The same has long been the case in Chemistry. As I have stated elsewhere[67\4], the chemical nomenclature of the oxygen theory was for a time very useful and effective. But yet it had defects which could not be overlooked, as I have already stated under Aphorism II. The relations of elements were too numerous, and their numerical properties too important, to be expressed by terminations and other modifications of words. Thus the compounds of Nitrogen and Oxygen are the Protoxide, the Deutoxide, Nitrous Acid, Peroxide of Nitrogen, Nitric Acid. The systematic nomenclature here, even thus loosely extended, does not express our knowledge. And the Atomic Theory, when established, brought to view numerical {360} relations which it was very important to keep in sight. If _N_ represents Nitrogen and _O_ Oxygen, the compounds of the two elements just mentioned might be denoted by _N_ + _O_, _N_ + 2_O_, _N_ + 3_O_, _N_ + 4_O_, _N_ + 5_O_. And by adopting a letter for each of the elementary substances, all the combinations of them might be expressed in this manner. [Note 67\4: _Hist. Ind. Sc._ b. xiv. c. 6.] But in chemistry there are different orders of combination. A salt, for instance, is a compound of a base and an acid, each of which is already compound. If _Fe_ be iron and _C_ be carbon, _Fe_ + _O_ will be the protoxide of iron, and _C_ + 2_O_ will be carbonic acid; and the carbonate of iron (more properly carbonate of protoxide of iron), may be represented by (_Fe_ + _O_) + (_C_ + 2_O_) where the brackets indicate the first stage of composition.