ON EDISON'S TALKING-MACHINE. 723 



the distance of the source of the sound from the membrane, and the 

 same compound sound will form an infinite number of different traces 

 as we gradually increase the distance of its place of origin from the 

 membrane ; for, as you increase this distance, the waves of the compo- 

 nents of the compound sound are made to strike on the membrane at 

 different periods of their swings. 



For example, if the compound sound is formed of six harmonics, 

 the removal of the source of the sonorous vibrations, from the mem- 

 brane to a distance equal to of a wave-length of the 1st harmonic, 

 will remove the 2d, 3d, 4th, 5th, and 6th harmonics to distances from the 

 membrane equal respectively to , f, 1, 1, and 1^ wave-lengths. The 

 consequence evidently is, that the resultant wave-form is entirely 

 changed by this motion of the source of the sound, though the sono- 

 rous sensation of the compound sound remains unchanged. 



The above facts are readily proved experimentally by sending a 

 constant compound sound into the cone of Konig's apparatus, while 

 we gradually lengthen the tube between the cone and the membrane 

 next to the flame. This is best done by the intervention of one tube 

 sliding in another, like a trombone. These experiments I have recent- 

 ly made with entire success, and they explain the discussions which 

 have arisen between different observers as to the composition of vocal 

 and other composite sounds, as analyzed by means of Konig's vibrat- 

 ing flames. 



These facts also show how futile it is for any one to hope to be 

 able to read the impressions and traces of phonographs, for these 

 traces will vary, not alone with the quality of the voices, but also with 

 the differently-related times of starting of the harmonics of these 

 voices, and with the different relative intensities of these harmonics. 



It is necessary to give to the cylinder a very regular motion of 

 rotation while it receives and reproduces the vibrations made in sing- 

 ing ; for even slight irregularities in the velocity of the cylinder 

 destroy the accuracy of the musical intervals, and cause the phono- 

 graph to sing falsetto. Even the reproducing of speech is greatly 

 improved by rotating the cylinder by mechanism which gives it uni- 

 formity of motion. If you make the machine talk by giving it a more 

 rapid rotation than it had when you spoke to it, the pitch of its voice 

 is raised ; and by varying the velocity of the cylinder the machine 

 may be made to speak the same sentence in a very bass voice, or in a 

 voice of a pitch so high that its sounds are really elfish and entirely 

 unnatural. 



Recent experiments seem to show that the nearer the diaphragm 

 A approaches to the construction of the drum-skin of the human ear 

 by " damping " it, as the hammer-bone does the latter, the better 

 does it record and repeat the sonorous vibrations; for the motion of a 

 membrane thus damped is ruled alone by the aerial vibrations falling 

 on it. 



