of Edinburgh, Session 1881-82. 729 
it came through, in considerable quantities, thus causing a great 
deal of waste. 
One of these cells was put on a voltameter consisting of two 
copper plates, “both weighed,” separated by two strips of wood 
tied together with string, and immersed in a beaker of CuS0 4 sol. 
This I found forms a very good voltameter for rough purposes, 
and as the plates are very close together its resistance is low, which 
is of course important. By weighing both plates, and taking the 
mean, the results are fairly accurate. I used ordinary sheet copper, 
but no doubt electro-deposited copper would be better. I describe 
this form of voltameter, because the form described in Wiedemann’s 
Galvanismus , namely, a platinum crucible with a silver rod in the 
centre, is only of use in very special and small experiments. It is 
very expensive and troublesome to work with. In this experiment, 
then, I found that after the cell had run six hours, the difference 
between the weight of zinc dissolved and the equivalent weight of 
copper deposited represented a loss of 15 per cent. 85 per cent, of the 
zinc dissolved had been returned as deposited copper, notwithstand- 
ing that I had used linen, which caused considerable waste. I then 
measured the internal resistance of a cell of 28 square inches 
surface, and having a distance of J inch between the plates, and 
found it to be T5 ohm. I have reduced the resistance very 
considerably below this in recent cells. 
I also made up twelve primary cells of 28 square inches surface 
each, and used them to illuminate a small Swan lamp, measuring 
the current and the candle power. 
These experiments also gave very satisfactory results. 
When I began to charge one of these cells, however, measur- 
ing the current on a tangent galvanometer, I found to my 
astonishment, that the current, starting at the strength to be 
expected from the EMF of the external battery and the resistance 
of my cell, rapidly fell off, so that in a few seconds the current 
indicated an increase of resistance in the cell from a few tenths 
of an ohm to as much as thirty ohms. At this point the current 
would remain stationary. On now attaching the cell to the gal- 
vanometer a powerful current would flow from it, indicating that it 
had instantaneously returned to its normal internal resistance. At 
first I supposed that this was due to some polarization set up in the 
