(See Encycl.). Besides the effects of heat and cold on the length of the pendulum rod, and of course on its isochronism, it may certainly be worth while, in the construction of clocks intended to measure time with the utmost possible exactness, to take into consideration the resistance of the air, which, by its unequal density, varying the weight of the pendulum, must in a small degree accelerate or retard its motion. The celebrated David Rittenhouse, who paid particular attention to this subject, estimates the extreme difference of velocity, arising from this cause, at half a second a day; and he observes, that a remedy dependent on the barometer will not be strictly accurate, as the weight of the entire column of air does not precisely correspond with the density of its base. He proposes, therefore, as a very simple and easy remedy, that the pendulum shall, as usual, consist of an inflexible rod carrying the ball beneath, and continued above the centre of suspension to an equal (or an unequal) distance upwards. At this extremity is to be fixed another ball of the same dimensions (or greater or less, according as the continuation is shorter or longer), but made as light as possible. The oscillations of this upper ball will be accelerated by its buoyancy by the same quantity as those of the lower would be retarded; and thus, by a proper adjustment, the two effects might be made to balance and correct each other.

Our author made a compound pendulum on these principles, of about one foot in its whole length. This pendulum, on many trials, made in the air 57 vibrations in a minute. On immersing the whole in water, it made 59 vibrations in the same time; shewing evidently, that its returns were quicker in so dense a medium as water than in the air. (This is contrary to what takes place with the common pendulum). When the lower bob or pendulum only was plunged in water, it made no more than 44 vibrations in a minute.