350 PROCEEDINGS OP THE AMERICAN ACADEMY 



182° (C6n..,('S02Cl)3 melts at 184°), proving that the trisulphouic acid 

 had beeu formed, while (b) gave a potassium salt crystallizing in an 

 asbestos-like* mass, which yielded a viscous chloride, from which no 

 trace of the benzoltrisulphochloride could be obtained ; and, as both 

 the potassic benzoltrisulphonate and its chloride are very character- 

 istic substances, there can be no doubt that in this latter case no 

 appreciable amount of the trisulphouic acid was formed. In a second 

 experiment the benzolmetadisulphonic acid (made by heating the chlo- 

 ride with water to 150° in a sealed tube) was used, and the result was 

 the same ; the specimen heated with sulphuric acid and potassic sul- 

 phate gave benzoltrisulphonic acid recognized as in the previous 

 experiment, while that heated with sulphuric acid alone gave no trace 

 of this product. 



The experiments just described having proved that the presence of 

 potassic sulphate is necessary for the formation of the trisulphouic 

 acid, we next considered the question, " Cannot the trisulphonic acid 

 be formed by the action of acid potassic sulphate alone without an 

 excess of sulphuric acid ? " and found tliat potassic benzolmetadisul- 

 phonate fused with acid potassic sulphate remained unaltered, even 

 at a temperature where the mass began to char. It is evident, there- 

 fore, that the potassic sulphate serves only to render the sulphuric 

 acid more efficient, and it remained to determine by experiment, if 

 possible, the manner of its action. Since in Senhofer's method, the 

 only one beside ours by wliich the benzoltrisu]23honic acid has been 

 ol)tained, the action was due to the presence of a powerful dehydrating 

 agent (phos[)horic pentoxide), the most plausible theory was that the 

 acid potassic sulphate acted simply by the removal of water. To test 

 this hypothesis we substituted for the potassic sulphate in one experi- 

 ment argentic sulphate, and in another zincic sulphate, as, if the action 

 of the salt is due only to its attraction for water, like that of the phos- 

 phoric pentoxide in Senhofer's method, the latter should produce quite 

 as good a result as potassic sulphate, whereas the argentic sulphate 

 would be entirely without action. Upon heating three portions of 

 benzolmetadisulphonic acid with the same proportion of sulphuric 

 acid, (a) without the addition of a salt, (b) with addition of zincic 

 sulphate, and (c) with addition of argentic sulphate, we found that, 

 while no trace of the trisulphouic acid could be detected in (a), or (b), 



* We liave not succeeded in determuiing the nature of this potassium salt. 

 It does not resemble tlie potassic benzolmonosulplionate, which crystallizes in 

 glistening plates, nor have we found any disulphonate corresponding to it in 

 properties. We hope to return to this subject hereafter. 



