THE SOUNPS ()/• Sri.l.CIf 609 



form.iiu'f (if iM«-li in spi'akinjj tin- tliirtirn vowi-l sounds, anfl will l)r 

 list-fill in thi" discussion to follow; they arc shown in Tahlc III, hflow: 



TAUl.l'; III 

 Sfxtikers' Charailcristics 



These records were made without constraint Imposed on the speaker, 

 except that he had to start and stop within an interval of about one 

 second, and was requesteti to repeat the sound several times at what 

 he judjjcti to be constant loudness. The resulting variation in per- 

 formance may therefore be of some interest. 



Of 52 men's records the vowel sounds 35 records showed a "normal " 

 effect of progressive rise in pitch during the course of the record. (The 

 mode is taken as the normal effect, and follows the mean very closely.) 

 In 6 records out of 13, speaker MB showed an individual or biased 

 effect of slight fall in pitch toward the end. The women's records show 

 greater variation, 24 records out of 52 showing a "normal" effect of a 

 rise in (>itch, followed by fallinii pitch, during the course of the record. 

 The individual bias of speaker KB towartl progressive fall in pitch was 

 shown in 7 records; that of KD toward progressive rise in 4 records. 



The relative constancy in fundamental pitch shown b\- speaker MB 

 is best exemplified in Plate \o. .58. Speaker FD made 3 records of 

 constant pitch: N'os. 24, 40 and 48. Other records of constant pilch 

 arc Nos. 19 and 99, both by MC. 



In duration, the bias of speaker MB towards short records was 

 shown in 6 records which fell short bv .08 sec. or more of the mean 



