1884.] measurement of the capacity of a condenser. 177 



(iii) the observation can be most conveniently made ; 

 (iv) the calculations can be most conveniently effected. 



Under the first head we notice that (n) must not be so great 

 that the time of contact between P, and either 8 or R is too small 

 to allow of the condenser being fully charged or discharged. 



Again we must be careful that the resistances BSP when P 

 and S are in contact and DRP when P and R are in contact are 

 negligible. 



That this should be the case it is necessary that the contact 

 resistances between P and R and between P and S should be very 

 small. If P strikes R and S hard a better contact will be made 

 than if it strikes gently, and thus we will get better results, the 

 stiffer our spring and the greater the power of our driving battery, 

 both for the above reason, and because in that case we diminish 

 the time during which P is in contact with neither 8 nor R, and 

 thus we might increase (n) and still have our condenser fully 

 charged and discharged at each contact. 



Again we must be careful that (n) is not so small as to allow 

 the phenomena of electric absorption and residual discharge to be 

 produced, for in that case we should get too large a value for C. 



Under the second head we notice that the greater the electro- 

 motive force of our battery, the more sensitive our galvanometer, 

 and the greater the value of (ri), the more delicate the adjustment 

 necessary to produce a balance, and consequently the more accurate 

 the method. 



Under the third head we notice that, since the observations can 

 only be conveniently taken when the needle is at rest, it should be 

 made as steady as possible. We must particularly guard against 

 its hanging in too weak a field of force, for the effect of the 

 transient currents through the galvanometer is to diminish the 

 stability of the needle, as has been proved by Lord Rayleigh. 



Under the fourth head we notice that, our formula being : 



q = S!l Ti _ i " 



cd\_ (a + c + g){a + b + d) 



ab 

 1 + 



1+ <* 



d(a + c+; 



c (a + b + d) 



and g being large and a and b being small compared with the other 

 resistances, it becomes 



Hence we see that it is more convenient that (c) should be the 

 adjustable resistance, for then the factor (l — ^ j is a constant one. 

 The advantages of this method have been exhaustively discussed 



