for enlarged projections of
sound vibrations.

As soon, therefore, as I gave attention to the subject of telephony, I
was able, with a few preliminary experiments, to determine the proper
conditions for the transmission of speech in an electric circuit; and,
without the slightest knowledge of the mechanism which Prof. Bell had
used, I devised the following arrangement for a speaking-telephone.

[Illustration: FIG. 14.--MY FIRST SPEAKING TELEPHONE.]

[Illustration: FIG. 14.--END VIEW.]

My first speaking-telephone, Fig. 14, consisted of a magnet made out of
half-inch round steel bent into a U form, having the poles about two
inches apart. Over these were slipped two bobbins taken from an old
telegraph register, and were already fitted to a half-inch core. These
bobbins, two inches and a half long, were wound with cotton-covered
copper wire, No. 23, each bobbin containing about 150 feet. This magnet,
with the bobbins slipped upon its poles, was made fast to a post two or
three inches high. The steel was made as strongly magnetic as was
possible, and would hold up three or four times its own weight. In
front of the poles, a sheet of thin steel, one-fiftieth of an inch
thick, was made fast to an upright board having a hole cut through it
three and a half inches in diameter (Fig. 14, end view); the plate was
screwed tightly to this board, so as to cover the hole; and the middle
of the hole was at the same height as the two poles of the magnet. The
wires from the two bobbins were connected, as if to make an
electro-magnet; while the two free terminals were to be connected with
the line-wires. Of course there were two of these instruments, both
alike; and talking and singing were reproduced with these.

A very great number of experiments have been made to determine the best
conditions for each of the essential parts,--the size and strength of
the magnet, the size of the bobbins, as to length and fineness of wire,
the best thickness for the plate for absorbing the vibrations, &c.; and
it is really surprising, how little is the difference between very wide
limits. The following directions will enable any one to construct a
speaking-telephone with which good results may be obtained. The
specifications will be for only one instrument; though of course two
instruments made alike will be necessary for any purposes of speaking or
other signals.

[Illustration: FIG. 15.]

[Illustration: FIG. 16.]

Procure three