624 
Proceedings of the Royal Society 
containing the battery, so as to be acted upon inductively. When 
the carbon / is screwed tightly into the hole e, the circuit is com- 
pletely closed, and no sound uttered into M is heard at T. But 
when / is drawn gradually back, until small electric arcs are seen 
to pass between / and e, every sound uttered into M is distinctly 
reproduced in the telephone T. 
Here we have clearly only the air-effect acting, and that solely 
upon the small electric arcs passing between the carbons. X have 
found it, as yet, somewhat difficult to get the sounds to last for 
any length of time, in consequence of the arc distance soon 
getting too great for the current to pass and requiring re-adjustment. 
When the arc begins and ends a sharp click is heard in the 
telephone ; but in the interval during which the arc lasts the 
sounds heard in the telephone are distinct. 
As far as the tremor effect alone is concerned, it is obvious 
that the microphone action must depend, either (1) upon the 
variation of resistance due to variation of pressure, or (2) to 
variation in the extent of surface contact due to the elastic yielding 
of the carbons under pressure. 
To test the first of these causes, X made, about two years ago, 
some experiments on the effect of pressure upon the specific resist- 
ance of carbon. Por this purpose, I took a short length of carbon 
rod, and soldered wires to it at a short distance from each end. By 
means of these wires, the resistance of the carbon rod was balanced 
in the Wheatstone bridge. Pressure was then applied by means of 
a lever to the carbon in a longitudinal direction. Ho appreciable 
variation in the resistance was observed, even under considerable 
pressure, and it only became manifest when the pressure was 
sufficient to bend or crush the carbon. 
I have recently repeated ‘these experiments with the greatest 
care, and found the same results. X observe, also, that similar 
experiments, with the same result, have quite lately been made by 
Professor Sylvanus Thomson. Hence we can hardly, I think, 
believe that variation of specific resistance due to pressure can have 
the slightest effect in producing the microphone action. To test the 
second cause above mentioned, that is, the variation of resistance 
due to variation, in the extent of surface contact due to elastic 
yielding under pressure, X experimented as follows : — 
