Three or four years ago, wishing to make some experiments on time-keepers, and desiring for that purpose to have a compensation pendulum fitted to my clock, I turned my attention to pendulums composed of two pieces, having different rates of expansion.

Such compensation pendulums are now pretty common; the idea which has led to their construction being this,—that if the pendulum rod be made of material having a very small rate of expansion, and if the bob or weight be of a substance much more expansive, then the upward expansion of the one may be made to compensate for the downward expansion of the other, and so the going of the clock may remain unaffected by changes of temperature.