Dec, 20, 1877] 



NATURE 



H3 



I spoke of heat passing across a stratum of gas from one solid 

 surface to another " as though there were, in contact with each 

 solid surface, a layer of gas whose temperature is throughout the 

 same as [it would perhaps have been been better to have said 

 "determined by "] that of the contiguous solid." 



I am fully aware of the ease with which one may be led into 

 serious mistakes by trusting too implicity to such simplifying 

 assumptions, and also that some of the particular suppositions 

 made above would be inadmissible in a discussion of the general 

 problem of the conduction of heat in gases ; but I do not see any 

 fallacy in employing them for the special purpose which I had in 

 view in my last letter, namely, to show why I think that the flow 

 of heat across a thin stratum of gas must be facilitated by dimi- 

 nishing the pressure of the gas. Prof. Osborne Reynolds's argu- 

 ment that " if there were a layer of uniform temperature, no heat 

 would be transmitted," does not appear to me to be applicable 

 to the case in question. It seems conceivable, as an extreme 

 case, that, in a very thin layer of gas, between parallel solid 

 surfaces maintained at different temperatures, the molecular 

 movements might take place exclusively in the direction of the 

 perpendicular to the bounding surfaces. In such a case the 

 particles would move from side to side of the layer of gas with a 

 uniform velocity, though the velocity one way would be greater 

 than the velocity the other way, and heat would be transmitted 

 across a layer of gas having the same temperature throughout. 

 Such a condition, whether practically realisable or not, would, 

 if I understand him aright, be the limiting case in one direction 

 of what Mr. Stoney has called for shortness a " Crookes's 

 layer : " the limiting case in the other direction being the ordi- 

 nary condition of a gas, where the average velocity of the 

 molecules is independent of direction. I venture to think that, 

 in pointing out the results which must follow from the existence 

 of a predominating direction of molecular motion, Mr. Stoney 

 has made a very important contribution to the kinetic theory 

 of gases ; and I do not see that his conclusions are in any way 

 invalidated by its being shown that they are not in harmony 

 with " the generally-accep'ed laws of gases," inasmuch as these 

 Uws are deduced from suppositions which expressly exclude the 

 conditions he has investigated. G. Carey Foster 



December 17 



Allow me to say a few words on what I believe to be the 

 correct theory of the radiometer. This theory was given to me 

 by Prof. Osborne Reynolds during spring of 1875, and I have 

 found it capable of explaining every experiment on the subject 

 with which I am acquainted. 



The conservation of momentum is one of the laws of nature 

 which even molecules do not break, and that law puts some 

 restraints on the wonderful things ^yhich the shocks of molecules 

 can accomplish. Imagine a vessel full of gas at a certain tem- 

 perature. The centre of gravity of the gas and that of the vessel 

 are supposed to be at relative rest. Suppose now that I increase 

 the velocity of a certain number of molecules in a given direction, 

 the centre of gravity of the gas will move relatively to the centre 

 of gravity of the vessel, and no number of encounters between 

 the molecules can alter that motion until the momentum has 

 been taken up by the vessel. If in any gas we have a passage of 

 heat in a certain direction, we shall have a propagation of 

 momentum owing to the fact that the molecules move more 

 quickly in one direction than in the opposite one, and no 

 number of encounters can alter that propagation. Where the 

 momentum enters the gas and where it leaves it we observe 

 certain forces. This is Prof, Reynolds's theory of the radiometer. 

 It has been objected that an increased pressure on the cool side 

 of the vanes of a radiometer will counterbalance the force acting 

 on the blackened sides, when the dimensions of the vessel are 

 large compared with the mean path of a molecule, but I do not 

 think that such is the case. The following special case may 

 make this point a little clearer. If the forces on the vanes are 

 counterbalanced, the forces on the vessel must be counter- 

 balanced as well. In the case of an ordinary radiometer the 

 forces reduce to a couple, and I do not see how any crowding of 

 molecules in one part of the vessel more than in another can 

 produce a couple on the vessel. The whole problem is one of 

 conduction of heat. All the experiments made by Mr. Crookes 

 on cups, inclined vanes, &c., admit of the same easy explanation 

 as the fact that when a long and a short wire are connected with 

 the poles of a battery, the current in the shorter wire will be the 

 strongest. In a radiometer with inclined vanes, for instance, the 



temperature is the same on both sides, but the gradient of tem- 

 perature is much larger on one side, and hence more heat will 

 escape on that side. The dimensions of the vessel also have to 

 be taken into account in the same way as the length of a wire 

 has to be taken into account when the strength of an electric 

 current flowing through it has to be calculated. It is difficult to 

 say exactly what takes place within very small distances from 

 the hot surface, but it seems clear that any phenomenon, such as 

 Prof. Carey Foster supposes to exist, must affect the passage of 

 heat in the same way as the force on the vanes. As the careful 

 researches of Messrs. Kundt and Warburg have shown that 

 under great exhaustion the conduction of heat decreases and 

 does not increase, I do not see how an increase in the force can 

 take place. 



The scientific world will judge how far Prof. Stoney has suc- 

 ceeded in establishing'any new laws on the conduction of heat 

 through gases. In justice, however, to Messrs. Provostaye and 

 Dessains, whose experiments he calls to his aid, I wish to point 

 out that their numerous experiments, with two exceptions, are in 

 entire accordance with existing theories. At the time these experi- 

 ments were made, no distinction was drawn between convection 

 and true conduction. In order to deduce, therefore, the loss of 

 heat due to true conduction, Prof. Stoney is obliged to subtract 

 the effect due to convection currents. He draws, therefore, a 

 curve representing the loss of heat due to this cause. All his 

 conclusions must stand or fall with this curve, and I am afraid 

 they must fall. 



After Professors Clausius and Maxwell had deduced theoreti- 

 cally the coefficient of conductivity for gases, a series of cele- 

 brated experiments were made by Stefan, by Narr, by Plank, by 

 Winkelmann, and last, but not least, by Kundt and Warburg. 

 The influence of convection currents has been fully discussedjin 

 these papers and eliminated, and the conclusions arrived at by 

 all.these experimenters are fully in accordance with each other and 

 with theory. It appears, as was expected, that when the effects 

 of convection currents are eliminated, the coefficient of conduc- 

 tivity is independent of pressure until the dimensions of the 

 vessel are comparable with the mean free path of a molecule, 

 and that then the conductivity rapidly diminishes. It also 

 appears that at the pressures at which Messrs. Provostaye and 

 Dessains found that the loss of heat was independent of pressure, 

 convection currents must have ceased to be appreciable, and 

 therefore the great mass of their experiments is fully in agreement 

 with later researches. 



The only exception is found in the case of carbonic acid and 

 nitrous oxide. These abnormal results were not confirmed by 

 Messrs. Kundt and Warburg in the case of carbonic acid, the 

 only one of the two gases which they examined. Whoever reads 

 their account of the difficulty they had in excluding the last 

 traces of moisture, and considers the increased conductivity which 

 such an admixture would' produce as the pressure diminishes, 

 will have no difficulty in explaining the anomaly. At any rate I 

 do not think Prof. Stoney will be inclined to base important 

 conclusions on unconfirmed experiments on two gases in which 

 we should expect the effect, owing to their density, to be parti- 

 cularly small. The discovery of Master Gerald Stoney, who 

 found that a red hot wire was cooled when a tin can containing 

 water was brought sufficiently close might, I think, have been 

 foretold by the recognised theory. Prof. Stoney, no doubt, will 

 find on reading over the literature on the subject, that what he 

 calls penetration of heat, has hitherto been known under the 

 name of conduction of heat, that it takes place at all pressures, 

 and begins to disappear at the exact point at which he makes it 

 appear. 



The timely calculation of Mr. S. T. Preston in the August 

 number of the Phil. Mag., shows that any theory of the radio- 

 meter which makes the action depend on the comparatively large 

 ratio of the mean free path to the dimensions of the vessel, must 

 necessarily be wrong. Arthur Schuster 



The Proposed Channel Islands' Zoological Station, 

 Aquarium, and Piscicultural Institute 



I AM very anxious that this project ' should succeed, mainly 

 because of the facilities it will afford to inland aquaria, in 

 procuring living animals cheaper, better, more variedly, and 

 more systematically, than at present. This, I believe, will form 

 the most profitable part of the undertaking. 



« Referred to in Nature, vol. xvii. p. i«2. 



