July 25, 1878] 



NATURE 



341 



Table I. — Amplitudes of the Harmonic Constituents of the Vozvel Sound Sung at Various Fitches by the same Voice. 



12-1 7-1 07 0-1 0-5 0-4 



10-5- 



-6-9- 



■07- 



-0-3 0-2 0-3 



11-9- 



-7-G- 



•0-5- 



-ro 0-3 0-2 i 



11-0 • 

 7-0 — 



16-0- 



rs- 



■ro 1-0 07 



;2-6- 



rs- 



•r4 0-6 O'l 



7-5 • 



-18-5 



12-5- 



■19-0 



6-9- 



■10-3- 



■27- 



-0-6- 



-0-2- 



-0-2 



5-5- 



■14-0- 



■4-5- 



-0-4- 



-0-8 -Q-t 



7-9- 



16-5- 



7-4- 



-0-8- 



-ro 0*3 



4-4- 



13-4- 



■8-2- 



-1-6- 



-2-2- 



-1-0 •• 



■9-5- 



6-l^ 



-3-3- 



-0-3- 



2-5- 



1 5- 



37- 



1-5- 



•5-8- 



I-o- 



■2-8- 



■6-1- 



3-1 0*6 0*5 



-47 ' ' ' l-'l^-^6 : 



rs- 



-2-9- 



-0-6- 



-0-3 



w 



1-5- 



-4-0- 



-0-8- 



-0-4 



A B, B 



Y%'-f-f^'!ef%'''Y^'i>^-^'^^^T'%^'''Y''^^'-^''Y%'''V''''Y''^^''^''W'Y''''i 



1 



from_/"' to g the first and second partials are the main com- 

 ponents ; from/ to d the third partial is prominent, and about 

 c the fourth partial becomes strong. At G, in one voice, the 

 fifth partial was strong, whereas in the example now given it 

 was weak. 



When the several voices were compared in respect of the. 

 relative strength of the partials, it became clear that this rela- 

 tive strength might vary greatly without affecting the vowel 

 quality much. 



This is brought out by Table II., which contains the con- 

 stituents of o as sung by five different voices on five different 

 notes. The figures are proportional to the actual amplitudes of 

 each harmonic constituent, and therefore not only. show the rela- 

 tive value of each constituent in a given vowel, but also indicate 

 (although only very roughly) the strength of the several voices on 

 each note. Those constituents which are most considerable are 

 printed in large type. Those printed in small type may perhaps 

 be regarded as accidental and not essential to the voAvel sound. 



On / the vowel sound is composed of the two first par- 

 tials, but with different voices these were present in propor- 

 tions which varied from ^— ? to ^^. 

 102 34 



On c' the vowel sound was still composed almost wholly of the 

 two first partials, and with different voices these were present in 



proportions which varied from ^ to ^^. From this it is 



145 46 



evident that the vowel sound may vary very slightly as a vowel 

 when the proportions of its constituents vary largely. No doubt 

 it may be said that the vowels pronounced were different vowels, 

 but it must be remembered that each man was at least trying to 

 sing the same vowel and was controlled by the same observers, 

 who were satisfied that he did sing generically the same vowel. 

 On the other hand, the quality of the voices 'was extremely 



^"f" 



different. Voice No. i was a fine, powerfid, trained baritone. 

 No. 2 was a high-pitched voice with no notes below f, rather 

 harsh, and unaccustomed to singing. No. 3 was the trained 

 bass voice of a man of eighty. Nos. 4 and $ were voices of 

 average range and moderate power, with some musical training. 

 No. 6 was a powerful bass. The differences in the proportions 

 of the constituents appear to us therefore to indicate differences 

 in the quality of the voices rather than differences in the vowel. 



On c' the two first partials give a good J, whether the second 

 is one and a half times as big as the prime or the prime twice as 

 big as the second. 



The results on a and b do not differ much from those on cf ; 

 when, however, we come to g, the table shows us that while two 

 out of the five voices formed their by producing two partials as 

 in the case of the higher notes, three other voices introduced a 

 considerable third partial. On descending one tone lower, to 

 f, we find that all the voices agree in employing a strong third 

 partial. This third partial is sometimes less than half the prime, 

 but it is never so small as one-third, and in one case it is sensibly 

 larger than the prime. It would also be observed that it was 

 the voice No. 5, which had no sensible third on^which gave the 

 largest third on /. It should also be noted that the smallest third 

 is due to voice No. 4, which was at the end of its range, and 

 spoke or sang half a tone sharp. 



As we descend below / the third assumes more and more 

 prominence ; on d voices I and 5 gave the following series of 

 partials : — 



III. 

 8-2 

 5-6 



Table II. shows that by the time B is reached the fourth 

 partial has become very prominent. 



