266 HOLMES. 



These last experiments were also taken to indicate that the precipi- 

 tation of the alkaloid is, imder certain conditions, incomplete, a conclu- 

 sion already reached from other considerations, and it was suggested 

 that curves sho^vn in plates numbered 1 and 2 might to a large extent 

 be merely solubility curves of the periodides. It has been shown that 

 the apparent atomic ratio of iodine to base could be diminished in a 

 number of ways, and as, for instance, by increasing the concentration of 

 potassium iodide in the solution, which would increase the solubility 

 of the precipitate. But it has also been demonstrated (see experiments 

 numbered 68 to 72) that the atomic ratio of iodine to base could also 

 be diminished by such means even when no precipitate was present, and 

 also that the recovery of free iodine from such soluble compounds was 

 impossible. While, therefore, it must be admitted that with low concen- 

 trations of iodine the precipitation is incomplete, it is very questionable 

 whether any great error is introduced by reason of that fact and it 

 seems probable that the curves give a fairly accurate representation of 

 the actual reaction. 



It has been shown that a portion of the additive iodine expended in 

 the reaction can not again be recovered with thiosulphate, the presump- 

 tion being that it is bound to the alkaloid in some intimate manner. It 

 is hardly possible that this bonding should be of the simple nature 

 represented by the theory. A compound such as alkaloid ■ HI • I^. would 

 not be expected to retain its additive iodine with such tenacity, nor 

 would it be colorless. It is apparently necessary to assume that the 

 reaction^ is more complicated than has been supposed. 



In view of these experiments the attempt to employ Wagner's reagent 

 for the quantitative estimation of these alkaloids in acid solution would 

 appear to be impracticable. It is not dithcult to understand why this 

 method has given rise to so much controversy, and the statement em- 

 phasized by Prescott and Gordin ^- that in acidulated aqueous solutions 

 morphine is always precipitated as a tetraiodide is not easy to under- 

 stand. 



The results obtained with neutral salt solutions of these alkaloids are 

 also variable, as is indicated by experiments numbered 69 to 72. A few 

 qualitative tests moreover served to show that under these conditions, 

 just as in acid solutions, the amount of iodine in the precipitates in- 

 creases with the concentration of iodine in the solution. 



That certain alkaloids do form definite crystalline compounds in a 

 quantitative manner with this reagent appears to be true and the ten- 

 dency in a number of cases seems to be to form the tri- and penta-iodides. 

 In this connection it is interesting to note the results of the investigations 

 of Bray and MacKay -* upon aqueous solutions of iodine in potassium 

 iodide. 



"Journ. Amer. Chem. 8oc. (1898), 20, 717. "Ibid. (1910), 22, 914. 



