90 



MR. K. K. SMITH OK THE CONSTRUCTION OF 



we have P + P r = 1 '0184 ohms approximately, so that P r = 0-0184 ohm, a value sufficiently accurate, 

 easily evaluated, and determined immediately after the observation for P. 



The coils a. and /J were now removed ; the battery lead at the junction of R and P r replaced to the 

 junction of P and P r , and the galvanometer connection made at tlQ, ft s shown in fig. 12. Under such 



W 



W 



Fig. 12. 



conditions QR/S = 1 -000218 + P r /1000 = 1-000236 ohms. Balance was obtained when W = 7055 ohms. 

 Hence P + d = 1-000236 (1 + 1000/7055) = 1-1419 ohms approximately. Thus d = 0-142 ohm, 

 a value certainly correct to 1 per cent. 



P. was measured by substituting it for d. Under such circumstances W = 63690 ohms to obtain 

 a balance. Hence P + P = 1-000236(1 + 1000/63690) = 1-0160 ohms approximately, and thus 

 P. = 0-01 60 ohm. 



P may now be accurately evaluated. Substituting the found values of P r , P. ad d in the expression 

 for P, we have 



P = 1-000218/1-000134 + O'0184/IOOO - x 0-142 (0' 0160/100 - O'0184/IOOO - 0'00004). 

 Thus resistance of standard tube X = 1 00009s ohms. 



Such was the general method adopted for the measurement of resistance by the 

 Kelvin double bridge. As will be readily realised, the whole of the observations 

 occupied but a very few minutes. 



A correction has to be applied for the platinum foil introduced. The thickness of 

 this was 0'002 ceutim , and although at the junction the cross-section was practically 

 infinite, the limitation of the stream lines results in the resistance of the junction being 

 equal to that of a mercury column of length 0'002 centirn., and of the same cross- 

 section as the united ends of the tubes. In consequence, the resistance as measured 

 from the mid-points of the junctions will be greater than the calculated resistance in the 

 ratio of Lj + 0'002 to L,, where L, is the length of the tube in centimetres. In the 

 case of X, L, is approximately 65 "6 ceutims., so that the measured resistance must be 



