
HARMONIC ANALYSIS OF CERTAIN VOWEL SOUNDS. 765 
we satisfied ourselves by frequent trials that in order to get a good @ the curve 
had either to be of the duplex or the simple type. The change required to be 
sudden. Anything between the two forms came out 6. And the same voice, 
in singing @ on this critical note, did actually sometimes produce the duplex 
form and sometimes the simple form. The two forms could be made to over- 
lap.* We found the same thing true of other voices. Fig. 8 shows that voice 
1 gave the duplex form as high as bb. Another voice, whose range was 
extremely limited, gave the simple form on every note it could take, and carried 
it as low as f, which was this voice’s lower limit. 
Table XII. gives the analyses of the curves in fig. 8. It may be mentioned 
here that although comparatively few examples of « curves have been tran- 
scribed and analysed, the simplicity of the curves was such that their form 
could often be observed fairly well by mere inspection of the tinfoil record ; 
and experiments with several voices have given us confidence that the examples 
given above are really representative of the results to be obtained from this 
letter. 
By way of further testing the vowel sound «@ at the critical point where it 
changed from the simple to the duplex form in a given voice, attempts were 
made by voice 5 to slur both up and down past this pitch ; that is to say, to 
go on pronouncing @, and gradually raise or alter the pitch so as to pass the 
pitch a. In all these attempts, however, one or other of two things happened. 
Either the vowel quality as appreciated by the ear and the form of the curves as 
recognised by the eye changed to 6 for an instant as the critical pitch was 
being passed, or else there was an interval of almost complete silence, and the 
curves sank to an insignificant size. This, as well as the evidence already 
given, goes to prove that an abrupt change must take place from the duplex to 
the simple form of @ if the vowel quality is to remain pure. 
As some doubt may be felt whether this smgular change in the wave-form 
of a may not have been due to some peculiarity in the instrument, we may 
repeat that, being fully alive to this danger, we tried the experiment with 
changed mouthpieces, changed vibrating discs, and changed springs, but always 
with the result that on or about the pitch @ certain voices made a sudden 
alteration in the constitution of their @. Further, it must not be forgotten 
that another voice, using the same instrument, continued to give the simple 
form as low asf. 
The following is a general summary of the results established by our 
experiments in the vowel sound 7@. , 
(1.) The generic character of @ from d to,f' is given by the prominence of 
* The following note is taken from our diary :—“‘ May 20. Voice No. 5 did duplex and single a’s 
on bb, a, and g. The duplex @ on bb was very 6-ish. The others were good. The single form on a was 
the best of all.” 
VOL. XXVIII. PART III. 90 
