We have already mentioned the experiments in which it was found that heat passed less readily through solids, when they were of a spongy or porous texture, than through those that were more dense and compact. The same kind of ef fect was found by Rumford to take place in fluids when substances were mixed with them, which, by their viscidi ty, or other analogous property, would prevent the motion of their particles, and thus put a stop to the circulation, which we have described as conveying the heat through their different portions. He compared the time which was necessary for a quantity of pure water to change its tem perature, by a certain number of degrees, with the same bulk of water when mixed with clown or with starch ; and he found the times to be as the numbers 6, 91, and 11. He afterwards went on to skew, that, by increasing the quanti ty of matter in the fluid, the difficulty ()Idle passage of heat through it was proportionally increased. Atli part of its weight of down added to water produced a retardation equal to 75, while of the same substance produced an effect equal to 95 : (Essays, vol. ii. p. 221. et seq.) The general principle, therefore, appears to be fully establish ed, that fluids transmit heat, or suffer it to pass among their parts, in a different way from that in which heat is transmitted through solids, not by its being given off from one particle and received by another, but by all the parti cles coming successively to the source of caloric, and indi vidually receiving it from the heating body. Whatever impedes the intestine motions of the fluid, and prevents this circulation of its particles from taking place, stops the passage of heat through it, and confines its effect to the portion which first received it.
The facts brought forward by Rumford may be consider ed as very decisively proving the general principle ; but it has still been questioned, whether there is in fluids that absolute and complete non-conducting power which he at tributes to them. Some experiments to prove that they really possess a small degree of a proper conducting pow er, have been performed by Dr Hope, Mr Nicholson, Mr Murray, Dr Thomson, and Mr Dalton. Their plan was to communicate heat to fluids in a direction different from that in which the supposed currents would act, as by applying it to the upper surface, and by their using every precau tion to prevent its being carried downwards by the sides of the vessel containing the fluid, or by any other counteract ing cause. The experiments appear, upon the whole, to prove their position, and thus to modify the conclusions of Rumford ; but perhaps in all of them there may still be some sources of inaccuracy, which it is very difficult to guard against. The process of Mr Murray appears, upon the whole, to be the most unexceptionable. He formed a hollow cylinder of ice, which served as a vessel in which to contain the subject of his experiment ; for ice, as long as it remains unmelted, being a perfect non-conductor of heat, the objection must be removed, which depends upon the conducting power of the vessel itself. The other arrange ments made by Mr Murray were very ingenious, and ap pear to be well adapted for preventing the passage of heat in any manner except through the actual fluid ; yet heat seemed to be certainly transmitted from the upper surface to the lower part, so as to affect a thermometer fixed near the bottom : (Nicholson's Journal, vol. i. p. 421.) If, then, we are brought to conclude, that fluids possess some de gree of conducting power, it next remains for us to inquire, what is the relative degree in which it exists in different kinds of fluids? This, however, is a question, which it is impossible for us to answer with any accuracy ; because we have no correct means of learning how far the commu nication of heat to them depends upon their proper conduct ing power, and how far upon the motion between their par ticles. It may be inferred, from some experiments of Rum
ford's, that mercury is a better conductor of heat than oil or water ; and this might be expected to be the case. It is natural to suppose that mercury, so far as it is a fluid substance, conveys caloric, by internal motions among its particles, like other fluids, yet that it still retains its me tallic property, and may conduct it from particle to par ticle.
The subject of the conducting power of liquids has very recently been investigated by Dr Brewster, who has re moved every doubt that could remain respecting the exist ence of this property in fluid bodies, and has pointed out a simple method, by which the conducting power of all trans parent fluids may be rendered visible to the eye, and easily measured, without the aid even of a thermonieter. The apparatus by which these experiments were made is shewn in Plate CCLXXXIX. Fig. 8. where ABCD is a tin ves sel about a foot long, having two openings E, F, in which two pieces of parallel glass are firmly cemented. A heat ed iron GIT is suspended by a wire LM, and a stand RT, having a small circular aperture P, about one-eighth of an inch in diameter, is placed very near the plate of glass F, so as to be seen distinctly by an eye at E. This aperture is capable of being raised and depressed by a finger screw S. The vessel is now filled with water, or any other fluid, near ly to the top, and the eye being placed at E, and the aper ture P raised so as to be seen through the upper stratum of fluid, it will appear of a perfectly circular form. Let the heated iron be now suspended, as in the Figure. In a few seconds the upper stratum of fluid will expand, as the heat penetrates the mass, and the circular aperture will have an elliptical form, as shewn at a, Fig. 9. By depressing the aperture P, so as to allow the rays which diverge from it to traverse the fluid strata at different depths, the aper ture will have the appearance shewn at b, c, becoming per fectly circular at the point c, to which the heat has advanced. These appearances are the necessary consequence of the propagation of the heat downwards, which diminishes the density, and consequently the refracting power of the up per strata. A very curious phenomenon now takes place upon removing the hot iron GH. The uppermost stratum n, Fig. 9. which was formerly the hottest and the least dense, now gives out its heat to the superincumbent air, and to a certain depth c, Fig. 10. the fluid diminishes in density, while from c to f its density increases. If, in this situation, we again examine the circular aperture, it will be found to be extremely elliptical at a, Fig. 10. having its larger axis horizontal. The ellipticity will gradually di minish towards c, where its form will be circular, and be low c it will again become elliptical, the larger axis of the ellipse being now in a vertical direction. In these experi ments the heat propagated down the sides of the vessel pro duced no effect, and if it had produced any, it could have been easily distinguished from that which was occasioned immediately below the hot iron.