A thin film of radium was obtained on a plate _A_ by evaporation of a radium solution containing a known weight of radium bromide. Some hours after evaporation, the activity of the radium, measured by the α rays, is about 25 per cent. of its maximum value, and the β rays are almost completely absent. The activity measured by the α and β rays is then slowly regained, and recovers its original value after about a month’s interval (see chapter XI.). The experiments were made on the active plate when its activity was a minimum, in order to avoid complications due to the presence of β rays. The film of radium was so thin that only a very small fraction of the α rays was absorbed. [Illustration: Fig. 33.] The active plate _A_ was insulated in a metal vessel _D_, and was connected to one pole of the battery, the other pole being earthed. The upper electrode, which was insulated and connected with a Dolezalek electrometer, consisted of a rectangular copper vessel _BC_, the lower part of which was covered with a thin sheet of aluminium foil. The α rays passed through the foil, but were stopped by the copper sides of the vessel. This arrangement was found to reduce the secondary ionization produced at the surface of the upper plate. The outside vessel _D_ could be connected with either _A_ or _B_ or with earth. By means of a mercury pump, the vessel was exhausted to a very low pressure. If the rays carry a positive charge, the current between the two plates measured by the electrometer should be greater when _A_ is charged positively. No certain difference, however, between the currents in the two directions was observed, even when a very good vacuum was obtained. In some arrangements, it was found that the current was even greater when the lower plate was negative than when it was positive. An unexpected experimental result was also noticed. The current between the parallel plates at first diminished with the pressure, but soon reached a limiting value which was not altered however good a vacuum was produced. For example, in one experiment, the current between the two parallel plates, placed about 3 mms. apart, was initially 6·5 × 10⁹ amperes and fell off directly as the pressure. The current reached a limiting value of about 6 × 10⁻¹² amperes, or about ¹⁄₁₀₀₀ of the value at atmospheric pressure. The magnitude of this limiting current was not much altered if the air was replaced by hydrogen.