LECTURE IX.] HISTORY OF CHEMISTRY. 159 



with which the salts are formed by the replacement by metals 

 of the hydrogen of the (hydrated) acid, is not to be rejected 

 unconditionally. The accuracy of this idea being premised, 

 the acids would not contain any ready-formed water, and they 

 could not be any longer regarded as consisting of anhydride 

 and water, any more than the salts were compounds composed 

 of acid (anhydride) and base. 



Liebig finds an important support for the latter hypothesis 

 (in accordance with which the metals, as such, should be 

 assumed in salts) in the behaviour of tartar emetic at a 

 high temperature. According to the analysis, the formula 

 C 8 H 8 KSb 2 Oi4 represents the compound dried at 100. It 

 was assumed to contain an atom of anhydrous tartaric acid, an 

 atom of potash, and an atom of oxide of antimony, so that its 

 formula was written C 8 H 8 O 10 + KO + Sb 2 O 3 (assuming that the 

 formula of tartaric acid was doubled). According to Liebig 

 this substance, on being heated to 300, loses two more atoms 

 of water, a property which it does not share with any other salt 

 of the same acid. The assumption of the presence of water in 

 the acid, hitherto regarded as anhydrous, appears objectionable 

 to Liebig on account of the consequences which would follow 

 from it, and so he believes that nothing remains but to ascribe 

 the formation of water to the reduction of the oxide of 

 antimony. The actual existence of a base, in the metallic 

 condition, combined with an oxygen acid (even if only for 

 certain compounds) would no longer require to be regarded as 

 a mere supposition. 20 



On another occasion, when discussing these relations, 

 Liebig writes the formula for tartaric acid, C 8 H 4 O 12 .H 8 , and 

 that of tartar emetic which has been heated to 300, 



C 8 H 4 O 12 < , and I must not omit to draw attention to the 



fact that the displacement of three atoms of hydrogen by one 

 atom of antimony, is here assumed. 21 



20 Arnalen. 26, 159. 21 The paper in question is by Dumas and 

 Liebig, Comptes Rendus. 5> 863. 



