3G6 Dr. C. Barns on the Relation between the Thermoelectric 

 only call attention to the following. In the thick bars expe- 

 rimented upon the value of rp ^ ° R is usually too large — a fact 



which is easily accounted for, as the unavoidable resistances 

 of contact above referred to will in this case have a relatively 

 great effect, the resistance of the bars themselves being very 

 small. In the rods included in Table VI. it was impossible to 

 secure a uniform redness throughout, the ends invariably re- 

 maining darker. As, however, T. E. H. depends principally 

 on the warm end, and AS on the mean hardness of the 

 whole bar, we have thus a second cause for an overlarge ratio. 

 So much, however, I think I have fully established, that 

 T. E. H. and specific resistance of steel throughout their varia- 

 tion are very simple functions one of another. T. E. H. and 

 specific resistance must therefore be looked upon as effects of 

 the same cause, as phenomena having some very intimate 

 connexion. 



d. Particular attention must here be called to the remark- 

 able result that the specific resistance of steel can by a pro- 

 cess of glasshardening be increased to nearly three times its 

 value in soft steel *. As this datum far exceeds that deter- 

 mined by Mousson (about 25 per cent.), it is not without some 

 hesitation that I make it public. The care bestowed on the 

 experiments, however, together with the regularity observable 

 in the variation of the results, I believe, sufficiently ensure 

 their correctness. See moreover § VI., Ef. 



e. As deserving special notice, I will further add that the 

 thermo -current always passes from the bar with greater to the 

 bar with less specific resistance. The few exceptions to this 

 fact in the Tables were afterwards found to be referable to 

 errors of experiment by direct observation J. 



* It is to be observed, however, that the difference between the specific 

 resistance of steel in the soft and hard states is dependent on the compo- 

 sition, increasing with the quantity of carbon contained from a very small 

 value in soft iron to the very large value above announced for steel. 



f Chwolson reports the increase of resistance due to glasshardening to 

 be only 0-6 per cent. This I can only explain by supposing the results 

 of this observer to have been obtained from wires suddenly cooled at a 

 temperature below that referred to in VII., c. 



% I would here again refer to the fact that, according to Magnus, 

 Thomson, and Mousson, drawn steel wire and hard-tempered steel are 

 on different sides of soft steel, both with respect to their thermoelectric 

 properties and their specific resistance. Thomson furthermore finds 

 transversely magnetized steel electronegative towards soft steel, this again 

 towards longitudinally magnetized steel ; Auerbach (Wiedemann's Ann. 

 v. p. 316, 1878), analogously, that the specific resistance of hard steel con- 



