Barus — Displacement Interferometer . 79 



12. Experiments with the disk electrometer. — The experi- 

 ments "were begun with but a small distance, D, between the 

 plates of the electrometer. In such a case the disk can not carry- 

 very high potential before it is drawn across to the condenser 

 plates fand short circuited. In fact, though the mica disks 

 may be made very light, the annoyance of short-circuiting is 

 correspondingly increased and the great advantage of sensitive- 

 ness cannot for this reason be realized unless the idiostatic 

 method is used. 



The disk was charged with the storage battery to 101 volts 

 and a number of small potentials measured at the plates. The 

 computed and uncorrected results were about 3 times as large 

 as the true voltages. The sensitiveness is here surprisingly 

 large, almost 8 drum parts or 2XlO~ 3 centimeter per volt, i.e., 

 TO rings per volt. The distribution of results is, nevertheless, 

 linear and proportional to the true voltages. To account for 

 this result by aid of the equation 



V *~ IV, r* 



A/2>)' 



is not impossible, if the asymmetry is such as to make the 

 terms in A/D essentially negative. Yet in the case of a large 

 value of A the results are so lacking in probability that a want 

 of uniformity or variation in the field is more liable to be in 

 question. 



On spacing the plates further, D=1'04: centimeters, and 

 carefully adjusting the disk, it carried a charge equivalent to 



^=250 volts easily, so long as V 3 did not exceed about 20 

 volts. 



A new and lighter disk was thereupon introduced, weighing 

 but *607 gram. As it failed to carry a charge quite as large as 



F 1= 250 volts, the only available smaller one, F 1 =104 , 5 volts, 

 was given to it. The uncorrected results for V 3 are here 

 much nearer the true values, i. e., about 1*2 times too large, 

 than were those of the preceding case, so that the correction 

 for asymmetry is probably applicable. The mean trend of the 

 locus is through zero. Sensitiveness is naturally much lower 

 than above, showing AN = 5 X 10"" centimeter per volt. The 

 experiment indicates well how markedly dependent the con- 

 stants of the instrument are on the position of the disk. 



The same disk was now weighted with a rider, making the 

 total mass 1*086 grams, with the object of charging it to 250 

 volts. The uncorrected results, while maintaining proportion- 

 ality, were about 2 times too large. 



The data obtained idiostatically, however, where the deflec- 

 tions are essentially small, fig. 7a, show an excellent agree- 

 ment with the true voltages. Discrepancies are as liable to be 



