553 
of Edinburgh, Session 1879 - 80 . 
and bands it was also made to rotate slowly. A phonograph mouth- 
piece, with a very thin disc of wood or mica, was next placed, so 
that a fine wire with a, sharp point bent at a right angle, and with 
its other end attached to the centre of the disc just pressed with its 
sharp point on the convex surface of the bismuth cylinder. A 
current of four Bunsen cells was now passed through the wire and 
cylinder, and also through the violin microphone. When the violin 
was played the tune was heard faintly proceeding from the mouth- 
piece even when the bismuth cylinder was stationary. This arose 
simply from the loose contact of the wire and bismuth. The sound 
was, however, very greatly increased when the cylinder was rotated 
slowly, — so loud indeed, that it could be distinctly heard all over 
an ordinary room. I have been able to transmit singing very 
clearly, but not speaking clearly enough to be understood. This 
instrument is analogous to the loud-speaking telephone of Mr 
Edison ; but the explanation of their action must be very different 
if electrolysis, as is usually supposed, be the cause of the variation 
in the slipping of the platinum point on the chalk cylinder, which 
is characteristic of Edison’s instrument. Quite recently the electro- 
lytic action has been questioned, and a different explanation given 
by Professor Barret of Dublin. It is evident that electrolysis can 
in no sense come into play when the cylinder and rubbing point are 
both metallic. In that case two probable explanations of the action 
readily suggest themselves. The one is that there is more or less of 
an actual sticking of the metals together, arising from their fusion 
by the heat of the current. If this be so, then, the loose contact is 
alternately made a very good one, and then one actually broken. The 
other is the action of the Trevelyan rocker. Here, however, we 
have clearly only an analogous, and not by any means an identical 
effect. In the Trevelyan rocker the heat passes from a large mass 
of hot metal through two points of contact to a cold block, whereas, 
in the other case the heat is only produced at the surfaces of 
separation, the temperature of the rest of the metals being almost 
unaffected. Still it appears to me that the variations of the heat at 
this point has a great deal to do with the actions of all microphones, 
and in general with all sounds transmitted from one loose contact to 
another. This is shown by substituting cylinders of different metals 
for the bismuth cylinder above mentioned, all other things remain- 
vol. x. 3 t 
