QUALITATIVE ANALYSIS. 51 



to resemble them; it does not in any way resemble [o] or [i], but the ear 

 fails to suggest anything more definite. The ear hears an [h] sound at 

 the beginning of each of the two laughs, but although it is fairly clear 

 that there is some sound after the vowel, at the end the ear can not decide 

 whether it is Pi] or not. The sounds of laughter are not so frequent as 

 the words of ordinary speech and their characters are not so typically 

 fixed; the ear is left, therefore, to judge from the sound itself with less sug- 

 gestion from the context or past experience. It is legitimate to ask, If 

 the ear fails us so when it receives no help from suggestion, does it not 

 falsify when the suggestion is present? I must add that I submitted 

 the record to five persons, who all showed the same indecision. 



The curves themselves are surprising. In the middle of the first line 

 the waves are characteristic vowel curves. Toward the right the form 

 of the wave changes steadily — that is, gradually from wave to wave — 

 but yet with such rapidity that there is as much change in five waves as 

 in two or three times the number in the American vowels we have con- 

 sidered ; between the middle and end of the vowel (last quarter of the 

 line) there are at least three utterly different types. The portion just 

 to the left of the middle exhibits a peculiar phenomenon ; there is a ten- 

 dency to weakness of the maximum amplitude in alternate wave-groups. 

 If dividers are placed over the middle portion so as to include two waves 

 it will be found that when applied to the earlier portion they will mark 

 off vibrations that appear to be almost single wave-groups. They can not 

 be single wave-groups, because they would then have an extremely low 

 pitch, whereas the laugh is on a high pitch; also because if that were 

 the case there would then be a sudden jump of an octave in the middle 

 of the laugh — a jump that the glottis can not make from one vibration 

 to the next, and finally because even in the middle of the laugh itself 

 (but not in the latter half) there is a fainter indication of the same 

 phenomenon. 



Let us now consider the laugh from beginning to end. At 11mm. 

 from the end of the first fine there appear faint vibrations which become 

 quite marked at about 40mm. further. They seem to fall into groups of 

 two faint waves with two vibrations each, and then a stronger wave with 

 likewise three subordinate divisions. Beyond this point the so-called 

 subordinate vibrations can not be grouped satisfactorily. We might be 

 inclined to take four subordinate groups for the fourth wave and two 

 each for the six following ones; but this is not allowable, because such 

 large groups would indicate a lower pitch of the voice than was present 

 in this case, and because the pitch can not change suddenly from group 

 to group. We are forced to conclude that each of the so-called subor- 



