EXPERIMENTAL STUDY OF HEAT LEAKAGE. 767 



i. Resistance Measuremenis. 



All resistance measurements have been made by means of a Carey- 

 Foster bridge and depend ultimately on Wolff standards certified by 

 the Physikalisch-Technische Reichsanstalt. All contacts on the 

 bridge except those in the galvanometer and battery branches are 

 mercury contacts. The bridge has two slide wires, both of manganin, 

 the finer (0.054 ohm/cm.) being used for measuring the resistance of 

 the high side thermometer, balanced against one or more secondary 

 standard coils. The other wire (0.00445 ohm/cm.) was used in measur- 

 ing the difference of resistance of the two thermometers, balanced 

 against each other. The galvanometer was sufficiently sensitive to 

 indicate a motion of the slider of 0.1 mm. from the point of balance 

 on the coarse wire, so that the order of the precision of the differential 

 measurements is 0.0445 ohm, or about 0°.0022 C. The precision of 

 the measurement of high-side temperature, in so far as it depends on 

 the resistance measurement, is about 0°.03 C. 



IV. EXPERIMENTAL RESULTS ON THE JOULE-THOMSON 



EFFECT. 



L Preliminary Discussion. Methods of Eliminating the 



Effect of Heat-leak. 



Before beginning the consideration of the results obtained with the 

 various types of throttling apparatus which have been described, it 

 seems desirable to outline briefly a few points concerned with the 

 general question of heat leakage and the methods which have been or 

 may be used to eliminate its effect from the directly measured ratio 

 of temperature drop to pressure drop. If we call the last mentioned 

 quantity the 'apparent Joule-Thomson effect' and denote it by m'> 

 we shall have 



(1) • ^T = ^.^p-J^^ 

 and hence 



(2) " =>'-fApC, 



in which fjL is the true Joule-Thomson effect, / is the flow (mass of fluid 



