chamber; but both are open to many objections, not the least being that the passages get choked and are certainly not under control, and can neither be cleaned nor closed when desired.

With those systems of forced draught in which the ashpits are closed, the funnel damper can be set so as to retard the draught above the bars to such an extent that just the right amount of air is admitted, even when the doors are wide open.

But

Flames. Interesting experiments have been made to show that on mountain tops, where the air pressure is much reduced, and also in partial vacuums and bad atmospheres, ordinary flames lose their luminosity, while under high pressures they grow smoky, and even the hydrogen and carbonic oxide flames grow luminous. From this it has been argued that draught-retarders and the closing of the funnel damper, when using forced draught, will improve combustion. those very experiments (E. Frankland, 1877, p. 876) go far to show that the combustion is not accelerated; besides, the increase of pressure attainable by these means is so slight in comparison with the atmospheric pressure variations, that the influence would have shown itself before now if it were true that more heat can be got out of coal with a high than a low atmospheric pressure.

Some interesting facts as to the igniting temperatures of various substances will be found in the chapter on Fuels and Combustion.' There will also be found a good deal of information on the heatingpower of fuels, and on various methods of measuring it, as well as for determining, from the funnel gases, whether the combustion is perfect or not. Some of these tests are so simple that they can easily be carried out at sea.

Furnace Doors. As regards economy of fuel, furnace doors unquestionably rank amongst the most important attachments to a boiler, and innumerable are the patents in connection with them; but the very fact of numberless ideas having been published makes it impossible to deal with any of them as exhaustively as might be wished, and only those will be mentioned which have found their way into the stoke-hole.

α

The door may consist of a single plate, fitted with hinges and a latch, and perforated, chiefly at the upper edge, or fitted with a gridiron or other contrivance for regulating the admission of air. The objection to this arrangement is that, on account of the direct radiation from the fuel, the door very soon gets excessively hot and warps, and even cracks. The presence of an air regulator is an advantage, if properly used, but the very reverse if its action is not understood (see p. 5). Some doors are hinged, so that they can be kept partly open. In order to protect the door from the heat, an inside screen should be fitted, which can be renewed when burnt away; for, as it cannot be kept as cool as the outer one, it suffers more severely. A simple arrangement is to rivet a plate, a, fig. 5, having a number of holes, to a solid one, b. The air then enters at the circumference and passes through the various holes, as shown by the arrows.

FIG. 5

In other cases the internal plate is so fitted that the current of air is directed either upwards or downwards, according to the views of the respective engineers. (See figs. 6, 7.) An idea seems to prevail that by leading the air through complicated passages it collects heat, thereby facilitating combustion; but this warming is so slight that it does not justify expensive arrangements. Wide dead plates in combination with doors which admit air only at the top keep the latter cool. As already mentioned, hinges are sometimes constructed so as to keep the doors partly open; and certainly they should all be arranged so that they will keep quite open when coaling, particularly during rough weather. Some of the contrivances used for this purpose are very simple, as well as fairly efficient. All of them should be strong, and capable of being worked in the easiest possible manner, for a fireman's chief tool is his shovel. Besides, on account of the heat, anything belonging to a fire door cannot well be touched by hand. One way of keeping the doors open is to balance them by either weights or springs, but with most of these arrangements the ashpit gets closed during firing, whereby that part of the air supply which passes through the fuel is lessened, and that which passes through the door increased, doing harm where it chills the various plates.

FIG. 6

FIG. 7

Door Frames suffer in the same way as the doors to which they are attached, unless they too are properly protected from the heat, either by baffle plates with air admissions at the back, or more generally by firebricks. It is never good to make the door frame in one piece, as it is sure to crack. With large furnaces, of 40 ins. diameter and above, two doors are sometimes fitted. Attempts have also been made to fit feed heaters at these points, but evidently without success.

Fire Bars. It cannot be said that this subject has been neglected by inventors, for the patents in connection with it are innumerable; but it is certainly very unsatisfactory that, after trying various novelties, engineers always fall back on the old pattern, viz. length 2 to 4 ft., air spaces to in., depth 3 to 5 ins., and thickness to 1 in. at top, tapering 1 to 14 in. per foot of depth. Even for forced draught, if not excessive, the above dimensions give good results, although the Admiralty, who use wrought iron or steel instead of cast iron, make the bars about 3 in. thick, while the air spaces are reduced to ĝ and even in.

It has been argued that by reducing the upper surface of the bars a smaller area is exposed to the heat of the fire, and that such bars will keep cooler; but a glance under the grate of a furnace is sufficient to convince anybody that as much if not more heat is received by radiation by the sides of the bars as by their upper surfaces. A closer examination will also show that each bar is surrounded by a visible layer of trembling hot air, which is moving upwards and seems to be about in. thick. It is the heating of this thin film of air

which keeps the bars cool. These facts might lead to the conclusion that a distinct advantage would be gained by reducing the thickness of each bar, and fitting more of them, because thereby relatively more cooling surface is obtained. But if, as seems necessary, the air spaces are left as wide as before, each bar will receive an extra amount of heat, so that the thin bars will probably grow quite as hot as the thick ones, and if that is the case they are at a great disadvantage, for being thin they are sure to get bent sideways. In fact, the only way to use them is to pack them tightly into the furnace, so that they can neither bend nor twist.

Where great trouble is experienced water in trays is placed in the ashpits. This seems capable of abstracting sufficient heat, but the air channels under the bars are seriously reduced, and salt water in the furnaces is not a desirable object, as it causes a lot of corrosion. On forced-draught trial trips it is often necessary to keep hoses playing sea water into the ashpits. No doubt the accumulation of ashes and red-hot small coal in the ashpits keeps the bars hotter than they should be, and, as they also seriously interfere with the draught, it would be a great advantage if means could be devised for removing them.

Another plan for keeping the bars cool is to make them deeper. Heat travels so very quickly in metals that the small extra distance which it has to go before reaching the lower edge hardly affects the result, which, roughly stated, is, that the temperature of the bars is inversely proportional to their depths, or more correctly to their exposed surface, and that their rigidity (horizontally) is proportional to their depth and to the square of their thicknesses. The horizontal deflection of a bar, to which a definite curvature has been given, is proportional to the square of its length. These views lead to the following formulæ, with whose help the small table has been compiled. If not numerically correct, it can at least be used for making comparisons.

I. For a given coal consumption, and for a given length of fire bar, the sum of the sectional areas of the bars contained within 12 ins. of the furnace diameter should be a constant value.

n.t.d. = C1.

II. For a given coal consumption the sum of the sectional areas of the bars contained within 12 ins. of the furnace diameter should be proportional to their lengths.

n.t.d. = C.l.

III. For equal lengths of fire bars the square of their depths should be proportional to the weight of fuel burnt per square foot per hour. d2. = C3.Q.

In the above formulæ the various letters have the following meanings:

n stands for number of fire bars per foot of furnace diameter.

t stands for thickness of fire bars at their upper edges.

d stands for depth of fire bars at their centres.

7 stands for length of fire bars.

C1, C2, C3 are constants.

Q is consumption of coal per square foot per hour.

Qis

Values of the Products n.t.d. (This is the smallest permissible Sum of the Sectional Areas in Square Inches of Cast-Iron Fire Bars which must be contained within One Foot of Furnace Diameter.)

With bars whose air spaces are one-half of their thickness-i.e. 4 ins. per foot of furnace front-the depths would be found by dividing any of the numbers in the table by 8 ins. Thus, with bars 2 ft. 6 ins. long, burning 40 lbs. per hour, the number is 28, and the minimum depths of such bars would be 3 ins., while the thickness might be made in., with 3-in. air spaces, or in. with 1-in. air spaces.

In cases of forced draught these two values are very often equalsay 3-in. bars and 3-in. air spaces. To obtain the depth in their case, the number in the table would have to be divided by 6 ins., so that with a consumption of 40 lbs. the 2 ft. 6 in. bars would have to be 4 ins. deep, and 18-in. bars would have to be 3 ins. deep.

Naturally these values are only approximate, and depend very much on the fuel, but they may serve as a guide when making alterations.

Furnace Diameters.-It will be noticed that as soon as the usual practice is departed from, either by increasing the length of bars or the coal consumption, then their depth grows so great that it seriously interferes with the draught. This influence is particularly noticeable in boilers with small furnaces. Compare, for instance, two flues, the one being 33 ins. in diameter and the other 48 ins. If, as is usual, the lines of the dead plates pass through their centres, then the sectional areas below and above these lines are 3 sq. ft. in the one furnace and 6.3 sq. ft. in the other. With five bars which are 3 ins. deep, the ashpit areas are reduced to 2:3 and 5·3 sq. ft. respectively, or 85 and 1:33 sq. ft. per foot of furnace front. Under ordinary conditions this means that in the one case the air entering the furnace has to travel with a velocity of 12 ft. per second, in the other case its velocity is only 7 ft., and the resistances would be as 3 to 1. An extra inch added to the depth of the bars would increase the one resistance 25%, and the other only about 6%, so that in cases where the performance is low there is much less chance of efficient alterations if the furnaces are small than if they are large.

The sum of the air spaces between the fire bars usually amounts to 33% of the width of the grate, so that if the length is 5 ft. we have 13 sq. ft. of air passage for every foot of furnace front. This being about twice as large as the ashpit area of the small furnace, little improvement would be effected in the draught and combustion of the small furnace by giving wider air spaces, for the depth of the bars would also have to be increased, whereas in the large furnace, where

the ashpit area is sufficient, the alteration might increase the combustion.

On trial trips, and whenever it is desired to obtain the highest performances, the ashpits should be kept clear of ashes at all times, for every inch of piled-up material restricts the draught. Ash ejectors might be used with advantage.

With forced draught the case is different. Except on trial trips, where the air pressure is limited, there is practically no restriction as to the pressure which may be applied, and all the above reasons for allowing large air spaces fall to the ground, and there seem to be no objections against reducing them to the very narrowest limits. In fact, very good results are said to have been obtained in some foreign vessels where the air spaces have been reduced to in. and a high-pressure blast introduced into the ashpits. The necessity for opening the lower doors to remove the ashes does not exist, for none can fall through; they are all fused and form clinker on the bars. Of course such bars could not be used for natural draught; but natural draught in the Navy is simply another name for a very low fan pressure, which could be increased if desired.

The following advantages should not be overlooked :-By reducing the air spaces till they are mere slits, the chief resistance to the air passage is found at these points and not in the fuel, and the draught will hardly be affected by considerable variations of thickness of the fires. The chief combustion takes place just over the slits, and little, if any, over the centre of the bars, which remain covered with comparatively cool fuel or ashes, and are therefore not exposed to so much heat as the narrower bars with wide air spaces. An attempt to indicate this difference is shown in figs. 8 and 9.

The Burning of Fire Bars is often the consequence of irregularities in the upper surface of the grate. It is but natural that if several air spaces are blocked (fig. 10), either by clinker or when the bars are bent, then the air can cool only one side of such bars, and these must grow hotter than the others. If, in addition, fuel should get wedged into the wide space, nothing will prevent the corners of the bar from burning, and as these waste away the fuel sinks lower and lower, and it is only a question of time as to when it will have effected the destruction of the bars; naturally the adjoining ones, which are then exposed to the same action, will suffer in the same way.

An examination of the grates of a boiler which has been worked hard will show that the bars have all sagged more or less, and that