear up this point. 



The method of procedure was the following: 



Standard Solutions of the estcr and of hydrochloric acid 

 were made, and warmed up in the thermostat to the required 

 temperature. Definite volumes of each Solution were then mix- 

 ed, and titrati(?ns of the amount of free acid made at various 

 intervals, the end titre being taken when it was quite certain 

 that all the ester had been saponified. 



In order to get reliable results various precautions had to 

 be taken. The raost exact method is that used by Hemptinne, 

 viz. using very dilute Solutions and dividing the reaction mixture 

 into a number of small fläsks and titrating the contents of each 

 flask as a whole. This was found to be impracticable for several 

 reasons. In the first place, with dilute Solutions the saponifi- 

 cation would go on so slowly at 0" and at ordinary temperatures, 

 that it would be very difficult to measure; secondly, if Solutions 

 of a strength great enough to make the velocity of reaction 

 appreciable were used, it would not do to divide the reaction 

 mixture into a number of smaller fläsks, as the loss of ester by 

 evaporation during the process of transference would be consider- 

 able and difFer in every case, thus making the results obtained 

 worthless. The error that the above method seeks to avoid is 



