272 PROCEEDINGS OF THE AMERICAN ACADEMY 



changed to e on increasing the speed to a rate considerably above 

 that at which the former vowel was impressed upon the recording 

 cylinder, and fell to a when the speed of the cylinder was carried 

 considerably below that at which the record was made.* Various 

 other results of the same character were secured. 



The accuracy of these results was questioned, but they were shortly 

 afterwards confirmed by A. G. Bell and F. Blake, f and others. The 

 only form of phonograph in existence at that time, however, remark- 

 able as it then seemed to be, was extremely crude and imperfect, 

 the tin-foil employed as a medium for receiving the record not be- 

 ing well fitted to receive and retain the delicate impressions of the 

 sounds of the human voice. Upon the commercial introduction of 

 the modern phonograph of Edison, in which a cylinder of wax re- 

 places the tin-foil, and which leaves little to be desired so far as 

 clearness of articulation is concerned, it appeared to be desirable to 

 repeat the early investigations just referred to, with the improved 

 instrument. After our work was well advanced, our attention was 

 called to investigations in the same direction by Hermann. $ But 

 his very valuable researches on vowel tones, and their study by the 

 aid of the phonograph, have mostly employed methods other than 

 the one under consideration, and only some general results by this 

 particular method seem to have been published by him. These 

 are in accordance with the observations described in the letter to 

 Nature already mentioned. 



The method followed in the present series of experiments is 

 identical with that used in the earlier studies referred to. As the 

 cylinder of the modern phonograph is rotated by an electro-motor 

 furnished with a good speed-governor, it was easy to vary the speed 

 within moderately wide limits, and to keep it at a tolerably definite 

 and known rate, which was done in most of our experiments. In 

 some of these, however, only general results were sought for, and 

 no attempt was made to measure the speed. 



We proceed to give a detailed description of a few of our more 

 general preliminary experiments, which is followed by a statement 

 in tabular form of the later and more precise observations. In the 



* In tins paper the sounds of the different vowels are denoted by the con- 

 ventional signs employed in the Century Dictionary. Certain other signs used 

 are explained later. 



t See American Journal of Otology, Vol. I. p. 163. 



| Cent. f. Physiologie, 1890, Vol. IV. p. 242 ; Pfliiger's Archiv, 1890, Vol. 

 LXXIV. p. 42. 



