1888.] Prof. Tait on Thermoelectric Properties of Iron. 127 
5. Oa the Thermoelectric Properties of Iron. 
By Prof. Tait. 
For some time Signor Battelli has been engaged, with remarkable 
success, in measuring directly the amount of the “ Thomson effect ” 
in various metals. 
With the exception of iron, the common metals have given him 
results coinciding as closely as could be expected with those I found 
in 1872 by an indirect method. Among other particularly satisfac- 
tory things, he has directly verified the first of the two changes of 
sign of the Thomson effect in nickel. And I think it will be 
allowed that what I introduced long ago as a mere working hypo- 
thesis, — that the Thomson effect is directly proportional to the 
absolnte temperature, — if it was not completely established as a fact 
by my own experiments, has been made absolutely certain by the 
recent work of Campbell and of Battelli. 
In his paper on iron, however, he finds that the specific heat of 
electricity in that metal is by no means closely proportional to the 
absolute temperature. I had long ago met with the same difficulty, 
and in fact I have never found two specimens of iron— even when cut 
from the same hank of wire — which agreed well with one another. 
Even Matthiessen’s pure iron does not give a straight line on the 
thermoelectric diagram (for temperatures under low red heat) ; and, 
as will be seen in PL IX., Trans. R.S.E., vol. xxvii., the lines for 
various kinds of iron are all sinuous, but all unlike one another. 
In 1870 I stated that, iron being one member of the thermoelectric 
couple, “ the parabola was slightly steeper on the hotter than on the 
colder side.” This implies that, if the line of the other metal be 
straight, the line of iron is concave downwards, as the “ diagram ” 
is drawn. Sig. Battelli gives the following values of the Thomson 
effect, which agree with this statement, viz. : — At 53° C. - 9*2.10"® ; 
at 108° C.- 12*15.10"®; at 242° 0.-17.10“®; and at 308° C. 
-21*4.10“®. The fact seems to be that not only is iron never 
obtained pure, but it is one^ of those metals on whose physical 
properties a very slight impurity produces a marked effect. 
Sig. Battelli kindly sent me one of the two iron rods employed in 
his experiments ; and Messrs Shand and Morison have determined 
