cated by the water-gauge, the flame flares, becoming much shorter and broader, like a little Indian club (Fig. 4, y), and producing a low roaring sound, due to the escape of unburned gas. Let the pressure now be diminished until this flaring barely ceases. The 
Fig. 4. flame is now in its most sensitive condition. Sounds of low or even medium pitch have no effect upon it; but on blowing a shrill whistle, or rattling a bunch of keys anywhere within thirty or forty feet, it flares. Perhaps the most beautiful illustration of its sensitiveness is given by placing an open watch near the nozzle but not touching it; every tick causes a momentary sinking and spreading of the flame, so that the effect may be seen across an audience-room. If the audience applauds with clapping of hands, the flame shrinks in acknowledgment.
A very sensitive flame, but not so convenient as that of Prof. Barrett, and not visible at so great a distance, may be obtained with no pressure greater than that of the street mains, by causing the gas to issue from a small tube, over the orifice of which, at a height of an inch or two, is placed a piece of wire gauze (Fig. 5). The mixture of coal-gas and air is ignited above the gauze, and a glass tube may be used to protect the flame from currents of air, though this is not usually necessary. Very little adjustment is needed to find the distance between nozzle and gauze at which the flame is most sensitive. This arrangement was devised independently by Prof. Govi, of Turin, and Mr. Barry, of Ireland. The flame is deficient in brightness, and is only a few inches high at its best, but has the advantage of not requiring any appliances that may not be easily supplied in any town. If Barrett's flame is available, however, it is decidedly preferable to anything else.
With such a flame as Barrett's it becomes possible to explore the air and detect regions of relative noise and silence just as a delicate thermometer enables us to determine variations of temperature in different layers of air or water. If the pitch be too high for the ear to estimate or even detect it, the sensitive flame is more delicate than the ear. Armed with a whistle yielding a pitch of twelve or fifteen thousand vibrations per second, and with a good flame, many beautiful analogies between sound and light may be exhibited with entire satisfaction to an audience of deaf-mutes, if the lecturer's fingers are fairly nimble, since there is no necessity for the sounds to be heard. Most of the experiments about to be described were devised by Lord Rayleigh, the suc-
