12 NEWBERY, The Theory of 'Overvoltage. 



compared with that obtained when K % is disconnected 

 and K 1 depressed, K x being in connection with the standard 

 Weston cell IV. 



Two points in connection with the commutator need 

 special attention here. 



The brushes on the commutator must make a clean 



sharp break in the primary current before connecting up 



the galvanometer circuit. Any drag in the break due to 



a loose wire in the brush will produce remarkable results 



in the measured overvoltage which may be highly mis- 

 ts j & ^ 



leading. 



If the commutator is rotated at a speed greater than 

 1,000 per minute, no change is produced in the observed 

 overvoltage by variation of the speed above this limit. 

 At lower speeds, a slight fall in the observed overvoltage 

 is produced with some metals, but in most cases it is not 

 more than two per cent, lower at 500 revs, per min. than 

 at 1,000 revs, per min. This is important as it shows that 

 the fall of the back E.M.F. in i/2000 th part of a minute is 

 negligibly small under the given conditions. It was other- 

 wise in Le Blanc's oscillograph experiments, owing to the 

 fact that he not only cut off the current, but reversed it. 

 Hence the accumulation of gas capable of setting up a 

 long continued back E.M.F, was prevented. The con- 

 ditions of Le Blanc's experiments are not present in 

 regular electrolytic work, and therefore his results, though 

 of great theoretical interest, are not likely to be of the 

 same practical value as those obtained by the method 

 just described. 



Observations of the effect of much slower commutator 

 speeds on the measured value of the back E.M.F. may 

 give some useful information on the rate of depolarisation 

 of the electrodes. 



With this apparatus therefore, the true value of the 



