^{page 381 note *} Birch's History of the Royal Society, vol. IV. p. 263.

^{page 394 note *} This experiment may perhaps be thought to give countenance to the opinion of the very ingenious Count Rumford, that fluids cannot conduct heart, and that no interchange of heat can take place between the particles of bodies in a fluid slate, seeing that for two days the fluid at the bottom of the vessel never fell below 39°, though the surface was at 32°.

From the circumstances detailed in his seventh essay, the Count concluded, that heat cannot descend in a fluid. From the present, it might with equal justice be inferred, that heat cannot ascend.

Had I not the fullest conviction that this celebrated philosopher has pushed his ideas too far, I might be disposed to consider this experiment as according well with the hypothesis.

Soon after the interesting speculations of the Count appeared, I began to investigate the subject; and, by a pretty long train of experiments, which I have annually taken an opportunity of detailing in my lectures, satisfied myself that he assigned to fluidity a character that does not belong to it. Though since the date of these experiments, the public has become possessed of several series, well devised, and, in my opinion, of themselves conclusive, it may yet be worth while to state the tenor and result of them, by which the value of their testimony in favour of the conducting power of liquids may be estimated.

The experiments were of two descriptions.

The one set, of the same nature nearly with those of Count Rumford, was designed to examine, Whether heat, when applied to the surface, can descend in a fluid; and the other to discover, Whether, on the mixture of different portions of fluid at different temperatures, an interchange of caloric takes place between the particles;—Water, oil and mercury, having been the subjects of the Count's experiments, were employed for the first set.

To explore the conducting power of water and oil, the apparatus which I used consisted of two vessels of tinned iron, both cylindrical, and the one somewhat larger than the other. The larger had a diameter of eleven inches, and into it were poured the subjects of the trial, to different depths on different occasions. The smaller was ten and a half inches in diameter. By three hooks it was suspended within the larger pan, in such a manner, that the bottom of it exactly reached and came in contact with the surface of the fluid. This smaller vessel became the source of the heat, by being filled with boiling hot water. The water was changed frequently, care being taken to avoid, by the use of a syphon, all agitation and disturbance.

In experiments of this nature, the difficulty is to prevent the conveyance of caloric by the sides of the vessel. I attempted, and, I think, I succeeded, in overcoming this difficulty, by encircling the larger vessel, at a height exactly corresponding with that of the surface of the fluid within, with a gutter or channel about half an inch in depth; and by causing a stream of cold water to slow constantly through a syphon into this gutter, while from the opposite side it ran off by a small spout.

The water was several degrees colder than the subject of the experiment; and keeping cool the portion of the vessel with which it was in contact, it intercepted the heat that would otherwise have travelled by this route to the bottom.

For mercury I had recourse to vessels of glass.

In all the experiments a thermometer bore testimony that the caloric descended from the surface to the bottom of the fluid, and demonstrated, at least to my conviction, that fluids can conduct heat.

The progress of the heat, however, was very slow, and attested the important fact, for which we ought to be thankful to the Count,—That fluids are very bad conductors.

The second set of experiments was calculated to examine, in a very different manner, the position, That all interchange and communication of heat between the particles of fluids is impossible.

When a hot and a cold fluid are mixed together and well agitated, very soon an uniform temperature is produced. This equality must proceed either from a communication of heat from the warmer to the colder fluid, agreeably to the common opinion, or from a perfect intermixture of hot and cold particles, according to the notion of Count Rumford. To which cause it ought to be attributed, I conceived I might discover, by ascertaining whether, after such an intermixture, any separation of the hot and cold portions took place. If the equilibrium of temperature be owing to intermixture without interchange of caloric, the hotter particles, as soon as the agitation ceases, ought, by reason of their greater rarity, to accumulate, to a certain degree, at the surface, and there exhibit a temperature above the common one.

I first tried water, and mixing this fluid boiling-hot, with an equal quantity nearly ice-cold, in a stoppered glass jar, I shook them well for a short time.

I then noticed the resulting temperature, and raising the ball of the thermometer towards the surface, I had an opportunity of observing, that it never rose the smallest portion of a degree above the common temperature which had been established.

I next made a similar experiment with alcohol, selecting it on account of its remarkable dilatability. I shook well, for half a minute, a mixture of equal parts of alcohol at temperature 40° and at temperature 170°. The resulting temperature of the mass was 104°.

Now, if this was a mixture of particles at 40° and at 170°, as the difference of specific gravity between the fluid at these temperatures is very considerable, some separation of the warmer and lighter particles from the others, ought, I conceive, to have taken place. The temperature of the top, however, never indicated the arrival of warmer particles. It never ascended above the point of equilibrium.

From these experiments I concluded, that the uniformity of temperature was established by an actual communication and interchange of heat between the particles.

It may not, however, be improper to state, that Count Rumford, with whom several years ago I had the pleasure of conversing upon this subject, alleged, that the intermixture might be so complete as to prevent any separation whatever.

If it be a property essential to fluidity, that heat cannot pass from one particle to another, the particles of different fluids ought to be equally, incapable of imparting caloric mutually to each other. Unfortunately, however, for the speculation, the caloric is so communicated. Though, à priori, I entertained no doubts respecting the result of the experiment, I poured a quantity of olive oil which had been heated by immersion in a vessel of boiling water for half an hour, upon an equal volume of water of 38°, and agitated the mixture, by shaking for a quarter of a minute. The common temperature produced was 78°, and the heat had gone from the oil into the water; for when the fluids separated, and had arranged themselves according to their specific gravity, both of them had the same temperature of 78°.

The experiments of the two descriptions now recorded, left on my mind little doubt that the Count had overstrained his conclusions.