ON INSTRUMENTAL AIDS FOR DEAFNESS 257 



Lord Rayleigh has also investigated this same problem by 

 two entirely different methods ; one depending on the blowing 

 of a whistle by a stream of air working at known power, the 

 observer going so far away that the whistle is only just audible, 

 and the second depending on the estimate of the rate of output 

 of sound energ}' by a tuning-fork at the instant at which it 

 ceases to be audible though still vibrating. The former method 

 gave o - 8 x io -7 cm.,and the latter 1*2 x io -7 cm. as the minimum 

 amplitude of vibration of the air for audibility. The figures 

 of Rayleigh are several times larger than those of Shaw, but 

 taking the higher mean figure, say io~ 7 of Rayleigh, this corre- 

 sponds to an energy density of only 0*000043 er g s P er sec - P er 

 sq. cm., or 1*3 x io -9 ergs per c.cm. If we might neglect the 

 existence of bone conduction of sound through the skull, and 

 if we take the opening of the external meatus of the ear as 

 about 1 sq. cm., then provided vibrational energy reach the 

 observer at the rate of 0*000043 er g s P er sec, or say 6 x io -15 

 of a horse power, the sound will be audible to a person of normal 

 acuity of hearing. For sounds used in ordinary conversation 

 this figure would have to be increased some ten times. 



It is, however, a question not merely of collecting enough 

 vibrational energy and conveying this to the ear of the person 

 with defective hearing, but it is just as essential that the 

 quality or timbre of the sound shall not be so far altered as to 

 render the sounds unintelligible. Now in ordinary speech the 

 consonants merely serve as momentary interruptions between 

 vowel sounds, the interruption being produced at lips, teeth, 

 or back of hard palate, the timbre of the sound is that given 

 by the succession of vowel sounds. 



Much research has been expended in the study of the con- 

 ditions for vowel-production, pioneer work being done by 

 Willis and by Wheatstone. It appears from the present state 

 of our knowledge that, to produce a vowel sound of given 

 character, it is necessary that the sound shall contain from 

 six to eight separate partial tones, each of definite pitch, irrespec- 

 tive of the fundamental tone to which the vibration of the vocal 

 chords give rise. These partial tones are widely separated in 

 pitch, and do not form a harmonic series. If, then, in any piece 

 of apparatus designed to act by the principle of resonance it 

 is possible for these partial tones and no others to be reinforced, 

 there is reason to hope that such a piece of apparatus may 



