Accordingly, we see that the lighter the gas the faster sound travels in it, pressure and temperature being the same. If the atmosphere we breathe consisted of hydrogen instead of a mixture of oxygen, nitrogen, and many other gases, a clap of thunder would follow a flash of lightning much more quickly than it does in our present air, supposing the storms to be at the same distance. Under present circumstances, if 20 seconds elapse between the flash and the peal, it indicates that the storm is about 4 miles away, but if the atmosphere were of hydrogen, for a storm at the same distance the thunder would follow the lightning in about 5 seconds. Furnished with these facts about the propagation of air waves, it is now possible to point out some interesting consequences. It will be in your recollection that in the first chapter it was pointed out that a wave on water could be reflected by a hard surface, and that it could be refracted, or bent, when it passed from a region where it was moving quickly to one where it was moving more slowly. It will be necessary now to prove experimentally that the same things can be done with sound, in order that a body of proof may be built up in your minds convincing you that the external cause of sound-sensation must be a wave-motion in the air. In the first place, I must describe to you, somewhat in detail, the nature of the arrangements we shall employ for producing and detecting the sound waves which will be used in these experiments. It would not do to rely upon the ear as a detector because you cannot all be so placed as to hear the sounds which will be produced, and we shall, therefore, employ a peculiar kind of flame, called a _sensitive flame_, to act as a detector.