In the case of most transparent substances the æther waves which constitute light are transmitted with different velocities, the longer waves moving faster than the shorter ones. Hence we have the familiar result of the decomposition of a ray of white light into its different constituents by a glass prism. We cannot, however, perform a similar experiment on a complex series of waves of sound by means of a carbonic acid prism. In other words, a sound-prism refracts, but does not disperse sound waves of various wave-lengths. One thing, however, should be pointed out before dismissing this experiment, and that is that to show successfully the experiment with the prism, the length of the sound waves used must be small compared with the dimensions of the prism. The reason for this is that otherwise there would be too much bending of the waves round the obstacle. When a train of waves, no matter whether waves in air or waves in water, meets with an impervious body, there is always a certain bending of the waves round it, which is technically called _diffraction_. We may see this effect on a large scale when sea waves, rolling in, pass by some large rock standing up like an island out of the water. The waves meet it, pass round it, and, so to speak, embrace it and continue on the other side. If there is to be any calm water on the leeward side, the island must be large compared with the length of the waves. The same thing holds good with regard to air waves. In order that an object may form an acoustic or sound-shadow, it is necessary that the construction shall be large compared with the length of the wave. Thus the hand held in front of the mouth does not much obstruct the waves of the speaking voice, because these waves are about 2 to 4 feet long. But as you have seen when using sound waves only 1 inch long, the hand will form a very well-marked sound-shadow, as shown by its effect when held between a whistle and a sensitive flame. In order to complete our proof that the agency which affects our ears as sound is really due to air waves, it is necessary to be able to show that we can produce _interference_ with air waves, as in the case of waves on water. The nature of the effect called interference by which one wave is made to annihilate another has been already fully explained. I will now endeavour to exhibit to you the interference of two sound-wave trains in an experiment due to Lord Rayleigh, the apparatus for which he has kindly lent to me. It consists, as you see, of a stand, to which is fixed a jet, from which we form a tall sensitive flame. Behind the flame is placed a sheet of glass, which is held vertically, but can be slid towards or from the flame. At a little distance we place a _bird-call_, or sort of whistle, which produces, when blown with air, a note so shrill as to be inaudible to human ears.