Fig. 71 represents the tracing which corresponds to the letter _a_ when pronounced as in _bat_, in the three systems of registration. That corresponding to line A is an enlarged reproduction of the tracing left on the tinfoil; that corresponding to line B represents its profile on the sheet of smoked glass. Finally, line C shows the outline of Koenig’s singing flames, when the same sound is produced quite close to the membrane of the register. It must be quite close, since the form of the tracing produced by a pointer attached to a vibrating membrane, when influenced by composite sounds, depends on the distance intervening between the membrane and the source of sound, and an infinite variety in the form of the tracing may be obtained by modifying the distance. In fact, when this distance is increased, the waves of sound which result from composite sounds react on the membrane at different moments of their emission. For example, if the composite sound is formed of six harmonics, the displacement of the source of vibration from the first harmonic by ¼ the length of a wave will respectively remove the second, third, fourth, fifth, and sixth harmonics ½, ¾, 1, 1¼, 1½ the length of a wave, and consequently the outline resulting from the combination of waves will no longer be the same as it was before the displacement of the source of sound, although the perception of the sounds remains the same in both cases. This principle is clearly demonstrated by Koenig’s instrument, by lengthening and shortening an extensible tube, inserted between the resonator and the vibrating membrane, which is placed close to the flame; and this explains the disagreement of physicists as to the composition of vocal sounds which they have analysed by means of the singing flames. Mr. Mayer adds that these facts further show that we cannot hope to read the impressions and tracings of the phonograph, which not only vary with the nature of the voice, but also with the different moments at which the harmonics of the voice are emitted, and with the relative differences in the intensities of these harmonics. Notwithstanding this assertion, we reproduce (fig. 72) an extremely curious tracing sent to us by Mr. Blake, which represents the vibrations produced by the words ‘Brown, University: how do you do?’ They were photographed by means of an index fastened to a vibrating disk on which a ray of light was thrown. The word ‘how’ is particularly remarkable for the combined forms of the inflections of the vibrations. Recent experiments seem to show that the more the vibrating membrane of the phonograph resembles the human ear in its construction, the better it repeats and registers the sound vibrations: it should be stretched, as far as possible, in the same way as the tympanum is stretched by the hammer of the ear, and moreover it should have the same form, since the vibrations of air are in this case much more effective. [Illustration: FIG. 72.]