As the skill and knowledge of the experimenter augment he is often able to simplify his experimental combinations. Thus, in the present instance, by the suitable arrangement of the source of sound and the sensitive flame, it was found that not only twenty-five layers, but three or four layers of coal-gas and carbonic acid sufficed to still the agitated flame. Nay, with improved manipulation, the action of a single layer of either gas was rendered perfectly sensible. So also as regards heated layers of air, not only were sixty-six or twenty-two heated platinum wires found sufficient, but the heated air from two or three candle-flames, or even from a single flame, or a heated poker, was found perfectly competent to stop the flame’s agitation. The same remark applies to vapors. Three or four heated layers of air, saturated with the vapor of a volatile liquid, stilled the flame; and, by improved manipulation, the action of a single saturated layer could be rendered sensible. In all these cases, moreover, a small, high-pitched reed might be substituted for the bell. My assistant has devised the simple apparatus sketched in Fig. 147, for showing reflection by gases, vapors, and heated air. At the end A of the square pipe A B is a small vibrating reed of high pitch, the sound of which violently agitates the sensitive flame _f_. To the horizontal tube _g g′_ are attached four small burners, and above them four chimneys, through which the heated gases from the flames can ascend into A B. When the coverings of the chimneys are removed and the gas is ignited, the air within A B is rendered rapidly non-homogeneous, and immediately stills the agitated flame. [Illustration: FIG. 147.] The pipe A B may be turned upside down, an orifice seen between A and B fitting on to the stand which supports the tube. The conduit _t_ leads into a shallow rectangular box, which communicates by a series of transverse apertures with A B. When air, saturated with the vapor of a volatile liquid, is forced through these apertures, the atmosphere in A B is immediately rendered heterogeneous, the agitated flame being as rapidly stilled. [Illustration: FIG. 148.] In the experiments at the South Foreland, not only was it proved that the acoustic clouds stopped the sound; but, in the proper position, the sounds which had been refused transmission were received by reflection. I wished very much to render this echoed sound evident experimentally; and stated to my assistant that we ought to be able to accomplish this. Mr. Cottrell met my desire by the following beautiful experiment, which has been thus described before the Royal Society: A vibrating reed B (Fig. 148) was placed so as to send sound-waves through a tin tube, 38 inches long, and 1-3/4 inch diameter, in the direction B A, the action of the sound being rendered manifest by its causing a sensitive flame placed at F′ to become violently agitated.