A few experiments of this kind with the flame in various positions are sufficient to show that the sound-beam is reflected by the glass in accordance with the law of reflection of wave-motion, viz. that the angle of incidence is equal to the angle of reflection. We can in the same way reflect the sound-beam by a wooden board, a piece of cardboard, a looking-glass, or a sheet of metal. We can reflect it from a wet duster, but not very well from a dry handkerchief. If we place the flame in the direct beam, it is easy to show that all the above good reflectors of sound are opaque to a sound-ray, and cast an acoustic shadow. In fact, I can prevent the flame from roaring by merely interposing my hand in front of it. A wet duster is found to be opaque to these sound waves, but a dry linen handkerchief is fairly transparent. The collodion film used in making the lens and prism is also exceedingly transparent to these short air waves. We may then go one step further, and show that these air waves are capable of refraction. It will be in your remembrance that, in speaking of water ripples, it was shown by experiment that, when water ripples passed over a boundary between two regions, in one of which they travelled more quickly than in the other, a bending of the direction of ripple-motion took place. We can show precisely the same thing with these air waves. The collodion prism has been filled with a heavy gas called carbonic acid. This gas is about half as heavy again as air, and it is this heavy and poisonous gas which, by accumulating in old wells or brewers’ vats or in coal-mines after an explosion, causes the death of any man or living animal immersed in it. It has already been explained that the velocity of sound waves in different gases varies inversely as the square root of their density. Hence the speed of a sound wave in carbonic acid gas will be less than that in air in the ratio of the square roots of the densities of these gases. The density of carbonic acid gas is to that of air as 1·552 is to 1. The square root of 1·552 is 1·246, or nearly 1¹⁄₄. Accordingly, the speed of a sound wave in carbonic acid gas is to the speed in air as 4 is to 5. A sound wave in air will therefore travel 5 feet or 5 inches in the same time that it travels 4 feet or 4 inches in carbonic acid gas. Let us now consider what must happen if a sound wave falls obliquely upon the face of our carbonic acid prism. [Illustration: FIG. 50.—The refraction of a wave by a prism.]