Nickel) and Nickel-Steel at Various Temperatures. 81 



overcome by using a platinum wire to compensate the heating 

 effect on the iron, the same wire serving for this as for the 

 measurement of temperature. The platinum wire was con- 

 nected in the arm of the bridge next the iron, instead of the 

 usual balancing resistance. As the temperature-coefficient 

 of platinum is much smaller than that of iron, the compen- 

 sation thus effected could only be partial. By placing, 

 however, a resistance in parallel with the iron wire, and 

 another in series with it in the arm of the bridge, the effect 

 of the heating of the iron on the total resistance of the arm 

 could be made as small as we pleased, and by adjusting these 

 resistances it could be made exactly equal to the change in 

 the other arm containing the platinum. 



The condition for balance under change of temperature is 



IS^ = K'P.A 

 (I + S)- 2- K B ' 



I = resistance of the iron, 

 P = ,, „ platinum, 



K = temp, coefficient of iron, 

 K' = ,. „ of platinum, 



S = shunt on iron, 

 A, B = resistance of auxiliary coils. 



In thus diminishing the effect of temperature on the resistance 

 of the arm we are also diminishing the effect of the magnetic 

 change, and the bridge is to this extent less sensitive, but 

 this disadvantage is far more than counterbalanced by the 

 greater accuracy with which the readings can be taken when 

 this arrangement is adopted. The exact formula for reduc- 

 tion of the readings in this case is : 



, y ,( I + S) 3 P.A ,,K 



da = da c ,„ ' — ii - = da „,, 



O" 1 -D IV 



where da' is the; actual reading and da the corresponding 

 reading if there were no extra resistances. 



r> 



Results. 



Nickel. — The results obtained for nickel are shown by the 

 curves in figs. 2, 3, 4. It will be seen that as the tempera- 

 ture is increased the change of resistance diminishes, the 

 diminution being proportionately greater at high fields, so 

 that the highest pair, of the curve becomes a straight line 

 parallel to the axis of PL 



Phil. Mag. S. 6. Vol. 9. No. 49. Jan. 1905. G 



