370 PROCEEDINGS OF THE AMERICAN ACADEMY 



Constitution of Benzoltrisidphonic Acid. 



We have determined the coustitution of the acid iu the two follow- 

 ing ways. 



First Method. — Dry potassic benzoltrisulphonate was mixed thor- 

 oughly with exactly the theoretical amount of potassic cyanide (1 grm. 

 to 0.45 grm.), — if an excess of the cyanide is used, the yield is very 

 much diminished, — and heated by means of a gas-furnace in a piece 

 of combustion-tubing closed at one end; the white crystalline subli- 

 mate which collects in the cold part of the tube was removed, and 

 saponified by boiling with alcoholic potassic hydrate, a process which 

 lasts several days. When the evolution of ammonia ceased, the con- 

 tents of the flask were acidified with hydrochloric acid, and shaken 

 with ether ; the ether extract was an acid which melted at 320°-325°, 

 and this melting point remained unaltered after crystallization from 

 water. It was converted into the silver salt, which gave the following 

 result on analysis. 



0.1528 grm. of the salt gave on ignition 0.093 grm. of metallic 

 silver. 



Calculated for CcH3'C02Ag)3. Found. 



Silver 61.02 60.86 



No other acid could be detected in the product of the reaction. The 

 melting points of the three tribasic acids derived from benzol are, 

 hemimellitic acid, 185°; trimellitic acid, 216°; trimesic acid, above 

 300°. So there can be no doubt that our acid is trimesic acid., and 

 therefore, if there has been no intramolecular transformation, the ben- 

 zoltrisulphonic acid has the symmetrical constitution 1, 3, 5. 



Second Method. — The benzoltrisulphochloride was heated with an 

 excess of phosphoric pentachloride, according to the method of Bar- 

 baglia and Kekule * for replacing a sulpho-group by an atom of 

 chlorine. For this purpose 3 grm. of the chloride were sealed with 

 5.2 grm. of phosi^horic pentachloride (one molecule of the sulpho- 

 chloride to six of the phosphoric pentachloride), and heated for some 

 hours to 200°-210°. Upon treating the product with water long 

 needles were deposited, which were purified by washing with sodic 

 hydrate, and several crystallizations from alcohol, when they melted 

 at 63°. The yield was about 62 per cent of the theoretical. The 

 melting point of this substance would indicate that it is the symmetri- 

 cal trichlorbenzol 1, 3, 5, wliich melts, according to Beilstein and 



* Ber. d. di. G., 1872, p. 875. 



