And the worst feature about the movement is, that as the teeth and pallets move, the leverage of the teeth on the pallets alters, and thus the bobs on the verge are under the influence not of a uniform or duly regulated force, but of a constantly varying one, and one that varies in a very complicated and erratic way. It would be hopeless to expect much time-keeping from such a contrivance. The most that could be expected would be by putting on a very big weight to reduce to comparative insignificance the friction, and then hope that the swings would be uniform, so that whatever went on in one swing would go on in the next, and thus the time-keeping be regular. But any course tending to diminish the driving force, such as the thickening of the oil, would greatly affect the going. It was for this reason that Huygens turned the verge into a pendulum by removing one of the bobs, and letting gravity thus act on the other. For watches, however, a different plan was contrived. One end of a slender spiral spring was affixed to the verge. The other end of the spring was made fast to the clock frame. The verge was now, therefore, chiefly under the action of the acceleration of the spring. To make the acceleration of the teeth of the ’scape wheel less embarrassing, the teeth were so shaped as only to give a short push at stated intervals, and not interfere with the free swing of the verge under the alternate to-and-fro accelerations and retardations of the spring. By this means the verge became in every way an excellent pendulum, not dependent on gravity, and permitting the watch to be held in any position. The verge thus fitted was turned into a wheel, and became a “balance wheel.” It was compensated for heat expansion by a cunning use of the unequal expansion of brass and steel, in a manner analogous to the way this unequal expansion of metals had been employed to compensate the pendulum, and became the beautiful and accurate time-measurer that we see to-day, with its pivots mounted in jewels to diminish friction, and with screws round the rims of the balance wheel to enable the centre of gravity to be exactly adjusted to its centre of rotation, and with a delicate hair-spring of tempered steel that is a marvel of microscopic work. But the escapement of the early watches left much to be desired. In order to make it clear how imperfect that early escapement was, we have to turn back and remember what has been said about the dead beat escapement. It will then be remembered that it was shown that for small arcs the pendulum would keep good time provided you let it have as much swing as it wanted to use up the force which the escapement had applied to it, _but not otherwise_, so the pendulums only acted really well when the impulse was given about the middle of the swing, and they were free to go on and stop when they pleased, and turn back at the end of it.