December 26, 1901 J 



NA TURE 



18 



the vowels accordintj to the number of harmonics which they 

 contained, in the following table :— 



In sounding a the mouth is widely opened and the fundamental 

 and eight harmonics are produced ; in the third series, on 

 the contrary, there is only one harmonic sounded, which is 

 more and more acute as we pronounce the vowels in the order ou, 

 II and /. The vowels of the second series, o, cii and e, are 

 transitional between the first and the third. Thus we pass from 

 a 10 ott by 0, from a to u by eii, and from a to i by e. 



Donders' showed that the cavity of the mouth, as arranged 

 for the giving forth of a vowel, was tuned as a resonator for a 

 tone of a certain pitch, and that different pitches coresponded 





Fig. 16. — Vibrating flame of cyanogen photographed by Gerhardt. 



to the forms of the cavity for the different vowels. This he dis- 

 covered by the peculiar noise produced in the mouth when the 

 different vowels are whispered. The cavity of the mouth is 

 then blown like an organ-pipe and by its resonance reinforces 

 the corresponding partials in the rushing wind-like noise. 

 Tlien the question was taken up by Helmholtz.- He attacked 

 it both by analysis and by synthesis. He analysed the vowel- 

 tones by his well-known resonators, aided by his own singularly 

 acute ear, anc^ he attempted to combine, by means of tuning- 

 forks, the tones which he thought existed in a vowel, so as to 

 reproduce the sound of the vowel. In the latter part of th; 

 investigation he was by no means successful. These investiga- 

 tions led Helmholtz to put forward in succession two theories as 

 to the formation of vowels. The first was that, as in all 

 musical instruments, the quality or timbre of the vowel depends 

 on the fundamental tone, reinforced by certain partials or over- 



soon after its invention. Donders sang the vowel tones to the 

 instrument, and then asked the operator to vary the speed of 

 the cylinder during reproduction. Then the vowel a became o, 

 and e became on. Thus while the phonograph reproduces in a 

 wonderful way the tones of musical instruments without change 

 of quality, it cannot transpose vowel-tones without altering 

 their character. This special ch.aracter or quality cannot, then,. , 

 depend on the overtones reinforced by the oral cavities being 

 simple multiples of the fundamental tone, and Helmholtz's first 

 theory had to be abandoned. 



This led Helmholtz to advance a second theory as follows : — 

 Each vowel is characterised by a certain harmonic or partial 

 tone, of constant pitch, whatever may be the pitch of the note 

 on which the vowel is sung or spoken. Attempts were then 

 made, notably by Helmholtz and Konig, (o-fix the pilch of the 

 characteristic partial tone or vocable, and there appeared to be 



Fig. iS.— Blake's mirror 



considerable differences in the results of the two distinguished 

 observers, differences amounting to as much, in some cases, as 

 three semi-tones. 



The next step was, as has already been explained, to tran- 

 scribe the marks on the wax cylinder of the phonograph, made 

 on singing or speaking a vowel, into sinuous curves and to 

 subject these to harmonic analysis. It is not difficult, in com- 

 paratively simple cases, to obtain a curve which is the algebraic 

 sum of the ordinates of several sinusroidal curves, but it is. not: 

 so easy to do the reverse operation, namely, to analyse the curves. 

 Fleeming Jenkin and Ewing, afterwards Schneebeli, Hensen, 

 Pipping and Hermann, have done this in accordance with the 

 theorem of Fourier and the law of Ohm. In particular, Her- 

 mann, by a beautiful and ingenious method, has analysed the 

 curves obtained by his photographic device, and has modified 

 the theory of Helmholtz. His statement is that the oral cavity 



tones, of which a number are produced by the vocal cords 

 along with the fundamental lone, the reinforcement depending 

 on the resonance of the cavil ies above the vocal cords. This 

 theory was upset by the use of the phonograph. If a vowel is 

 sung to the phonograph while the cylinder is travelling at a 

 certain speed, the vowel-tone will be reproduced with exactly 

 the same quality if the cylinder is driven at the same speed, but 

 if it is driven faster, then the quality of the vowel will be 

 changed, so much so as to be scarcely recognisable. M. 

 Marey narrates that Donders and he first made this observa- 

 tion when it so happened that the two savants were 

 present in Paris at a public demonstration o( the phonograph 



' Donders: " De physiologie der Spraakklanken " (1870). 

 ■-' Helmholtz: " Ueber de Vokale." .-frMm. f. d. Holland. Beiti 

 Heilk. (Utrecht. 1S57). See ..iher references given in Sch.Hfer-: 

 ook," vol. ii. p. 1217 (footnote). 



, Nat. 



NO. 1678, VOX. 65] 



produces independently a harmonic or partial tone which has no 

 definite relation to the fundamental tone emitted by the larynx. 

 A vowel, according to him, is a special acoustic phenomenon, 

 depending on the intermittent production of a special partial, 

 or " formant," or "characteristique. " The pitch of the " for- 

 mant " may vary a little without altering the character of the 

 vowel. For a, for example, the " formant " may vary from fa^ 

 to la^, even in the same person. He has also attempted, but 

 not with complete success, to reproduce the vowel-tones by 

 synthesis. 



There are thus three theories : (l) the first of Helmholtz, now 

 abandoned, that the pitch of the partials is represented by 

 simple multiples of the vibration periods of the fundamental ; 

 (2) the second of Helmholtz, that the pitch of the characteristic 

 partial is always fixed, but has a definite relation to the pitch 

 of the fundamental ;_and (3) that of Hermann, that the pitch 



