SCIENCE AND PRACTICE. 355 



measurement. The arrangements of this system have been 

 already explained. 



Having found the copper resistance, r, for the whole con- 

 ductor of the length /, in knots, at a temperature of t Gels., 

 it remains still to find the resistance per knot, at some fixed 

 degree of temperature, from these data. The resistance r t of 

 the conductor, per knot, at the temperature at which the 

 measurement was made, is found by simply dividing the 

 resistance, r, by the length, I, 



. 



When r t has to be reduced to some other temperature, say r 

 Cels., in order to compare it with the resistance of some 

 standard knot of wire at that temperature, the quotient is 

 multiplied by 1 + 0,0003765 (r f), and the reduced value, 

 r r , becomes 



r T = JLfl + 0, 0003765 (r t)\ 



This comparison contains, however, an element which may 

 lead to error ; for, although r, I, and t may be known exactly, 

 yet the copper may be of different sectional area to that of 

 the standard, and its resistance appear too great or too little, 

 when, in reality, its different dimensions are in fault, and its 

 conducting power unexceptionable. 



To avoid this, it is preferred to reduce the conducting power 

 of the material from the data already obtained, and the weight, 

 W, of the / knots of copper strand. 



Having the value of W in Ibs., I in knots, and r in units, 

 the conducting power c, compared with that of pure mercury 

 at 0C., is, by the formula given before, 



c = 



in which a, |3> and a are constants : 



a is the length of a knot in centimetres =185200, 

 ]3 is the weight of a Ib. in grammes =453,6. 

 a is the specific gravity of drawn copper =8,899. 



A A 2 



