The saturation current through air was found to be 1·2 × 10⁻⁸ amperes, _i.e._ 36 E.S. units, for parallel plates 4·5 cms. apart, when ·45 gramme of radium of activity 1000 times that of uranium was spread over an area of 33 sq. cms. of the lower plate. This corresponds to a production of about 10¹¹ ions per second. Assuming, for the purpose of illustration, that the ionization was uniform between the plates, the volume of air acted on by the rays was about 148 c.c., and the number of ions produced per c.c. per second about 7 × 10⁸. Since _N_ = 3·6 × 10¹⁹, we see that, if one molecule produces two ions, the proportion of the gas ionized per second is about 10⁻¹¹ of the whole. For uranium the fraction is about 10⁻¹⁴, and for pure radium, of activity one million times that of uranium, about 10⁻⁸. Thus even in the case of pure radium, only about one molecule of gas is acted on per second in every 100 millions. The electrical methods are so delicate that the production of one ion per cubic centimetre per second can be detected readily. This corresponds to the ionization of about one molecule in every 10¹⁹ present in the gas. =40. Size and nature of the ions.= An approximate estimate of the mass of an ion, compared with the mass of the molecule of the gas in which it is produced, can be made from the known data of the coefficient _K_ of inter-diffusion of the ions into gases. The value of _K_ for the positive ions in moist carbon dioxide has been shown to be ·0245, while the value of _K_ for the inter-diffusion of carbon dioxide with air is ·14. The value of _K_ for different gases is approximately inversely proportional to the square root of the products of the masses of the molecules of the two inter-diffusing gases; thus, the positive ion in carbon dioxide behaves as if its mass were large compared with that of the molecule. Similar results hold for the negative as well as for the positive ion, and for other gases besides carbon dioxide. This has led to the view that the ion consists of a charged centre surrounded by a cluster of molecules travelling with it, which are kept in position round the charged nucleus by electrical forces. A rough estimate shows that this cluster consists of about 30 molecules of the gas. This idea is supported by the variation in velocity, _i.e._ the variation of the size of the negative ion, in the presence of water vapour; for the negative ion undoubtedly has a greater mass in moist than in dry gases. At the same time it is possible that the apparently large size of the ion, as determined by diffusion methods, may be in part a result of the charge carried by the ion. The presence of a charge on a moving body would increase the frequency of collision with the molecules of the gas, and consequently diminish the rate of diffusion. The ion on this view may not actually be of greater size than the molecule from which it is produced.