Let us assume a plain iron core, Fig. 7, magnetized as indicated, so that its poles, N, S, complete their magnetic circuits by what is called free field or lines in space around it.  Let a coil of wire be wound thereon as indicated.  Now assume that the magnetism is to be lost or cease, either suddenly or slowly.  An electric potential will be set up in the coil, and if it has a circuit, work or energy will be produced or given out in that circuit, and in any other inductively related to it.  Hence the magnetic field represents work or potential energy.  But to develop potential in the wire the lines must cut the wire.  This they can do by collapsing or closing on themselves.  The bar seems, therefore, to lose its magnetism by gaining it all, and in doing so all the external lines of force moving inward cut the wire.  The magnetic circuits shorten and short-circuit themselves in the bar, perhaps as innumerable molecular magnetic circuits interior to the iron medium.  To remagnetize the bar we may pass an electric current through the coil.  The small closed circuits are again distended, the free field appears, and the lines moving outward cut across the wire coil opposite to the former direction and produce a counter potential in the wire, and consequent absorption of the energy represented in the free field produced.  As before studied, the magnetism cannot disappear without giving out the energy it represents, even though the wire coil be on open circuit, and therefore unable to discharge that energy.  The coil open-circuited is static, not dynamic.  In such assumed case the lines in closing cut the core and heat it.  Let us, however, laminate the core or subdivide it as far as possible, and we appear to have cut off this escape for the energy.  This is not really so, however.  We have simply increased the possible rate of speed of closure, or movement of the lines, and so have increased for the divided core the intensity of the actions of magnetic friction and local currents in the core, the latter still receiving the energy of the magnetic circuit.  This reasoning is based on the possibility in this case of cutting off the current in the magnetizing coil and retaining the magnetic field.  This is of itself probably impossible with soft iron.  That the core receives the energy when the coil cannot is shown in the well known fact that in some dynamos with armatures of bobbins on iron cores, the running of the armature coils on open circuit gives rise to dangerous heating of the cores, and that under normal work the heating is less.  In the former case the core accumulates the energy represented in the magnetic changes.  In the latter the external circuit of the machine and its wire coils take the larger part of the energy which is expended in doing the work in the circuit.  In this case, also, the current in the coils causes a retardation of the speed of change and extent of change of magnetism in the iron cores, which keeps down the intensity of the magnetic reaction.  In fact, this retardation or lag and reduction of range of magnetic change may in some machines be made so great by closing the circuit of the armature coils themselves or short-circuiting them that the total heat developed in the cores is much less than under normal load.