254 HOLMES. 



Tlie generally accepted theory of the composition and mechanism of formation 

 of these precipitates was that they were di-iodides of the alkaloidal hydriodide, 

 deriving their "superiodide atoms" from the free iodine of the solution and their 

 hydriodic acid by metathesis from potassium iodide and whatever acid was 

 present. Schweissinger's results with strychnine were in general agreement with 

 the theory. 



In 1895 Kippenberger ^ investigated the availability of Wagner's reagent in 

 the estimation of a number of alkaloids. He obtained precipitates which cor- 

 responded to the tri-iodide formula by the addition of a slight excess of — iodine 



"20 

 solution to morphine in neutral salt solution and to acid solutions of the hydro- 

 chlorides of a number of alkaloids. He assumed that all of the iodine of the 

 periodides was derived from the free iodine of the solution, the hydriodic acid 

 radicles being produced according to the reaction 2I+2HoO=2HI+Hj02. Con- 

 siderable emphasis was laid on the proper concentration of potassiimi iodide in 

 the reagent. When the smallest possible amount was used, the results were 

 irregular, the precipitates containing free iodine in greater or less degree depend- 

 ing on the excess of reagent employed. Kippenberger's final method consisted 

 in adding a solution of silver iodide in potassium iodide to that of the alkaloid 

 which contained the smallest possible excess of acid, employing only an amount 

 of the iodide equivalent to, or slightly in excess of the acid used. The alkaloid 



N 

 was then precipitated with a small excess of an — iodine solution. In this 



manner tri-iodides were obtained with a number of alkaloids, although quinine 

 gave a pentiodide. 



A few years later M. Stolz " declared Kippenberger's method to be worthless, 

 his analytical data to be untrustworthy and his theory of the formation of 

 periodides to be unsoimd. 



A lengthy controversy ensued," in the course of which Kippenberger stated 

 that the discrepancy between his work and that of Stolz was due to the employ- 

 ment by the latter of a solution containing too much potassium iodide. He 

 maintained, in defense of his theory, that the iodine of the hydriodic acid group 

 came from the free iodine of the solution, that the periodide of narcotine may 

 be produced by the action of iodine without the presence of potassium iodide. 

 He admitted the inaccuracies of his method, but still recommended it as practical 

 in case the reagent was standardized against the alkaloids under conditions 

 approximating those of the subsequent determinations. 



In 1896 Gomberg'' brought forward a method for the determination of caffeine. 

 Tarnit" had already shown that caffeine in a solution of mineral acid is pre- 

 cipitated by Wagner's reagent even at very great dilution and Shaw" had 

 obtained similar results with theobromine. Gomberg showed that in the presence 

 of mineral acids the caffeine precipitates, even under widely varying conditions 

 of precipitation, were of a constant composition, corresponding to the pentiodide. 



^ Ztschr. f. anal. Chem. (1895), 34, 317. Ibid. (1896), 35, 10. Ibid., 422. 



'Arch. d. Pharm. (1889), 237, 71. 



"Kippenberger, Ztschr. f. anal. Chepv. (1899), 38, 230. Stolz. Ibid., 278. 

 Kippenberger, Ibid., 280. Kippenberger, Arch. d. Pharm. (1900), 238, 135. 

 Stolz, Ibid., 301. 



'^Joiirn. Am. Chem. 8oc. (1896), 18, 331. 



"Journ. d. Pharm., 28, 433, and 490. 



"Journ. Chem. 8oc. London (1896), 69, 102. 



