PART I.—HISTORY AND DESCRIPTIVE ACCOUNT OF THE STEAM-ENGINE.

SECTION I.

It appears highly probable that the ancients knew more of the phenomena of steam than has been generally admitted; and one evident cause of this mistake is, that no specific term equivalent to the word steam was generally used by them; and water, when heated, was said to be converted into air. Hero of Alexandria, in his *Pneumatics*, written more than 120 years before the Christian era, collected the science and inventions of the ancients, along with some of his own, into a systematic treatise. His book commences with a lucid dissertation on the properties of air as a medium for the communication of pressure and motion, and especially upon the nature and effects of a vacuum, subjects to be thoroughly understood by all who would master the theory of the steam-engine.

The following description of the manner in which the force of steam issuing from a boiler may be applied to support a weight is given in the *Pneumatics*:

"A boiler (fig. 1) perforated on the top is placed on the fire. From the perforation there proceeds a tube on whose extremity is fixed a hollow hemisphere perforated in like manner. If then we place a light ball in the hemispherical cup, it will follow that the vapour rising up from the boiler through the tube will support the sphere, and it will appear to dance."

There is another apparatus in the *Pneumatics* for producing a revolving motion by the action of steam from a boiler, issuing through a couple of bent arms, from orifices at their extremities (fig. 2); the arms being free to revolve horizontally, do so in obedience to the action of the escaping jets of steam.

The same apparatus, on similar principles, is next applied by Hero to the construction of a machine still more curious. The agent mentioned in this case is rarified air, although the action is of precisely the same nature. Here the science of the philosopher appears to have been degraded to assist the priesthood in deceiving the populace by the resemblance of miraculous interference. "A fire having been kindled upon an altar, living figures will appear to lead a choral dance, even although the altar itself be of transparent glass or horn. Through the epipyrrus a pipe is to be let down to the base of the altar, where it is to revolve on an iron pin, the other end being passed through a tubular fitting attached to the epipyrrus. And this pipe is to have other little cross-pipes attached to it, and perforated (fig. 3), so as to communicate with it (which are to radiate opposite to one another around it), and turned alternately in opposite directions in the ends. There shall likewise be a drum attached to it, upon which the figures of the dance are to be set. Then by the action of the kindled fire, the air, being warmed, will proceed into the pipe, and from its being driven out through the bent tubes in the base of the altar, will turn round the pipe and its drum."

There is considerable reason to suppose that, to their knowledge of the elements of machinery, the Egyptians added some acquaintance with the power of steam, applied however, only to the degraded service of superstition. The statue of Memnon is said to have emitted sounds which Pausanias compares to those produced by the snapping of the strings of a harp. Strabo expressly states that he heard them; and Philostratus states, that when the sun shone strongly on the statue, sounds issued from its mouth similar to those of a stringed instrument.

The Romans appear to have done little for the mechanical arts, and nothing for the improvement of steam apparatus. It was not until the dawn of knowledge succeeded the darkness of the middle ages, that the light reflected from the works of Hero and the older mechanicians, rekindled the flame of mechanical invention. The works of Archimedes and of Hero were read with great avidity, and formed some of the most popular productions of the young art of printing.

Giambattista Della Porta, an Italian, was one of the ablest expositors of the principles of pneumatics. In his work published at Naples in 1601, he describes a beautiful and simple expedient designed for determining the philosophical question, How much aqueous vapour is formed by a given quantity of water? And it shows an acquaintance with the fact, that water heated by fire is converted into aqueous air, with sufficient force to raise water above its level and form a running stream.

Salomon de Caus, a French engineer, published at Frankfort, in 1615, some account of his works, constructed for the gratification of "gentle curiosity," in which he states that the violence with which water is dissolved into air by means of fire is very great, and that a ball of copper containing water, if placed in a fire, would certainly burst. He afterwards shows how a jet of water may be made to rise above its level, and play in the air by means of fire; a copper ball \(a\) (fig. 4) has an orifice \(b\), at which water is poured in, and it is then closed with a stop-cock. Another tube \(c\) is closely fitted to the same ball, but passes down near to the bottom, where it is open in the water; the pipe terminates above in an orifice for a small jet of water, which is regulated by a stop-cock. In fact, the apparatus is precisely that which Hero uses for raising a jet of water, but without the aid of heat, when he forces air above the water, till it raises a jet of water by its pressure. Of course the heat produces the same elastic force in De Caus's machine as the compression of air does in Hero.

The direct emission of steam from an orifice of the boiler, which had been used by Hero to sustain the ball in the air, was applied by Branca, an Italian architect and engineer, to impress a revolving motion on the vanes of a wheel like a common mill-wheel, and this communicating with a series of toothed wheels gave motion to a series of pestles in mortars. This and many other machines were published by him at Rome in 1629. This period appears to have teemed with curiosities of mechanical invention—Perpetual motions were very common; wings for enabling men to fly in the air, mechanical chariots for a similar purpose, conveyances to the moon, and engines for making continual and cheap music by mills or by fire, for rocking of cradles and turning of spits, were favourite subjects of design; and many of these curious contrivances, without serving any definite purpose, form elegant and curious pieces of apparatus.

The Marquis of Worcester invented and constructed the first actual steam-engine, announced in the Century of Inventions in 1663; none of the contrivances before his time were provided with means to make their action continuous. The instrument of De Caus had merely the power of emptying itself of the boiling-water with which it was filled; the engine of Worcester was provided with two vessels to act successively and continuously, in raising water. "Having found a way to make my vessels so that they are straightened by the force within them, and the one to fill after the other, I have seen the water run like a constant fountain, 40 feet high. One vessel of water rarified by fire driveth up 40 of cold water; and a man that attends the work is but to turn two cocks, that one vessel of water being consumed, another begins to force and refill with cold water, and so successively." The Marquis enumerates many important practical uses to which his engine might be applied, and many of them it certainly would have been competent to compass. Lord Worcester's "Water-commanding Engine" was the first machine moved by fire of efficacy and permanence: he actually erected one of his engines of about 2-horse power on the banks of the Thames, at Vauxhall, and it was employed in supplying the town with water. Not that any great merit is to be attached even to that contrivance, for it is deducible without any violent stretch of the imagination from the schemes of Hero; and there is no doubt, as Mr Bourne suggests, that the ancients would have realized an effectual steam-engine, if they only had possessed mines that required to be drained, and coal to bestow on such a purpose.

The next forward step of invention was the accomplishment of a vacuum by the agency of steam, by Captain Thomas Savery, who exhibited a model of his engine to the Royal Society in 1669. The Marquis's engine appears to have been placed on or below the level of the water to be raised, and, his vessels being filled, their contents were raised by the elastic force only of the steam; Savery, on the other hand, erected his engine at a height of nearly 30 feet above the level of the water. A large close vessel was filled with steam; this steam was reconverted, by cooling the outside of the vessel, into water, leaving the large space it had formerly occupied vacuous; into this vacuum water was raised, as into the vacuum of a common sucking-pump, by atmospheric pressure, and so within the limit of atmospheric pressure, raised 28 or 30 feet. After this was accomplished, the water was further raised by the elastic force of the steam, just as in the engine of the Marquis of Worcester. But the improvement was great, and led to the general use of steam as a means of raising water. Fig. 5 represents this engine as applied to draw water from deep mines. It is placed under ground, on a platform from 20 to 30 feet above the level of the water at \(a\). The chimney ascends in the shaft of the mine along with a pipe, through which the water is forced to the surface. The engine consisted of two boilers, in which the necessary steam was generated, and two receivers with valves. The process of pumping was effected by admitting steam into one of the receivers \(a\), and then cutting off the connection with the boiler. The steam was suddenly condensed by means of a jet of cold water, which, forming a vacuum, the water to be lifted rushed up the pipe \(b\), to refill the receiver. Steam was then admitted from the boiler to press upon the water in the receiver, and all connection with \(b\) being cut off by a valve, the water was forced up the pipe \(c\), and discharged into the trough \(d\). The steam in the receiver \(a\), being again condensed, the process was repeated; and thus by the alternate action of two receivers, a continuous stream was maintained.

One of the first uses of Savery's engine, proposed by himself, was to raise water to fall on a mill-wheel, turning machinery as by a common fall of water. Several engines were erected by a Mr Joshua Rigley at Manchester, and throughout Lancashire, to impel the machinery of some of the earliest cotton mills and manufactories of the district. One was erected at St Pancras, London, at the manufactory of a Mr Kier, where it long continued to turn lathes, &c.

Dr Denis Papin, in 1690, proposed a scheme for producing a vacuum under a piston,—first of all by gunpowder, and afterwards by steam. He had a cylinder \(a b c d\) (fig. 6), containing a piston \(p\), below which he placed a fire, and so as to generate steam from a little water in the bottom of the cylinder. This steam raised the piston, which was secured in its elevated position by a catch; then the fire was removed, the steam... condensed, and the piston was released and forced down by the pressure of the atmosphere. This plan, though crude, contained the earliest suggestion of a vacuum under a piston by the agency of steam.

But Papin halted there. His proposed modes of procuring a vacuum by removing the fire was a step in retrogression from Savery's practice, which was to pour water over the receiver, and operate by surface condensation. In Savery's engine, however, there was the waste of steam consumed in re-heating the receivers and the upper stratum of water, every time it was admitted for the purpose of forcing up the water. It was reserved for Newcomen and Cawley of Dartford to combine the good points of Savery and Papin—a separate boiler and furnace, and a separate cylinder and piston, with a distinct water-pump connected with the piston through the medium of a beam; a means of rapid condensation by injecting water into the cylinder in direct contact with the steam to be condensed, and a system of self-acting valves, which were opened and closed by the reciprocations of the beam, without the usual agency of an attendant. The history of the discovery of the two last-named features is thus concisely given by Dr Desaguliers:—"They were surprised to see the engine go several strokes and very quick together, when after a search they found a hole in the piston which let the cold water in to condense the steam in the inside of the cylinder, whereas before they had always done it on the outside. They used before to work with a buoy in the cylinder enclosed in a pipe, which buoy rose when the steam was strong, and opened the injection-pipe and made a stroke, whereby they were capable of only giving six, eight, or ten strokes in a minute, till a boy named Humphrey Potter, who attended the engine, added what he called a scoggan, by which the beam of the engine always opened and shut its own valves, and then it would go (entirely without the attendance of a man) fifteen or sixteen strokes in a minute. But this being perplexed with catches and strings, Mr Henry Beighton, in an engine he had built at Newcastle-on-Tyne in 1788, took them all away, the beam itself supplying all much better." Thus, for more than half a century, the engine remained in general use without any material change.

Newcomen, Potter, and Beighton had rendered the atmospheric steam-engine an independent self-acting mechanical power of so great perfection in its principle of action, and its minor details, as to be very generally introduced as a substitute for the power of animals in draining mines and collieries, and to confer very great advantages in those important and primary sources of national industry and wealth. The saving of money from this change was so great as to be continually opening up new avenues of mining enterprise, by the rapid progress of which the capabilities of the engine were soon put to the severest trial. The cylinders, which had been originally of 12 and 16 inches diameter, were gradually increased to 60 inches. Along with this dimension, the other parts required to be increased in a still higher proportion; and at last the structure became so gigantic as to demand an amount of science and practical skill which was rare in that period. The man suited to the emergency at last arose in the father of civil engineering, the justly celebrated Smeaton, who brought to bear on this subject endowments and accomplishments seldom united. He conferred upon the atmospheric steam-engine all the extent and variety of application of which it was capable, and all the perfection of proportion and execution which the state of the mechanical arts could then afford. The most magnificent of Smeaton's works in this department is his great Chasewater engine, of which the details are also given in his reports, and which abound in ingenious contrivances and judicious arrangement. This engine was of 150 horse-power, turning out 880 hogsheads of water per hour, by the heat of 163 bushels of coal. The cylinder A B fig. 7, is 72 inches in diameter, the stroke 10 feet 6 inches. The great lever or beam of the engine D D consists of twenty large bulks of timber, the four nearest the centre being each a foot square, and the whole firmly joggled together with heart of oak, and bolted with iron, forming a very elegant but ponderous beam. The beams F F upon which the cylinder rests are kept in their place by being entered into the side-walls of the house, and are joggled and framed together similarly to the great lever; G is the boiler, H the furnace, I B the steam-pipe, J the injection-pipe, K the injection cistern, fed by a pump L, which is wrought by the great lever, M the waste-pipe for the condensed steam. On the left edge of the figure is the spear or rod of the great draining-pump, wrought by the engine; P P is the plug-tree, suspended from the main beam and carrying plugs, which in their upward and downward progress act on the levers which open and shut the regulator and injection-cocks. The date of this engine is 1775. Its working gear, which is very simple and good, is represented on a larger scale in fig. 8, for those who are conscientious enough to study it.

A collateral scintillation of invention was exhibited by Jacob Leupold in 1725, who inverted the ordinary Newcomen's engine, causing the engine to do its work against the atmosphere. He introduced comparatively high pressure steam below a piston, and raised it against the pressure of the atmosphere, plus the active resistance of the column of water forced up by a pump. His machine (figs. 9 and 10) was a true water-pumping high-pressure steam-engine, and might be efficiently used without any alteration at the present day, only the modern machines do the work with less fuel. Two pumps T V for raising water are directly worked by steam by connecting the handles G H of these pumps with the pistons of two high-pressure cylinders \( c \) and \( d \) in such a manner that, when the pistons are raised by the steam, the water is forced up in the pump-pipe \( k \). In fig. 9, at \( a \), the steam is entering the cylinder \( c \), and pushing up the end of the lever \( g \), so as to force the water; and in fig. 10 the steam is shown entering the cylinder \( d \) to work it. This change is effected by turning round the disc \( a \) into the position which reverses the passages. Thus, while the steam is entering the cylinder \( c \) (fig. 9) through \( a \), the steam from the cylinder \( d \) is escaping through \( e \) into the open air; and in fig. 10 the steam is passing into the cylinder \( d \) through \( a \), and out of the cylinder \( c \) through \( e \). The action of this four-way stop-cock is very simple and beautiful, and deserves to be carefully studied. By continually turning it in one direction, communications are simultaneously effected between the boiler and each of the cylinders alternately, and between each cylinder and the open air.

Hitherto the fire-engine, even in Smeaton's hands, was so imperfect, that it wasted a large quantity of fuel and of steam in doing what was useless, viz., heating the cylinder, which was cooled alternately in each stroke by the cold water injected into it. In Long Benton colliery engine, out of 63 cubic feet of steam 32 were thus wasted, and the remaining 31 feet alone performed useful work. There remained, therefore, one-half of the power of the steam and expense of the fuel to be saved by future improvements, provided the useless heating and cooling of the cylinder could be avoided. The vacuum formed below the cylinder was also far from being perfect. In this state James Watt found the atmospheric fire-engine in the hands of Smeaton, and produced from it the pure steam-engine, which he left to us in its present state of high improvement. He was the man who turned the scale of expense in favour of the fire-engine, when it was a more costly power than horses, except when fuel was extremely cheap. In his hands it ceased to be an atmospheric engine, and became wholly a steam-engine, capable of being applied to any purpose, on a much larger scale, and at much less expense than the power of horses. He reflected that, "in order to make the best use of the steam, it was necessary, first, that the cylinder should be maintained always as hot as the steam which entered it; and secondly, that when the steam was condensed, the water of which it was composed, and the injection itself, should be cooled down to 100°, or lower where that was possible." The means of accomplishing these objects occurred to Mr Watt in 1765.—The separate condenser, distinct from but in connection with the steam-cylinder, into which the steam from the cylinder would flow, and in which all the operations of condensation could be performed by surrounding it with cold water, or by injection, or both. The water which would necessarily accumulate in the condenser, Mr Watt proposed to remove by means of a pump. "It next occurred to me," says Mr Watt, "that the mouth of the cylinder, being open, the air which entered to act on the piston would cool the cylinder, and condense some steam on again filling it. I therefore proposed to put an air-tight cover upon the cylinder, with a hole and stuffing-box for the piston rod to slide through, and to admit steam above the piston to act upon it instead of the atmosphere. There still remained another source of the destruction of steam, the cooling of the cylinder by the external air, which would produce an internal condensation whenever steam entered it, and which would be repeated every stroke. This I proposed to remedy by an external cylinder containing steam, surrounded by another of wood, or of some other substance which would conduct heat slowly.

"When once the idea of the separate condensation was started, all these improvements followed as corollaries in quick succession; so that, in the course of one or two days, the invention was thus far complete in my mind, and I immediately set about an experiment to verify it practically. I took a large brass syringe \( a \), 14 inches diameter and 10 inches long, made a cover and bottom to it of tin-plate, with a pipe \( s \) to convey steam to both ends of the cylinder from the boiler; another pipe \( r \) to convey steam from the upper end to the condenser (for to save apparatus I inverted the cylinder). I drilled a hole longitudinally through the axis of the stem of \( r \) the piston, and fixed a valve at its lower end, to permit the water which was produced by the condensed steam on first filling the cylinder to issue. The condenser used upon this occasion consisted of two pipes \( a b, c d \) of thin tin-plate, 10 or 12 inches long and about \( \frac{1}{4} \) th-inch diameter, standing perpendicular and communicating at top with a short horizontal pipe \( h \) of large diameter, having an aperture on its upper side which was shut by a valve opening upwards. These pipes were joined at bottom to another perpendicular pipe \( p \) of about an inch diameter, which served for the air and water pump, and both the condensing pipes and the air-pump were placed in a small cistern \( c \) filled with cold water.

"The steam-pipe was adjusted to a small boiler \( n \). When steam was produced it was admitted into the cylinder, and soon issued through the perforation of the re-heating it, which was unavoidable in Newcomen's engine. We have, then, the boiler or generator, with its appendages; the cylinder or applicator, with its appendages; and the refrigerator or condenser, with its appendages—the function to be discharged by the first of these being altogether the reverse of the last; the first producing steam by heat from water, the last producing water from steam by cooling. Papin's scheme was possible but impracticable; Newcomen's system was practicable but wasteful; Watt's system was practical, economical, and complete.

Mr Watt's first engines are in general form represented by fig. 13, in which a represents the cylinder of his earlier engine, b the boiler, and c the condenser, each with its various appendages. The appendages of the boiler b are, of course, f, the furnace; g, small pipes for showing the height of the water in the boiler; h, a pipe for supplying the boiler with water as it passes off in the form of steam; s, a pipe for conveying the steam to the cylinder. The appendages of the cylinder are—p, the piston, fitting accurately the inside of the cylinder, and surrounded with hemp-packing, soaked with tallow and oil, so as to be steam-tight; the jacket or casing e, e, e, which prevents the cold air of the atmosphere from entering into and cooling the cylinder at the expense of afterwards heating it by the steam; and instead of allowing such air to enter at the top of the cylinder a press down the piston, as in Newcomen's engine, the hot steam is substituted, which, being of an elasticity equal to the force of the atmosphere, presses it towards the bottom of the cylinder. On this being accomplished, the handle of the valve e is raised so as to admit the steam below as well as above the piston, which equilibrium of upward and downward pressure allows the piston again to rise, in consequence of a counter weight connected with the top of the piston-rod r, and this opening, of what is called the steam-valve v, continues until the piston again reaches the top of the cylinder, when it is closed. The eduction valve v, which is at that moment opened, permits the steam to escape suddenly into the condenser, when it becomes water, and leaves the space below the cylinder vacuous, so as to give free space for the piston to be carried down into the cylinder by the pressure of the steam upon the top of the piston. These, the casing, piston, piston-rod, steam-valve, eduction-valve, and communicating passages are appendages of the second great member of the machine, viz., the cylinder, by which the power of the steam is applied to give the required motion to whatever solid machinery may be placed in connection with the piston-rod. The appendages of the condenser c of Mr Watt are as follows:—First of all, a large cistern w of cold water is provided, and furnished continually with fresh supplies of cold water either from a running stream or by means of a pump in or wrought by the engine itself; in this is placed the condensing chamber \( c \times x \), wholly surrounded by cold water, but perfectly empty, excepting that a small jet of cold water from the exterior is admitted through a regulated aperture to play in the inside, by which injection it has always been observed that the condensation of the steam is more efficient than when a casing of metal intervenes between the cold water and the steam. The eduction-pipe \( e \times e \) conducts the steam out of the cylinder by the valve \( v \) into the condenser \( x \times x \), where it is reduced back into the water from which it had been originally generated. Now, it must be obvious on a little consideration, that the water which is injected into the condenser must rapidly accumulate there, becoming at the same time warmed by mixing with the steam, and so would impede the progress of condensation by ultimately filling up the interior of the condensing chamber, which should be kept vacuous; and further, that the steam itself, becoming reconverted into water, would soon accumulate in the condenser and choke it up. Hence a principal appendage of the condenser is a large pump, which is essential to its long-continued efficient action, and which withdraws a portion of the accumulated warm water from the interior of the condenser, and keeps it vacuous; and because there is generally air in the water, and because, also, air is very apt to insinuate itself by many chinks or crevices into the condenser, this clearing pump must be capable of pumping out air as well as water. This appendage of the condenser, represented in the preceding figure by \( y \times y \), is generally termed the "air-pump," a name which but imperfectly expresses its functions.

Fire being placed under the boiler, its heat, communicated to the water, rapidly expands that water, and rarifies it into steam, by the addition of more than five times the heat which would raise its temperature from the freezing to the boiling point. This combination of heat and water, forming the steam, rushes along the steam-pipe into the cylinder casing, and is admitted into the interior of the machine, filling all its chambers and pipes with steam; but that portion of the steam which is in communication with the condenser, being instantly chilled by the jet of cold water and the cold sides of the vessels in the cold well, is condensed, and then the valve \( v \) being closed so as to admit no more steam into the condenser, and the valve \( v \) being closed so as to admit no more steam into the lower part of the cylinder below the piston, there remains the elastic force of the steam above pressing it towards the bottom of the cylinder with a force proportioned to the pressure of the steam and the extent of the cylinder. Thus, a moving power is generated in the cylinder by the steam which may be conveyed through the piston-rod \( r \), and applied through various mechanism of application to the performance of the required work. The steam which has thus pressed down the piston is now admitted below to neutralize the force of that which remains; and having thus done its duty, is again annihilated by the opening of the communication into the condenser, into which it rushes; and being almost instantly deprived of the heat which gave it power and magnitude, there remains nothing except the few spoonfuls of water from which all that volume of steam had arisen, now lying inert at the bottom of the cylinder. This dead water is not yet cold. It is evident that in the primary generation of steam in the boiler, the supply of water must be rapidly diminished by this boiling off, and this water must somehow be supplied. Now, here lies an opportunity for economy; this waste, instead of being supplied by cold water, may be better replenished with the water of the condenser, which is highly heated in condensing the steam from the cylinder.

Mr Watt's engine was first used as a substitute for the engine of Newcomen in pumping up water or draining mines. In 1788 it had attained the form represented in fig. 14, as placed within the walls of a building, the anterior portion of which is omitted, to show the machine. On the left stands the boiler outside the building; and on the right also outside the house, is the large pump, by which the water is raised and the work of the engine performed. Nearly in the middle stands the cylinder, with its appendages, and below these are the well and condensing apparatus.

Beginning with the apparatus for generating steam. It is the boiler of what is called the wagon shape, set in a furnace of brickwork immediately over the fire, which rests on the fire-bars at \( p \), over a deep ash-pit; the flame passes under the concave bottom of the boiler to the further end, and there, instead of proceeding at once up the chimney, returns by \( o \) on the left side of the boiler, through the brick channel or flue, giving out additional heat to the water, and after passing across the front of the boiler, proceeds along the right hand flue \( o \) in the chimney. The draught of the chimney is regulated by the damper \( r \), which is lowered into the flue or raised out of it in any degree by the attendant, and so permits the gases to rush with greater or less ease up the chimney. A tube \( t \), regulated by a stop-cock, comes from a small pump \( u \), on the right hand side of the cylinder, which raises the warm water discharged by the air-pump, and sends it into the boiler, so as to replenish its waste; this pipe and pump being generally named the feed-pipe and tube. The two little tubes proceeding from the water in the boiler are open at both ends, and have external stop-cocks, which are always shut, except when the attendant wishes to ascertain the height of the water in the boiler. He then opens these gauge-cocks, and observing whether water or steam issues from them, forms his judgment accordingly. \( v \) is the steam-pipe which carries the steam from the boiler to the cylinder.

The second great member of the machine, the working-cylinder \( a \), is placed in the engine-house. It contains the moving piston \( n \), which communicates the force impressed on it by the steam, through the piston-rod \( c \), and the chain \( f \) to the end of the great lever or working-beam \( f \times a \times e \), which is forced up and down around the fixed centre or iron gudgeon \( b \), and so raises or depresses the other end of the lever on the right hand side of the figure, and thus... gives the required motion to $hj$ the piston and rods in the barrel of the great pump, in which the work of raising water to a height or from a mine is the useful labour or duty to be performed by the engine. Returning to the cylinder at $A$, we have now to examine the mechanism by which the steam is admitted alternately above and below the piston, through the openings or ports which may be observed on the right hand side of the cylinder at top and bottom. $F$ is the steam-pipe which brings steam from the boiler to the top of the valve-passages, and the pipe $r$ conducts it down to the bottom valves and port at $K$, and the pipe $J$ forming the eduction-pipe, conducts the steam into the refrigerating apparatus, where it is finally condensed. In commencing to work the machine, the duty of the attendant is to allow the steam to pass freely into all the pipes, passages, and ports $F$, $G$, $I$, $J$, &c., filling the cylinder $A$, the condenser $M$, and passing out at an aperture $O$, closed by a valve, called the blow-off valve, by means of which operation the whole of the parts being filled with steam, are rendered vacuous from air, and this preparatory process is called blowing through. At $G$ is a steam nozzle and valve, or regulator, which allows the steam to enter the cylinder at the upper part whenever it is opened, by raising the metallic cover or valve from the opening of the nozzle immediately below, which it exactly fits. At $K$ is a similar contrivance, called the equilibrium valve and nozzle, which admits steam through the pipe $I$ into the bottom part of the cylinder, and the third, or exhaustion valve and nozzle or aperture, allows the final egress of the steam into the condenser. After the engine has been wholly filled with steam, the piston $N$, being at the top of the cylinder, the injection-cock $X$ is suddenly opened, and the cold jet of water playing amongst the steam condenses it instantaneously, forming a vacuum into which the steam from the cylinder instantly rushes, and is in like manner annihilated, leaving the cylinder below the piston equally vacuous; and, of course, the steam from the boiler, on being admitted by the valve $G$ to the upper side of the piston, instantly presses it down into the vacuum below with a force proportional to the perfection of that vacuum and to the pressure of the steam. Thus, the engine makes its first stroke, and raises the water of the great pump on the right of the figure, and the weight of the chain, rod, and bucket, and also a counterpoise $H$, added for restoring the beam to its former position, which it does in the following manner. The equilibrium valve $K$ is opened, and the steam getting admission below the piston, as well as above it, ceases to urge it in either direction, and being thus in equilibrio, the piston would remain passively in its place at the bottom of the cylinder; but the counterpoise $H$, and the weight of the pump, rods, and bucket in the large pump on the outside, draw down the outer end of the great lever or working-beam $f a e$, and so raise the anterior end $f$ and the piston $N$ to the top of the cylinder. The equilibrium valve is then closed at $K$, and the eduction-valve $L$ is opened, so as to allow the steam below the piston to rush down into the condenser and leave a vacuum under the piston, into which it is immediately forced down by the pressure of the steam above $A$, as at first, and raising water at the other end of the beam through a second stroke; and thus, by the continual opening and shutting of the valves by the attendant, the engine performs its work. But we have still to consider the mechanism by which the machine shuts and opens its own valves; for this purpose we have given in fig. 15 two separate and enlarged drawings of one of the valves and its working gear; $l i l$ is a part of the air-pump rod, formed of wood, called the plug-frame or plug-tree, on which are two projecting plugs of wood to work the upper and lower valves. One of these plugs is seen at $i$. As the pluggate moves up and down, the plugs strike the handles or working gear of the valves, and open or shut them at the proper instant. The valve $D$ is called a conical valve, because the small cover $P$ which closes the opening of the nozzle is slightly tapered downwards, so as the more readily to fit its seat, from which it is lifted by a small-toothed rack and pinion $C$, moved by a spindle from without, and communicating by rods with the valve gear at $r$, or at $z$ and $y$ in figure 14. When the plug frame $l i l$ descends, the valve $D$ is closed by the plug $i$, and the valve $K$ is shut, and the valve $L$ in fig. 14 opened by the plug $y$.

Returning to figure 14, where the condensing apparatus and its appendages are placed almost immediately under the cylinder, and to the right of it. The eduction-pipe $J$ conducts the steam into the condensing chamber $M$, which is in the middle of the cold well, wholly surrounded by cold water, and through the regulated aperture $N$ a jet of cold water, pressed in by the atmosphere, is allowed to play in the inside of it amongst the steam. If the air-pump is also placed in the cold well, surrounded by water; $Q$ the piston or bucket of the air-pump, is worked up and down by the piston-rod $q y z g$ from the great lever. The valve $R$ closes when the piston descends, and opens on its ascent, allowing water and air to pass into the air-pump, but preventing their return; and the upper valve of the air-pump allows the escape of water and air outwards, but prevents their return; this valve $S$ leads to the hot well $T$, from which the feed-pump $U$ supplies water for the boiler.

The great advantage of Mr Watt's form consists in avoiding the excessive waste of steam formerly occasioned by condensing in the cylinder itself. The cylinder now is always hot, and therefore perfectly dry. By the time Mr Watt had completed these improvements, his experiments on steam had given him a pretty accurate knowledge of its density; and he found that the quantity of steam employed did not much exceed what would fill the cylinder, so that very little was unavoidably wasted. But before he could bring the engine to this degree of perfection, he had many difficulties to overcome. He enclosed the cylinder in another containing steam, and that in a wooden case at a small distance from it, which effectually prevented all condensation in the inner cylinder from external influence; and the condensation by the outer cylinder was very small.

In order to regulate the power of the engines when working light loads, Mr Watt introduced the variety which he called the expansive engine, the principle of which had first occurred to him in 1769, and was applied by him about 1776. "The steam-valve is always allowed to open fully; the pins of the plug-frame are regulated so that that valve shall shut the moment the piston has descended a certain portion, suppose one-fourth, one-third, or one-half of the length of the cylinder. Thus far the cylinder is occupied by steam as elastic as common air. In pressing the piston farther down, it behoves the steam to expand and its elasticity to diminish. It is plain that this can be done in any degree we please, and that the adjustment can be varied in a minute by shifting the plug-pins.

"In the meantime, the pressure on the piston is continually changing, and consequently the accelerating force. The motion, therefore, will no longer be uniformly accelerated. It will approach much faster to uniformity; nay, it may be retarded, because, although the pressure on the piston at the beginning of the stroke may exceed the resistance of the load, yet, when the piston is near the bottom, the resistance may exceed the pressure."

Hitherto we have considered the condensing steam-engine of Watt as applied to work the large pumps used to draw water from mines, or to supply reservoirs from a lower level. This, indeed, was the most obvious and immediate application of the steam-engine, which was at first introduced as a substitute for the atmospheric pumping-engine of Newcomen. But it had always been matter of regret that one-half of the motion was unaccompanied by any work. It was a very obvious thing to Mr Watt, that as the steam admitted above the piston pressed it down, so steam admitted below the piston would press it up with the same force, provided a vacuum was made on its upper side. This was easily done, by connecting the lower end of the cylinder with the boiler, and the upper end with the condenser. The steam-engine of revolution of Mr Watt was an invention subsequent to the mining steam-engine, or "water commanding machine." Previously to the time of Watt, indeed, there had been a few attempts to produce a revolving motion by steam, such as the case where the engines of Savery and Newcomen drew up water to turn a wheel. There had also been many attempts to apply the old pumping-engine directly to this purpose. Jonathan Hulks, Keane, Fitzgerald, Mr Oxley, John Stewart, and Matthew Wasbrough, had all contrived some means of producing a revolving motion from the reciprocation of the great beam; but Watt's engine alone was capable of being rendered an efficient and economical motive-power for driving machinery of continual motion. The earliest record of the invention was in 1774. The Albion Mill

engine (figs. 16 and 17) were amongst the earliest rotative engines made for sale.

The steam-pipe r conveys steam from the boiler n to the cross-pipe or upper steam-nozzle o, and by the perpendicular steam-pipe t, to the lower steam-nozzle x. In the nozzle o is a valve, which, when open, admits steam into the cylinder above the piston b (fig. 16), through the horizontal square pipe at its top; and in the lower steam-nozzle k there is another valve, which, when open, admits steam into the cylinder below the piston. In the upper exhaust-nozzle h is a valve, which, when open, admits steam to pass from the cylinder above the piston into the exhaust-pipe j, which conveys it to the condensing vessel m, where it meets the jet of the injection from the cock s, and is reduced to water; and in the lower exhaust-nozzle l there is also a valve, which, when open, admits steam to pass out of the cylinder below the piston, by the eduction-pipe into the condenser m.

"The piston being at the top of its stroke, the valves o and t are to be opened, and the fly-wheel x turned by hand about the eighth of a revolution, or more, in the direction in which it is intended to move; the steam which is then in the cylinder will pass by l into the condenser, when, meeting the jet of water from the injection-cock, it will be converted into water, and the cylinder thus becoming exhausted, the steam, entering the cylinder by the valve g, will press upon the piston, and cause it to descend, while, by its action upon the working-beam through the piston-rod, &c., it pulls down the cylinder out of the beam and raises up the outer end and the connecting-rod h, which causes the planet-wheel i to tend to revolve round the sun-wheel j; but the former of these wheels being fixed upon the connecting-rod, so that it cannot turn upon its own axis, and its teeth being engaged in those of the sun-wheel, the latter, and the fly-wheel, upon whose axle or shaft it is fixed, are made to revolve in the desired direction, and give motion to the millwork.

"As the piston descends, the plug-tree z also descends, and a clamp or slider q, fixed upon the side of the plug-tree, presses upon the handle l of the upper y shaft, or axis, and thereby shuts the valves o and t; and the same operation, by disengaging a detent, permits a weight suspended to the arm of the lower y shaft to turn the shaft upon its axis, and thereby to open the valves x and h. Just before the opening of these valves, the piston had reached the lowest part of its stroke, and the cylinder above the piston was filled with steam; but as soon as h is opened, that steam rushes, by the eduction-pipe j, into the condenser, and leaves the cylinder empty above the piston. The steam from the boiler entering by t and x, acts upon the lower side of the piston, and forces it to return to the top of the cylinder. When the piston is very near the upper limit of its stroke, another slider a raises the handle q, and in so doing disengages the catch, which permits the upper y shaft to revolve upon its own axis and open the valves o and t, and the downward stroke recommences, as has been related.

When the piston descends, the buckets n, r of the air-pump p and hot-water pump r also descend. The water which is contained in these pumps passes through the valves of their buckets, and is drawn up and discharged by them through the lander or trough by the next descending stroke of the piston. Part of this water is raised up by the pump for the supply of the boiler, and the rest runs to waste.

The history of the steam-engine in a great measure ends with the history of Mr Watt's labours. There remain, of course, important improvements in points more nearly of detail; and it would be unfair to ignore the doings of Watt, and others, in the article of boilers. For a consideration of these and other points arising for discussion, the reader is referred to the succeeding portions of this treatise.

The names of many individuals who have in earlier times distinguished themselves by ingenuity directed to the steam-engine have been omitted in this notice, not because we consider their labours either undeserving of notice or uninteresting to the general reader, but because they have not contributed towards the production of the modern steam-engine, and because an account of their works would rather serve to illustrate the possible varieties of the machine and the fertility of the human mind in mechanical devices, than either to conduct the reader along the stream of historical succession, or render him better acquainted with the nature and mechanical peculiarities of the steam-engine itself.

For the important purpose of converting reciprocating rectilineal into rotary motion, Mr Watt had early designed the application of the crank. A crank is an elementary machine, which has been used from the earliest times for converting a revolving into a rectilineal motion, or the reverse. A crank is merely a handle to a wheel, by which it may be turned round. In fig. 18, let ax be an axis of a wheel bed, and abp the usual bent (or crooked) handle by which it is turned round by the man, whose arm first pushes it from him, and then draws it towards him, and so continually turns the wheel round, then the part ax radiating from the centre, is called the crank, the axis ax is called the crank-axle, and the straight part pb is called the crank-pin.

The idea of the crank had very early occurred to Mr Watt; "but my attention," says he, "being fully employed in making and erecting engines for raising water, it remained in petto until about the year 1778 or 1779, when Mr Wasbrough erected one of his ratchet-wheel engines at Birmingham, the frequent breakages and irregularities of which recalled the subject to my mind, and I proceeded to make a model of my method, which answered my expectations; but having neglected to take out a patent, the invention was communicated by a workman employed to make the model to some of the people about Mr Wasbrough's engine, and a patent was taken out by them for the application of the crank to steam-engines. This fact the said workman confessed, and the engineer who directed the works acknowledged it, but said, nevertheless, the same idea had occurred to him prior to his hearing of mine, and that he had even made a model of it before that time, which might be a fact, as the application of a single crank was sufficiently obvious. In these circumstances I thought it better to endeavour to accomplish the same end by other means, than to enter into litigation, and if successful, by demolishing the patent, to lay the matter open to everybody. Accordingly, in 1781, I invented and took out a patent for several methods of producing rotative motions from reciprocating ones, amongst which was the method of the sun and planet wheels, above described."

Oliver Evans, of Philadelphia, United States, devoted himself to the development of the non-condensing, high-pressure steam-engine, in which the condenser was superceded, and the steam was treated as in Leupold's engine, and exhausted into the atmosphere. Before 1785 he had erected Valves and made experiments upon a high-pressure engine, which seems to have been in all essential respects similar to our own. Indeed, it appears that the Americans have taken the form and arrangements of their engines from Evans, as implicitly as in this country we have adopted those of Watt. The history of Evans consists almost entirely of the romance of real life. Sanguine and energetic, he continually encountered difficulties only to overcome them, and to encounter renewed disaster and disappointment; till he at length died of a broken heart. To him may be attributed the rapid advancement of America, at the commencement of the present century, in all that relates to the introduction of the steam-engine in its multifarious applications, and especially in steam navigation. He had awakened in that nation a lively sense of the advantages they were likely to derive from the power of steam, and placed in their hands an instrument well fitted for their use, and which they were not slow to adopt and apply.

SECTION II.—CYLINDER-VALVES AND THEIR MECHANISM—HISTORICAL AND DESCRIPTIVE ACCOUNT.

The modern non-condensing steam-engine is in principle the same as Leupold's engine; and if the latter be supposed to be consolidated into one cylinder, double-acting, analogous to the conversion of the single-acting condensing engine by Mr Watt into double action, its operation would be represented by the figure 19, which comprises an automatic valve, or four-way cock, opened and closed by the reciprocating movements of the beam. A vertical rod, ii', being suspended from the lever, with two plugs by which the handle n of the valve is raised; in that position the steam enters at s, and passes up the superior passage into the top of the cylinder, forcing the piston down, while the steam, acting below the piston, finds free egress along the inferior passage b, through the valve, and escapes by the eduction-pipe e into the open air. Just before the piston gets to the bottom of the stroke, the plug i strikes the handle ii into the position ii', when the direction of the passages is reversed; the steam enters below the piston, and escapes from above. The four-way cock was superseded by Murray's "slide-valve," now in ordinary use, by means of which the distribution of the steam was more simply effected. In this case all the four passages are united in a square box called a valve-box, or valve-chest, as in fig. 20; s, e, a, n being the steam, eduction, upper and lower passages. Into this box is introduced a small valve or cover p (fig. 21), which is of such a size as at one time to leave open only one of the three openings on the right; so that, by covering two of the openings a and e, as in fig. 22, the steam from s can only find its way through n into the lower part of the engine, while the steam already in the upper part of the cylinder can find its way, below the valve p, into the eduction-pipe e, so as to escape into the air. The valve is next shown in fig. 23, in its middle position, where all the three passages are closed, preparatory to reversing the direction of the steam, as in the third position when it slides from the upper port a, as is shown in fig. 24, so as to allow the steam to Cylinder-Valves and their Mechanism.

In the figs. 27 and 28, the valve is shown in the working position. In fig. 28, the steam from s rises up along the centre of the slide, and enters the upper port A, while the steam in the under part of the cylinder has free egress through B to the eduction-pipe E. In fig. 27, the steam has free access to the lower port N, while the steam already above the piston has free egress through the upper port A to the eduction-pipe E. In this species of slide, there is scarcely any loss of steam in the passages, as it is cut off close to the cylinder.

Instead of the long D slide, which is very heavy, on a large scale, two short slides, similar to its two ends, and connected together by bars, have been used in the following form. Fig. 29 is a section of the slide; fig. 30 a face view; and fig. 31 a section of the cylinder with the valves in their places. In this case, however, there are two eduction-pipes E, E instead of one, as formerly, and the steam-pipe s enters between the valves.

A cylindrical slide-valve of the following form has been used in a considerable number of engines, and works well. The valve-chest is an upright cylindrical pipe P Q, the inside of which is bored truly cylindrical, and is exactly fitted by two metallic cylindrical plugs, which are ground so smooth in their places as to be steam-tight. It will be apparent from the figure that these two plugs, being raised and depressed by the valve-rod which connects them, will effect the same purpose as the former valve. (Fig. 32.)

The conical-valve, spindle-valve, or button-valve, as it is variously designated, is a species introduced by Mr Watt, and improved by his assistant Mr Murdoch, from whom the steam-engine of Watt has received many valuable ap-

passage B, under the valve D, by the eduction-pipe E. This valve, named from its figure the D valve, is also worked by the machine itself, either by some of its moving parts striking plugs on a rod which is fixed to the valve, or by some of the other apparatus, which will afterwards be described.

Another form is that called the long slide or long D valve, the invention of Mr Murdoch, which gives the advantage of shutting off the steam close to its ingress into the cylinder, and so saving what in the common short D slide is lost in the passage from A and B to the ends of the cylinder. It is formed thus:—The valve-chest extends along the side of the cylinder. It is shown in fig. 25 without the valve. In fig. 26 the long D slide valve is shown separately. It is a sort of pipe, extending along the whole length of the cylinder. Towards the ends, this pipe is almost semicircular, with two narrow flat plates capable of covering the openings or ports of the cylinder. This pipe is left open, and perfectly clear from the one end to the other, so that the steam may traverse it freely lengthwise. The steam-pipe is represented as entering the valve-chest from below at S, and the eduction-pipe in the middle as at E. In this valve-chest are placed packing-boxes, as they are called, immediately opposite the ports of the cylinder. They contain soft, elastic hemp, soaked in oily matter, the object of which is to press against the truly cylindrical and polished outside of the slide-valve when it is placed, and make steam-tight partitions in the valve-chest, to prevent communication between the middle and the two ends. pendages and much of its practical perfection. It has been applied in two forms. Mr Watt's own form, the earlier one, is given in the following figures. For a single engine four valves are required. One of them is represented separately in figures 33, 34, which are vertical sections through the valve, at right angles to each other. The valve is shown open in fig. 33, and shut in 34. s is the entrance of the steam, a the port, v the conical valve, and x the seat or nozzle which it covers. On a cursory glance, it is evident that when the conical cover v of the aperture x is up, as in the first diagram, the steam has free entrance; and when it is closed, the steam will merely press the valve down into its seat, without obtaining an escape from the nozzle.

The last valve which may here be described is the crown-valve, or equilibrium-valve, which is in use on the Cornish engine, and has also been introduced into rotative engines. Its value consists in effecting a large opening; and requiring little force to work it, while large valves of the common sort are heavy, or are so much pressed in one direction by the steam as to require great force to work them.

The crown-valve is so named from its resemblance to a diadem. Conceive a chamber, fig. 35, out of which an aperture a leads into the cylinder, and into which a pipe s brings steam. The aperture a is surrounded by an upright ring or collar rising a few inches into the chamber, which ring is on all sides perforated by slits of considerable size, but closed at the top. Figure 36 represents the crown or cover of this valve, which is also a ring attached to a steel rod or spindle, by which it is raised or depressed. All round at the top and bottom, the collar in the chamber and the crown-valve are ground so as accurately to fit each other. Fig. 37 shows the valve on its seat and closed on all sides, so that no steam can find admittance; and fig. 38 represents it open or raised up from its seat, with steam entering freely on every side. There are two ways in which valves are worked by the steam-engine itself. The first of these is by the agency of some part of the engine that happens to move up and down, or performs a reciprocating motion, and the other is by the agency of some part which revolves. The following is a simple method, which has been applied to the short b slide, already described. In fig. 39, a is the cylinder, p the piston, acting on the end l of the great lever l r, raising and depressing it alternately; while the other end l r, united by the connecting-rod l r to the crank of the fly-wheel, turns it round. The manner in which the steam-valves are moved, is by the long vertical bar or plug-rod rr, suspended from the lever l r, so as to move up and down with it. This bar rr carries two projecting plugs of wood ss upon it, which strike alternately up and down upon the handle h at the bottom and Valves and their Mechanism.

In the second system of valve apparatus, by which the steam-engine is rendered automatic, the steam-valves are worked by the revolving of one of the shafts or wheels. Of the various methods in which this has been done, the following are some examples. On the axis o of the crank (at the bottom of the right side of fig. 40), which is turned round by the rod l r during each alternate ascent and descent of the piston, is placed a cam or projection. A square frame s s s s encloses this cam. As the axis turns round, the cam comes into such positions as to bear upon the sides s s s s of the frame successively, and so pushes the frame towards the right and left alternately. s e v is a bell But that modification of this principle which is in by far the most general use, is in the form called "the eccentric;" which is a circular disc, or ring of metal, placed upon the shaft or axis, turned by the crank. In fig. 44, \( \phi \) is the centre of the shaft or axis, to which revolution is given by the crank \( R \) of the steam-engine. On this axis the circular disc \( E \) is placed, but eccentric to it, so that its centre \( d \) moves round the axis. The distance of the centre \( d \) of the disc from the centre \( o \) of the axis is called the eccentricity, and it is equal to half the throw or range of the motion of the valves to be moved by the eccentric. The rod, fig. 45, is called the eccentric rod, and is attached to a hoop, or circle, that exactly fits the eccentric disc. The various positions which the eccentric will take during the revolution of the engine is shown in figs. 46 to 49.

In the application of valve-gear to railway locomotives, it appeared that the gearing must be capable of being readily reversed. In the early engines of the Killingworth Railway, before 1820, the valves were moved by means of Valves and a "square box or tumbler, which was superseded by a loose eccentric, fitting the driving-axle (fig. 50), a motion which exactly answered the purpose, and was employed till the revolutionary year in railway history, 1829. A lever was fixed upon, and revolved with the driving-axle, formed with a stud, which entered and slid freely in a concentric groove cut in the body of the eccentric. The stud found its way to one end of the groove, and determined the position of the eccentric on the axle, for the fore or back gear. The small end of the eccentric rod was permitted some longitudinal play in the eye of the intermediate lever, adjustable by nuts. With an adjustable eccentric, of sufficient throw, and the adjustable limits of the travel of valve, the valve was quickly opened and closed, and its movement was equally good for both directions. The motion so derived was obviously similar to that of the ancient tappet frame, or the more modern cam. Loose eccentric gear was employed by Hackworth, and in the original inside cylinder engines of the Liverpool and Manchester Railway. The two eccentrics were cast in two pieces, and bolted together into one mass, capable of sliding laterally on the axle between the cranks. No play was permitted between the eccentric rods and the valve-levers; a necessary precaution in high-speed engines. The eccentrics were engaged in fore and back gear, by two snugs or catches fixed on the axle, behind each crank; the locking of the eccentrics being accomplished by a forcible lateral movement. Separate mechanism was employed to control the small ends of the eccentric-rods, formed with gabs to disengage them when it was necessary to work the valves by "hand-gear," which was occasionally required at starting, as in Bury's valve-gear, fig. 51. But the system of loose eccentrics was, for locomotives, cumbrous, complicated, abrupt, and easily deranged. It was, therefore, readily abandoned for a simpler and firmer plan. Two fast eccentrics were substituted for the loose ones, one to each cylinder. This was partially a reversion to the primitive plan of the fixed cam; and the same difficulty of working equally well in fore and back gear was encountered. Mechanism suitable for working each valve with one fixed eccentric, affording the required lead both ways, was invented by J. and C. Carmichael, of Dundee, in 1818, and has been variously applied by them. For locomotives (fig. 52), the eccentric rod was finished with a double fork, to gear with the pivots of a double spanner on the traverse shaft of the valve, and was

lower or upper spanner. The spread of the forks enabled them to engage the pivots of the spanner in all positions of the valve, and to bring them home to the gabs. This motion, duly proportioned, preserved the lead of the valve both ways—a matter of fact which for many years proved a pons asinorum for ingenious youths. It was a real advance upon the system of the loose eccentric, as one handle sufficed for working the gear, and all the parts were solidly put together.

The inconvenience of combining in one eccentric the functions of two, was removed by the adoption of four fixed eccentrics, of which the authorship is uncertain. Two were provided for each cylinder, for working respectively in fore and back gear. Each eccentric had its own fork; and though the plan entailed the use of four eccentrics, and additional bearings, the increased workmanship was compensated by precision and certainty of action. The first application of four eccentrics to locomotives was made by the Hawthorns of Newcastle, in 1837. It has been ascertained, however, that some time previously, an ingenious mechanic, in private circumstances, residing at Newcastle, had contrived and constructed an efficient model of valve-gear for locomotives, in which four eccentrics were employed, in the manner afterwards wrought out in practice.

In all the forms of double-eccentric gear, it was necessary to effect simultaneously the disengagement of one pair of eccentrics, and the engagement of the other. Various plans of reversing were adopted. Stephenson employed two transverse shafts (fig. 53), the principal of which was worked by the reversing handle, and commanded the four gabs; the secondary shaft was linked to, and worked by, the principal shaft, and it had charge of the back gabs. All the forks geared from below, and one movement of the reversing handle elevated one pair and lowered the other. Thus, the manipulation was simple and easy, as the reciprocal action of the gear balanced the weight of one pair of rods with that of the other. To consolidate the gearing by dispensing with the second shaft, Buddicom employed a system for the outside-cylinder engines of the Paris and Rouen Railway, in which the forks were opposed, and worked by one lever on the reversing shaft, which was placed below, and left clear head-room under the boiler.

In these plans an intermediate traverse shaft was necessary to transmit the motion to the valves, as the valve-spindles could not, according to the prevailing arrangement of the steam-chest over the cylinder, be in the same horizontal plane with the driving-axle. A modification was introduced, however, by which the valve-chests were removed from the top and placed between the cylinders, uniting into one capacious chest, with vertical valve faces, which brought the valve-spindles to the level of the driving-axle, to dismiss the intermediate shaft, and to work the valves directly. The forks were transferred to the valve-rods, and the eccentric rod-ends formed with plain pins, and linked. This motion, as it was the most direct, and involved the fewest parts, was the best of all that had yet appeared.

EXPANSION VALVES.

The conical valves of Watt, operated by tappets, which opened and closed them at any desirable point of the stroke of the piston, afforded ample range for expansive-working; and equal facilities were afforded by the use of cams. But when slide-valves and eccentric motion came in, the benefits of expansive-working and a clear exhaustion of steam were sacrificed to the simplicity and smoothness of action of the new mechanism of distribution. The essential proportions of the primitive slide-valves, and the steam-ports over which they travelled, are shown in section in fig. 54, in which it is apparent that the length of the valve was a very little greater than the extent of the ports \(a\), \(a\), so as barely to lap one-sixteenth inch over these openings, just as an assurance that the steam should not enter at both ends of the cylinder at the same time; the body of the valve was made just so long that the walls of the cavity \(c\) just closed the steam-ports \(a\), \(a\), on the inside, when the valve stood midway. Engineers desired chiefly to insure timely and free admission of steam, overlooking the much greater necessity that existed for an early and liberal exhaustion; this valve, with such properties, was in common use until 1838. The steam-passage was opened to the exhaust immediately after it was closed for admission, and both events took place at the termination of the slider. Thus, no time was allowed for exhaustion previously to the beginning of the succeeding stroke; and so little were the defects of this valve understood in locomotives, where the highest speeds of piston were practised, that when, in 1836, according to Mr Edward Woods, short-stroke passenger-engines were introduced on the Liverpool and Manchester Railway, to run at high speed, their greater consumption of fuel was ascribed to the supposed mechanical disadvantage of the short stroke. Mr John Gray, then locomotive-foreman at Liverpool, appears to have been the first to have suspected the cause of the excessive consumption of fuel; and after having made a few preliminary experiments on the valves of the stationary engine, and satisfied himself that the evils complained of could be traced to defective exhaustion, he lengthened the valves of one of the locomotives, which had originally a "lap" over the steam-port at each end, of only \(\frac{1}{2}\) inch to \(\frac{3}{4}\) inch. The effect was, that the eccentric being shifted and advanced on the shaft, so as to cause the valve, with the additional lap, to be just open at the beginning of the stroke, the inside was open at least \(\frac{3}{4}\) inch to the exhaust, whereas, previously, it was not open at all. The consumption of Expansion fuel was, in consequence of the timely and more efficient exhaust, reduced about one-fourth. Further experience led the way to the adoption by Mr John Dewrance of 1 inch of lap at each end of the valve, with a travel of 4 inches, as depicted in fig. 55. By the use of the long valve, which just opened for steam at the beginning of the stroke, and stood 1 inch open for the exhaust, the steam was cut off at 79 per cent. of the stroke, expanded in the cylinder to 95 per cent., or 5 per cent. from the end of the steam-stroke, and at this point it was exhausted. The waste steam which had previously been choked up in the cylinder, owing to the difficulty of escape, and so causing excessive back-pressure on the piston, was freely released, less steam was thus consumed, and other natural advantages accrued.

The necessity and advantage of lap on the valve was thus established in railway practice, both as to its facilities for affording a free exhaust, and for working steam expansively. Without lap there could be no expansion; and though it was introduced primarily for the purpose of an efficient release, its advantages as a means of working expansively became likewise apparent. It was quickly adopted on other lines of railway, as a specially good thing for high-speed engines, and by slow degrees circulated also in stationary and marine practice, in which, certainly, pistons moved at a more moderate velocity. In order to vary the degree of expansive-working of steam, according to circumstances, two classes of mechanism have been employed for slide-valves—first, those mechanisms which operate upon single valves by varying their travel; second, those in which two valves are employed, one of which is specially designed for varying the expansion. Increase of expansion, it must be observed, is obtained simply by causing the valve to cut off the steam earlier in the course of the stroke. In mechanisms of the first-class, the travel is varied by means of mechanism external to the valve-chest, of which the merit of the first application was made by Mr John Gray, on the Liverpool and Manchester Railway in 1839. In his plan, the pin at the end of the eccentric-rod slides in a segmented lever, curved to the radius of the rod, the upper end of the lever being linked to the valve-rod; thus the travel of the valve could be varied by raising or lowering the eccentric-rod end in the segmented lever. The celebrated "link-motion" was originally introduced in 1843, by Robert Stephenson and Co., on locomotives, and is now universally applied to them; it is also extensively applied in marine and winding engines. It is shown at fig. 56, as applied by that firm to locomotives, and is specially adapted for situations where the engine requires to be quickly and easily reversed, whilst its capacity for advantageously working by variable expansion, in varying the travel of the valve, is unquestionable. The two eccentrics, fore and aft, usually provided, have the ends of their rods connected by a slotted link, the slot of which embraces the end of the valve-spindle from which the valve is worked. The link is commanded by a reversing lever, and may be raised or lowered, so as to bring the valve and spindle into gear with, and receive the motion of, the fore or the back eccentric; and, while the link would partake of the two motions jointly of the two eccentrics, its horizontal motion would be smallest at the centre of its length, and would be extended towards the extremities. By shifting the block or end of the valve-spindle towards the centre, the travel of the valve is reduced, and variable expansion obtained.

Variable-expansion gear of the second-class, with two or more valves, has been applied in many different forms. The use of a secondary valve, in addition to the principal valve for expansive-working, is now limited to stationary engines. For locomotives, it has often been tried and abandoned, except in the United States, where, however, the link-motion with a single valve is steadily superseding the double-valves.

SECT. III.—HISTORICAL NOTICE OF STEAM-BOILERS.

During the first period of the history of the steam-engine, the danger of bursting the boiler, and the difficulty of making it strong enough to resist the internal force acting towards explosion, and also of making the joints tight against the leakage of highly elastic steam, formed the chief obstacles to the introduction of steam as a mechanical mover. The first important point in preparing a steam-boiler is to secure strength without unnecessary expense. The globular or spherical shape was very early adopted as one of greatest capacity, as a shape in which, the pressure at every point being equal, there remained no force tending to produce flexure, or destroy the equilibrium of strength and strain at any point. A fire was lighted below the boiler, and the steam confined until the heat had raised it to the required pressure. This form was accordingly adopted by Hero, Savery, and others, as already noticed.

It was soon found that a spherical boiler, when set upon an open fire, required an enormous consumption of fuel to raise a small quantity of steam, the heat being copiously communicated not only to the water in the boiler, but also in very great quantity to the surrounding objects, besides being rapidly carried off by the air. To surround the spherical boiler with non-conducting substances, and to keep the flame throughout its whole extent in contact with the boiler, so as to prevent radiation to surrounding objects, and also to diminish the size of the fire by making it wind round the boiler, were the first steps towards improvement; and we accordingly find in the work of Dr Desaguliers the form of boiler (fig. 57), which is built into a brick casing, with the fireplace below, and the flue winding round the boiler on its way to the chimney.

The form next in simplicity to the spherical boiler is the cylindrical. From the facility with which a cylinder is made, it was introduced at a very early period. It stood upright like a bottle, as in fig. 58, the fire being placed at the bottom, and the flue winding round that part of the sides covered with water. This form of boiler was found, however, to be weak in the bottom part.

For the prevention of these two evils, the cylindrical form of boiler was very soon modified and improved by two opposite expedients, one applied at the top and the other at the bottom of the cylinder. The top being made hemispherical, possessed all the advantages of a spheri... Historical boiler; and the bottom being arched upwards, so as to present a large concave dome to the impact of the flame.

This dome being sustained by the cylindrical belt round its spring, a very strong and extensive surface was obtained, as in fig. 59.

In this cylindro-spherical boiler it was found that the action of the flame on the upright round sides produced but a very slight effect in raising heat. It was therefore desirable that the flame should be brought somewhat under the sides by inclining them a little outwards. From the form the boiler then assumed, and which has since become very common, it has not inaptly been named the hay-stack boiler, fig. 60. The same effect was next obtained in many of the boilers of Newcomen in the way represented in fig. 61, so that the flame in the flues impinged upon a surface directly over them, the flues in this case forming a recess in the sides of the boiler, instead of being built around it by the brick-work alone.

In process of time, boilers of much larger size came to be required, and the spherical shape was found cumbrous and too capacious; that is to say, contained an enormous mass of water, which it required much time and fuel to heat to the boiling point before any steam could be raised. The diameter, also, of the boiler was so great when much steam was required, that the enormous dome became weakened. To make a stronger boiler, and one which should at the same time cover a large fire, the waggon-boiler (fig. 62), so named from its form, was introduced by Mr Watt. It was made of considerable length, and its transverse section resembles that of the old cylindrical boiler. In this form the boiler was long made by Messrs Boulton and Watt. It was afterwards improved by hollowing the sides (see figs. 63 and 64) in order to bring them more immediately over the flame. These forms of boiler, although very convenient, are weak; they are very different from the spherical or cylindro-spherical boilers in strength and safety, and it is necessary to place in them strong iron stays, which are essential to strength and security in boilers having large surfaces, concave outwardly, or perfectly flat. To avoid the use of stays, and to secure great strength without any other metal than the shell of the boiler itself, is the object of the cylindrical boiler with two hemispherical ends (fig. 65), laid with its axis nearly horizontal, and below it, at one end, is placed the fire, enclosed by brick, as usual. The flame traverses the bottom of the boiler, beating directly upon its under horizontal surface, till it reaches the end furthest from the fire. The flame and hot air then, in some examples, return along the one side of the cylinder, being confined in a brick flue, and passing along in front of the end which is over the fire, traverses the other side towards the chimney, which it enters after having thus traversed the length of the boiler three times, and applied its heat successively to every point of the cylinder, which is covered with water. This is a boiler that requires no stays, and is valuable where room is not important. It contains much water, requires much heat to raise its temperature after being cooled at night, and is very bulky.

The spherical, cylindrical, and waggon shaped may properly be denominated the simple boilers. But some hundred kinds of boilers have been invented for different purposes; almost all of them designed to save either bulk, weight, or fuel. For these purposes one great object of improvements in boilers has been to increase as much as possible the extent of heating surface without increasing the general dimensions. Thus Boulton and Watt have inserted a flue in the middle of the large waggon-boiler in the manner shown in figs. 66 and 67, so that, after the flame has passed along the bottom of the boiler to the further end, it returns along the flue in the middle of the water to the front, and then makes an entire circuit of the outside of the boiler before entering the chimney. The same plan has been employed in cylindrical-boilers, the flame and hot air being made to traverse a hollow tube or cylinder in the interior of the boiler; sometimes several such flues have been used. In these boilers, a large surface is still exposed to the cold air; and the brick-work, in which the fire is placed, radiates off a considerable portion of heat, which is lost. To remedy this evil, the furnace has been so contrived that the fire is in the inside of the boiler. This was probably done for the first time by Smeaton, who succeeded in producing almost as high a proportion of steam from fuel as engineers of a more modern date. His portable engine-boiler is represented in figs. 68, 69. The interior of this hay-cock boiler contains a hollow ball of cast-iron, in which the fuel is burned. Air enters by an aperture at the bottom; a large cast-iron pipe leads through the water to the door, and another pipe, in the opposite direction, passes through the water, conducting the products of combustion to the chimney, immediately round which are introduced the fresh supplies of cold water for replenishing the boiler. But a much better boiler than this, and one which indeed might bear comparison with many boilers of the present day, is given by Mr Farey as the invention of an unknown author. In the centre of a large old-fashioned hay-stack boiler, figs. 70, 71, is placed a large round furnace, from which there passes a simple rectangular flue, winding round and round the boiler in spiral circuits till it reaches the outside, and thence passes to the chimney. In the same way, it has often been provided, that the furnace should be in the interior of a cylindrical boiler, by placing another cylindrical tube, of large dimensions, in the interior of the outer-case, as in fig. 72, to serve at once as furnace and flue.

This was probably first done by Trevithick, the advocate of high-pressure engines in this country.

To the great central flue there have been sometimes added lateral flues on each side, for the return of the products of combustion, fig. 73. Thus, again, this internal flue has been made elliptical, fig. 74, a weak, and therefore dangerous, form. It is one of the faults of the boilers that have their fires in the internal tubes, that the ash-pit and interior of the furnace over the fuel are so confined as to prevent that perfect combustion of fuel which may be obtained by a deep ash-pit, a large expanse of fire-grate, and a deep and wide furnace. These evils may, in some measure, be obviated by an internal flue of large dimensions, or by the following species of boiler, figs. 75, 76, where the fire is still surrounded by water, and gives ample room for the most perfect combustion. In this species of boiler, the tube opens out at the front, so as to leave a semi-cylinder above the fire, and two vertical spaces, or "water-legs," as they are called, which cover the fire on both sides; thus obstructing the heat that would otherwise pass away into the brick-building, and, at the same time, covering a large and wide space of furnace-bars, a deep ash-pit, and so insuring adequate combustion. The internal surface of this boiler has been still further increased, by substituting for this single tube a number of smaller ones, which in some cases are not more than two inches in diameter, as in figs. 77, 78. After passing through all these tubes, the flame and hot gases again return along the bottom and sides on the right of the boiler, and pass back on the other side to the chimney. The Butterfly boiler (figs. 79, 80), and similar to this, has been much used in Lancashire. It has the large internal flue, but wants the fire-legs, and, in this respect, is inferior to the former. PART II.—THE MODERN STEAM-ENGINE.

SECTION I.—THE MECHANICAL ORGANS OF THE MODERN STEAM-ENGINE—THEIR DYNAMICAL ACTION.

CHAP. I.—ENUMERATION AND DEFINITION OF THE ORGANS OF THE ENGINE PROPER.

The steam is conducted to the engine through the steam-pipe, in which the stop-valve is placed; also the throttle-valve, or regulator, for adjusting the supply of steam to the cylinder; the supply may be regulated by hand or by a governor. The steam-pipe sometimes contains also the cut-off or expansion-valve, for cutting off the steam at any part of the stroke, and may be controlled by the governor.

The cylinder may be single or double-acting. When single-acting, or operating in one direction only under the steam pressure, the piston is made to return in the opposite direction by the action of a weight or counterpoise. In a double-acting cylinder, the steam is admitted on the piston at both ends alternately. The steam passes into the cylinder through steam-passages, or steam-ways, or ports, the entrances to which are sometimes called specifically the ports, sometimes nozzles, opened and closed by the induction and eduction-valves. When these are in one, it is called a slide-valve, which is placed in the valve-chest.

In non-condensing engines, conventionally called high-pressure engines, the waste steam discharged from the cylinder escapes into the atmosphere through an exhaust-pipe, or blast-pipe as in locomotives. The cylinder-cover has a stuffing-box to pass the piston-rod, which, if tubular, is called a trunk. A grease-pan is fixed to the cylinder, or the cover, for the lubrication of the piston. An escape-valve, held by a spring, may be placed at each end of the cylinder, or blow-through, or cylinder-cocks, for the escape of water collected in the cylinder, either by condensation or by the priming of liquid water from the boiler. To prevent condensation in the cylinder, it is cased in a jacket, filled with steam from the boiler, or hot air; and the jacket is covered with clothing, or cleading, of felt and wood.

Double-cylinder engines have two cylinders, the steam being admitted from the boiler to the first cylinder, and filling the second by expansion from the first. The condenser is a steam and air-tight vessel, into which the steam from the cylinder is discharged and condensed by a shower of cold water from the overheated injection-valve. In land engines the injection-water is supplied from the cold well surrounding the condenser, which is filled by the cold water pump. In marine engines the water enters direct from the sea. In the surface condenser the steam is condensed within tubes, or other passages, surrounded by cold water or air. The blow-through valves connect the cylinder with the condenser, and there is a snifting-valve opening to the atmosphere; through these valves steam may be blown to expel air from the cylinder and condenser before the engine is set to work. The vacuum-gauge on the condenser shows how much the pressure in the condenser falls below the atmospheric pressure. Residual steam, air, and water are extracted from the condenser by the air-pump, and discharged into the hot well, from which the boiler is supplied with water. The surplus is discharged into a pond to cool, in land engines; in marine engines it is delivered into the sea.

The cross-head of the piston-rod is guided by a parallel motion, to move in a straight line; except in trunk-engine, guided by the stuffing-box; oscillating-engines, in which the cylinders oscillate on trunnions, and the piston-rod is connected direct to the crank; Mr Hunt's Z-crank-engines, disc-engines, and rotatory-engines. The reciprocations of the piston are either transmitted through a beam, and a connecting-rod, to the crank, and the crank-shaft, for double-acting rotative-engines; or the beam may be dispensed with, and the piston-rod coupled to the connecting-rod, forming a direct-action engine. The fly-wheel on the crank or main shaft equalizes the motion. The mechanism to work the valves is called the valve-gearing, or valve-motion.

A pair of engines work together on the same shaft; they have a pair of cylinders, a pair of pistons, a pair of cranks. Engines are properly in pairs, when they are designed simply to equalize the action of the power.

CHAP. II.—ENUMERATION AND DEFINITION OF THE PARTS AND APPENDAGES OF BOILERS.

The shell of the boiler, or outer part, commonly of iron, is spherical, cylindrical, or flat in figure, or a combination of these forms. The steam-chest, or dome, on the upper side of the boiler, is a reservoir, whence the steam is supplied to the engine by the steam-pipe, which is fitted with a stop-valve. The furnace is the chamber for the combustion of the fuel; when within the shell, it is called a firebox. The flues, or conduits for the products, are either external or internal to the boiler; cylindrical metal flues are flue-tubes, and they are fixed at the ends into tube-plates. The man-hole is the entrance to the boiler for inspection, &c. Mud-holes are placed at or near the bottom for the discharge of sediment, &c. The water is supplied by the feed-apparatus; its level is indicated by a float. The boiler is emptied by the blowoff-cock; the surface of the water is cleared by the scum-cock. Brine-pumps may be used instead of blowoff-cocks to draw off the brine from marine boilers. Sediment-collectors receive the solid impurities floating in the water. Surplus steam escapes by the safety-valves. Vacuum-valves admit air into the boiler, when the pressure is less than that of the atmosphere, to prevent collapse. Fusible plugs are inserted in the boiler, over the fire, which melt and give vent to the steam when the pressure and temperature in the boiler become excessive and dangerous. The pressure is indicated by the pressure-gauge. The water-gauge shows the level of the water; it may be a glass-tube, or it may be gauge-cocks. The boiler is strengthened by stays, which may consist of rods, bolts, or gussets. The boiler is covered with clothing or cleading.

The fire-grate carries the fuel, and consists of fire-bars or grate-bars, supported by cross-bearers or bar-frames. The mouth-piece is the entrance to the furnace, and rests on the dead-plate; the fire-door is fitted to and hung by it. The heating-surface is the surface of the boiler exposed to the flame and hot gases from the furnace. The boiler-room, or internal capacity of the boiler, is divided into the water-room, occupied by water; and the steam-room, occupied by steam.

CHAP. III.—THE CRANK AND FLY-WHEEL.

The introductory historical account of the steam-engine—in particular the engine of Watt—prepares the way for a more systematic investigation of the modern steam-engine, with its manifold adaptations to the demands of engineering practice. The cylinder and piston, the crank and the main shaft, are the primary elements of steam-engines; the cylinder to contain the steam, the others to receive and transmit its energy. The following chapter is to be devoted to a consideration of the movements and action of the mechanical organs of the steam-engine, geometrically con- The Crank and Fly-wheel.—One of the most important appendages of the steam-engine is the crank, by means of which the force of steam, although at first producing motion only upwards and downwards in the straight line of the axis of the cylinder, is nevertheless rendered capable of exerting that force equally well in a circular direction. When the steam-engine is only employed for some such purpose as pumping up water no crank is necessary; but as some of the most usual and valuable applications of the steam-engine are those where it turns wheels of mills, of cotton machinery, of steam-vessels, or locomotive engines, the crank, by which this is accomplished in an admirable and simple manner, which has superseded every other plan of transmission, is entitled to very minute consideration. On examining in detail the action of the crank, it is to be observed that the force exerted by the steam is neither constant in direction nor in action. If the steam be admitted first below the piston, it forces it to the top of the cylinder; it is then cut off, preparatory to its being admitted above the piston; and in the interval it has no motive action. When admitted above the piston, it forces it to the bottom of the cylinder; and again there is a cessation in its action during the change in the position of the valve. Now it is evident that this recurring cessation of action between the alternating impulses would interrupt the continuous revolution in the wheel, but for the power of the wheel itself to continue the motion, by what is termed the momentum of its mass. When the steam, during a stroke of the machine, is acting most powerfully on the piston, part of its power is spent in accelerating the wheel; and when, at the end of the stroke, it ceases for a time to act, the wheel gives out the power which it had gained, and continues its motion until the next stroke gives it a fresh accession of power.

A wheel of this kind, when attached to an axle for equalizing motion, is termed a fly-wheel; and to obtain the full benefit of its equalizing power it is made of large diameter, that its rim may move rapidly, and it is made of great weight, being formed of iron, that it may acquire momentum to render the motion as uniform as possible.

From the circumstance that at one period the steam possesses no power of turning the crank, it has been imagined that some considerable loss of the power of the steam takes place during its transmission through the crank. This is a grave error. Figs. 81 to 84 represent the crank in different positions. In figs. 81, 84, the connecting rod and crank are in the same straight line, technically called the position "on the centre," or passing the line of centres, in which the action of the steam neither tends to turn the crank in the one direction nor the other. Again at x and n (figs. 82, 83), where the crank is acted upon at right angles by the connecting-rod, it is plain that the whole force transferred through the rod is acting to turn the crank; while in the intermediate positions there are two efforts, one acting on the centre of the crank, and another to turn it round. For examining the proportion of these forces to each other, we may use the two following diagrams. Fig. 85 represents the circle of the crank, the arrows showing the direction in which the crank-rod would require to act, in order that all its force should be undivided, and produce alone the single effect of causing revolution. Fig. 86 indicates the deviation which the actual motion of the crank exhibits from this hypothetical condition. The arrow a indicates the direction of the action of the connecting-rod, which at divisions 10 and 20 is acting only towards the centre of the circle, with no effect in producing revolution. At divisions 5 and 15 the whole effect takes place in producing revolution only. Through the first half of the circle the pressure of the rod acts wholly downwards, and through the latter half of the circle wholly upwards. The circumference of the circle being divided into twenty equal parts, the analysis of the force is given in the figure at several of these points. At the second division, a represents the direction of action of the crank-rod; b is parallel to the direction of the circumference (or tangent) of the circle at that point, while the line c is directed to the centre; a indicating the direction of the whole force of the connecting-rod, b representing the effect produced in the direction of the tangent to turn it round, and c the effect of the force of the connecting-rod acting to produce pressure on the centre of the crank; but as the centre of the crank is fixed and prevented from moving, none of the moving power of the crank is given out in producing motion towards the centre, but only in producing motion in the circumference. At the fourth division of the circumference, it may be observed that the effect of the connecting-rod is differently distributed. The whole force a is now more nearly in the direction of b, and c is comparatively small; showing that, as we approach the end of the first quarter's revolution, the force of the connecting-rod is producing much less pressure in the centre of the crank, and pressing in a higher proportion in the direction of the revolving effect, until at last the connecting-rod being at right angles to the crank, its whole pressure acting to turn round the crank, none of it is directed towards the centre. After passing the quadrant point 5, the crank-rod still presses downwards, as shown by the arrow a at point 7; but, of its two effective pressures, one represented by b still acts in turning round the crank, while another, represented by c, instead of acting towards the centre, as in the upper quadrant, now produces a pressure which would draw the crank away from the centre; but as the crank is fixed, none of the motive-power is employed in producing any motion of the crank away from its centre. Similar alternating effects are produced through the other quadrants; so that while the pressure of the steam, acting through the connecting-rod upon the extremity of the crank, is divided into two effects, one of these is prevented from expending the moving force of the engine by the fixedness of the crank-centre, and the whole motive-power is given out only at the circumference of the crank-circle in turning it round, but in a proportion of pressure that is continually varying from 0 to a maximum, and from a maximum to 0, through every successive quadrant of the circle. In order to simplify the inquiry, it proceeds, meantime, on the assumption that the pressure transmitted through the connecting-rod is constant. In general practice, with expansive-working and other circumstances, it varies, and the influence of this The Crank variation will be another subject for consideration. The variation of pressure circumferentially on the crank of a steam-engine may be conveniently represented by curves.

Let the circumference of the circle described by the crank be represented by the straight line \( \alpha x \) (fig. 87), and divided into any number of equal parts; let straight lines \( y_1 y_2 y_3 \), &c., be drawn to represent the amount of pressure converted into the direction of the motion of the crank, according to the line \( b \) in fig. 118, being the amounts represented in the line of figures, then the curved line \( A \), passing through the summit of all these lines, will represent the variation in the power of the crank at each instant of time, each ordinate \( y_1 y_2 y_3 \) being the pressure, and the area of the whole figure will represent the whole motive-power, having a maximum at \( y_1 \) and \( y_3 \), and a point of change of direction from pressure one way to pressure the opposite way at \( y_2 \).

Now, one method of equalizing the rotative pressure on the crank has been proposed, and is very generally adopted,—viz., to make two steam-engines act on the same axis by means of two cranks at right angles to each other, so that when the one ceases to exert force the other may be at its point of greatest force.

Thus, in figs. 88 to 91, two cranks are represented as coming from two cylinders, and attached to the same axis, so that when the one of them is at 0 the other is at 5, when the first is at 5 the second is at 10, and so on; so that while either is on the line of cessation of force, the other is at the point of maximum.

The joint effect of two such cranks may be represented by curves in the following way:—Let the circumference of each crank-circle be represented by the lines \( \alpha x \) and \( \alpha^2 x^2 \) (figs. 92, 93), as formerly, each semi-circumference being divided into eight parts, and let the pressure be calculated from a table of sines, where each will be found as the sine of the arch of the circumference to which it corresponds; the numbers thus obtained being set off on the base lines, the varying quantity of force, but without regarding the reversion of direction.

If, now, we place these curves together, as in fig. 94, their whole ordinates taken across from the one curve to the other, will represent the amount of the sum of the forces and its variation; and if we place all these ordinates from a fourth axis, we shall have represented by the new curve (fig. 95) the variations of the sum of the forces of the two cranks.

When a lever intervenes between the crank-rod and the piston-rod, new irregularities are introduced. The variation in the direction of the connecting-link, and in the position of the lever-ends from a straight line, introduces modifications of these effects of a serious nature, but not of a large amount. It is worthy the attention of practical men to consider these variations, and the manner in which they affect the uniformity of the pressure. They affect it by way of increase at the beginning and end of the stroke. By proper arrangements these very obliquities may be rendered very considerable improvements in the working of the engine. It should also be observed that the stroke of the piston and crank will not remain of the same length. The agency of the crank in transmitting a force parallel to the piston-rod has been represented by the curve of sines, as in fig. 96. But if we represent in a similar way the pressures produced by the obliquity of the connecting-rod, we shall find the form become that given in the following figures. Fig. 97 represents the variation of pressure with a connecting-rod of four times the length of the crank; fig. 98 with a connecting-rod of double the length of the crank; and fig. 99 with a connecting-rod equal to the length of the crank. It is obvious that with the shortening of the connecting-rod, the irregularity of the motion becomes very great. Two maxima rapidly succeed each other, and these are wide apart from the next pair. Thus two violent pressures succeed at a short interval, and a long pause intervenes when the force is very small.

Instead of using two cranks for applying the force of two steam-engines to the same axis of revolution, two engines have been used with their cylinders laid at right angles to each other, and having their connecting-rods applied to the same cranks.

**CHAP. IV.—THE CONNECTING-ROD AND PARALLEL MOTION**

**—THE GOVERNOR.**

In considering the agency of the crank in modifying the force and velocity of steam, so as to connect its direction, and distribute its force in the manner required to produce a rotative motion in the machinery from the original reciprocating motion of the piston in the cylinder, we have hitherto avoided the introduction of another important element, by which a further variation of force and of motion is produced. The connecting-rod is a rigid bar of metal, which conveys the motion of the piston from the piston-rod to the crank, either immediately or through the interposition of the lever or beam; and as the connecting-rod, in doing so, takes various directions, different from those either of the piston-rod or of the crank, there is an obliquity of pressure. produced at both extremities of the connecting-rod, which gives rise to a variation of force and of direction which must be practically provided for, and carefully appreciated in quantity, in so far as it may affect the ultimate operation of the machine.

There are two ways in which the motion of the piston-rod is most commonly transferred to the crank,—either immediately through the connecting-rod, as in fig. 100, or through the medium of the great lever, as in fig. 101, both ends of that lever describing circles around its middle fulcrum as a centre, and the head of the piston-rod being connected with the one end of the lever by means of an iron strap or connecting-link. From inspection of the figure it becomes plain that the connecting-rod or link is never, except at two points, in the same straight line with the piston-rod, so as to propagate its remodified force to the crank; but that in these oblique positions it would produce a lateral motion in the end of the piston-rod, which would not only be a waste of power in producing motion in a place where it is useless, but would have the effect of continually bending the piston-rod in opposite sides, so as either to break it or materially to impair its working. In the first of these figures, \( p \) being the direction of the piston-rod, \( p \) being that of the crank, the force in the piston-rod in the direction \( p \) becomes resolved into two parts \( p \) and \( p \), \( p \) being effective in the direction of the crank-rod, and \( p \) tending only to give lateral motion to the piston-rod, or else to bend it or break it across. And so also in the second figure there is a similar separation of pressure. To prevent these oblique pressures from wasting the power of the steam, by producing lateral, useless, or injurious motions, is the object of a series of contrivances called parallel motions or parallel guides. The most notable of these we owe to Mr Watt.

Let it be supposed that we desire to prevent the top of the piston-rod \( p \) (fig. 102) from being moved by the obliquity of the connecting-rod \( p \), either towards the right or the left, then it is accomplished in the following way:—A fixed support \( s \) is found on one side of the piston-rod, and another on the other \( s' \), at equal distances from it, and two parallel bars \( g \) and \( g' \) are placed between the piston-rod and these points, so that it may be steadied between them. These parallel bars are made so as to revolve freely round the points \( s, s' \), as centres, each of the ends \( g, g' \) describing the circles \( g, g', g', g' \); from which it is evident that, if these rods were directly attached to the piston-rod at \( g \) and \( g' \), they should have the effect of keeping the point \( p \) in the straight line \( o g g' p \). As these bars \( s g \) and \( s g' \) must describe circles round \( s \) and \( s' \), they would, in the positions \( s g^2, s' g' \), deviate altogether from the straight line of the piston-rod; but as the one will act nearly as much in the one direction as the other in the opposite, it occurred to Mr Watt that, by connecting their extremities with a link \( g g' \), and attaching the piston-rod, not to the ends of the guide-bars, but to the middle of this link, the point \( p \) might be prevented from deviating to any appreciable extent from the straight line. This is accordingly produced in a very simple way. The following figures, 103, 104, 105, show the effect of these links in various positions.

This elegant and simple contrivance is not, however, absolutely perfect. At the best, only a part of the line which it describes makes an approximation to a straight line, of scarcely sufficient length, and beyond which the stroke of the piston cannot be increased without being seriously deranged. Nor can this be remedied but by constructing the apparatus on a scale so large as to be highly objectionable. Thus, in the above arrangement, the point \( p \) is not kept perfectly in a straight line, but is, on the contrary, compelled to deviate from it so as to describe a looped curve. The nature of this deviation will become very evident if we suppose the parallel motion to be altogether detached from the piston-rod, and the motion of the parallel bar and link carried to its extreme, as in the following figures, 106, 107. A pencil being used to trace the motion of the middle point, \( p \) will describe, not a straight line, but a curve \( p x y \). When we carry the rods up to the position represented in fig. 106, where the bar \( g s \) comes into the straight line with the link \( g g \), the point \( p \) deviates from the straight line by turning to \( p' \); and this is reversed in the opposite extreme. In figure 107 the deviation is much greater when the link \( g g \) comes into the same line with the other bar \( g s \), and is also reversed in the position at the bottom of the figure. By the time the links have been re- turned to their primitive position, they have described the curve \( x p y \).

It is important to diminish this deviation, which increases more rapidly than the square of the length of the stroke. Having ascertained the greatest deviation at the end of the stroke, and also at \( \frac{1}{4} \)th part of the stroke from the middle, bring the centres \( s \) and \( s' \) nearer each other by a quantity equal to the deviation at the said eighth part, and the greatest deviation will now be reduced to less than one quarter of its former amount; the curve will now become a line of the sixth (eighth?) order.

The parallel motion of one point having thus been secured, it is easy to transfer it to any other point. This is most commonly done by a pointed parallelogram. Thus, to transfer it to a point in connection with \( s g \) prolonged to \( t \) (figs. 108 to 110), take a second link \( t q \), equal to \( g g \),


and a second bar, called the parallel bar \( g q \), equal to \( gt \), the corner \( q \) of the parallelogram will give a motion \( tg \) similar to \( p \). Figs. 111, 112, show the parallel motion transferred to a point still farther from the original point.

Another form of Mr Watt's invention consists in placing two bars in the same direction, with such a difference in their length as may afford the means of compensation. Suppose that the point \( p \) (fig. 113) is to be guided to move


in the straight line \( p g g' \); \( s s' \) are points on the same side of the required direction of motion, and \( s g, s' g' \) are the differential bars connected by a link \( g g' \), which is prolonged to \( p \). The dotted lines of the figures show the bars in different positions. The point \( p \) does not describe a straight line, but a curve, like figs. 106, 107. The motion of the point \( p \) may be transferred to a distance, as in the former instance, by a jointed parallelogram \( g p t q \) (fig. 114). All these parallel motions may be inverted, and, indeed, generally are inverted, in steamboat engines. For practical examples of them the reader may consult the plates.

All these motions, as well as the first, being imperfect, various plans have from time to time been adopted for remedying the evil. In American steam-engines, Watt's parallel motion has been to a great extent abandoned, because in them long strokes and long cranks are preferred; and because, in such cases, the deviations of the point \( p \)—that is to say, of the piston-rod from a straight line—would,


with Watt's method, become excessive. Watt and his assistants and followers were perfectly aware of this, and hence were led to construct beams, and connecting-rods, and parallel motions, of very great length, so as to diminish the evil as far as possible. This has, of course, the effect of rendering the whole engine both bulky and expensive, and is, therefore, in many cases inexpedient. The American engineers, therefore, use the sliding parallel motion; that is, they have substituted for the radius bars of the parallel motion of Mr Watt a sliding-bar or groove, in which the top of the piston-rod is guided. The head of the piston-rod \( p \) (figs. 115, 116) is enclosed between two flat surfaces, or between two parallel iron bars, which are kept in the vertical position by means of stiff framing; on these it slides, or, to diminish the friction, wheels may be added; but there are reasons why such wheels do not, in practice, work very well, and the plain slide is therefore preferred.

In fig. 115 we have represented this motion as applied to an engine of the simplest form; and in fig. 116, to a beam engine. In locomotives and some classes of marine engines, the guide-bars are universally employed.

Another species of parallel motion was, we think, first adopted in America; but it has also been used in this country. It is the engine with vibrating pillar. The pillar, which supports the beam or lever, instead of being fixed in an upright position, has a joint at the bottom, as will be seen in fig. 117, on which it, and the beam, and the crank-rod perform a joggling motion backwards and forwards during each stroke. The motion is of the following nature:—The points (fig. 118) is fixed; so is \( s \); \( s g \) and \( s' g' \) are moveable bars; \( p g \) is \( \frac{1}{2} \) of \( p g' \). The point \( g \) describes a circle round \( s \), and \( g' \) round \( s' \); hence \( p \) describes the curve \( p s p' \) of the sixth order. The oscillation of the moving mass of the engine in alternate directions, with a sudden jolt at the end of the stroke, renders this a bad engine when made. on a large scale; and it is obvious that the deviation of the piston-rod from the straight line is very great.

But the principle which furnishes the most perfect parallel motion is one which, although not new, we have never seen applied to practice. It is well known that the locus of the extremity of a straight line, the middle of which moves in a circle, the other end being confined to one straight line, is also another straight line at right angles to the former. Let a straight bar \(xy\) (fig. 119) be placed with one end \(y\) confined in an horizontal groove \(as\), and let a pin in the middle \(g\) be allowed to slide in circular groove \(ygx\); then the end \(x\) will always describe a straight line \(sx\) perpendicular to the first. Or it may be thus modified:—If the arc of a semi-circle have one of its extremities placed in a given straight line, while it moves along a given fixed point, the other extremity of the arc will describe another straight line at right angles to the former. Let a semi-circular round bar \(by\) (fig. 120) be allowed to slide through a fixed centre at \(s\), the one end \(y\) sliding in a groove, or along a bar \(sy\), then the point \(x\) will describe the perpendicular \(sx\), a perfect straight line.

To put this in practice in a form which shall not deviate widely from received forms of construction, is not difficult. The semi-circular groove and the semi-circular bar are not good constructive expedients. But if we take a radius bar \(sy\) (figs. 121, 122, 123), fixed at a centre \(s\), so that its end \(g\) describes a circle freely round it; and if we take a rigid bar \(py\) of double the length of \(sg\), and united to it at \(g\), then the middle of \(py\) being thus constrained to move in the circle round \(s\), we have only to permit \(y\) to slide freely in an horizontal groove, and the point \(p\) being carried up and down, will describe the straight line \(psp\). Fig. 124 shows the application of this motion to the simple engine, and fig. 125 to the beam engine.

Such is the mechanism which the obliquity of the direction between the connecting-rod, or link, renders necessary to prevent any of the motion, propagated through them, from being expended in producing oblique transverse motion in the top of the piston-rod. Still, however, the motion of the piston-rod is modified by transference in an oblique direction, and we have now to consider the nature of that modification. With an indefinitely long connecting-rod, of which the angularity is inconsiderable, the relation of the motion of the crank and the piston is represented by the annexed diagram (fig. 126), in which \(ac\) is the stroke of the piston, and \(abc\) the half revolution of the crank-pin simultaneously described. Let the path of the crank-pin be divided into equal parts at the points 1, 2, 3, 4, and draw verticals from the points of division to the line \(ac\); then, as the angular speed of the crank is uniform, and the divisions of the circular path \(abc\) are equal, the line \(ac\) will be divided by the perpendiculars already drawn into segments representing spaces described by the piston in equal times; and therefore, also, the varying average velocity of the piston in the same spaces. Whence it is obvious, that the speed of the piston, during one stroke, begins and ends at nothing at the extreme or dead points \(a, c\); that it accelerates towards \(b\), the position, at half-stroke, when it reaches a maximum, and that beyond this point it is retarded till it gains the end of its stroke. The two halves of the stroke are described in equal times; and in these halves the variation of the velocity of the piston are exact counterparts.

The obliquity of the connecting-rod destroys the symmetry here observed. In a stroke of the piston there are three cardinal points—the commencement, the middle, and the termination of the stroke. According to the preceding diagram, these three points are arrived at by the piston simultaneously with the horizontal and vertical positions of the crank. But the angularity of the connecting-rod at half-stroke of the piston virtually shortens its length, and the crank-pin is by as much short of its midway position. As the crank is presumed to move with a uniform angular velocity, it follows that the piston describes the two halves of its stroke with different average velocities, and in unequal times. In an engine, for example, with a stroke of 22 inches, and a connecting-rod 5 feet long, or six times the length of the crank, we find from the annexed diagram (fig. 127) of the relative positions of the piston and the crank, that, at half-stroke of the piston, the connecting-rod \(ab\) falls short of the vertical centre line of the crank by the amount \(or\), fully 1 inch. Dividing the stroke of the piston into three equal parts, the connecting-rod being in the relative positions, \(ad, ef\), the distances of the points \(d, f\), from the centre line, are \(os, ot\), respectively \(4\frac{1}{2}\) and \(2\frac{1}{2}\) inches. The corresponding angular positions of the crank are, for the half-stroke of the piston, \(6^\circ\) with the vertical, and for the one-third of the stroke respectively \(26^\circ\). and the complements of these $75^\circ$ and $64^\circ$ are the angular motions for the extreme thirds. The average speeds of the piston, therefore, in describing the successive thirds of its stroke in the direction $AB$, are inversely as $64, 41, 75$, or directly, as $6, 9, 5$, nearly; and the two halves of the whole stroke are described with average speeds inversely as $84$ to $96$, or directly, as $8$ to $7$. The shorter the connecting-rod, the greater is the irregularity so introduced into the motion of the piston. The general effect, therefore, of the connecting-rod on the motion of the piston is, that the piston anticipates the position which it would occupy if the connecting-rod were "indefinitely" long at any point throughout the whole of the front stroke, which is described towards the crank; and that throughout the back stroke the piston is in the same degree behind the position due to the crank alone.

The Governor.—The governor is an appendage to a steam-engine of much value in regulating all its applications to the production of uniform revolving motion. It is merely a modification of an apparatus similar to the pendulum, and by which Huyghens once attempted to regulate a time-keeper instead of the common pendulum. If we suppose the axis $AX$ (fig. 128) to revolve along with the ball $B$, hung by a thread from $x$, and also with two pieces of iron $XC$, $XC'$ bent so as to form cheeks, of a form called the cycloidal curve; then, when the string $AX$ comes in contact with those cheeks, the ball will perform each revolution in the same time as it would make two oscillations if merely swinging as a common pendulum; that is, if there be $39\frac{1}{2}$ inches from the centre of $N$ to $X$, the pendulum will revolve once in two seconds. If, however, the ball $B$ be suspended from $X$ by a straight bar, such as $AX$ in fig. 129, the line $BX$, in deviating from $AX$, will describe the circular arch $AN$ instead of a cycloid as formerly, and the time of oscillation will vary as the ball recedes from $A$, the revolutions being more rapid at $B$ than at $N$. If, in the position $B$, the perpendicular height $XC$ be $39\frac{1}{2}$ inches, then will the revolution be performed in two seconds, or at the rate of 30 per minute; while at $B$ they will be performed in less time, and between $B$ and $N$ more slowly. The height $CX$ for any required number of revolutions is equal to the length of a simple pendulum, which will give double the number of vibrations in the same time.

The regulation of the engine by the rate of the admission of steam to the cylinder is effected in the following manner:—The balls $NN$ (fig. 130), showing Watt's original governor, are suspended by rigid bars from the fixed point or centre $X$. These bars, being prolonged to $KK$, are joined by links $KY$ and $KY'$ to a moveable socket $Y$, which can slide up and down the axis. The straight lever $YP$ is acted on at one end by $Y$, and at the other it draws up or pushes down the handle of a circular disc $V$, so as either to close or open it to different degrees; this disc, called the throttle-valve, is placed within or in connection with the steam-pipe, that supplies the cylinder, so that if the engine should at any time move too slowly from having too much work to do, the balls will collapse, raise up $Y$, and open $V$ to the fullest extent, as at $4$ in the small sectional figure to the right, $3$ being the mean position; while, on the other hand, should the engine, from its work being taken off, go too quickly, the balls would fly off from the axis, bring down $Y$, and close the valve to $2$; or if it had happened, as by an accident, that the load was suddenly withdrawn, close the valve altogether, as at $1$ in the side figure. Instead, however, of the bars being always, as here, prolonged above $X$, they have frequently the points $KK$ placed below $X$, and the socket $Y$ below these again, as in several of the plates and woodcut illustrations.

SECTION II.—THE DISTRIBUTION AND BEHAVIOUR OF STEAM IN THE STEAM-ENGINE.

CHAP. I.—Introductory.

It is within the cylinder that the power of steam is exerted and its work consummated, and our next business is to study the behaviour of steam in the cylinder, and to discover the conditions upon which its good working properties may most efficiently be turned to advantage.

Steam, as it is commonly employed, is drawn directly from the boiler, and is in the condition of maximum density for the pressure and temperature, under which it is generated and delivered over to the cylinder. The slightest reduction of temperature, or abstraction of heat, does, therefore, unavoidably incur the condensation of a portion of the steam. It is this peculiarity of ordinary or saturated steam—its sensitiveness to cold and susceptibility of condensation—which defeats the ordinary expedients for increasing its efficiency; and it is essential that the nature and extent of this distinction should be understood and appreciated, in order to show how it may be prevented.

From the detailed description already accorded to Watt's engines, it may be gathered that steam operates in the cylinder in a twofold manner—first, it is admitted freely from the boiler into the cylinder, following the piston, and exerting pressure upon it throughout a greater or less portion of the whole stroke; second, the communication from the boiler to the cylinder being cut off, the volume of steam thus enclosed within the cylinder continues, though isolated, to press upon the piston and to follow it to the end of the stroke, or at least for so far as the steam is therein confined—that is to say, the steam is "worked expansively." The energy resident in the steam, and of which only a part is utilised in following the piston direct from the boiler, is further utilised in virtue of the inherent elasticity of the body of steam cut off into the cylinder. Strictly speaking, the whole process is one of expansive action, as the steam admitted direct from the boiler flows into the cylinder by its elastic or expansive force, the boiler constituting the fulcrum, or "point d'appui," of action. The process is continued on a more limited scale within the cylinder, after the steam is shut off, the steam continuing, in virtue of the same elastic force, its expansive action against the piston, the end of the cylinders constituting the fulcrum. The boiler, of course, exercises independently its function of generating fresh steam to replace that which is expelled.

Although, however, there is no generic distinction between the admission of steam direct from the boiler to the cylinder, and the subsequent continuation of duty under the name of expansive working—seeing that it is but elastic action in both cases—yet it is convenient in practice to draw a line of demarcation between the period of the stroke during which steam is admitted into the cylinder and the period during which it is simply expanded therein. The different conditions of the pressure or elastic force during these two periods is usually quite apparent in the indications of the internal pressure which, by means of suitable instruments, may readily be observed and registered, where proper proportions subsist in the mechanism employed for the distribution of the steam; but in certain conditions the distinction is obliterated, and the steady uniform pressure with which the entering steam ought to take its place in the cylinder merges frequently in the descending pressure characteristic of simply expanding steam. This descending pressure, indicated while yet the communication between the boiler and the cylinder is not finally cut off, is the result of what is expressively called "wire-drawing" the steam, the current of steam into the cylinder being partially intercepted at the "port" or entrance, by the nearly-closed valve or slide, and thus wire-drawn, "throttled," or spun into steam of reduced density and pressure. Thus, then, in consequence of what is technically considered an imperfection of mechanism, interrupting the flow of steam before it is fairly shut off, the process of "expansive working"—an economising process nominally initiated when the cylinder is fully closed,—may be anticipated at an earlier part of the stroke.

In describing the cycle of events known as the "distribution," or the ordering of the steam admitted to and subsequently discharged from the cylinder, it should be noted, by way of recapitulation—speaking of engines as ordinarily formed—that with the cylinder is associated the valve-chest or steam-chest, into which the steam from the boiler enters previously to its passing into the cylinder,—an ante-room where the steam waits in readiness to enter the cylinder when admitted. The form and position of the chest or chamber varies indefinitely with the design of the engine. From this chest three passages are formed, one leading to each end of the cylinder, and the third passage leading to the condenser or to the atmosphere, or otherwise for the exit of the steam from the cylinder. The orifices of these three passages or thoroughfares are known as ports, and are usually brought together and placed parallel, terminating in a flat surface on the side of the cylinder, on which the valve reciprocates. The function of the valve is to distribute the steam, for which purpose it is impressed with a simple reciprocating motion, by which it alternately covers and uncovers each port leading to the cylinder, admitting the steam from the chest, suppressing or cutting it off, and ultimately releasing it from the cylinder by opening a means of exit by the third port already mentioned. The reciprocating motion of the valve is derived from an eccentric, in the simpler forms of mechanism, fixed on the driving-axle, and revolving with it. The linear motion derived to the valve from the eccentric is, on a smaller scale, exactly similar to that of the piston in connection with the crank.

That the steam may gain admission to the interior of the cylinder at the commencement of each stroke, the eccentric is so set on the axle, in advance of the crank, as to have the valve moved sufficiently aside at that juncture, that the steam-port may be uncovered by a small amount known as "lead," at the beginning of the stroke. When the piston has described a portion of the stroke, the valve returns in obedience to the return of the eccentric, and closes the port, thereby shutting off the further supply of steam to the cylinder behind the advancing piston; and confining what has been admitted, during an additional portion of the stroke. As the valve continues in its retrograde motion, it uncovers the steam-port on the inside, while the piston is still some distance from the end of the stroke, and opens the way out of the cylinder for the steam within, from which accordingly it emerges, and rushes into the condenser or the atmosphere. This external communication continues open, not only to the end of the "steam-stroke" through which the course of the piston has been traced, but also during the greater part of the "return-stroke," while the steam from the valve-chest is busy on the other face of the piston. Shortly before the completion of the return-stroke, the valve, in the regular course of the motion prescribed for it by the eccentric, closes the port to the atmosphere, and, finally, at a very small distance from the end of the return-stroke, the port is again opened, and the valve obtains the necessary lead in timely preparation for the entrance of steam from the valve-chest, before the commencement of the next steam-stroke, and the development of the full steam-pressure on the piston for another cycle of duty.

The periodical and contemporaneous operations of the piston, the valve, and the steam, just described for one end of the cylinder and one face of the piston, take place independently for the other face of the piston; so that, two performances are proceeding together in one cylinder, and the engine is thence denominated double-acting. Four distinct events take place in consecutive order with respect to each end of the cylinder—first, the admission of the steam at or just before the beginning of the stroke; second, the suppression of the steam; third, the release or exhaust of the steam; and finally, the lock-up, or compression of the exhaust steam, prior to the opening of the port for admission. These four events together constitute the "distribution" for the cylinder, and their durations, measured in parts of the stroke, are the "periods of the distribution." By the aid of the indicator, which, as its name implies, is a sort of stethoscope for the observation of what transpires within the cylinder—a simple instrument for receiving and registering the tension of the steam—a minute and accurate picture of the operations within is transferred by pencil to paper, affording valuable and indeed indispensable data for the measurement of the power and efficiency of the steam in the cylinder. But before proceeding with this part of the inquiry, the movement and action of the slide-valve, in its relation to that of the piston, had better be explained by the process of geometrical illustration.

CHAP. II.—GEOMETRICAL ILLUSTRATION OF THE ACTION OF THE SLIDE-VALVE.

As the path of the crank-pin is represented by a circle, and the stroke of the piston by a straight line, equal to the diameter of that circle, so also the path of the eccentric is represented by a circle, and the travel of the slide-valve by a straight line, equal to the diameter of the eccentric circle—assuming, for the sake of illustration, that the valve is actuated in direct connection with the eccentric. If, then, two circles be described on a common centre c (fig. 131), for the crank-path and the eccentric-path respectively, their diameters ab, ac, are the stroke of the piston, and the travel of the valve. When the piston is at one end of the stroke at x', the valve is opening the port at a', and is just as much in advance of its middle position over the ports as is needed to stroke. The position of the eccentric, then, (represented by its own revolving-centre), must be at the point \(a'\), which is in advance of its position at half-throw, in the line \(DE\), by as much as the lap plus the lead. As the axle revolves, the valve is further opened by the retreating eccentric, till it falls into the line \(AB\), when the crank is getting on to half-throw. When the crank has attained to half-throw, in the position \(CD\), the eccentric is on its way returning, and the motion of the valve is reversed on the way to close the port. The port is actually closed sometime before the valve and the eccentric return to their midway position—the former over the ports, and the latter in the line \(CE\),—in virtue of the lap on the valve. Further, when the valve and the eccentric do arrive at their middle positions, half-travel and half-throw, the edge of the cavity of the valve coincides with the inner edge of the steam-port, and the opening of the port to the exhaust-passage, through the medium of the cavity, is forthwith established by the progressive motion of the valve.

All that has just been described of the operation of the eccentric on the valve is effected before the crank completes a half-revolution; that is, before the crank-pin arrives at the point \(N\), and, consequently, before the piston completes its stroke. The valve, indeed, makes another change before this; it opens the other port at \(B'\), for the other end of the cylinder, the lap having been completely withdrawn, and an additional movement for the lead effected, just as the stroke is completed.

The successive positions occupied by the valve during a revolution of the crank may be graphically represented, as in fig. 132, where \(AB\) and \(ab\), being the circles of the crank and the eccentric respectively, are divided into any equal number of parts, of which the points \(A, A'\), are respectively the positions of the crank and the eccentric at the beginning of the stroke; the other points, \(B, B', D, D'\), show the simultaneous positions of the crank and eccentric at intervals of one-fourth of a revolution. Draw the lines \(e, f, g, h, i, k\) parallel to the centre line \(AB\), spaced apart at intervals, equal to the exhaust and steam-ports and the intervening bridges; \(ef\) and \(ik\) being the steam-ports, and \(gh\) the exhaust-port. Place the valve on the perpendicular at \(A\), in the right position with respect to the parallels \(e, f, h, k\), for the commencement of the stroke, showing the requisite lead at \(v\), and set off its positions on the face of the diagram, according with the positions of the crank. The elliptic lines, traced so as to connect these positions, represent the linear motion of the valve relative to the ports and to the crank, and, with the aid of a little shading, they clearly show the successive periods and changes of the distribution, subject, of course, to correction for the angularity of the connecting-rod. The shaded space \(G\) shows the period of admission, terminating at \(g\); and the shaded space \(H\), the period of exhaustion, commencing at \(h\). The shaded space \(I\) shows the exhaustion for the alternate end of the cylinder, \(K\) the compression, and \(L\) the short period of pre-admission of steam for the following stroke.

On the same system, diagrams of motions may be constructed for any other proportions, or other species of valve, whether double or superposed valves, conical valves, moved by cams, or with conditions otherwise varied. The link-motion, as a variable expansion gear, operates by varying the travel of the valve; the extra expansion, and diminished period of admission, being affected by the shortening of the travel,—the result being precisely the same as if an eccentric of correspondingly smaller throw were substituted for an eccentric of greater throw.

CHAP. III.—THE INDICATOR-DIAGRAM—THE GENERAL BEHAVIOUR OF STEAM IN THE CYLINDER.

Proceeding to the investigation of the behaviour of steam in the cylinder, it may be noted that the action is fundamentally the same with condensing as with non-condensing engines; the difference being chiefly in degree. The receptacle into which the steam from the cylinder is discharged, is, in one case, the artificial atmosphere of the condenser, which causes an absolute pressure of 1 lb. per square inch, less or more; in the other case, the natural atmosphere, having a higher absolute pressure of 14½ lbs., or, in round numbers, 15 lbs. per square inch. The action of steam is developed, in its most simple form, in the non-condensing engine, in which the question of the vacuum has no part; and the writer will, therefore, open the inquiry with a summary of his experimental analysis of the behaviour of steam in non-condensing engines, chiefly of the locomotive class, first published in 1851, in Railway Machinery.

In illustration of the function and utility of the indicator, by means of which most of the writer's observations were conducted, examples of indicator-diagrams, obtained by the writer from one of the cylinders of a locomotive, are illustrated in fig. 133. The base line \(AB\) is the line of atmospheric pressure, and represents the stroke of the piston; and the rectangular space above it may be supposed to be the interior of the cylinder. The heavily lined figure is a diagram of the indicated action of the steam, when the piston moved in the cylinder at an average slow speed of Behaviour of Steam in the Cylinder during Admission.

In the flow of steam from the boiler to the cylinder it meets with hindrances to its passage which usually operate to cause a considerable reduction of pressure when it reaches the cylinder, even if all the passages be thrown wide open. The actual charge of steam transmitted through an irregular passage of considerable length, and of a given sectional area, is, in all cases, less than what can be passed through an aperture, of a very short length, as in a thin plate of the the commencement of the stroke, one-fifth of the length of the steam-port is sufficient. When the lead is excessive, the steam is admitted so readily as to be momentarily compressed, and to cause, in some cases, an unfavourable pulsatory action of the steam. The total absence of lead likewise occasions an unsteady pulsatory pressure in the cylinder. If lead is deficient, or wanting, the maximum pressure of steam in the cylinder is not attained until after a portion of the stroke is travelled by the piston.

CHAP. V.—THE BEHAVIOUR OF STEAM IN THE CYLINDER DURING EXPANSION.

When steam is admitted into the cylinder, while the latter is comparatively cold, or colder than the steam, a very sensible condensation of the steam takes place during admission, in the process of heating the cylinder to the temperature of the steam, which continues to a certain extent during the period of expansion. A portion of this heat, though but a small part, passes off and is lost; the remainder is retained by the cylinder until it is re-absorbed by the precipitated steam during the expansion of the remaining steam, if it be long enough continued—that is, until the temperature of the latter falls below that of the cylinder. This is a destructive process, occasioning an absolute loss of steam, and the amount of steam thus injuriously precipitated, and but partially revived, increases rapidly in proportion as the steam is earlier cut off and expansion extended. In the cylinders of ordinary steam-engines, the extra consumption and waste of steam devoted to the heating of the cylinder in the first part of the stroke is above 12 per cent. of the whole steam consumed for a period of admission of one-third of the stroke. In exposed locomotive cylinders, the loss has been proved to amount to nearly 40 per cent. of the whole steam consumed, when cut off at an eighth of the stroke.

This important species of loss is inseparable from the attempt to work steam expansively where there is no provision for the heating of the cylinder, and maintaining it at a suitably high temperature, equal, at least, to the initial temperature of the steam. The magnitude of the loss is so great as to defeat all such attempts at economy of fuel and steam by expansive-working, and it affords a sufficient explanation of the fact, in engineering practice, that expansive-working has been found to be expensive working, and that, in many cases, an absolutely greater quantity of fuel has been consumed in extended expansive-working, while less power is actually developed.

With respect to the ratio of pressure to expansion of steam in cylinders, observed in ordinary practice, it may be sufficient to remark in this place, that the quantity or weight of steam in the cylinder is the same throughout the process of expansion, estimated in terms of the pressure and the volume of the steam, as saturated at different points of the stroke, when the steam is dry, and the temperature of the cylinder is properly maintained; and that, consequently, the pressure of expanding steam in a cylinder, under such circumstances, may be determined with sufficient accuracy, for any degree of expansion, in terms of the ascertained density of saturated steam. On the contrary, in cylinders imperfectly heated, when the steam is partially precipitated during admission, and during the first part of the expansion, the expanding pressure at first declines more rapidly than would be due to the maintenance of a constant quantity of steam, and, afterwards less rapidly, rising above the expanding line of pressure, proper for a constant weight of steam, equal to that contained in the cylinder at the commencement of expansion. This want of conformity is exemplified in a diagram taken from an outside-cylinder locomotive, with a stroke of 24 inches, at a low speed (fig. 135), in which the dotted lines show the expansion-curve which... Behaviour would have been described with a constant weight of steam. This process of successive condensation and re-evaporation is distinctly indicated; for, no sooner is the steam cut off at \( A \), than condensation is made visible by the vertical sinking of the expansion-curve below the standard or normal curve, until the temperatures of the steam and the material of the cylinder become equal, when, as the pressure continues to fall, and the temperature of the steam with it, the curve rises and crosses the normal curve at \( C \), in virtue of a partial re-evaporation of the steam previously precipitated, caused by the cylinder itself, which, at first colder than the steam, and heated by it in the first stage of the expansion, is then relatively hotter, and partially restores the heat of which it had previously robbed the steam. The process of restoration of heat goes on to the end of the expansion, as farther proved by the increasing excess of the indicated above the normal pressure at the point \( D \), amounting to above 10 lb. per square inch at the point of exhaustion.

That the condensing power of an unprotected cylinder is something very considerable, is rendered very obvious by an indicator-diagram (fig. 136), taken from the same cylinder, in full gear, at a low speed, shortly after starting with a train. It shows that the pressure could not be maintained in the cylinder, as condensation in heating the cylinder proceeded faster than steam could be supplied through the opening of the port. Had the cylinder been hot, the pressure would have been fully maintained, according to the dotted line.

**CHAP. VI.—THE BEHAVIOUR OF STEAM IN THE CYLINDER DURING EXHAUSTION.**

In no part of the distribution is the advantage of time more apparent than during the period of exhaust. It is plain, by reference to figs. 133 and 134, that the steam does not discharge itself instantaneously from the cylinder at the point of release, as the piston, in all the diagrams, has visibly to go some distance before the pressure falls to a minimum. In fig. 133, at the lower speed, the piston moves 3½ inches from the point at which the steam is released to the point at which the pressure falls to the atmospheric line. At the higher speed, the steam only reaches the minimum pressure of 2 lb., when the piston has attained to the end of the stroke, through 5 inches of the cylinder. These are elementary proofs of the benefit of time for insuring a good exhaust.

As the velocity of steam, escaping uninterruptedly, would practically suffice to evacuate the cylinder in good time, to prevent the evil of back-pressure, there is no doubt that the back-pressure which does actually arise is owing to the circumstantial hindrance of mixed water, strictures, bends, and friction. The retarded motion of the piston towards the end of the stroke, in virtue of the action of the crank, is peculiarly favourable for the exhaustion of the steam, as it allows time for its escape before the piston returns upon it. At the higher speeds, however, the escaping steam may be overtaken, and driven before the piston into the atmosphere, should its remaining elasticity prove insufficient, and then an opposing back-pressure is established. The wider the lead for the exhaust, the less is the back-pressure, on account of the increased facility for escape. For the usual speeds of pistons of stationary engines, about 300 feet per minute, the back-pressure is practically unimportant, if the cylinder be properly heated and the steam dry. On the contrary, the back-pressure is very great when the steam is condensed within the cylinder, or if it be loaded with water by priming. Fig. 137 shows, in contrast, the evil of condensation in a cylinder, in causing back-pressure. In the one case, the steam admitted at 80 lb. pressure above the atmosphere is cut off at one-sixth of the stroke, and exhausted at half-stroke, yet it is so loaded with water, when discharged, that it incurs a back-pressure of 12 lb. per inch in being expelled by the piston, moving at an average speed of 430 feet per minute. In the other case, illustrated in the same figure, the steam was admitted in a much greater volume, though half the stroke, at a speed of piston of 580 feet per minute, whilst the exhaust-pressure fell to about 2½ lb. per square inch. The cylinder had been previously heated by hard work, the steam was comparatively dry, and the opposing pressure was, consequently, almost entirely removed.

The ordinary effect of the priming of muddy water from a locomotive boiler is illustrated by fig. 138, in which are shown indicator-diagrams, taken from the cylinder immediately previous and subsequent to blowing off the boiler when the water had been unusually impure, at the same speed of piston, about 600 feet per minute. The full line was described before, and the dot-line after the boiler was supplied with clean water; and other circumstances being the same, the back-pressure fell from 9 lb. to about 1½ lb. per square inch above the atmosphere. In some cases, the priming of water into the cylinder has been found to reduce the effective pressure more than a half.

The blast-pipe of locomotives should be larger in sectional area than the steam-passages in the cylinder, in order to discharge the steam with facility. The orifice of the blast-pipe should not be less, and it need not be greater than the smallest part of the passage from the cylinder; if smaller, it increases back-pressure.

**Summary of data as to Back-pressure.**—If the steam working in steam-engines could escape freely, without resistance, the back-pressure would be simply the pressure of the atmosphere, in non-condensing engines; and, in condensing engines, it would be the pressure corresponding to the temperature in the condenser—what Professor Rankine calls the "pressure of condensation." The mean back-pressure, however, always—sometimes considerably—exceeds the pressure of condensation. One cause of this, in condensing engines, is the presence of air mixed with the steam, which causes the pressure in the condenser, and also the back-pressure, to be greater than the pressure of condensation of the steam. The ordinary temperature in the condenser, in proper working order, is about 104° Fahr., for which the pressure is 1·06 lb. per square inch; whilst the actual pressure in the best condensers of ordinary engines may be scarcely ever less than 2 lb. on the square inch.

The principal cause, however, of increased back-pressure is resistance to the escape of the steam from the cylinder, amounting to from 1 to 3 lb. per inch greater than the pressure in the condenser. Professor Rankine thus summarises the ordinary results of observation on the back-pressure in condensing engines:

| Ratio of expansion from | Mean Back-pressure | |------------------------|-------------------| | 1 to 3 | 5 lb. per square inch | | Ditto | 4 to 7 | 4½ to 3½ lb. | | Ditto | 8 to 15 | 3½ to 3 lb. |

There is no doubt that, practically, in condensing engines, the back-pressure increases with the speed of the engine, and also with the density of the exhausted steam, and with Relative the size of the exhaust-ports. In non-condensing locomotive engines, Mr. D. K. Clark, in *Railway Machinery*, has found that the excess of back-pressure above the atmospheric pressure varies nearly as the square of the speed;—as the pressure of the exhaust-steam at the commencement of the exhaust;—and inversely as the square of the area of the orifice of the blast-pipe; that it is less, the greater the ratio of expansion; that it is less, the longer the time during which the exhaustion of the steam lasts; and that it is increased by the presence of liquid water amongst the steam.

**CHAP. VII.—RELATIVE PRESSURES OF STEAM IN THE BOILER, THE VALVE-CHEST, AND THE CYLINDER.**

As in the cylinder, so in the intermediate thoroughfares between that and the boiler, the movements of steam are affected by the conditions of dryness, as well as by the size and form of the passages. In well-protected cylinders, with dry steam, and ports \( \frac{1}{2} \)th of the area of the piston, the fall of pressure, or wire-drawing, in the cylinder-passages is about 15 per cent. of the pressure in the valve-chest, at a speed of piston of 600 feet per minute; and with a port \( \frac{1}{3} \)th the area of the piston, the fall of pressure does not exceed 10 per cent. at any speed. An example of the relative indicated pressures in the valve-chest and the cylinder of a locomotive is shown in fig. 139, when the former is represented to have oscillated above 5 lb. per square inch.

In imperfectly protected cylinders, when the steam is not dry, the fall of pressure varies from 20 to 40 per cent. The fall of pressure in passing through pipes less than \( \frac{1}{3} \)th of the piston in area, may be from one-third to one-fourth; in a pipe of \( \frac{1}{3} \)th the piston, the fall is inconsiderable. If the steam be highly dried, a pipe of ordinary length, \( \frac{1}{3} \)th the area of piston, is sufficient to transmit the steam undiminished in pressure.

The greatest useful opening of the regulator does not exceed \( \frac{1}{3} \)th the area of the piston. A greater opening is not found to add to the facility of transmission.

**SECTION III.—MEASUREMENT OF POWER—HORSE-POWER.**

**CHAP. I.—MEASUREMENT OF WORK DONE IN THE CYLINDER FROM THE INDICATOR-DIAGRAM—FRICTION-BRAKES.**

*Definition of Work.*—Work is an exertion of pressure through space. The unit by which quantities of work are measurable is the labour necessary to raise one pound weight through the height of one foot—that is, a foot-pound. The rate at which work is done is expressed in horse-power, and one horse-power is equivalent to work done at the rate of 33,000 lbs. raised through one foot in one minute of time, or 33,000 foot-pounds—that is, one horse-power is expressed by the performance of 33,000 units of work per minute. The rather odd number 33,000 was adopted by the fathers of steam-engines, for the measure of power, because it was actually the measure of the performance of an average working-horse, under favourable circumstances; and though it was a useful and convenient measure at the time of its adoption, when horses rivalled steam-engines, the number 33,000 is only now retained because it is inconvenient to alter what has continued for so long a period to be the standard measure of power.

The indicator-diagram represents the active pressure and the back-pressure per square inch exerted by the steam on one face of the piston. During one stroke of the piston, therefore, if the steam-pressure be supposed to be exerted uniformly throughout the stroke, and if there be no back-pressure, the work done would be simply expressed by the product of the whole pressure on the piston in lbs. into the stroke in feet. The diagram expressing such an uniform exertion of force would necessarily be rectangular, and the product of its length in feet by its weight in lbs., measured by scale, would be an expression of its area, and would express the work done for one stroke per square inch of piston. If an uniform back-pressure were indicated, it would form a deduction from the useful pressure; and the difference of height representing power and resistance on the diagram, would be the effective pressure, which, multiplied by the length, would express, as before, the useful area of the diagram, or the useful work done per square inch of piston.

But, in practice, the lines of positive-pressure and back-pressure on the diagram are not straight but curved. These conditions involve a preliminary process of reduction before the area of the diagram can be estimated, and the work for one stroke determined. Let the diagram be divided vertically into a sufficient number of parts by a series of parallels, drawn at equal distances across the figure transversely to the atmospheric line, representing the stroke of the piston; measure the mean effective pressure in each section by the scale, take the sum of the pressures thus found, and divide by the number of divisions of the diagram. The quotient is the effective mean pressure per unit of surface for the whole diagram. Multiply the effective mean pressure per unit of surface thus found by the area of the piston in the same units of surface, say square inches, and by the length of stroke in feet, and the resulting product expresses the whole effective work done upon the whole surface of the piston for one stroke. Reduce, in a similar manner, a second indicator-diagram for the other face of the piston, and take the mean of the two expressions of work done; or assume for a double-acting engine that the same quantity of work is done in the alternate stroke, multiply by the number of single strokes of the piston per minute, and divide by 33,000; the resulting quotient expresses the indicator horse-power of the engine.

If there be two or more cylinders to the engine, the power must be found in the same way for each cylinder, and the whole added into one sum, to express the entire indicator-power. But the simplest way in dealing with two or more cylinders, of equal diameters and strokes, working on a common main-shaft, is, first, to find from indicator-diagrams the average effective mean pressure on the whole surface of the piston, for all the cylinders; second, to multiply that by the length of the stroke, and by 2, and by the number of revolutions of the shaft per minute, and by the number of cylinders; third, to divide this product by 33,000. The quotient is the indicator horse-power.

On the contrary, should the cylinders be various in their functions, as in compound engines, in which a non-condensing and a condensing cylinder are combined, or should they be differently connected to the main-shaft, or have different diameters or lengths of stroke, it is better to find the power for each cylinder singly, and subsequently to sum the powers together, to find the whole. There is an advantage in this method of procedure, that the contribution of each cylinder, or class of cylinder, to the aggregate of power may be distinguished.

The friction-dynamometer measures the useful work done by a prime mover, by causing the whole of that work to be expended in overcoming the friction of a brake. The magnitude of the work done may be measured by a weight or a spring, as in Prony's dynamometer, in which the moving shaft or pulley is embraced by friction-blocks commanded by a lever weighted at the end. Horse-power of Engines.

Horse-power is the term employed to express the capacity, magnitude, or power of an engine. It sprung from the actual measure of horse-power originally adopted by Mr Watt—namely, the performance of work at the rate of 33,000 foot-pounds or units of work per hour; or the raising of 33,000 pounds, 1 foot high, per hour. From the desire to give over-measure, just as the hundredweight rose from the original 100 lb. weight to 112 lbs., as well as from the scanty means of exactly gauging the real power of engines, the horse-power rose to two, three, four, or five times that of actual horse-power; and thus, commercially, real or actual horse-power was altogether disregarded, and was replaced by what is called "nominal horse-power," the estimation of which is based simply upon the dimensions of the cylinder. "Nominal horse-power" is not, then, in any sense, horse-power, but is a commercial unit of capacity or power of performance, to fix the magnitude of the engine, and the price which is to be paid for it. But, even nominal power is not estimated by any uniform standard; individual manufacturers and others adopt different measures, and a uniform and universally recognised measure of power is a desideratum.

Different rules are applied to condensing and to non-condensing engines, the effect of which is to give a larger allowance of capacity to the former than to the latter. For condensing engines, the following are a few of the rules in use:

Boulton and Watt's Rule.—Assume the speed of the piston to be 128 feet per minute, multiplied by the cube root of the length of stroke in feet; and the mean effective pressure to be 7 lb. per square inch. Then, nominal horse-power = \( \frac{\sqrt[3]{\text{stroke in feet} \times \text{diameter}^2}}{60} \) inches. This rule is much in use in the south of England, and to some extent in Manchester.

Manchester Rule.—The common rule is to allow 23 square inches of piston per nominal horse-power; or, nominal horse-power = \( \frac{\text{area of piston in inches}}{23} \). Occasionally, 30 is used for the divisor.

Leeds Rule.—This rule is taken in terms of circular inches of piston, or the simple square of the diameter, allowing 30 per nominal horse-power; thus, nominal horse-power = \( \frac{\text{diameter}^2}{30} \) inches. This is, practically, almost identical with the Manchester rule, as 30 circular inches are equal to 24 square inches.

For non-condensing engines, the usual rules are as follow:

Manchester Rule.—Ten square inches of piston are allowed per nominal horse-power; or, nominal horse-power = \( \frac{\text{area of piston in inches}}{10} \).

Leeds Rule.—Sixteen circular inches are allowed; thus, nominal horse-power = \( \frac{\text{diameter}^2}{16} \) inches.

Glasgow Rule.—Square the diameter of piston in inches, and point off the unit figure. The result is the nominal horse-power. This is the same in form as the Manchester rule, as it is essentially a process of division by 10. The Leeds rule is more liberal than the other, as 16 circular inches are equal to about 13 square inches.

For compound engines, having both non-condensing and condensing cylinders, it is the custom in Leeds to throw in the small or non-condensing cylinder, taking no account of it, and to rule from the condensing cylinder only.

The elements of a sound and comprehensive rule for the power of steam-engines must comprise the mean speed of the piston, its area, and the mean effective pressure. The mean speed is based on the stroke and the number of revolutions in a given time; and as a given number of revolutions is usually required, the formula should contain these two data individually. The effective pressure, also, is the result of the positive and negative pressures, in the boiler on the one part, and the exhaust-pressure on the other part, with the additional element of expansive-working.

Actual horse-power, reckoned according to the principle with which Mr Watt originally started, of dealing with actual quantities, should constitute the basis of a revised code of rules for nominal power. There are two forms in which it may be reckoned: as the indicator horse-power, or the power communicated to the piston, undiminished by friction, measured by the indicator; or as the power delivered at the main or crank-shaft, measured by the brake, which is less than the indicator-power by as much as the friction of the mechanism of the engine employed in transmitting the power from the piston. For commercial purposes, the friction of the engine should be a consideration.

The indicator horse-power of a cylinder, reckoned in terms of the effective mean pressure in lbs. per square-inch on the piston, the diameter in inches, the stroke in feet, and the revolutions per minute, is computed thus:

\[ \text{Ind.-h.p.} = \frac{\text{pressure} \times \text{diam.}^2 \times 7854 \times \text{stroke} \times 2 \times \text{No. of turns}}{33,000} \]

\[ \text{Ind.-h.p.} = \frac{\text{pressure} \times \text{diam.}^2 \times \text{stroke} \times \text{No. of turns}}{21,000} \]

\[ \text{Ind.-h.p.} = \frac{\text{pressure} \times \text{area} \times \text{stroke} \times 2 \times \text{No. of turns}}{33,000} \]

CHAP. III.—Horse-power of Boilers.

The nominal horse-power of boilers is reckoned variously—from the lineal, and from the superficial, dimensions. The old waggon boiler was reckoned at the rate of 1 foot in length per nominal horse-power. For ordinary cylindrical boilers without internal flues, \( \frac{5}{6} \) to \( \frac{6}{6} \)—usually 6—square feet of horizontal area is allowed per nominal horse-power, the area being taken as the product of the diameter by the length in feet. For boilers having one or two internal fire-tubes, or "Cornish boilers," the diameter of the fire-tube is added to that of the shell to find the proper multiplier. Galloway allows only \( \frac{5}{6} \) to 4 square feet of horizontal area in his boilers, as they have upright water-tubes in the flues, available as heating surface.

Again, reckoning by superficies, 15 square feet of heating surface, and 1 square foot of fire-grate, are commonly allowed per nominal horse-power; the heating surface comprising the actual area of the shell exposed to heat, but only the upper half of the internal flues. For multitubular boilers, of the locomotive type, 16 to 20 square feet of the total heating or exposed surface, and \( \frac{1}{3} \), \( \frac{2}{3} \), or \( \frac{3}{3} \) square foot of fire-grate, are allowed per nominal horse-power; that is, more heating surface, because it is not considered so effective, and less grate used, because the draft is stronger than in Cornish boilers.

SECTION IV.—Classification of Steam-Engines, with Examples.

The modern steam-engine, comprising under this general designation, the engine proper, and the boiler and furnace, has been moulded into many varieties of form, according to the systems on which it has been constructed, and the various uses to which it has been applied. Irrespective of the uses to which engines are applied, they are distinguished into the two great classes of condensing engines, Stationary and non-condensing engines; the former exhausting the used steam into an artificial atmosphere of low pressure; the latter exhausting into the natural atmosphere of 147 lb. pressure per square inch. The former class is usually more efficient in the production of power from fuel than the latter; but the latter has fewer parts and is simpler in design.

The second classification is based on the uses to which steam-engines are applied, thus—First, stationary-engines, placed in permanent situations, for driving factory-machinery, pumping water, &c. Second, portable-engines, placed on wheels, or transportable, to be shifted to the scene of operations, to do the work of stationary-engines. Third, locomotive-engines, for propelling vehicles on land, chiefly on railways. Fourth, marine-engines, for propelling vessels on water. With the first, second, and third, it is the province of this article to deal; the fourth is reserved for a separate article.

CHAP. I.—STATIONARY BEAM-ENGINES.

Stationary beam-engines are usually made with condensers, and may be typified by the large engines erected by Messrs William Fairbairn and Sons, Manchester, to drive the machinery for preparing, spinning, and weaving alpaca fabrics, at the Saltaire Mills, near Bradford, the property of Mr Titus Salt. The engines are arranged in two pairs, to obtain the requisite uniformity of action, and are placed in large engine-houses on either side of the front entrance to the buildings; and they are supplied with steam from boilers placed in a boiler-house beneath the surface of the ground, and a short distance in front of the mills, according to the general plan and sectional elevations of the engine and boiler houses, Plate XVIII. The following is Mr Fairbairn's account of the Saltaire engines and boilers:

The Engines.—Plate XIX. contains a side elevation of one of these engines, giving a general view of the arrangement of the parts. The power generated in the cylinders c, and transmitted through the working beam n n, to the large spur-flywheel w, 24 feet in diameter, is taken direct from its circumference by the pinions r, r, which give it off at the required velocity to the shafting of the mill.

The working beam is supported on two massive columns e, 16 feet high, 14½ inches in least diameter, and 1¾ inch thick of metal; these columns are bolted down beneath the whole mass of masonry supporting the engine. The heavy entablature e bolted to each column, and to the columns of the adjoining engine, is firmly fixed in the walls of the engine-house on each side, and the spring-beams a, a, over this and at right angles with it, are similarly attached to the cross-beams, b, b. In this way an exceedingly strong and rigid support is secured for the main centre of the engine, which, resting in its pedestal a, has to sustain the principal strain of working. The spaces between the spring-beams and walls, excepting where the main beam vibrates, are filled with ornamental perforated iron-plates forming the beam-room, approached by the staircase f, for the purpose of oiling the centres, repairs, &c. The working-beam receives its motion from the piston-rod g, through the parallel motion h h, and transmits it by the connecting-rod x and crank o to the fly-wheel w.

The steam is brought from the boilers through a prolongation of the tunnel in which the smoke passes to the chimney, and enters the engine-house by the pipe p. Having thence been admitted to the cylinder through the valve-chests k k, it repasses, after it has completed its work, to the condenser n, through the eduction-pipe z, in the usual way. The condenser is supplied with cold water from the river Aire, by the pipe k k, which communicates with the cold water cistern j; the injector through which the water enters is in these engines 6 inches in diameter, but the supply of water may be diminished if necessary, by the injection gear hereafter described. Beside the condenser is the air-pump, for pumping out the water and the air which enters with the water into the condenser, and is worked by the rod ll from a part of the parallel motion. A pump to supply the cold-water cistern is worked by the rod n, and another pump is worked by the rod p p, by which part of the hot water from the condenser is pumped back again for the supply of the boilers, in proportion as the water in them is decreased by the evaporation into steam. The supply of steam to the engine is regulated by the governor x acting on the throttle-valve q, and thus the speed of the engine is kept uniform. A shaft s s, receiving motion from a bevel-wheel on the crank-shaft, works the equilibrium valves in the valve-chests k k, as will be described. rr is a flooring or stage by which access is gained to the cylinder covers for oiling and cleaning. The cylinder is 50 inches in internal diameter, and has 7 feet stroke; it stands on the circular cylinder bottom c', which is firmly bolted to the masonry by the long holding-down bolts rr.

"Plate XIX. contains also a plan of one engine-house, showing the relative arrangement of one pair of engines. The length of this house is 50 feet, and its breadth 24 feet. It will be seen that the two engines are combined so as to act in concert upon the same crank-shaft and fly-wheel, the cranks being placed at right angles to each other, that, when one engine is passing its top and bottom centres, and exerting least power, the other is in mid-stroke, and exerting its whole power upon the full leverage of the crank. In this way the action of the engines is equalised, and the motion rendered smoother than is possible with an independent engine, whilst, in case of accident to either of the pair, its fellow may be employed alone until the damage is made good.

"Plate XX. exhibits a half elevation and half section of the valve-chests, condensers, air-pumps, &c., showing the valves and the manner of working them. As before, c c are the cylinders, c' c' the cylinder-bottoms, k k the upper, and k' k' the lower valve-chests, fixed right over the cylinder ports, and communicating by the side-pipes t t. D d steam-pipe, h h condensers, l l air-pumps; m hot-well, into which the air-pump lifts the water accumulating in the condenser; this water passes away by the overflow-pipe m; p p, p p feed-pipes for supplying the boilers, with an air-vessel p for equalising the pressure and preventing injury to the pipes from sudden shocks; u injection-cock and injector, the quantity of water admitted being regulated by the injection-cock, worked by the hand-wheel f, through the medium of the small shafts and bell-cranks n n.

"The valves in these engines are of a peculiar construction, being modifications of the double beat or equilibrium valve, invented by Mr Hornblower, and generally employed in the mining engines in Cornwall, where the high price of coal has led to that rigid economy for which its engineers have long been justly famous. We believe Mr Fairbairn was the first to introduce and advocate the present system of high steam worked expansively for factory engines, which he accomplished (not without considerable opposition on the part of those for whose benefit it was designed) by perseverance, and with the assistance of an active trade competition, driving the manufacturers to greater economy. It has now become general, much to the advantage of manufacturers, where the same amount of work is accomplished with half the quantity of fuel. Most of the appliances for using steam expansively in rotative engines are open to the objections—first, of wire-drawing the steam; second, of cutting it off too slowly; and third, of leaving too much space between the cut-off valve and the cylinder, whereby much steam is wasted without producing any. Stationary effect. To remedy these defects, Mr Fairbairn employs the shafts and expansion apparatus shown in the plates, which are applicable to all rotative engines working expansively, whether with high or low pressure-steam.

"The steam entering the upper steam-chest \( x \), through the stop-valve \( a \), has free access also to the lower steam-chest \( x \) through the side-pipe \( t \), whilst the exhaust steam has also clear access to the condenser through the other side-pipe \( t \). The steam is admitted to the cylinder from the valve-boxes by means of the valves \( x \) and \( x' \), and after having completed its work, it passes through the exhaust-valves \( y \) and \( y' \) to the condenser, these valves being opened and shut alternately at the right instant by an apparatus yet to be described. Each of the valves consists of two single conical valves, carefully secured together and accurately fitting their seats, the lower valve being slightly smaller than the upper. The steam is admitted on the upper and lower side of each of these pairs of valves, and presses in opposite directions, so that the downward pressure on the upper valve is neutralized by the upward pressure on the lower, excepting that a slight preponderance is given to the former in consequence of the difference of area in the valves, in order to aid in keeping the valves firmly pressed upon their seats when released by the cams. Hence they lift with the greatest ease, and expose any required opening for the admission of the steam.

"The mode of working these valves is very simple; a shaft (\( xz \), Plate XIX.) receives motion from the crank-shaft, and imparts it by the bevel-wheels \( bb \) to the horizontal-shaft \( cc \); this in turn gives motion to the valve-spindles \( dd \), which pass continuously through bearings in the valve-chests, and are supported on footsteps on the brackets \( ee \). Upon each of these spindles are fixed two discs \( gg \), carrying cams upon their upper surfaces so arranged as to lift and release each valve at the proper instant of time. This is effected by a direct and simple action; the height of the cam corresponds with the lift of the valve, its length with the duration of the lift, and its position on the cam ring, which revolves at the same rate as the crank-shaft regulates the instant of time in the course of each stroke at which the valve is opened. The action of the cams is transferred to the valves through the medium of friction pulleys \( k k k k \), fixed upon small cross-heads, which are guided in their upward and downward motion by the brass standards on which they work. In the case of the steam-valve, these pulleys are capable of adjustment by sliding along the cross-head, so as to bring them over any section of cam that may be required, and thus the steam may be cut off at \( \frac{1}{2} \), \( \frac{1}{3} \), \( \frac{1}{4} \), or any required portion of the stroke, the remainder being effected by expansion.

"The exhaust steam being a constant quantity requiring a full opening into the condenser, it is desirable to retain the exhaust-valve open during the whole length of the stroke. By the present arrangement, this is effected with a greater degree of certainty than by any other means hitherto proposed. The exhaust-valves rise suddenly upon the short incline planes of the cams, and having allowed time for the escape of the steam through a wide passage to the condenser, they fall with equal celerity by their own weight, and thus a more complete vacuum is formed under the piston than is perhaps possible to obtain by any other process.

"The stop-valve \( a \) is a simple conical valve, worked by a lever and hand-wheel \( z \), fixed by a bracket to the side of the steam-chest, and is chiefly used for shutting off the steam from the engine.

"The indicator-diagrams (figs. 140 and 141) were taken from these engines on May 4, 1859. The engines were then working at 25 revolutions per minute, and one pair with part of the load off. The maximum vacuum varies from 130 to 132 lbs. per square inch below the atmosphere; the average vacuum is 12½ lbs. per square inch. Diameter of cylinder, 50 inches; area, 1963·50 square inches; speed of piston, 350 feet per minute.

"From these diagrams we get—

| Engine A. | Mean pressure of steam | =7·1684 lb. | | --- | --- | --- | | | Deduct for friction, air-pump, &c. | =2·0000 " | | | Effective pressure | =5·1684 " | | | Actual horses power | =107·63 |

| Engine B. | Mean pressure of steam | =7·3646 lb. | | --- | --- | --- | | | Deduct for friction | =2·0000 " | | | Horses power | =111·46 |

| Engine C. | Mean pressure of steam | =13·301 lb. | | --- | --- | --- | | | Deduct for friction | =2·000 " | | | Horses power | =235·34 |

| Engine D. | Mean pressure of steam | =12·946 lb. | | --- | --- | --- | | | Deduct for friction | =2·000 " | | | Horses power | =227·95 |

"With a higher pressure of steam, however, or a shorter expansion, these engines will work to a considerably higher power."

"The Saltaire Boilers.—Plate XVIII. contains a general plan of the boiler-house and boilers at Saltaire; and Plate XX. contains enlarged sectional views of one of the boilers, \( c' c' \), are the boilers, ten in number. They are 24 feet long and 7 feet in diameter, and are of the description known as multitubular. At \( b' \) the upper half of one boiler has been removed so as to expose the flues in the interior, in which the heat is generated; and at \( a' \) the whole boiler has been removed to expose the arrangement of the flues beneath, and the direction of the currents of heated gases as they pass from the furnace to the chimney. The fire generated in the fire-tubes \( d, d' \) passes into the mixing-chamber \( b' \), where the air is thoroughly mixed with the flame and smoke, to ensure as perfect a combustion as possible, and thence the gaseous products pass through 109 small tubes, 3 inches in diameter, and descend at \( b', b' \), into the brickwork flues beneath the boilers; thence following the direction of the arrows, they return at \( c' \), and enter at \( d' \) the great tunnel \( x' x' \), through which they are conveyed to a chimney 240 feet high, at some distance on one side of the mill.

"In the tunnel is placed the large steam-pipe \( s' s' s' \), which communicates with all the boilers by means of the pipes \( e', e', e' \); by this arrangement the hot gases round the pipe not only prevent the radiation of heat and condensation of the steam, but in fact in a slight degree superheat it, and render it more effective in working. The steam then passes to the engines by the pipes \( t', t', t' \). To supply the waste of water in the boilers caused by evaporation, they are supplied from the engines by means of the feed-pumps, from which the water is conveyed by the feed-pipe \( q' q' \) to the tunnel \( r' r' \), in which it is heated by the waste gases in an apparatus known as Green's fuel economiser. This apparatus consists of a series of upright tubes intro- Steam-Engine.

Stationary beam engines duced in the tunnel between the boilers and the chimney, through which the water is made to pass on its way to the boilers, and is there heated to above the boiling point. The practical difficulty in the employment of this apparatus was the formation of a coating of soot upon the pipes, which effectually prevented the absorption of heat by the pipes and the water. To obviate this, Mr Green employs a number of rings or scrapers encircling the pipes, and kept in motion by means of chains and pulleys, driven by a shaft from the engine; these scrapers traverse the whole length of the pipes, and thus prevent any accumulation of soot upon their surfaces. It has been found that, when the waste gases from a boiler escape at a temperature of only 400° to 500°, the feed water can be heated to an average of 225°, and the temperature of the gases on leaving the pipes is reduced to about 250°. To produce this effect, 10 square feet of heating surface in the pipes are required for each horse-power. In this way a considerable economy of fuel is said to be effected, amounting to 17 to 25 per cent. After passing through this apparatus, the heated feed-water re-enters the boiler-house by the pipe r' r', and is thence distributed to the boilers. The safety-valves shown at f', f', f', consist, for each of these boilers, of two valves, each of 12 inches area, loaded with fixed weights to the maximum working pressure of the boiler, and of a third valve of five inches area, attached to a spring-balance, so that any required pressure may be secured by its adjustment. These valves are fixed upon a common valve-seating; m', m', m', are the man-holes by which access is gained when requisite to the interior of the boiler, for the purpose of examination, cleaning, &c. At p' there is a self-acting elevator for raising the ashes and depositing them at once in railway waggons above. The coal is brought by railway and dropped at once from the waggons into the boiler-house at e', e'.

Referring to the sectional views of the boiler and boiler-house, a', a', are the fire-tubes and fire-grates, and n', n', the bridges; b' the mixing-chamber, in which three vertical tubes p', p', p', will be noticed, which strengthen the flue in its weakest part, viz., where it is elliptical in form, and aid in absorbing the heat by increasing the heating surface; b' the descending flue; d' d' the entrance to the common flue or tunnel t'; o' e' the steam-pipe, with shut-off valve at y', entering the common steam-pipe s' in the tunnel; r' the feed-pipe, with valve for regulating the supply to each boiler, communicating with the interior by the pipe x'; v', v', v', stays or gussets for strengthening the ends of the boiler; f' safety-valves, with levers, weights, and spring-balance g'; h' the glass water-gauge; i' cock for drawing off the water when requisite; and k' additional man-hole for cleaning the boiler beneath the flues.

The principal dimensions of the boiler are as follows:— Shell, 7 feet diameter, 24 feet long. The length is divided thus—Fire-tubes, 9 feet long; mixing-chamber, 8 feet; flue-tubes, 7 feet; total, 24 feet. The fire-tubes are 2½ feet in diameter; grater, 2½ feet by 6½ feet long; mixing-chamber, 3 feet 10 inches diameter; 109 flue-tubes, 3 inches diameter.

The heating surface in each of these boilers may be stated as follows:

| Area of fire-tubes | 135 square feet | |-------------------|----------------| | Area of mixing-chamber | 102 " | | Area of vertical tubes | 28 " | | Area of 3-inch tubes | 550 " | | Total | 815 " | | Area of fire-grate (16½ × 2) | 324 " |

"The chimney is 250 feet high above the level of the ground. It is built in the style of an Italian campanile, and consists of an outer ornamental stone casing, and an inner and perfectly distinct parallel brick flue, the space between the flue and the casing being free for the expansion and contraction of the flue. The flue is 6½ feet square, having 42½ square feet of sectional area, to the summit. The external casing is 20 feet square outside at the base, and tapers to 10 feet square at the summit."

The following are additional particulars and illustrations of the Saltaire engines and boilers, with reference also to the foregoing detail illustrations:—The cylinders (fig. 143) are 50 inches diameter, with 7 feet stroke, of metal, 1½ inches thick, on a pedestal of the same thickness, bolted down to the foundation. The ports are 20 inches wide by 6 inches deep, with a sectional area ¼ of the area of the piston. The equilibrium valves have the upper disc 12 inches diameter, area 113 square inches; the lower disc is 10½ inches diameter, with 86½ square inches area; difference of areas, 26½ square inches. Lift of steam-valves, 1½ inches; of exhaust-valves, 1½ inches; steam-branches, 12½ inches diameter; exhaust-pipe, 13 inches; stop-valve, 12½ inches diameter. Condenser, 40 inches diameter; air-pump, 33½ inches, by 3½ feet stroke; foot-valve, 30 inches by 6½ inches, 195 square inches, or ¼ of the sectional area of the air-pump; discharge-valve, 31 inches by 7 inches; injector, 7 inches diameter. The pistons and air-pump buckets (figs. 145 and 146) are packed with Goodfellow's notched V rings, in one piece. The beam is 21 feet 6 inches long between end centres, or over three times the length of the stroke; it is 3½ feet deep at the middle, or ¼ of the length, and the web and the rim are 2½ inches thick; the main-centre (fig. 147) is 12 inches diameter in the beam, and 9 inches at the bearings; the connecting-rod (fig. 148) is about 20 feet long, of cast-iron, of a cruciform section in the body, measuring 22 inches across, and 1¾ inches thick at the edges. The spur-flywheel is 24 feet 5 inches in diameter, with 230 cogs in the rim, 14 inches broad, 4 inches pitch; Stationary the rim is in 10 segments, and has a sectional area of 200 square inches. In fig. 149, the cover of the centre is removed in order to show how the arm of the wheel is inserted and bolted into the recess prepared for it.

The boilers are made of best Staffordshire iron-plates, \( \frac{3}{8} \) inch thick for the shell, and \( \frac{1}{8} \) inch for the ends; \( \frac{1}{8} \) inch Lownmoor plates for the flues, except for the furnaces, or upper part of the fire-tubes. The pressure varies from 25 lb. to 30 lb. per square inch, according to the load; it is calculated that the pressure may safely amount to 40 lb. Eight boilers are in steam at one time to supply the mill and the engines when working with their full load; but nine boilers, and sometimes all the ten, are at work when inferior coal is used. In addition to the large engines of 400 nominal horse-power, there are two 25 horse-power blowing-engines, also worked by the boilers, to heat the mills. Each boiler is estimated at 50 nominal horse-power, allowing about \( \frac{1}{2} \) square foot of fire-grate surface per horse-power, and is stated to be sufficient to supply steam for 150 indicator horse-power.

The consumption of fuel averages from 3 lb. to 3\(\frac{1}{2}\) lb. of good coal per indicator horse-power per hour, and 4\(\frac{1}{2}\) to 4\(\frac{3}{4}\) lb. of slack per indicator horse-power.

The total weight of cast-iron in each pair of engines, exclusive of steam-pipes, is 157 tons, 7 cwt., 1 qr.; of wrought-iron, 11 tons, 8 cwt., 3 qrs.; of brass, 2 tons, 7 cwt., 1 qr.; total weight of metal in one pair of engines, exclusive of connections, 171 tons, 3 cwt., 1 qr.; in two pairs, 342 tons, 6 cwt., 2 qrs. The weight of one boiler complete is 10\(\frac{1}{2}\) tons; of ten boilers, 105 tons. The gross weight of metal in all the four engines and ten boilers, exclusive of connections, is 447 tons, 6 cwt., 2 qrs.

The peculiar design of non-condensing beam-engines (fig. 152) is the production of the Corliss Steam-Engine Company, Providence, in the United States. The beam is uncommonly deep and rigid; the cylinder is coupled to the beam, not by the usual parallel motion, but by a connecting-rod, from the cross-head of the piston-rod, which is guided by an upright slide. The beam is supported by a cast-iron standard, stayed diagonally by wrought-iron rods. The engine is worked by Corliss's valve-gear. On this system, the steam-passages are reduced to the shortest length prac- ticable, and a valve is applied to each passage, two for the admission of steam, and two separate passages for exhaust; four in all. The gearing is designed to work the steam with various degrees of expansion, and is operated on by the governor, so that the speed is regulated, not by wire-drawing, as is usual, but by varying the degree of expansion, admitting steam of full pressure into the cylinder before cutting off. The valves are segmental, and rotative-reciprocating in their action; they are worked from a "wrist-plate," shown in the figure, deriving its motion from the eccentric. They are easy to work, and are operated on by the governor, without any resistance to the latter, and thus a steady equable motion of the engine is insured. The governor operates through an ingenious contrivance for disengaging and re-engaging the valve-gear, by means of an inclined-plane motion. The indicator-diagrams produced by this motion are good, and the practical results of its working are excellent.

CHAP. II.—MODIFIED ARRANGEMENTS OF BEAM-ENGINES.

The non-condensing engines made by Messrs Mather, Dixon, and Co., erected to work the inclined plane on the Liverpool and Manchester Railway, at Liverpool Station, Plate XXI., shows how, by simple inversion of the engine, the power is transmitted from the cylinder through the beam, at once to the main-shaft under ground, which carries the large grooved pulley to work the rope. There is a pair of engines to work the main-shaft, connected at right angles. The centres of the beams LL rest at a level of 3 feet above the floor of the engine. A is the cylinder, in section in fig. 2; L P is a side-rod from the cylinder cross-head to the lever, and the connecting-rod L K descends to the crank-shaft X. SS the steam-pipes, DD the valves, and Z Z the valve-gear. The foundations are solid red-sandstone rock, in which excavations are made for the shafts and ropes.

Another arrangement of beam-engine, with low-hung beams, Plate XXII., was erected by Messrs Wm. Fairbairn, in a mill at Staley Bridge. The whole of the pair of engines stands on an independent metal sole-plate, and is similar in form to the well-known type of marine engines. The beam LL is, as it were, split in two, one of the halves being placed on each side of the engine, but united at the middle by a large guide-pan, or main-centre, and at the ends by cross-heads, side-rods, and connecting-rods, to the piston and the crank. A the cylinders; B the cross-heads; K the crank; LL the beams; W the spur-flywheel, driving the mill-shaft Y through the smaller spur-wheel G; M the parallel-motion; Z Z the motion for actuating the valve-gear; D a counterpoise to the valves; W the governor; H the cross-head to work the air-pump.

CHAP. III.—COMPOUND BEAM-ENGINES.

Mr Wm. M'Naught, of Manchester, introduced an important addition to the ordinary beam-condensing engine, in the application of an additional cylinder, non-condensing, to co-operate with the condensing cylinder, under the same beam, but on the other side of the main-centre. This system is illustrated by the diagram (fig. 153) of a pair of engines, erected at the Gutta Percha Company's Works, City Road, London. A, A, are the ordinary condensing cylinders at one end of the beam, 34 inches diameter, 6-feet stroke; B, B, are the additional, and smaller, non-condensing cylinders, 31 inches diameter, 3-feet stroke, placed under and connected with the other half of the beam, near the crank-end. "High-pressure" steam is admitted from the boiler to the small cylinders B, thence to the large cylinders A, thence to the condenser C, by means of suitably-arranged pipes and gearing; the steam is admitted to the former at a pressure of from 40 lb. to 50 lb. above the atmosphere, and is cut off by means of an ordinary lap-valve, at Compound Beam-Engines. from 1/4 to 1/8, but usually at half-stroke; it is expanded to the end of the stroke, and then is exhausted into the large cylinder, which commences its stroke at the termination of the stroke of the first cylinder, and the two cylinders perform respectively their steam-stroke and return-stroke at the same time, the steam continuously expanding from the smaller into the larger, and evacuating the former precisely as if it was exhausted into the atmosphere. At the end of the steam-stroke of the larger cylinder, the steam is exhausted in the usual way into the condenser; and at the end of the return-stroke of the smaller cylinder, the steam from the boiler is admitted into it for the next steam-stroke. The indicator-diagrams (fig. 154) show the conjoint working of the coupled cylinders of one of Mr M'Naught's engines, erected at a grain-mill at Drogheda, in which the high-pressure cylinder is 31½ inches diameter, with 37-inch stroke; and the low-pressure cylinder is 34 inches diameter, with 74-inch stroke, making 20 revolutions per minute. The nearness of the exhaust-line of the former to the expansion-line of the latter, proves the proper action of the system, and that there is no material loss of pressure by the passage of the steam from the one cylinder to the other. The steam is admitted at 33 lb. sensible pressure, and there is a vacuum of 11½ lb. The indicator horse-power obtained in the small cylinder is 79½ horse-power; and in the larger, 77 horse-power, nearly equal, making together 156¾ horse-power. The engine works day and night, and consumes 2½ lb. of Welsh slack per indicator horse-power per hour. In other cases of M'Naught's engines, a fall Compound of 2 lb. to 3 lb. pressure may be observed in the passage of the steam from the small to the large cylinder, arising from the probable condensation and friction in the passages. The usual consumption of fuel by this class of engine in mills is about 3 lb. per indicator horse-power per hour, which includes the necessary steaming of the mill; in one case it was only 2½ lb., and if ¼ lb. be deducted for other purposes, the usual consumption of M'Naught's engines would be at the rate of 2½ to 2 lb. per indicator horse-power per hour. With very inferior fuel, the gross consumption has amounted to 5·28 lb. per horse-power. The system has been applied to upwards of 400 engines, and, besides the economy of fuel, it effects a perfectly steady motion, by the equal distribution of the moving force at the two ends of the beam, and reduces the violent strain, or "lift," usual in single-cylinder engines, working expansively at the main-centre and the crank-shaft bearings.

Beam-engines are occasionally compounded by placing two cylinders side by side, under the same end of the beam, exhausting from the top of the small cylinder to the bottom of the large one, and vice versa. On this system the passages between the two cylinders are shorter than on M'Naught's; but there is an obvious disadvantage in the united action of the two cylinders on the beam and the bearings in the same direction.

Mr Charles Swift, Blackburn, constructs a single slide-valve, for exhausting from the high to the low-pressure cylinder, shown in detail in figs. 155, 156. The exhaust takes place close to the low-pressure cylinder, as shown in the fig. 156, which represents in plan a pair of cylinders, high-pressure and low-pressure, with separate beams and cranks on one shaft. Thus the full pressure of the steam exhausted from the high-pressure cylinder is obtained on the low-pressure piston.

**CHAP. IV.—DIRECT ACTION HORIZONTAL ENGINES.**

Non-condensing engines are now almost invariably constructed "direct-action," or without the intervention of a beam—the connecting-rod uniting the cross-head of the piston-rod to the crank-pin. They are made horizontal and vertical; but the former disposition is most common, and is adopted where space is not limited, and other arrangements permit, as the whole of the engine lies low, under easy inspection, and may be firmly placed on a solid cast-iron base, rested on a plain foundation. The non-condensing horizontal engine, made by Messrs Carrett, Marshall, and Co., Leeds, of 30 nominal horse-power, is shown in figs. 157 and 158. The cylinder is 21 inches in diameter, with 42-inch stroke, with short ports, and is fitted with double valves, to effect variable expansion, the expansion-valve being placed upon the back of the ordinary valve, and worked by a separate eccentric on the main-shaft, adjusted to cut off at various points of the stroke. There is also a feed-water heating apparatus placed over the cylinder, consisting of a flange of small tubes placed in an upright cylindrical vessel; the steam is exhausted through the tubes, and the feed-water is forced into the vessel around the tubes, by which it is heated from the steam, on its way to the boiler. The advantage of so heating the water on its way from the pump is, that it may be lifted cold by the pump from considerable depths, with more certainty than when hot. The sole-plate of the engine is 20 feet long, in one piece.

Another variety of direct-action horizontal engine, non-condensing, of 10 nominal horse-power, by Messrs Clayton, Shuttleworth, and Co., Lincoln, is represented by figs. 160 and 161. The cylinder is 10 inches diameter, 14-inches stroke, and the engine makes 110 revolutions per minute. The distinguishing feature of this engine is the apparatus for heating the feed-water, which consists of a large pipe underground, 9 feet long, 5 inches diameter, into which the steam is exhausted from the cylinder at one end, and from which it escapes at the other end. The cold water is pumped by the engine into a 2-inch tube, which traverses the heating-pipe from end to end, where it becomes heated on its way to the boiler. The engine is placed on a sole-piece, 8½ feet long.

Condensing engines of moderate power are frequently made horizontal direct-acting. They are distinguished in appearance from the non-condensing engine by the addition of the air-pump and condenser on the same base-plate. Two examples are illustrated in Plate XXIII. The first is an engine of 20 nominal horse-power, by Messrs Barrett, Exall, variable expansion, worked by two eccentrics, one to each; the expansion is varied by altering the position of the expansion-eccentric on the shaft, by means of a screw and hand-wheel moving the eccentric on a diagonal key fixed on the shaft. The piston-rod is extended through the bottom of the cylinder to force the air-pump rod, the air-pump being situated within the condenser, at the extreme left of the engine. The injection-water is discharged from the condenser through an overflow-pipe, and the injection-water rises to the condenser in virtue of the vacuum. The base-plate of the engine is 15 feet long.

An engine of this class was subjected to a ten hours' continuous trial by Messrs Barrett and Co., at their works. It had a 21-inch cylinder, 30-inch stroke, jacketed on the sides and ends; it cut off steam at 3 inches, or 9 per cent., or about \(\frac{1}{4}\)th of the stroke, besides a clearance of 1 inch of the stroke, with 50 lb. steam in the boiler. The boiler evaporated 9 lb. of water per lb. of coal, fed from the hot-well at a temperature of 82° Fahr., the cold water being supplied at 38° Fahr. The average speed of the engine was 60-03 revolutions, or 300 feet of piston per minute, and the power measured at the main-shaft, by the friction-dynamometer, was 40 horse-power. The indicator-diagrams (fig. 159) were taken from the two ends of the cylinder during the trial, and show an indicator-power... of 47 horse-power, from which it appears that the resistance of the engine consumed 7 horse-power, or 15 per cent. of the whole indicator-power. The fuel consumed was 3-06 lb. per dynamometric horse-power, or 2-6 lb. per indicator horse-power. The figure shows an average vacuum of nearly 12 lb. below the atmospheric line, whilst the vacuum in the condenser was 27 inches of mercury, or $12\frac{1}{4}$ lb. per square inch.

The compound-engine, by Messrs Carrett, Marshall, and Co., Leeds, Plate XXIII., comprises a variety of new and useful features. The high and low cylinders, 9\(\frac{1}{2}\) and 16 inches diameter respectively, with 20-inch stroke, are placed side by side on one base-plate, and work to cranks, diametrically opposed on the same shaft. Thus the pistons do not, as is usual in compounded twin-cylinders, work together on the same cross-head, making their forward and backward strokes together; but they alternate, and by this means the steam is exhausted by the medium of a single slide, straight across from the high-pressure to the same end of the low-pressure cylinder, through the shortest possible passage. Thus, also, the reciprocating mechanism balances itself, and a perfectly steady motion is obtained.

By the cross sections it is shown, that the valve-faces of the cylinder are at right angles to each other, connected by a single slide, and are at 45° with the horizontal plane of the cylinders. The air-pump, which is double-acting, and the condenser, are placed apart behind the low-pressure cylinder; the pump is within the condenser, and is worked by a prolongation of the low-pressure piston-rod. In this way, it is apparent, the weight of the heavier piston is carried in stuffing-boxes, and on slide-blocks, at both ends of the cylinder. The feed-water is heated in a simple, ingenious, and effective way, by injecting water from the hot well through a perforated pipe into the interior of the exhaust-pipe, from the low-pressure cylinder, where it acquires the temperature of the steam, and whence it is pumped into the boiler. The base-plate of the engine is about 10 feet long, and does not include the condenser.

Mr B. Goodfellow, of Hyde, constructs compound horizontal engines, working on cranks at right angles on the main-shaft, in which the high-pressure piston is worked at a higher speed than the low-pressure, by the expedient of giving it a longer stroke; and the air-pump and condenser are placed on the base-plate between the low-pressure cylinder and its crank, being worked by the continued piston-rod of the cylinder. The condenser and air-pump for horizontal engines are frequently placed beneath the base-plate, and worked by bell-cranks; but the arrangements illustrated are now generally preferred.

**CHAP. V.—DIRECT ACTION VERTICAL ENGINES.**

Where space is confined, and compactness an object, upright engines are serviceable. They are made variously; but the chief distinctions are—first, with the cylinder above and the crank below, near the ground; second, with cylinder below and the crank overhead. Of the first class, the engine (figs. 162 and 163) of 16 nominal horse-power, by Messrs Carrett, Marshall, and Co., Leeds, is a simple and effective example. The cylinder is 17 inches diameter, 2 feet stroke, and is supported by a strong frame directly on the cast-iron base, which carries also the crank-shaft; the valve-gear, feed-pump, and governors, also are directly worked; all the parts are easily accessible, and the whole rests on one foundation, 6 feet by about 5 feet, excepting the outer bearing of the crank-shaft. The stuffing-box projects into the cylinder, and thus prevents the leakage of water from it; and the lubrication of the piston-rod is effected by a receptacle in the gland for lubricating packing, distinct from the ordinary packing. In addition to the ordinary valve there is a separate expansion, or cut-off valve, on the back of it, worked by a separate eccentric, by means of which the steam may be advantageously cut off at any point, from \( \frac{1}{8} \)th to \( \frac{3}{8} \)ths of the stroke—the lead remaining constant, and the exhaust unvarying. With heated feed-water, engines of this class work with from 3 to 4 lb. of good coal per indicator horse-power. In some cases this kind of engine is made with a condenser.

The vertical engine, by Messrs Williamson Brothers, of Kendal (figs. 164 and 165), exemplifies the reverse arrangement, where the crank is overhead and the cylinder below. Here a stiffer and more solid erection is needed for resisting the swing and action of the crank, and it is made in the form of a hollow pillar, equally stiff in every direction, and combining strength with lightness. The strain is confined to one casting, and there are few separate parts. The engine is fitted with an equilibrium-valve, controlled by the governor, designed without arms or rings, to prevent sticking in any part; the governor can be disconnected instantly by a lever, which shuts off the steam at the same time. The same form of engine is applied to underlying crank-shafts, when the engine is required to stand independent, and to drive only by a belt from the fly-wheel, or from a pulley on the main shaft.

In situations where a medium level of crank-shaft is to be observed, the "steeple" form of engine is occasionally used, exemplified in fig. 166, as applied by Messrs Dunn and Co., Manchester, to the driving of sugar-mills. Here the cylinder is below, the cross-head and guide-bars at the top, and the crank-shaft is half-way up, worked by a connecting-rod depending from the cross-head. To effect this arrangement, the piston-rod must be united to a triangular frame, wide enough to clear the vibrations of the connecting-rod.

Oscillating engines, also, are occasionally applied vertically, so as to diminish the weight of the engine; and for small power they are useful and convenient. The Example (fig. 167), made by Mr Michie, of Nine Elms, London, has a 4½-inch cylinder, 12-inch stroke, and makes 100 revolutions per minute. The slide-valve is not worked by an eccentric, but derives its motion from a fixed stud in the framing, to which is jointed a link and double lever, passing to the spindle of the slide in the side of the cylinder; thus the oscillation of the cylinder gives the action to the valve. Engines in this style are made up to 12 horse-power, and they work well, with few repairs.

**CHAP. VI.—PUMPING-ENGINES.**

Pumping water from mines, or for the service of towns, or for other purposes, has, until of late, been almost universally performed by direct lifts, after the style introduced by Newcomen and Cawley, and adopted by Smeaton and by Watt (figs. 7 and 14). As the duty was simple—the lifting of a dead-weight—so was the engine, the steam cylinder at one end, the pump at the other, and the beam between to connect them, with no superfluous rotative or crank-motion. Lately, double-acting rotative pumping-engines, with crank and fly-wheel, have been introduced. They are more convenient and more compact than the direct-lift engines, but they are not so efficient. The "duty," as it is technically called, is inferior, owing to the greater internal friction of the rotative engine, and to the less nearly perfect adaptation of the action of the engine to the nature of the material to be dealt with and the work to be done. The term "duty" was first employed by Mr Watt to signify the net effect resulting from the consumption of a given quantity of coal; and was expressed by the product of the weight in pounds of water raised, by the height through which it was lifted in feet, divided by the weight in bushels of the fuel consumed in performing a given quantity of work. The duty of the best of Smeaton's engines was, in 1772, 9,450,000 foot-pounds per cwt. of coal. On the expiration of Watt's patent, about the year 1800, the highest duty of his engine amounted to 20,000,000, or more than double the former duty, which may represent the economic value of the improvement effected by Watt under his various patents. The reported duty of Cornish pumping-engines, by the consumption of 94 lbs. of coal, rose from an average of 19,500,000, and a maximum of 26,500,000 in 1813, to an average of 60,000,000 and a maximum of 96,000,000 in 1843. The average duty in 1856 was only 47,000,000, probably on account of inferior qualities of coals used.

It is necessary to bear in mind the distinction between the duty of a bushel of coals, 94 lb., and the duty of a cwt. of coals, 112 lb., both of which are employed as measures of fuel and of duty. In terms of 112 lb. of coal, the average duty of Cornish engines, in 1843, would be 71,500,000; and, in 1856, 56,000,000; and the maximum duty, in 1843, would be 114,500,000.

The functions of pumping-engines, for the supply of water to towns, &c., are chiefly distinguished from those of Cornish engines in the relatively shallow lifts, or depths from which the water is pumped, together with the height to which it is elevated, or the "head" of pressure against which the engine works. Thus Austin's engine, Fowey Consols, in Cornwall, in 1856, lifted the water through a height of 183 fathoms, or 1100 feet. In the London water-bearing district, the depth from the surface of the ground to the level of water does not exceed 100 to 120 feet. No water-works engines are employed to lift water from a greater depth than 15 or 16 feet below the pump; but at the East London Water-Works, the combined lift and head of pressure is about 140 feet. The extreme variety of duty of course involves a variety of proportions, chiefly in the ratio of the diameter of the cylinder to that of the "plunger-pole" of the pump. The Fowey Consols engine has an 80-inch cylinder, with 10 feet 4 inches stroke; the pump has 9 feet 3 inches stroke, drawing six lifts—that is to say, pumping by six successive stages: three of the plunger-poles are 15 inches, one 13½ inches, and two 12 inches diameter—the diameter of the plunger-poles being about ¼th that of the cylinder. On the contrary, the pole of the "Victoria" pumping-engine at the East London Water-Works is 50 inches diameter, with 100-inch cylinder, being one-half the diameter. The obvious ruling element is the "head" against which the engine is to work; the greater the head, the smaller is the pole in proportion to the cylinder, other circumstances being alike.

The duty of a pumping-engine may be variously defined. Preliminarily, there is the gross work done by the steam in the cylinder, or the "total load," measured by the indicator, comprising the whole work done of whatever kind, useful and useless—water lifted, friction of pump, and friction of the engine itself. The "steam-duty" is the duty performed within the pump, measured by an indicator, comprising the whole work done in lifting the water into the pumps and expelling it into the mains—the steam-duty being equal to the gross work of the steam, minus the friction of the engine. The pumping, or effective duty, is the duty realised in the mains; this is measured by the head of pressure outside the pump, measured by the guage, and is equal to the steam-duty, minus the resistance of the pump. The effective duty is that which is usually understood to be signified by the "duty" of an engine.

At the East London Water-Works there are four engines, particulars of the working of which have been kindly contributed by the engineer, Mr C. Greaves. The leading dimension is respectively as follows for the four engines:

| Engine | Diameter of Cylinder | |------------|----------------------| | Ajax | 72 inches | | Cornish | 90 | | Wickstead | 90 | | Victoria | 100 |

These engines are worked with 35 lb. steam in the boiler, and the initial pressure in the cylinder is 20 lb. to 25 lb. per square inch on the piston. The cylinders are steam-jacketed, and the steam is usually cut off at about ¼th of the stroke, and expanded during the remainder. The following are particulars of regular working results in 1857. The relative effective steam pressures per square inch on the piston, or the "total load," and the work done in the pump, reduced to the equivalent pressure on the steam-piston, being elements respectively of the total load, or gross work of the steam, and of the steam-duty, are as follow:

| Cylinder | Total Load in lbs. per Square Inch | Pumping Work Done in lbs. per Square Inch | Difference, or Friction of Engine in lbs. per Square Inch | |----------|-----------------------------------|------------------------------------------|----------------------------------------------------------| | 72-inch | 15-00 | 14-09 | 1-00 or 6-64 per cent. | | 80 | 14-38 | 12-73 | 1-65 or 11-47 per cent. | | 90 | 15-58 | 14-10 | 1-48 or 9-51 per cent. | | 100 | 16-58 | 15-01 | 1-57 or 9-47 per cent. | | Mean | 15-38 | 13-96 | 1-42 or 9-23 per cent. |

Showing a mean friction of engine equal to 1-42 lb. per square inch of piston, being 9½ per cent. of the gross work of the steam; and showing that the steam-duty is 90½ per cent. of the gross work of steam. The performance of three of these engines is as follows:

| Cylinder | Usual Length of Stroke | No. of Strokes per Minute | Duty per 112 lbs. of Fuel | Effective H.P. | |----------|------------------------|---------------------------|---------------------------|---------------| | 80-inch | 9 | 6-48 | 86,576,976 | 104 | | | | 8-08 | 78,675,829 | 135-4 | | | | 5-38 | 86,749,117 | 120-5 | | 90 | 10 | 7 | 93,595,265 | 133-9 | | | | 5-97 | 78,667,331 | 165-5 | | | | 5-05 | 79,848,876 | 180 | | Mean | | 5-95 | 84,018,899 | 140 |

Showing that, with a mean of 6 strokes per minute, 140 effective horse-power was realised, with a duty of 84,000,000, the maximum duty being about 93,600,000. The horse-power developed at different points was as follows, that in the cylinder being estimated in terms of the friction-ratios previously deduced: showing that the effective horse-power,—the "pumping duty," is 81 per cent., or about \( \frac{4}{5} \)ths of the indicator horse-power exerted in the cylinder—the "total load;" and that the remaining 19 per cent., or \( \frac{1}{5} \), is absorbed almost in equal proportions by the friction of the engine, including that of the pole on one part; and the friction and resistance of the pump on the other part, being each nearly 10 per cent. of the "total load."

The East London pumping-engines are supplied with steam from Cornish boilers, with single internal flue, of the following dimensions:

| Cylinder | No. of Boilers | Diameter of Shell | Diameter of Flue | Length | Total Area of Fire Grate | |----------|---------------|-------------------|-----------------|--------|-------------------------| | 80 | 3 | 6 in. | 5½ in. | 30½ in.| 27 sq. ft. | | 90 | 3 | 6 in. | 4 in. | 34 in. | 74½ sq. ft. | | 100 | 5 | 5 in. | 3 in. | 30 in. | 116 sq. ft. |

It is found from long-continued observation, that these boilers evaporate 10 lb. of water per lb. of gas coke, or Welsh coal, and the duty of the engines has already been stated according to this ratio. The following are the proportional consumptions of fuel:

| Cylinder | Fuel per Hour | Per Foot of Grate per Hour | Per H.P. Gross Load per Hour | Per H.P. in Pump per Hour | Per H.P. Effective per Hour | |----------|---------------|----------------------------|------------------------------|---------------------------|-----------------------------| | 80 | 266 | 4½ | 2½ | 2¾ | 2½ | | 90 | 382 | 6½ | 2½ | 2½ | 2½ | | 100 | 407 | 6½ | 2½ | 2½ | 2½ |

The following are the proportional consumptions of water from the boilers:

| Cylinder | Water per lb. of Fuel | Water Per Hour | Per Boiler Per Hour | Per Foot of Grate Per Hour | Per Stroke | |----------|-----------------------|----------------|--------------------|---------------------------|------------| | 80 | 10 | 42½ Cubic ft. | 14½ Cubic ft. | 68 Cubic ft. | 6½ | | 90 | 10 | 61½ Cubic ft. | 20½ Cubic ft. | 98 Cubic ft. | 8½ | | 100 | 10 | 75½ Cubic ft. | 18½ Cubic ft. | 104 Cubic ft. | 10½ |

Showing that an average of 4½ lb. of fuel is consumed per square foot of grate per hour, and 2½ lb. per cylinder horse-power per hour, or \( \frac{2}{5} \) lb. per effective horse-power per hour; that an average of 60 cubic feet of water per hour is evaporated for each engine, and 17 cubic feet by each boiler, or about \( \frac{1}{5} \) cubic feet per square foot of grate per hour; and 11 lb., or above a gallon per stroke is consumed.

As the average cylinder or indicator horse-power is 173 horse-power, performed with an average of 3-6 boilers, the gross actual horse-power of one boiler averages only 47 horse-power. About \( \frac{1}{5} \) cubic foot of water is consumed per actual horse-power.

The air-space through the grates averages from \( \frac{1}{4} \)th to \( \frac{1}{4} \)th of the area of the grate. The force of the draft at the base of the chimney varies from \( \frac{1}{10} \) to \( \frac{1}{5} \) inch of water. In the side-flues the draft was \( \frac{1}{10} \)th inch for the boilers of the 80 and 90 inch cylinders, and \( \frac{1}{5} \) to \( \frac{1}{5} \) inch for the 100-inch boilers. In the side-flues the temperature varied from 280° to 415° Fahr., according to the opening of the damper and the strength of the draft; and in the chimney, at a level 10 feet from the top, the temperature varied from 260° to 318°—the average for 24 hours being 274°. The temperature of 35 lb. steam, of 50 lb. total pressure, is 281°, which approaches the lower limit of temperature in the side-flues.

The condensation of steam in the jacket of the 90-inch cylinder, cutting off at \( \frac{1}{4} \)th, amounts to \( \frac{1}{4} \)th, or about 4 per cent. of the steam consumed in the cylinder, and would average about 2 cubic feet per hour for this engine. The steam-jackets are thickly clothed with sand or ashes, to prevent loss of heat by radiation, which, when the clothing is omitted, causes a loss of more than 10 per cent. of the efficiency.

The expenditure of fuel necessary to keep up the steam in one boiler—that is, to keep up the heat of one boiler and the surrounding building—was less than 1 cwt. of coke per day of 24 hours. To keep up the steam in the jacket additionally, the consumption was 2 cwt. per day.

The duty of double-acting rotative pumping-engines, having a crank and fly-wheel, averages from 40,000,000 to 50,000,000; and, in general, their efficiency is not above two-thirds of that of single-acting engines. A pair of rotative engines, with two fly-wheels, for pumping, were found to absorb about 30 per cent. of their power in driving themselves, leaving 70 per cent. for the interior of the pump, or the "steam-duty."

Single-acting pumping-engines are of two general classes:—beam-engines, represented by Plate XXIV., and direct-action, or bull-engines, represented by Plate XXV. In the former, the cylinder is at one end of the beam, the pump at the other; in the latter, the beam is dispensed with and the pump placed under the cylinder, and directly connected to it by the piston-rod. In the beam-engine, the steam from the boiler is admitted to the upper side of the piston and worked expansively, forcing it down, and lifting the weighted pole of the pump at the other end of the beam. The impetus communicated to the moving mass by the initial charge of steam, together with the expansive action of the steam, carries it through the remainder of the stroke, —the velocity of the piston being gradually retarded till the end of its course, while the resistance remains the same. The uplifting of the pole causes a vacuum in the pump into which the water flows. At the end of the steam-stroke the exhaust-valve closes, and the equilibrium-valve opens, admitting the steam above the piston to the under side, and the pressure is equalized above and below the piston; then the weighted pole descends and forces the water out of the pump; the piston ascends also to the top of the cylinder, and, just before reaching the top, the equilibrium-valve closes, and compresses the portion of steam left above the piston, forming a cushion, to bring the piston to a state of rest. The outlet-valve from the lower side of the piston to the condenser is then opened, and fresh steam from the boiler admitted above the piston, when the piston performs the next descending stroke. The speed of the engine is regulated by a "cataract," a mechanical appliance connected with the gearing, which is set in action at every stroke of the engine, and is adjusted to expend itself, and, at a suitable interval, to release the detents by which the movements of the valve-gearings are controlled. The de- tents being disengaged, the steam-valve opens, by means of the action of the treadle-weights, and the steam is admitted for the next steam-stroke. All the valves fall open by means of the treadle-weights, and are closed by means of tappets.

The "Victoria" engine, made for the East London Water-Works, by Messrs Harvey and Co., to the specifica- tion of Mr Greaves, the Engineer of the company, is repre- sented in Plate XXIV. It is a single-acting beam-engine. The cylinder \(a\) is 100 inches in diameter, and is capable of making a stroke of 12 feet. The beam \(b\) is 36 feet long and 7 feet 6 inches deep at the middle. The pole \(c\) is 50 inches diameter, same stroke as the piston; \(d\) is the pole- case, 68 inches diameter, with the water supply \(w\), and the discharging main \(e\), upwards of 4 feet in diameter; \(p\) is the piston; \(i\) the inlet valve; \(o\) the outlet valve; \(t\) the tappet rod, moved by the beam, having tappets to shut the valves, by means of levers and rods; \(c\) the condenser; \(a\) the air- pump; \(h\) the hot well; \(f\) the feed-pump; \(w\) inlet into water- pump; \(x\) exit from pump into mains; \(v\) air-vessel on the mains, to equalise the head of pressure; \(d\) detents, or cross- bars on the beam, to limit the stroke of the engine. For particulars of the boilers, the reader is referred to previous statements.

The direct-acting pumping-engine, made by Messrs Harvey and Co., for the Grand Junction Water-Works at Campden Hill Station (Plate XXIV.), has a 70-inch cylinder, with 10 feet stroke; pole 32½ inch diameter, with 3 feet water-pipes. The letters of reference are the same as for the beam-engine. The pump and valve-gear are worked by means of the lever \(z\), actuated by connection with the pole. The engine is supplied with steam from six boilers \(b, b\), shown in section longitudinally, and endwise in a separate view on the same plate. The boilers are 5 feet 9 inches diameter, with 3 feet 6 inches flues; and are expanded to 6 feet 1 inch, with 3 feet 10 inch fire-places at the furnaces; \(s\) is the steam-chest, to collect the steam from the boilers. The fire-grates are 6 feet long.

There are many varieties of pumping-engine on a small scale. Horizontal direct-action engines are in use. Upright their ordinary upright engine. A new and good system, by Messrs T. Cowburn and Co., of Manchester (fig. 168), deserves notice. The steam-cylinder and the pump are fixed against a cast-iron pillar, which carries the crank- shaft on the top, and a fly-wheel, from which power may be taken off if wanted. The plunger is novel; it is hollow, so as to admit a second plunger, or ram, to work within it. Thus a continuous delivery of water is effected.

CHAP. VII.—PORTABLE ENGINES.

The term portable engines, formerly signifying self-con- tained engines, is now used to signify engines placed on wheels, capable of being moved from place to place, either by their own power of self-propulsion, or by horse-traction. They are in general use for agricultural purposes, and are carried on four broad wheels; they are made light enough to be handy and manageable by the staff of an agricultural establishment. Their boilers are of the locomotive type, with inside fire-box and flue-tubes, and the cylinder, or cylinders if two in number, are fixed on the top of the boiler at one end, and the shaft at the other. The 8 horse- power portable-engines, made by Messrs Clayton, Shuttle- worth, and Co., of Lincoln, are shown in elevation and sec- tion in Plate XXVII., in two varieties, from which it is appar- ent that, practically, the boiler constitutes the foundation or basis of the whole structure. Small boilers are proportion- ally stronger than large ones; in the latter, independent frame-work, is needed. Messrs Barrett, Exall, and An- drewes, of Reading, formerly made their portable-engines on a bed-plate fixed to the boiler, which, with several advan- tages, had the disadvantage of increasing materially the weight of the engine, and was subsequently abandoned. In Clayton and Co.'s engine there is one cylinder \(a\), 9 inches diameter, with 12-inches stroke; and \(n\), the band-pulley on the crank-shaft is 5 feet in diameter. The boiler \(c\) has a fire-box 1 foot 6 inches long, 2 feet 7 inches wide inside, and 2 feet 1 inch in height above the grate. There are 30 flue-tubes 2¼ inches outside diameter, 6 feet 4 inches long within the boiler. The following are the amounts of sur- face:

| Heating surface in the fire-box | 18·5 square feet | |-------------------------------|----------------| | Do. do. tubes | 124·5 " |

Total inside surfaces: 143·0 " Area of grate: 3·87 "

The cylinder is placed directly over the fire-box, from which it is supplied with steam, and the exhaust-pipe from the cylinder is carried through the boiler into the chimney. The chimney is made with a joint to fold back over the boiler, and rest in a crutch, when not required. In the whole design of this engine there is freedom of access to and inspection of all parts. Agricultural engines are, in general features, much alike.

Messrs Clayton and Co. make engines with the cylinders within the hot smoke-box, and jacketed, from which, no doubt, the cylinders derive benefit, leading to economy of fuel.

CHAP. VIII.—LOCOMOTIVE ENGINES.

The general features and characteristics of modern loco- motive practice are represented in Plates XXVI., XXVII., and XXVIII., showing English and American engines. The passenger-locomotive, Plate XXVI., by Messrs Robert Stephenson and Co., is a type of the prevailing kind of engines used in this country for passenger-traffic; and it is specially adapted for drawing express trains at high velocity. It has two cylinders coupled to cranks at right angles on the driving-axle. They are 15¾ inches diameter, and have 20 inches stroke. The driving-wheels are 6 feet 2 inches diameter; so that, at a speed of 50 miles per hour, the piston would move at a mean speed of 760 feet per minute; and at 60 miles per hour, it would have a speed of 910 feet per minute, or three times as much as ordinary stationary engines. There are 170 flue-tubes 2 inches diameter, and 11 feet 4 inches long, and the heating surface is as follows:

Inside heating-surface of tubes: 915 square feet. Do. do. fire-box: 50 " Total: 965 " Area of fire-grate: 1363 "

The goods-engine, with six coupled wheels, Plates XXVI. and XXVII., made by Messrs Beyer, Peacock, and Co., Manchester, represent in like manner the ordinary form of goods engine used in this country. The sections and plans show in detail the construction of the engine. The cylinders are 16 inches diameter, with 24 inches stroke; the wheels are 5 feet diameter. The fire-box is 4 feet 3 inches long, by 3 feet 6 inches wide; there are 191 flue-tubes, 2 inches diameter outside, and 11 feet 7 inches long. The following are the surfaces:

Heating surface of fire-box: 863 square feet. Do. do. tubes: 10524 " Total: 11388 " Area of grate: 149 "

The following literal references in Plate XXVII. will explain the details of the goods engine. A, the cylinder; B, the driving, or crank axle; C, the driving-wheels; D, the fire-box; E, the fire-box shell; F, the barrel of the boiler; G, the flue-tubes; H, the steam-pipe from the boiler, proceeding from the regulator K, in the dome L, to the smoke-box M, and thence by branches N, to the cylinders; O, the blast-pipe; P, the chimney; Q, the valve-gear, handled by the reversing lever R; S, the regulator-handle; T, the safety-valves; U, the feed-pumps.

The following formulas, based on experimental data, have been given for the resistances of engines and trains on railways, at various speeds. Resistance is divided into two parts—a fixed or constant quantity, and a variable quantity, increasing as the square of the speed. The conditions under which the formulas are applicable, are—

1. A good sound road. 2. A straight and level road. 3. An average side wind. 4. Engine, tender, and train, in good working order.

Let \( v \) = the speed of the train in miles per hour, and \( R \) = the total resistance per ton of weight moved. Then, first, to find the total resistance per ton gross, of the engine, tender, and train, there is a constant resistance of 8 lb. per ton, and

\[ R = 8 + \frac{v^2}{171}. \]

Second. To find the resistance per ton, \( R' \), of the train alone, there is a constant of 6 lb. per ton, and

\[ R' = 6 + \frac{v^2}{240}. \]

Third. For the resistance per ton (\( R'' \)) of the engine and tender alone, including machinery-friction, let \( W \) = the weight of the engine and tender in tons, \( w \) = the weight of the train in tons, then

\[ R'' = (6 + \frac{v^2}{240}) + \left(2 + \frac{v^2}{600}\right) \times \left(\frac{W + w}{W}\right), \]

in which the first member is the resistance of the engine and tender as carriages simply, and the second member is the whole machinery-friction, the amount of which varies with the weight of the train.

Note 1. On inclines, the resistance is increased or diminished by gravity, according to the ratio of the incline, and according as the train ascends or descends.

2. The results obtained by the above formulas must be increased one-half more, in order to find the resistance under the ordinary conditions of railways.

The following are examples of the tractive resistance per ton, gross, of engine, tender, and train, on a level, at various speeds:

| Speeds | Resistance under superior Conditions | Resistance under ordinary Conditions | |--------|--------------------------------------|-------------------------------------| | Miles per hour | Ib. per ton | Ib. per ton | | 10 | 8-6 | 13-0 | | 20 | 10-3 | 15-5 | | 30 | 12-2 | 20-0 | | 40 | 17-3 | 26-0 | | 50 | 22-6 | 34-0 | | 60 | 29-0 | 43-5 | | 70 | 36-6 | 55-0 |

The prevailing types of American locomotives are shown in Plate XXVIII. for passengers and for goods traffic in the United States. It is remarkable with what a degree of unanimity American engineers have matured the designs of their engines.

For an account of the history of locomotives, their peculiarities, their performances, and other information, see Railways.

The annexed cut (fig. 169) shows a simple design of balanced slide-valve, by Mr Gregory of Lisbon, for locomotives. The back of the valve is formed cylindrically with a groove, and a packing ring is let into the groove or circular channel, with an elastic tube of india-rubber between. By the partial compression of this tube the ring is pressed against the back of the valve-box, and steam-tightness is produced, preventing the steam from pressing on the protected portion of the back of the valve. This is an object of considerable importance.

SECTION V.—CLASSIFICATION OF STEAM-BOILERS, WITH EXAMPLES.

In the course of the preceding section on steam-engines, a few examples of boilers have been illustrated and described, and performances recorded. It is the object of this section to complete the survey of steam-boilers. Boilers may be classed, generally, as horizontal and as upright boilers, in which the direction of the heating surface is respectively horizontal and vertical. Or they may be classed as open flue-boilers, and as multi-tubular boilers. But it is not necessary to follow any particular classification; a consecutive notice of each kind will suffice.

The Waggon Boiler.—This form of boiler has already been illustrated and described in Part I. It is suitable for low pressure only, and was for a long time the most generally used of all boilers.

The Cylindrical Egg-end boiler, fig. 65, has been in extensive use, but is now used chiefly at collieries, where simplicity is a particular object, and fuel of small value. This boiler is 4 feet 6 inches diameter, and 30 feet long. The grate is 6 feet 6 inches long, and 3 feet wide, and the flue proceeds direct to the other end. The form of this boiler is favourable for strength and safety, but it requires great length of boiler to provide sufficient heating surface. The steam is taken from the receiver at the centre, with a stop-valve above. There are two safety-valves towards the front, and a man-hole behind; and the feed-water is introduced at the far end.

The Retort Boiler, the production of Messrs Dunn and Co., Manchester, is composed of a series of independent cylinders placed side by side (figs. 170), about 19 inches in diameter, and upwards of 9 feet long, with cast-iron ends. They all communicate upwards with a steam-receiver above This class of boiler is undoubtedly safer than the ordinary large boiler, and is equally efficient in evaporative performances. It is compact and portable.

The Cornish Boiler has already been illustrated and described in the previous section.

The Cylindrical Double-Flue Boiler, Plate XX., is in general use in the north of England. It has two fires, one in each flue \(a\), which are fed with fuel alternately, so that the smoke from one furnace mixes with, and is consumed by, the hot gases and air from the other furnace. The products of combustion descend and pass through the flues \(c\), twice along the whole length of the boiler, and thence by the tunnel \(t\) to the chimney. The steam is taken from a dome \(d\), in order to insure its dryness, and is conveyed by the steam-pipe \(s\) to the engine. The other arrangements are similar to those in the Saltaire boilers.

\(m\), man-hole; \(k\), additional man-hole; \(f\), safety-valves; \(g\), spring-balance; \(h\), glass water-gauge; \(i\), blow-off cock; \(r\), feed-pipe; \(v\), gussets, or stays.

In this drawing the flues \(a\), \(b\), \(c\), are represented with the strong ribs of angle iron \(b\), \(b\), \(b\), as recommended by Mr Fairbairn, for obviating the dangers of collapse. The flue, as shown, is in accordance with the laws deduced in the experimental researches recently conducted by him, according to which the strength is three times that of the same flue constructed in the ordinary method without intermediate supports; and the load thrown upon the subject by these experiments is likely to cause a considerable revolution in the construction of flues. (See Experimental Researches on the Collapse of Boiler-Flues, by Mr Fairbairn, 1858.)

Multitubular Horizontal Boilers.—Of these, the Saltaire boilers, introduced by Mr Fairbairn, Plates XVIII. and XX., and already described, are examples of one kind—the double flue in conjunction with the multitubular arrangement. An effective and compact plan of boiler, introduced and constructed by Messrs Smith and Coventry, Manchester, is shown in figs. 171 and 172, in which the boiler consists of a cylindrical shell, and furnace beneath, with return multitubular flues through the boiler, ending in a smoke-box at the front. The safety-valve of this boiler, detailed in fig. 173, is worthy of note. It is placed on knife-edges, and the valve, being spherical, is free to suit itself to the seat.

Agricultural boilers, and locomotive boilers, multitubular, have already been described.

The Butterly boiler, already described (figs. 79, 80), is frequently used.

Varieties of Boilers.—Galloway's boiler (fig. 174) is made with a large oval flue within the boiler, having a number of vertical conical water-passages uniting the top and the bottom; so that a free circulation of water is maintained, and a considerable addition to the heating surface is effected. It is a species of "water-tube" boiler, and there is no doubt it produces good results. There is a double furnace, so as to allow of alternate firing, and the presence of the water-tubes in the thoroughfare of the draft promotes the mixture and combustion of the gases. A steam-reservoir is placed upon the boiler.

In Cowburn's cellular boiler (fig. 175), the internal flue consists of a number of compartments or cells, connected by short wrought-iron pipes to each other, and to the lower There are two distinct furnaces, one within the flue, and one below the boiler, and the flame and smoke from the two furnaces commingle within the cells, and become effectually consumed. The flue, by being subdivided, as shown, yields to expansion elastically endwise; and as, in addition to this advantage, the boiler is more uniformly heated by the distribution of the fires inside and outside, the over-straining of joints is prevented, and the chances of leakage and rupture are diminished.

The safety-valve of this boiler, detailed in figs. 176 and 177, is worthy of note. It is direct-acting, without levers; is freely suspended, so preventing adhesion or sticking to its seat, and readily adjusts itself to the seat, the bearing surfaces being spherical. The fusible plug, or valve (figs. 178, 179), to give warning in case of low water, is another good feature. It consists of an inverted brass cup, fixed on the top of the hottest part of the flue, with a cap screwed on, and numerously perforated, screwed, and filled with block-tin. When the water in the boiler falls, accidentally, below the level of the cap, the plugs are melted out by the heat, and the steam rushing into the furnace, gives the needful warning of danger.

Mr Goodfellow of Hyde, in order to prevent the unequal expansion and straining of flue-boilers, contracts the flue at each end conically (figs. 180), so as to admit of elastic action. It is difficult to perceive how, with such a gentle taper, there can be sufficient scope for the elastic movement. The steeper taper employed in the Cornish boilers, exemplified in Plate XXV., is more likely to be of service, as in fact it has been proved to be sufficient to prevent injurious overstraining of those boilers.

**Upright Boilers.**—The ordinary form of upright boiler is cylindrical, with the axis vertical. An internal fire-box is placed near the bottom, and a faggot of small tubes proceeds from the crown of the fire-box to the roof of the boiler, and thence to the chimney. The direct upward draft thus occasioned, though it quickens combustion, and excites a high temperature, is wasteful, insomuch as the heat is too rapidly conveyed, and much of it is carried away into the chimney. Expedients for detaining or deflecting the current are found to operate with advantage. Messrs Dunn and Co. have produced an effective boiler (figs. 181), on the principle of down-draught; the boiler is cylindrical, with a plain domed fire-box within, divided transversely by two water-partitions; there are two grates, one on each side, and the gases ascending from them meet at the crown, are deflected, and descend through the vertical flue. The shell of the boiler is 4 feet 6 inches diameter, and 10 feet high; the fire-box is 3 feet 10 inches diameter, with a total height of 7 feet of Thermodynamics. The segmental grates have a versed sine of 15½ inches, and Dynamics have a combined area equal to 7 square feet.

The upright boiler of Messrs Armstrong and Bowman, the former of whom is the well-known authority on boilers, contains a number of transverse double-cone steam-generators (figs. 182), within which water circulates, which baffles and diverts the ascending hot currents; the ends of these generators project into the surrounding water space of the boiler, and so throw off the steam which is discharged from them, out of contact with the vertical sides of the fire-box. This boiler is found to work effectively.

An objection to ordinary conical safety-valves, directly weighted, fig. 183, or by a lever (fig. 184), is, that they do not rise with a sufficient degree of freedom to let escape the surplus steam, and the pressure in the boiler consequently rises considerably above that to which the valve is weighted. Mr R. Bodmer has removed this objection, in the design (fig. 185), by admitting water or steam, at the whole pressure in the boiler, under the valve, through an independent tube, free from the interference of the rushing steam. Thus the valve is lifted well off its seat, and a free exit provided for the steam. The valve is in the form of a hollow cylinder, fitting exactly, but easily, over a piston or plug, on which it slides vertically.

**SECTION VI.—THE WORK OF STEAM IN THE STEAM-ENGINE.**

**CHAP. I.—PRINCIPLES OF THERMO-DYNAMICS.**

The following general outline of the principles of thermodynamics—a science of recent origin, and underlying the whole theory of heat-engines—is abstracted from Professor Rankine's work on the *Steam-Engine and other Prime Movers*, which contains the first systematic exposition of the science. **Thermo-dynamics defined.**—It is a matter of ordinary observation, that heat, by expanding bodies, is a source of mechanical energy; and, conversely, that mechanical energy, being expended either in compressing bodies or in friction, is a source of heat. The reduction of the laws, according to which such phenomena take place, to a physical theory, or connected system of principles, constitutes what is called the science of thermo-dynamics.

**First Law of Thermo-dynamics.**—Heat and mechanical energy are mutually convertible; and heat requires for its production, and produces by its disappearance, mechanical energy in the proportion of 772 foot-pounds for each British unit of heat (Joule's equivalent); the said unit being the amount of heat required to raise the temperature of one pound of liquid water by one degree of Fahr., near the temperature of the maximum density. This law may be considered as a particular case of the application of two more general laws, namely,—1. All forms of energy are convertible. 2. The total energy of any substance or system cannot be altered by the mutual actions of its parts.

**Dynamical Expression of Quantities of Heat.**—All quantities of heat, such as the specific heat of any substance, or the latent heat corresponding to any physical effect, or any other of the quantities of heat treated of, may be expressed dynamically, that is, in units of work, by multiplying their values in ordinary units of heat by Joule's equivalent. The following are examples of this mode of expressing quantities of heat, which is by far the most convenient in treating of thermo-dynamical questions:

| Latent heat of evaporation of 1 pound of water | 745,812 | |---------------------------------------------|---------| | from and at 212° | | | Total heat of combustion of 1 pound of carbon | 11,194,000 |

**Total Actual Heat.**—Let a substance, by the expenditure of energy in friction, be brought from a condition of total privation of heat to any particular condition as to heat. Then, if from the total energy so expended, there is subtracted, first, the mechanical work performed by the action of the substance on external bodies, through changes of its volume and figure, during such heating; secondly, the mechanical work due to mutual actions between the particles of the substance itself during such heating; the remainder will represent the energy which is employed in making the substance hot, and which might be made to reappear as ordinary mechanical energy, if it were possible to reduce the substance to a state of total privation of heat. This remainder is the quantity called the *total actual heat* of the substance, being the total energy, or capacity for performing work, which the substance possesses in virtue of being hot.

**Second Law of Thermo-dynamics.**—If the total actual heat of a homogeneous and uniformly hot substance be conceived to be divided into any number of equal parts, the effects of those parts in causing work to be performed are equal.

This law may be considered as a particular case of a general law, applicable to every kind of actual energy; that is, capacity for performing work, constituted by a certain condition of each particle of a substance, how small soever, independently of the presence of other particles (such as the energy of motion).

**Absolute Temperature—Specific Heat, Real and Apparent.**—Temperature is a function depending on the tendency of bodies to communicate the condition of heat to each other. Two bodies are at equal temperatures, when the tendencies of each to make the other hotter are equal. All substances absolutely devoid of heat are at the same temperature. Let this be called the *absolute zero of heat*; and let the scale of temperature be so graduated, that for a given homogeneous substance each degree shall correspond to an equal increment of actual heat. This mode of graduation necessarily leads to the same scale of temperature for all substances. The amount of actual heat expressed in units of work, which corresponds in a given substance to one degree of absolute temperature, is the *real dynamical specific heat* of that substance, and is a constant quantity for all temperatures. The total quantity of mechanical energy required to raise the temperature of unity of weight of a substance by one degree, generally includes, besides the real specific heat, work performed in overcoming molecular forces and external pressures. This is the *apparent dynamical specific heat*, and may be constant or variable. Joule's equivalent is the apparent dynamical specific heat of liquid water at and near its maximum density; and it is probably equal sensibly to the real specific heat of that substance. The real specific heat of each substance is constant at all densities, so long as the substance retains the same condition, solid, liquid, or gaseous; but a change of real specific heat, sometimes considerable, often accompanies the change between any two of these conditions. From the mutual proportionality of actual heat and absolute temperature there follows—

**The Second Law of Thermo-dynamics,** expressed with reference to absolute temperature. If the absolute temperature of any uniformly hot substance be divided into any number of equal parts, the effects of those parts in causing work to be performed are equal.

The first and second laws virtually comprise the whole theory of thermo-dynamics.

**Of Heat Potentially, and Thermo-dynamic Functions.**—The second law of thermo-dynamics may also be expressed in the following form:—The work performed by the disappearance of heat during any indefinitely small variation in the state of a substance, is expressed by the product of the absolute temperature into the variation of a certain function, which function is the rate of variation of the effective work performed with temperature; that is to say, in Professor Rankine's notation, let $U$ = the effective work performed, $\tau$ = the absolute temperature, $J$ = Joule's equivalent; $H$ = the quantity of heat in common thermal units, $k$ = the real dynamic specific heat of the substance; and make

$$\frac{dU}{d\tau} = F;$$

then the work performed by the disappearance of heat is $\tau dF$. This function $F$ has been called the *heat potential* of the given substance for the kind of work under consideration. Now, let the substance both perform work and undergo a variation of absolute temperature $d\tau$; then the whole heat which it must receive from an external source of heat to produce those two effects simultaneously, is

$$JdH = k\tau d\tau + \tau dF = \tau d\phi,$$

in which

$$\phi = k \times \text{hyp. log } \tau + \frac{dU}{d\tau}.$$

$\phi$ is called the *thermo-dynamic function* of the substance for the kind of work in question, and sometimes the *heat-factor*. The above equation is the *general equation of thermo-dynamics*.

**Principal Applications of the Laws of the Expansive Action of Heat.**—The relation between the temperature, pressure, and volume, of 1 pound of any particular substance being known by experiment, the principles of thermo-dynamics serve to compute the quantity of heat which will be absorbed or rejected by 1 pound of that substance under given circumstances; and, conversely, in some cases, when the quantities of heat absorbed or rejected under given circumstances are known by experiment, the same principles serve to determine relations between the temperature, pressure, and density of the substances. The chief subjects to which the principles of the expansive action of heat are applicable are the following:—Real and

When the temperature of the fluid is again to be raised, it is passed through the interstices of the economiser in the contrary direction, and the heat which it had previously given out is in part restored to it. It is impossible to perform this process absolutely without waste. In some experiments by Mr Siemens on air, the waste of heat at each stroke was about $\frac{1}{5}$th part of the heat ultimately abstracted from and restored to the air; and in the air-engines of the ship Ericsson about $\frac{1}{7}$th.

Air-Engines.—The ease with which air is obtained in any quantity, and its safety from explosion at high temperatures, have induced many inventors to devise engines in which it is the working fluid. Very few, however, of these engines have been brought into practical operation, owing chiefly to the difficulty of obtaining a sufficiently rapid convection of heat to and from the mass of air employed, and to the necessity for using a more bulky cylinder than is required for a steam-engine of the same power, and with the same maximum pressure.

Such, very summarily, are the principles and the scope of the modern science of thermo-dynamics, as propounded by Professor Rankine, and investigated by him at considerable length in his work on Prime Movers. His treatise embraces also, in detail, the applications of the power of steam in the steam-engine, and the results of his investigations will be referred to in the sequel as occasion may arise.

CHAP. II.—PRELIMINARY.

In calculations of the duty of steam in a cylinder, the back-pressure on the piston, whether arising from the atmosphere or from a condenser, is an essential element. The piston of an engine, in fact, works between two pressures, and continues in motion, or has a tendency to do so, as long as the pressure in the boiler is greater than that in the condenser, or in the exhaust passage; and when steam is very greatly expanded in a condensing-engine, a low pressure in the condenser is no less necessary than a high pressure in the boiler. If all losses and difficulties incidental to, and perhaps in some degree inseparable from, the use of steam of very high pressure, be neglected, then it must be maintained that the highest pressure in the boiler, coupled with the lowest pressure in the condenser, would give the highest duty for a given quantity of heat, provided the steam is expanded in the cylinder from the greater pressure down to, or nearly down to, the lower pressure.

The term "vacuum," it may be remarked, is liable to a double interpretation, signifying either the absolute pressure in the condenser, or the difference between this and the atmospheric pressure. Now, in questions affecting the quantity of work of steam, and its efficiency in the steam-engine, there are the total pressures respectively in the two separate vessels which require to be considered; that is to say, the initial pressure in the cylinder, and the total pressure in the condenser, into which the exhausted steam is propelled by the superior pressure on the other face of the piston. If the pressure of the atmosphere were 10 lb. or 30 lbs. in place of 14.7 lb. per square inch, as it is, it would not at all affect the action of a condensing engine, farther than slightly diminishing or increasing the force required to work the air-pump, and causing a greater or less weight to be placed upon the safety-valve, in order to obtain the same total pressure in the boiler. When the mercury in an ordinary barometer is observed to stand at a height of 30 inches, and the mercury in another tube communicating with the condenser of a steam-engine, at a height of 5 inches, instead of describing the conditions of the case as representing a vacuum of 25 inches of mercury, it would afford a clearer conception of the matter to consider that the total pressure in the condenser is equal to 5 inches of Work of Steam without Expansion.

The development of heat into pressure and motion by the media of the water in the boiler and the piston of the engine may, sinking details, be thus concisely illustrated. Let \( abcd \) (fig. 186) be a tall cylindrical upright vessel, open at the top, and having one square foot of area of base or cross section; and \( e \) a piston or disc, without weight, exactly fitting the vessel, and capable of moving up and down without friction. Let there be 1 pound weight of water at the bottom of the vessel, with the piston resting upon it. If a fire be lighted beneath the vessel, and heat communicated to the water, the temperature of the water will be raised to 212° Fahr., before any steam is generated, being subject to the atmospheric pressure on the piston, of 14-7 lb. per square inch. When the temperature reaches 212° Fahr., the heat of the fire being continued, it will not rise higher, but, instead of an ascending temperature, steam will be formed and disengaged under the piston; the piston will be raised, with its atmospheric load of 14-7 lb. per square inch, or 2116-8 lb. on the 1 square foot area of the piston, through successive stages, 1 foot high, to the positions \( e', e'', e''' \), until it reaches an elevation of 26-36 feet above the bottom of the cylinder, when the whole of the pound of water will be evaporated, having a constant elasticity throughout the process, of 14-7 lb. per square inch, and a temperature of 212°. In this instance, the boiler and the engine are represented by one vessel, in which the piston and the water are brought into direct contact, and the intervention of pipes or passages for the steam is dispensed with. The work done or duty performed is the raising of a weight of 2116-8 lb., through a height of 26-36 feet, equal to 55,799 foot-pounds, by 1 pound of saturated steam at atmospheric pressure.

Work of Steam without Expansion and without Condensation.—This duty or work done is performed by the steam pushing its way into space, and repelling or pushing aside whatever resisting medium is opposed to its development,—the resistance in the case under consideration being the pressure of the atmosphere. Suppose the experiment to be repeated, with the addition of a weight on the piston equal to the pressure of the atmosphere, making, say, in round numbers, a total incumbent pressure of 30 lb. per square inch, or 4320 lb. absolute weight on the square foot, the area of the piston. On the application of heat, the temperature of the water will be gradually raised to 250°4 Fahr., when the temperature will become stationary, and evaporation will commence, and will proceed until the whole of the water is evaporated. At this stage, the piston, with its load, will have been raised to a height of 13-46 feet; and the work done will be the raising of a weight of 4320 lb., though a height of 13-46 feet, equal to 58,147 foot-pounds, by 1 pound of saturated steam at a total pressure of 30 lb. per square inch, of which about a half is only atmospheric resistance, or exactly 14.7 lb. \( \times 144 \times 13-46 \text{ feet} = 28,492 \text{ foot-pounds}, \) having a balance of 29,655 foot-pounds of tangible work done, which will be all the available useful duty of the pound of steam, supposing it then to be discharged. Generally, whatever be the load imposed on the piston, a deduction must be made from the total duty of an amount necessary for repelling the atmosphere, to find the available useful duty.

Such are precisely the conditions of a non-condensing steam-engine worked without expansion, the steam being admitted behind the piston throughout the whole of the stroke, and then discharged into the atmosphere. The cylinder of a steam-engine, it is true, is placed at a convenient distance from the boiler, and not immediately over the water, as in the experimental apparatus; the difference is, however, only circumstantial, the steam is generated at one end of the boiler, and it goes out simultaneously at the other.

To find how much of the whole quantity of heat consumed in generating the steam, is, by this non-condensing non-expansive process, converted into useful effect, let it be assumed that the water is supplied to the boiler at a boiling temperature, 212° Fahr.; then, according to the formula for the total heat of steam, of 30 lb. pressure per square inch, and 250°4 Fahr.

\[ H = 1113.4 - 212.9 + 3054 = 900.5 + 3054; \]

and the total heat given to the boiler for 1 pound of water evaporated, is 9769 degrees, or units of heat. On the contrary, the tangible work done, or useful effect, has been found to amount to 29,655 foot-pounds, or to 29,655 + 772 = 38.4 units of heat. The loss of heat, that is to say, the unappropriated heat, is therefore 9769 - 38.4 = 938.5 units thrown away, and the proportion of heat utilised is only about \( \frac{1}{3} \)th of the total heat expended.

The amount of heat expended in repelling the atmosphere, expressed in force, was found to be 28,492 foot-pounds; or, expressed in units, it is 28,492 + 772 = 36.9 units. If this be added to the heat utilised, the sum 36.9 + 38.4 = 75.3 units, expresses the total exertion of the steam in forcing its way into existence behind the piston. Of course, the generation of the steam demands the same exertion, whether it takes place behind a piston, or under a safety-valve loaded to the given pressure, through which the surplus steam escapes; and the whole amount of heat supplied to the boiler in the formation of 1 pound of steam of 30 lb. total pressure per square inch, must, therefore, be distributed as follow:

| Units of Heat | |--------------| | 1. In overcoming the molecular attraction, and separating the particles, 696,020 foot-pounds + 772 = 901.6 | | 2. In repelling the incumbent pressure; in other words, to raise a load of 4320 pounds through a height of 13-46 feet = 58,147 foot-pounds + 772 = 75.3 |

The total heat expended = 976.9

Of all this heat, there is, as already estimated, only about a half of the second and smaller portion utilised by a common non-condensing non-expansive steam-engine. Even this twenty-fifth part of the total is too favourable an estimate, as applied to engines in ordinary condition, because there is usually a portion of the steam condensed in the cylinder, or, it may be, burdened with priming, and a back-pressure is caused which still further diminishes the efficiency. The annexed indicator-diagram, fig. 187, was taken from a small non-condensing steam-engine, from which it appears that, of the whole sensible pressure, 20 lb. above the atmosphere, exerted on the piston, 5 lb. was destroyed by back-pressure, leaving only an effective pressure of 15 lb. on the piston, or only three-fourths of the whole sensible pressure.

If, under the same conditions, steam of 45 lb. total pressure be raised from water supplied at 212°, and worked at that pressure in the cylinder, it is found, by a similar calculation, that for each pound of steam consumed, 77 units of heat are absorbed in repelling the incumbent pressure; of this only 25 units are expended on the atmospheric resistance, and the remainder, 52 units, may be utilised on the piston, being one-third more than with 30 lb. steam, and constituting one-nineteenth of the whole of the heat, 984°, supplied to the water in generating steam.

Again, with steam of 60 lb. total pressure, by the same calculation, 78.5 units of heat are absorbed in repelling incumbent pressure, of which only 19 units are expended on the atmospheric resistance, leaving 59.5 units for useful work, or a little over one-half more than with 30 lb. steam.

There is a common notion, that in proportion as steam of higher pressure exceeds the atmospheric pressure, the efficiency for useful work increases in the same ratio. This is true only of the capacity of a cylinder of given dimensions for delivering useful work. But it is very far from the truth if applied to the efficiency of a given weight of steam; for, in the examples above, it has been found approximately that—

For 1 pound weight of steam of 30, 45, 60 lb. total pressure. The effective useful pressures are 15, 30, 45

But the useful quantities of work are only.............. 38.4, 52.0, 59.5 units of heat.

Only in the ratio of about.............. 1, 1.3, 1.5

Showing, that with three times the excess of pressure above the atmosphere, there is only one and a half times the useful work, with equal weights of steam. And, in order to get double the useful work from the same weight of steam, worked, of course, in the manner which has formed the subject of this chapter, it would be necessary to introduce high-pressed steam of 180 lb. absolute pressure per square inch, compared with steam of 30 lb. per inch.

Work of Steam without Expansion, but with Condensation.—That is to say, the steam is supposed to be condensed within the cylinder (fig. 186), after the piston has been raised through the height due to its volume, against the resistance of the atmosphere; and the development of useful duty from the steam, heretofore terminated on having raised the piston with its incumbent load, comprising the unprofitable atmospheric resistance, may be further promoted. Let the material weight placed on the piston be removed, and the piston be attached to a rising weight over a pulley nearly equal to the amount of atmospheric pressure on the piston, or 2116.8 lb.; then, if an absolute vacuum be established within the cylinder, the piston will descend to the bottom, and the suspended weight will rise through a height equal to the descent. This weight may, for present purposes, be supposed to equal the atmospheric pressure, and the additional work done will be equal to 2116.8 lb., multiplied by the descent of the piston, or the ascent of the weight, in feet; being, in fact, equal to the work before unprofitably expended in opposing the atmosphere. The work thus redeemed was found to be, for 1 pound weight of 30 lb. steam under the piston, 28,492 foot-pounds, or 36.9 units; for 45 lb. steam, 25 units of heat; for 6 lb. steam, 19 units.

These reclaimed quantities of work being added to the useful work previously done in raising the piston, the total useful work done would be—

For 30 lb. steam........................................... 75.3 units. For 45 lb. steam........................................... 77 " " For 60 lb. steam........................................... 78.5 " "

being, in any case, just about one-thirteenth of the heat supplied to the boiler.

The above conclusions are directly applicable to the steam-engine; and it appears, that when steam is worked in a common steam-engine, in direct continued connection with the boiler, non-expansively, that is to say, without applying its inherent expansive energy, there is only a small fraction, about a thirteenth, of all the heat given to the boiler, converted into useful work on the piston, with the advantage of a perfect vacuum behind the piston; and the proportion of useful work is still less than that if the steam be exhausted into the atmosphere. These deductions are based on otherwise favourable conditions.

CHAP. IV.—THE WORK OF GASEOUS STEAM WITH EXPANSION.

The action of gaseous or superheated steam, maintained at a constant temperature within the cylinder, is based on comparatively simple conditions. The total pressure of gaseous steam, when the temperature is constant, falls in proportion as the volume by expansion is increased, the pressure being inversely as the volume; and, consequently, the product of the diminishing pressure by the increasing volume, at all points of the stroke, is constant; and, for 1 pound weight of steam, it is equal to 85.4 times the absolute temperature; or, $PV = 85.4T$. This constant relation of the pressure and volume of expanding steam may be represented diagrammatically on a base line $AB$ (fig. 188), representing volume, which is supposed to be developed in a straight line, in a cylinder, of which $AB$ is the length of stroke. Neglecting clearance for the present, let steam be admitted into the cylinder at a total pressure of 10 lb., represented by the ordinate $AC$, and for a space of 1 foot in length represented by $AD$, then the rectangular area $ACED$ represents the product of the pressure and volume of the steam admitted into the cylinder. Let the steam be expanded at a constant temperature into double the volume $AD$, and to half the pressure $DE$; the area of the elongated rectangle $ADE$ will be equal to that of the initial rectangle $ACE$. Expanding, further, to three volumes $AF$, and to the third part of the initial pressure $DE$, the still more extended rectangle $AE'$ is equal to each of the others. The expansion may be further extended to four and five volumes, and to a fourth and a fifth of the pressure, constituting the rectangles $AE'$ and $AE''$; and while the products of the pressure and volume are always the same, represented by the successively elongated rectangles, the elevations $DE', DE'', \ldots$ etc., will correctly represent the successively reduced pressures for the several increments of expansion. It is obvious, that any number of intermediate pressures may be interpolated by the simple process of dividing the product of the initial pressure and volume by the distance traversed by the piston, and that they may be joined by a curve $AE''$, traced through their summits, representing the continuous expansion curve of the steam, which would be traced by the pencil of an indicator applied to the cylinder. This curve of expansion is, according to the principle of its construction—the equality of the rectangles—a portion of a hyperbola; and it may be indefinitely extended at either end, to embrace on the one part intense pressures and small volumes, and, on the other part, very low pressures and large volumes. In the extension upward of the curve into the higher pressures, it would only be correct as an exponent of the mutual relation of pressure and volume, on the understanding that a proportionally high uniform temperature should be adopted, to ensure the gaseous condition of the steam. A curve constructed on this principle, though indefinitely extended in either direction, does not meet or touch the base line \(a_b\), or the vertical \(a_c\).

As the rectangular area \(a_e\), in the foregoing figure, expresses the work done by the steam in entering and occupying the cylinder, the hyperbolic area \(d_e\) likewise expresses the work done by the steam by expansion against the piston. This area, and consequently the quantity of work done, may be computed from the known relations of hyperbolic surfaces with their base lines, according to which, if the base lines \(a_d\), \(a_d'\), \(a_d''\), etc., extend in a geometrical ratio, the successive areas \(d_e\), \(d_e'\), etc., increase in an arithmetical ratio. If, for example, the lines or volumes \(a_d\), \(a_d'\), etc., are as 1, 2, 4, 8, 16, etc., the areas \(d_e\), \(d_e'\), etc., are as 1, 2, 3, 4, etc. On the principles of logarithms, which represent, in arithmetical ratio, numbers naturally in geometrical ratio, special tables of hyperbolic logarithms are compiled, to facilitate the calculation of areas of power due to various degrees of expansion,—such logarithms being, in fact, direct expressions of the proportion borne by the expansive duty pertaining to different degrees of expansion, to the work done by the initial volume of steam. If, for example, the initial volume be expressed by 1, and the total volumes by expansion to different degrees be represented by

\[ 2, \quad 4, \quad 8, \quad 16, \quad \text{etc.} \]

the hyperbolic logarithms of these numbers are,

\[ 693, \quad 1386, \quad 2079, \quad 2772, \]

which are in arithmetical proportion thus—

\[ 1, \quad 2, \quad 3, \quad 4, \]

and express the actual ratios of the entire expansive duties successively to the initial duty of the steam in entering and occupying the cylinder previous to expansion. The logarithms, it will be noted, are, in each case, the logarithm of the number expressing the ratio of expansion, or the number of times the initial volume is expanded.

It is needful, of course, to make a deduction for the unavoidable back-pressure in the condenser, to find the effective duty of the steam. Suppose a cylinder 5 feet long, with 1 square inch area of piston; steam being admitted, as before, at 10 lb. pressure; let the back-pressure be taken as a uniform resistance of 2 lb. throughout the stroke, and applied to the diagram of duty, constructed as before, as in fig. 189, where the 2 lb. zone of resistance is shaded. It is plain, in the first place, that if the steam, after following the piston through the first foot of the stroke \(a_d\), be exhausted into the condenser, with a residual pressure of 2 lb., the effective pressure is \(10 - 2 = 8\) lb., and the work done is equal to 8 foot-pounds. But let the steam be expanded through the remaining four-fifths of the stroke; then—

At the end of the.............. 1st, 24, 34, 4th, 5th foot of stroke The total pressures would be 10 5 34 24 2 lb. per sq. inch. The back-pressures.............. 2 2 2 2 2 lb. do. do. The effective pressures........... 8 3 1 0 0 lb. do. do.

Then the total work done by expansion up to the end of each foot of stroke is represented by the hyperbolic logarithm of the ratio of expansion, when the work done during admission is represented by unity, thus—

At the end of the.............. 1st, 24, 34, 4th, 5th foot of stroke The steam is expanded into 2 3 4 5 volumes. Of which the hyp. logs. are 0 69 110 139 161 The initial duty being as... 1 1 1 1 1 unity. And the total duty as..... 1 69 210 239 251

showing that the total duty of the steam, by expanding it to five times the initial volume, is increased to fully 2½ times the initial duty without expansion, if no deduction be made for back-pressure. To make this necessary deduction at the rate of 2 lb. per square inch, and also to express the relative duty in foot-pounds, the whole initial duty is, as already said, 10 foot-pounds on a piston of 1 square inch area, moved through 1 foot, with a pressure of 10 lb.; the total work done in the first and second feet of the stroke is \(10 \times 1.69 = 16.9\) foot-pounds; and so on. Then there is a drawback of 2 foot-pounds on each foot of the stroke. The relations of the work of the same weight of steam in the cylinder are therefore as follows:—

At the end of the.............. 1st, 24, 34, 4th, 5th foot of stroke The total work done is... 10 69 210 239 251 foot-pounds. The total resistance is... 2, 4, 6, 8, 10 do. The total effective work is 8, 12, 15, 18, 21 do. And the gain by expansion is... 0 61 87 99 101 per cent.

Here it is to be observed, that by expanding the steam five times, the effective work is doubled, that is, it is increased from 8 foot-pounds to 161 foot-pounds; and that when the expansion is extended to the utmost useful limit—to the point at which the pressure of the expanded steam becomes equal to the back-pressure—the total resistance of the back-pressure in foot-pounds amounts to as much as the total work, which is done by the steam previous to expansion,—namely, 10 foot-pounds. It follows, therefore, that the total effective work of the steam expanded down to the pressure in the condenser is just equal to the total work developed by expansion alone,—that is, 161 foot-pounds is equal to 261—10 foot-pounds. The initial performance of the steam may be said, conventionally, to go for nothing, being balanced by the resistance, and the whole of the useful work may be said to be accomplished by the residual expansive force alone. This conclusion applies generally wherever the steam is expanded down to the pressure in the condenser. It may further be concluded, that the utmost ratio of expansion, or limit to which expansion can be carried efficiently,—regard being had simply to the operations of the steam, as indicated in the cylinder,—is measured by the number of times which the total pressure of the vapour in the condenser is contained in the total initial steam in the cylinder.

For further illustration, let the total initial pressure \(a_c\) of the steam in the cylinder (fig. 189), during the first foot of stroke be 75 lb. per square inch, and suppose it to be expanded five times as before, and then exhausted into the atmosphere, with a back-pressure of 15 lb. per square inch throughout the stroke, indicated by the shaded zone. Then, as before, if the steam be maintained at a uniform temperature, the pressure at the end of the stroke will be \(75 - 4.5 = 15\) lb. per square inch, or down to atmospheric pressure, beyond which it would be useless to expand it. On a piston of 1 square inch area, the ratios of work done will be as before—

At the end of the.............. 1st, 24, 34, 4th, 5th foot of stroke The total work done is... 75 517 157.5 179.2 195.7 foot-pounds. The total resistance is... 15 30 45 60 75 do. The total effective work is... 60 487 112.5 119.2 120.7 do. And the gain by expansion is... 0 61 87 99 101 per cent.

The proportions of work, in this case, where steam of five atmospheres is expanded five times, and exhausted into the atmosphere at a pressure of one atmosphere, are the same as when steam of 10 lb. pressure per inch is also expanded five times, and exhausted at a pressure of one-fifth, or 2 lb. per inch; and they indicate an equal efficiency of the steam in the way it is applied. It is observable from the foregoing illustration, that when the opposing pressure is not less than one-fifth of the initial pressure in the cylinder, or, say in the boiler, there is no material extra duty gained by expanding to more than four volumes; in further expanding to five volumes, the proportion of further gain is only 0·2 upon 15·9, or 1¼ per cent.

The clearance which must be allowed between the piston and the end of the cylinder, for safety of working, and also the space or passage from the valve to the cylinder, must necessarily be filled with steam of the initial pressure, at the commencement of every stroke, which, doing no work, must be reckoned as entirely non-effective in non-expansive engines, and, in that respect, as wasted. Where the steam, however, is expanded within the cylinder, the clearance-steam contributes its quota of effect by expansion; and, therefore, in estimating the work done by steam, a distinction must be drawn between the quantity of steam effectively engaged in the cylinder before the suppression or cut-off, and the quantity after suppression—that is to say, before expansion and during expansion. For simplicity of calculation, the volume of the clearance may be expressed in units of the stroke, and added to the period of suppression, in order to express proportionally the initial volume of the steam submitted to expansion.

In the construction of a formula for the relative efficiency of gaseous steam worked expansively, viewed simply as a problem of absolute duty, irrespective of drawbacks on account of the resistance in the condenser and otherwise, the length of the stroke, the clearance, the period of admission, and the absolute initial pressure, are essential elements; and whereas the ratio of expansion is commonly expressed by the number of times the period of admission is contained in the length of stroke, or otherwise by the fraction of the stroke at which the steam is cut off, the actual ratio for the calculation of efficiency must be formed upon the addition of the amount of the clearance.

Let the duty of steam, without expansion, when the clearance is neglected, be expressed by unity, or 1; then the additional duty by expansively working it, is expressed by the hyperbolic logarithm of the ratio of expansion, and the entire duty is expressed by $1 + \text{hyp. log. ratio of expansion}$. When the clearance is added, the initial duty is proportionally less than unity, as the clearance adds to the volume consumed; and the duty is expressible by the period of admission divided by itself plus the clearance. The ratio of expansion also is less, insomuch as the clearance augments the initial volume proportionally more than it does the expanded volume, and is, in fact, equal to the initial volume plus the clearance, divided by the final volume plus the clearance. Let $L$ = the length of stroke, $l$ = the period of admission, and $c$ = the clearance. Then the relative expression of the entire duty, neglecting clearance, is $1 + \text{hyp. log. } \frac{L}{l}$.

If the clearance be added, the expression becomes

$$\frac{l}{l+c} + \text{hyp. log. } \frac{L+c}{l+c};$$

in which the first fraction expresses something less than 1, or unity (which is the assumed datum or measure of the entire non-expansive duty of steam without clearance), being the depreciation caused by the presence of clearance; and the second fraction is the modified ratio of expansion.

Now, the period of admission plus the clearance, or $(l+c)$, is the measure of the volume of steam actually admitted to the cylinder for the performance of work, to which the foregoing general expression of entire duty is related; and if it be multiplied by the area of the piston ($a$), the product will be the volume of the steam admitted, or $a(l+c)$. The entire duty of which the steam actually expended would be capable, without expansion and without clearance, would be equal to this volume multiplied by the pressure ($p$) per unit of surface, or to $ap(l+c)$; and it is Steam with to this standard measure of duty that the general expression of entire duty must be applied by multiplying the one by the other, in order to show how the work of steam is affected by the clearance. If $W =$ the actual entire duty, then

$$W = ap(l+c)\left\{\frac{l}{l+c} + \text{hyp. log. } \frac{L+c}{l+c}\right\}.$$

For non-expansive engines, the actual entire duty is expressed by the equation—

$$W = ap(l+c)\frac{l}{l+c} = apl,$$

in which $l$ becomes equal to $L$, and signifying that the actual duty is, of course, equal to the product of the volume described by the piston, or the area multiplied by the stroke, and the pressure per unit of surface.

To indicate the importance of the clearance as an element in the efficiency of steam, take a locomotive-cylinder 15 inches in diameter and 20 inches stroke, with a clearance of $\frac{1}{4}$ inch off each end of the cylinder, and steam-passages 10 inches wide by $1\frac{1}{4}$ inch broad, and brought up in the usual manner to a single valve-face in the middle of the cylinder. Taking each passage at 12 inches in length, its cubic contents are $10 \times 1\frac{1}{4} \times 12 = 150$ cubic inches, which, divided by the area of piston, gives $\frac{3}{4}$ inch as the equivalent length of cylinder; and $\frac{3}{4} + \frac{1}{4} = 1$ inch, or one-twentieth of the stroke, is the total clearance or steam-room at each end of the cylinder. Now, if the steam were cut off at one-fifth, or 4 inches, and expanded to the end of the stroke, it would not properly be expanded so much as five times; for there are 5 inches of steam to be expanded into 20 inches, and the real expansion would be only four times. In cylinders, of course, where higher ratios of expansion are practised, the influence of clearance is still greater; and if, with a clearance of one-twentieth of the stroke, the steam be cut off at a twentieth, the volume admitted would be two-twentieths, and the steam would only be expanded into ten volumes, and not twenty. The advantages in this respect of short and direct passages from the valve-chest to the cylinder are obvious—insuring a higher initial pressure, a higher ratio of expansion, and a less proportion of inactive steam in the cylinder than with long passages.

Professor Rankine proposes the following formula for the work $U$ of 1 pound of gaseous steam, or "steam-gas," in which $P_1$ and $V_1$ are the initial pressure and volume, $r$ is the ratio of expansion, and $P_2$ is the back-pressure:

$$U = P_1 V_1 \left(4\cdot29 - 3\cdot29 \frac{1}{r}\right) - P_2 r V_1.$$

He estimates that the gain of efficiency by superheating steam, of 34 lb. total pressure, in an ordinary engine, cutting off at one-fifth, by the heat of the furnace, is about 15 per cent., and that if the heat for this purpose be derived from the flues, which would otherwise be wasted, the saving is about 23 per cent. The latter result agrees very nearly with the general results of recent practice.

The preceding discussion comprehends the ordinary hypothesis on which the work of steam by expansive working is usually estimated. It assumes too much, however—that steam, as used in practice, is entirely gaseous; and it does not consist with the true ratio of expansion in a steam-cylinder. The exemplifications of its application, in estimating the particular gains by expansive-working, are, nevertheless, important, as they measure, with a sufficient degree of exactitude, the value of expansive-working generally, in various degrees, as a means of increasing the efficiency of steam, and they point to the special significance of back-pressure in neutralising in an important degree the expected benefit of prolonged expansion. In the following chapter is treated the work of dry saturated steam,—the condition in which steam exists in actual good practice.

The "regenerative steam-engine" of Mr C. W. Siemens is worked with gaseous steam, maintained at a nearly constant temperature, by placing the cylinder over a furnace; but the steam on its way to and from the space below the plunger of the cylinder traverses a "regenerator," the effect of which is, that nearly the whole of the heat employed to raise the temperature of the steam above the boiling point corresponding to its pressure is obtained at each stroke from the regenerator in which that heat has previously been stored by steam leaving the hot end of the cylinder. Mr Siemens states, that in some of his experiments with this engine, the consumption of fuel was only 1½ lb. per indicator horse-power per hour.

CHAP. V.—THE WORK OF DRY SATURATED STEAM WITH EXPANSION.

The pressure of saturated, or ordinary steam, expanding in a jacketless or unheated cylinder, it has already been stated, falls more rapidly than that of steam of equal temperature, during the first stages of expansion; but ultimately less rapidly, if the expansion be sufficiently prolonged. This has been proved by the writer, from his observations on locomotive-engines, and by others on stationary-engines, to arise from the evaporation of water from the heated sides of the cylinder, when the expanding steam has fallen to a relatively low pressure and temperature—on the same principle as the spontaneous ebullition of hot fluids in vacuo—the water having been formed by the precipitation of steam during the admission to the cylinder, and in the first stages of expansion, upon the relatively cold surface of the cylinder.

There is another, but secondary source of the production of water in the cylinder,—the condensation of ordinary steam during the process of expansion behind the piston, owing to the conversion of part of the heat into work, and the precipitation of the water thus forsaken. This is a species of condensation which arises independently of the relative temperature of the cylinder, and would exist even if the material of the cylinder were a perfect non-conductor. Its existence was originally demonstrated, contemporaneously and independently, by Professor Rankine and Professor Clausius, in 1849.

This condition of things supplies an explanation of the whole mystery of expansive-working of steam. It fully accounts for the failure of attempts to economise fuel by expansive-working, without any provision for preventing the condensation of the steam within the cylinder, arising from the extreme alternations of temperature of the steam itself, during a double stroke of the piston, and the sympathetic condition of the cylinder, heated and cooled by the steam. In ordinary practice, the higher the initial pressure in the cylinder, and the greater the extent of expansive-working to which the steam is subjected, the more difficult it is to prevent the pressure in the condenser from rising; because, as before stated, if the cylinder be exposed, and the integrity of the steam be affected by the alternations of temperature within, there is a considerable amount of supplementary evaporation of water within the cylinder towards the end of the steam-stroke, as well as during the return stroke, when the cylinder is open to the condenser, which partially overpowers the latter, and excites additional back-pressure; and yet, if the back-pressure be not subdued, the gain by expansive-working is neutralised. The loss of expansive efficiency from this cause is important also in non-condensing engines.

The means of preventing the existence of water in the cylinders, thus caused, is provided by supplying heat to the cylinder from an independent source—an external jacket of hot steam, for example—in virtue of which the cylinder is not only maintained at a temperature as high as that of the steam within it, but may also communicate a portion of its heat to the expanding steam. And the question arises, is it sufficient that the steam should be simply preserved from being diminished in quantity by condensation during expansion: being supplied from the jacket while expanding, with just sufficient heat to prevent condensation? Or, is it preferable that the steam should be superheated into the gaseous condition, either within the cylinder from the jacket, or by other means, before it is admitted, so as to act as a perfect gas, considerably removed from the state of saturation? It can be shown that the best result is obtainable when there is just a sufficient supply of heat communicated to the steam in the cylinder, to prevent the precipitation of any portion of it as water. If the steam be discharged from the cylinder in a superheated state, it carries with it a greater quantity of rejected heat than if discharged at the saturation-point.

The liquefaction of steam in the cylinder, by the legitimate process of the conversion of heat into work, is not in itself a loss; it is, in fact, an expenditure of steam representing its equivalent of duty, and if the water of condensation thus precipitated could be instantly removed from the cylinder as soon as formed, so as to leave the cylinder dry at the end of the stroke, a still better result would be obtained per unit of heat expended, than when heat is absorbed from the jacket to prevent its condensation. But the objection to this liquefaction of steam in the cylinder is a practical one, as water in the cylinder, by whatever cause it may be formed, cools the cylinder by evaporation from its surfaces—an evaporation doing but little good for the time, that is to say, during the remainder of the steam-stroke, and otherwise positive harm.

It is believed that, practically, the use of the steam-jacket does not extend materially beyond the maintenance of the steam in a state of dry saturation while expanding; the steam in the cylinder receiving just enough of heat from the steam in the jacket to prevent any appreciable part of it from condensing without superheating it. Professor Rankine, in adopting this view of the case, bases the assumption on the fact, "that dry steam is a bad conductor of heat, as compared with liquid water, or with cloudy steam;" and that after cloudy steam has received enough of heat to make it dry, or nearly dry, it will receive additional heat very slowly. The assumption," he adds, "is justified by the fact, that its results are confirmed by experiment." Professor Rankine investigated, and published in 1859, exact and also approximate formulas for the work of dry saturated steam, under such conditions, and for the expenditure of heat in producing and using that steam. Adopting his notation, let $T =$ absolute temperature in degrees of Fahrenheit = temperature measured from the ordinary zero + 461.2°; $p =$ pressure in pounds on the square foot; $v =$ volume of 1 pound of steam in cubic feet; $t_p$, $P_t$, $v_t$, refer to the admission of steam into the cylinder; $t_e$, $P_e$, $v_e$, to the end of the expansion; let $r = \frac{v_e}{v_t}$, the ratio of expansion; $p_e$, the pressure of exhaustion, or back-pressure; $t_w$, the absolute temperature of the feed-water; $J =$ Joule's equivalent, or specific heat of 1 pound of liquid water, = 772 foot-pounds per degree of Fahrenheit; $W =$ work of 1 pound of steam in foot-pounds; $H =$ expenditure of heat per pound of steam in foot-pounds; $a = 1,109,550$ foot-pounds; $b = 540$ foot-pounds per degree of Fahrenheit; efficiency of steam, $W + H$.

Let the curve $b c k$ (fig. 190) be the expansion-curve of a body of dry saturated steam. Let $OA = p_t$, and $AB = v_t$, represent the pressure and volume of admission; also let $OD = p_e$, and $DC = v_e$, represent the pressure and volume at the end of the expansion; and let $OF = p_e$, the pressure of exhaustion; then $\frac{v_e}{v_t} = r$ is the ratio of expansion, Work of Dry Saturated Steam with Expansion.

\[ \frac{17}{r} - \frac{16}{r^2} \]

By efficiency is meant the proportion which the actual useful effect bears to the whole expenditure of heat communicated to the steam.

In the foregoing formulas, the initial volume for the effective cut-off is measured to the point at which the curve of expansion, if produced upwards, intersects the horizontal line of admission. It is thus that the action of wire-drawing is accounted for, and an example of the operation is shown in fig. 134, where, by an extension of the expansion-line upwards, the shaded area D is enclosed, and the point of effective cut-off is placed 2 inches in advance, at 5 inches of the stroke, instead of 7 inches, to which the valves are set.

There is a small loss of effect by the release of the steam from the cylinder before the end of the stroke, shown also by shaded spaces in fig. 134. The compression or cushioning of the exhaust steam, though it reduces the area of the diagram, and the work done, does not, unless excessive, reduce the efficiency of the steam.

The clearance, as before mentioned, adds to the space to be filled with steam for each stroke, and reduces the ratio of expansion. This effect is more or less counterbalanced by the wire-drawing, the influence of which is to increase the ratio of expansion.

When the feed-water is delivered at other temperatures than 104° Fahr., or thereabouts, for which temperature the approximate formula for the expenditure of heat was obtained, an allowance must be made for the greater or less expenditure of heat, in proportion to the total heat of the steam generated. For example, if the water be heated to 204° Fahr., 100° higher than assumed in the formula; and if the steam be generated at 100 lb. total pressure, having 1181.4° total heat, the addition of 100° to the temperature of the feed-water reduces the expenditure of heat by

\[ \frac{100}{1181.4} = \frac{1}{11.814} \]

and the coefficient of the formula

\[ H = 15\frac{1}{2} p_1 v_1 \]

is reduced by \(15\frac{1}{2} \times 11.814 = 1.3\), to 14.2; and the expression becomes

\[ H = 14.2 p_1 v_1 \]

The following example of the application of the formulas to actual engines, and of the comparison of their results with those of experiment, is given by Professor Rankine:

**Example.—Double-cylinder engines of 744 indicator horse-power, calculated by the exact formulas:**

**Data.**

| Bottom of Cylinders. | Top of Cylinders. | |----------------------|-------------------| | Pressure of admission, \(p_1\) = 144 | 337 | | Back-pressure, \(p_2\) = 144 | 49 | | Ratio of expansion, \(r\) = 4.1 | 4.1 | | Ordinary temperature of feed-water, 104° Fahr. | |

**Calculated Results.**

| Bottom. | Top. | |---------|------| | Final volume of 1 pound of steam, \(v_1\) = \(r^2\) | 50.375 | | Final pressure, \(p_2\) = 144 | 7.367 | | Work of 1 pound of steam, \(W\) = 109552 | 117338 | | Mean effective pressure in pounds on the inch | 15.1 | | Mean of both results | 13.03 | | Mean effective pressure, as observed | 13.1 | | Difference | 0.07 |

Available heat expended per pound of steam:

- Bottom: 906889 foot-pounds - Top: 925678 foot-pounds

Pressure in lbs. per square inch, equivalent to heat:

- Bottom: 125 - Top: 86.4

Mean Efficiency:

- Bottom: 0.121 - Top: 0.127

Mean, 13.03—105.7: 0.123 The same example calculated by the approximate formula:

**Data**

- Mean pressure of admission, \( p_1 = 144 \) lb. per inch. - Mean back-pressure, \( p_2 = 144 \) lb. per inch. - Mean cut of \( \frac{v_1}{v_2} = \frac{0.24 + 0.16}{2} = 0.2 \)

**Results**

- Mean gross pressure, \( 34 \times 505 = 17.17 \) - Mean effective pressure, \( 17.17 - 4 = 13.17 \) - Observed, \( 13.10 \) - Difference, \( 0.07 \) - Pressure equivalent to expenditure of heat, \( 105.4 \) - Efficiency, \( 0.125 \)

Both the exact and the approximate formulas appear to possess a great degree of accuracy in their practical application.

Mr Brownlee, in July 1859, published the following formulas for the work and heat of dry saturated steam, founded, like Rankine's approximate formula, on the fact of the uniform variation of the density, with a constant power of the pressure; in fact, the '941 power. (See Steam.)

Conversely, the pressure varies as the \( \frac{1}{941} \) power of the volume inversely, or as the \( \frac{17}{16} \) power, or as the \( 1.0627 \) power inversely, which are equivalent expressions of the power; and

\[ p_2 = p_1 \left( \frac{v_1}{v_2} \right)^{\frac{1}{941}} = p_1 \left( \frac{v_1}{v_2} \right)^{1.0627} \]

If the steam be discharged after expansion at the same terminal pressure \( p_2 \), as the pressure at which the steam is condensed or exhausted, the total available work done, in foot-pounds, including clearance, and other drawbacks, is

\[ W = \frac{1.0627}{0.627} P_1 v_1 \left[ 1 - \left( \frac{v_1}{v_2} \right)^{0.627} \right] \]

Let \( H_i \) = the total heat of the initial steam, and \( H_e \) = the total heat at the end of the expansion; then the heat required to be supplied from the jacket to prevent condensation, while the steam expands from \( p_1 \) to \( p_2 \),

\[ W = \frac{W}{J} = (H_i - H_e) \]

in which \( J = 772 \), Joule's equivalent.

If again, the steam is permitted to flow into the condenser, or any space in which the pressure is less than that, \( p_2 \), at which expansion is terminated, and the pressure in the cylinder is thereby reduced to \( p_2 \), at which pressure the remaining vapour is finally expelled, there is obtained the additional work \( v_2 (p_2 - p_1) \); and the whole available work of the steam is

\[ W = 16499 P_1 v_1 \left[ 1 - \left( \frac{v_1}{v_2} \right)^{0.627} \right] + v_2 (p_2 - p_1). \]

The exhaustion from the cylinder to the lowest pressure \( p_2 \) has a slightly cooling effect on the cylinder, as the exhausted vapour falls in temperature, and partially condenses while expanding and expelling that portion which flows off. Neglecting, in the meantime, this small expenditure of heat, Mr Brownlee takes for illustration the case of 1 pound of steam at 100 lb. total pressure per square inch, or 14,400 lb. per square foot, expanded to ten volumes, and maintained in a state of dry saturation by the addition of heat from the jacket. The terminal pressure will be

\[ p_2 = 100 \times \left( \frac{1}{10} \right)^{1.0627} = 8.656 \text{ lb. per square inch}, \]

with a temperature of 185°5; \( p_1 = 14,400 \) lb. per square foot; \( v_1 = 43351 \) cubic feet; \( \frac{v_1}{v_2} = \frac{1}{10} \) and \( 1 - \left( \frac{1}{10} \right)^{1.0627} = 0.344 \). If the back-pressure be equal to the pressure of final expansion \( p_2 \), the whole available work will be

\[ W = 16499 \times 14400 \times 43351 \times 0.344 = 42200 \text{ ft.-pounds}. \]

Then, assuming the water to be supplied at 100°, the total heat \( H_i \), of 100 lbs. steam, reckoned from this temperature, will be 1113.4° (see Steam), whilst at 185°5 and 8.656 lb. pressure, the total heat \( H_e \) will be, 1013.4° + (305 × 185.5) = 1070.2°; and \( H_i - H_e = 1113.4 - 1070.2 = 43.2° \).

Also, \( W = 142200 \)

\[ \frac{W}{J} = \frac{142200}{772} = 184.2 \text{ units}. \]

Therefore, the heat absorbed from the jacket, necessary to prevent condensation, is 184.2 - 43.2 = 141 units of heat. Hence, with water supplied at 100°, the total heat necessarily expended per pound of steam = 1113.4 + 141 = 1254.4 units, and the work done for each unit of heat expended, is

\[ \frac{142200}{1254.4} = 113.4 \text{ foot-pounds}. \]

But, further, let the steam, after expanding to 8.656 lb. pressure per inch, be exhausted to 2 lb. back-pressure in the cylinder during the return-stroke. To find the additional work, \( v_2 (p_2 - p_1) \), thus done, \( v_2 = v_1 \times 10 = 43351 \times 10 = 43351 \) cubic feet; and \( p_2 - p_1 = 6.656 \) lb. per square inch; then the additional work is

\[ 43351 \times 6.656 \times 144 = 14540 \text{ foot-pounds}; \]

and the total available work is 142200 + 14540 = 156740 foot-pounds for 1 pound of steam; or, for each unit of heat expended, it is

\[ \frac{156740}{1254.4} = 125.4 \text{ foot-pounds}. \]

It appears that, in this case, nearly 11.4 per cent. of the whole heat consumed within the cylinder would be supplied to the steam while expanding, to maintain it in a state of dry saturation; and the efficiency of the engine is measured by the fraction \( \frac{146.4}{772} = \frac{1}{5.3} \) or 18 per cent., showing that 82 per cent. of the heat expended is rejected.

Formulas have been constructed for the work of high-pressure steam, worked expansively by means of the link-motion, based on the indicator-power of locomotives, detailed in the work on Railway Machinery. The cylinder from which the data were obtained, was 18 inches diameter, 24 inches stroke, and had 1.8 inches total clearance, or \( \frac{1}{20} \)th of the stroke at each end, with steam-ports 13 by 2 inches, or about \( \frac{1}{20} \)th of the piston-area. The cylinders were placed in the smoke-box, and the steam maintained in good condition. The maximum pressure in the cylinder varied from 70 to 90 lb. per square inch above the atmosphere. The volume of steam admitted was estimated in terms of the indicator-pressure at the point of suppression, and is now recalculated from the most recent investigations of relative volume. The formulas give the nett performances, allowing for drawbacks of every kind, as clearance, back-pressure, wire-drawing; and it is remarked that the incidental advantage by the increased wire-drawing of steam at greater speed during admission, is practically equivalent to the loss by increase of back-pressure. Thus, the formulas are applicable with accuracy for all the different observed speeds of piston, from 200 to 800 feet per minute.

Let \( r \) = the percentage of the stroke at which the steam is cut off, then the consumption of steam in lbs. per horse-power per hour

\[ = 24r + 15; \]

and dividing the result by the evaporative ratio of the fuel, the quotient is the consumption of fuel in lbs. per indicator horse-power per hour.

Again, as one actual horse-power is equal to 1,980,000 foot-pounds per hour, the work of 1 lb. of steam

\[ \frac{1,980,000}{(24r+15)} \]

The effective mean pressure in lbs. per square inch in the cylinder, for a given maximum pressure \( P_1 \) and percentage of admission, \( r \), is

\[ \text{Eff. mean pressure} = P \times \frac{135\sqrt{r}}{r - 28}. \]

The above three formulas are applicable without material error for periods of admission from 10 to 80 per cent., and for pressures of 60 lb. to 150 lb. per square inch. For high pressures, the results they give are slightly under the actual results,—a negative error. The effective mean pressure, by the last equation, is slightly too small for the lower speeds, and slightly too great for the higher, but is exact for a speed of piston of 560 feet per minute.

**CHAP. VI.—THE WORK OF SATURATED STEAM UNDER NORMAL CONDITIONS, WITH EXPANSION.**

In the two preceding sections the work of steam in a cylinder has been considered, first, as gaseous steam, where the steam has been super-heated above its temperature of saturation, and maintained at a uniform temperature; second, as dry saturated steam, where it receives, during expansion, just as much heat as is sufficient to prevent the condensation of any portion of it. It remains to consider the action of steam, when it neither parts with nor receives heat during any part of the stroke. This condition supposes a normal state of things, and implies that the material of the cylinder is a perfect non-conductor, and that the water of condensation during expansion is discharged with the steam at the end of each stroke. Such a condition of things does not, and perhaps cannot, exist in practice, but the consideration of it as an elementary question will be useful.

It has been stated that, whatever the fluid employed, the proportion of heat utilised, in working over the same range, or between the same pair of temperatures, is the same, whatever the fluid employed, when all the heat derived from the heater or boiler is communicated to the fluid at the maximum temperature, and when the entire fall of temperature is effected by the useful expansion of the fluid. On these principles it is necessary, first, that the substance should be raised from the temperature of the condenser to that of the boiler, without directly abstracting heat from the source of heat,—for example, the feed-water should be heated independently to the temperature in the boiler; second, that the fluid should be expanded down to the temperature of the condenser. Under these circumstances, the ratio \( \frac{T_1 - T_2}{T_1} \) expresses the efficiency of the heat expended, or the proportion of heat utilised, and the remainder, \( T_1 - T_2 \), being the measure of the heat utilised, and \( T_2 \) the measure of the heat rejected, the relative proportion of the latter can never, under any conceivable circumstances, be less than \( \frac{T_1}{T_2} \). For example, let the absolute temperature of the condenser, \( T_2 = 600^\circ \), and that of the boiler \( T_1 = 800^\circ \); then the heat utilised cannot be more than \( 800 - 600 = 200^\circ \), or the fraction \( \frac{1}{4} \), and the heat rejected cannot be less than \( 600^\circ \), or the fraction \( \frac{3}{4} \). The ultimate maximum of work, in foot-pounds per unit of heat expended, would, with any vapour or gas, equal \( 772 \times \frac{T_1 - T_2}{T_1} \); and suppose, again, \( T_1 = 718^\circ \), and \( T_2 = 563^\circ \); the heat utilised would be \( 718 - 563 = 155^\circ \), and there would be \( \frac{155}{718} \) parts of the heat converted into work, and \( \frac{563}{718} \) parts rejected, or transferred to the condenser. The work done for each unit of heat expended would be \( 772 \times \frac{155}{718} = 166.6 \) foot-pounds. This is for a perfect heat-engine.

Now, reverting to the steam-engine, whilst the maximum of duty is obtained by the expansion of the steam within the cylinder down to the pressure in the condenser, prior to its being discharged, it is obviously impossible to comply with the first condition of a perfect engine, namely, the heating of the feed-water to the temperature in the boiler, without abstracting heat from the furnace, or from the boiler itself; the only other source of heat is in the condenser, which is only at the temperature due to the pressure—that is, for say 1 lb. absolute pressure, the sensible temperature is 102° Fahr., and the feed-water supplied from the condenser cannot exceed this temperature. The heat expended, under such circumstances, must therefore be increased by a quantity arising from the inferiority of the sensible temperature of the feed-water at 102°, to that of the steam in the boiler, expressed by \( T_1 - T_2 \) degrees; so that, if the water at 102° be converted into saturated steam at 102°, and the temperature and pressure of the resulting steam be raised to those of the saturated steam in the boiler, though \( T_1 - T_2 \) or 155°, the quantity of heat thus absorbed is measured by \( 305(T_1 - T_2) \), or say, to allow for the varying specific heat of water, \( 3(T_1 - T_2) \). The denominator of the ratio of efficiency must therefore be increased by this quantity; and putting \( h \) the greatest proportion of the heat expended, it is possible to convert into work, by expanding the steam down to a temperature of 102°, with the condenser at this temperature, and an absolute pressure of 1 lb. per square-inch. Then we have Mr Brownlee's formula,

\[ h = \frac{T_1 - T_2}{T_1 + 3(T_1 - T_2)}. \]

Let, for example, as before, \( T_1 = 718^\circ \), for a total pressure of 337 lb. per square inch; then \( T_2 \) being 102° sensible, or 563° absolute, \( (T_1 - T_2) = 155^\circ \), and

\[ h = \frac{155}{718 + (3 \times 155)} = \frac{155}{764.5} = 0.2027, \]

or about one-fifth, being the ratio of efficiency; and the work done for each unit of heat expended is \( 772 \times 0.2027 = 156.5 \) foot-pounds.

The comparison, in the particular cases under consideration, between the efficiency of the most perfect engine conceivable, and that of an ordinary condensing engine, working under the conditions specified, may thus be stated:

| Extreme absolute temperatures | Extreme total pressures | Work done per unit of heat expended | |-------------------------------|-------------------------|-----------------------------------| | Perfect engine,...718° to 563° | 33-7 lb. to 1 lb. | 166.6 foot-pounds. | | Ordinary do...,718° to 563° | 33-7 lb. to 1 lb. | 166.5 |

It would appear, then, that the ordinary engine would yield 10.1 foot-pounds, or only 6 per cent. less duty than the perfect engine, when working over the same range, or between the same pairs of temperatures.

It has been said that, in a perfect engine working between the same temperatures, the total work performed for a given expenditure of heat is the same, whatever the substance or vehicle of the heat may be. But in ordinary working,—confining the comparison to condensing engines for the present,—the liquid is supplied directly from the condenser at a lower than the initial temperature; and the work done with different liquids is not the same for equal quantities of heat expended, although working similarly between the same pair of temperatures, and regularly expanding from the highest to the lowest. A steam-engine, for instance, with the boiler at 212°, and the condenser at 104°, the steam expanded in the cylinder from the former to the latter temperature, and the feed-water supplied at 104°; would convert into work 1526 of the heat expended, represented by unity. With a sulphuric ether engine, under the same conditions, the maximum utility would be expressed by 1·406. In an air-engine, let the air be supplied to the heater at the constant temperature 104°, and therein raised at constant pressure from 104° to 212°; thence admitted to and expanded upon a piston till the temperature falls to 104°, and then discharge at this temperature, then it cannot utilise more than 0·9 of the heat expended.

In order to compare, by an example, the action of steam in an ordinary engine, where it neither gains nor loses heat while expanding with its action in a jacketed cylinder,—being supplied with just as much heat while expanding as to prevent condensation of any kind in the cylinder—the example already detailed, in illustration of Mr Brownlee's formula for the action of steam in a jacketed cylinder may be assumed. It was found in that case, that 121·6 foot-pounds of work was done for each unit of heat expended, with initial steam of 100 lb. total pressure per square-inch expanded to 10 volumes, with a final pressure in the cylinder of 8·656 lb., and the same pressure in the condenser. The initial and final absolute temperatures due to the pressures were 788·6° and 646·5°; then \( T_1 - T_2 = 142·3 \), \( T_1 + 3 (T_1 - T_2) = 831·5 \); \( 142·3 / 831·5 = 171·1 \); and \( 772 + 171·1 = 131 \) foot-pounds of work per unit of heat expended, when heat was neither given to nor taken from the steam in the cylinder, to be compared with 121·6 foot-pounds in the jacketed cylinder. Again, for 1 pound of steam, the expenditure of heat in the act of expansion was found to be 1028 units, exclusive of the heat derived from the jacket; and 1028 x 131 = 134,700 foot-pounds would be the total action of 1 pound of steam in the unjacketed cylinder, to be compared with 142,200 foot-pounds, which comprises the additional work arising from the absorption of heat from the jacket. Hence it appears that, in the unjacketed non-condensing cylinder, there was less work per pound of steam, but more work per unit of heat expended, than in the jacketed cylinder. Of course this comparison, though just, is hypothetical, as it supposes that no part of the heat of the steam is absorbed or wasted through the cylinder, which is not a state of things that can be said ever to occur in practice. The jacketed cylinder, then, is not essentially better than the unjacketed, but is so only because of practical imperfections arising from the conducting power of metals, in the latter case, and the consequent precipitation and unprofitable re-evaporation of steam.

SECTION VII.

CHAP. I.—COAL:—ITS COMPOSITION AND DECOMPOSITION.

COKE.

The combustion of fuel consists in the chemical union of the combustible elements with atmospheric oxygen, and the practical management of the process consists, therefore, in so disposing of the fuel in layers on the grate, and so administering the air as to effect their intimate mixture in order to effect their final combustion. The process is simple or complex, according to the composition of the fuel. Coke consists, substantially, of but one combustible, carbon, and the whole process is comprised in the union of the carbon with oxygen in the proper saturating proportion. But coal is not a homogeneous substance; the combustible matter of it is found, by analysis, to consist mainly of carbon and hydrogen:—carbon, dull and unevaporable;—hydrogen, the most elastic gas known:—at the extreme of the scale of existence. The question of the prevention of smoke is intimately associated with that of the combustion of coal. They are, indeed, essentially one question; for, if coal be completely burnt, there cannot be any smoke; and, otherwise, if there be smoke, the coal is certainly not completely burnt. Now, as coke and anthracite coal give no smoke during combustion, although composed of the very matter of smoke—carbon,—it follows that the property of smoke-making is in some way associated with the hydrogen of the coal, which is at the same time the most perfect type of gasosity known in chemistry. The constituent hydrogen is associated in chemical union with a portion of the carbon, forming with it a complex group of compounds, known collectively as hydro-carbons, which present themselves in various forms, when separated or distilled, by the application of heat. At the lowest temperature, the products obtained are chiefly oils, resins, and like distinctive compounds, vaporizable at temperatures under red heat. A somewhat higher temperature brings off fluids of volatile character, as naphtha. A higher still, the third stage of temperature, produces the rich illuminating gas, olefiant gas, or bi-carburetted hydrogen. The fourth stage discharges the common gas, carburetted hydrogen, which continues to be given off after the coal has reached low red heat. But as the temperature rises, pure hydrogen also is given off; until, finally, in the fifth and last stage, hydrogen gas alone is discharged. What remains is the fixed or solid carbon of coal—the coke; with earthy matter—the ash of the coal.

These hydro-carbons, especially those which are given off at the lowest temperatures, which are richest in carbon, constitute the flame-and-smoke making part of the coal; and the greater their entire weight, as compared with that of the fixed carbon, the more highly is this character developed. When subjected to degrees of heat much above the temperatures required to vaporize them, they become decomposed, and pass successively into more and more permanent forms, by precipitating portions of their carbon. At the temperature of low redness, none of them are to be found, and the olefiant gas is the densest type that remains, mixed largely with carburetted hydrogen and free hydrogen. It is in these transformations that the great body of smoke is produced, when the precipitated carbon passes off uncombined; even olefiant gas, at a bright red heat, deposits half its carbon, changing to carburetted hydrogen, and this gas may deposit the last remaining equivalent of carbon, at the highest furnace-heats becoming pure hydrogen.

Throughout all the primary and secondary conditions of the hydro-carbon compounds, raised by distillation from coal, the hydrogen maintains the first claim to the oxygen present above the fuel; until it is satisfied, the precipitated carbon remains unburnt.

There are very great individual differences in the chemical composition and properties of coals, and their varieties are very numerous. The proportion of fixed carbon in coal ranges from 30 to 93 per cent.; of hydro-carbons, from 5 to 58 per cent.; of water, or oxygen and hydrogen, in the proportions to form water, from a mere trace to 27 per cent.; and of ash, from 13 to 26 per cent. The varieties of coal may be arranged into five classes:—

1. Anthracite, or blind coal, consisting almost entirely of free carbon. 2. Dry bituminous coal, having from 70 to 80 per cent. of carbon. 3. Bituminous coking coal, having 50 to 60 per cent. of carbon. 4. Long flaming or cannel coal, differing from the last in containing more oxygen; and in some varieties it does not cake. 5. Lignite, or brown coal, containing 27 to 50 per cent. of carbon.

The following summary presents the mean composition and characteristics of English, Welsh, and Scotch coals, derived from the Report on Coals suited to the Royal Navy, by Dr Playfair and Professor De La Beche:— It appears from this summary that the composition of British coals averages about 80 per cent. of carbon, 5 per cent. of hydrogen, 8 per cent. of oxygen, 1½ per cent. of nitrogen, 1½ per cent. of sulphur, and 4 per cent. of ashes. Also, that the coke, or fixed carbon, as distinguished from the volatilized carbon, averages a little over 60 per cent. of the weight of the coal; the volatilized carbon averaging 20 per cent. Of the volatile portions of coal, then, generated in the preliminary stage of decomposition and distillation, prior to combustion, those which arise in chemical combination are, say,

Carbon........................................... 20 per cent. Hydrogen........................................ 5 " " Oxygen.......................................... 8 "

One part of the hydrogen is united to the 8 parts of oxygen, in the chemically combining proportions forming steam, and of the remaining 4 parts of hydrogen, 3½ parts unite in chemical proportion with the 20 parts of carbon, forming carburetted hydrogen, and ½ of one part passes off as pure hydrogen. The elements of the decomposed fuel, prior to combustion, may thus be summarised:

| 100 lbs. of Coal. | |-------------------| | Fixed carbon...... | 60 | | Volatilized ditto.. | 20 | | Hydrogen........... | 5 | | Oxygen............. | 8 | | Nitrogen, ash, &c. | 7 |

Coke exists in various degrees of purity. The purest coke contains about 97½ per cent. of carbon, according to the following analysis of the best Newcastle coke:

- Carbon................................. 97.6 per cent. - Sulphur................................. 0.85 " - Ashes.................................... 1.55 " - Total.................................... 100.00

Coke is classified into three kinds: good, middling, bad; of which the good contains an average of 94 per cent. of carbon; the middling 88 per cent.; and the bad 82 per cent. Coke usually contains more sulphur than coal, and that is the most injurious element in its effects on the metallic substances exposed to the products of combustion.

Peat, or turf, is in considerable use in Ireland. Perfectly dry peat, of the best quality, contains, in 100 parts—

- Carbon................................. 68 per cent. - Hydrogen............................... 6 " - Oxygen................................. 31 " - Ash...................................... 5 "

In its ordinary state of dryness, it contains 25 to 30 per cent. of water.

**CHAP. II.—THE CHEMISTRY OF THE COMBUSTION OF COAL.**

One pound of hydrogen unites with and requires 8 lb. of oxygen for its combustion; measuring by volume, 1 cubic foot of hydrogen requires just half a cubic foot of oxygen for combustion; the product being steam, aqueous vapour, or water. Oxygen is sixteen times as weighty as hydrogen, and so hydrogen combines with eight times its weight, and but half its volume, of oxygen. In round numbers, 1 pound of hydrogen is 200 cubic feet in bulk, at 62° Fahr., and the combining volume of oxygen is 100 cubic feet.

One pound of carbon unites with 2½ lb., or 32 cubic feet, of oxygen for its complete combustion, forming carbonic acid.

Atmospheric air is composed of oxygen and nitrogen, in the proportion of 1 lb. of the former to 3½ lb. of the latter; or, by volume, 1 cubic foot of oxygen to 4 cubic feet of nitrogen. Nitrogen is a neutral gas in combustion, and is present as a diluent simply; and for every cubic foot of oxygen required in combustion 5 cubic feet of air must be supplied.

It follows, that for the combustion of 1 lb. of hydrogen 500 cubic feet of air are required; and for the complete combustion of 1 lb. of carbon, 160 cubic feet of air are required.

These are the combining proportions of hydrogen and carbon with oxygen, and air, in combustion. In practice, the presence of an excess of oxygen above that which is chemically appropriated, is essential to the completeness of the combustion of the volatilised portions; but the amount of excess necessary becomes less as the general temperature in the furnace becomes greater. It is not needful, for present objects, to entertain the question of excess of air specifically, nor the relative demands of the varieties of hydro-carbons generated from coal. It suffices to show, generally, the proportions of air required for full chemical union with the volatile and the solid portions of the fuel; and thus illustrate the relative importance of the claims of the gases upon the general oxygen fund. It was shown that the average proportion of volatilised hydrocarbons was 23½ per cent. by weight of the whole body of coal, of which the hydrogen constituted 3½ per cent. and the carbon 20 per cent.; and there remained 60 per cent. as solid carbon. For illustration, take 100 lb. of coal; then the relative quantities of air chemically consumed in completely burning the combustible elements are as follows:

- Volatile Hydrogen, 4 lbs. consumes 2000 cubic feet of air. - Carbon, 20 do. do. 3200 do. do.

Fixed Carbon, 60 do. do. 9600 do. do.

85 do. do. 14,800 do. do.

It may be assumed, in round numbers, that, for the complete combustion of 100 lb. of coal, that is, of its combustible elements, 15,000 cubic feet of air is chemically consumed, or 150 cubic feet for 1 lb. of coal. And, of this supply of air, the volatile and fixed elements consume respectively, for the volatile, about one-third, and for the fixed, two-thirds.

If allowance be made for the excess of air practically required for the complete combustion of the gases, say double the chemical equivalent, the total supply of air required for the combustion of 100 lb. of coal would be as follows:

Volatile elements, 10,400, or say, 10,000 cubic feet of air. Fixed element, 9,600, or say, 10,000 do., do.

Total for 100 lb. of coal, 20,000 cubic feet of air.

Thus, finally, it is estimated, that for the complete combustion of 100 lb. of coal of average composition, 15,000 cubic feet of air are chemically consumed; and that 20,000 cubic feet of air are required in practice, or 200 cubic feet of air for 1 lb. of coal, of which one-half is devoted to the fixed portion of the fuel, and one-half to the volatile portion.

The importance of the share in the business of the furnace, taken by the volatilised parts of the fuel, evinced by the large proportion of air allotted to it, is enhanced by the reflection, that the development of heat by combustion is generally in the ratio of the quantity of oxygen chemically combined in the process; and that thus the heat developed by the complete combustion of the volatile elements is one-third of the entire quantity of heat generated. There is, then, scope for the economization of fuel, as well as for the prevention of smoke; but, as there is no doubt that much of the air consumed in burning the volatile elements is drawn through the grate, in company with that which is devoted to the solid portion, there remains so much the less fresh air to be thrown in above or beyond the fuel. There are, doubtless, empiricists who do not believe in the increase of economy by smoke-prevention. Unquestionably, however, when the prevention of dense smoke is not accompanied by a material saving of fuel, there must be a want of adjustment of the appliances, and necessarily imperfect combustion, or waste of means. Carbon, as already stated, may combine with oxygen in one of two proportions, forming carbonic oxide and carbonic acid, of which both are colourless; but, in the production of the former of these, much less heat is developed than in that of the latter.

Besides the combustibles, hydrogen and carbon, the other elements of coal—oxygen, in union with its equivalent of hydrogen, with nitrogen and sulphur—are driven off in the gaseous form. Ammonia is the product of the union of hydrogen and nitrogen, and may possibly be driven off direct from the coal. However the chemical details of combustion may be developed, it is certain that hydrogen from coal practically monopolises a considerable portion of atmospheric oxygen; for the supply of which, therefore, to the hydrogen, provision must be made without prejudice to the requirements of the fixed carbon, in order to effect complete combustion.

CHAP. III.—THE HEAT OF COMBUSTION OF COAL AND COKE.

The total heat of combustion of the elements of coal, simple and compound, has been the subject of elaborate experiment by several physicists, amongst whom are MM. Favre and Silbermann, who employed in their observations the mercurial calorimeter, an instrument from which great accuracy may be expected. The following table of the results arrived at by the above-named experimentalists shows the total heat of combustion with oxygen, of one pound of each of the substances named, in British thermal units, and also in pounds of water evaporated from 212° Fahr., with the weight of oxygen required to combine with one pound of each of the combustibles.

The result arrived at by MM. Favre and Silbermann, as the total heat of combustion of carbon, is higher than that of Dulong, 12,906 units, and that of Despritz, 14,010 units; but it is preferred, as the highest result in experiments of this kind is the most likely to be correct.

It is to be observed, that the imperfect combustion of carbon, making carbonic oxide, produces less than one-third of the heat yielded by the complete combustion of it; also, that the heating power of hydrogen gas is 4½ times that of carbon. It may be deduced, further, that the total heat of combustion of any compound of hydrogen and carbon is the sum of the quantities of heat which the hydrogen and carbon contained in it would produce separately by their combustion. It has also been inferred from other experiments, that the presence of oxygen and hydrogen in fuel, in the proportion to form water or steam, does not affect the total heat of combustion; and that it is only the excess of hydrogen that can be made serviceable as a source of available heat. It would appear, however, that one essential element in the question of the total heat of combustion of coal has not yet been determined, namely the quantity of heat absorbed from the general stock, or rendered latent, in gasifying the hydro-carbons, which is to be charged against the hydrogen, as the evaporative efficiency of the hydrogen must be measured by the quantity of heat developed in its union with oxygen, minus the heat absorbed in the preliminary gasification of it. As Professor Rankine concisely puts it, "the heat obtained is the excess of the heat produced by the combinations, above the heat which disappears in consequence of the decompositions." Sometimes, also, the heat produced is subject to a further deduction, on account of heat which disappears in melting or evaporating some of the substances which combine, either before or during the act of combination." One thing is clear, that in order to make the best of it, the hydrogen, once volatilised, should be oxidized, as well as the carbon associated with it, in order to realize the large measure of heat generated by their combustion. There is a favourite theory, having at least the merit of simplicity, that the heating power of coal is just equal to that of the coke derived from it, which is manifestly absurd. It would be nearer the truth to say, that the heating power of coal is measured by that of its constituent carbon. The evidence of the evaporative powers of coals, abstracted in the table, page 629, appears to support this mode of estimation, in so far as the evaporative efficiency varies generally with the percentage of constituent carbon. The percentages of constituent hydrogen vary within narrow limits, and do not afford data for any marked comparison; but it may be suggested, that generally, the evaporative efficiency is less, as the constituent hydrogen is greater in quantity. Neither the variations of the hydrogen nor those of the carbon, however, suffice to account for the comparatively wide differences of efficiency; but, on referring to the next column, of the constituent oxygen, it is remarkable, that the efficiency of the fuel decreases regularly as the percentage of oxygen in the fuel increases. Welsh coal, with about 84 per cent. of carbon, and 4 per cent. of oxygen, evaporates 9-05 lb. of water per lb. of fuel; whilst Derbyshire coal, with about 80 per cent. of carbon and 10 per cent. of oxygen, evaporates only 7-58 lb. of water per pound of fuel. The difference of carbon does not sufficiently account for the difference of evaporative efficiency; nor does the difference of hydrogen, which is practically the same in both cases. The prime cause, apparently, is the oxygen, which is in great excess in the inferior coal; and an explanation readily occurs. All this oxygen must, in the first place, be volatilised, and it must absorb a portion of heat, which is thus diverted from the business of evaporation; and though, no doubt, it may subsequently restore the heat thus temporarily abstracted in combining with the hydrogen as a gas, yet, as compared with the atmospheric oxygen, which, in the absence of solid oxygen, supplies its place, the solid oxygen is at a disadvantage, in so far as atmospheric oxygen is yielded at once in the half converted, desirable condition of a gas.

It appears, then, that the evaporative efficiency of coal varies directly with the quantity of constituent carbon, and inversely with the quantity of constituent oxygen; but that it varies, not so much because there is more or less carbon, as, chiefly, because there is less or more oxygen. The percentages of constituent hydrogen, nitrogen, sulphur, and ash, are practically constant, with individual exceptions, of course, and their united influence should be so also. Practically, then, treating the question as one of evaporative efficiency, the solution of it lies between the carbon and the oxygen.

The theoretical estimate of the calorific value of coal of average quality, namely, the evaporation of 14-67 lb. of water, from and at 212°, is confirmed by the results of an apparatus constructed by Mr Wright of Westminster; so contrived, that a portion of coal is burned under water, and the products of combustion actually passed through the water, so that the whole of the heat generated is absorbed. By means of this apparatus, the following calorific values, or total heats of combustion, were obtained, the particulars of which were published in the Reports on the Use of the Steam Coals of the Hartley District of Northumberland, by Messrs Armstrong, Longridge, and Richardson:

| Coal Type | Total Heat of Combustion per Pound | |---------------|-----------------------------------| | Welsh coal | 14-30 lb. from 212° | | Hartley coal | 14-63 lb. |

Mean (by experiment).............. 14-46 lb. By theoretical estimate.......... 14-67 lb.

The total heat of combustion of coke is deducible directly from that of carbon, which constitutes the combustible matter of coke. The percentage of carbon in coke determines the percentage of its heating power. Thus:

| Coke | Constituent Carbon | Total Heat of Combustion per Pound | |--------|-------------------|-----------------------------------| | Good | 94 per cent. | 13,620 units | | Middling | 88 | 12,700 ,, | | Bad | 82 | 11,890 ,, |

Total Evaporative Power in Pounds of Water from 212°

| Coke | Constituent Carbon | Total Heat of Combustion per Pound | |--------|-------------------|-----------------------------------| | Good | 94 per cent. | 14-0 | | Middling | 88 | 13-2 | | Bad | 82 | 12-3 |

CHAP. IV.—PHYSICAL CONDITIONS OF THE COMPLETE COMBUSTION OF COAL.

It has been seen that coal undergoing combustion is exhibited in two forms, solid and gaseous, of which the solid (coke) rests on the grate, and the gaseous (hydro-carbons and hydrogen) rise from the solid portion. To get the air into immediate contact and mixture with these elements, so as completely to burn them individually and in detail, is the important problem, the solution of which has been the study of engineers. The mode of so introducing and mixing the air depends on the circumstances of the furnace. If the grate be very large, and the combustion very slow and uniform, as is usual in the practice of Cornish pumping-engines, all or nearly all the air needed for effecting the complete combustion of the fuel may be passed through the grate. If, on the contrary, the grate be small, and the combustion rapid and irregular, as in locomotive practice, a large proportion of the air needed to accomplish the entire combustion of the fuel must be introduced above the fuel, there to mix with and consume the gases. The introduction of air through the grate is, in ordinary practice, the fundamental condition; the admission of air otherwise above the fuel is auxiliary or supplementary to it, supplying just the additional quantity of air requisite to complete the combustion.

Again, the temperature should be maintained at a sufficient elevation, or, more correctly, it should not be lowered by external causes, during the combustion of the hydrocarbon gases, in order to effect the union of the carbon element with its full proportion of oxygen. In chemical order, the hydrogen discharges the associated carbon, and unites with oxygen; by this union intense heat is generated, which envelopes the separated carbon-particles, and raises them to a white heat. Becoming thus luminous, and being the matter of flame, the carbon, in its intensely heated state, is prepared to unite with its saturating proportion of oxygen, and the union is effected the instant they meet, should the carbon retain its temperature until it gets into contact with oxygen. Upon this contingency depends the final condition of the precipitated carbon, whether as unburnt, uncombined particles, the colouring matter of smoke, or as the product of combustion, carbonic acid. Should the carbon-particles miss the opportunity of uniting with oxygen whilst yet at the high temperature which qualifies them to unite, it becomes practically impossible to restore them to the combining condition, and they inevitably pass away as smoke.

It follows, further, as a condition essential to the complete combustion of coal, that the combustible gases should be thoroughly mixed with their supply of air. When the streams or columns of hydro-carbon gases rise, undisturbed or unbroken, from the body of the fuel, they are decomposed in bands or films at what may be conceived as their surfaces of contact, when the oxygen of the surrounding air unites, in the first place, with the hydrogen of the decomposed distillation, and, in the second place, with the carbon-particles. So far the combustion is complete. Watery vapour and carbonic acid are generated as the results of the union of the distilled gases and the oxygen. At this stage, however, a contingency arises, and a partial re-action may ensue; for in the ordinary course of the circulation of the gases, the film or stratum of burnt gas mixes with and loses itself amongst the neighbouring hydro-carbon gases, and should there not be present a sufficiency of fresh atmospheric oxygen to continue the combustion in that quarter, the newly formed carbonic acid would be attacked by the hydrogen of the hydro-carbon, and resolved into its elements, of which the oxygen would be appropriated by the hydrogen, and the carbon would be re-precipitated simultaneously with the carbon separated from the newly formed hydrogen. It may be observed, however, that notwithstanding such occasional action and re-action to which the carbon is subjected, and chargeable to the superior affinities of the associated hydrogen, the carbon may be in condition, as at first, to again unite with oxygen, and to be ultimately and completely burned—in the condition, namely, of a sufficiently elevated temperature. It is clear that a continuous process of intermixture is necessary to the completion of the combustion of coal, bringing together successively fresh portions of the combining elements, and throwing the separated carbon in the way of fresh oxygen for its own proper combustion.

The complication that usually characterizes the burning of coal is, then, both physical and chemical; physical, be- Physical cause an intimate mixture, and a suitable proportion of the elements concerned, is essential to the completeness of their conversion; chemical, because, unfortunately for the special object of the furnace, which is to generate heat, the least important element, hydrogen, is precisely that which commands the preference, and must have its share of oxygen before the claims of the staple element, carbon, can be really entertained and satisfied. The hydrogen must be driven off before the main business of the furnace commences. The occasional presence of oxygen and nitrogen in considerable quantity further complicates the process, as they must be volatilised and driven off in the due course of distillation.

Nevertheless, both coal and coke are, in good practice, effectually burned in furnaces properly formed. The completeness of the combustion of coke in locomotive boilers has been proved by chemical and by mechanical analysis. The combustion of coal has been for many years diligently studied and practised by Mr Charles Wye Williams, and has been by him reduced to successful and varied practice. His leading principle, as published in his Treatise on the Combustion of Coal, is the introduction of air above the fuel, to mix with and consume the combustible gases, on the principle of the argand lamp, through a great number of small orifices in the furnace-door, or at the bridge, or otherwise. In its application, of course, the details are various, and in some classes of furnace having powerful draught, as locomotive fire-boxes, it has not been found necessary to subdivide the air so minutely as Mr Williams' practice indicates.

On the system of steam-induced air-currents, the air is introduced in several streams, which, when necessary, are propelled into the furnace amongst the smoke, by means of jets of steam properly directed through the apertures, which are quite efficient in effecting the object.

There is an opinion that there is advantage in burning fuel in furnaces with long flues, slightly moist, and that ash-pits should be supplied with water, from which steam may be generated by the radiant heat from the fire, and passed through the grate. There is no actual gain of heat by this expedient; but it may be that the radiant heat, otherwise lost, is utilised in making steam, which may increase the efficiency of the fuel. The access of water to the fuel lessens the "glow-fire" or flameless incandescent fuel, and increases the quantity of flame by forming carbonic-oxide and hydrogen gases in its decomposition into its elements, oxygen and hydrogen, and the reduction by the oxygen of the carbonic acid already formed in the furnace. The presence of moisture, even in coke, creates flame, and reduces the intensity of the heat in the "glow-fire;" on the same principle of deferred or distributed combustion, moist bituminous coal is found to be most effective in furnaces with long flues, as in Cornish boilers, where they are 100 or 150 feet in length. These considerations point to the distinctive qualifications of coke and coal in furnaces. Whilst, under steam-boilers, coal and coke may be rendered equally efficient in the generation and communication of heat, per pound weight; on the contrary, it has been found, as recorded by Mr Apsley Pellatt, who gives the results of many years' practice, that in glass furnaces, where intense local heat is required, 13 cwt. of coke is practically equivalent to 21 cwt. of coal. Coke is of the most effective value where local intensity of heat is needed, as in glass furnaces, and in others with short flues and rapid draft, where the flame cannot be used. Coal is generally more effective where carried heat is in demand, the process of combustion being deferred and distributed, and the active development of flame and heat being sustained in the flues; and the benefit of a limited proportion of moisture in the coal depends upon the length of the flues and the time allowed for combustion, when the heat taken up in the glow-fire is given out again in the flame.

The circumstances under which coal is burned under steam-boilers are as various as the qualities of coal itself. The rate of combustion varies from 2 lbs. to 120 lbs. of fuel per square foot of fire-grate surface per hour; the combustion may be equally perfect for all rates, and in good practice it is so. The rate depends chiefly on the strength of the draught, which, if by chimney, does not, at its best, enable the furnace to consume above 30 lbs. of coal per square foot of grate per hour. For higher rates of combustion, the blast-pipe or the fan is employed. The rates of combustion in boiler furnaces may be thus classified:

| No. | Description | Rate per Square Foot per Hour | |-----|------------------------------|-------------------------------| | 1 | The slowest rate of combustion in Cornish boilers with Welsh coal | 2 to 2½ | | 2 | The slowest rate of combustion in Cornish boilers with Newcastle coal | 4 | | 3 | Factory boilers | 12 to 16 | | 4 | Marine boilers | 16 to 24 | | 5 | Dry coal, quickest rate | 20 to 23 | | 6 | Coking coal | 24 to 27 | | 7 | Locomotives | 40 to 130 | | 8 | Do. ordinary variations | 50 to 100 |

The temperature of the products of combustion at the instant of their formation varies of course with the quantity of cold air in dilution. Professor Rankine estimates the temperature to be as follows:

| Condition | Temperature (°F) | |---------------------------|------------------| | Fuel, undiluted with air | 4580 | | If diluted with an excess of half the air consumed | 3215 | | If diluted with an excess equal to all the air consumed | 2440 |

The temperature of the escaping products of combustion in the chimney does not usually exceed 600° Fahr.; it is frequently lower, occasionally not above 300°, and at times barely sufficient to boil water. But this sort of excellence is not to be found except with the largest class of boilers and the slowest combustion. In passenger-locomotives on ordinary duty, the temperature usually rises to 600° and upwards, and in goods locomotives it probably reaches to 1000°.

CHAP. V.—THE EFFICIENCY OF STEAM-BOILERS.

Of the total heat of combustion of fuel, perfectly consumed in the furnace, a portion is absorbed through the heating surface of the boiler, and the remainder passes off without being utilised in the formation of steam. The proportion of utilised heat, or the efficiency of the boiler, varies with every circumstance, but it is regulated chiefly by the area of the fire-grate, the area of the heating surface, and the rate of combustion per unit of grate-surface, say per square foot of grate. If the grate be large and the combustion slow, a greater extent of heating surface is necessary to insure the same efficiency, than if the grate were small and the combustion quick, supposing in the two cases the same total quantity of fuel be consumed per hour. The reason is, that the intensity of combustion increases with the rate at which it proceeds per square foot of grate, the general temperature is higher, as there is a less excess of air in dilution, the radiant heat is also greater, and consequently there is a more rapid absorption of heat into the boiler. The importance of the last condition as to radiant heat is proved by the experiments of Peclet, who found that more than half of the heat of incandescent fuel is radiated from the mass. The relative proportions of the grate area, the heating surface, and the rate of combustion, constitute a question of great practical importance. It was investigated with respect to locomotive-boilers, and the results published in Railway Machinery, in 1852, and also in the Proceedings of the Institution of Civil Engineers, in 1853, when it was shown that, in order to insure the same efficiency, the following rela- Efficiency of Steam-Boilers.

For a given area of grate, the total hourly consumption of fuel should vary as the square of the total heating surface; that is to say, for example, if the heating surface were doubled, the total consumption of fuel might be increased to four times, whilst the same evaporative efficiency would be maintained.

2d. For a given extent of heating surface, the total hourly consumption should vary inversely as the area of grate; for example, if the grate-surface were increased to twice the area, the total hourly consumption of fuel should be absolutely reduced to one-half, in order to maintain the same efficiency.

3d. For a given hourly consumption of fuel, the area of fire-grate will vary as the square of the heating surface, in maintaining the same efficiency; for example, if twice the heating surface be employed, the grate may be extended to four times; conversely, if half of the heating surface be removed, the grate must be reduced to one-fourth of its area.

The mutual relations of heating surface, grate-area, and consumption of fuel, thus announced, are remarkable; and they bear upon the well-known sensitiveness of furnaces to superfluous enlargement of grates, in the reduction of efficiency. For it appears, that if a grate be doubled in area, only half the fuel can be burned on it, and only half the steam generated, if the same degree of efficiency, or evaporative power of the fuel, is to be maintained; so that only a fourth of the fuel can be, with equal utility, burned upon a square foot of the enlarged grate. It is apparent, then, that a superfluous size of grate is detrimental to the power of the boiler, unless at a sacrifice of fuel. On the contrary, an extension of heating surface adds in a still greater proportion to the power of the boiler, whilst the same efficiency of fuel is maintained. The general equation embodying these relations is,

\[ F = C \frac{H^2}{G} \]

in which \( F \) = the quantity of fuel consumed per hour, \( H \) = the area of heating surface, \( G \) = the area of fire-grate, and \( C \) = a co-efficient, which is constant for similar boilers, but may vary for different kinds of boilers. Let the fuel be expressed in pounds, and the areas in square feet, and let the efficiency be represented by the evaporation of 9 pounds of water per pound of fuel, then the constant for locomotive-boilers, burning good coke, is equal to 0.154, and the formula becomes

\[ F = 0.154 \frac{H^2}{G} \]

Assuming 10 square feet area of grate, the hourly consumption of fuel, for different heating surfaces, capable of evaporating 9 lbs. of water per pound of coke, are as follow:

| Area of Grate | Area of Heating Surface | Consumption of Coke per Hour | |---------------|-------------------------|-----------------------------| | Square Feet | Square Feet | | | 10 | 450 | 312 | | 10 | 500 | 385 | | 10 | 600 | 554 | | 10 | 700 | 755 | | 10 | 800 | 985 | | 10 | 900 | 1247 |

If the consumption be expressed at per unit of grate area, one square foot, the expression in the general formula becomes

\[ F_i = C \frac{H^2}{G^2} = C \left( \frac{H}{G} \right)^2 = Cr^2 \]

in which \( r \) = the ratio of the heating surface to the grate, and showing that the consumption of fuel, for constant efficiency, per square foot of grate, varies as the square of that ratio; in which case the formula for locomotives, as before, is

\[ F_i = 0.154 r^2 \]

The relative consumption of water evaporated is nine times the weight of coke, or \( 0.154 r^2 \times 9 = 1.384 r^2 \). The following are examples of the relative quantities, reduced to units of measure:

| Ratio of Heating Surface to Fire-grate | Consumption of Coke per Square Foot of Grate per Hour | Evaporation of Water per Square Foot of Grate per Hour | |---------------------------------------|------------------------------------------------------|------------------------------------------------------| | Grate - 1 | 45 | 289 | | | 50 | 346 | | | 60 | 498 | | | 70 | 678 | | | 80 | 888 | | | 90 | 1121 |

This statement shows how very much the active value of the heating surface, per unit of its area, in evaporating water with the same degree of efficiency, may be increased by relatively reducing the grate, or by extending the heating surface itself. These results are based upon the conditions of the ordinary locomotive-boilers, with fire-box and small flue-tubes, about 2 inches diameter, and with sufficient clearance between the tubes, say \( \frac{1}{2} \) or \( \frac{3}{4} \) inch for 160 tubes, and, of course, under good management. The same proportions stand good for coal in the locomotive. The evaporative standard of 9 lbs. of water per pound of fuel is assumed, which would be equivalent to the utilisation of, say, 10,000 units of heat per pound of fuel, or about \( \frac{1}{4} \)ths of the total heat of combustion. For other evaporative standards, though, no doubt, the same general equation is appropriate, different co-efficients would be required; and with respect to locomotive-boilers, it may only now be remarked, that, generally, within certain limits, the greater the rate of combustion on a given area of grate, with a given heating surface, the less is the evaporative efficiency. In locomotive boilers, an evaporative efficiency of above 10 lbs. of water per pound of fuel has been attained. It must be admitted, however, that in general practice, an evaporation of only 8 lbs. of water per pound of fuel is effected; and the efficiency fluctuates within the ordinary limits of 7 lbs. and 9 lbs. per pound of fuel. In locomotive practice the fuel is very commonly consumed at a rate varying from 60 lbs. to 80 lbs. per square foot of grate; but it descends as low as 40 lbs. per foot of grate, with passenger-trains, and amounts to 100 lbs., 120 lbs., and even above that, per foot of grate per hour, in working goods-trains.

It follows, that the "long-boiler" type of locomotive-boiler, introduced and made by Messrs Robert Stephenson and Co., with a moderate grate, and long flue-tubes, is in general the most efficient in evaporative performance.

The general equation of efficient evaporative power is applicable to all classes of boilers, the value of the co-efficient probably differing for each class, and, of course, for each sort of fuel. The doctrine is powerfully illustrated in the practice of Cornish boilers, in which, with very large grates, the rate of combustion, as practised in Cornwall, is only from 2 to 4 lbs. of fuel per square foot of grate. It is by slow combustion, they say, that they obtain their economical results, namely, a high evaporation of water—about 10 lbs. per pound of coal. The system, of course, demands immense boilers, with long flues and extensive heating surface, to absorb at leisure the heat produced so languidly from the grate. The very low rate of combustion practised is in accordance with the nature of the fuel, commonly in small pieces, and in the form of dross; it requires to be deposited in layers on the grate and left undisturbed, so as to burn off gradually and uniformly. If forced, or disturbed, the smoke is disengaged, and there is a loss of efficiency. Thus there is a necessity for a large grate, in order to effect the complete combustion of the fuel; and there is a necessity for a large boiler and extensive surface, in consequence of the large grate, in order to absorb the heat.

Professor Rankine has announced the following approximate formula for expressing the efficiency of a furnace:

$$\frac{E'}{E} = \frac{BS}{S + AF}$$

in which $E'$ denotes the theoretical evaporative power, and $E$ the available evaporative power of 1 pound of a given sort of fuel; $S$ the area of heating surface in square feet per square foot of grate, and $F$ the number of pounds of fuel burned per foot of grate per hour; $A$, a constant, which is to be found empirically, probably proportional approximately to the square of the quantity of air supplied per pound of fuel; $B$, an empirical constant. In applying the formula to different classes of boilers, different values of $A$ and $B$ are proposed; for locomotive-boilers Professor Rankine computes the values of $S$ and $F$ from Mr Clark's formula, and arrives at near approximations for rates of consumption of about 60 lb. of fuel per square foot of grate per hour.

With respect to forcing a boiler, that is, urging the fire by frequent stirring and poking, discharging smoke and wasting fuel—a practice strongly deprecated by Mr Fairbairn—Professor Rankine justly observes that "the economy of fuel depends very much on the proper adjustment of the rate of combustion per square foot of grate to the draught of the furnace;" and that "it is best, in practice, to make the grate-area at first rather too large, and then to contract it by means of fire-bricks, until the smallest area is obtained upon which the required quantity of coal can be burned without incomplete combustion."

**CHAP. VI.—STRENGTH OF STEAM-BOILERS.**

English boiler-plates of iron are usually classed as Yorkshire plates, and Staffordshire, indicating the localities where they are manufactured. Cast-steel plates, also, are successfully manufactured of great strength, tenacity, and toughness; and good boilers may be made of that material. American boiler-plate appears to be made generally of superior quality. The following are the average ultimate tensile strengths of boiler-plate; that is to say, the weights necessary to tear the material asunder, per square inch of section:

| Type of Plate | Per square inch | |---------------|----------------| | Yorkshire iron plate, best quality | 25 tons | | Staffordshire iron plate, do. | 20 " | | American iron plate, do. | 31 " | | Do., ordinary | 27 " | | Cast-steel plates | 40 " |

The breaking or ultimate tensile strengths of welded and riveted joints, making that of the solid plate = 100, are as follow:

1. The solid plate ........................................... 100 2. Scarf-welded joint ........................................ 100 3. Double-riveted double-welt joint ...................... 80 4. Double-riveted lap-joint .................................. 72 5. Lap-welded joint .......................................... 68 6. Double-riveted single-welt joint ....................... 65 7. Single-riveted lap-joint .................................. 60

It thus appears that scarf-welded joints are equal in strength to the solid plate, and that lap-welded joints, or such as have the edges of the plates merely superposed, have only $\frac{2}{3}$ of the strength. Double and single riveting, as the terms imply, signify the union of two plates by two rows or by one row of rivets, respectively. A welt is a band of metal applied over the seam on one side; a double-welt is applied on both sides. The above proportions are to be accepted as correct, for plates not exceeding $\frac{3}{4}$-inch in thickness; for thicker plates the relative strengths are smaller, and in some cases thicker plates are positively weaker at the joints than thinner ones. The "working strength" is measured by the maximum strain prescribed by cautious engineers for practice, and does not usually exceed one-sixth of the ultimate strength, though occasionally one-fifth.

To apply these data for finding the strength of a cylindrical boiler—for example, one of 6 feet in diameter, and Staffordshire plates $\frac{3}{4}$-inch thick; the strength is to be expressed in lbs. per square inch. The united thickness of two opposite sides of the boiler is $\frac{3}{4} \times 2 = \frac{3}{2}$-inch, and there is $\frac{3}{2}$ square inch sectional area of metal to resist the strain on 1 inch length of the boiler. The ultimate strength of Staffordshire plate is 20 tons per square inch, or 15 tons on $\frac{3}{2}$ square inch. In 6 feet diameter there are 72 inches, and 15 tons $\div 72 = 466$ lb. per square inch, the ultimate strength of the solid metal of the boiler expressed in steam-pressure; and 466 $\div 6 = 78$ lb. per square inch, the working strength of the solid metal. If the plates be united by single-riveted lap-joints, of which the strength is only 60 per cent. of that of the solid plate, then $78 \times \frac{60}{100} = 47$ lb. per square inch is the working pressure per square inch.

Large iron tube flues are not capable of resisting so great a pressure externally against a collapse as they can bear internally, and they require careful staying to prevent collapse. Mr Fairbairn has worked out that question successfully, and he finds that the power of resistance of a plain tube to external pressure varies inversely on the unsupported length of the tube, inversely on the diameter, and inversely on the square of the thickness. The application of angle-iron ribs for stiffening long flues, recommended by Mr Fairbairn, is shown in his double-flue boiler, already described, Plate XX.

**CHAP. VII.—RUPTURES AND EXPLOSIONS OF BOILERS.**

A steam-boiler fails, when the internal pressure of the steam overpowers the strength of the boiler to withstand the pressure. The boiler is said to rupture when the failure is not accompanied by a sudden or extraordinary development of elastic force; the material of the boiler giving way, by cracking or splitting open, and affording an outlet for the contained water and steam. The boiler is said to explode when the failure is accompanied by an extraordinary development of elastic force; the boiler being rent and torn asunder at strong places and weak places, frequently without distinction. When merely ruptured, a boiler rests in its place; when exploded, it is usually torn from its bed and projected—more or less entire, or in fragments—to considerable distances from its bed, carrying with it, or repelling, whatever opposes its progress, moving large masses, and affording other evidences of the enormous power suddenly developed as a consequence of the failure of the boiler. Various theories of explosions have been proposed; but though, to a greater or lesser extent, they serve to explain special phenomena, yet, by none of those hitherto proposed has the essential distinction between ruptures and explosions been accounted for. Many of them are based upon considerations derived from the sciences of electricity, magnetism, and chemistry:—electrical discharges, and the decomposition of steam into its elementary gases, oxygen and hydrogen; and the subsequent explosion of the hydrogen. Suffice it for the present to say, that discharges of electricity have only been proved to take place outside boilers, not inside; that steam can only be decomposed at a temperature equal to that of iron at a white heat, and that when decomposed, the oxygen unites with the material of the boiler, forming a metallic oxide, and the hydrogen alone remains. But without a sufficient combining equivalent of free oxygen, which is nowhere to The ordinary explanation of explosions assign them to original weakness of boilers; to weakness produced by gradual corrosion of the material of which the boiler is made; to wilful or accidental obstruction or overloading of the safety-valves; to the sudden production of steam of a pressure greater than the boiler can bear, in a quantity greater than the safety-valve can discharge, of which the primary cause is said to be the overheating of a portion of the plates of the boiler, uncovered by water, and exposed to the intense heat of the furnace, so that a store of heat is accumulated, which is suddenly expended when water is suddenly placed in contact with the overheated plates, in the production of a large quantity of steam at a high pressure. These arguments, of course, point generally to the existence of overpressure—excessive with relation to the strength of the boiler to resist it, as the general cause. Mr Fairbairn, in his Useful Information for Engineers, takes the same view of the causes of the failure of boilers. The argument of simple overpressure, however, backed by the weighty authority of Mr Fairbairn, fails to explain the generic distinction between the causes of ruptures and those of explosions.

The essential distinction appears to be indicated by the symptoms, namely, the non-production of extraordinary elastic force in the case of ruptures, and the obvious production of it in the case of explosions. The phenomena of explosions are to be ascribed to the projectile force acquired by the particles of water and steam within the boiler, by reason of the suddenly developed expansive force of the steam, spontaneously generated in the boiler, concurrently with the sudden fall of pressure produced by the sudden escape of steam from the boiler, at the locality of the original failure; or by its enlargement of volume from any cause, as the collapse of a flue, for example. The particles of water and steam are projected against the shell of the boiler at an enormous velocity, like as many bullets, or small shot, making up in velocity what they want in mass or weight; and they communicate their centrifugal momentum to the material of the boiler, necessarily straining it, and when powerful enough, tearing it asunder, and thus effecting an explosion. Space, or latitude, is required for the generation of such projectile velocity, and the momentum due to the velocity; and if the water were confined within tubes entirely filled by it, there would be no projection of particles, and no explosive momentum, until the level of the water descended below the point of rupture, which explains the fact of the greater safety of water-tube boilers. The action of the projectile force in question is manifested by the bulging outwards of the flue-tubes usually to be observed in exploded locomotive boilers, due to the sudden generation of steam amongst the tubes; and the new theory affords an explanation of the frequent explosions of boilers that occur immediately after the starting of the engine, when, in the first place, the pressure in the boiler is lowered by the sudden admission of steam to the engine, and its rapid consumption by condensation in blowing through condensing engines, succeeded by the immediate spontaneous generation of fresh steam, and a discharge of projectiles within the boiler. The famous explosion on board the Great Eastern steamship may be similarly explained: the feed-water heater consisted of a cylinder surrounding the chimney, forming an annular space, filled with water and steam at a high temperature and pressure. The chimney, being unstayed, collapsed like a flue, under the pressure on its outer surface, and thus occasioned a sudden enlargement of volume within the heater, and a simultaneous reduction of pressure; the steam suddenly and spontaneously generated, expanded against the collapsed flue, and rebounded on the casing, projecting water with it; the moving mass of steam and water thus suddenly arrested, expended its momentum on the casing, and rended it asunder.

Explosions are probably in many cases preceded by ruptures. Circumstances determine whether a rupture shall be followed by an explosion. The limits of this article forbid the extended discussion of such circumstances, but perhaps the occurrence of an explosion, in ordinary cases, mainly depends upon the locality of the rupture, under water, or above water. If under water, water only is expelled, causing very little enlargement of volume, reduction of pressure, or internal commotion, and therefore unlikely to incur explosion. If, on the contrary, the rupture takes place above water, steam issues with the enormous velocity due jointly to its comparative lightness and its tension—1800 to 2000 feet per second—which, of course, would cause the evacuation of the steam-room of an ordinary boiler instantaneously through a very small opening, and, therefore, an instant fall of pressure and an instant generation of steam, and the projection of water and steam at the rate of many hundred feet per second within the boiler.

This new theory, which may be called the projectile theory of explosions, is susceptible of a variety of illustrative, collateral, and confirmatory evidence; and whilst it is compatible with other and more partial theories, it serves to account for much that remains otherwise unexplained, with respect to the various forms of the failure of boilers.

For the prevention of explosions, proper precaution must be exercised in the choice of materials, and the design and construction of the boiler. It should be subjected to frequent and careful inspection, and should have a sufficiently free action of the safety-valves, one or more of which should be placed beyond the control of attendants, and should suffice to liberate surplus steam sufficiently freely to prevent an excessive rise of pressure, under all ordinary circumstances. The sudden production of steam, in excess of the ability of the safety-valves to discharge the surplus, is most likely to be prevented by avoiding the forcing of the fires, which makes the boiler produce steam faster than the rate suited to its size and surface; by a regular, constant, and sufficient supply of feed-water; and by abstaining from the sudden introduction of feed-water, should the plates have become overheated, which would produce an explosion, and by drawing the fires, so as to allow the boiler and its contents to cool down, before refilling it with water.

CHAP. VIII.—DEPOSITS IN BOILERS.

The impurities of the feed-water are precipitated and deposited on the internal surface of a boiler; sometimes as a hard crust of the minerals contained in the water, chiefly sulphate of lime; sometimes as mud or sediment, which settles loosely. Hard deposits, by resisting the conduction of heat, impair the efficiency of the boiler, as well as its durability and safety; and they may be prevented, either by administering chemical re-agents with the water, neutralising the ingredients of the water, or converting them into innocuous or loose compounds, which fall as sediment; or, what is better, by purifying the water before it is admitted into the boiler. The practice of surface-condensation in steam-engines, by which the condensed steam may be collected and returned to the boiler, insures a supply of pure water.

(Besides the authorities mentioned in the body of the foregoing article, the following have been consulted:—The Engineer; Proceedings of the Institution of Mechanical Engineers; Bourne's Catechism of the Steam-Engine.)