Thus, for the first time in science, the idea appears that the atom is not the smallest fraction of matter to be considered. Fragments a thousand times smaller may exist which possess, however, a negative charge. These are the electrons, which other considerations will again bring to our notice. § 3. HOW IONS ARE PRODUCED It is very seldom that a gaseous mass does not contain a few ions. They may have been formed from many causes, for although to give precision to our studies, and to deal with a well ascertained case, I mentioned only ionisation by the X rays in the first instance, I ought not to give the impression that the phenomenon is confined to these rays. It is, on the contrary, very general, and ionisation is just as well produced by the cathode rays, by the radiations emitted by radio-active bodies, by the ultra-violet rays, by heating to a high temperature, by certain chemical actions, and finally by the impact of the ions already existing in neutral molecules. Of late years these new questions have been the object of a multitude of researches, and if it has not always been possible to avoid some confusion, yet certain general conclusions may be drawn. The ionisation by flames, in particular, is fairly well known. For it to be produced spontaneously, it would appear that there must exist simultaneously a rather high temperature and a chemical action in the gas. According to M. Moreau, the ionisation is very marked when the flame contains the vapour of the salt of an alkali or of an alkaline earth, but much less so when it contains that of other salts. Arrhenius, Mr C.T.R. Wilson, and M. Moreau, have studied all the circumstances of the phenomenon; and it seems indeed that there is a somewhat close analogy between what first occurs in the saline vapours and that which is noted in liquid electrolytes. There should be produced, as soon as a certain temperature is reached, a dissociation of the saline molecule; and, as M. Moreau has shown in a series of very well conducted researches, the ions formed at about 100°C. seem constituted by an electrified centre of the size of a gas molecule, surrounded by some ten layers of other molecules. We are thus dealing with rather large ions, but according to Mr Wilson, this condensation phenomenon does not affect the number of ions produced by dissociation. In proportion as the temperature rises, the molecules condensed round the nucleus disappear, and, as in all other circumstances, the negative ion tends to become an electron, while the positive ion continues the size of an atom.