balance was effected in the exhausted receiver, he determined the amount of the preponderance. In another experiment he compressed air in a bladder by tying a pack-thread round it, balanced it from one arm of his balance in the open air; then, pricking the bladder so as to relieve the pressure, he found that with the escape of the compressed air the weight diminished. One of the most important of his experiments with the air-pump was the following. He placed within the receiver the cistern of a mercurial barometer, the tube of which was made to pass through the central hole in the brass plate, from which the stopper had been removed. The space around the tube was filled up with cement, and the receiver exhausted. At each stroke of the pump the mercury in the barometer tube descended, but through successively diminishing distances, until at length it stood only an inch above the mercury in the cistern. The experiment was then repeated with a tube four feet long and filled with water. This constituted the nineteenth experiment referred to later on. A great many strokes of the pump had to be made before the water began to descend. At length it fell till the surface in the tube stood only about a foot above that in the tank. Placing vessels of ordinary spring-water and of distilled rain-water in the receiver, he found that, after the exhaustion had reached a certain stage, bubbles of gas were copiously evolved from the spring-water, but not from the distilled water. On another occasion he caused warm water to boil by a few strokes of the pump; and, continuing the exhaustion, the water was made to boil at intervals until it became only lukewarm. The experiment was repeated with several volatile liquids. He also noticed the cloud formed in the receiver when the air was allowed rapidly to expand; but the mechanical theory of heat had not then made sufficient progress to enable him to account for the condensation by the loss of heat due to the work done by the expanding air. The very minute accuracy of his observations is conspicuous in the descriptions of most of his experiments. That the air is the usual medium for the conveyance of sound was shown by suspending a watch by a linen thread within the receiver. On exhausting the air, the ticking of the watch ceased to be heard. A pretty experiment consisted in placing a bottle of a certain fuming liquid within the receiver; on exhausting the air, the fumes fell over the neck of the bottle and poured over the stand on which it was placed like a stream of water. Another experiment, the thirty-second, is worthy of mention on account of the use to which it was afterwards applied in the controversy respecting the cause of suction. The receiver, having been exhausted, was removed from the cylinder, the stop-cock being turned off, and a small brass valve, to which a scale-pan was attached, was placed just under the aperture of the tube below the stop-cock. On turning the latter, the stream of air