OF ELECTROLYTES — MCKITTRICK. 383 



With an actual mirror and magnet of finite moment of inertia, 

 no appreciable motion can be observed till the current has 

 acted for some time If the adjustment of the arms be such as 

 to indicate a resistance for the cell only slightly greater than its 

 real value, the first current flowing through the Galvanometer 

 will be very weak, and, before it will have had time to produce 

 an appreciable deflection of the mirror, the electrodes will have 

 become polarized, and our only deflection will be one to the left. 

 If, then, we take the resistance of the cell to be that indicated 

 when the double deflection has just vanished, we make an error, 

 and conclude the resistance to be greater than it really is. 

 Whether, with any magnet and mirror yet manufactured, this 

 error can be made small enough to be neglected, can be deter- 

 mined only by experiment. 



In their experiments, Ewingand MacGregor assumed that the 

 light mirror and magnets of Lord Kelvin's Dead-Beat Galvano- 

 meter would give a close approximation to accurate results. In 

 order to reduce polarization, they used as current generator only 

 one Grove cell, and introduced large resistances into the arms of 

 the bridge. Their determinations, however, when compared with 

 those given by more elaborate and accurate methods, were found 

 to be too large, differing from Kohlrausch's, in some cases, by as 

 much as 12 per cent. 



As their experiments were made when they were students and 

 had acquired little experience in experimental work, the discre- 

 pancy may have been due to other causes than the defects of the 

 method. Hence Prof. MacGregor, in order to test the method, 

 made a series of experiments, described in the paper cited above, 

 comparing the resistances of solutions of zinc sulphate as given 

 by this method and by the use of non-polarizable electrodes. To 

 reduce polarization, he also used weak currents and, in addition, 

 electrodes of large area. He found that, if the mere vanishing 

 of the double deflection were taken as the test of adjustment, 

 the resistances determined would be too much greater than their 

 real values to be regarded as a sufficiently close approximation. 

 He noticed, however, that, when he could no longer observe a 

 double deflection, he yet could notice a distinct hesitation at the 



