384 MEASUREMENT OF THE RESISTANCE 



beginning of the single deflection. That this should be so is 

 evident. For, after the double deflection has just vanished, the 

 first current passing through the Galvanometer still tends to cause 

 a deflection to the right, but is overpowered by the polarization 

 current. Hence the mirror does not begin to move as soon as 

 contact is made but, for a short time, remains at rest. Clearly 

 then, if the adjustment of the arms be changed until we can 

 observe not only no double deflection, but also no hesitation, we 

 shall get more nearly accurate results. By observing the vanish- 

 ing of the hesitation, Prof. MacGregor was able to get results 

 which, in the case of high resistances, differed from their true 

 values only by from 0.1 to 0.4 per cent. 



I began, therefore, by endeavoring to observe the vanishing of 

 this hesitation, testing my results in the same way as Prof. Mac- 

 Gregor had done in the paper cited above. 



The electrolytic cell which I used consisted of strips of window 

 glass cemented together with marine glue. Its length was about 

 16 cm, breadth, 8 cm., depth, 9 cm. It was divided transversely 

 into two equal compartments by a glass partition cemented in so 

 as to be water tight. Passing through the centre of this partition, 

 and cemented at right angles to it, was a glass tube with its ends 

 open. It length was 1 cm., the diameter of its bore 0.3 cm. At 

 each end of the box I placed my platinum or amalgamated zinc 

 electrodes, as the case might be. They were each about 7 cm. 

 square and had narrow strips projecting above the box. To these 

 strips of the zinc electrodes thick copper wires had been soldered 

 without the use of acid ; to the platinum electrodes, thick platinum 

 wires had been welded. By these wires the cell could be joined 

 up as an arm of the bridge. 



The electrodes were, in all cases, placed as close as possible to 

 the ends of the box. Since, however, the resistance of the 

 column of liquid in the tube was about 1,000 times that of a cross 

 section, 1cm. in length of the liquid in the box, slight differences 

 in placing the electrodes made no appreciable difference in the 

 resistance of the box. 



The part of the box above the surface of the liquid (especially 

 that of the glass partition) was kept clean and dry, and, during 



