separated, it would follow that the weakest electrical attraction was stronger than, if not the strongest, yet very strong chemical attraction, namely, such as exists between oxygen and hydrogen, potassium and oxygen, chlorine and sodium, acid and alkali, etc., a consequence which, although perhaps not impossible, seems in the present state of the subject very unlikely.
243. The view which M. de la Rive has taken (225), and also MM. Riffault and Chompré (221), of the manner in which electro-chemical decomposition is effected, is very different to that already considered, and is not affected by either the arguments or facts urged against the latter. Considering it as stated by the former philosopher, it appears to me to be incompetent to account for the experiments of decomposition against surfaces of air (198, 205) and water (231), which I have described; for if the physical differences between metals and humid conductors, which M. de la Rive supposes to account for the transmission of the compound of matter and electricity in the latter, and the transmission of the electricity only with the rejection of the matter in the former, be allowed for a moment, still the analogy of air to metal is, electrically considered, so small, that instead of the former replacing the latter (198), an effect the very reverse might have been expected. Or if even that were allowed, the experiment with water (231) at once sets the matter at rest, the decomposing pole being now of a substance which is admitted as competent to transmit the assumed compound of electricity and matter.
244. With regard to the views of MM. Riffault and Chompré (221), the occurrence of decomposition alone in the course of the current is so contrary to the well-known effects obtained in the forms of experiment adopted up to this time, that it must be proved before the hypothesis depending on it need be considered.
245. The consideration of the various theories of electro-chemical decomposition, whilst it has made me diffident, has also given me confidence to add another to the number; for it is because the one I have to propose appears, after the most attentive consideration, to explain and agree with the immense collection of facts belonging to this branch of science, and to remain uncontradicted by, or unopposed to, any of them, that I have been encouraged to give it.
246. Electro-chemical decomposition is well known to depend essentially upon the current of electricity. I have shown that in certain cases (m) the decomposition is proportionate to the
