118 PROCEEDINGS OF THE SCIENTIFIC ASSOCIATION, 
exclusion and at the same time insure a correct determina- 
tion of prussic acid, a special mode of operating had to be 
adopted. A description of the process used is therefore 
reproduced from the above-mentioned paper. 
Five hundred grammes of the root were quickly grated 
into 500cc. of water. The mixture was allowed to stand in a 
well-closed vessel for about two hours, and the liquid was 
then squeezed through a linen cloth into a flask. The flask 
being corked, the starch was allowed to deposit, and then 
200cc. of the upper portion of the liquid were decanted and 
distilled. The retort was connected by a caoutchouc joint to 
the condenser, the end of which dipped beneath the surface 
of water made alkaline with soda, and contained in a closed 
receiver. 
The quantity of prussic acid found in the 200cc. 
of liquid represented one-fourth of that actually present 
This estimate is arrived at in the following manner : both 
bitter and sweet cassava were found by a number of de- 
terminations to contain close upon 60 per cent, of water. 
The 500 grammes of root taken, therefore, would furnish, 
approximately, SOOcc. of water, which, augmented by the 
500cc. added, would make an aggregate of about 800cc. 
All the prussic acid yielded by the sample would of course be 
found in solution in this quantity of liquid. The amount 
present in the 200cc. removed would thus represent ~ or 
one -fourth of the total quantity. 
The prussic acid in the alkaline distillate was estimated 
either volumetrically or by precipitating and weighing as 
silver cyanide. The results were occasionally verified by 
combining the methods — first determining the prussic acid by 
volumetry, then adding excess of the silver solution to the 
