occupies immediately disappear. For our purposes the manner of this disappearance of the tubes of force is of no consequence: it may be described as a shrinking of the tubes in the conductor, so that the ends which leave the plates of the condenser move toward the middle of the conductor and at last meet. The disappearance of these tubes from the conductor relieves those lying immediately around it of the lateral pressure which maintained them in equilibrium, and they are accordingly driven into the conductor and in turn disappear. This process is continued until all the tubes of force have disappeared. The discharge of the condenser may therefore be represented as the lateral movement of the tubes of force originally in the field and their disappearance in the conductor. The discharge is not really so simple as it is here supposed to be. We have supposed the process to cease when the difference of potential between the two plates becomes zero, but this is in fact not the case: the discharge is really oscillatory, the difference of potential being alternately positive and negative, rapidly diminishing in numerical value until it disappears. In order to account for this, something analogous to inertia must be ascribed to the tubes of force. While the discharge is going on, a magnetic field exists around the conductor. If the discharge be thought of as being merely the transfer of charge along the conductor, there is no apparent mechanism connecting the discharge with the magnetic force, but on the view now being presented the magnetic field may be thought of as due in some way to the movement through the dielectric of the tubes of force.
If the discharge pass through a compound body, capable of decomposition by it, a portion of that body will be resolved into two constituents. On the view that the discharge is a mere transfer of charge, these constituents must be supposed to serve as carriers of that charge, but this view cannot represent the connection between these constituents and their charges, nor the conditions which enable them to give up their charges. On the other hand, by supposing each unit of the constituent to be invariably associated with a tube of force, we may describe the discharge through such a chemical compound in terms of the changes which
