the Thermal Conductivity oj Copper. 127 



lower end of the outflow tube 0, the current of water can be 

 easily varied. With the apparatus in use in my own class 

 the difference of level is never greater than 1 metre. 



The mass of water which flows through per second is 

 readily found by observing the time taken to fill a 250 c.c. 

 flask. The water is. of course, allowed to flow continuously, 

 and the time is counted from the instant when the flask is 

 brought up to receive the water. 



If, when the conditions have become steady, the tempera- 

 ture of the water in H be 4 and that of the water in G 

 lie 6 Z . and if M grammes of water flow through in t seconds, 

 then heat is taken away by convection from the end Y at 

 the rate of 



M(fl 3 -ftQ 



t 



thermal units per second. 



§ 4. To determine the temperature-gradient along the part 

 of the rod between the steam-jacket and the cooling-tube, 

 we must know the temperatures at two points on the bar 

 separated by a measured interval. These temperatures are 

 found by means of the thermometers A, B, which fit loosely 

 into holes bored in the stout copper rods E, F nearly to the 

 lower ends of the rods *. The lower ends of these rods are 

 cut away so as to fit into saw-cuts made in the bar at 

 U and V, and the rods are secured to the bar by solder or, 

 preferably, by brazing. If no heat escapes from the copper 

 tubes, the thermometers placed within them will indicate the 

 temperatures of the bars at U and V. The distance between 

 U and V is about 10 cm. in my apparatus. The isothermal 

 surface- passing through U and V are supposed to be plane. 

 To secure this approximately, the distances of U and V from 

 the -team-jacket and the cooling-tube respectively are not 

 less than the diameter of the bar. 



If 6 X be the temperature at U and 2 that at V, when the 

 conditions have become steady, and if the cross-section of 

 the bar be A scp.iare cm. and the distance between the 

 centres <>f the saw-cuts at U and V be I cm., then the rate 

 at which heat flows along the bar is 



i 



* The temperatures at U and V could be determined electrically by 

 measuring the resistances of two coils, each formed of a few turns of 

 ailk-covered platinum wire wound round the bar at (' and V ; or a pair 

 of thermo-junctions, insulated from the bar, inighl be used, 



