of Edinburgh, Session 1877-7 8 . 
631 
sudden. Then, in a mercury break, the spark that ensues on the 
separation of the dipper from the surface of the mercury is con- 
ducting, and it may only be the break of the attenuated current 
when the spark ceases that has telephonic power. This would 
bring us down to a low scale of effective action, when working 
with a water cell interrupted six times per second. It cannot, 
however, be with certainty alleged that this kind of interruption 
is strictly single, as there may be pulses in it preceding the final 
stoppage. 
The relation of resistance to sound can be shown by an experi- 
ment like the following : — I measured off a yard of iron, German 
silver, copper, and silver wire, of diameters respectively of *008, '0086, 
*007, and ’006 of an inch. The resistance of each yard was re- 
spectively 3, 8 '7, *7, and 1*4 ohms (the silver alloy was exceptionally 
hard). From these again I cut off a foot and soldered them end to 
end in a successive chain, and stretched them out (fig. 8) in a line, 
with their junctions firmly clumped to prevent the sound of the one 
running into the other. I then attached four similar mechanical 
telephones to them, and found, at least so far as the ear could judge, 
that the sounds emitted by each, when the same broken current 
passed through all, were proportionate to the resistance such offered, 
in the case of the last three metals, whilst the sound given out by 
the iron distanced all the others. The German silver here was much 
louder than the silver, which again was, one could fancy, twice as 
loud as the copper. This seems to clash with what we have stated 
with reference to the conducting discs on or near the core of the 
magnet of the telephone. It may be somewhat hazardous to venture 
an explanation, but there is one that turns up so readily at the first 
consideration of the matter that I may mention it. In the case of 
the wires the same current passed through all, but in the discs the 
currents induced would be, as is generally received, relatively in pro* 
