y made, are liable to this objection. The best connection, no
doubt, is that of the screw coupling.

The insulation of the rod from the building is an expense not only
without the least advantage, but the contrary. Harris (_Thunderstorms_,
pp. 129, 131) says: "This practice is not only useless, but
disadvantageous, and is manifestly inconsistent with the principles on
which conductors are applied." Dr. Franklin says: "The rod may be
fastened to the wall, chimney, etc., with staples of iron. The lightning
will not leave the rod to pass into the wall through these staples. It
would rather, if any were in the wall, pass out of it into the rod, to
get more readily by that conductor to the earth." The practice may have
gained vogue from an observance of the use of glass knobs as insulators
of telegraph-wires. Many intelligent people have failed to apprehend the
vast difference between the low tension of voltaic electricity and
frictional electricity, lightning being in the nature of the latter. The
fact that when lightning strikes the telegraph-wire it jumps from the
wire to the posts, often tearing in pieces half a dozen in a row, ought
to be conclusive in regard to insulating lightning-rods.

The same considerations will also effectually dissipate the fallacy by
which the horizontal lightning-rod has duped so many people in certain
portions of the West; for if the wire be cut off from the
ground-connections (in which condition it accords with the conductor in
question) the posts (which answer to the building thus "protected") must
suffer still greater damage. So far from being insulated, the rod should
be connected with all considerable masses of metal in the building,
these having also a good connection with the earth. Frequently during a
thunder-shower--sometimes even on the approach of one--all metallic
objects will be electrified, and those of considerable size will often
yield a spark; and this without the building containing these objects
being struck. When struck, the larger masses of metal might occasion a
dangerous explosion from induction, though at some distance from the
rod: for this reason, as before stated, they should be connected with
the ground. Being then liable to receive a part of the current from the
conductor in case this be too small, they should be connected with it,
as otherwise the current would cause damage in its passage. In a word,
therefore, all metal bodies in a building should, as far as pos