CERTAIN ALKALOIDAL PERIODIDES. 255 



His method has apparently given general satisfaction, although Kippenberger " 

 maintains that a considerable amount of free acid is essential and even then, 

 unless a marked excess of iodine is used, complete precipitation can not be 

 obtained. He states that if the caffeine is first converted into its hydriodide the 

 tri-iodide instead of the pentaiodide is precipitated. 



In 1898 Prescott and Gordin" published a paper on the volumetric estimation 

 of alkaloids as higher periodides. It wa.s already known from the work of Jorgen- 

 sen and others that many alkaloids were capable of forming more than one 

 periodide under different conditions. Previous quantitative methods had con- 

 sisted of adding the iodine solution in slight excess to the solution of the alkaloid. 

 If, according to Prescott and Gordin, the alkaloidal solution is added to that 

 of the iodine, keeping a large excess of iodine always present, the higher periodide 

 as a rule is formed. Applying this general method they state that strychnine, 

 bruoine and aconitine are precipitated as heptiodides and atropine as an ennea- 

 iodide. Morphine, whatever the conditions of precipitation, according to these 

 authors always forms the tetriodide in aqueous solution. 



In the course of their investigations, they showed that in the precipitation 

 of atropine enneaiodide, the potassium iodide of the reagent took an active part 

 in furnishing the iodine for the hydriodic acid group of the periodide moleciile. 

 In 1899 Gordin '" gave a method for the alkalimetric estimation of salt-forming 

 alkaloids, based on the supposition that, although periodides contain variable 

 amounts of additive iodine, they all have one molecule of hydriodic acid for each 

 molecule of alkaloid. The method consisted in dissolving the alkaloid in a known 

 amount of dilute mineral acid, precipitating the periodide with an excess of 

 Wagner's reagent, and titrating the acid remaining in the filtrate with phenol- 

 phthalein as indicator. Several alkaloids gave good results, but it was found 

 neither colchicine nor beriberine, which does not carry down any acid, could be 

 estimated. According to Kippenberger," results by this method are so profoundly 

 influenced by the proportion of free acid as well as that of potassium iodide in 

 the reagent that they are useless for quantitative purposes, and attention was 

 also called to the fact that many of the impurities which in actual work would 

 inevitably be present would combine with the acid and be estimated as alkaloids. 

 Recently Garsed and Collie ^* found that a black, tarry precipitate was thrown 

 down on adding a one per cent neutral solution of cocaine hydrochloride to 

 Wagner's reagent in which the molar ratio of cocaine to the equivalent of iodine 

 varied with the excess of reagent used from 1 : 3.4 to 1 : 5. A considerable amount 

 of iodine could be extracted from these compounds with ether, leaving a crystalline 

 residue corresponding in composition to the tri-iodide. With less concentrated 

 solutions the molar ratio became as great as 1 : 2.3. However, if the reagent 

 were added to the cocaine solution, a brown flocculent precipitate was obtained 

 which had the composition of the tri-iodide. In order to obtain good results it 

 was found essential to add the iodine solution slowly. 



It is plain to judge from this historical review of the more im- 

 portant investigations relating to the employment of periodides in the 



^'Ztschr. f. anal. Ghent. (1900), 39, 435. 



"Journ. Am. Chem. Soc. (1898), 20, 706. Arch. d. Pharm. (1899), 237, 380. 

 ^'Pharm. Arch., 2, 313. Arch. d. Pharm. (1900), 238, 335. Ibid. (1901), 

 239, 645. 



" Ztschr. f. anal. Chem. (1903), 42, 101. 

 ^^Journ. Chem. Soc. London (1907), 79, 675. 



