48 THE STUDY OF SPEECH CURVES. 



or not, as the reader pleases — shows a long series of waves whose ampli- 

 tude rises and falls, but whose pitch, as can be seen directly in the length- 

 ening of the wave-groups, steadily falls. The latter part of [ai] falls 

 steadily to a very low pitch, as is usual at the end of a phrase. The [i] 

 is very weak. 



After an interval of 504mm. = 0.8s. the following stanza is begun with 

 the curve for "Who." The weak vibrations for [u], hne 12, fade in three 

 groups into [k] of " caught," which extends to the first part of line 13. 

 The unusually strong vibrations for [o] change their form in the middle 

 of line 13 ancl indicate slight diphthongization. The vowel fades in three 

 or four groups of faint vibrations, which, in my opinion, are to be con- 

 sidered as a sonant beginning of [t]. The limits between [t] and [h] can 

 not be given; the two sounds are very brief. The short [i] of " his," line 

 13, is followed by vibrations for [z], lines 13 and 14; those for [1)] of "blood" 

 are too weak to be seen. The vibrations for [1], hne 14, are weak. The 

 [a] of "blood" begins with very strong vibrations, line 14, which grad- 

 ually become weaker in line 15, and fade into [d]. The GOmin. omitted 

 belong partly to [d], but mostly to the pause. The curve for [ai] of " I," 

 hues 15-16, resembles that for [ai] of "I," hues 6-7, above, but is longer. 

 The [s] of " said" is represented by the straight line in line 17. The vowel 

 [s], hne 17, is followed by weak vibrations for [d]. Then comes a line for 

 [5], followed by the vibrations of short [d] of " the," in line 18. The [f] of 

 " fish," hne 18, is rather long; the vowel [i] is shown in line 19. 



Owing to the comparative condensation along the horizontal axis 

 (1mm. = 0.0016s.) the amphtudes of the vowels appear more character- 

 istically to the eye than in the Depew plate (1mm. = 0.0007s.). Several 

 general characteristics of the American vowel can be at once estab- 

 hshed. We note first that all but one of the vowels on the Cock Robin 

 plate is of circumflex amphtude, that is, the amphtude rises steadily to 

 a maximum and then falls, in a regular crescendo-diminuendo. This is the 

 fundamental form of amplitude in the American vowel; the variations 

 in suddenness and extent of the circumflexion are elements of expres- 

 sion. The vowel [i] in "fish," line 19, has a double circumflex; it should 

 be perhaps treated as an intensity diphthong with two hke elements, 

 or as [ii]; if not so short, it would be heard as distinctly diphthongal 

 as [ai]. 



It is to be noted that the two kinds of [ai] differ also in the course of 

 amphtude; the [ai] of "I" and "my" is not only a quahty diphthong 

 with two different types of wave-group, luit is also an intensity diphthong 

 with two maxima of circumflexion; the [ai] of "eye" and "die" is a qual- 

 ity diphthong with only one circumflexion of amplitude. An investigation 



