August 2 2, 1878] 



NATURE 



455 



same. There remains the fact that speakers and hearers were 

 iincoiifcious of any generic change in the vowel il when the 

 pitch of the strongly-reinforced partial changed by a whole 

 octave. 



On the other hand, it is equally clear the voice, in singing a 

 given vowel at various pitches, does not simply produce a 

 certain constant group of relative partial tones. Possibly, 

 indeed, the ear might recognise a single tone, especially if very 

 feebly accompanied by higher harmonics, as a kind of il outside 

 the region within which the human voice forms il in that way. 

 Thus Helmholtz, in his Tonempfindungen, says that the single 

 tone B(j, when sounded alone, gave a very dull il, much duller 

 than could be produced by the voice. This tone is an octave 

 below the plr.ce ^^•here voice I ceased to make il by rein- 

 forcing the prime. Quite similarly it is conceivable that the 

 gi-oup consisting of a prime and its octave might be recognised 

 as o eve:? when produced below the limits within which the 

 human voice does produce this simple harmony in singing o. Our 

 own impression as to the result of running the phonograph slower 

 when it is speaking than when it is spoken to, supports this view, 

 but we do not desire to base any inference on that. Certainly 

 the low produced in this manner is not the human 0. 



Moreover, we find a very decided resemblance in the relative 

 constituents of J at a low pitch and cC or a at higher pitches. 5 

 in the neighbourhood of Bp, and a in the neighbourhood of / 

 and g, are pretty similarly constituted. Our experiments on a° 

 and a are not sufficiently extended to allow any very general 

 conclusions to be drawn, but they are sufficient to show that 

 between certain vowels the main distinction must lie in the abso- 

 solute pitch of the reinforced group of partial tones. 



We are thus brought back to our original statement that in 

 distinguishing vo\\els the ear is aided by two factors, one 

 depending on the harmony or group of partials, and the other on 

 the absolute pitch of the constituents. It seems not a little sin- 

 gular that the ear should attribute a distinct unity to sounds so 

 dissimilar in their relative and absolute composition as those 

 represented by the curves of Fig. i . 



We are forced to the conclusion already adopted by Helmholtz 

 and Bonders that the ear recognises the kind of cavity by which 

 the reinforcement is produced ; that although the sounds which 

 issue differ so much that we fail when they are graphically repre- 

 sented and mathematically analysed to grasp any one prominent 

 common feature, nevertheless by long practice the ear is able to 

 distinguish between the different sorts of cavities which are 

 formed in pronouncing given vowels. Something of the same 

 kind may be observed with other sources of sound than the 

 human voice ; the resonating cavities of various musical instru- 

 ments aid greatly in allowing each particular species to be 

 recognised at once, though their effect must be widely different at 

 different parts of the scale. It i-, moreover, no mere inference 

 that we recognise the cavity. Prof. Crum Brown's gutta percha 

 bottle, described before, proves that we do, and that it is a 

 group of tones reinforced by a particular kind of cavity that we 

 call a particular vowel. But we have to consider what light the 

 experiments throw on the kinds of cavity that are required for 

 certain vowels. The cavities are clearly distinguished in virtue 

 of two distinct properties : first, the pitch of their maximum 

 resonance or strongest proper tone, and second the range of 

 reinforcement which they are capable of producing. This latter 

 property has, we believe, been hitherto much neglected. 



Prof. Crum Brown's bottle proves that a constant cavity is 

 capable of produchig the constant vowel over a large range of 

 pitch. On the other hand our experiments with various human 

 voices singing o appear to exhibit a tun'ng of the cavity by 

 which new partials are sometimes introduced somewhat abruptly ; 

 and for the sound u it seems certain that the cavity is tuned, that 

 is to say, that the pitch of its proper tone is not the same when 

 the vowel is sung on different pitches. The appreciably strong 

 fourth partial in all the duplex u!% of Table VI. may here be 

 noticed as favouring the view that in each of these examples the oral 

 cavity had been adjusted so as to be in unison with the second 

 partial. We may describe the il cavity as an adjustable cavity 

 with a very limited range of resonance, whose effect is 

 to reinforce strongly only one partial lying above a. It is 

 possible that this cavity may keep itself constant throughout the 

 very limited range of pitch employed in ordinary speech, but 

 when the range is increased as in singing, a certain tuning seems 

 indispensable. 



If we assume that the J cavity is absolutely constant, we must 

 dsscribe it as a cavity capable of reinforcing more or less 



strongly tones lying anywhere belv.cen g and /". This cavity 



y 



■ aAA/V/V\AA/V\AA 



'' -VAnA-WWVWWWV. 



•' WvvA/wWwAa/vaAaA 



'% kAaAaAaMMJVxPAJ 



" a/WWWWa/WW 



■9 \y\/\/\J\/\Aj\/\f\J\r 

 ^ XAA/VV\AsAA/VV 



- AAxA/VVVWWV 



Fig. i.^ — Wave-forms of Sung by the same Voice at Various 

 Pitches, 



' We reproduce this week the figure which acsompanied Profs. Jenkin 

 and Ewing's first paper, on "The Wave-Form of O," p. 342. It shows the 

 delicate forms of the curves with greater ezactnesc, and will enable the 

 reader to understand more |clearly the value of the conclusions come to by 

 the authors. — Ed. , ^ 



