[ 448 ] 

 LVI. Intelligence and Miscellaneous Articles. 



ON THE THERMAL CONDUCTIVITY OF HARD AND SOFT STEEL. 

 BY F. KOHLRAUSCH. 



TU"E know from Mousson *, and more particularly from the 



' thorough investigations of Barus t, that the electrical con- 

 ductivity of steel depends on its hardness, and, as Barus found, so 

 closely that the hardening of a soft steel bar can increase its con- 

 ductivity two or three times. 



If, now, the relation proved by Wiedemann and FranzJ to exist for 

 different metals, that a metal conducts heat and electricity almost 

 equally well, holds also for the influence of mechanical preparation, 

 or molecular aggregation, it is to be expected that the thermal conduc- 

 tivity of steel will be powerfully influenced by its degree of hardness. 



The probability of an affirmative answer to this question follows 

 from the fact that the statements respecting the conductivity in 

 iron and steel lie much further apart than with other materials. 

 Kirchhoff and Hanseinann find in three kinds the numbers 0*096, 

 0-137, and 0*142 grm. cal./cm. sec. The authors do not discuss the 

 possible causes of this great difference, and seem to assign it 

 principally to the different proportions of carbon and silicon. An 

 observation, however, from which follows that the magnetic coercive 

 force of the worse conducting iron was the greater, points to an 

 influence of the hardness on the conducting power §. 



I used two cylindrical-turned and well polished bars of 1*2 cm. 

 diameter, 30 cm. in length, and 270 grammes in mass ||. They 

 were cut from the same piece ; one was heated and cooled slowly, 

 while the other was glass-hardened. 



Sensitive hands can at once tell by holding the cold bars, that 

 the soft steel conducts better than the hard. A lecture-experiment 

 on the rate of melting of wax or something similar is sufficient to 

 show the difference %. 



In order to determine the numerical relations approximately, I 

 made a few measurements by Despretz's method, which was also 

 that adopted by Wiedemann and Franz. The end of the bar was 

 heated by steam, and after the temperature was stationary the 

 excess of temperature u over the surrounding air was measured in 

 three equidistant sections **. 



* Mousson, Neue Denkschr. cler ScMoeiz. Geselhchaft. xiv. p. 1 (1855). 



t Barus, Wiedemann Ann. vol. vii. p. 399 (1879). 



X Wiedemann and Franz, Pog<?. Ann. vol. lxxxix. p. 531 (1853). 



§ Kirchhoff and Hansemann, Wied. Ann. vol. xiii. p. 417 (1881). 



|| The soft bars weighed 271, and the hard one 268 grammes. 



•fl The following method is perhaps the most convenient. The bottoms of 

 the bars to he compared are placed in a freezing-mixture, of alcohol and 

 snow for instance, and the height is observed up to which a deposit of mois- 

 ture or of ice takes place. 

 ** As the conductivity of iron from 0° to 100° does not vary more than 

 2 per cent, according to Lorenz, it was needless for my purposes to take 

 this into account. 



