720 



Mr. A. Campbell on Direct 



portion *, and the arm AD is set to give a balance at some 

 known temperature (say 0° C), the compensating leads, 

 increased in like ratio, being between B and the slider. 



For example, for the thermometer already mentioned 

 [equation (5)] we could make the total resistance of the 

 loop 603'4 ohms, and the resistance sc Q equal to 24*46 ohms. 

 Then each Ol ohm moved by the slider will accurately 

 represent 1° C. Without altering these resistances, the 

 combination could be used with fair accuracy with any other 

 thermometer of nearly similar platinum. For the best work, 

 however, the simplest wa}^ is to have two small resistance- 

 coils adjusted to correspond to each thermometer, one repre- 

 senting .i' , and the other n— sc — 100. 



The arrangement shown in fig. 6 is convenient for the 

 Rectifying Loop, the two adjusted coils being at P and Q. 



Here A and B, the ends of the bridge-arm, are connected to 

 turning-sliders on the ten 10-ohni and ten 1-ohm coils 

 respectively, while the central turning head is a double one. 

 altering the upper and lower resistances by equal and opposite 

 steps of 01 ohm at a time. For clearness I have put this 

 head (M) in the middle of the figure, but in the actual 

 construction the dials should be in their proper order of 

 magnitude, so as to give the temperature at a glance. 



Appendix. 



On Composite Thermocouples. 



For the measurement of moderate or small differences of 

 temperature, up to 150° C, sensitive thermo-junctions of 

 several kinds have been successfully used. For example, in 

 the measurement of the small temperature-differences due 

 to the Peltier Effect, I found Iron-German silver thermopiles 

 very convenient (Proc. Roy. Soc. Edin. 1882). The great 



* A similar procedure is required to give equal bridge-arms. 



