o52 Ji'ew Form of Platinum, Resistance- Thermometer. 



the other end the two tubes are o£ unequal length, with the 

 inner projecting several centimetres beyond the outer. The 

 copper leads are bent over the outer tube and bound securely. 

 The two tubes are closed at this point by shellac or marine 

 glue, or, where temperatures higher than that of Ijoiling water 

 are to be measured, the space may be closed by binding around 

 a rubber strip. 



The length and diameter of the bulb will depend on the 

 fineness of the thread and o£ the wire, and on the length of 

 wire required to give the requisite resistance. As a rule_, 

 6-mil wire is suitable for most thermometers ; but we have 

 used 4-mil wire with very satisfactory results. The thermo- 

 meter has a small heat capacity, and as the liquid passes 

 through the centre of the bulb as well as over the outside, it 

 is exceedingly quick reading. 



It is particularly suited for measuring steam-points. When 

 it is placed in a hypsometer the steam passes out through the 

 thermometer and keeps the entire length of stem at the steam 

 temperature. It has been shown by one of us"^ that the length 

 of stem exposed at the high temperature is an important con- 

 sideration for very accurate measurement. The use of copper 

 wire as leads, which is much to be recommended, has one 

 disadvantage in having a high conducting-power for heat. 

 With the mica -frame thermometer, unless a great length of 

 stem is immersed in the steam-jacket, the conduction of heat 

 from the air-space around the bulb by the leads produces a small 

 error. 



With the cored thermometer no such stem-correction is 

 needed, since the leads are heated for their entire length by 

 the flowing steam. 



In taking an ice-point the water around the ice in which 

 the bulb is immersed is drawn up through the thermometer 

 and discharged again into the mixture. This can be easily 

 arranged with a small laboratory w^ater-pump. 



The sulphur-point is more difficult to obtain ; but, as a 

 rule, for accurate temperature measurement it is not necessary 

 to go as high as that. If it is desired to determine the con- 

 stant of the wire, where this is not known, a mica-frame 

 thermometer should be tested which has been made from a 

 piece of the same sample of wire. Temperature measurements 

 by the platinum thermometer are now^ so universal that it is 

 easy to obtain the pure wire with known constants. 



We have tested a number of these cored thermometers of 

 different resistances, and find them exceedingly quick and 

 accurate. 



McGill University, June, 1903. 



* Phil. Trans, vol. cxcix. p. 195 (1902). 



