g over the cavity of the hook.

In the shuttle the tension was obtained by threading through holes in the
shell, or beneath a tension plate, as in Howe's machine. This tension, so
long as the reel ran between spring centers, was never constant. The
variation was chiefly due to the angular strain set up when unwinding from
the reel. This strain varied according to the point of unwinding. It was
light in the middle of the reel and heavy at either extremity. These
drawbacks caused immense anxiety to the first makers of sewing machines,
and numerous attempts to overcome them led to little improvement. With
reference to high rates of speed, the older shuttle, requiring a long and
noisy reciprocation, had its disadvantages.

The only effective remedy for these drawbacks was a radical one. It was
necessary to substitute depth of reel for length. Hence, several attempts
have been made to construct disk or ring shuttles. Many forms of those have
been tried. They all depend upon the principle of coiling up the thread in
a vertical plane, rather than in horizontal spirals. Some makers placed the
disk in a horizontal plane, and caused it to revolve. Nothing could be
worse, as will be seen, if we follow the course the enveloping loop must
take in encircling such a shuttle. But a complete solution of the
difficulty of employing a ring shuttle has been achieved in the oscillating
form, invented by Mr. Phil. Diehl, and known as Singer's (Fig. 1). A short
examination of it may profitably engage your attention. The shuttle itself
is sufficiently well known, but certain features of it, and to which it
owes its efficiency, appear to call for some explanation. Its introduction
dates back some years, during which time it has undergone certain
modifications.

[Illustration: FIG. 1.]

It consists of a thick disk bobbin of thread, _h_, fitting loosely in a
case constructed in the form of a bivalve, _a_ and _d_. This case is
furnished with a long beak, usually forming a continuation of the
periphery. The beak is intended to enter and detain the loops of upper
thread, and lead them so that they ultimately envelop the shuttle, a motion
of the thread which is chiefly due to the oscillation of the shuttle in a
vertical plane. The oscillating movement is to the extent of 180 degs. of
the circle, which suffices to cast the loops freely over the shuttle. The
center of oscillation is not coincident with the center of the shuttle; but
it is nearly so