98 



Profs. Dewar and Fleming on the 



all 



r temperature-measurements in platinum tern-* 

 peratures, as defined by one standard platinum thermometer ; 

 and all the measurements in this present thermo-electric 

 research are given in terms of the same standard platinum 

 thermometer, which we shall hereafter mention as our 

 standard platinum thermometer P x . 



If at any time it becomes necessary, these temperatures 

 can be reduced to other scales of measurement. 



For several reasons it was necessary to construct a platinum 

 thermometer of special form for this thermo-electric work ; 

 and this was made as follows : — A small cylinder or tube of 

 copper 3' 2 centim. long, and having a longitudinal* hole 

 through it '64 centim. in diameter, was wrapped round with 

 two other half-cylinders of copper. These half-cylinders 

 had bevelled edges. One half-cylinder was soldered to the 

 central tube, and the other half-cylinder was insulated from 

 the tube and from the first half- cylinder by a slip of mica. 

 The whole was formed into a fairly true cylinder on the 

 outside. Two wires of platinoid No. 18 s.w.G. and 30 

 centim. long were then soldered to the two insulated half- 

 cylinders. These platinoid wires were kept apart by distance- 

 pieces of ebonite. A thin sheet of mica was then wrapped 

 round the cylinder, and over this was wound 100 centim. 

 of pure annealed silk-covered platinum wire. One end of 

 this wire was soldered to one copper half-cylinder, and the 

 other end to the second insulated half. It will thus be 

 seen that we had a platinum wire connecting the ends of the 

 platinoid wires but wound on a highly conducting copper 

 cylinder, the over-all diameter of the bobbin being about a 

 centimetre and a quarter. This wire constituted the platinum 

 thermometer which will hereafter be spoken of as the working- 

 thermometer. A section and elevation of the thermometer is 

 given in PL III. fig. 1. To the ends of the platinoid wires 

 were soldered long stranded and well-insulated flexible copper 

 connecting leads. A duplicate was then constructed in 

 exactly the same manner, save that it had no platinum wire 

 but had the two copper half-cylinders connected metallically 

 together. The object of constructing this duplicate was to be 

 able to eliminate from all measurements the resistance of the 

 platinoid and copper wires forming the connectors of the 

 platinum thermometer wire. 



Our platinum thermometer consisted therefore of a platinum 

 wire and of connecting conductors partly of platinoid and 

 partly of stranded copper wire, and a blank or duplicate com- 

 posed of a loop of an identical platinoid and copper wire. The 

 reason for interposing the platinoid wire was as follows. As 

 will Le seen presently, our method of p rocedure to obtain 



