Method of Determining Thermal Conductivity. 593 



by turning a lap the steam would leave the jacket J\ by 

 the spiral as well as by the main tap at the bottom of the 

 jacket. 



Both tin jackets were surrounded with cotton-wool which 

 was also bound carefully round the experimental glass tube, 

 the latter being also shielded from the jackets by a thick- 

 walled asbestos enclosure. A thermometer hung in this 

 enclosure gave the temperature of the air around the ex- 

 perimental tube. Fig. 1 (Plate VII ) gives a general view of 

 the apparatus. 



(5) Method of Experiment. 



The apparatus having been filled with pure mercury, ice 

 was placed in the jacket J 2 and steam was injected into 3 X . 

 When the top of the glass tube T was warm enough and 

 there was no danger of breaking it, steam was allowed to 

 circulate through the spiral coil. 



In about 3 hours the steady state was obtained and the 

 temperatures were noted. The head H was then raised till 

 a suitable flow up the tube T and out of the capillary re- 

 sulted. The weight flowing out in a definite time — usually 

 15 minutes — was taken, and when in two or three more 

 hours' time the new steady state was attained this flow was 

 remarkably constant, the weighings not differing by more 

 than *5 gm. or *6 gm. in 300 gms. In this connexion it is 

 interesting to note that there is an automatic tendency to 

 maintain constancy in the flow ; for any slight increase of 

 head — due for example to vibration — temporarily causing a 

 faster flow would introduce a higher counterbalancing 

 pressure in the experimental tube the mercury in which on 

 the whole would become slightly colder and heavier. Thus 

 the equilibrium is stable. 



Under the new steady state the temperatures are again 

 noted as well as the temperature of the enclosure. All the 

 readings were corrected so as to give the actual excesses 

 above the temperature of the surroundings. To do this 

 readings were taken when all the thermometers were in silu 

 and at the same temperature. They then showed slight 

 differences due to various causes, one of which was the hydro- 

 static pressure of the mercury. 



It was found that if the ordinary stationary state was not 

 required the steady state under How could be more quickly 

 obtained by starting the flow, when, in the variable state, the 

 top thermometer reached a temperature near that desired. 



If a flow from hot to cold was desired the head H and 

 capillary C were interchanged. 



Phil. Mag. S. 6. Vol. 19. No. 112. April 1910. 2 Q 



