ANALYSIS OP BTITTBR-FAT. 47 



nearly so, over the naked flame and constantly stirred with 

 the glass rod, to intimately mix the oil with the alkaline 

 solution. 



The water constantly evaporating was from time to time 

 replaced by distUled water, until a perfectly clear and trans- 

 parent soap solution was obtained, which was usually the 

 case in about two or three hours. Dilute sulphuric or hydro- 

 chloric acid was then added, the fatty acids which separated 

 were allowed to melt, and collected on a weighed filter, 

 washed with a large quantity of boUing water, dried and 

 weighed. 



But very shortly after the publication of the results thus 

 obtained we were made aware that our figures, as might have 

 been expected, were for the most part somewhat too low, 

 notably so our average of 85 '85 per cent. During an ebulli- 

 tion of two or three hours'- duration, in an open evaporating 

 basin a slight loss by spurting cannot be entirely prevented, 

 even when the greatest possible care be taken, the less so, 

 because every bubble of steam, which rises from the aqueous 

 liquid, bursts in the oUy layer which floats at the top, and 

 carries not rarely a minute fraction of the fat with it, which 

 in most cases falls back into the basin, but which in others is 

 lost. As ,we have mentioned, the process of saponification 

 cannot be conveniently carried out in a covered glass flask, 

 as the violence of the boiling in most cases causes loss of tlie 

 liquid, or even shatters the glass to fragments. This loss by 

 spurting was of course a very serious drawback, and the 

 great trouble and labour and the unceasing attention neces- 

 sary during saponification rendered the process a by no means 

 easy or elegant one. But aU these inconveniences and the 

 source of error were at once removed by a modification intro- 

 duced by Dr. G. Turner, Medical Officer of Health and Public 

 Analyst for Portsmouth. This modification consisted in the 

 employment of alcohol as a solvent for the alkali necessary 



