own 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.

[Illustration: Fig. 7.]

I wish now, in closing, to refer briefly to phenomena of moving lines
of force, and to the effects of speed of movement. In order to
generate a given potential in a length of conductor we have choice of
certain conditions. We can vary the strength of field and we can vary
the velocity. We can use a strong field and slow movement of
conductor, or we can use a weak field and rapid movement of the
conductor. But we find also that where the conductor has large section
it is liable to heat from eddy currents caused by one part of its
section being in a stronger field than another at the same time. One
part cuts the lines where they are dense and the other where they are
not dense, with the result of difference of potential and local
currents which waste energy in heat. We cannot make the conductor move
in a field of uniform density, because it must pass into and out of
the field. The conditions just stated are present in dynamos for heavy
current work, where the speed of cutting of lines is low and the
armature conductor large in section.

But we find that in a transformer secondary we can use very large
section of conductor, even (as in welding machines) 12 to 15 square
inches solid copper, without meeting appreciable difficulty from eddy
currents in it. The magnetic lines certainly cut the heavy conductor
and generate the heavy current and potential needed. What difference,
if any, exists? In the transformer the currents are generated by
magnetic field of very low density, in which the lines are moving
across the conductor with extreme rapidity. The velocity of emanation
of lines around the primary coil is probably near that of light, and
each line passes across the section secondary conductor in a
practically inappreciable time. There is no cause then for differences
of potential at different parts of the section heavy secondary. Then
to avoid eddy currents in large conductors and generate useful
currents in them, we may cause the conductor to be either moved into
and out of a low density field with very great speed, or better, we
must cause the lines of a very low or diffused fie