162 M. J. A. Angstrom on the Cenducting-Power of 



k is constant for good conductors of heat, such as copper, but 

 variable with the temperature in the case of the other metals. 



Indeed, the question cannot be answered in the way taken by 

 the above-named experimenters, viz. by a determination of the 

 temperature at different points in a bar which is heated at one 

 end to a constant temperature. For, in the first place, the 

 observed departures from the geometrical law may be due to an 

 altered value of the radiating power of the surface as well as to 

 an alteration in the conducting-power of the bar. And in the 

 second place, in the case of the better conductors of heat, the 

 method is too little sensitive to indicate with accuracy slight 

 differences in conducting-power. We see this from the last 

 experiments of Despretz, wherein he took bars of a bad con- 

 ductor (marble), and gave to them so great a diameter that the 

 heat could only very slowly pass away along the bar, and found 

 the departure from the geometrical law to be insensible, at least 

 when the temperature of the bar did not exceed that of the room 

 by more than 30° or 40° C. 



Lastly, it is also clear that if the proportionality between the 

 conducting-power for heat and that for electricity found by 

 "Wiedemann and Franz is to have any theoretical importance, and 

 not be a merely accidental coincidence, it will not suffice for the 

 metals to deport themselves similarly in regard to their conduct- 

 ing-powers for heat and electricity, but they must do this at all 

 temperatures; that is, the coefficient of conducting-power for 

 one and the same metal must be the same for heat and for elec- 

 tricity. Since, up to the present time, no body has been found 

 which does not vary in conducting-power for electricity with 

 alteration of temperature, it follows that this must also be the 

 case with regard to the corresponding relation to heat. 



Having made these remarks, I go on to the experiments 

 themselves. 



§ 2. The bars which were used were 1178 miliims. long and 

 35 miliims. thick. The distances between the holes for the 

 reception of the thermometer amounted to 200 miliims. ; and 

 the diameter of each hole was 2 miliims. in the experiments 

 with the copper bar and in the first experiments with the iron 

 bar. The holes in the iron bar were afterwards enlarged in 

 order to permit of the use of more sensitive thermometers. The 

 duration of the periods in the experiments with the copper bar 

 was, with o-ne exception, twenty-four minutes, and that in the 

 experiments with the iron bar thirty-two minutes. 



The apparatus for raising and lowering the temperature was 

 altered, so that cold water and vapour of water were not alter- 

 nately made to surround the bar in one and the same bath, but 

 for each of these agents a special bath was provided. Inas- 



