198 EDWARD DIVERS AND TAMEMASA HAGA; 



which was almost certainly a double salt. Jacquelain did very 

 completely and carefully analyse his, and the result shows it to have 

 been a barium ammonium salt, as we explained in our first paper. 

 Berglund found the normal barium salt to gïve up its 5 mois, of water 

 almost completely at 100° ; we maintain the accuracy of our statement 

 that it loses water only very slowly even at 115° (see further on this 

 point our account of the strontium salt). Barium imidosulphonate is 

 soluble in a solution of ammonium chloride. 



Tiro-thirds normal barium imidosulphonate was fully described by 

 Berglund, his account agreeing with ours. 



Double salts of barium with alkali* received peculiar treatment by 

 Berglund. He did not formally recognise their existence and relegated 

 to foot-notes observations which, he admitted, made their existence 

 probable. He could not satisfactorily formulate the composition he 

 found them to have, and for him they remained as impure barium salt 

 only. Yet our own work confirms the results of his and makes 

 it possible to give formulae to his preparations. First, there is 

 Ba 1 ,K 8 H(NSo0 6 ) 1I , HHoO, which he constantly obtained when he added 

 normal potassium imidosulphonate to barium chloride. Had he 

 thought of the device, here adopted, of introducing the atom of 

 hydrogen into the formula, he would probably have not looked upon 

 his analytical results as incapable of being interpreted. À sodium salt 

 described by us, comes very near to this salt, being Ba^a^NSoOe)^, 

 loH L ,0 ; for if we subtract a mol. of the two-thirds normal barium salt, 

 BaHNS 2 6 , from his formula and write Nu for K, we get ours. 

 Evidently, either salt is mainly BaK(or Na)NS 2 6 with a little 

 Ba^NSoOc).. By adding two-thirds normal potassium salt to ammonia 

 and barium chloride, he got Ba 12 K 7 (NH 4 ) 2 (NS 2 6 ) n , 79H 2 0. Lastly, 

 thesalts, Ba 6 (NH 4 ) 2 H(NS 2 6 ) 6 , 2lH 2 and Ba 7 (NH 4 )(NS 2 G 6 ) 5 , 2211,0 ; 

 all three having an obvious relation to the first. We have described a 



