
HARMONIC ANALYSIS OF CERTAIN VOWEL SOUNDS. 773 
The vowel producing resonance cavities are clearly distinguished in virtue of 
two properties—first, the absolute pitch at which they produce a maximum 
reinforcement ; and second, the area of pitch over which reinforcement acts. 
The latter property, when it is extensive, is very probably due to the existence 
of subordinate proper tones not far from each other in pitch. This property of 
the oral cavity has, we believe, been hitherto much neglected. In order com- 
pletely to define the properties of the oral cavity, the relative reinforcing power 
at each pitch over the whole range of reinforcement ought to be determined ; 
our experiments, however, give only a rough idea of this relative power 
throughout the range. 
Professor Crum Brown’s gutta-percha cavity, which produced the vowels 6, 
a, ad, and 2 at very different pitches, proves conclusively that a constant cavity 
of irregular form and tolerably soft material is capable of producing any one of 
these vowels over a wide range of pitch. We have then to inquire whether the 
experiments show that these vowels (or more particularly the vowel 6, which 
has been most fully investigated), when produced by the human voice, are 
actually the result of a constant oral cavity ; and further, whether the action of 
a constant cavity can explain the phenomena observed for the vowel «, which 
was not one of those spoken by the bottle. 
It does not appear possible that the reinforcement which we have observed 
for the sound « can have been the effect of a constant oral cavity. There is, on 
the contrary, every evidence that, as the pitch on which the vowel was sung 
descended, the proper tone cavity was adjusted so as to bring its proper tone 
into unison, first with the prime tone, and then afterwards with the second 
partial. The proper tone of the cavity seems to have fallen note by note with 
the pitch of the vowel until it reached d, or thereabout, when it suddenly rose 
an octave, and then again went on falling as before. This, at least, seems to 
be much the most natural view to take of the causes which produced the 
excessive prominence, first of the prime and afterwards of the second partial 
when % was sung down the scale. It will be observed in Table XII. that in the 
duplex #s, when, according to this view, the maximum pitch of resonance of 
the cavity is in unison with the second partial, the fourth partials are appre- 
ciably strong, as might be expected to be the case. 
We should then describe the @ cavity as an adjustable cavity, with a very 
limited range of resonance, whose effect is to reinforce strongly only one partial 
lying above the pitch a. 
It is, indeed, possible that, so far as the small changes of pitch throughout the 
range of ordinary speech are concerned, even the # cavity may be constant. We 
have no evidence either for or against such a view. But, when the range of 
pitch is extended as it is in song, the analyses show that the cavity can no 
longer be constant. We may add, as a mere conjecture, that one point of 
VOL. XXVIII. PART III. 9Q 
