538 MR A. CRICHTON MITCHELL ON THE THERMAL CONDUCTIVITY 



tubes having been carefully calibrated before filling, the standard points 0° C. 

 and 100° C. were directly determined in the usual manner. But the positions 

 of 200° C. and 300° C. were determined by taking successive portions of the 

 tube whose volume (cold) corresponded to that of the portion (also cold) from 

 0° C. to 100° C. I have not the means of making allowance for this defect, 

 which will probably mar all experiments of the kind until suitable air- 

 thermometers are employed. 



5. The fact that the conductivity deduced from experiments on the iron bar, 

 when its full length is employed, differ so considerably from those obtained 

 when it is artificially cooled in the middle, appears to be intimately connected 

 with a remark made in my paper (§ 14) that " in measuring conductivity, at 

 whatever temperature, things ought to be arranged so as to avoid any slow flux 

 of heat." It seems that, even after the lapse of eight hours, the steady state of 

 temperature has not been reached in the colder parts of the long iron bar. 



6. As the numerical data, concerning specific gravity and specific heat, 

 which Mr Mitchell has (in default of better) been obliged to employ, are 

 only rough estimates, I asked him to test them by finding the ratios of 

 the rates of cooling of copper and iron at various common temperatures. The 

 surface material was the same in the two bars, and their dimensions equal, so 

 that the amount of heat lost in a given (short) time must have been the same 

 for each at the same temperature. The ratio of the rates of cooling should 

 therefore be constant for all temperatures if, and only if, the rate of change of 

 specific heat with temperature be the same for each of the two materials. The 

 result does not seem to favour the accuracy of the assumed data, but the 

 process employed is not by any means an accurate one. 



7. As my determinations of the relative electric conductivities of the bars 

 had been verified by Mr D'Arcy Thompson, there is no necessity for their 

 repetition. But, using them, with Mr Mitchell's results for thermal con- 

 ductivity, my comparative table (Trans. R. S. E., 1878, p. 739) should be 

 altered (subject, of course, to correction for improved values of specific gravity 

 and specific heat) to something like the following : — 



Copper (Crown), . 



„ (G), .... 

 Forbes' Iron, .... 

 Lead, ..... 

 German Silver, 



Thermal. 



Electric. 





15 



1-729 





10 



1000 





0-23 



0264 





012 



0149 





013 



0117 



P. G. Tait.] 



