THERMOELECTRICITY 77 



less pay, but I had a great deal more leisure for private work. In fact I have barely 

 time for any private work during the winter session now-a-days. 



However, I have got some students who are able and willing to work and I have 

 handed over my apparatus to them to make the best of it. At present I am entirely 

 engaged with " 1'effet Thomson " if you know what that is the so-called specific heat 

 of electricityjin different conductors, which I think I have proved both experimentally and 

 theoretically to be proportional to the absolute temperature. This has led me to 

 construct a thermometer depending on two separate thermoelectric circuits working 

 against one another, so as to give galvanometric deflections rigorously proportional to 

 differences of absolute temperature through all ranges till the wires melt. I hope to 

 get the specific heats and melting points of various igneous rocks, &c., &c., true to a very 

 few degrees. 



My Holtz machine perhaps about the last thing that Ruhmkorff sent out of 

 Paris 1 is a splendid success ; 2-inch sparks from a jar with \ square yard of 

 coated surface at intervals of 4 seconds. 



Tait was now in the heart of his thermoelectric investigations, which 

 for several years dominated the work of the Physical Laboratory. The 

 difficulties encountered and the methods by which they were overcome are 

 discussed in a series of short papers communicated to the Royal Society of 

 Edinburgh, afterwards worked up into the great Transactions paper of 1873. 

 In the earlier pioneer work Tait was helped by May and Straker, and a little 

 later by John Murray and R. M. Morrison. In the summer of 1873 he 

 instructed C. E. Greig and myself, who had spent one winter in the Laboratory, 

 to investigate by one and the same method the thermoelectric properties 

 of some twenty different metals paired in a sufficient number of ways ; and 

 these experiments which were made in the Natural Philosophy class room 

 formed the basis of the " First Approximation to the Thermoelectric Diagram." 

 The hot junctions were heated in oil up to a temperature of nearly 300 C. 

 Meanwhile Tait himself had been working with iron at still higher temperatures, 

 and making the first of what proved to be the most novel of his discoveries 

 in thermoelectricity, namely, the remarkable changes at certain temperatures 

 in the thermoelectric properties of iron and nickel. 



Nearly all pairs of metals up to the temperatures of their melting points 

 have the thermoelectromotive force a parabolic function of the difference of 

 the temperatures of the junctions. When, however, iron or nickel is one of the 

 metals forming the thermoelectric couple this rule breaks down. Nevertheless 

 between particular limits of temperature the parabolic law is satisfied, so that 

 the relation between electromotive force and temperature can be fairly well 



1 That is, before its investment by the German troops. 



