370 Mr Wilson, On the Measurement of the Earth- Air 



the compensator readings constructed. It is not however necessary 

 to have available a whole series of cells, the whole calibration may 

 in fact be made with a single cell by slightly modifying the 

 method. The compensator reading corresponding to a charge 

 equal to that induced on the test-plate by a single cell is deter- 

 mined as before. The compensator however, instead of being 

 drawn back after this observation to its standard position, is left 

 untouched ; the test-plate and condenser plate are earthed and the 

 operation described above repeated, the only difference being that 

 the displacement of the compensator is from the point where it 

 was left at the end of the previous determination. This process is 

 repeated till the whole scale of the compensator has been traversed ; 

 at each displacement of the compensator an additional charge is 

 drawn into the compensator rod equal to that induced on the test- 

 plate when it is at zero potential and the condenser plate is con- 

 nected to the positive terminal of the cell. A combination of the 

 two methods was in practice found convenient, the first method 

 being used till six cells had been inserted, and the second method 

 with steps of six cells at a time being used for the rest of the 

 calibration. 



For a given compensator reading the charge held on the 

 compensator rod (when kept at zero potential) is proportional 

 to the potential of the compensator tube. The form of the 

 calibration curve is thus the same whatever the potential of 

 the compensator tube. When once a calibration curve has been 

 constructed for one value of the potential of the compensator tube 

 (chosen for its convenience in giving a suitable number of points 

 on the curve with the cells available) we may adjust the charge of 

 the Leyden jar attached to the compensator to suit the magnitude 

 of the charge it is desired to measure. The ordinates of the cali- 

 bration will merely have to be multiplied by a different factor, 

 which can readily be found by comparing the displacements of the 

 compensator required to raise the potential of the gold-leaf system 

 by a definite amount as indicated by the displacement of the gold- 

 leaf; the measurement in each case being made with the ordinary 

 cover over the test-plate so that the capacity of the system may 

 be the same. 



The calibration curve for deducing from the compensator 

 reading the charge which it was desired to measure was of the 

 form shown in Fig. 3, which represents the results of observations 

 made on April 20. The unit of charge of the table and curve is 

 that held on the upper surface of the test-plate kept at zero 

 potential when the condenser plate is at a potential of 2 volts. 

 The reading 30*2 is that of the compensator when drawn fully 

 out, the reading when it is pushed in against the stops being 0. It 

 will be noticed that the displacement corresponding to a given 



