390 RESISTANCE OF ELECTROLYTES — MCKITTRICK. 



As stated above, Professor MacGregor's error in the measure- 

 ments of high resistances varied from 0.1 to 0.4 per cent. With 

 low resistances, however, his error was as great as 6 or 7 per cent. 

 The first three entries of the above table show, then, that,, 

 with a comparatively weak battery as current generator, the 

 errors in the resistances determined by employing the vanishing 

 of the double deflection as test of adjustment, are comparatively 

 large, considerably larger than they would be according to Pro- 

 fessor MacGregor's experience, if the vanishing of the hesitation 

 were taken as the test. The last seven show that, with a strong 

 current, the employment of the vanishing of the double deflection 

 as test of adjustment gives results which, in the case of high 

 resistances, are considerably more accurate, and in the case of 

 low resistances very much more accurate than those given by a 

 weak battery with the vanishing of the hesitation as test. 



It would thus appear that the above modifications render this 

 method capable of measuring all resistances with increased 

 accuracy and low resistances with greatly increased accuracy, 

 while they render it also capable of application without any pre- 

 vious training of the eye, and diminish very materially the time 

 required for its application. The error in the measurement of 

 high resistances is still greater than that of Kohlrausch's method, 

 viz., 0.05 percent. The only mirror, however, which was avail- 

 able for my experiments, though it weighed but 0.035 grms, was 

 much heavier than some which I believe are now manufactured. 

 With a lighter mirror the inevitable error would doubtless h& 

 still further diminished. 



