D. W. Shea — Calibration of an Electrometer. 205 



tial between the positive pole of any one of these boxes of 

 cells and the ground, with the negative pole to ground, was 

 about forty volts. 



In making the calibration, a battery of small gravity cells 

 was used, with circuit closed through an external resistance of 

 ten thousand ohms. One and the same end of the set of resis- 

 tance coils was to ground in all observations. Points on the 

 box of coils such that one, two, three, etc., thousand ohms were 

 included between them and the point to ground were success- 

 ively connected to one pair of quadrants, while the other pair 

 of quadrants was to ground, and the quadrant connections were 

 alternated in order to get readings on the right and left of the 

 zero, which was placed at the middle point of the scale. The 

 needle was charged by one, two, three, etc., boxes of cells 

 successively, and the number of cells in the gravity battery 

 was decreased, as the charge of the needle was increased, so as to 

 keep the difference of potential between the ends of the set of 

 resistance coils such that the spot of light always remained on 

 the scale. 



It was found that the form of the curve for a given charge of 

 the needle did not long remain constant, and that even the di- 

 rection of curvature changed. The change in the form of the 

 curve was greatest when the charge of the needle was smallest, 

 and it was not until the whole water-battery of four hundred 

 cells was employed that a curve was obtained which changed so 

 little as to admit of accurate work. The change in the curves 

 was most rapid when the temperature of the room was chang- 

 ing rapidly between certain limits. Beyond these limits there 

 was little change in the curves. But for any given tempera- 

 ture between the limits the form of the curve was not con- 

 stant, even though the temperature of the room had been so 

 constant for several hours that all parts of the electrometer 

 could reasonably be supposed to have the temperature of the 

 room. 



The changes in the jprm of the curves for various charges of 

 the needle were followed through the range of temperature at- 

 tainable, at the time, in the room where the electrometer was 

 set up. The following tables will serve to show these changes. 



In all the observations the scale was at a distance of 126 cm 

 from the mirror. 



The curves shown in the plate were plotted by taking the 

 potentials of the quadrants as abscissas and the deflections of 

 the needle in degrees as ordinates. Fig. 1 shows curves when 

 charge of needle was forty volts; fig. 2, when charge was 

 eighty volts ; fig. 3, when charge was one hundred and twenty 



