the potential is reversed. Let [Greek: iota]_{1} be the charge per sec. communicated to the upper electrode when the lower plate is charged positively and ι_{2} the value when charged negatively. Then
[Greek: iota]_{1} = [Greek: iota]_{0} + ne,
[Greek: iota]_{2} = [Greek: iota]_{0} + ne;
adding we get ne = ([Greek: iota]_{1} + [Greek: iota]_{2})/2._{0} in one of the above two equations]
Now in the third column of the above table it is seen that ([Greek: iota]_{1} + [Greek: iota]_{2})/2 has the values ·39, ·41, ·43 for 2, 4, and 8 volts respectively. The numbers are thus in fairly good agreement. Similar results were obtained when a brass plate was substituted for the upper electrode shown in the figure. Taking into consideration that the magnitude of ne is independent of the strength of the magnetic field above a certain small value, and the good agreement of the numbers obtained with variation of voltage, I think that there can be no doubt that the positive charge communicated to the upper electrode was carried by the [Greek: alpha] particles. This positive charge was not small, for using a weight of ·48 mgrs. radium bromide spread in a thin foil over an area of about 20 sq. cms., the charge communicated by the particles corresponded to a current 8·8 × 10^{-13} amperes, and, with the Dolezalek electrometer employed, it was necessary to add a capacity of ·0024 microfarads to the electrometer system.
In these experiments, the film of radium bromide was so thin, that only a very small percentage of the [Greek: alpha] particles was stopped by the radium itself. Assuming that each [Greek: alpha] particle carries the same charge as an ion, viz. 1·1 × 10^{-19} coulombs, and remembering that half of the [Greek: alpha] particles are absorbed in the lower plate, the total number N of [Greek: alpha] particles expelled per second from one gram of radium bromide (at its minimum activity) can be deduced. In two separate experiments where the amount of radium used was ·194 and ·484 mgrs. respectively, the values of N were in close agreement and equal to 3·6 × 10^{10}. Now it will be shown later that in radium there are three other products in radio-active equilibrium, each of which probably gives out the same number of [Greek: alpha] particles as radium itself. If this is the case, the total number of [Greek: alpha] particles expelled per second from 1 gram of radium bromide in radio-active equilibrium is 4N or 1·44 × 10^{11}. Assuming the
