714 
Proceedings of the Pogal Society 
5. On a Class of Determinants. By Mr J. D. H. Dickson, 
Tutor of St Peter’s College, Cambridge. 
6. On the Wave-Forms of Articulate Sounds. By Professor 
Fleeming Jenkin, F.E.S., and J. A. Ewing, B.Sc. 
{Abstract.) 
By the help of the phonograph we have continued the investiga- 
tion described in a previous Communication (Proc. R.S.E., p. 582), 
and have now obtained about two hundred magnified traces of the 
phonographic records of vowel sounds spoken and sung by various 
voices, and of these sixty-five have been already subjected to har- 
monic analysis, extending as far as the sixth partial tone. In each 
case the results have been accepted as satisfactory only when, after 
the magnified trace had been obtained, the record on the tinfoil of 
the phonograph still gave the vowel sound satisfactorily. Our 
attention has hitherto been almost exclusively directed to the vowels 
u (the vowel sound in “ food,”) and o (as in “ oh,”) both of which 
are well spoken by the phonograph. The results, which are still 
incomplete, are briefly as follow : — 
When a vowel sound is continuously sung without change of 
pitch or quality, the wave-form produced is of remarkable constancy, 
showing that the compound sound does not contain any, or at least 
any important, constituent which is inharmonic to the prime tone. 
A naturally high man’s voice, with the comparatively small range 
/ to f saying u at any pitch throughout its range, produced a wave- 
form which was substantially a simple harmonic curve of length 
corresponding to the pitch. The upper partials, when present at all, 
were present only very feebly. Thus u sung on iS gave a prime 
whose amplitude was 25 (in the unit of measurement used), the second 
partial was only 0*8, the third IT, and the others inappreciably 
small. 
When the same voice spoke o anywhere throughout the same 
range, it produced a trace which in every case consisted of a strong 
prime and a strong second partial (that is to say, the octave of the 
prime), the higher partials being feeble or absent. Experiments on 
this part of the scale with various voices proved that the proportion 
which the amplitude of the prime bore to the amplitude of the 
