. Just as he
seizes the pattern drawing lever with his right hand, he presses with
his left on the head of a compression valve shown at the left side of
top of machine, thus admitting air to the pneumatic vibrator already
referred to.

[Illustration: FIG. 3.--POWER DRIVEN VIBRATOR MACHINE.]

Fig. 3, a rear view of the machine, shows at the top center, with its
inlet hose hanging to it, this vibrator, which is shown in section in
Fig. 4. It consists simply of a double acting elongated piston having
a stroke of about 5/16 inch in a valveless cylinder and impacting upon
hardened anvils at either end at the estimated rate of 5,000 blows per
minute.

[Illustration: FIG. 4.--SECTION THROUGH VIBRATOR.]

The method of communicating the rapid yet small oscillations of the
vibrator to the patterns and yet keeping them from being transmitted
to the rest of the mechanism is this:

A frame, called a vibrator frame, to which the pneumatic vibrator is
bolted and keyed, is shown in Fig. 5. To this frame the plate carrying
the patterns, often, in cases of patterns having irregular parting
lines, forming one and the same casting with the patterns, is fastened
by the four machine screws, the small tapped holes for which are shown
in the corners. In fact, in changing patterns, the process consists of
simply removing these four machine screws, taking up the pattern
plate and screwing to the vibrator frame the new pattern plate. The
vibrator frame itself is secured to the machine structure by the four
larger bolts, the holes for which are shown in the inner corners.
These bolts are, as shown in Fig. 7, surrounded by thick bushings.
These bushings are elastic to such a degree as to absorb the sharp
vibrations of vibrator frame and patterns, while so firm and well
fitted as to hold patterns accurately to their position.

[Illustration: FIG. 5.--VIBRATOR FRAME.]

The action of the vibrator is such as to give to the entire pattern
surface an exceedingly violent shiver, making it impossible that any
sand should adhere to this surface, while the magnitude of the actual
movement of the pattern is so slight that it is found to fill the
mould so completely that it is impracticable to draw it a second time
without rapping. Yet, so truly are the patterns held and so little
disturbed from their original position, that it is perfectly
practicable to return patterns to a mould having the finest ornamental
surface in the ordinary practice of "prin