28 OECHSLIN. 



quinine constituent, namel)' to 1 to 800. A very small percentage of 

 unoxidized product therefore markedly increases the toxicity. 



However, mercuric oxide, because it is practically insoluble, reacts 

 too irregularly with larger quantities of material ; not only is the oxidation 

 incomplete, but it also causes decomposition. Hydrogen peroxide was 

 finally found to be most satisfactory as an oxidizer in hydrochloric acid 

 solution. This reagent has been used before * in slightly alkaline or 

 acetic acid solution. A preliminary experiment proved that dilute 

 hydrogen peroxide does not act upon quinine. 



A few drops of the reagent was added to a solution of the alkaloid and 

 the titanic acid test for hydrogen peroxide was then applied. Even after 

 two days this very small quantity of hydrogen peroxide had not disap- 

 peared. When the reagent was added to the arsenic-quinine derivative, it 

 was at once used for the oxidation, and the titanic test was positive only 

 after the oxidation had become comjDlete, as was proved by the test with 

 living trypanosomata. 



After the oxidation is complete, sodium hydroxide and carbonate pre- 

 cipitate a white, heavy, rather sticky substance, soluble in an excess of the 

 alkali. A considerable quantity, in better fonn, precipitates on standing. 

 The precipitate, filtered and washed, is dissolved in a very small 

 amount of alcohol and reprecipitated by ether as a white, heavy 

 powder, then filtered and dried with ether. It contains no chlorine: 

 C^oH.gN^O • CO, • CeH, • AsCh — > C2oO,3^T,0 • CO, • C^H^ • AsOjH^. 

 The substance is readily soluble in acids, even acetic acid, and in alkali. 

 It shows fluorescence in acid, but not in alkaline solutions. 



II. THE DIQUININE ESTER OF DIBENZAESINIC ACID. 



Di-pamidodiphenylarsinic acid was prepared by heating aniline arse- 

 nate with twice its weight of aniline to 185° in an oil bath at 200° 

 for 1 to 2 hours. Mono-^amidophenylarsinic acid are fonned in varying 

 quantities. A smaller amount of a substance, which differs from di- 

 pamidodiphenylarsinic acid in being much more soluble in 50 per cent 

 acetic acid and crystallizing much slower, was also formed. It contains 

 arsenic and has not been identified. The first two substances were sepa- 

 rated by means of the differing solubilities of their sodium salts in 

 alcohol," the second and third l^y repeated crystallization from 50 per cent 

 acetic acid solution. 



Twenty grains of the di-paniidophenylavsinie aeid in 400 cubic centimetei'S 

 hydrochloric acid were diazotized in the usual way with 9.4 grams of sodium 

 nitrite, the solution was then added to a cuprous cyanide solution, prepared 

 from 37 grams copper sulphate and 41 grams potassium cyanide. The clear, 

 bro%vn solution was filtered from the cuprous chloride, and, after being rendered 



'Michaelis, Ann. d. Chem. (Liebig) (1902), 320, 299; Ehrlich and Berthenn, 

 loc. cit. 



° Benda, Ber. d. deutschen chem. Ges. (1908), 41, 1672. 



