This is just a detail that I found interesting — early regulator clocks tended to slow down as their pendulums lengthened in warm conditions. One solution, offered by George Graham in 1721, was to attach two vials of mercury to the pendulum — as the pendulum warmed and expanded, so did the mercury, creeping upward in its vials and, at least in theory, preserving the pendulum’s center of mass.

One difficulty is that the mercury tends to warm more slowly than the pendulum itself, but the system worked well enough to persist into the 20th century.