Velocity of Electric Waves in Air. 1 23 



proceeding from the oscillator and preceding those oscillations 

 which give rise to the stationary waves. Whatever the 

 cause may be, it has the effect of lessening the resistance 

 between the coherer-globules. This phenomenon would 

 appear to throw some light upon the real cause of the action 

 of a coherer. It seems to show that the metal particles are 

 attracted nearer to each other, and there held in contact by 

 the electrostatic effect produced upon them by the passage of 

 the electric waves. Such an effect being different at different 

 distances from the reflector, and most marked near the 

 oscillator, would very much alter the readings really given 

 by the waves we are endeavouring to measure. Accordingly 

 this effect had to be eliminated. The elimination was thus 

 accomplished. Upon first closing the coherer circuit and 

 allowing the needle to come to rest, and then closing the 

 oscillator circuit, the above-described phenomenon having 

 taken place, the sparking at the oscillator was continued for 

 a few seconds, causing the needle to creep up to a maximum 

 reading; then the oscillator circuit was opened, and the 

 coherer circuit was opened and closed four or five times in 

 succession , each time the needle returned to the above- 

 determined maximum reading. Next, the coherer circuit was 

 opened, the oscillator circuit closed while the sparking was 

 continued for some time, and finally the coherer circuit, was 

 closed. The above-determined maximum reading of the 

 needle is in all cases (except when the coherer is at a node) 

 augmented; the amount of this augmentation increasing or 

 decreasing as we move away from the reflector towards the 

 oscillator. It is only such waves as the vigorous and pro- 

 longed sparking at the oscillator sets up and maintains that 

 are able to increase the first-determined maximum reading. 

 Thus the effects of the above-observed phenomenon are 

 obviated, since they are powerless to alter the first -determined 

 maximum reading. Such a precaution demanded a great 

 amount of additional time and patience ; nevertheless it suc- 

 cessfully eliminated the disturbance, which was xery detri- 

 mental to the accuracy of the observed readings of the 

 needle. All the measurements herein given were taken by 

 the above method. The readings thus taken are very small, 

 the highest not exceeding three small divisions of the milli- 

 amperemeter scale, but they are otherwise perfectly regular, 

 showing a distinct increase or decrease with change of position 

 of the coherer with regard to the reflector. Readings can be 

 taken with a fair degree of rapidity when all conditions are 

 favourable, as many as forty having been taken in three hours. 

 Measurements are begun at as short a distance from the 



