ABSOLUTE THERMAL CONDUCTIVITY OF NICKEL. 



377 



the temperature of the cold water (at 0) going to the bar ; 3°, the time at which the 

 empty flask, Q, was put to catch the overflowing water ; 4°, the temperature of the water 

 leaving the bar (at P) was read every half-miaute while the flask was filling ; 5°, the 

 time the flask was exactly filled to the fiducial mark ; 6°, the temperature of the water 

 entering the cap at ; 7°, the thermometers in the bar. All these readings varied little 

 in the course of one evening, and the rate at which heat was given out at the end of 

 the bar varied within 2 per cent. The following table gives the readings (uncorrected) 

 taken on 6th January 1898. 



Average 



Temperature 



of Water. 



Average Rate of 



Flow of Heat in 



Calories per 



Second. 



Temperature 

 of Air. 





Temper 



itures of Holes 



in Bar. 





No. 1. 



No. 2. 



No. 3. 



No. 4. 



No. 5. 



21-7 



4 85 •{ 



13-2 



396 



61-65 



88-35 



122-2 



168-2 



13-2 



39-6 



61-7 



88-4 



122-4 



168-1 



21-8 



4-86 { 



13-2 



39-6 



61-7 



88-4 



122-4 



168-1 



133 



39-8 



61-8 



88-5 



122-45 



168-2 



21-85 



4-88 { 



13-3 



39-65 



61-9 



88-55 



122-5 



168-25 



13-3 



39-65 



61-8 



88-5 



122-55 



168-45 



21-86 



4-83 | 



13-3 



39-65 



61-8 



88-5 



122-55 



168-45 



13-3 



39-7 



61-85 



88-6 



122-6 



168-35 



Mean 21-8 



486 



13-3 



39-65 



61-8 



88-5 



122-45 



168-25 



§ 10. Correction of the Thermometers. — The corrections of the thermometers were 

 lot found in the ordinary way, as there is always more or less doubt attached to any 

 llowance that may be made for stem exposure on account of the impossibility of know- 

 ng the exact distribution of temperature along the stem of the thermometer. The 

 aethod adopted was simple and allowed the testing to be carried out with the ther- 

 Qometers in as nearly as possible the same circumstances as they are in during the 

 xperiments. 



The thermometers were tested at three different temperatures, at 0° C, about 100° C, 



nd about 218° C. At 0° C, the correction was found by hanging the thermometers in 



vertical position with their bulbs, and as much of their stem as was under the surface 



f the bar, embedded in powdered ice washed with distilled water. At the other two 



^mperatures the apparatus, a vertical section through the centre of which is shown in 



g- 3, was used. A piece of brass tubing of nearly the same diameter as the bar 



f nickel was cut into three lengths, E, F, and G, and these were brazed together as 



lown. The end, H, was closed, and three small tubes, A, B, and C, were brazed in the 



liddle piece F. These tubes were about half full of Wood's alloy. The piece E was 



alf-filled with water which was heated by the burner D, which was adjusted until just 



small quantity of water vapour escaped at the open end K. The tubes, A, B, and C, 



ere of about the same depth as the thermometer holes in the bar, and during the test 



ie thermometers were suspended in a vertical position with their bulbs near the bottom 



the tubes. Asbestos screens were fitted up at L and M to shield the thermometers 



VOL. XXXIX. PART II. (NO. 12). 3 L 



