water.

In another case, from a deficiency of water (through design or neglect), the boiler exhibited the ordinary appearance of having been overheated: some plates were softened, bulged, and ruptured (one of which plates I have now in my possession), and the seams and rivetings, not only along the bottom, but extending to the sides and crown, were deranged, requiring new riveting and caulking. Although accidents of this sort are of daily occurrence in the manufacturing districts, the present was attributed to some imaginary expanding and contracting influence, under an ingeniously supposed alternate heating and cooling process; for notwithstanding the entire bottom and flues were exposed to an uniform stream of heated products of combustion from the furnace, the theory assumed that there was a body of air at one time driving the flame against the boiler bottom and causing it to expand; and again, that the same air caused the same part of the boiler suddenly to contract, until the rivets were dragged in opposite directions (like a man attempting to pull his own arms off), and the boiler so became leaky! The ingenuity of this mode of making the boiler leaky, might however have been spared, had the engineer for a moment considered, that, as this boiler, a new one, had been leaky from the beginning, and even to a considerable extent; and there was no water gauge for exhibiting the height of the water within, the deficiency which led to the overheating and injury, might, without any great stretch of fancy, have been occasioned in this natural way. The air, in this instance, no doubt, was "crude air (vulgo, pure air), and doubtless would not have produced such dire effects had it been "diluted with nitrogen and steam," as recommended by Mr. Armstrong, to whom the above ingenious theory of expanding and contracting is attributed.

When practical men will thus strain after new and speculative sources of injury, while they overlook natural and ordinary causes, it is time that further inquiry be made, and the subject taken out of the hands of quacks and pretend

ers.

A closer view of the principles which practically govern the conduction of heat through metallic bodies, will help to clear away those erroneous notions, and

BY C. W. WILLIAMS, ESQ. 183

bring the question within narrower and better defined limits.

Hitherto, I have examined the subject with reference to the illustrations which practice presented: I will now draw some from the statements of others.

"If a metallic bar," says Professor Brande, speaking of conduction, "be placed in connexion with a constant source of heat, and we wait till it has taken up a permanent state of temperature, we shall find that for distances from the source, taken in arithmetical progression, the excess of temperature above the surrounding medium, will form a geometrical progression."

We have here a defined connexion between those rates of progression, and a permanent state of temperature " in the conductor bar. Now this "permanent" state corresponds with what I have termed the statical heat of the bar, and which indicates the degree in which the metal will be affected by heat, injuriously or otherwise. When Mr.

Brande uses the term 66 permanent state," it is not to be taken as referring to any particular temperature, but merely to the condition (as to temperature) in which the conductor may then be placed; and only cæteris paribus, as regards the surrounding state of things; inasmuch as each, and every change, will induce a new and varying state, or statical temperature. This, however, will be more apparent as we proceed.

The point now under consideration is this, how far the nature of the recipient will influence this permanent state of statical temperature? The bar and its state, as mentioned by Mr. Brande, we see had reference solely to one kind of recipient for the conducted heat, namely, the air. If, however, it be brought into connexion with a different class of recipients, as oil, mercury, or water, a new and different pro tempore, though "permanent state," will be established. In other words, the statical heat will vary as the circumstances which govern it, and which I am endeavouring to show are solely attributable to the nature and properties of such recipient.

What then are the circumstances which modify or govern the statical heat? I here prefer using the term statical, rather than permanent, as it avoids confusion, and, without any apparent contradiction, involves the idea of a temperature, though still defined, yet varying

according to circumstances. This statical heat, then, must be considered under two relations, namely, as indicating, first, the amount of power exercised by the recipient in carrying away the heat received from the conductor, and secondly, the amount of influence which such power exercises on the conductor itself.

With respect to this direct connexion I find but little notice has been taken. Professor Kelland, of Cambridge (now of Edinburgh), has however, in his "Theory of Heat," distinctly recognised it as demanding attention. Under the head "Convection," he observes, "A mode of loss of heat analogous to conduction, is that to which Dr. Prout has applied the term convection. When a hot body is in contact with the air, the part next the body becoming more elastic (rarefied), flies off and is supplied by colder portions; thus the heat of the body is conveyed away more rapidly than it would be if the air were not in motion. It is obvious that this circumstance will materially affect all experiments on the motion of heat, in which it is hardly possible to estimate the effects due to this cause." Now it is this very circumstance, as it affects the condition of the metals employed in evaporative processes, that I am desirous of examining; inasmuch as this "convection," or carrying away, as the term imports, will, in practice, be found to be the primary source of good or evil. Conduction being but the secondary, or induced cause.

Professor Kelland, further on, in examining the mode by which heat is transferred from one part of a body to another, observes, "When speaking of solids, this is called conduction. It is clear, from the term itself, that we do not include, under this head, the transfer of heat by radiation; nor do we include that transfer which takes place amongst the particles themselves, carrying with them the caloric they have acquired, which we designate convection." It is manifest from this clearly defined distinction between conduction and convection, that as the former refers to solids, so the latter refers to fluidsaeriform or liquid; inasmuch as the "carrying with them" the caloric they have received, involves a mobility among the particles which is inconsistent with the nature of solids, while it is a correct definition of that of fluids.

Practically speaking, then, this distinction is peculiarly applicable to our present inquiry, the whole turning on this point, whether conduction or convection be the prime mover in producing those fluctuations of temperature in the conductor which lead to useful or injurious results. Let us now examine, practically, the effect of this convection, or power of carrying away the heat, and the extent to which it influences the heat conducted, both as regards quantity and rapidity.

Let us suppose that a given quantity of heat is passing through a metallic bar, with a rapidity due to its maximum power of conduction; and that it is taken up, or absorbed, by the recipient, with a corresponding rapidity. The temperature of the conductor will then truly indicate that permanent state, referred to by Mr. Brande. If, however, this absorbing power on the part of the recipient be diminished, the rate at which the heat passes from the conductor will also be diminished in the same ratio; and as a necessary corollary, the rate of the current of heat through the conductor will be reduced in a corresponding degree. The practical question then is, to what extent will this " permanent state" or statical temperature of the conductor be affected by such diminished power of convection in the recipient; for this statical heat will ever influence the question whether such conductor be under or overheated. The following experiment will illustrate some of the relations which affect this inquiry.

The annexed engraving, it will be seen, exhibits a state of things, corresponding in principle, to that referred to in your Number 967, in which the thermometer indicated the statical heat of the conductor bar. I have now extended the illustration by lengthening the bar and introducing three thermometers, thus, to indicate the heat at three different sections, and mark more accurately the varying temperatures.

In this engraving, as in the former one, let A represent the conductor bar; B the vessel to contain the recipient, water, or air, or whatever it may be; C the cock for letting off the liquid, when employed as a recipient; D1, D2, D3, the three thermometers, indicating the statical heat at their respective portions of the bar; E the constant source of heat, being that from a powerful gas-burner, furnished with a metallic dome, on the

principle of the "Solar Lamp ;" and F, the protecting shield. To these I have now added the chamber G, which was filled with charcoal powder, to act the part of a non-conductor. By this means, radiation from the bar was prevented; the thermometers received heat alone from their points of contact with the bar -the internal streams of conducted heat being confined, as though it were water in a canal or tube.

By the following table, No. 1, we find that in fifty minutes the water in the vessel B was raised to the boiling point, by the issue of what I have called the stream of dynamical heat, from the end of the bar at A. A "permanent state" of temperature was then attained; the thermometers at their respective distances from the source of heat, in arithmetical progression, indicating the temperatures, in geometrical progression, (or sufficiently near for our purpose, for accuracy in this respect was not attempted), of 400°, 309°, 242,-the extreme difference being 158°. Let us now suppose this to be the state of things in which the maximum power of conduction of the metal was brought into action. It is here manifest that the amount of statical heat in the conductor will bear some ratio to the joint power of conduction in the bar and convection in the recipient water.

On the recipient being changed, by letting out the water and letting in the air, a new state of things is induced. The power of convection will be diminished, (air being a worse-recipient than water); the quantity of heat received from the conductor in each unit of time will also be diminished; and the rate at which the stream of conducted heat passes through the bar will be influenced in a corresponding degree. The source of heat, however, remaining constant, the inevitable result is that accumulation will take place in the bar, and the thermometers instantly indicate an increase of statical heat. This I believe to be the rationale of the process, and this is exactly what we find confirmed in practice. Now, if the velocity of the convection of the new recipient, air, be ascertained, as well as the conductive power of the metal bar, and both taken as constants, we shall be enabled to approximate to the amount of statical heat in the conductor, and the degree in which the metal will be affected.

It is worthy of notice here, that contrary to what might have been expected

By inspection of this table, No. 1, we find that between the periods, 55 and 60 minutes, (say in five minutes), the thermometer No. 3, (the farthest from the heat), had risen 42°, that is, from 242° to 284; whereas, No. 1, (that nearest to the heat), rose but 13° in the same five minutes, say from 400° to 413°; thus proving, that what may be called the wave of statical increase had flowed backwards towards the source of heat, and in a manner strictly analogous to what would take place if, instead of a stream of heat, it had been a stream of water in a tube or canal.

The bearing, practically, of this illustration is, that the temperature of a conductor plate or bar will be influenced, not so much by the quantity of heat imparted to the conductor, as by the absorbing or convecting power of the recipient. This also shows the practical error of attributing injury to what is going on outside the boiler, or in the plates themselves, while we neglect what is taking place within the boiler, where the real source of injury exists.

This table further shows, that, by reason of the restricted power which the air possesses of carrying away the heat from the conductor bar, the temperature of the latter rapidly increases, until a second "permanent" state is established; and when the three thermometers stand, respectively, at 476, 415°, 362°, the extreme difference here is but 114°, whereas with water as the recipient it was 158. This relation of 114° to 158°, then, indicates the ratio of the conductive powers of the bar brought into action by the influence of the respective recipients, air and water. Again, we see that this new statical heat of the bar is in the ratio of the convecting power of the recipient, the thermometer nearest to the source of heat rising from 46° to 476, while that most removed rose from 242° to 362°. Many other instructive relations might here be noticed, if time and space permitted.

The more accurately to observe the varying relations between the statical and dynamical heat, under a different state of things, I reversed the experiment, and began the operation with air as the recipient; then changing it to water, so soon as the first permanent state of temperature was established, and which, by the Table No. II., we see took place in forty-five minutes. The results here are equally interesting and instructive.

187

mained stationary, such being the indicated amount of statical heat. (It should here be observed, that these numbers cannot be taken as giving the absolute temperature of the bar, although they approach sufficiently near to give the relative temperatures, under the influence of the recipients, air and water. It will hereafter be shown how near they approach to such absolute temperature.) The recipient being now changed from air to water-the latter being poured into the vessel B-we see the superior convecting power of the latter at once brought into action, by the immediate lowering of the temperature of the bar. On the water being again raised to the boiling point, a permanent state is again established, as in the preceding example, whatever variance has taken place being accidental. It is here important to observe how the analogy between the current of conducted heat and a current of water is maintained; for we see that, although thermometer No. 3 arrives at its minimum, 228°, in 60 minutes, as marked with an *, No. 2 does not reach it until after 70 minutes, and No. 1 until 75 minutes. These experiments were made with great care, although the adjustment of the apparatus required much exactness, to produce uniform results. Throughout the whole we see sufficient to justify the observation, that much remains yet to be done, before this complicated subject be exhausted. The practical inference I draw from the above experiments is, that they afford sufficient proof of the position before stated, namely, that it is not to the furnace or draught, or activity of the fire, we are to look for that accumulation of heat in the plates of ordinary boilers which produces overheating and rupture, but to the recipient, and its powers of carrying away and absorbing the heat which the conductor plate or bar is capable of imparting.

Purposing to continue this subject on a future occasion,

m.nutes. No. 1.

No. 2.

. No. 3.

0

54

54

54

5

170

132

100

10

270

220

178

15 350

288

237

20