,66 Prof. L. Lorenz on the Determination of 



If we now observe that electrical conductivity for the various 

 pure metals is nearly inversely proportional to the temperature 

 calculated from absolute zero, while their thermal conductivity 

 approaches more to constancy, and further that in the case of 

 iron the deviations for both kinds of conductivity are in the 

 same direction, then we may say that the preceding facts esta- 

 blish, with as great closeness as can be expected, the law that 

 the conducting -power of a pure metal for heat and for electricity 

 is proportional to the temperature calculated from absolute zero. 



This ratio must obviously more or less vary in several cases. 

 If, for instance, the metal is not homogeneous, or contains ad- 

 mixtures of foreign metals, and, generally speaking, in cases in 

 which, owing to unequal heating in the interior, thermo-electric 

 currents may be formed, then the thermal con ducting-power will 

 probably be increased, or at all events the relation between the two 

 kinds of conductivity will be altered. This is doubtless the case 

 when heat can be transmitted as radiant heat in the interior of 

 bodies. In this kind of transmission we must seek the reason 

 why thermal conductivity for all transparent and translucent 

 media, and generally for all non-metallic bodies, is apparently 

 far greater than that which would correspond to their electrical 

 conductivity. Lastly, if the body is liquid, the ratio must alter, 

 owing to the mobility of the parts. If, for instance, a column 

 of liquid is heated from below, this mobility will, of course, in- 

 crease the observed conductivity. If it is heated from above, 

 currents in the interior cannot be entirely avoided ; for every 

 part of the liquid in the same horizontal section cannot have 

 quite the same temperature; the colder parts will then sink, 

 and the warmer ones rise toward the source of heat ; and the 

 conducting-power must therefore be diminished in consequence 

 of the motion of the parts. 



We must hold, then, that the law, if it is at all valid, is pro- 

 bably only absolutely so for pure homogeneous and solid metals. 

 In fact even an unequal heating will make the metals heteroge- 

 neous and will give rise to thermo-electrical currents. 



From the foregoing observations I shall now attempt to de- 

 duce the ratio between the conducting-power of metals for heat 

 and for electricity in absolute units. We shall obtain from this 

 the remarkable result, that for a pure homogeneous solid metal 

 this ratio is equal to the temperature calculated from the absolute 

 zero expressed in the above-defined absolute units. 



To determine thermal conductivity in absolute measure, we 

 must know how great is the quantity of heat which traverses 

 each unit of surface of a plate of given thickness and with a 

 given difference of temperature at the two sides. The older expe- 

 riments on this subject have, for very intelligible reasons, led to 



