wings by another electromagnet fixed vertically under it (fig. 29). As
the pendulum approached the vertical and was say half an inch from its
lowest point, the current would be switched on, and switched off as soon
as the pendulum got to its lowest point. A very small attraction with
this arrangement, probably about a grain weight, acting through the 1/2
in. would drive a heavy pendulum. A switch would have to be worked in
connexion with the pendulum. A strip of ebonite with a small face of
metal on the end of one side, such as a b (fig. 29) might be pivoted at
one end on the pendulum with a weak spring to keep it where free along
the rod. As the pendulum swung by this would be swept on its journey
from left to right against a fixed pin P. This would complete the
electric circuit down through the pendulum rod, round the coil on the
bottom of the pendulum, through the switch into the pin P, thence
through the fixed electromagnet, and so back to the battery. On the
return journey no contact would be made because only the ebonite face of
the switch would touch P. The pendulum would thus receive an impulse
every other vibration. We have described this switch, not to advocate
it, but to warn against its use. For the contact would be quite
insufficient. In order that the switch might not unduly retard the
pendulum it must be light, but this would make the pressure on P too
light to be trustworthy. Moreover, the strength of the impulse would
vary with the strength of the battery, and hence the arc would be
repeatedly uneven.

In contrast with this, let us consider a clock that is now giving
excellent results at the Observatory of Neuchatel in Switzerland on
Hipp's system (_La Pendule electrique de precision_, Neuchatel, 1884 and
1891). The pendulum (fig. 30) consists of two rods of steel joined by
four bridges, one just below the suspension spring, the next about 12
in. lower, the next about half way down, and the last supporting a glass
vessel of mercury which forms the bob. On the third of them is placed an
iron armature, which works between the poles of an electromagnet fixed
to the case, and by which the pendulum is actuated. The circuit is
closed and broken by a flipper, which is swayed to and fro by a block
fixed to the pendulum at the second bridge. As long as the flipper is
merely swayed, no contact takes place, but when the arc of vibration of
the pendulum is diminished the flipper does not clear the block but is
caug