1805.] On the Calibration of the Capillary Electrometer, 



21 



With, a capillary of different form, this relation will no longer 

 hold. The velocity of the meniscus may be different, for the same 

 acting P.D., at the points p and q. But the value of the subnormal at 

 q is at once given by the formula 



§L±^ . k . = x'. 



The constant Jc is most conveniently determined by measuring the 

 subnormal at the point p on a third photograph in which a normal 

 excursion of known value starts from a zero point within the field of 

 view. 



Peactice of the Method. 



It may be convenient to describe briefly the entire process of 

 calibrating an electrometer. In order to ensure that the tube shall 

 always occupy the same position, the image of the tip of the capil- 

 lary is adjusted to a mark on the back of the exposing shutter. It 

 is of course not perfectly in focus, but is sufficiently sharp for the 

 purpose. This mark serves for all experiments with the same 

 capillary. 



Calibration of the Scale Readings. 



A millimetre scale on glass is fixed to the focussing screen so that 

 its zero coincides with the upper limit of the slit. The slit and the 

 cylindrical lens are then removed, and the image of the meniscus 

 adjusted to zero on the scale. 



The E.M.IYs necessary to bring the mercury down to 5, 10, 15, &c, 

 mm. on the glass scale are then measured with a potentiometer, care 

 being taken to short-circuit between each observation lest the zero 

 should alter. With very delicate instruments, or if the pressure 

 tubes are not air-tight, it is sometimes easier to measure the length 

 of the excursion produced by a small constant P.D. in various parts 

 of the tube, and obtain the value of the scale-readings by calcu- 

 lation. 



Calibration of the Subnormal. 



Two normal excursions in opposite directions are photographed. 

 Each must be large enough to traverse the full length of the slit 

 during the passage of the plate, but not much more, otherwise the 

 curves will be too steep for accurate measurement. 



A non-inductive, non-polarisable resistance of at least 10,000 ohms 

 must be included in the circuit. The subnormals to these curves 

 are then measured at points 5, 10, 15, &c, mms. from the reference 

 circle. 



As an illustration I give the results obtained with an electrometer 

 which had been in use for a long time. The tube had been cali- 



