Il6 CARNEGIE INSTITUTION OF WASHINGTON. 



therefore alter the form of the puff (implying that the vowel involves 

 not only special innervations to the mouth muscles, but also to those 

 of the larynx). Interesting details concerning vowels of different 

 speakers were found — e. g., the strong chest tones of Jefferson and 

 Depew, the strong mouth tone of Mitchell, etc. 



Attempts were made to imitate the vowel curves by apparatus and 

 by calculation on the assumption that if the results were good coun- 

 terfeits the principles used in the apparatus or the computation must 

 be valid for the vowels themselves. Principle after principle was 

 tried. Good counterfeits were finally obtained. The principles 

 found were used in developing the vowel theory just mentioned. 



An apparatus was made to produce artificial curves on a gramo- 

 phone disc. A speech curve of an}^ kind could be taken and engraved 

 on the disc. On placing the disc in the gramophone the sound is 

 heard. This can be used to test any published curve. A zinc etch- 

 ing is made from it. This is used in the apparatus, and the sound 

 is heard. This apparatus opens up an utterly new field, namely, 

 that of the acoustic analysis of a vow^el. Each wave for a vowel 

 curve is engraved separately in repetition, and its sound is heard. 

 Thus the sound of the vowel at each of its waves is produced. This 

 apparatus produces a series of acoustic sections of the sound of a 

 vowel, just as a microtome makes a series of sections of a tissue. 



The studies of speech melody showed that the fundamental form 

 for the American sentence is a circumflex melody, and that this is 

 modified to suit each expression of thought or emotion (parenthesis, 

 religion, ceremony, humor, etc.). Curves of the interjections illus- 

 trated how the melody was changed to express the emotion . Records 

 of a German poem proved that it had a general type of melody — a 

 specific melody — of its own, which showed itself in spite of the 

 different melodies of different dialects of Germany. 



These investigations form a unit and have to be carried along to- 

 gether. Every one of them is the first attempt to enter a new field 

 either in any way or with accurate experimental records. Owing to 

 the funds available and the concentration of effort, these investiga- 

 tions are now so far advanced that they can not be duplicated else- 

 where for many years. The stimulus of the work is being felt in 

 Germany. The melody investigations have been taken up in the 

 University of Leipzig, and work on curves — obtained by Professor 

 Scripture's apparatus — is being carried on in the University of Berlin. 

 It is intended to make a unit of three years' work and publish the 

 results in a volume of text with an atlas of plates containing speech 

 cur\-es. 



