OF ARTS AND SCIENCES. 119 



The ziucic salt seems to be soluble, the baric salt insoluble,* while 

 the silver salt is probably very unstable, as curcumin is decomposed 

 when boiled tor more than a minute with argentic nitrate and alcohol. 



The fact that only one atom of the hydrogen contained in curcumin 

 can be replaced by the potassium of potassic carbonate would point to 

 the existence of one, and only one, carboxyl group in its molecule ; 

 the presence of this group is confirmed farther by the power of de- 

 composing calcic carbonate possessed by curcumin, f The replace- 

 ment of a second atom of hydrogen when curcumin is treated with 

 potassic hydrate in excess indicates the existence of a hydroxyl group, 

 probably a phenol hydroxyl, and it would seem, therefore, that curcu- 

 min is a diatomic monobasic acid. 



JEsters of Curcumin. 



Although the analyses of the potassium salts had agreed with the 

 formula of curcumin derived from the analysis of the original sub- 

 stance, it seemed desirable to confirm this formula still farther by the 

 study of some derivative of curcumin more stable and easily handled 

 than the salts ; we accordingly took up the investigation of the esters, 

 but found that the ethyl ester made by the action of ethyliodide on 

 the dipotassic salt was a disagreeable brownish-black tarry substance, 

 that could not be obtained in a crystalline condition. We therefore 

 abandoned the stud}' of this substance and turned our attention to the 

 monopai-abrombenzyl ester, which we preferred to the benzyl ester, in 

 the first place because of the great tendency of the parabrombenzyl 

 compounds to crystallize, and secondly, since the presence of bromine 

 increased the difference between the percentages of carbon in the two 

 formulas by more than five tenths of one per cent, and also gave a 

 third element whose quantity could be determined. 



Monoparabromhenzyl Ester of Curcumin C^^Hj3(Cj.HgBr)0^. 



To an alcoholic solution of the monopotassic salt of curcumin an 

 excess of parabrombenzylbromide was added, and the mixture allowed 



* As our work on the potassium salts had achieved the end for which we 

 undertook tlie study of tlie salts of ourcuniin, we tliought it not worth wiiiie 

 to purify any of the otlier salts for analysis. 



t Some experiments on the action of phosphorous trichloride upon mono- 

 ethylcurcumin confirmed the presence of a carboxyl group, as far as they went, 

 but the product was too ill-defined to repay a thorough study. 



