VOLATILE CHLORIDES WITH AMMONIA. 45 
But the results of the experiments above detailed show that 
this is not the case. Both Regnault’s compound, as well as 
the sulphate of the chloride of sulphur I have prepared, take 
up more ammonia than if they were sulphuric acids in which 
a part of the oxygen was replaced by chlorine. 
Regnault’s compound is S Cl? + 2 S, or according to his 
view, S Cl. If we were to regard it as a sulphuric acid in which 
a portion of the oxygen is replaced by chlorine, and in which 
the latter fully occupies the place of the oxygen, then, since 
one atom of sulphuric acid takes up one atom of ammonia, 
to form sulphat-ammon; S Cl? + 2 S should take up three double 
atoms, and S Cl one double atom of ammonia. But according 
to Regnault’s own researches, in the first case six, in the second 
case two double atoms of ammonia are taken up by the com- 
pound; twice as much therefore as, according to the theory of 
substitution, should be expected. The combination of the chlo- 
ride of sulphur with sulphuric acid, which I have prepared, is 
SCEB+58. According to the theory of substitutions this com- 
bination would be regarded as S + 23 O + Cl, or rather as S? 
O° Cl. In the first case it ought, according to this theory, to 
combine with six; in the second with one ; and in the third 
with two double atoms of ammonia. But experiments have 
shown that, in the first case, it takes up nine, in the second one 
and a half, and in the third, three double atoms of ammonia. 
I believe I may hence conclude that all the volatile compounds 
of chlorine which have hitherto for some time been considered 
as combinations of acids with chlorides, must still be so re- 
garded, and not as acids in which a portion of the oxygen is re- 
placed by chlorine. 
Liebig has brought forward an important reason in favour of 
the theory of substitutions*. He draws attention to the isomor- 
phism discovered by Mitscherlich between the hyperchlorates 
and hypermanganates, from which it results that chlorine may 
be replaced by manganese. But in properties chlorine and 
manganese have, in fact, far less similarity to each other than 
chlorine and oxygen. Even, however, if no resemblance can 
* Annalen der Pharmacie, Bd, xxxi. p. 119. 
