ABSTRACTS OF SCIENTIFIC PAPERS 367 



XXXVI. On the Harmonic Analysis of Certain Vowel Sounds. By 

 Fleeming Jenkin and J. A. Ewing. ' Transactions of the Royal 

 Society of Edinburgh,' 1878, vol. xxviii. p. 745. 



This group of papers (XXXTL-XXXVI.) refers to an experi- 

 mental investigation of the wave-forms of articulate sounds, under- 

 taken with the aid of the phonograph, which had then been newly in- 

 vented by Mr. T. A. Edison. In experimenting with the phonograph 

 the authors had observed that the quality of vowel sounds spoken by 

 the instrument did not vary so much as Helmholtz's theory led 

 them to expect, when the barrel of the phonograph was turned faster 

 or slower than its normal rate. This led them to undertake an 

 extensive analytical examination of records embossed on the tinfoil 

 of the phonograph, in the hope of finding that common feature in 

 the wave-forms corresponding to any one vowel sound, by virtue of 

 which the vowel is recognised when spoken or sung by different 

 voices or at different parts of the scale. The phonograph records 

 were magnified for the purpose of analysis by a system of light 

 levers terminating in a pen formed of an electrified capillary tube 

 which worked like the siphon of Sir W. Thomson's recorder, and 

 traced a magnified version of the wave forms on a moving strip of 

 paper. This magnified transcript was taken without injuring the 

 original tinfoil records, which were still available for reproducing 

 the spoken sounds. The wave-forms traced on paper were then 

 subjected to harmonic analysis to determine the amplitudes of the 

 constituent simple tones. By this means tables were constructed 

 showing the relative amplitude of the prime tone and the first five 

 overtones in the vowel sounds 6, a, a, and u, sung at various pitches 

 by several voices. The results were compared with Helmholtz's 

 theory of vowel sounds as depending on the reinforcement of certain 

 overtones by the oral cavity. They showed that the quality of a 

 vowel sound does not depend either on the absolute pitch of rein- 

 forcement of the constituent tones alone, or on the simple grouping 

 of relative partial tones independently of their absolute pitch. The 

 authors' conclusions are stated at length in the last paper, where the 

 method of experiment and the numerical results arrived at are also 

 fully described and a number of the magnified traces of sound waves 

 are reproduced by photo-lithography. The other papers are for the 

 most part preliminary notices of the last paper, but contain some 

 incidental observations regarding the reproduction of other sounds 

 than vowels by means of the phonograph. Amongst these is the 

 curious observation that any single element of spoken sound, whether 

 vowel or consonant, is phonetically reversible, as nearly as the 



