34 THE STUDY OF SPEECH CURVES. 



its character somewhat Uke the sound we usually represent by " sh". 

 The second puff was louder, hollower, and of lower pitch. The third 

 was still louder, but again of higher pitch. Like all noises, these puffs 

 are characterized by irregular vibrations in which there are indications 

 of regularity. We can even pick out portions that are regular enough to 

 justify the statement that tones are present. Thus in the fourth line it is 

 clear that the waves with the period 9mm. = 0.0036s. represent a distinct 

 tone, about c'**. In the last puff two strong tones are evidently present. 



The first line of figure 37 gives the curve produced by striking two 

 blocks in rapid succession to imitate the gallop of a horse; the waves of 

 the first blow fall into groups of rather long period. The period of the 

 group is 18mm. = 0.0072s. or about c"*; the period of the waves is 3mm. 

 = 0.0012s., or about gr'^*. There is also in the latter part of this noise a 

 vibration that steadily shortens its period, namely, 8, 7, 6mm. —0.0032, 

 0.0028,0.0024s.; a small rise in pitch regularly accompanies tones of decreas- 

 ing intensity, as in fading tuning-fork tones; here the rise is rapid. The 

 suggestion presents itself that one of the essential characteristics of a noise 

 of this kind is the rapid dying away of the vibrations (with the accompany- 

 ing rapid rise in pitch) due to the high degree of damping (internal friction). 

 The second noise in this hne shows a long period of 9.5mm. = 0.0039s., or 

 about c' ; this represents a tone nearly an octave higher than that in the 

 first noise. The stronger blow of the blocks thus produced a sound not 

 only greater in intensity but also higher in pitch. The last four lines of 

 figure 37 give four blows of the same blocks in rapid succession, with 

 increasing intensity, the last blow being specially emphasized ; the period 

 diminishes as the intensity of the blow increases — this seems to be a 

 general physical law for such noises. 



The whistUng curve in figure 38 is from a professional whistler; it 

 gives a very small portion of a long stretch that varied only in the length 

 of the waves. The period at this point is 1.05mm. = 0.00042s., the fre- 

 quency 2381, and the note about d*. 



Figure 39 gives a note whistled to a piano accompaniment; the eye 

 readily selects the portions where the piano vibrations occur alone and 

 those where the high note of the whistling is imposed upon them. It 

 will be noticed that the note of the whistUng is harmonic to that of the 

 piano; that is, that one wave-group from the piano is evenly filled out 

 by a number of whistling waves. 



Figure 40 gives the vibrations produced by a plucked string of a Chinese 

 musical instrument; the pitch of the string is high. We note the pecu- 

 liar phenomenon of pseudobeats — or perhaps real beats, if such can be 

 produced by a single string; on account of the brevity of the sound they 



