274 Mr. W. Duddell [March 16, 



WEEKLY EVENING MEETING, 



Friday, March 16, 1906. 



His Grace the Duke of Northumbeeland, K.G. D.C.L. F.R.S., 

 President, in the Chair. 



W. Duddell, Esq. 



Hoiv to Im'prove Telephony. 



In my discourse to-night I propose to strictly limit the word 

 " telephony " to the art by means of which sounds and speech are 

 electrically transmitted to a distance. The loudness, and articulation 

 or clearness of the transmitted speech, and the distance over which it 

 can be transmitted, are the main directions in which improvement is 

 required. The questions, how Mr. A. in London shall get connected 

 to Mr. B. in Glasgow with the minimum of loss of time and temper, 

 and how to find the number of the person you wish to communicate 

 with, when his entry in the directory, in small type, is sandwiched in 

 between two advertisements, are questions with which I do not pro- 

 pose to deal — not that there is any lack of room for improvement in 

 these directions. 



Before proceeding with our subject, it will be necessary to consider 

 briefly what constitutes sound, and more especially articulate speech, 

 so as to form a clear idea of what we want to transmit. The sensa- 

 tion of sound, as is well known, is produced by the vibration back- 

 wards and forwards of the particles of the air about their position of 

 rest, and tlie character of the sound depends on the quickness and the 

 form of the vibrations. Thus, in the case of a musical note, the air 

 particles vibrate in a perfectly regular manner, and the number of 

 complete vibrations in a second, or the frequency determines the 

 pitch ; and the amplitude, or distance the air particle moves from its 

 position of rest, determines the loudness of the note. In speech, 

 however, the vibrations are very complex, and in order to form any 

 clear mental idea of their character, it is necessary to represent tlie 

 movements as curves, which I will call sound patterns. Various ob- 

 servers have made records of these patterns ; among the earliest and 

 best, are those ol)taiiied l)y Fleeming Jenkin, and Ewing (Edinburgh 

 Phil. Trans.), who magnified the impression obtained on the cylinder 

 of a tin-foil phonograph. Figs. 1 to 6 are some typical sound pat- 

 terns obtained by another method, to be explained later, which iUus- 



