The different conducting power of bodies produces a great difference in their action upon the nerves in exciting the sensations of heat and cold, although the bodies indi cate the same temperature to the thermometer. Thus it is well known, that in the same apartment, when it con tains no source of local heat, the different articles of furnitut e will convey to the hand very different degrees of warmth. Metals feel the coldest, stone or marble is the next in degree, while wood, and still more woollen stuffs, produce little or no feelings of any change of temperature. The effect in this case depends upon the different conduct ing power of these substances ; the human body being al most always wanner than the objects that are contiguous to it, its heat is abstracted by them ; but this operation goes on in proportion to their conducting power, i. e. to the respcc tit e velocity with which they are enabled to remove the heat from our body. It obviously follows, that when we wish to retain the heat in any substance that is warmer than the surrounding medium, we must enclose it in had conductors of heat ; and that exactly the same method must be pursued, if we wish to keep any substance at a lower temperature than the atmosphere, or other contiguous bodies. Thus we envelope ourselves in woollen cloth or furs when we wish to retain our natural warmth ; and we should employ the same method to prevent the melting of ice or snow.
We have already observed that one cause, which obvi ously tends to affect the conducting power of bodies, is their possessing a spongy texture, by which portions of air become, as it were, entangled in its pores, and thus seem to prevent the communication of heat. But we ob serve a very marked difference in the conducting power of bodies that are perfectly solid, and where no air, or at least no quantity that can be supposed capable of produ cing any effect of this kind, is present. \Vhat is it, in this case, that causes the difference in the conducting power of bodies ? Is it an attraction which the heat pos sesses for the particles of the solid, or do the experiments of Rumford and others lead us to regard the operation as of a more mechanical nature, as if there was something in the arrangement or shape of the particles, which re tards the passage of the heat along them ? These are ques tions which at present it appears to be beyond our power to answer. Those cases in which a greater or less degree of density, or of aggregation of the parts of bodies, pro duces a considerable effect upon their conducting power, would induce us to suppose, that the worst conductors should be regarded as merely the most effectual retarders of heat, and the best conductors as simply those that have the least power in retaining the heat that has been impart ed to them. But this view of the subject seems scarcely to apply to metals and other solids of a similar kind. The radiation of the surface may be supposed to have some in fluence in these cases ; and M. Poisson goes so far as to conjecture, that the conducting power of solids, generally, is to be regarded as a kind of radiation from particle to particle, operating at very small distances : Phys. t. lxxx. p. 434. et seq.) Upon this point, however, it does not appear that we have any data which can enable us to form a decided judgment.
\Vithout, however, entering upon any abstruse theory on the subject, for which the present state of our infor mation does not seem to afford a sufficient foundation, we may assume, as the most natural deduction from the facts, that the conducting power of bodies depends principally upon three circumstances. It is affected partly by the mechanical relation of their particles to each other, partly by an attraction between the heat and the particles of which the body is composed, and partly by the radiating power of the heat. The heat which escapes from the surface will tend to draw from the interior a portion of its remain ing heat, in order to supply what has been lost from the external parts. There are also other causes, which, al
though perhaps less efficacious, are not to be neglected. The consequence of adding heat to a body is, to expand it in all its dimensions ; but, by this expansion, it appears to acquire a greater capacity for retaining heat, so that it will become more disposed to carry it from other bodies, and to diffuse it over its own substance, and thus to have its conducting power increased. The effect which caloric has in altering the state of bodies, may likewise materially affect their conducting power, according to the nature of this change. Thus, if we throw a certain quantity of heat into a metal, it is reduced to the fluid state, by which its relation to heat, and the manner of conducting it, will be much affected. There is some reason to suppose, from an experiment of Pictet's, that heat passes more readily up wards than in the contrary direction. He enclosed a me tallic bar vertically in a vacuum, and, heating it exactly in the centre, observed the effect produced upon thermo meters attached to each end ; when it appeared, that the one at the upper end was affected sooner than the one at the bottom : on Fire, § 33.) The experiment is ingenious : but there are some points connected with it, which render the result rather dubious.
We now proceed to consider the manner in which heat passes through fluids. Fluids differ essentially from solids in their particles being movable among each other ; and this circumstance, as we shall find, acts a very important part in the transmission of heat. Fluids, like solids, are expanded by caloric, and of course become specifically lighter ; and therefore, when heat is partially applied to them, in consequence of this alteration of gravity, the parts change places with respect to each other, the lighter or heated part rises to the surface, while the colder part sinks to the bottom. It is obvious, then, that heat may be supposed to pass through fluids in two ways ; it may either be transmitted from particle to particle, as is the case with solids, or it may be conveyed by a change of gravity in the substance itself, producing a species of circulation among its parts. Many facts that fall under common observation, shew that this circulation takes place in partially heated fluids, and prove that it is much easier to cause heat to pass' upwards through them, than in the contrary direction ; but the impossibility of heat being propagated downwards through fluids, was not admitted until after the experiments of Rumford. These experiments, which are perhaps the most ingenious of any which he ever performed, have been generally regarded as sufficient to prove the general prin ciple to which we have already alluded, that when any por tion of a fluid is heated beyond the temperature of the re maining part, it rises to the surface, in consequence of its comparative levity ; and if the heat be applied to the bot tom of the vessel, which is usually the case, the successive portions of the fluid, as they receive the heat, rise in their turn to the upper part, until the whole acquires one uniform temperature. On the contrary, if the heat be applied to the top of the fluid, it is only the upper stratum which be comes heated, the remainder retaining its former tempera tore. Rumford. placed a portion of ice on a certain quan tity of boiling water, and found that it was melted in about 3 minutes ; hut when the same substances and apparatus were used, except that the ice was fixed at the bottom of the water, several hours elapsed, and yet the ice was not completely thawed. To render the operation still more striking, it was so managed that water was made to boil in the upper part of a cylindrical vessel, while the lower part remained full of ice. Essays, vol. ii. p. 241, et seq.