02 DRS. J. A. HARKER AND P. CHAPPUIS ON A 



might be able to fix upon tbe best magnitude of the current to be employed for the 

 thermometric measurements. 



For this purpose we made a number of determinations of the api>urent resistance 

 of thermometer K.8 in ice using different battery-currents. 



For these a curve was constructed showing the increase in apparent resistance of 

 the wire with increasing energy absorbed in the coil, and a value calculated I'm what 

 the resistance would be, if the current through it (and, consequently, the heating 

 effect) were vanishingly small. Our measurements conclusively showed that, within 

 the limits of accurate experiment, the heating effect was directly proportional to the 

 watts in the wire, and that the heating per milliwatt for K.8 was about 0'OOG. 



In some of the earlier experiments, made before the heating effect was investigated, 

 we employed a total resistance in the battery circuit of 150 ohms for measurements 

 at 0; the heating due to the current in this case being 0'024. For all the subse- 

 quent experiments, however, by increasing the external resistance the heating was 

 diminished to 0'014 in ice. 



Although we only made direct determinations of the magnitude of the heating 

 effect at 0, we have assumed, in the absence of further data, that for a thermometer 

 coil the heating due to a given amount of energy expended in it is the same at all 

 temperatures. As this is only approximate, some of our results may subsequently 

 require small modifications; but the value we give later for the boiling-point of 

 sulphur would not be affected, as it is expressed on the scale of the gas thermometer, 

 the platinum thermometers being only used as an intermediary. 



We calculated a table for each of the principal platinum thermometers, giving the 

 resistance to be inserted in the external circuit for different temperatures.* In the 

 example of a platinum temperature calculation given later, this number is referred to 

 as the battery resistance " B.R. = 317 ohms." 



XXIII. DETERMINATION OP THE CENTRE OF THE BRIDGE. 



The index-error of the scale was determined from time to time during the work by 

 reducing the resistances between dC^ and P,P 2 (fig. 1, p. 41) to zero, fixing all the 

 contact pieces firmly in position, and determining the point of balance of the bridge. 

 Should this not fall strictly at the centre of the scale, a correction for " bridge- 

 centre " is applied in each resistance measurement. 



* It was afterwards found that the formula used to calculate the table referred to was not strictly 

 correct, but made the external resistance at high temperatures greater than it should have been. As, 

 however, the total current heating at was only - 014, and less than this at higher temperatures, the 

 correction to be applied to the results, on account of the adoption of wrong external resistances, is 

 probably well within the limits of experimental error, especially seeing that the error introduced is 

 already partly compensated by its effect on the fundamental intervals as well us on the platinum tempera- 

 tures found. 



