March 31, 1881] 



NATURE 



5Q5 



was led by the letter to suppose that Prof. Ayrton himself might 

 have something further to say regarding his views as soon as he 

 returned to England, but mainly because I did not see any point 

 in it specially requinng an immedi.ite re[>ly. I find however that 

 a considerable amount of cautious scepticism and suspense of 

 judgment still prevail on the subject — a scepticism which I'rof. 

 Herschel's enthusiastic letter of a month ago (p. 383) has not 

 gone far to remove, because, though there can be no doubt of his 

 confirmation of the fact that ice in a hot vacuum is infusible 

 and disappears slowly, there is nothing in his letter confirming 

 the hypothesis that it is hot, which is the only point under 

 discussion. 



Now for my own part I fully and unreservedly accept this as 

 a fact, not only on account of Dr. Carnelley's experimental 

 evidence, but also because 1 imagine myself to perceive exactly 

 why it occurs, and indeed that it might conceivably have been 

 conjectured as probable beforehand. 



My present communicatiim therefore is merely to remove as 

 far as possible any sense of mystification which Prof. Perry's 

 letter may have tended to produce, and to indicate the ground of 

 his error. 



Professors Ayrton and Perry, with their stiff paper models, 

 start, if I am not mistaken, on the assumption that the ordinary 

 equations deduced from the two laws of thermodynamics will 

 apply to the case : and this is exactly how I started myself. I 

 considered that it was necessary to investigate the behaviour of 

 a substance whose properties were defined, not by two inde- 

 pemlcnt variable^, as is usual, but by three ; the pressure, 

 quantity of solid, and temperature, being all three arVvitrary and 

 independent of each other in the Carnelley experiment ; and I 

 extended Clausius's general equations to suit this case. But it 

 was very soon evident that they did not apply at all, and for this 

 reason, that the second law is only true f r processes that are 

 reversible, and the sublimation f>f hot ice is essentially an irrever- 

 sible process. This is indeed the whole gist of the matter, and it is 

 entirely due to this that the ice gets hot. Ordinary evap^-ration 

 of a liquid below its boiling-point against a presure less than 

 its "vapour-tension" is an irreversible process, and accordingly 

 the temperature is perfectly indefinite, and depends on the rate 

 of supply of heat and on the rale of evap ^ration. .So also with 

 ice above the boiling-point, that is, ice subliming under a less 

 pressure than the vapour-tension; its temperature depends 

 simply on the rate of supj4y of heat and on the rate of evapo- 

 ration. So far everything is perfectly simple and absolutely 

 certain. 



The only possible question that can arise is whether internal 

 disintegration of the solid will not set in and prevent its rising 

 above the boiling-point ; whether in fact a solid cannot boil as a 

 liquid does. I have given reasons for believing that in a solid 

 formed in vacuo, or without air- bubbles, and constantly rising 

 in temperature, this will not occur ; and I deny that under these 

 circumstances it is in a particularly unstable condili m analogous 

 to that of superheated water on the point of "boiling by 

 bumping." 



This however 1 fully admit is a p jint distinctly open to dis- 

 cussion, and I imagine that without an experiment one could not 

 feel at all certain about it. But personally I feel that the evi- 

 dence already given us by Dr. Carnelley, together with the 

 theoretical probability indicated in my former letter (p. 264), is 

 sufficient and con--lu~ive. 



It was no doubt sonjewhat staggering to learn (NATURE, vol. 

 xxiii. p. 341) that Prof. McLeod, with his well-known experi- 

 mental skill, should have hitherto failed to repeat the experiment, 

 or to get the ice at all ab ne zero;' but 1 take this as an 

 instructive example of those r.are cases where refined experi- 

 mental appliances are obstructions rather than aid-, for I believe 

 the failure to he simply due to the fact that Prof. McLeod's 

 vacuum «as far too perfect, aud the evaporation thtrefore so 

 rapid that the ice did not have a fair chance of showing its 

 willingness to rise in temperature ; it could not in fact get even 

 as high as o" C. But if Prof. McLeod will discreetly spoil 

 his vacuum until the |iressure is only just below the vapour- 

 tension corresponding to the temperature shown by his thermo- 

 meter, I have no doubt that he will see the ice rise to any 



^ Since this was in type I have received, by the kindness ofM. Boutler^w, 

 a. copy of a paper read by him before the St. Petersburg Academy of Sciences, 

 iri which he summarises the views which have appeared on the subject, relates 

 his failure to repeat the experiment, and confesses himself a sceptic. It 

 would not be d.jing justice to M. Boutlerow's carefiiliy-wroiight memoir to 

 discuss it in a foot-note, but it is my impression that his failure is due to the 

 same cause as that which I have ventured to suggest above as accounting for 

 Pro£ McLeod's, viz. tjo perfect apparatus and too great experimental skill. 



temperature l.e likes, and he will find that when it is crossing 

 zero it will be utterly regardless of the fact. 



The same kind of statement applies to solid carbonic acid, on- 

 which 1 have made a few experiments with a view to raising its 

 tempeniture. 1 squeezed it into the ice form in a hydraulic 

 press (to diminish the evaporating surface), put a thermometer in 

 it, and held it over a fire. The evaporation is so excessiv«ly 

 rapid, however, that it remains apparently just as cold as 

 before. 



I have not time to follow it up just now, but the obvious 

 thing is to [jut it undi,r pressure, so as to diminish the rate of 

 evaporation, and then heat it. Prof. McLeod informs me that 

 the boilii;g point of CO2 continues below its melting-i oint 

 (which is given by Frankland as -57" C), until the pressure is 

 four atmospheres ; so that anything just ui der four atmo pberes 

 may be a | -plied to this substance with impunity, and it will then 

 be exactly in the most favourable condition for the Carnelley 

 experiment ; and I have not the slightest doubt that it can then 

 be « armed, and if at the same time the pressure be judiciously 

 and gradually increased, that it can be made as warm as one 

 pleases until it has all disappeared. 



Experiment with suostances other than water however are 

 likely to be more difficult, simply bee 'use few substances have 

 such a l.nrge latent heat both in the liquid and gaseous condition, 

 and therefore few sub-tances will be anything like so permanent 

 and outlive the evaporation so long, Oliver J. Lodge 



17, I'arkhurst Koad, N. 



The announcement made some time since by Dr. Carnelley 

 that ice in vacuo could be raised to a temperature far above its 

 ordinary melting-point, seemed so thoroughly in opposition to 

 the exitrience derived from the great work of Renault on 

 the ten-ions of vapours ; and as it called for a complete 

 change of idens in a field in which I am much interested, and as 

 Dr. Carnelley asked others to repeat his experiments, I was 

 induced to examine for myself the experim<nts on which .'o 

 curious a statement was founded. 



I used two different methods : the Torricellian vacuum and 

 the Sprengel vacuum. As the experiment, as conducted by the 

 TorrieeDian method, can easily be repeated by any one, and is 

 much Minpler in form than Dr. Carnelley's, I shall detail it. In 

 the first place I wished to obtain a clear continuous piece of ice 

 round the thermometer, as Dr. Carnelley's method gave flaky 

 ice, which I found might lead to errors, owing to its discon- 

 tinuity leaving the thermometer bare in parts. To obtain clear 

 ice the following method was used : — .Some di'-tilled water was 

 boiled ill a test-tube A fitted with a two-holed stopper, with a ther- 

 mometer through one hole dipping into the water; when all the air 

 was expelleil, a glass plug was pressed into the other hole asfainst 

 theu-suing steam, and the whole allowed to cool, and then frozen 

 in a freezing-mixture. A long-necked " German Florence 

 flask " was then rin ed with distilled water and filled with mer- 

 cury, anil aLs ) placed in a freezing-mixture. The tube A was 

 then gently warmed wiih the hand, and the plug of ice adhering- 

 to the iher.nomcter withdrawn. The glass plug in the second- 

 hole in the stopper was then replaced by a marine barometer- 

 tube of ai)i'Ut forty inches in length, having been drawn out 

 about four inches from the top to facilitate sealing. The pluij 

 of ice round the thermometer was then inserted into the neck of 

 the flask full of mercury, and the stopper pressed home. This 

 caused the mercury to rise in the barometer-tube, and the whole 

 was then inverted as at B ; and when the ir.ercury had all 

 run out, the fall tube was melted through at the con^tric 

 tion" IS, leaving a Torricellian vacuum ab-^ve. The fla k wa.s 

 now laid on its side in a freezing-mixture and well covered over 

 w'yCa ice and salt as at c. After a few minutes, to allow the 

 receiver to cool, heat was applied to the neck of the flask with 

 aBunen lamp, and even with a blowpipe, till the glass softened, 

 but the temperature of the thermometer did not rise until some 

 part of it became denuded of ice, or until air had been admitted. 

 The experiment was repeated a;^ain and a-^ain, but in no case 

 while the vacuum was intact could the temperature of the ice be 

 raised materially above that of the receiver. If the tem|ierature 

 oft the receiver «as - 12°, then the ice was a little over - 12% 

 say a'lout - 11°, but never more than two degrees above the 

 receiver, although the glass almost in contact with the ice was at 

 its softening point. This is exactly what we would expect froin 

 Regiiault's exi.eriments ; the temperature of the receiver deter- 

 mines the vapour-tension, and therefore the "lioiling point" of 

 the ice. The ice was certainly never hot, and was not even 



