The air-vibrations so generated are at the rate of 33,000 per second, which is beyond the limit of audition. Hence, even when blown strongly, you hear no sound from this appliance. It produces, however, as you can see, a very violent effect upon the sensitive flame. Hence this flame _hears_ a note which we cannot hear, and it suggests that perhaps some animals or insects may have a range of hearing quite beyond the limits fixed for our human ears. Such being the case, you will see that if the glass plate is placed behind the flame at a certain distance, the flame at once stops flaring and becomes quiescent. If, however, the plate is moved to or from the flame by a very small distance equal to about the one-twelfth part of an inch, the tall flame at once drops in height and begins to flare. If we move the plate steadily backwards by equal small distances, we find the flame alternately quiescent and waving. The explanation of this effect is that it is due to the interference between the direct and reflected sound-rays. The waves of air are turned back when they meet the glass in such a manner that the crests of the arriving waves are made to coincide with the hollows of the reflected waves, or, to speak more correctly, the zones of condensation of one are coincident with the places of rarefaction of the other. When the glass is adjusted so that this happens, all air-wave motion just in front of it is destroyed, and hence the sensitive detecting flame remains quiescent. If, however, the glass is moved nearer to or further from the flame, then the condensations of the reflected wave may be made to fall in the same places as the condensations of the arriving wave, and in that case the disturbance is doubled, and not destroyed. [Illustration: FIG. 52.] A little model may be made which will help the reader to grasp this point. Cut out a piece of paper in the form shown in Fig. 52 to represent a wave. Bend back the paper on itself at the dotted line _ab_, and let one half represent the arriving wave, and the other the reflecting wave. It will be seen that in this case the crests of the incoming wave are obliterated by the hollows of the returning wave. If, however, the paper is bent back at _cd_, then the crests of the reflected and incident waves conspire, and there is no interference. Whenever we can produce _interference_ in this manner between two sets of sound-rays, or light-rays, or rays of any other kind, we have the strongest possible proof that we are concerned with a _wave-motion_; because in no other way that we can understand is it possible that a destruction of sound by sound can take place by, so to speak, superimposing two sound-rays, or a destruction of light by bringing together two rays of light.