What has now been said relates merely to the facts ob served. In chemical philosophy it is further requisite to try if an account can be given why the facts are thus and not otherwise. Some memoirs of M. Berzelius hat e had in view to examine the cause of chemical proportions.
as Whenever we begin to consider this matter," he ob serves, " it is evident, at first sight, that the cause cannot be any thing else than of a mechanical nature; and the idea which seems most probable and best suited to the views suggested by experience, is, that bodies are com posed of molecules or atoms, which combine one with one, with 2, 3, 4, &c. and the laws of chemical proportions appear to result from this principle, in a manner so clear and evident, that it seems strange how an idea so simple and rational should have failed, not only to be adopted, but even to be stated till our own age." This hypothesis gains an additional degree of credibility when applied to those electro chemical facts, by which we have just learned that all the phenomena of affinities are, in truth, nothing more than phenomena of an electric action between the bodies mutually combined or decomposed. Supposing these atoms or elementary molecules, of which bodies are formed, to be endowed with an electric polarity, by which their affi nities are exercised, we are enabled to comprehend how the forces, named chemical affinities, may be the same as the opposite electric states, named E and — E. By this means, the phenomena of chemical proportions will not be difficult to understand, if we admit that molecules com bine one atom or molecule with one or several elementary molecules ; and a corpuscular theory that shall not omit the forces on which the combinations of molecules de pend, will henceforth constitute the basis of chemistry and physics; whether, in fact, this theory be a true expo sition of the nature of things, or only a mode of repre sentation, enabling us to know and comprehend what otherwise must have remained inexplicable and undisco vered. Now, if experience has begun to ratify such a re presentation of the intimate composition of bodies, the se cond step will be to attempt discovering the number of molecules belonging to each element in each combination. Researches of this kind are doubtless extremely difficult; their first results will perhaps at best be doubtful, yet still it is plain, that any supposition as to this point, if taken up at hazard, cannot have the smallest value. Mr. Dalton was the first who attempted to compute the molecules of the elements existing in several compound inorganic bo dies. He set out, on this investigation, from a principle altogether artificial. We have already mentioned, that, when there is but one known combination of two element ary bodies, Mr. Dalton considers it as containing a mole cule of each element; but that when there are two or more combinations, he allows himself to be guided by the pro portion subsisting between them. Experience, however, daily shows that we are not yet acquainted with all the de grees of combination; and when it happens, that of seve ral possible degrees of combination, we have discovered but a single one, nothing can assure us, that this must be precisely the degree which contains only a single molecule of each of its elements. It is impossible, then, according to Bei zelius, that Mr. Dalton's application of the corpus cular hypothesis can ever give satisfactory results. To dis cover the number of molecules in oxydized bodies, Al.
Berzelius made use of two circumstances. He examined (a) the different degrees of oxydation in some one radical. Suppose that of the several possible oxvds of this radical but two ale known; their quantities of oxygen being in the ratio of 2 to 3, of 3 to 4, or of 4 to 5, he infers that these uxyds contain also this number of molecules, for other wise (granting always the hypothesis of molecules) one of them would contain a fraction of a molecule, which is not admissible. Combining (b) any oxyd with another oxyd, which may serve as a (saline) base to it, or which, in re spect of it, is electro-negative—most frequently it happens that the oxygen contained in the electronegative oxyd is a multiple by 2, 3, 4, &c. of the oxygen contained in the clectro-positive oxyd. It follows next that the number must be such as to introduce no fraction of a molecule into the electronegative oxyd. An oxyd of A, for example, which contains 3 molecules of oxygen—if it be combined with another B, su that A contained four times as much oxygen as B—would always presuppose a fraction of a molecule to exist in the radical of A. But in examining the combina tions of oxyds with each other, in their different degrees of mutual saturation, Berzelius has lately found that both the circumstances (a and b) lead always to the same result; and that, from their coincidence, the number of molecules in the greater part of oxyds, may be inferred with consider able certainty. It follows, that most of the oxyds, which, according to Dalton, contain only 1 molecule of oxygen, contain in reality 2 or 3. Yet there are oxyds, the number of whose molecules cannot be determined by any experi ment; such are potass and the earths in general. Berze bus has attempted to supply this deficiency by a train of analogical reasoning, which, though it. no doubt occasion ally misleads, is, in this case, the only clue we have to guide us. The analogy of potass to soda induces a suppo sition, that in each of those two alkalis the number of atoms is the same ; but we know, from analyzing the su peroxyd of soda, that this alkali must contain 2 molecules of oxygen ; from which we may perhaps conclude, that potass contains 2 likewise. Besides, on comparing toge ther those two metallic oxyds which are electropositive, (that is to say, which give salts with acids,) it appears that such as contain 2 molecules of oxygen form always the strongest saline base, and show the closest analogy to alkalis. Such as contain but one molecule, either do not at all combine with acids, or constitute a separate class of saline bases; they have a strong mutual resemblance in their general properties, and their salts differ in a charac teristic manner front those of oxyds containing 2 molecules of oxygen: such are the minimum oxyds of copper, of mercury, platina, rhodium, and gold. With regard to such as contain 3 molecules of oxygen, they likewise. in gene ral, form weaker saline bases, and even frequently, in the state of electronegative oxyds, themselves combine with saline bases. From all this, Berzelius concludes, that, for the present, the fixed alkalis and the earths may, with most plausibility, be looked upon as containing 2 mole cules of oxygen. As to alumina, on examining the pro portions in which this earth is found combined with po tass in alum and feldspar, with magnesia and the oxyd of zinc, in the spinelle and gahnite, Berzelius imagined it to contain 3 molecules of oxygen. The sime is probably true of glucina.