319. It does not follow that because coke in England, anthracite in Pennsylvania, or wood in New England, is the most economical fuel, that either of the above will be so in Ohio, Indiana, or Illinois, or because wood is the cheapest in some parts of a State, that it is so throughout, or even that one fuel should be applied to the whole length of a single road.

The heat used to evaporate water in the locomotive boiler is developed by combustion; combustion is produced by chemically combining the oxygen of the air with the carbon of the fuel; whence, that material containing in a given cost the largest amount of carbon will produce heat the most economically.

From the table on page 320, we see that, by bulk, thirteen of coke are equal to sixty of wood; that one pound of coke evaporates eight and one half pounds of water; that one pound of wood will evaporate two and one half pounds of water. Tables of specific gravity give as an average weight per cubic foot of hard wood, thirty pounds. A cord of wood, by very careful measurement, contains one hundred cubic feet solid, or one hundred twenty-eight feet as piled, taking the average size of wood; whence a cord will weigh three thousand pounds. And we have as the relative evaporative efficiency of a cord of wood and a ton of coke,

Now if the cost of a cord of wood is to the price of a ton of coke as 7,500 to 19,040, it is immaterial which we use.

As an example of the use of the above proportion, when the absolute cost of wood, coal, coke, and labor are known, take the following.

If wood, cut and ready for burning, costs $3.00 per cord, how much may be given for a ton of coke?

From the same proportion we form the following table.

In the comparison above, the maximum evaporative power of wood has been used, 2½ lbs., and the ordinary power of coke, 8½ lbs. of water per pound of fuel.

320. In making coke in large quantities, the ovens should be at the mines, as we thus save transporting the extra weight of coal over coke.

The cost of making coke, exclusive of the cost of the coal, is approximately as follows:—

or in round numbers, thirty cents per ton; and if coal is $1.50 per ton, adding twenty-five per cent. we have $1.87 as the cost of coal that will make one ton of coke, to which add the cost of making per ton, thirty cents, and we have as the whole cost of one ton of coke $2.17; and from the rule on page 327 we see that wood must not cost over $0.85 per cord to be as economical as coke at $2.17; of course inferior qualities of coal will give less good coke and change the comparison.

COMBUSTION.

321. The combustible element in all fuels is carbon; the heat necessary for steam producing, is obtained by combining the carbon of the fuel with the oxygen of the air, forming carbonic acid gas.

Carbonic acid gas consists of

Atmospheric air consists of

Whence, for the combustion of one pound of carbon, we require

But to obtain 2.66 of oxygen from the atmospheric air, we also use nitrogen in the proportion of 28 nitrogen to 8 oxygen; whence, for converting one pound of carbon to carbonic acid, we require

From careful observations on the gases passing through the chimneys of well-constructed boilers, oxygen is found free, varying in amount from one quarter to one half of the quantity necessary for combustion; this is owing to the mechanical obstructions to the perfect conversion of the air arising from leakage through the fuel.

More than the above 11.97 lbs. of air should, therefore, be applied to the fire for each pound of carbon consumed. Twenty-five per cent. is found by experience to be a sufficient surplus allowance to convert the carbon.

Air weighs .075 lbs. per cubic foot, whence 15
.075 or 200 cubic feet of air are necessary for the proper combustion of one lb. of pure carbon.

Knowing the necessary amount of air for one lb. of carbon, and also the percentage of carbon in the different kinds of fuel, it becomes a simple arithmetical operation to fix the bulk of air required for any species of coal, coke, or wood. The result of such a calculation is shown in the seventh column of the table on page 320.

“There are two causes why all the heat which fuel may furnish is not obtained. First, that the inflammable gases evolved by the heat are not all consumed from want of a sufficient supply of oxygen, the air drawn through the fire being only sufficient to decompose more fuel than when decomposed it could burn, or supply with oxygen. The thick smoke that escapes from a chimney when fresh fuel is thrown on a hot fire, is unconsumed gas; decomposed from the fuel, but without oxygen enough to burn—although there may have been a sufficient supply of heat. From this cause it is, perhaps, that flame is seen coming from the top of a steamboat chimney which appears to be continuous from the furnace; but which, in fact, is ignited by contact with the air, having retained sufficient heat for that purpose.

“All smoke-consuming furnaces are simply means of admitting fresh air to the unconsumed gases above the fire, which, in a common chimney, will effect the object, as so large a mass of smoke retains the necessary amount of heat. This only prevents the nuisance of smoke. To render the gases thus reheated useful in evaporating water, this supply of oxygen must be added while the gases are yet in the flues.” This might seem difficult. Mr. McConnell (England) divides the flues of his locomotives into two parts, connecting the front ends of the first part and the back ends of the second part by a space of twelve or fifteen inches, (called by him a ‘combustion chamber,’) into which he admits any required amount of fresh air. (See appendix E.)

“A second cause why the full value of the fuel produced is not obtained is, that so much is abstracted from the gases in passing through long tubes, that there is not enough left to continue combustion, although the inflammable gas is still there. That a tube or any substance in the way of the hot gases does absorb the heat enough to prevent the burning of the gas, is proved by the action of Davy’s Safety Lamp; this is a common light surrounded by a wire gauze, which so absorbs the heat from the flame as to extinguish the latter at the wire; by applying fire above the gauze, the gas is again kindled, showing plainly that want of heat above had quenched the flame.” See Stöckhardt’s Chemistry; translation by C. H. Peirce, M. D., Cambridge, Mass., 1852, page 105.

We require, then, in every boiler, first, to have a sufficient supply of oxygen to decompose the fuel; next, a quantity above the fire to consume the produced gases; third, such an arrangement of communicating surface that so much heat shall not be abstracted from the gases as to deaden their combustion, until just as they are discharged, at which period they ought to be consumed. (See appendix E.)

GENERATION OF STEAM.

322. The means of producing the power is of course of the first importance.

The heat generated in the fire-box is conducted through the tubes to the exhaust chamber; during which passage it is imparted to the metal, and from thence absorbed by the adjacent water, which being thereby made lighter, rises to the surface and gives place to a new supply. The duty of the furnace is to generate, and of the tubes to communicate, heat.

The power of a plain surface to generate steam, depends upon its position and on the ability of the material to transmit heat An experiment recorded in Clark’s Railway Machinery, gave the following results: A cubic metallic box submerged in water and heated from within, generated steam from its upper surface more than twice as fast as from the sides when vertical, while the bottom yielded none at all. By slightly inclining the box the elevated side produced steam much faster, while the depressed one parted so badly with it as to cause overheating of the metal.

Acting upon this result, most builders of engines of the present day give an inclination of from one inch to one quarter of an inch per foot to the sides of the inner fire-box. That the heat should be applied in the most effectual manner to the water, the latter should circulate freely around the hot metal, carrying off the heat as soon as it reaches the surface. As the heat is applied to the inside of the furnace and tubes, it must, therefore, be the inside dimensions which determine the amount of heating surface.

Note.—If we multiply the interior surface of a tube by the intensity of heat applied, and divide the product by the exterior surface, we shall have the intensity at the outside. We also apply more heat to the outside of a tube, which, passing to the inner surface, augments in intensity per unit of area.

The area of the inner fire-box is not all available for heating, but requires to be reduced as follows:—

The area is, therefore,

TUBES.

323. The tubes or flues, varying in number from one hundred to three hundred, in diameter from one and a half to three inches, and in length from eight to sixteen feet, furnish the real communicating surface. The amount of heating surface thus obtained for any length, number, and diameter, is given in table 10, Chapter XIV., Part I. The surface of a single tube is found by the formula

The efficiency of circular tubes is a matter not yet fully understood. They certainly give a large amount of surface in a small boiler. Pambour considered the value of tube area per unit of surface, in terms of the furnace area, as one third only; that is, three square feet of tube surface as equal to one foot of furnace area, in power of generating steam. D. K. Clark makes no distinction between the two surfaces, but observes “there is reason to believe that in the upper semicircular part of each tube the efficiency principally resides. The winding progressive motion, observable in tubes of considerable diameter, confirms this conclusion, as it is with much probability due to the cooling of the upper portion of the gases of combustion, which, as they cool also, become heavier and descend laterally, to make room for the hotter smoke next the bottom of the flue; the general result of which is the spiral motion of the current in its progress onwards.” Certainly the upper half of the tube would part much easier with the steam than the under one, even supposing the applied heat to be the same.

At page 340 of “Overman’s Mechanics,” is the following: “The application of heat to a concave surface is wrong in principle. The heat in gases is conducted to other bodies, and among themselves by convection only. This quality of gases causes the convex form of a vessel to be the most profitable in absorbing the heat of ascending gases, because the motion of the gas causes a constant change of particles on the convex body. On a concave surface exposed to the influence of moving gases, but little effect is produced; because the particles of gas in the concavity are at rest. A plane surface is for the same reason an imperfect form for absorbing heat; it must be exposed at an angle of 45° to obtain the best effect of the heating gases. In all cases if we wish to obtain the best effect from the fuel, we should expose a convex surface to the current of hot air. The direction of the motion of the hot gases decides the position of the metal which is to absorb the heat; if the current is horizontal the pipes must be vertical. Gases do not convey heat by radiation. Tubes and other vessels containing water must be so placed that the hot gases play around the outside.

“If we lead a current of hot air around a cylinder we shall observe that a particle of air plays but a short time upon its surface, when it gives way to another; the particles play almost around the cylinder, and a concentration or increase of density behind the pipe is the result. The relative position of pipes in the range is not indifferent, and the distance of one from the other must be related to their diameter.”

The conducting power of the metal composing the fire-box and tubes, is one condition which limits the rate of evaporation, when the heat is abundant on the one side and circulation free on the other, as the water certainly carries off the heat as fast as it arrives at the outer surface.

All the heat should be extracted if possible from the gases before they enter the smoke box. We should so arrange the flues, that without so much contracting the passage for the exit of the gases as to need too strong a blast, yet to confine the gases until their full value is extracted.

Several attempts have been made to apply the ideas of Clark and Overman, but as yet they have been very indirect and have met with only moderate success. (See Appendix, E.)

EVAPORATION, PRESSURE, TEMPERATURE, AND DENSITY.

324. The character of work to be done determines the nature of the steam to be used.

The quantity of work to be done shows the amount of steam to be produced.

The amount and character of the steam required, fixes the dimensions and proportions of the boiler.

A cubic foot of water, at a temperature of 62°, weighs 62.321 lbs.

A cubic foot of steam, generated at 212° Fahrenheit, under the atmospheric pressure (14.7 lbs. per square inch) weighs .03666 lbs.

Whence one cubic foot of water boiled at 212°, makes 1,700 cubic feet of steam.

The total heat of saturated steam (steam produced in contact with the water), consists of two parts at all temperatures; the latent and the sensible. The sensible heat is that shown by the thermometer, and varies with the pressure. The latent heat absorbed during the generation of steam, amounts to three fourths of the whole. As the temperature at which the steam is produced increases, the bulk produced from a given unit of water decreases, but the pressure and the total heat increase. (See C. R. M. p. 59, 61, Regnault’s experiments.)

Table 8, Chapter XIV., Part I., gives the properties of saturated steam, produced under pressures varying from fifty to one hundred and fifty pounds per square inch.

The steam produced over water is called saturated, and an application of heat to an isolated volume of this steam, raises both the temperature and pressure, the volume and density remaining the same. The saturation is then no more, and the steam is surcharged. If the heat be withdrawn, pressure and density fall, and a precipitation of water takes place. The priming of steam in a cylinder is an illustration of this. D. K. Clark, in Railway Machinery, urges the necessity of thoroughly drying the steam before applying it to the pistons in this manner, he says, ten per cent. may be gained at low velocities, and in some cases forty per cent. at high speeds.

MOTION OF STEAM IN PIPES.

325. Steam may flow from any vessel into a vacuum, into the open air, or into steam of a less density. The velocity of efflux of steam is the same as that of a stream of water flowing under a pressure equal to that of the steam. Steam flowing into the atmosphere of course has 14.7 lbs. per inch resistance to meet, which is equivalent to a reduction of 14.7 lbs. of its pressure. The following numbers show the velocity of efflux of steam into the open air under different pressures.

326. The loss of power suffered by the steam during its motion from the boiler to the cylinder is caused by condensation in passing through cold pipes, and by friction and sharp bends. The total fall that may be caused by a combination of circumstances is from ten to fifteen per cent. at low velocities, and from fifty to sixty per cent. at high speeds. The fall of pressure decreases as the square of the velocity of motion, that is, the fall at a velocity of 1,600 feet per second is four times as great as the fall at a velocity of eight hundred feet. By well protecting the steam pipes and cylinders, and by drying, it may be worked at nearly its initial pressure.

APPLICATION OF STEAM.

327. The steam being generated in the boiler, and conveyed to the cylinders, is admitted alternately to the opposite sides of the piston, by which its reciprocations are produced. The first valve applied to regulating the admission of steam to the cylinder was so arranged that the steam was admitted during the whole stroke; at the end of which, ingress stopped and egress commenced at the first end, and ingress commenced at the second end simultaneously; this caused an unnecessary resistance to the return movement, by preventing the quick escape of the first cylinder-full, which had to be pushed out, instead of flowing out. The continuance of the full pressure upon the piston also, until the end of the stroke, caused a dangerous momentum to be given to the reciprocating machinery.

These evils are obviated by causing the exhaust passage to open, and the entering port to close a little before the end of the stroke. This is effected by moving the valve bodily forward.

Now it is well ascertained, that with very free steam entrances, if we allow the cylinder to be only partially filled, and then cause the steam to expand itself, more work is accomplished with a given bulk than when the cylinder is completely filled. That the steam may have time thus to expand itself, the return of the piston must not take place until after the suppression (the stopping of admission).

328. There are four positions of the valve during each half stroke, and three distinct actions of steam in the same period, which are as follows:—

The longer the time between suppression and release, of course the more complete will be the expansion. The entire force of the steam should not (even if possible) be extracted, as a certain force is necessary to produce a blast.

The time of expansion is regulated by the proportions of the valve cover; which may be so adjusted as to fix suppression or release at any desired part of the stroke.

By the above means any rate of expansion may be established, but when once fixed will remain the same, the valve being invariably connected with the eccentric, and thus partaking of its motion.

329. The great step which has been taken in locomotive construction since 1840 is the invention of the “link motion,” by Williams, which, perfected by Howe, admits of varying the travel of the valve, and thus using the steam under any desired rate of expansion. By this arrangement, the power of regulating the force applied to the piston, according to the work to be done, is placed in the engineer’s hands, to be used at any time under whatever conditions the engine may be working.

By this arrangement, two eccentrics to each cylinder are required, (and in some dispositions of the link, only one). Fig. 150 shows the general plan of varying the expansion. A fixed relation evidently exists between the points A and B, two distinct motions are communicated by the eccentrics C and D through the rods E and F, to the two ends G H, of the curved link L; the eccentrics are so adjusted upon the driving axle as to cause the two ends of the link to move in opposite directions, whence at some point midway there is no motion; the link is movable (vertically) upon the suspended point L, so that by bringing L to one end or the other, the motion given to the rod m partakes of the motion of that eccentric which is nearest to it. Thus the movement of the valve may be checked, or even reversed in a second, while the engine is in motion, and that without sudden shocks.

The link is moved by the levers n n′ n″ terminating in the bar O, placed at the foot board of the engine in reach of the engineer. Applied to this is an iron sector h h′ h″ made fast to the frame of the engine. Now when the point L is in such a part of the link as to place the valve in a position admitting steam for any fraction of the stroke, let the point at which the bar O stands upon the sector be marked for that admission; and so also for any number of different degrees of expansion. It is plain that the engineer may thus, by fixing the lever O, use any percentage of admission that is required; and may always know just what duty the engine is doing. Five minutes’ examination of the reversing gear upon an engine will render the operation plain.

330. If we cut the steam off at half stroke and then allow it to expand, of course the mean pressure during the whole stroke is less than that at entering. The effective mean pressure obtained by any degree of expansion is shown by the following formula, deduced from a mean of forty-nine experiments with the Great Britain locomotive, (Great Western Railroad, England,) having cylinders 18 × 24.

where a is the percentage of admission.

From this formula, table 11 is made.

331. Mr. Clark deduces as general results, from a very extensive and carefully conducted system of experiments, the following.

That the maximum useful admission is seventy-five per cent.

The minimum useful admission is ten per cent.

The greatest possible gain by working expansively is one hundred per cent., which is effected by an admission of ten per cent.

The best admission for engines having ports 1
14 of the area of the piston, and blast area from 1
13 to 1
16 of piston, at high speeds (from thirty to sixty miles per hour) and with considerable loads, is from sixty to sixty-six per cent. With a wider port and blast area, the best admission is seventy-five per cent.

The resistance due to the back pressure of the blast, varies as the speed squared, and inversely as the square of the area of blast orifice.

332. From the experiments made by Daniel Gooch, with the engine “Great Britain,” the following results appear.

The loss of fuel at seventy-five per cent. admission, the blast orifice being from ⅒ to 1
11 of piston at sixty miles per hour, is from ⅓ to ⅒; at thirty or forty per cent. admission, the loss is from ⅛ to 1
50; and at thirty miles per hour, (seventy-five per cent. admission,) from 1
11 to 1
40.

The resistance from steam compressed in the cylinder, increases with the speed, and also with the degree of expansion; it varies from eight per cent. in full gear, (seventy-five per cent.,) to twenty-eight per cent. at an admission of forty per cent.

At the highest velocities, the whole resistance from back pressure is nearly the same for all expansions; for compression increases as blast pressure decreases.

The above deductions hold good for speeds under forty miles per hour, with steam ports at least 1
14, and blast orifice from 1
12 to 1
15 of the piston area.

OF BOILER PROPORTIONS.

333. The dimensions of American locomotives seem to depend more upon the shop whence they come, than upon any special duty required of them. It is not surprising that the utmost economy is seldom attained when a railroad president orders a lot of locomotives, from the cheapest builder, to suit his own ideas of an engine; or when engines are ordered by a superintendent of machinery who does not know the difference between a sixty foot grade and a level. It is the affair of the company’s agent and not of the machinist to know just what a railroad needs. It is a common, and most absurd practice, for a man who is completely ignorant of machinery to order five or ten engines, without the least regard to the character of the road or of the traffic.

334. The particular characteristics of each class of engines is entirely a matter of figures. There is no reason why a general table should not be formed embracing all divisions, orders, and classes of locomotives, in which the requirements and general dimensions corresponding thereto should be laid down for machine shop reference. Such a table would at once establish a mutual understanding between railroad companies and builders. Such a general classification is shown hereafter. The dimensions of engines are not given, as it was thought best to let each person fill it up according to his own ideas. By so doing some valuable general proportions may be arrived at.

335. Thus far experience has been the only guide to proportion (in America at least). Practice, in many things, is the only correct path to the right results, but locomotives are too expensive for philosophical apparatus; correct experiments upon imperfect machines will lead to the means of avoiding errors. The following is the modus operandi of D. K. Clark in his “Railway Machinery.”

A number of engines of different proportions are chosen, and observations made upon the amounts of fuel and water consumed upon the work done, and under what conditions. These results are so tabulated as to show the effect in difference of construction upon the performance of the engine, whence the proportioning of parts becomes a simple arithmetical operation. The reduction of experiments to tables, and the deduction from tables of formulæ, is a simple operation compared with the skill and care required in observing the operation of a machine, subject to so many disturbances as a locomotive engine in rapid motion. None have had a better opportunity of observing, have conducted experiments with more care, or have obtained results which show fewer discrepancies than the English engineers Clark and Gooch, and the French and German observers Le Chatlier and Nollau.

336. Three essential parts of the locomotive are the grate area, heating surface, and cylinders. No two writers upon this subject arrive at the same dimensions to perform the same work. They not only differ, but differ widely. They cannot all be right; all but one, or all must be wrong. American builders have fixed the dimensions of their engines by observing the performance of constructed machines, not by rules deduced from any systematic experiments, but upon a system of remedying visible errors. If a chimney diameter of ten inches is found too small and twenty too large, fifteen has been assumed as about right.

337. As an example of the difference in the results obtained by different authors, take the following:—

An engine to do the same work must have, according to

From these figures, the work done being the same, Mr. Clark gives forty per cent, more grate area than either Colburn or Norris, an easier blast, and greater heating surface. Norris makes the steam and water room equal, while Colburn makes the latter almost double the former. It is to be observed that Colburn gives only rules adopted by different builders, not vouching for their correctness, while Norris lays down his rules as fixed and right. The engines used by the English experimenters in their observations, vary in dimension between the following wide limits, whence the universal application of their results.

338. The result of some sixty experiments upon forty-five different engines (detailed in Clark’s Railway Machinery, page 156), gives the following formula, expressing the relations which ought to exist between grate area, heating surface, and consumption of water; that evaporation may be carried on in the most economical manner.

From which we deduce the value of a or c thus,

The maximum evaporation which should be carried on per square foot of grate is found, by Mr. Clark, to be sixteen cubic feet per hour. Thus, if we wish to evaporate 160 cubic feet of water per hour, we must have a grate area of at least 160
16 or ten square feet.

339. The above formula for the grate area gives the dimension for a coke-burning furnace. Locomotives burning wood or coal require a modification of the above, as follows:—

To produce a given amount of heat, a certain amount of carbon must be burnt. As wood contains much less carbon than coke, a correspondingly larger bulk must be burnt, and a larger grate is necessary; not, however, larger in proportion to the larger bulk of fuel, as we may have a deeper wood than coke fire. The relative depth of fire being as the stowage bulk, and the actual depth of a coke fire being 1.9 feet, that of a wood fire will be 2.5 feet.

Now let A be the number of lbs. of coke per foot of water evaporated.

B the number of lbs. of coal per foot of water evaporated.

C the number of lbs. of wood per foot of water evaporated.

Call d the depth at which if is the most economical to burn coke; d′ the same depth for coal, and the depth for wood d″. Then will the area of a coke grate be

Of a coal grate

And of a wood grate

To be able to fix the proper grate area for any fuel, we must know its evaporative power, and a depth of a layer in the furnace. Knowing the absolute value for coke, it remains only to obtain the relative value for any other. Thus far we have disregarded the difference in time of burning wood and coke. To produce a given amount of heat, we burn a certain chemical value of fuel; a much larger bulk of wood than of coke is needed. If we burn wood and coke at the same depth and in the same time, the grate areas would be proportional to the bulks of fuel to produce the same heat; but, first, we burn fuel in a depth proportioned to the economic stowage bulk, or as 2.5 to 1.9, which decreases the wood area; and, second, a layer of coke 1.9 feet deep burns in one hour, while a layer of wood 24 feet deep burns in fifteen minutes; whence 60 m. divided by 15 m. = 4 layers of 2½ feet deep each, or in all ten feet, which into the bulk (equal to a mass of coke 1 foot square × 1.9 high) or 1 foot square by 14 high, gives 14 ÷ 10 = 1.4; or, finally, the area of the wood grate should be 1.4 times that of a grate to burn coke.

OF THE SIZE AND USE OF THE SMOKE BOX.

340. The smoke box is the general termination of the flues, and the place where the vacuum is produced, which causes the draft. The size of the boiler being the same, the vacuum varies directly as the blast pressure. The power of the blast is of course affected by the capacity of the smoke box. Mr. Clark fixes the capacity of the exhaust chamber at three cubic feet per square foot of grate. The vacuum in the furnace varies from one to two thirds of that in the smoke box. The less the resistance to the hot gases experienced in the flues, the less may be the vacuum. Upon the vacuum depends the amount of air drawn through the grate; upon the bulk of air drawn through the grate depends the combustion; upon the combustion the evaporation. Whence the evaporation cet. par. depends the vacuum in the smoke box.

The velocity of any fluid depends upon the power applied to it, (being as the square root,) the pressure applied to the gases in the furnace of a locomotive is the vacuum in the smoke box; thus the combustion or rate of evaporation is as the square root of this vacuum. To double the evaporation it is necessary to quadruple the vacuum.

BLAST PIPE.

341. The blast pipe conducts the waste steam from the cylinder, which drives the air from the chimney and produces the vacuum in the smoke box; its form should permit the freest escape of the steam from the cylinder. The blast pipe area should nowhere be smaller than the exit port, except at the contraction at the top. “Too much care,” says Mr. Clark, “cannot be taken to adjust the blast pipe concentrically with the chimney; one half inch has been known to spoil the draft of a locomotive.” “The area of orifice is the most critical and most important item in the composition of the locomotive.”

For the form, dimensions, and influence of this important member, the reader is referred to Clark’s Railway Machinery.

As the grate area increases, the blast may decrease. The greater the flue area the easier may be the blast; decrease of smoke box capacity and of chimney diameter, both allow a milder blast.

342. The following proportions are collected from the work of Mr. Clark. The order in which the different parts of the engine stand in importance with relation to the blast, is shown in column 1. The figures show the ratios (the best) which may be had under the most favorable circumstances.

The vacuum in the smoke box is somewhat regulated by a damper placed in front of the ash pan, by a valve in the chimney, or by a Venetian blind covering the front ends of the tubes.

TUBE SECTION AND LENGTH.

343. The section of the tubes (crosswise) is the space through which the hot gases pass off. By increasing the length or decreasing the diameter, we of course require a stronger blast.

That the steam may escape as soon as generated, there must be a certain clearance between the tubes, which Mr. Clark fixes as follows:—

Divide the number of tubes by thirty and the result is the clearance in eighths of an inch; or algebraically