618 



BELL SYSTEM TECHNICAL JOURNAL 



resonances could be identified and reproduced in models." It is 

 interesting to compare some observations made by him on /, ng, n, 

 m, and reported in liis second memoir. Working according to the 

 method previously described (§1) Paget has constructed resonators 

 which, under certain conditions, will produce transient forms of the 

 four sounds we are discussing. Their tone constituents are identified 

 by him as follows: 



Reson.ant Frequencies, Semi- Vowel Sounds 

 (Paget: Reference 9b) 



' Throat ' 



228-406' 



203-228 



203-228 



271 



'Middle" (Nasal) 



683 (faint) 



683 



541-724 



1217-1366 

 1217-1448 

 1217-1448= 



"Upper" (Oral) 



1625-1932' 

 1448-2169' 

 2298-2579 

 861-1722' 

 2434-2579 (faint) 



' Varying and finally approximating a characteristic region of resonance o( the 

 associated vowel. 



' Varj'ing with the associated vowel. 



.Siu(l\ing Paget's results in comKction with those ot Fig. 14, we 

 note that the energy spectra clearly show the "throat" resonances 

 for all four sounds in the neighborhood of 25(5 cycles. In the case of 

 « the nasal resonance at G83 cycles (Paget) is one of the prominent 

 tones centering around a frequency of 512 in the spectrum diagram. 

 This resonance also appears prominently in the spectrum for m though 

 Paget did not notice it. The higher middle resonances (1217-1448 

 cycles) which appear in Paget's table for the last three sounds appear 

 also in the spectra for these three sounds according to Fig. 14. Allow- 

 ing for the variation stated in notes (1) and (2) above, it appears 

 that the upper (oral) resonances for the four sounds, as noted by Paget, 

 are essentially the same as those that appear in all four spectra in the 

 diagram in the range of 2048-2896 cycles. 



With regard to Paget's observations on the transient character of 

 these sounds (he classifies them as consonants) and on the variability 

 of some of their components (Notes 1 and 2 of table above), depending 

 on the associated vowel, there is room for some difference of opinion 

 and the reader may form his own conclusions after a detailed inspection 

 of the records shown. Taking the sound 1 for example, and studying 

 first the three records loo, lee, la by M A and then the three correspond- 

 ing records by M B it seems to the writer that such variations as are 

 noted in characteristics are due not so much to change in the associated 



