34 PHYSICS. 
arc, take place in a horizontal circle, and always in the same direction. .To 
this end, however, the pendulum rod must be capable of moving about the 
point of suspension, not in a single plane only, but in any direction at pleasure. 
In the material pendulum there is still a circurastance which affects the 
oscillations, namely, the influence of temperature, which, when elevated, 
lengthens the pendulum, and when lowered, shortens it. This circumstance 
is especially injurious not so much in particular experiments, where the 
length may be regulated each time, as in the application of the pendulum to 
clocks, where the slightest variation in its length must affect the rate. In 
this latter case the pendulum, to be an accurate regulator of motion, must 
' first regulate itself; and to this end, many combinations have been devised, 
of which Hlarrison’s compensation or gridiron pendulum, and Graham’s 
mercurial pendulum, will alone be mentioned here. 
The gridiron pendulum ( pl. 16, fig. 21) was invented in 1725 by Harri- 
son, for which, in connexion with his chronometer, he received a premium 
of £25,000 sterling from the British parliament. It consists of five steel and 
four brass rods, which alternate with each other, so that the central rod to 
which the disk of the pendulum is attached, is of steel. These brass and 
steel rods are so fixed in the heads, aa, bb, that while the expansion of the 
steel rods produces a tendency to elongation in the pendulum, that of the 
brass rods, which press upwards the head to which the pendulum rod is 
attached, produce a tendency to contraction. If, now, the lengths of these 
brass and steel rods are to each other in the proper proportion of their co- 
efficients of expansion, or as 61: 100, the expansion of one set will elevate 
the pendulum just as much as it is depressed by the other, and the actual 
length will be invariable. 
This pendulum, philosophical and beautiful as it is in theory, is diminished 
in practical value by the following considerations: 1. That it is difficult to 
make the rods sufficiently accurate ; 2. It is difficult to give them their pro- 
per proportional lengths; 3. That it is more exposed to the resistance of 
the atmosphere. Other metals may be employed instead of steel and brass. 
The mercurial pendulum ( pl. 16, fig. 22) invented by Graham in 1715, 
has a brass rod, aab, which carries below a cylindrical glass ike from 
13—14 inches long, and two inches in diameter. This vessel, , filled 
up to 12 inches with mercury, forms the ball of the pendulum, and qe the 
expansion of the rod should be too great for that of the ball, the quantity of 
mercury in the latter may be varied. By the influence of temperature, the 
rod is expanded; the mercury is expanded at the same time, however, and 
its centre of gravity is elevated : the pendulum is thus shortened again, and 
by trial a very accurate compensation may be obtained. ‘The single influ- 
ence operating against this pendulum is that the mercury sometimes begins 
to-expand before the rod; the variation, however, rarely amounts to more 
than one eighth of what takes place in good common pendulums. The disk, d, 
serves for the general regulation of the pendulum. 
After Galileo had developed the laws of the pendulum, Huyghens deter- 
mined the centre of oscillation of the material pendulum, and thereby made 
possible an accurate measurement of time, by applying the pendulum to the 
208 
