difference between the cathode and anode depended greatly upon the metal used for the electrode; the pressure being the same in all cases.
To test this point further I used a tube like that shown in fig. 6, where a, b, c are cathodes made of different metals, the anodes being in all cases platinum wires. The cathodes were disks of aluminium, iron, lead, tin, copper, mercury, sodium amalgam, and silver chloride; the potential-difference
Fig. 6.
between the cathode and anode was measured by Lord Kelvin's vertical voltmeter, and also by measuring the length of spark in air which, when placed parallel with the anode and cathode, seemed to allow the discharge to go as often through the spark-gap as through the tube. With this arrangement the pressures were the same for all the cathodes. The potential- difference between the anode and cathode and the equivalent spark-length depended greatly upon the nature of the cathode. The extent of the variation in potential may be estimated from the following table:-
| Cathode. | Mean Potential-Difference between Cathode and Anode. | |
| Aluminium..... | 1800 volts. | |
| Lead............. | 2100" | |
| Tin................ | 2400" | |
| Copper......... | 2600" | |
| Iron.............. | 2900" | |
The potential-difference when the cathode was made of sodium amalgam or silver chloride was less even than that of aluminium.
The order of many of the metals changed about very capriciously, experiments made at intervals of a few minutes frequently giving quite different results. From the abrupt way in which these changes take place I am inclined to think that gas absorbed by the electrode has considerable influence on the passage of the discharge.
I have much pleasure in thanking Mr. Everitt for the assistance he has given me in the preceding investigation.
Cambridge, Aug. 7, 1897.
