Direct -reading " Platinum Thermometer, 71 



six hours' boiling in sulphur vapour, this copper rod became 

 circular in section ; it had a central perforation, as could be 

 proved by passing a wire into it. The colour of the surface 

 was a deep " beetle " blue. 



In order to show the amount of expansion which the copper 

 experiences during the change to sulphide, a second piece of 

 copper, of square section, was sawn out of the commutator-bar, 

 half the length being turned down to a circular section. This 

 was put into sulphur vapour for six hours, with the result that 

 the diameter of what had been the square part had increased 

 from 215 to 386 mils ; the circular part had increased from 

 120 to 232 mils ; and a third piece, of ordinary drawn copper 

 wire, increased from 31 to 77 mils. During the transforma- 

 tion from copper to sulphide the diameter in all cases is ap- 

 proximately doubled. The length is only slightly increased; 

 all sharp edges disappear ; and there is a general tendency 

 towards the circular section. The fracture is crystalline, 

 resembling that of an aerolite. 



A piece of Delta metal submitted to the same test came out 

 of the sulphur vapour without appreciable change. It was 

 discoloured but otherwise unaltered, and I propose to adopt 

 this metal in future for the electrodes of platinum thermo- 

 meters. 



Summary. 



The foregoing results indicate that platinum-coil thermo- 

 metry may be reduced to a simple operation, and that the 

 simplified method is sufficiently precise for general work. 

 The exact conditions which determine the numerical limits of 

 accuracy, generally, cannot be very concisely stated. The 

 possible causes of error are well known, and with a little care 

 may be nearly eliminated. The first precaution is to main- 

 tain the bridge-coils at some constant temperature just above 

 the maximum air-temperature. For this purpose the oil-bath 

 and glow-lamp will be found satisfactory. 



If the temperature of the slide- wire is likely to vary con- 

 siderably, r may be redetermined ; or a correction- factor 

 may be applied. This will seldom be necessary ; no such 

 correction was used in the above tests. 



There must be some easy arrangement for calibrating the 

 slide-wire. The auxiliary-coil test is expeditious. The wire 

 now supplied by good makers is sufficiently uniform ; and, 

 except where great precision is essential, it will not require 

 to be corrected for differences of diameter ; it should be of 

 platinum-silver. 



