separates and insulates these plates, and they are bound together with
the ring into a firm structure by a tube of hard rubber, having a
shoulder and knob at the top, and at the lower end a screw-thread
engaging with a thin nut soldered to the upper side of the bottom plate.
When the cover is in place, its lower plate is even with the top of
the cell; and the contained water, which nearly fills the cell, is
surrounded by polished, nickel-plated, brass plates 0.01 inch thick,
insulated trom other metal by interposed hard rubber. The spaces between
the cell and case (a single space if the partitions are omitted), the
space above the hard rubber rings, and the space or spaces in the cover
are all filled with eider-down, which costs $1.00 per ounce avoirdupois,
but a few ounces are sufficient. Soft, fine shavings, or turnings of
hard rubber, are said to be excellent as a substitute for eider-down.
Heat cannot be confined by any known method. Its transmission can be in
some degree retarded, and in a greater degree, perhaps, regulated. Some
heat will be promptly absorbed by the sides, bottom, and cover of the
cell, and by the agitator; but this does no harm, as its quantity can
be accurately ascertained and allowed for. Some will be gradually
transmitted to the eider-down, filling the spaces, and through this to
the outer casing; but this can be reduced to a minimum by rapid and
skillful manipulation, and its quantity, under normal conditions, can
be ascertained approximately, so as not to introduce large errors. But
varying external influences, such as currents of air, caused by opening
doors, or by persons passing along near the apparatus during
the progress of an experiment, which would introduce disturbing
irregularities, can best be guarded against by such spaces as I have
described, filled with the poorest heat-conductor and the lightest
_solid_ substance attainable. Air, although a poor heat-conductor, and
extremely light, is diathermous, and offers no obstruction to the escape
of radiant heat.