26 



NATURE 



[September 5, 191: 



the conclusion that the character of the radiation 

 emitted during- the recombination of the ions will be 

 a series of pulses, each pulse containing the same 

 amount of energy and being of the same type as very 

 soft X rays. If the X rays are really corpuscular, 

 these definite units or quanta of energy generated 

 by the recombination of the ions bear a close 

 resemblance to the hypothetical molecules of caloric. 



It may be objected that in many cases of friction, 

 such as internal or viscous friction in a fluid, no 

 electrification or ionisation is observable, and that the 

 generation of caloric cannot in this case be attributed 

 to the recombination of ions. It must, however, be 

 remarked that the generation of a molecule of caloric 

 requires less energy than the separation of two ions ; 

 that, just as the separation of two ions corresponds 

 with the breaking of a chemical bond, so the genera- 

 tion of one or more molecules of caloric may corre- 

 spond with the rupture of a physical bond, such as the 

 separation of a molecule of vapour from a liquid or 

 solid. The assumption of a molecular constitution 

 for caloric follows almost of necessity from the mole- 

 cular theories of matter and electricity, and is not 

 inconsistent with any well-established experimental 

 facts. On the contrary, the many relations which are 

 known to exist between the specific heats of similar 

 substances, and also between latent heats, would 

 appear to lead naturally to a molecular theory of 

 caloric. For instance, it has often been noticed that 

 the molecular latent heats of vaporisation of similar 

 compounds at their boiling-points are proportional to 

 the absolute temperature. It follows that the 

 molecular latent caloric of vaporisation is the same 

 for all such compounds, or that they require the same 

 number of molecules of caloric to effect the same 

 change of state, irrespective of the absolute tempera- 

 tures of their boiling-points. From this point of view- 

 one may naturally regard the liquid and gaseous states 

 as conjugate solutions of caloric in matter and matter 

 in caloric respectively. The proportion of caloric to 

 nl„t-^er varies regularly with pressure and temperature, 

 and there is a definite saturation limit of solubility at 

 each temperature. 



One of the most difficult cases of the generation of 

 caloric to follow in detail is that which occurs when- 

 ever there is exchange of heat by radiation between 

 bodies at different temperatures. If radiation is an 

 electro-magnetic wave-motion, we must suppose that 

 there is some kind of electric oscillator or resonator 

 in the constitution of a material molecule which is 

 capable of responding to the electric oscillations. If 

 the natural periods of the resonators correspond suffi- 

 ciently closely with those of the incident radiation 

 the amplitude of the vibration excited may be sufficient 

 to cause the ejection of a corpuscle of caloric. It is 

 generally admitted that the ejection of an electron 

 may be brought about in this manner, but it would 

 evidently require far less energy to produce the emis- 

 sion of a neutral corpuscle, which ought therefore to 

 be a much more common effect. On this view, the 

 conversion of energy of radiation into energy of caloric 

 is a discontinuous process taking place by definite 

 molecular increments, but the absorption or emission 

 of radiation itself is a continuous process. Prof. 

 Planck, by a most ingenious argument based on the 

 probability of the distribution of energv among a large 

 number of similar electric oscillators fin which the 

 entropy is taken as the logarithm of the probabilitv, 

 and the temperature as the rate of increase of energy 

 per unit of entropy), has succeeded in deducing his 

 well-known formula for the distribution of energy in 

 full radiation at any temperature ; and has receiitly, 

 by a further extension of the same line of argument, 

 arrived at the reniark.able conclusion that, while the 

 absorption of radiation is continuous, the emission of 

 NO. 2236, VOL. go] 



radiation is discontinuous, occurring in discrete 

 elements or quanta. Where an argument depends on 

 so many intricate hypotheses and analogies the possible 

 interpretations of the mathematical tormulse are to 

 some extent uncertain ; but it would appear that Prof. 

 Planck's equations are not necessarily inconsistent with 

 the view above expressed that both emission and 

 absorption of radiation are continuous, and that his 

 elementa quanta, the energy of which varies with their 

 frequency, should rather be identified with the mole- 

 cules of caloric, representing the conversion of the 

 electro-magnetic energy of radiation into the form of 

 heat, and possessing energy in proportion to their 

 temperature. 



.\mong the difficulties felt, rather than explicitly 

 stated, in regarding entropy or caloric as the measure 

 of heat quantity is its awkward habit of becoming 

 infinite, according to the usual approximate formulae, 

 at extremes of pressure or temperature. If caloric is 

 to be regarded as the measure of heat quantity, the 

 quantity existing in a finite body must be finite, and 

 must vanish at the absolute zero of temperature. In 

 reality there is no experimental foundation for any 

 other conclusion. According to the usual gas formulje 

 it would be possible to extract an infinite quantity of 

 caloric from a finite quantity of gas by compressing 

 it at constant temperature. It is true that (even if 

 we assumed the law of gases to hold up to infinite pres- 

 sures, which is far from being the case) the quantity 

 of caloric extracted would be of an infinitely low order 

 of infinity as compared with the pressure required. 

 But, as a matter of fact, experiment indicates that the 

 quantity obtainable would be finite, although its exact 

 value cannot be calculated owing to our ignorance of 

 the properties of gases at infinite pressures. In a 

 similar way, if we assume that the specific heat as 

 ordinarily measured remains constant, or approaches a 

 finite limit at the absolute zero of temperature, we 

 should arrive at the conclusion that an infinite quantity 

 of caloric would be required to raise the temperature 

 of a finite body from o° to i° absolute. The tendency 

 of recent experimental work on specific heats at low 

 temperatures, by Tilden, Nernst, Lindemann, and 

 others, is to show, on the contrary, that the specific 

 heats of all substances tend to vanish as the absolute 

 zero is approached, and that it is the specific capacity 

 for caloric which approaches a finite limit. The theory 

 of the variation of the specific heats of solids at low 

 temperatures is one of the most vital problems in the 

 theory of heat at the present time, and is engaging 

 the attention of many active workers. Prof. Linde- 

 mann, one of the leading exponents of this work, has 

 kindly consented to open a discussion on the subject 

 in our section. We are very fortunate to have suc- 

 ceeded in securing so able an e.xponent, and shall await 

 his exposition with the greatest interest. For the 

 present I need only add that the obvious conclusion 

 of the caloric theory bids fair to be completely 

 justified. 



A most interesting question, which early presented 

 itself to Rumford and other inquirers into the caloric 

 theory of heat, was whether caloric possessed weight. 

 While a positive answer to this question would be 

 greatl}' in favour of a material theory, a negative 

 answer, such as that found by Rumford, or quite 

 recently by Profs. Poynting and Phillips, and by Mr. 

 L. Southerns working independently, would not be 

 conclusively against it. The l_atter (f^f vers found 

 that the change in weight, if any, c§5^ ily did not 

 e.xceed i in lo' per i° C. If the mass^c. a molecule 

 of caloric were the same as that generally attributed 

 to an electron, the change of weight, in the cases 

 tested, should have been of the order of i in lo' per 

 1° C, and shoi^ld not have escaped detection. It is 

 generally agreed, however, that the mass of the elec- 



