446 On definite Proportions. 
are however so far of material value, as they léad us witts 
some certainty to the traces of the Jaws of nature. 
That the ammonium of the alkali is not here an integral 
multiple of that of the acid, depends on the same cause as 
the apparently anomalous progressions of several combus- 
tible bodies, of which I have partly spoken already, and 
shall partly have occasion to speak more particularly in 
treating of the vegetable acids. 
The : analysis of the nitrate of ammonia may be considered 
as a formal proof of the existence of oxygen in hydrogen: 
for there can be no other reason why ‘the oxygen of the 
acid, when considered as ‘formed from nitrogen, should not 
bea ‘multiple of that of the base by a whole “number. But 
this analysis seems at the same time to show, ihat [ have 
estimated the quantity of oxygen, in the First Continuation 
of my Essay, much too high, and perhaps at Jeast four times 
as much as the truth. 
Since I have reason to believe from my investigation of 
the composition of ammonia, and of the neutral 1 nitrates, 
that nitrogen is to be considered as a higher degree of oxi- 
dation of the same radical that forms ammonia, I think it 
follows from the same grounds, although not with the same 
force of evidence, that “hydrogen ‘must consist of the same 
radical in a lower degree of oxidation. It appears however, 
that the oxygen of the hydrogen should on this supposition 
be a multiple of that of any body with which water can 
combine, by some whole number, which lies between those 
of the oxygen in water and in the other substance, exactly 
-as we shall find that of the oxygen In nitrogen with respect 
to the neutral nitrates. But this cannot be, if water really 
contains only 113 per cent. of hydrogen. If on the other 
hand water really contained more than 113 per cent. of 
hydrogen, as Mr. Gay-Lussace has concluded from some 
reasons with which I am unacquainted, the Jast-mentioned 
number, with which most of the analyses of bodies con- 
taining water agree best, would belong to the metallic am- 
monium, and the difference between the quantity of hydro- 
gen and ammonium in water would depend on the oxygen 
of the hydrogen. If however, as is most probable, future 
analyses of water, performed with accuracy proportional to 
the present state of the investigations, should still make 
the quantity of hydrogen in water accurately or very nearly 
11:75 per cent., it will be difficult to reconcile the existence 
of oxygen in hydrogen with the calculations relating to 
multiple proportions. Since however this, as we shall here- 
after find, happens also in some cases with respect to the 
oxygen 
