HAVING now explained in a general way the scientific principles concerned in the use of steam as a motive power, we have next to exhibit the practical application of those principles; and we begin with the subject of boilers. And it may not be superfluous to repeat, that, however useful an acquaintance with scientific principles may be, it would be a most unsafe course to take them as a sufficient guide in the actual operations of engineering; for they have in almost every case to receive material modifications, before they can be considered representative of a judicious and successful practice. At best, scientific principles give only general conclusions, and afford an approximation to the position in which the truth lies; but practice reveals many conclusions which science could not anticipate, and which may modify or even reverse its decisions. Precepts, therefore, derived from theory alone, are but of little value until confirmed by the results of experience, as embodied in successful practice; and to this more important division of the subject we shall now direct our attention.

We have already given a theoretical rule for the dimensions of a land engine chimney. Boulton and Watt's practical rule for the dimensions of the chimney of a land engine is as follows :-multiply the number of pounds of coal consumed under the boiler per hour by 12, and divide the product by the square root of the height of the chimney in feet; the quotient is the area of the chimney in square inches in the smallest part. A factory chimney suitable for a 20 horse boiler is commonly made about 20 in. square inside, and 80 ft. high; and these dimensions are those which answer to a consumption of 15 lbs. of coal per horse power per hour, which is a very common consumption in factory engines. If 15 lbs. of coal be consumed per horse power per hour, the total consumption per hour in a 20 horse boiler, will be 300 lbs., and 300 multiplied by 12 = 3,600, and divided by 9 (the square root of the height)=400, which is the area of the chimney in square inches. It will not answer well to increase the height of a chimney of this area to more than 40 or 50 yards, without also increasing the area, nor will it be of utility to increase the area much without also increasing the height. The quantity of coal consumed per hour in pounds, multiplied by 5, and divided by the square root of the height of the chimney, is the proper collective area of the openings between the bars of the grate for the admission of air to the fire. In steam vessels Boulton and Watt allow 81 square inches of area of chimney per horse power, and in marine flue boilers they allow 18 square inches of sectional area of fiue per horse power; but this proportion appears to be about one-third greater than what is allowed by many other makers, whose boilers, however, are scarcely so conspicuous for an abundant supply of steam. The sectional area of the flue in square inches is what is termed the calorimeter of the boiler, and the calorimeter divided by the length of the flue in feet is what is termed the vent. In marine flue boilers of good construction the vent varies between the limits of 21 and 25, according to the size of the boiler and other circumstances the largest boilers having generally the largest vents; and the calorimeter divided by the vent will give the length of the flue in feet. The collective area for the escape of the smoke and flame over the furnace bridges in marine boilers is 19 square inches per horse power, according to Boulton and Watt's proportion. In waggon and tubular boilers very different proportions prevail, yet the proportions of every kind of boiler are determinable on the same general principle. In waggon boilers the proportion of the perimeter of the flue which is effective as heating surface, is to the total perimeter as 1 to 3, or, in some caseɛ, as 1 to 2·5; and with any given area of flue, therefore, the length of the flue must be from 3 to 2.5 times greater than would be necessary if the total surface were effective. If then the vent be the calorimeter divided by the length, and the length be made 3 or 2.5 times greater, the vent must become 3 or 2.5 times less; and in waggon boilers accordingly the vent varies from 8 to 11 instead of from 21 to 25, as in the case of marine flue boilers. In Boulton and Watt's 45 horse waggon boiler, the area of flue is 18 square inches per

horse power, but the area per horse power increases very rapidly as the size of the boiler becomes less, and amounts to about 80 square inches per horse power in a boiler of two horse power. Some such increase is obviously inevitable if a similar form of flue be retained in the larger and smaller powers, and at the same time the elongation of the flue in the same proportion as the increase of any other dimension is prevented; but in the smaller class of waggon boilers the consideration of facility of cleaning the flues is also operative in inducing a large proportion of sectional area. Boulton and Watt's 2 horse power waggon boiler has 30 square feet of surface, and the flue is 18 in. high above the level of the boiler bottom, by in. wide; while their 12 horse waggon boiler has 118 square feet of heating surface, and the dimensions of the flue similarly measured are 36 in. by 13 in. The width of the smaller flue, if similarly proportioned to the larger one, would be 6 in., instead of 9 in., and, by assuming this dimension, we should have the same proportion of sectional area per square foot of heating surface in both boilers. The length of flue in the 2 horse boiler is 19.5 ft., and in the 12 horse boiler 39 ft., so that the length and height of the flue are increased in the same proportion. The Nile steamer, with engines of 110 horse power, by Boulton and Watt, is supplied with steam by two boilers, which are, therefore, of 55 horse power each. The height of the flue winding within the boiler is 60 in., and its mean width 16 in., making a sectional area or calorimeter of 990 square inches, or 18 square inches per horse power of the boiler. The length of the flue is 39 ft., making the vent 25, which is the vent proper for large boilers. In the Dee and Solway steamers, by Scott and Sinclair, the calorimeter is only 9.72 square inches per horse power; in the Eagle, by Caird, 119; in the Thames and Medway, by Maudslay, 11-34, and in a great number of other cases it does not rise above 12 square inches per horse power, but the engines of most of these vessels are intended to operate to a certain extent expansively, and the boilers are less powerful in evaporating efficacy on that account.

The calorimeter of each boiler of the Dee and Solway is 1,296 square inches; of the Eagle, 1,548 square inches; and of the Thames and Medway 1,134 square inches; and the length of flue is 57, 60, and 52 ft. in the boilers respectively, which makes the respective vents 22, 25, and 21 in. Taking then the boiler of the Eagle for comparison with the boiler of the Nile, as it has the same vent, it will be seen that the proportions of the two are almost identical, for 990 is to 1,548 as 39 is to 60, nearly; but Messrs. Boulton and Watt would not have set a boiler like that of the Eagle to do so much work. The evaporating power varies as the square root of the area of the flue, if the length of the flue remain the same; but it varies as the area simply, if the length of the flue be increased in the same proportion as its other dimensions. The evaporating power of a boiler is referrible to the amount of its heating surface, and the amount of heating surface in any flue or tube is proportional to the product of the length of the tube and the square root of its sectional area, multiplied by a certain quantity that is constant for each particular form. But in similar tubes the length is proportional to the square root of the sectional area; therefore, in similar tubes, the amount of heating surface is proportional to the sectional area. On this area also depends the quantity of hot air passing through the flue, supposing the intensity of the draught to remain unaffected, and the quantity of hot air or smoke passing through the flue should vary in the same ratio as the quantity of surface. A boiler therefore to exert four times the power should have four times the extent of heating surface, and four times the sectional area of flue for the transmission of the smoke; and if the same form of flue is to be retained, it should be of twice the diameter and twice the length; or twice the height and width, if rectangular, and twice the length. As then the diameter or square root of the area increases in the same ratio as the length, the square root of the area divided by the length ought to be a constant quantity in each type of boiler, in order that the

same proportions of flue may be retained; and in waggon boilers without an internal flue, the height in inches of the flue encircling the boiler divided by the length of the flue in feet will be 1 very nearly. Instead of the square root of the area the effective perimeter, or outline of that part of the cross section of the flue which is effective in generating steam, may be taken; and the effective perimeter divided by the length ought to be a constant quantity in similar forms of flue and with the same velocity of draft, whatever the size of the flue may be. It is clear, that with any given area of flue, to increase the perimeter by adopting a different shape, is to diminish the length of the flue; and, if the extent of the perimeter be diminished, the length of the flue must at the same time be increased, else it will be impossible to obtain the necessary amount of heating surface. In Boulton and Watt's waggon boilers the sectional area of the flue in square inches, per square foot of heating surface, is 5.4 in the two horse boiler; in the three horse it is 4.74; in the four horse 4:35; six horse, 3.75; eight horse, 4.33; ten horse, 3.96; twelve horse, 3.63; eighteen horse, 3-17; thirty horse, 2-52, and in the forty-five horse boiler, 2.05 square inches. Taking the amount of heating surface in the 45 horse boiler at 9 square feet per horse power, we obtain 18 square inches of sectional area of flue per horse power, which is also Boulton and Watt's proportion of sectional area for marine boilers with internal flues.

If to increase the perimeter of a flue is virtually to diminish the length, it is clear a tubular boiler, where the perimeter is in effect greatly extended, ought to have but a short length of tube. The flue of the Nile, if reduced to the cylindrical form, would be 35 in. in diameter, to have the same area; but it would then require to be made 47 ft. long, to have the same amount of heating surface. Supposing that with these proportions the heat is sufficiently extracted from the smoke, then every tube of a tubular boiler in which the same draft existed, ought to have very nearly the same proportions, so that a tube 3 in. in diameter ought to be about 4 ft. long, supposing the conducting power of the metallic surface through which the heat is transmitted, to be in each case identical. But the metal of small tubes being thinner than that of flues must conduct better, and a tube 3 in. in diameter should therefore be less than 4 ft. long, provided the draft remains such as is due to an area of 18 square inches per horse power. If the thinness of the metal attainable by the tubular form be supposed to increase the efficacy of the heating surface in the same proportion as the increase of surface due to the rectangular form, the length of a tube 3 in. diameter ought to be 3 ft. 3 in.; and it would be of no service to extend its length beyond this point, supposing the flue boiler to be properly proportioned, as by the time the hot air had traversed a length of 3 ft. 3 in. of tube, the heat of the air would have been as thoroughly extracted as in ordinary boilers appears to be beneficial. The tubes of tubular boilers, however, are usually about 6 ft. 6 in, long; but to make this excess of length influential in generating steam, the draft has to be made nearly twice greater than in flue boilers having a sectional area of 18 square inches per horse power, or in other words, the sectional area of tubular boilers must not much exceed 9 square inches per horse power when the tubes are of the length stated. The smaller the tubes are, the shorter they should be made, or the less the sectional area ought to be; and with a sectional area of 10 square inches per horse power, there will be no advantage in making the length of the tube more than from 26 to 32 times its diameter, which will make the tubes from 6 ft. 6 in. to 8 ft. long, when the diameter is three inches, and give from 74 to 8 square feet of beating surface of tubes per horse power. If the sectional area per horse power be increased, the length of tube should be diminshed in the same proportion; for the velocity of the draft varies with the sectional area of tube per horse power, and on the velocity of draft the length of the tube ought to depend.

In locomotive boilers where the velocity of draft is very great, long tubes are employed; but it is preferable to have the tubes of moderate length, and a draft of moderate intensity, as in maintaining a fierce draft by any process, there is a considerable expenditure of power. If, however, with the view of making the draft very slow, a proportion of sectional area approaching that of flue boilers be provided, the result will not be satisfactory, as the smoke will all pass through a few of the tubes, leaving the rest inoperative; though this defect may be in a great measure corrected by partially closing up the ends of the tubes, or even by partially closing the damper. The length of tube multiplied by the diameter, and divided by the area, is a constant quantity both in flue and tubular boilers, or at least nearly so; and when any of the elements are given, the rest can easily be computed by the aid of this proportion. Bury's locomotive with 14 in. cylinders contains 92 tubes of 24th in. external diameter, and 10 ft. 6 in. long, whereas Stephenson's locomotive with 15 in. cylinders, contains 150 tubes of 13ths external diameter, and 13 ft. 6 in. long. In Stephenson's boiler, in order that the part of the tubes next the chimney may be of any avail for the generation of steam, the draft has to be very intense, which in its turn involves a considerable expenditure of power; and it is questionable whether the increased expenditure of power upon the blast, in Stephenson's long tubed locomotives, is compensated by the increased generation of steam consequent upon the extension of the heating surface. When the tubes are small in diameter they are apt to become partially choked with pieces of coke, but an internal diameter of 13ths may be employed without inconvenience, if the draft be of medium intensity. The intensity of the draft may easily be diminished by partially closing the

damper in the chimney, and it may be increased by contracting the orifice of the blast. In most locomotives the velocity of the draft is such that it would require very long tubes to extract the heat from the products of combustion, if the heat were transmitted through the metal of the tubes with only the same facility as through the iron of ordinary flue boilers, and if it were required at the same time that the heat should be as thoroughly extracted. The Nile steamer, with engines of 110 nominal horses power each, and with two boilers having two independent flues in each, of such dimensions as to make each flue equivalent to 55 nominal horses power, works at 62 per cent. above the nominal power, so that the actual evaporative efficacy of each flue would be equivalent to 89 actual horses power, supposing the engines to operate without expansion; but as the mean pressure in the cylinder is somewhat less than the initial pressure, the evaporative efficacy of each flue may be reckoned equivalent to 80 actual horses power. With this evaporative power there is a calorimeter of 990 square inches, or 12-3 square inches per actual horse power, whereas in Stephenson's locomotive with 150 tubes, if the evaporative power be taken at 200 cubic feet of water in the hour, which makes the engine equal to 200 actual horses power, and the internal diameter of the tubes be taken at thirteen-eighths of an inch, the calorimeter per actual horse power will only be 11136 square inches, or in other words the calorimeter in the locomotive boiler will be 11'11 times less than in the flue boiler for the same power, so that the draft in the locomotive must be 11'11 times stronger, and the ratio of the length of the tube to its diameter 11'11 times greater than in the flue boiler, supposing the heat to be transmitted with ouly the same facility. The flue of the Nile, as already stated, would require to be 35 in. in diameter, if made of the cylindrical form, and 47 ft. long the tubes of a locomotive if 18 in. diameter would only require to be 22.19 in. long with the same velocity of draft, but as the draft is 11.11 times faster than in a flue boiler, the tubes ought to be 246-558 inches, or, about 20 ft. long, according to this proportion. In practice, however, they are one third less than this, which reduces the heating surface from 9 to 6 square feet per actual horse power; and this length even is found to be inconvenient. It is greatly preferable, therefore, to increase the calorimeter, and diminish the intensity of the draft.

:

The arched forms of the bottom and sides of the waggon boiler are not, as is sometimes erroneously supposed, given solely for the sake of additional strength, as the want of proper abutments to the arch, renders them only very slightly available for that purpose; and it is an easy thing to prove that greater strength would be obtained if made the reverse way, or convex outwards, instead of concave. Great strength, however, in a lowpressure boiler, working at only two or three pounds above the atmosphere, is not an object of much consideration: but there is another very important matter to be considered; and that is the position of the heating surface as respects the fire grate, or as a recipient of heat, and also as respects the contained water, to which the material of the boiler conducts that heat in order to generate the steam.

As on the proper arrangement of the heating and generating surface it is certain that the excellence of all kinds of steam boilers, both in respect to durability and to economy in fuel, most materially depends, we shall here endeavour to give such a popular investigation regarding it as we hope will serve in some measure as a first lesson or key to the art and mystery of boiler designing, as well as enable the student to appreciate the merits of the common waggon-shaped boiler.

With regard to the boiler bottom forming an arch over the fire, one reason for this construction is very obvious: the heating surface within a given width between the side walls of the furnace is hereby increased; and although this may also be said of a cylindrical boiler, where the arch is downwards, yet in the former the fire and flame are more enclosed within the water, and so far prevented from being expended on a large area of side walls to no useful purpose. The main reason for this construction, however, may be taken to be as follows:- The fire-grate is generally horizontal in its cross-section, and the fuel being generally spread equally thick on the bars in that direction, it follows that the temperature of the furnace chamber, so far at least as it depends on the radiation of heat from the redhot coal, must be somewhat higher immediately over the middle of the firegrate than at its sides; and the effect of the heat thus given out by the burning fire against the boiler bottom, supposing the latter to be disposed horizontally across, will be gradually diminished from the centre towards each side, the effect in all cases being directly as the temperature of the surface of the fuel at any given point, and in the inverse ratio of the square of its distance from the boiler bottom. Provided we knew the exact law of the decrease of temperature from the centre to the sides of the mass of the fuel, we could then easily obtain the true form of the curve describing the arch of the boiler bottom, so that the effect of the heat given out from a fire of uniform thickness would be equal over all parts of the heating surface, which would not then be liable to be unduly acted on or overheated in one part more than another. Although any abstruse determination of this is not necessary here, it having been long realized in practice to a much greater degree of perfection than it is possible to acquire in keeping the fire of a uniform thickness, it may, nevertheless, be useful to give a short illustration of the erroneous principle of the contrary practice ; that is, of arching the boiler bottom downwards, as in common high-pressure or cylindrical boilers. Supposing then that the temperature of the burning fuel is no greater in the middle than at the sides of the grate, and assuming

the average distance of the surface of the fuel below the boiler bottom to be six inches,say five inches in the centre, and seven at the sides a very common case,- the effect of the radiated heat against the central portion of the boiler bottom, compared to that of the sides of the furnace, will then be as the fractional numbers and, or nearly as two to one: and this great disparity of effect is of course still farther increased by the greater heat of the fire at the middle of the grate.

Under such circumstances we need not to be surprised at the very great liability of the bottom plates of cylindrical boilers to become gradually overheated and finally burnt out, while those towards the sides are very little injured. In waggon boilers, it is true, the converse of this does sometimes take place; that is, the seating plates, or those immediately adjoining them, are found occasionally to deteriorate and become injured sooner than those in the crown of the boiler bottom. But this result only corroborates the truth of the general principle we are contending for; at the same time proving the necessity of attending to those apparently unimportant mechanical details so commonly overlooked by the scientific engineer; for the defect always arises either from the crown of the arch being too high, or by having the fuel too thick on the bars, so that its radiating surface is too near the sides of the boiler bottom; or otherwise in consequence of the fire-grate itself being placed too high; whereas no situation of the grate whatever can prevent the injurious action of the fire against the inverted arch of a cylindrical boiler.

The manner in which the position of the heating surface affects the generation of the steam, is best understood by considering that it can only be by a certain tendency to overheat, or-to use the more ambitious language of philosophical theorists—to overcharge the substance of the heating sur

-

face with caloric, that any heat can be imparted to the water within the boiler; and the greater the difference of temperature between one side of the boiler plate and the other, the more rapid ought to be the communication in order to preserve the latter uninjured. Now, if instead of the heating surface being in a position approaching to the horizontal over the fire, as described above, let us suppose it to be placed in a perpendicular direction with respect to the fire grate, and let us farther suppose the burning fuel to be in immediate contact with it, as in fact it usually is, in most of those boilers with internal furnaces, such as marine and locomotive boilers, and in all those called fire-box boilers, the effect then must be that all the particles of water that come in contact with the inside of this heating surface, or generating surface, which are converted into steam, will rise to the surface of the water; and by doing so, if in sufficient quantity, they form a continually ascending current or thin stratum of vapour, interposing itself between the water and the plates of the boiler, thus creating two evils at once; first, the prevention of a sufficiently rapid transference of the heat of the plate to the water, excepting only at the lowest point of the effective heating surface, where the fire is in immediate contact with the plate, thereby diminishing the evaporative power of the boiler; and secondly, the effectual prevention of all proper access of the water to those portions of the generating surface immediately above the heated point. The plates in those parts are therefore sure to get overheated and burnt out in a comparatively short time. This is in fact precisely what takes place in many fire-box boilers, and, to a certain extent, in all boilers that have internal furnaces or flues with flat sides, unless those sides incline somewhat to one another, to facilitate the escape of the steam from the heated surface.

It is, of course, impossible that we should attempt any enumeration of the various kinds of steam engine furnaces that at different times have been brought before the public, for their name is legion, and very few of them have approved themselves of any utility: but we shall give a brief sketch of some of the principal examples. The purpose of all these schemes has been to consume the smoke generated in all ordinary furnaces in which bituminous coal is burned; and although the means by which this is proposed to be done are diverse and innumerable, yet the whole of the plans may be roughly classed under two general heads. In the first of these the smoke is proposed to be burned by passing it over or through a fire or other incandescent substance, and in the second the same end is proposed to be attained by admitting a stream of air into the flue or furnace, by which the combustion of the inflammable parts of the smoke may be effected. Each of these methods, however, presumes that there is enough oxygen mingled with the smoke for the purposes of combustion, but in the one case the necessary supply of air has to make its way through the fire-which it can easily do if the fire be thin, and the draught vehement-while in the other case an orifice is specially provided for its admission.

One of the first schemes for the prevention of smoke is that of Papin, who proposed to make the smoke descend through the fire;—the necessary draught for that purpose being maintained by means of his centrifugal blower. This is a very feasible scheme, and one which we are convinced would be found effectual if carried out judiciously in practice, but the impediments presented by the coking of the coal and the absence of an effectual exhauster has hitherto prevented its success. This scheme was revived by Delasme, and was afterwards carried into effect for house stoves by Franklin, but various inconveniences attendant on the lighting of the fire, and other technical points, has prevented Franklin's stove from reaching into any considerable adoption.

In 1785 Mr. Watt took out a patent for obviating the smoke of steam engines by placing the coal in an upright conical tube or hopper fixed in the brickwork of the boiler, immediately behind the furnace door, and causing a stream of air to rush through the furnace door for maintaining the combustion. In this plan, there are no fire bars, but the fire rests upon a brick arch, and the whole of the air that reaches the fire has to pass through the coal that has not yet entered into combustion. By this means the gases, evolved from the coal by the application of heat, mixed with air, are passed through and over the ignited fuel, and are thereby consumed; or they may be passed through hot funnels or pipes which will accomplish nearly the same purpose. This plan, though ingenious and in one sense successful, was relinquished by Watt, on account of the difficulty of dealing with caking coal, which by its concretion prevented the due admission of the air; and he adopted in its stead the plan of a dead plate between the furnace door and the fire bars, upon which the coal is first coked, and is then pushed back upon the bars to undergo combustion. This plan, which is still used by Messrs. Boulton and Watt, is very effectual in preventing smoke with careful firing: we give a drawing of it as at present used by Messrs. Boulton and Watt, in another part of the present chapter.

In 1796, Mr. William Thompson of Bow Lane took out a patent for consuming smoke by admitting a stream of air behind the bridge, which appears to have been the first plan brought into use on that principle. The same species of furnace was afterwards patented by Sheffield, Gregson, and others with trivial variations, and during the last few years a host of patents have been taken out for this kind of furnace, some of them under very imposing names, but none of them can be said to have been successful. One great impediment to their success has been the difficulty of apportioning the quantity of air admitted to the varying wants of the fire; for after the furnace has received a charge of coal a greater quantity of smoke is produced than at other times, and a considerable admission of air becomes necessary; but if the air valve be regulated so as to supply this quantity, it will supply too much when the smoke has passed away, and unless a more sedulous attention be given by the fireman than can be expected in practice the quantity of air admitted to the furnace will generally either be too great or tão little, and a defective performance must therefore be the consequence. Some of these facts we believe we have stated already; but it is necessary

to repeat them here to explain that when the fire is supplied with coal by means of Stanley's fire-feeding machine, or any other self-acting mechanism which makes the production of smoke uniform, it becomes possible to adjust the air valve with great nicety to the requirements of the fire; and even in other cases the due adjustment of the supply may be approximated to by means of a very ingenious mechanism invented by Mr. Murray of Leeds, and described in the London Journal for 1821. In this plan the air supplied to the fire for burning the smoke passes through a tube furnished with a throttle valve, which valve is opened by an attachment to the furnace door, and closed by means of a vane wheel, like a smoke jack, inserted in the mouth of the tube, and which is moved by the current of air passing through the tube into the furnace. When therefore the furnace door is opened and a charge of coal put upon the grate, the throttle valve is opened by its connection with the furnace door, and a sufficiency of air enters: but the entering air, by turning the vane, gradually closes the valve as the generation of smoke diminishes; and by a proper adjustment of the mechanism to the quality of the coal employed and the quantity introduced at a time into the furnace, a very nice regulation of the air may be accomplished. At the same time it must be stated that furnaces which admit air into the flue or furnace, even when provided with a mechanism of this kind, have not been successful in practice. Mr. Pritchard of Leeds, in 1821, proposed to accomplish the same end as was attained by Mr. Murray's mechanism, by means of a piston descending by gravity in a cylinder of air, the piston forcing out the air through a small orifice. The desired effect, it is obvious, can be produced in many ways, which will at once occur to our readers. and which it would, therefore, be superfluous to describe.

An experiment a short time since was tried at Soho, with the view of testing the efficacy of C. W. Williams's Argand furnace, as it is called, which is a furnace for burning smoke by the admission of air into the flue by a number of orifices, or for accomplishing the chemical combination of the gaseous constituents of coal with the oxygen of the atmosphere, as we suppose would be the definition in Mr. Williams's pompous phraseology. The ordinary boiler in use at Soho was fired with one kind of coal for a period of four months, and the effect and consumption were carefully noted. Mr. Williams's improvements were then applied, and with the same kind of coal, the same man firing, and all other circumstances as nearly as possible identical, the consumption, Messrs. Boulton and Watt inform us, was nearly a pound per horse power per hour more than before. The furnace in its ordinary form, which was with a dead plate fitted before the furnace as in Messrs. Boulton and Watt's usual plan of land boiler, produced no smoke; whereas after Mr. Williams's improvement a good deal of smoke was sometimes produced, though usually there was none.

One very ingenious method of consuming coal so as completely to obviate smoke consists in lighting the coal on the top; and a contrivance known as Cutler's grate was some years ago introduced for accomplishing this object in the case of house fires. There is a difficulty, however, in introducing fresh coal beneath the ignited mass, and the plan of a piston for raising it up to the fire is inconvenient. In 1815 Mr. William Moult contrived a furnace on this principle, in which he brought the flame over the coals, the coals being laid upon a dead plate; but it does not appear that this plan has met with much encouragement or success. In 1815 Mr. William Losh of Newcastle took out a patent for a combination of double furnaces; in which, by a suitable arrangement of dampers, the smoke from one of the furnaces was made to pass into the ash-pit of the other; and mixing there with the atmospheric air, and ascending through the fire, it was consumed. The furnaces were of course fired alternately, so that one was bright while the other was smoky. This is one of the most effectual methods of consuming smoke that has yet been contrived, but the shifting of the dampers is troublesome, especially as they require to be strong and heavy, to withstand the heat.

We pass over the plans of Roberton, Johnson, Parkes, Coombs, Stretton, and a host of others, as there is nothing in their schemes of much novelty or utility. In the report of the Smoke Committee of the House of Commons of 1819, a plan is given by Mr. John Walker, junior, of an engine furnace with a coke oven attached; the design being that the coal should first be coked in this oven, and then transferred to the grate to be consumed. K

This scheme is not very practicable, but we believe something of the kind is now being revived by a French engineer at Manchester, the details of which may be such as to give it a better chance of success.

The two most important projects in smoke-burning furnaces appear to be the revolving grate usually known as Brunton's and Stanley's firefeeders, which last are much employed in the manufacturing districts. The revolving furnace was described by Mr. John Steel, of Dartmouth, before the Smoke Committee of the House of Commons in 1819, and in December of the same year the plan was patented by Mr. Brunton; so that it does not appear very clearly to which of these mechanists the invention is due: but, whoever be the inventor, the plan is the best, in our opinion, yet devised for obviating smoke, and, if applied judiciously, can hardly fail to be productive of an important economy. Stanley's plan, which was patented in 1822, consists in the application of a hopper to the furnace into which the coal is thrown, the neck of the hopper being furnished with grooved rollers put into revolution by the engine, which seize the pieces of coal that exceed the size regulated by the distance asunder of the rollers, and which may be adjusted by a screw,-and drop them on an iron plate underneath, from whence they are projected by an arrangement resembling a revolving fan, which scatters the pieces equally over the fire. The gearing, which gives motion to this apparatus, is so contrived, that as the engine becomes quicker its effective speed becomes less, and the quantity of coal supplied to the furnace, therefore, just comes up to the demands of the engine for steam. A drawing of this machine, as applied at the present time to boilers in Lancashire, will be given in one of the plates of the present work, which will make the nature of the apparatus more intelligible than the most elaborate description could hope to effect. A sketch of the revolving grate in its present improved form, as applied by Messrs. Boulton and Watt a year ago to the steam furnaces of the Bank of England, is given at page 52. of the present work; and we must reserve what we have to say farther of the contrivance in question until we come to the description which that representation demands.

In 1824 a patent was taken out by Mr. Humphrey Jeffrey, of Bristol, for a plan of condensing smoke and metallic vapours by means of a shower of water, which has been introduced with success in many cases in practice. In order to carry into effect this plan, it is necessary to have two or more chimneys, each closed at the top, and connected together by a cross flue at the top, so as to form a combination resembling the Greek letter П (pi). The smoke or vapour ascends one of these stalks, and passes into the upper part of the other by the horizontal flue, where it meets a shower of water descending from a tank on the top of the second stalk, by which it is carried downwards, and it passes off with the water into a drain. We do not think it necessary to resort to this plan for the destruction of smoke, as we think that end attainable by more economical means; but we think the plan is one of much value for condensing the insalubrious vapours of vitriol works, copper smelting furnaces, and other such manufactories of asthma, and that its use should be made compulsory in such cases. Several newlyfledged projectors have been endeavouring, we find, to attract the public attention to this plan under some unimportant modifications, by pretending that it is a novelty of their own device: but the plan is an old one, of which the patent has for some time expired; and any one is free to use it who feels so disposed. It will be expedient, in some cases, to make the smoke or vapour ascend a stalk several times, or rather to make it ascend several stalks in succession, in order to receive in each returning stalk a fresh shower of water; and it is important that the final exit of the gaseous matter should be at the top of a stalk, which may be made higher than any of the others. A good method of forming a combination of this kind would be to cause the smoke and vapours to traverse a series of upright iron cylinders connected alternately at the top and bottom, with a shower of water in each descending cylinder, and the last one might be carried to a sufficient elevation to constitute a chimney.

In 1824 Mr. Evans, of Queen Street, Cheapside, took out a patent for obviating smoke by admitting steam into the furnace. This plan has been often tried: it was introduced into the Edinburgh gas-works many years ago by Mr. Nasmyth, of Patricroft, and has since been revived by Mr. Ivison, of Edinburgh: but we do not anticipate that the plan of admitting steam into the furnace or ash-pit will reach any high measure of success so far as relates to the prevention of smoke, though for other purposes its admission may in some cases be advantageous.

Of all the projectors in the field of smoke-burning there is no more assiduous veteran than Mr. John Chanter. He has a great number of patents for different kinds of furnaces; and, indeed, his plans vary so frequently, that it is difficult to know what they are; and we question whether he is perfectly confident on this head himself. Mr. Chanter, we should say, is a person of a wavering disposition, the effect in many cases of exuberant ingenuity; but, however this may be, we do not think any of his plans are likely to approve themselves of much utility in practice. Mr. Samuel Hall and Mr. Joseph Williams have plans for burning smoke by the admission of hot air into the flue or furnace, and Messrs. Drew, Rodda, and others, have plans which nearly resemble some of those that we have already described. Mr. Cheetham, of Staley Bridge, accomplished the combustion of the inflammable parts of the smoke by drawing it by means of a fan out of the upper part of the flue leading to the chimney, and sending it. mixed with atmospheric air, into the ash-pit, to ascend through the fire. The carbonic acid, which, by virtue of its specific gravity, occupies the

lower part of the flue, is not returned to the fire, but passes at once to the chimney, and thus maintains the draught.

About four years ago certain contrivances for obviating smoke were applied in some of the vessels of the Peninsular Steam Company, which we may here explain. Fig. 30. represents the method adopted in the steamer Fig. 30.

William Fawcett.

The whole of the furnaces were arranged in pairs, and an opening was cut through the waterspace intervening between each pair of furnaces, so as to permit the smoke from the one to pass through a pipe fitted at the furnace mouth into the ash-pit of the other, and mixing there with atmospheric air, and ascending through the incandescent fuel on the bars of the second furnace, it was consumed. One furnace thus operated as a retort while the other acted as a furnace, a damper extending between the double bridges shown in the horizontal section being closed when the furnace received a charge of coal so as to convert it into a retort for the time being. As soon as all the gas was expelled from the coal thus introduced the damper of this furnace was opened so as to convert it into an active furnace, and the damper of the other furnace was closed to be in readiness to receive a charge of coal, and no smoke was visible at the chimney while this action proceeded. It is troublesome, however, in practice to shift these dampers; and altogether the plan is not so good as many others that have been brought under the notice of the public. Fig. 31. represents Fig. 31.

the method of diminishing smoke introduced into the steamer Tagus. Two Venetian bridges were formed of tiles laid at an angle about three quarters of an inch apart, the bridges themselves being about a foot apart, with a vacant space between them. These tiles of course became very hot by the flame passing between them; and by keeping a thin fire on the bars, so as to enable a sufficiency of oxygen to find its way through the burning fuel to accomplish the combustion of the inflammable gases during their passage between these hot surfaces, the smoke was very completely extinguished. The bridges were built upon an arch closed by a door, so as to enable a person to get into the flues without taking the bridges down. We think that a succession of bridges of this description, with a furnace lined with fire-brick, a good draught, thin fire, and very narrow furnace-bars widely set, would accomplish the combustion of smoke very completely. It would be better if the tiles, instead of being flat, were semicircular, for the hot current would be thus made to reverberate; and conflicting draughts and eddies have a powerful effect in aiding the combustion of smoke. Fig. 32. represents the furnace of Mr. John Juckes, the peculiarity of Fig. 32.