THE SOUNDS or SrF.ECll f»2I 



in the rase of the sound t' (Plates 14") ami 14()) the hi^li frer|ucncies arc 

 IHTsisti-nt and in many rases of !a^^Je amplitude, both at the start and 

 during the course of the consonant soimd. Thi'se frequencies rise, 

 as we Ro throuijh this^roup. to values of 7000 and 8000 cycles in the 

 case of the sounds z and s, shown in the last four records. For a 

 full appreciation of these pronounced high frequency characteristics 

 reference must be matle to the records themselves, or the siniimary of 

 char.icteristics, in Table VII. Here again, in distinguishing these 

 sounds the remarkable pierformancc of the ear is manifest, and the 

 recording apparatus is used nearly to the limit of its utility. 



We may best conclude this discussion of the consonant records by 

 brief comments on some of the individual sounds, and a comparison 

 where possible with data given for them in Paget's second meinoir. 



B P.— (Plates 129-132). Both Paget (ref. 9b, p. 165) and Miller 

 (rcf. 3) have noted the essential impulsive quality of these sounds, and 

 have pnKluced them by sudden closing and opening of the mouth of a 

 resonator. Paget considers p to be the more suddenly released, i. e. 

 to have the steeper wave-front. From the records this is not evident; 

 following the voicing pericxl, the h would seem to be more suddenly 

 produced, as judged b>- tlie growth in anii)liiu(k- of tlu- <; siniiui lol- 

 lowing. 



D'T.— (Plates 133-131)). For both of these (see either Table \1 

 or the records themselves) we note a high frequency characteristic of 

 about 4000 cycles. Paget (9b, p. 168) observed "an upper resonance 

 5 to 8 semitones higher than that of the associated vowel, and a low 

 resonance of about 362." We note in the records a low frequency of 

 the order of 500 in the case of d. Paget notes a "greater amplitude 

 in / clue to higher air pressure" and the records show a greater ampli- 

 tude for the high frequency in the case of /, except right at the transi- 

 tion point, where d shows the high frequency of large amplitude. 

 No conclusion can be given as to relative steepness of wave-front, d 

 vs. /, because in both cases we note for speaker MB (Records 134, 

 136) a steeper wave-front than for MA (Records 133, 135). The 

 difference between d and / may depend entirely on the voicing and on 

 the complicated phenomena at the transition point. 



G K. — (Plates 137-140). k shows the characteristic transients 

 (1.500, 4000; Table IV, notes 4 and 5) to much more pronounced degree 

 than g. From the records it would seem that g, in addition to the 

 voicing, disclosed a steeper wave-front, the four transitional cycles 

 required for k (records 139-140) emphasizing this point. No other 



