the Thermal Conductivity of Water. 53 



Calibration of Heating- Coils as Thermometers. 



Attempts to calibrate these in stirred oil within the small 

 range of temperature they were to measure not being very 

 satisfactory, this was done by completely enclosing them in 

 metal water-jackets (J, J', fig. 4) arranged in series and sup- 



Fig. 4. 



plied with a steady stream of cold water. A difference of tem- 

 perature between the jackets was produced by heating the 

 spiral tube S, and the difference was measured by two mercury 

 thermometers immersed in them, and divided to tenths of a 

 degree. The opposing arms of the bridge were the coils b a, a' b' 

 of fig. 3, connected as in the actual conductivity measurements, 

 so that the bridge-wire itself could be calibrated directly into 

 degrees. This wire (A A') is of platinoid, 0*7 mm. thick, in 

 parallel wdth five thick wires of german-silver, and its resist- 

 ance is 0*0003280 ohm per cm. To increase its effective 

 length (necessary for the calibration) two other sets of wires in 

 parallel (B and C, fig. 4) were connected to it in series by 

 mercury commutators P and Q. These commutators were so 

 arranged that A A' could be inserted in the middle or placed 

 at either end of B and C. 



To eliminate thermoelectric effects the contact-maker D 

 was made in the form of a platinoid edge from which a pla- 

 tinoid wire led to a fixed binding-screw R. The connexion 



