MKTHODS OK ANALYSIS. 



and that long-continued heating product's no variation in the results 

 as compared with those given by the minimum time required to cli- 

 mate the acid. From this fact and the nature of the results obtained 

 there is no reason to believe that the figures given represent other 

 than the hydrocyanic acid existing in a free state. 



The accuracy of the method of analysis was further checked during 

 the season of 1905 at the Miami station, by making the following 

 determinations: 



Hydrocyanic acid gas was liberated from potassium cyanid and 

 absorbed in cold water, the amount of acid in this solution being defi- 

 nitely determined. Four samples of a variety of cassava which had 

 been found to contain less than 0.001 per cent of hydrocyanic acid were 

 selected for the experiment. The first sample was run as a blank, to 

 the second 2 cc of the water solution of the hydrocyanic acid were 

 added, 4 cc to the third, and 8 cc to the fourth. In a fifth retort, 

 containing no cassava, 8 cc of the solution were placed. All of the 

 determinations were made simultaneous!}' and according to the usual 

 method, and the results obtained showed a satisfactory agreement with 

 the theoretical figures, except in the instance when 8 cc of the solu- 

 tion were added to the charge, when the determined result was 0.018 

 and the theoretical figure 0.021. This is probably due to the fact that 

 the hydrocyanic acid solution was added by dropping it into a pocket 

 made in the pulped material and, being rapidly volatilized, was not 

 distributed throughout the charge, as would be the case when present 

 naturally. The results obtained are shown in Table I. All of the 

 retorts used were tested by adding a definite amount of hydrocyanic 

 acid water and distilling it off, the results in all cases agreeing with 

 those obtained by direct tit-ration. This experiment was made to 

 eliminate any error that might be due to the liberated acid combining 

 with the metal of the retort. 



TABLE I. ComjHirlxoii oftln'oreticn/ u<l actual determinations of hydrocyanic acid, 



1905. 

 [10 cc hydrocyanic acid water = 27 cc of tenth-normal silver nitrate.") 



The question was also considered as to the possibility of the hydro- 

 cyanic acid being present in combination with a base, and thus being 

 nonvolatile. To determine this point, the juice of two plants \\ns 

 expressed and an aliquot portion of each sample Avas distilled and the 



