202 



Profs. W. E. Ayrton and John Perry. [Mar. 21, 



electrometer are insulated from one another, any slight change in the 

 charge of the instrument causes a motion of the needle. 



The hand must be trained to insulate the plates from one another. 

 After three months' practice our insulations are so perfect that there is 

 seldom even a tremor on removing the plug, whereas in the beginning 

 of our experiments we often had in succession ten bad attempts at 

 insulation. A good insulation depends on the plug leaving both sides 

 of its seat simultaneously. This key gave better results than any of 

 the other arrangements at our disposal. 



We at first thought, from considering the change of capacity in the 

 quadrants (as the needle moved), that our measurements would be 

 more accurate when very small. This is not the case, but it led us to 

 try a method of observation which may suggest itself. We brought 

 the spot of light to zero in every experiment by acting inductively on 

 the insulated plates A and B by means of small plates placed over 

 them charged to a measurable difference of potential. We may 

 employ this method modified at some future time. 



The suspension fibres of the needle were tolerably near one another, 

 and the trap- door gauge was adjusted for as great a charge as it was 

 convenient to give to the jar, and as it was nearly unstable it indicated 

 extremely small changes of charge. The relative positions of the 

 mirror G, the lamp, lens, and scale* are shown in the lower right-hand 

 part of the Plate. The lens forms a virtual image of the wire of the 

 lamp-slit at 60 centims. from the mirror G', and the scale J*, which is 

 two metres from the electrometer, receives the image formed by the 

 mirror. We had supposed that no better definition of the slit could 

 be obtained by the use of a lens, but in reality great benefit was derived 

 from its use, our readings being much more accurate than they would 

 have been had the lamp-slit, without the lens, been put at 60 centims. 

 from the mirror. We consider that our readings (as far as the electro- 

 meter itself is concerned) are fairly accurate to the one two-thousandth 

 part of a volt. 



III. 



To find the relative values of deflections of the electrometer needle, 

 two brass plates were placed under A and B at the distance of 6'4 mm. 

 We were able to give to the plates a difference of potentials equal to 

 the electromotive force of a Daniell, or to any fraction of this. Reversal 

 is made by means of a key without any actual changes in the positions 

 of the plates. 



* Not represented in the Plate. 



