212 ' Professor Joseph Larmor [M^J Ir 



more strongly than by recalling that it is just this Thomson-Joule 

 intrinsic cooling effect of expansion without external work, yery slight 

 under ordinary conditions, due merely to mutual separation of the 

 molecules of the gas, that is the essential feature in the modern 

 continuous processes for liquefaction of eyen the most refractory 

 gases, by the expenditure of mechanical power to abstract the heat, 

 which have now become familiar. On the other hand, the great 

 economy of the reversed Carnot gas-cycle for ordinary refrigeration 

 was pointed out in 1852, and applied by his brother to the ventilation 

 of Belfast College. 



In their parallel developments of the subject, while Clausius kept 

 mainly to the theory of heat engines, applications over the whole 

 domain of physical science crowded on Thomson. Already in De- 

 cember 1851 he communicates to the Eoyal Society of Edinburgh 

 his Theory of Thermo-electric Phenomena, including the classical 

 prediction of the convection of heat by the electric current, the so- 

 called Thomson effect, which in the theory of electrons has a literal 

 title to its original name. The formulae of the printed abstract * of 

 this paper show that he must have been already in full command 

 (December 1851) of Carnot's principle in its most generalised form — 

 viz., as he expressed it in May 1854, but there introducing absolute 

 temperature T, then recently determined by himself and Joule — that 

 in a complete reversible cycle of change 2(H/T) vanishes, or in 

 differential notation /(^;?H/T) = 0, a form which was independently 

 given by Clausius in December 1854, and from which the transition 

 to Clausius' entropy-function (1856) is but a step. These advances 

 appeared in full in the memoir, 'Trans. R. S. Edin.,' 1854,"|* where, 

 in the way customary with him, he passes on to a lonu: digression on 

 the thermo-electrics of crystalline matter, including, after Stokes, the 

 full theory of rotational vector effects. This latter subject was 

 brought again into prominence many years after, when times were 

 riper for it, with reference back to the present exposition, on the 

 announcement by E. H. Hall of the discovery of an influence of this 

 kind in electric conduction in a powerful magnetic field. Here also 

 shines forth in a notable example what was always a main feature of 

 Thomson's theoretical activity, the utilisation to tlie utmost of models 

 and images of physical phenomena. He absolutely refused to deny 

 to matter, however continuous and uniform as to sense it might 

 appear to be, the possession of any property which he could imitate 

 in a lattice structure or other architectural model, however complex ; 

 clearly, in his view, one has no right to assign limits a priori to the 

 possible physical complexity of molecular aggregation. 



One type of such limits, indeed, the only ones a priori^ he vindi- 

 cated in one of his most refined theoretical advances, those, namely, 



* Math, and Phys. Papers, i. pp. 316-323. 

 t Loc. cit., pp. 232-261. 



