On definite Proportions. 175 
oxidation, merely multiples of the lowest degree of oxida- 
tion by 2 and 4, so that, in this lowest degree, 100 parts of 
carbon may be united only with 62°9 of oxygen. In the 
same manner we may expect, from the propoftions deter- 
mined for carbon and hydrogen, that 100 parts of hydro- 
en may be saturated at a minimum by 74°584 of carbon, 
and that the observed proportions are multiples of this by 
4and 8. We want therefore, in these combinations, be- 
sides the lowest degree of saturation, also the multiples by 
2 and 6, which we are to look for in the composition of 
organized bodies. I have attempted, in various ways, to 
separate, by means of the electrical column, the combusti- 
ble base or radical of the vegetable acids from the oxygen 3 
but I have never been able to succeed. I was induced to 
perform these experiments by the reduction of ammonia, 
the base of which I then considered as consisting of hydro- 
gen and nitrogen, and as standing in the same relation to 
the metals as the combustible radical of the vegetable acid 
bears to sulphur or phosphorus. Perhaps all these are com- 
binations which cannot exist independently. Probably also 
hydrogen is capable of lower degrees of oxidation than 
that in which it constitutes water, which in this case must 
correspond to divisions ot submultipies by 8,6,4,or2. If 
in the analysis of such bodies it happens, for example, that 
carbon and hydrogen are united with oxygen in a proportion 
which does not agree with the numbers appropriate to those 
substances, we may attempt to divide the oxygen between 
both the others ; and if we obtain, in this manner, propor- 
tious which can he reconciled with the general theory, we 
may be permitted to consider the body as composed of two 
different oxides. [shall here only adduce the example of 
the subsulphate of the oxide of iron, which, though very 
regularly constituted, contains for each 100 parts of iron, 
22 parts of sulphur; a number of which the quantity of 
sulphur primarily combining with the iron is no integer 
multiple. We shail learn from this and other similar cir- 
cumstances to understand how Nature, with all her sim- 
plicity, can still be so astonishingly diversified. 
SUPPLEMENT*. 
Tue possibility of determining the proportions of the 
component parts of chemical compounds by computation 
being once established, it becomes of importance to provide 
* Trom the German Original in Gilbert's Ann, 1812. vi. 
a very 
