558 



MR A. CRICHTON MITCHELL ON THE THERMAL CONDUCTIVITY 



APPENDIX I. 



In the determination of the rate of cooling, when the bar experimented upon 

 was heated to about 100° C. by radiation from an open fire, previous to its 

 being heated rapidly over the small Bunsen burners, it was found that the rates 

 of cooling of the bar so treated were, for temperature excesses above 150° C, 

 in a constantly decreasing ratio to those deduced from the experiments in which 

 the bar had been subjected to a prolonged heating, during which the smoking 

 action of the burners so far altered the brightness of the surface as to produce 

 the effect mentioned. In the final deduction of conductivity these lower rates 

 of cooling have been used. The following table shows, in the case of iron and 

 German silver, the difference between the two series of estimates of the rate of 

 cooling. In both copper bars the difference was noticeable only at the very 

 highest temperatures used in the cooling experiments, and as, besides being 

 comparatively small, such difference was at a temperature excess not reached 

 in the long bar experiment, its effect on the area of the " statical curve of 

 cooling " was almost inappreciable. 



It must be remarked, however, that even with the larger differences between 

 the two series of rates of cooling in the cases of iron and German silver, the 

 two corresponding estimates of the conductivity differ only by about 1 per cent. 







Iron. 





( 



jEKMAN Silvef 





Temperature 

 Excess. 





























Lower. 



Higher. 



Ratio. 



Lower. 



Higher. 



Eatio. 



150 



302 



302 



10 



35 



354 



•988 



160 



328 



3-28 



10 



379 



3-84 



•987 



170 



356 



357 



•997 



414 



417 



•992 



180 



3815 



3-88 



•983 



4-46 



451 



•989 



190 



411 



4-2 



•978 



4-79 



4-92 



•973 



200 



4-4 



452 



•973 



511 



5-28 



•967 



210 



47 



4-88 



•963 



5 48 



566 



•968 



220 



5 03 



5-28 



•952 



5-77 



605 



•954 



230 



537 



560 



•958 



611 



64 



•954 



240 



574 



613 



•936 



642 



6-79 



•945 



250 



612 



659 



•928 



6-78 



715 



•948 



260 



655 



709 



•923 



715 



755 



•947 



270 



693 



7-59 



•913 



... 



... 



... 



APPENDIX II. 



In the text of this paper it is stated that the error due to the circumstance 

 of the thermometric readings being those of a variably heated thermometer is 

 not taken into account in the final results given above. While the paper was 

 being printed, I applied to the tangents and areas the corrections given by 



