
PROFESSOR TAIT ON THERMAL AND ELECTRIC CONDUCTIVITY. Cat 
in the case of the iron, still it is felt to a certain extent in the case of all the 
metals tried. My results are all somewhat uncertain on this account. This 
uncertainty, and means of removing it, are discussed in § 13. 
Another reason for the delay that has occurred in producing the results has 
been my endeavouring—to a certain extent fruitlessly—to give the results in 
terms of absolute temperature, by the help of air-thermometers. Much time has 
been spent on that work, yet, even with the assistance of Dr Joule and others, 
I have not been able to get a really good set of determinations. The real 
difficulty lies in the fact that the holes cut in the bars for the insertion of the 
bulbs of thermometers are necessarily so small, that it is not possible to con- 
struct any efficient air-thermometer which can be made to take the place of the 
mercurial ones. 
I have been assisted in the experimental part of the work by several of my 
Laboratory students ; but most especially by my mechanical assistant, Mr T. 
Linpsay, who has been throughout the inquiry as valuable to me as was his 
father to ForBEs. 
§ 5. The results now given are founded, some of them on experiments 
made before 1871, and some on experiments made last year. The calcula- 
tions have all been carried out with care and accuracy by Mr Evans (who 
used the processes described by Forsss), and their results have been verified 
by myself, partly by graphical methods, partly by various devices for interpola- 
tion, and in the majority of instances by calculation also.* But, as will be 
seen, I content myself at present with the statement of probable values only. 
I have only now arrived at nearly definite conclusions as to the best mode of 
* One of these interpolation methods is so easily applied, and (in consequence of the usual nature 
of the statical curves) gives results so fairly approximate, that it must be mentioned here as of great use 
if only in checking the results of the more complex calculations. 
Let v,, %», Vz, ¥., be the observed temperatures shown by the four thermometers, placed at intervals 
of three inches on the long bar. Let w be the number of degrees lost per minute by the thermometer in 
the short bar, when its temperature-excess above the air is nearly that of $(v,+;). Then the conduc- 
tivity at the temperature 4(v,+v,), in terms of the units employed in § 15 below, is very approximately 
w 
8(v, —%,— 03 +%%4) 
[This formula assumes third differences of v, to vanish.] With a single bar of 20 inches, or so, in 
length (with four or more holes three inches apart), to be used alternately for the statical and for the 
dynamical experiment (in the former with its free end artificially cooled), I believe that very fair 
determinations of thermal conductivity may be made in a few hours by the use of the above formula. 
Had I known this ten years ago I should not have undertaken the repetition and extension of ForBrs’ 
experiments wnder conditions exactly similar to his. But, on the other hand, had I not undertaken this 
work, I should probably not have fallen upon this simple method. 
I believe that it may be found applicable even to stout wires or rods, the temperatures being 
observed by a thermo-electric process. Thus these determinations may be made for very rare metals, 
and also for substances of very low conductivity. I hope, with the assistance of a party of my Labora- 
tory students, to get a large number of metals examined by this method during next winter and summer 
sessions. 
VOL. XXVIII. PART III. 9c 
