Apeil 10, 1896.] 



SCIENCE. 



545 



enable the observer to determine which fall 

 gives rise to the louder sound. The objec- 

 tions to this method are many and obvious; 

 it answers well enough for a demonstration, 

 but not for exact research. A second 

 method consists in moving an object pro- ' 

 ducing a constant sound, such as a ticking 

 watch, or a tuning fork, uniformly towards 

 or away from the ear, and recording the 

 minimum change in position, that enables 

 the observer to determine whether the sound 

 has grown louder or lower. This has ad- 

 vantages over the falling ball, but is far 

 from satisfactory ; and both are alike limited 

 in the scope of their applicability. There 

 is also an electrical apparatus, an audi- 

 ometer, that is useful in determining the 

 sensitiveness to minimal sounds, but is not 

 so satisfactory for determining difiFerential 

 sensibility; the sound moreover is very 

 artificial, difficult to listen to, and difficult 

 to reproduce. A common defect of all the 

 methods is the difficulty of determining by 

 an objective test whether the sound pro- 

 duced by the apparatus on one occasion is 

 really the same in intensity as in a succeed- 

 ing trial. 



It was in the attempt to secure a means 

 of graduallj'' increasing the intensity of a 

 sound, just as the siren gradually changes 

 the pitch, that I succeeded in devising a 

 moderately satisfactory apparatus for this 

 purpose. No apparatus can be regarded as 

 completely satisfactory unless its operation 

 depends upon a principle that clearly estab- 

 lishes the relation between the physical 

 stimulus and the sensational result. Un- 

 fortunately the physicist is not as yet ready 

 to define and measure the variotis factors 

 contributing to the tones produced by the 

 the apparatus about to be described. In 

 the absence of such knowledge the' appa- 

 ratus can be proposed only as an empirical 

 solution of certain phases of the study of 

 sound intensities. The apparatus makes 

 use of the principle of the singing 



flame. A singing flame consists of a 

 very fine jet of gas, burning through 

 an aperture of about one millimetre, under 

 a long, narrow glass tube; the pitch of 

 the resulting tone varies in an inverse 

 sense with the size of the tube. (For de- 

 tails see Tyndall, Sound, Lecture VI.) 

 The sound is due to the vertical vibra- 

 tions of the flame, the pitch being de- 

 termined by their frequency and the in- 

 tensity by their amplitude. The ampli- 

 tude, however, can be directly observed; 

 the flame is first turned down until the 

 sound just ceases to be heard, and this point 

 is noted on a millimetre scale placed in 

 back of the flame ; when the flame is turned 

 up to any given point the intensity of the 

 resulting sound is clearly marked by the 

 amplitude of the flame, as determined by 

 the height of the flame above the zero point 

 just described. 



The other requisite of the problem is a 

 means of delicately regulating the flow of 

 gas and thus the intensity of the sound, 

 This was accomplished as follows : An or- 

 dinary steam valve was taken apart and 

 the coarse thread adjustment replaced by a 

 fine one {^ inch), at the same time giving 

 the end of the pin a delicate taper; the 

 handle of the valve was then firmly fixed 

 to the center of a wheel ten inches in diame- 

 ter ; this larger wheel was moved by the 

 friction of a smaller wheel one inch in di- 

 ameter, having at its center an index mov- 

 ing over a dial eight inches in diameter. In 

 this way a movement of the index along the 

 circumference of the dial magnified about 

 100 times the change in the height of the 

 flame. The height of the flame is first de- 

 termined for a few points by sighting it 

 through a lens, and the divisions of the dial 

 are then made accordingly. 



One further difficulty remained, namely, 

 to secure a constant pressure of gas. This 

 was accomplished with sufficient accuracy 

 by forcing the air out of a bell jar (fitted 



