1226 Otf VOCAL SOUNDS. 



rotated. This should obviously not be the case, were the relative pitch 

 theory correct. It is curious that even with competent observers there 

 should be such difficulty in deciding this apparently simple question of 

 fact, some asserting that there is no difference in quality, and others as 

 positively stating that there is a difference when the cylinder is caused to 

 move slowly. M'Kendrick, after many experiments, has come to the 

 conclusion that there is a difference, but not so great as to disguise 

 the quality of the vowel. To produce any appreciable change in 

 quality, the cylinder must be driven at an enormous speed. The ear 

 can always distinguish or A or E at different rates, but the quality 

 is undoubtedly altered. The sound of a vowel never passes into the 

 sound of another vowel. 



Hermann presents the fixed-pitch theory in a modified form, and 

 states that there is for each vowel a characteristic tone, a formant. Not 

 only does Hermann support the fixed-pitch theory, but he maintains 



FIG. 438. Examples of vowel curves obtained by Hermann. 



that the formant need not necessarily be a partial of the fundamental. 

 Sometimes it is such a note, but more often not. 1 The pitch of the 

 formant, however, may vary considerably ; indeed, with the same prime, 

 it may vary as much in certain cases as several semitones. Fig. 439 shows, 

 in musical notation, the pitch of the vowel, according to Hermann. 2 



Sauberschwartz 3 has investigated the subject by an ingenious 

 application of the laws of the interference of sound. Certain vowels 

 were sung into the mouthpiece of a long tube, to which other short 

 tubes of definite length were attached. By closing the outer ends 

 of certain of these tubes, various partials could be extinguished by 

 interference, and the listener at the other end of the tube observed an 

 alteration in the quality of the vowel. Thus, by extinguishing the 

 formant of A, the sound approximated to that of 0, and became 

 somewhat nasal. With E there are two characteristic formants, and 



1 Hermann, Arch.f. d. ges. PhysioL, Bonn, 1895, Bd. Ixi. S. 178. 



2 Ibid., Bd. xxxv. S. 42. 



3 Hermann, Ibid., 1895, Bd. Ixi. 



