220 Professor Joseph Larmor [May 1, 



physical properties. This use of the law of energy as a connecting- 

 principle afterwards became the note of Thomson and Tait's ' Treatise 

 on Natural Philosophy.' In revising for press a continuation of this 

 magnetic memoir, ' Phil. Mag.,' April 1855, where he is engaged in 

 deducing magnetic reciprocal relations in more elementary fashion by 

 use of a work-cycle, a thought occurred to him and was embodied in 

 a footnote under date March 26, which will be quoted in full.* 



" It might be objected that perhaps the magnet, in the motion 

 carried on as described, would absorb heat and convert it into 

 mechanical effect, and therefore that there would be no absurdity in 

 admitting the hypothesis of a continued development of energy. 

 This objection, which has occurred to me since the present paper was 

 written, is perfectly valid against the reason assigned in the text for 

 rejecting that hypothesis : but the second law of the dynamical theory 

 of heat (the principle discovered by Garnot and introduced by Clausius 

 and myself into the dynamical theory, of which, after Joule's law, it 

 completes the foundation) shows the true reason for rejecting it, and 

 establishes the validity of the remainder of the reasoning in the text. 

 In fact the only absurdity that would be involved in admitting the 

 hypothesis that there is either more or less work spent in one part of 

 the motion than lost in the other, would be the supposition that a 

 thermodynamic engine could absorb heat from matter in its neigh- 

 bourhood, and either convert it wholly into mechanical effect, or con- 

 vert a part into mechanical effect and emit the remainder into a body 

 at a higher temperature than that from which the supply is drawn. 

 The investigation of a new branch of thermodyniunics, which I intend 

 shortly to communicate to the Royal Society of Edinburgh, shows that 

 the magnet (if of magnetised steel) does really experience a cooling 

 effect when its pole is carried from A to />, and would experience a 

 heating effect if carried in the reverse direction. But the same 

 investigation also shows that the magnet must absorb just as much 

 heat to keep up its temperature during the motion of its pole icitli 

 the force, along AB^ as it must emit to keep from rising in tempera- 

 ture when its pole is carried against the force, along DC" 



The exposition of the new branch of thermodynamics here 

 referred to appeared in fact in the same month, April 1S55, in the 

 first part of the first volume of the ' Quarterly Journal of ]\Iathe- 

 matics,' under the title ' On the Thermoelastic and Thermomaa-netic 



* Elec, and Mag., § 672. In a less definite way this principle had been 

 effective already; cf. Mack, Principien der Wamelehre hist.-krit. entwickelt. 

 Early in 1849 James Thomson explains that it was his brother's pointing out 

 to him that, on Carnot's principle, water could be frozen isothermally without 

 requiring mechanical work, which set him on to the train of thought that 

 predicted the lowering of the freezing-point by pressure and calculated its 

 amount. As freezing is accompanied by expansion, a cycle involving freezing 

 at a high pressure and melting at a low pressure, in fact confronted him with 

 a perpetual motion, which he had to evade. 



