622 BELL SYSTEM TECIIXICAL JOURNAL 



generalizations seem warranted, on account of the complicated series 

 of events recorded. These sounds are treated at length by Paget (9b, 

 p. 171-173) who observes considerable variation in their resonant 

 ranges, depending on the associated vowel. It will be noted howc\-er, 

 that in these four records particularly, consonant characteristics are 

 persistent and of large amplitude before the vowel sound begins to 

 appear. 



DTH/TH.— (Plates 141-144). The high frequencies (2UU0, ;jOUO, 

 3200) culminating at the transition point seem to be the key to these 

 records. They are more persistent for dth, while th appears to show 

 the steeper wave-front. Paget states (9b, p. 158) that "in 5 [(///;] 



the middle resonance (1149-1932, his figures] is overblown, louder 



than the corresponding resonance in B [th]." He gives also an 

 "upper sibilant of 3444-5950," louder for dth than th, and "difficult 

 to identify." It will be noted that in one record for dth (no. 141) 

 there is during the voicing period a faint high frequency which has 

 been set down in Table \T as 4000 cycles. This faint "sibilant" 

 (which may always be audible though it fail to be recorded) establishes 

 a certain kinship between these two sounds and those following (the 

 fricative consonants) which are rich in sibiliant sounds. 



V/F. — (Plates 145-148). v shows a pronounced \oicing, and as 

 pre\iously noted, a less prominent high frequency component than 

 its partner/, or any of the other fricative consonants. Comparing 

 V 7 with dth/lh it seems from the records that the former pair are of 

 higher frequency (particularly/) and that for v/f as a unit the high 

 frequency characteristic is more pronounced; just the opposite con- 

 clusion to that reached by Paget (9b, p. 161-162). / may indeed 

 differ more from v than v from dth, thus raising difficulties of classifi- 

 cation both physically and phonetically, which cannot be resolved on 

 the basis of the few records available. The exceedingly fine distinc- 

 tion between the sounds v and dth could be no more strikingly shown 

 than it is in the records gi\en, for both speakers. 



J/CH. — (Plates 149-152). Some of the recorded phenomena of 

 this pair suggest correspondences between them and the pair g/k; 

 luit the pair j/cA shows a higher frequency characteristic during the 

 important mid-portion of its history. Of the pair, ch seems to show the 

 steeper wave-front, that is, the more rapid transition to the vowel sound. 



ZH/SH.— (Plates 153-156). With this pair we pass to the field of 

 pure sibilants, in which there is no evidence of impulsi\e action or 

 steepness of wave-front. The action seems to be that in the voiced 



