QUININE ESTERS OF PHENYLARSINIC ACID DERIVATIVES. 29 



alkaline and warming slightly, it was filtered from cuprous oxide. On boiling 

 with a large excess of sodium hydroxide, ammonia is given oflF and the dinitrile 

 saponified. After cooling, the solution is neutralized with hydrochloric acid ; 

 when acidified with acetic acid only part of the dibenzarsinic acid which is 

 formed is precipitated, the remainder stays in solution as the acid salt and 

 is decomposed only on heating with a large excess of acetic or with hydrochloric 

 acid." Precipitation with acetic acid carries down the dark-colored impurities, 

 leaving a slightly yellow solution; when treated with hot, concentrated hydro- 

 chloric acid, the impurities remain behind, and dibenzarsinic acid is formed; 

 this is precipitated on cooling and dilution with water. It is a grayish powder, 

 as is the precipitate from the original solution with hydrochloric acid. For 

 analysis it was reerystallized from hot, concentrated hydrochloric acid. 

 I. 0.2164 gram substance gave 0.0946 gram MgsASoO,. 

 II. 0.2349 gram substance gave 0.1012 gram MgoAs^Oi. 



Calculated for CnHnOeAs. Found per cent. 



per cent. I. II. 



As=21.4 21.1 20.7 



Before treating the substance with phosphorus pentachloride, it was reduced 

 for the reason given when benzarsinic acid was discussed. Two grams of the 

 acid were dissolved in a solution of alkali and 5 grams of potassium iodide were 

 added. Sulphur dioxide, passed through the warm solution and dilute sulphuric 

 acid added to it, are successively employed. Dibenzarsinic acid liberates iodine 

 from potassium iodide. After the reduction was complete, the acid was precipi- 

 tated by the addition of more sulphuric acid. A white precipitate, resembling 

 in everyway the original compound, appeared and was filtered off. The filtrate 

 was then treated with hydrogen sulphide in order to prove whether arsenic had 

 been split off', as in the case of other derivatives of arsenious acid, for instance, 

 pamidophenylarsenious oxide." A small quantity of a white precipitate, most 

 probably the less soluble dibenzthioarsinioiis acid,'- appeared, but not a trace of 

 j-ellow arsenic trisulphide was visible. 



However, the precipitate, after being thoroughly washed, gave a qualitative 

 test for iodine, and even repeated solutions in alkali and precipitation did not 

 remove the halogen. A certain amount of substitution by iodine seems therefore 

 to have taken place, although no arsenic was split off. A second experiment 

 with concentrated hydroiodic acid and amorphous phosphorus gave the same 

 result. 



As a result, instead of first reducing the acid it was treated directly with 

 phosphoruspentachloride. 



Five grams of dry dip-benzarsinic acid were added to 25 grams phospliorus 

 oxychloride. The body dissolved without visible reaction. After heating it 

 slightlj', 6 grams of phosphorus pentachloride were added. The development of 

 hydrochloric acid began at once and was completed by heating the solution to 

 90° to 100°. The greater part of the solution which contained dibenzoylchloride 

 arsinic trichloride (CICO • CeH< • ),AsCl3. etc., was then distilled in vacuo and 

 chloroform added; phosphoric acid precipitated, the acid chloride remained in 

 solution. Eleven grams of quinine (required 9.3 grams) in chloroform were 

 then added, and the mixture boiled on a reflux condenser for one hour, after 

 which most of the chloroform was distilled off. The thick, light brownish residue 

 was thoroughly worked through with ice wat€r in order to decompose ehlorphos- 



"La Coste, An)i. d. Chem. (Liebig) (1881), 208, 22.' 



"Ehrlich and Bertheim, Bet: d. deutschen chem. Ges. (1907), 40, 2292. 



"La Coste, loc. cit. 



