70 PROCEEDINGS OF THE AMERICAN ACADEMY 



and the light of the spark concentrated by a lens, quite a bright field 

 was obtained, against which the electrodes were seen projected, as sil- 

 houettes. The light, though intermittent, of course seemed continuous, 

 since the sparks were so numerous. When good definition was ob- 

 tained, the microphone was set in operation, usually by means of an 

 organ-pipe placed at a convenient and variable distance, and in some 

 experiments by the voice. The pipe was blown by a constant blast, 

 and great uniformity of intensity in the sound was secured. An open 

 Cg organ-pipe making approximately 256 double vibrations per second 

 was commonly used, its pitch being variable to a moderate extent by 

 shading the mouth or the opening at the top. 



So long as the pipe was an exact octave in pitch above the inter- 

 rupting fork, the electrodes of the microphone as seen with the micro- 

 scope appeared to be at rest ; but if the interval was slightly disturbed, 

 the stroboscopic effect was observed, and the electrodes seemed to 

 move slowly through their complete course. Tiie rate of this apparent 

 vibration was of course dependent upon the deviation of the pipe from 

 exactness in its interval with the fork, and could be varied at will 

 within quite wide limits. 



The extent of the motion of the electrodes could be determined by 

 observing the grains of dust which adhered to them, or some definitely 

 marked roughness on their surface. Measurements were made by 

 means of a spider-line micrometer, the wires of which were placed at 

 a convenient distance apart, and the amplitude of the motions of the 

 selected points of reference on the microphone was determined by esti- 

 mating the relation of their apparent motion to the distance separating 

 the wires of the micrometer. This last distance was frequently varied 

 to diminish the liability to error from a possible bias of the observer 

 towards an agreement with earlier results. 



In the experiments described in the present paper, the electrodes 

 were generally so adjusted that the motion of the anvil electrode was 

 too small to be observed. Under these circumstances the observed 

 motion of the hammer electrode, as measured by the micrometer, was 

 the motion of this relatively to the anvil electrode, which is of course 

 the quantity to be determined, rather than the actual excursion of tlie 

 hammer electrode. 



The microphone was placed in circuit with a battery and the pri- 

 mary of an induction coil, whose secondary contained a receiving tele- 

 phone. With this arrangement the effect on the ear of the electrical 

 variations due to the various values of the excursions of the electrodes 

 could readily be observed. In many case-? a second observer was sta- 



