188 Mr. J. Daniel on the Polarization upon a 



are indicated in fig. 2. A solid iron core was used in the 

 induction-coil to avoid the great difference in the direct and 

 inverse currents which exists when a wire core is used, and 

 which results in a polarization of the voltameter. The P.O. 

 box was used as a Wheatstone bridge, of which the voltameter 

 was one arm, and the movable coil of the electrodynamometer 

 was connected across the branches of the bridge in the usual 

 "null method" fashion, the fixed coils being in series with 

 the main circuit. Being thus used only as a galvanoscope, 

 its constant need not be discussed ; suffice it to say that the 

 arrangement was sensitive to several decimals, though resist- 

 ance was measured only to hundredths of an ohm. 



(c) The voltameter is simple, though its construction 

 involves the only novelty in the apparatus. The voltameter 

 first used was a glass vessel, 18 centim. long, 6 centim. wide, 

 8 centim. deep, with a hard rubber partition, sealed-in with 

 sealing-wax, in the middle. The opening in the vulcanite was 

 3 centim. wide, and extended to within 2 centim. of the bottom. 

 This frame for the partition was made by sealing two pieces 

 of vulcanite together with a rim of the same between, of the 

 same thickness as the glass plates bearing the metal partitions, 

 which were slipped down into the groove thus formed. This 

 was convenient, as the plates could be easily and quickly 

 changed ; but a quantitative test showed leakage around the 

 edge of the glass plate of nearly 20 per cent, of the conduc- 

 tivity when the plate was open, using solid and bored glass 

 respectively : hence the results obtained with this voltameter 

 are only relative, not quantitatively correct. This voltameter 

 and its results would not be mentioned here but for the fact 

 that I was so careless as to think that this tight-fitting glass 

 would not allow sensible leakage under the circumstances, 

 and so made a long series of measurements before testing 

 the leakage quantitatively, and, too, with some solutions 

 which I have not had time to repeat with the 



Voltameter as finally adopted. 



This consists of a glass vessel 20 centim. long, 12 centim. 

 wide, 11 centim. deep. In the middle is fixed by sealing-wax 

 a plate-glass partition 3 millim. thick, with a hole in the centre 

 2 centim. in diameter. A smaller glass plate, with hole 

 1*5 centim. in diameter, is sealed over this opening, the metal 

 partition being first sealed over the hole in the small glass. 

 This gives two water-tight compartments, which also offer no 

 electrical communication except through the metal under 



