254 PROCEEDINGS OF THE AMERICAN ACADEMY 



derivatives of assculetine, sesculine, and fisetine,* substances which we 

 have selected for calculation on account of the differences in their 

 molecular weights. 



Name and Formulas of Benzoyl Compound. Brombenzoyl Compound. 



Substance. Difiference in <^ of C. Difference in oj of Br. 



..^sculetine, 



CgHgOCOH), \ 1.97 3.60 



C,H,0,(OH); 

 JEsculine, 



Ci5H,„03(OH)e \ 0.90 1.50 



C,,H,,0,(OH), 

 Fisetine, 



C23Hio03(OH)g \ 0.78 1.66 



C,3H„0,(0H), 



The numbers given in the third column of the above table are far 

 enough removed from the maximum analytical error to leave no doubt 

 in regard to the composition of the substance; and we may add, that, 

 as far as our experiments go, the parabrombenzoates show a much 

 greater tendency to crystallization than the corresponding benzoates, 

 so that in most cases the error from incomplete purification can be 

 reduced to a minimum. 



In deciding which acid containing bromine was the best adapted to 

 our purpose, we rejected the fat acids, although, other things being 

 equal, they were to be preferred on account of the greater differences 

 in the percentages of bromine, because we feared the removal of a 

 portion of the bromine in the preparation or purification of the ester, 

 and accordingly selected from the aromatic acids that one which we 

 found could be prepared most easily, that is, the parabrombenzoic 

 acid. 



In the remainder of the paper we give the details of the methods 

 which we found most convenient for the preparation of the parabrom- 

 benzoic acid, and its chloride and anhydride, as well as a few experi- 

 ments on the formation of parabrombenzoyl esters, undertaken to de- 

 termine whether these bodies could be formed as easily as the corre- 

 sponding benzoyl compounds, but we have not considered it necessary 

 to extend these experiments to the less accessible substances, or to 

 multiply them for the commoner bodies, as those which we have tried 

 are sufiicient to prove the general application of the method; showing 

 that the parabrombenzoylchloride is as reactive as benzoylchloride, and 



* Schraid, Ber. d. ch. G., 1886, p. 1351. 



