1891.] Quadrant Electrometers. 57 



insulated stems are made of glass, and are under cover, protected 

 from dust and damp; pressure contact between the electrodes and the 

 quadrants is replaced by spirals of fine wire screwed to the quadrants 

 and to the electrodes; the needle, quadrants, and guard tube are 

 so shaped that, in whatever symmetrical position the quadrants be 

 placed, the deflection produced by a given P.D. between the quadrants 

 is directly proportional to the potential of the needle, and further, 

 this improved electrometer is at least ten times as sensitive as our 

 specimen of the White pattern when the instruments are adjusted to 

 be in equally trustworthy condition as regards definiteness of the zero 

 and of the deflected position of the spot of light. 



Next follows an account of some experiments made by us on a 

 White electrometer, the needle of which Mr. Boys had suspended 

 with a single quartz fibre. Although this instrument was in excellent 

 condition as regards definiteness, &c., the raising of the potential of 

 the needle to only 400 volts was sufficient to show that the sensibility 

 was not proportional to the needle's potential. 



Lastly, for the purpose of obtaining conclusive evidence as to 

 whether our idea was correct about the connecting pieces of the 

 guard tube in the White electrometer causing the sensibility of this 

 instrument to be in many cases actually less when the needle had a 

 high potential than when it had a low, we had constructed a little 

 collar with two legs. This collar could be clamped to the upper 

 portion of the guard tube of the improved electrometer with the legs 

 projecting down into the quadrants on each side of the needle, and 

 experiments showed that when this collar was attached to the guard 

 tube the improved electrometer, although not a bifilar instrument, 

 became as bad as the White pattern. For while before the attach- 

 ment of this collar the sensibility increased proportionately to the 

 potential of the needle, after the collar was attached the sensibility 

 first increased and then diminished again as the potential of the 

 needle increased, and with the same adjustment of the quadrants, 

 controlling magnets, &c., and with the needle charged to a potential 

 of 1300 volts, the mere attachment of this little collar reduced the 

 sensibility to one quarter. 



The paper concludes with a sketch of the mathematical investiga- 

 tion that we carried out, and it is explained that by taking into 

 account the electrical action of the connecting pieces of the guard 

 tube of the White electrometer, the diminution in this action as the 

 quadrants are pulled out, the alteration produced by the tilting of 

 the needle at high potentials on the magnitude of this electrical 

 action as well as on the rate of variation, per radian deflection of the 

 needle, of the coefficient of induction between the insulated pair of 

 quadrants and the needle, an expression was obtained for the deflec- 

 tion of the needle in terms of its potential and che P.D. between the 



