Both Professor Leslie and Count Rumford performed a number of experiments, similar to the one mentioned above, on the action of the surfaces of vessels in promoting or re tarding the escape of heat from bodies contained in them. We arrive at many conclusions which are singular and im portant ; but they so obviously follow from the facts which have been already noticed respecting the radiation of heat, that it will not be necessary to dwell very long upon them. In Professor Leslie's experiments, it was always found that the cooling of water in metallic vessels was accelerated by coating the vessel with paper, paint, isinglass, or other similar substance ; and while the coating was comparative ly thin, this acceleration appeared to be in proportion to its thickness. So that although in these cases the conducting power must have beenitijured, yet this defect was more than counteracted by an increase of radiation. It can, however, scarcely be doubted, that were we to go beyond a certain limit, we should End, that, by augmenting the thickness of the coating, the diminution of the conducting faculty would prevail over any advantage which might be derived from the increase of the radiating power. There are several reasons which induce us to suppOse that this latter quality is effected principally by the nature of the surface, while the conducting power must depend equally upon the na ture of the substance through its whole thickness. It does not appear that any experiments have yet been performed, by which we can ]earn in what point this balance of power exists. Rumford's general conclusions entirely coincide with Professor Leslie's. He found that the same metallic cylinder, which, when filled with water, required 55 minutes to cool through a certain number of degrees, required only 43 minutes when covered with a layer of glue, 36 when covered with linen, and 34 when coated with a varnish of lamp black. Phil. Trans. 1804, p. 90.
When bodies are surrounded either by a fluid or by air during their cooling, the rate of the process will be consi derably affected by the motions that take place among the particles of the investing medium. The action of a current of air in promoting refrigeration is too well known to re quire illustration ; and this most evidently be an effect en tirely independent of radiation, and in a great degree, at least of the conducting power of the air. Without decid ing the question, whether air be an absolute non-conductor, it is rendered probable, both from circumstances that fall under daily observation, as well as from the direct experi ments of Rumford and others, that air is not a good con ductor of heat, and that one of the most effectual methods of retaining the temperature of a body, is to surround it with a stratum of air, so confined, that no internal motion can take place among its particles.
Hitherto heat has appeared only as a simple substance, capable of being communicated through the parts of solid bodies with different degrees of velocity, according to their conducting power ; the presence of the heat being in every case marked by an expansion of the solid body corresponding with the temperature. These views, however, though uni
versally received both among chemists and natural philo sophers, have been completely disproved by some recent: •experiments made by Dr on the propagation of heat along glass, obsidian, scmiopal, muriate of soda, fluor spar, alum, gum copal, rosin, horn, amber, tortoise shell, and other substances. We shall endeavour to ;lay before our readers as brief and perspicuous an account as we can of the new properties of heat which were discovered in the course of these experiments, without anticipating any of the phenomena, which more properly belong to the subject of optics.
Let a plate of glass, ABDC, (Plate CCLXXX IX. Fig. 1.) having MN (Fig. 2 ) for its section, be placed with its edge CHID upon a hot iron, or be iat any stay exposed to a source of heat. The heat will be slowly communicated through the substance of the glass in the direction HG FE, and when the heat has reached E, the temperature will be greatest at H, diminishing gradually towards E. The glass therefore, according to the common doctrine, be in a state of expansion, as shewn in Fig. S. being most dilated at H, and least dilated at E. The case, however, is very different. When the heat has entered the glass at the edge CD, the parts of the glass between the edge and the dark line at G are in a state of expansion, diminishing towards G, and at the same instant the parts of the glass between E and F are thrown into a similar state of expansion, while the intermediate portions between F and G arc thrown into a state of contraction, the lines AFB and CGD being the limits between the contracted and expanded portions, where the glass has suffered no mechanical change. These op posite effects, which are shewn in Fig. 4. are distinctly pro duced before the heat has reached the poiQt F. When the height of the glass plate HE is very small, EP is equal to HG, and the contractions and expansions are almost simul taneously produced ; but when II h is two or more inches, EF is always much greater than HG, and the expansion between E and F is less distinctly seen, being spread over a greater space. I rout these results it follows, •1. That heat, in along a plate of glass, expands a part of the glass, where the heat does not exist in a sensible state.
2. That the heat also contracts an adjacent portion of the glass, where it does not exist in a sensible state.
3. That the heat contracts a part of the glass, where it does exist in a sensible state.
4. That the heat expands an adjacent portion of the glass, where it does exist in a sensible state.
If the plate of glass ABDC, instead of being heated, is brought to an uniform temperature by immersion in boil ing water, and is then allowed to cool in the air, all the ef fects, which we have described, are exactly reversed dur ing cooling, as shewn in Fig. 5. the parts of the glass which were formerly contracted being now expanded, and vice versa.