The apparatus being now arranged, the air is pumped out; and here, again, care must be taken not to shake the gold off the stages. When a proper vacuum has been obtained, which will be shown by the pump-gauge, the stop is withdrawn from one of the stages, and the gold and feather are seen to fall simultaneously to the air-pump plate. Another stage, with the gold-leaf and coin, may now be detached; both showing distinctly, that when the resistance of the air is withdrawn, all bodies, whether called _light_ or _heavy_, gravitate equally to the earth. Then, the screw at the bottom of the pump barrels [Page 16] being opened, attention may be directed to the whizzing noise the air makes on entering the vacuum, and when the air is once more restored to the long glass vessel, the last stage may be allowed to fall; and now, the gold coin reaches the pump-plate first, and the feather, lingering behind, loses (as it were) the race, and touches the plate after the gold coin; thus demonstrating clearly the resistance of the air to falling bodies. Another, and perhaps less troublesome, mode of showing the same fact, is to use a long glass tube closed at each end with brass caps cemented on. One cap should have the largest possible aperture closed by a brass screw, and the other may fit a small hand-pump. If a piece of gold and a small feather are placed in the tube, it may be shown that the former reaches the bottom of the tube first, whilst it is full of air, and when the air is withdrawn by means of the pump, and the tube again inverted, both the gold and the feather fall in the same time. [Illustration: Fig. 19. A B. Glass tube containing a piece of gold and a feather, which are placed in at the large aperture A. C. Small hand-pump.] For this reason, all attempts to measure heights or depths by observing the time occupied by a falling body in reaching the earth must be incorrect, and can only be rough approximations. An experiment tried at St. Paul's Cathedral, with a stone, which was allowed to fall from the cupola, indicated the time occupied in the descent to be four and a half seconds: now, if we square this time, and multiply by 16, a height of 324 feet is denoted; whereas the actual height is only 272 feet, and the difference of 52 feet shows how the stone was retarded in its passage through the air; for, had there been no obstacle, it would have reached the ground in 4-3/20ths seconds. [Illustration: Fig. 20.]