750 



That the speech movements made to produce oand o as distinguished 

 by the ear differ considerably', has been sufficiently proved in the above. 



Fig. 6. 



Pp. V. d. S. 



Fig. 6. 



Pp. R. 



E 



Fig. 6. 



Experimental phonetic analysis of the sounds. 



Also the characters of the sounds themselves proved to show a 

 difference which could be easily recorded. In the first place tiie 

 sounds can be easily registered on the kymogiaphion. It need hardly 

 be said that the taraboui' used for this purpose has to answer 

 special requirements. This part of the inquiry was made under the 

 guidance of Prof. Chi.umsky. The tambour had the same shape as 

 the recorder of a phonograph, the membrane was made of mica. 

 An aluminium "mouthfunnel" after Rousselot was connected with 

 this tambour by means of a wide rubber tube. Fig. 8 shows curves 

 of the two sounds as i-egistered in this manner. As a matter of 

 fact the vibrations of a membrane like this are not large, owing 

 to its stiffness; it is however partly due to this fact that we get 

 curves which are thoroughly characteristic of the sounds recorded. 

 So in our case there is a clear difference between the curve of 

 o and that of o. 



o • 



The sounds can also be registered by means of a phonograph. 

 A few monosyllabic words in which either of the two sounds 

 occur according to the meaning (e.g. bod and bot) as well as the 



