424 Mr. H. M. Tory on the 



3. Resistance Measurements. — The ratio-coils of the bridge 

 were connected permanently in the galvanometer-case, and 

 had a fixed resistance of about 10 ohms each, being made 

 exactly equal. The standard or balancing coil M, against 

 which the resistances of the pyrometer and box were balanced, 

 was also fixed permanently in connexion with the galva- 

 nometer, but in such a way that it could be kept in a test-tube 

 of oil, and its temperature taken after each observation. The 

 compensating leads C were in series with this coil. The 

 adjustable resistance-box B was inserted in the same arm as 

 the pyrometer, and readings were taken by the method of 

 substitution, so as to keep the resistance in circuit constant. 



Before the pyrometer was put in the circuit, the balancing- 

 coil M was adjusted exactly equal to the total resistance of 

 the box B at 20° C, so that when the pyrometer was inserted 

 in the same arm as the box its resistance could be at once 

 determined by reading the resistances of the box that were 

 short-circuited, and taking the deflexion of the galvanometer. 

 In addition to the usual coils in the box, namely, 640, 320, 

 160, 80, 40, 20, 10 units, two others were added, of 5 and 2 

 units respectively ; so that deflexions of the galvanometer 

 greater than 20 scale-divisions were never required to be 

 taken. Ordinarily it is not necessary to adjust the balancing- 

 coil M equal to the resistance of the box, as only the differ- 

 ences R — R° and R' — R° are required. It was convenient, 

 however, in order to get the temperature-coefficient of the 

 wire used. The temperature-coefficient of the platinum- 

 silver box coils was determined to be "000330 per degree, and 

 that of the manganin balancing-coil M to be '000045. All 

 the observations were corrected for both box and coil. 



The resistances of the pyrometers used for comparison, 

 owing to the large differences in the temperature-coefficients 

 of the wires, differed considerably, excepting at the points 

 where they were calculated to agree. The fundamental 

 interval, R' — R°, was usually calculated for one pyrometer to 

 be 100 approximately, and the other to have a nearly equal 

 resistance at 500°, the temperature-coefficients being roughly 

 determined before the pyrometers were made up. The 

 result of this difference of resistance was, that it was found 

 exceedingly difficult to take readings of the pyrometers in the 

 usual way so quickly as to be sure that the temperature was 

 constant to 1/10 of a degree during the process. The pyro- 

 meters were first connected in series, so that one could be 

 immediately short-circuited and the other read after adjusting 

 the resistance in the box. The time required to take a reading 

 in this way, on account of the swing of the needle of the 

 galvanometer, was about three minutes. 



