272 Prof. G. Wiedemann on the Methods 



Since, moreover, the sources of error mentioned above may 

 undoubtedly cause considerable deviations, such a one-sided 

 accuracy cannot be a guarantee for the exactitude of the final 

 results *. 



On account of the much larger number of determinations 

 which this method requires, I consider it much less suitable 

 for the final construction of a definite resistance than the 

 fourth and first of Weber's methods, of which again I give 

 the preference to the latter. 



After a conductor A of known absolute resistance has been 

 constructed, by the one method or the other, it has to be com- 

 pared with the resistance B of a mercury column of known 

 dimensions in order to determine the length of such a column 

 of 1 square millimetre section which possesses 1 ohm resistance. 



If we employ Weber's first method, we may introduce the 

 mercury column B directly between the inductor and multiplier, 

 and calculate the resistance of B from the decrease of inten- 

 sity ; otherwise the resistances A and B may be compared by 

 means of the differential galvanometer, or more conveniently 

 by means of the Wheatstone's bridge, according to the accurate 

 method already employed by W. Siemens in constructing 

 copies of his mercury unit. Since we already know the length 

 of mercury corresponding to 1 ohm very nearly, we take care 

 to arrange the resistance B so as to be as nearly as possible 

 equal to the resistance A. 



It is highly important, after calibration of the measuring- 

 wire, to secure perfect contacts by using freshly-amalgamated 

 thick copper connecting-pieces and cups full of pure mercuty, 

 the resistance of which is previously determined by including 

 several in a circuit. Plug-contacts are not sufficiently accu- 

 rate. Also the temperature must be carefully kept constant 

 throughout at the right point, as was the case in the construc- 

 tion of Siemens's copies. 



* If we employ in this and other methods a coil wound uniformly on a 

 closed ring, a neutral solenoid, as inductor, hut as induced coil a coil 

 surrounding the former at one point, the conditions are so far simpler 

 that the current due to opening or closing the current in the solenoid, only 

 the linear dimensions of the solenoid, its interior and exterior diameters 

 cl and d v the number of coils of the induction-coil and of the solenoid, or in 

 the case of a solenoid of rectangular section the height a, are concerned 

 (compare Roiti, Atti di Torino, vol. xvii., 30th April 1882). Nevertheless 

 the employment of such a solenoid is attended with the great difficulty 

 of winding the wire exactly upon it, and of measuring its dimensions with 

 sufficient accuracy. The latter is the more difficult, since the section of 

 the solenoid can only be taken relatively small, and the formula contains 



the expressions s/d- Vd x for a solenoid with circular section, a log -4 

 for a solenoid with rectangular section. 



