12

1

PARTS OF THE GOVERNOR.

shown in section with sizes in fig. 288., slides, and the curved guide, fig. 289., is fixed. From the top of the spindle, fig. 285., the arms, fig. 286., are suspended, with the balls at the end, the arms moving in the slit in the curved guides, fig. 289. The divergence or collapse of the balls causes the collar to slide up or down on the spindle; and through a slit shown in the spindle (which is hollow), this movement is communicated to a rod sliding within it, which, by a suitable attachment, moves the throttlevalve. This arrangement of governor is very neat and convenient, and commends itself to general adoption.

Messrs. Boulton and Watt have supplied us with a statement of the speed of the piston in their engines of different powers, and it is as follows:Feet. Ft. in. Feet. Ft. in. Feet. Feet. Feet. Feet. Feet. 2 26 3 3 6 4 5 6 7 No. of strokes per minute 40 34 0 30 27 O 25 21 19 17 16 Feet travelled per minute 160 170 0 180 189 0 200 215 228 245 256

Length of stroke

In this table the pressure is taken to vary slightly with the size of the engine; but Messrs. Boulton and Watt notify to us that, though such was the manner in which they formerly treated the subject, they now adopt a uniform pressure of 7 lbs. on the square inch, as a preferable element of computation. The effect of this substitution is to make small engines of a somewhat greater nominal power than they formerly were reckoned at, which, since the nominal power is now a commercial unit, as well as a scientific one, is convenient in approximating the price per horse-power of the different sizes. The speed of the piston in feet per minute is about 128 times the cube-root of the stroke; and, as has already been mentioned, the nominal horse-power of an engine may be found by multiplying the square of the diameter of the cylinder in inches by the cube-root of the stroke in feet, and dividing by 47. To find how many millions of pounds are raised 1 foot high by the consumption of a bushel or 84 lb. of coal:Divide 166.32 by the number of pounds of coal consumed per hour by

each horse power: the quotient is the number of millions of pounds raised 1 foot high by the consumption of 84 lbs. of coal. A bushel of Newcastle coal will weigh about 84 lbs., but the Welsh coal is heavier.

If a cubic inch of water be supposed to produce a cubic foot of steam, and the latent heat of steam at 212° be taken, with Mr. Watt, at 960°, or, in other words, the cubic foot of steam be supposed to contain as much heat in the latent form as would raise the temperature of the cubic inch of water, if it could be prevented from expanding, 960°, then the sum of the latent and sensible heats will be represented by 1172°. The temperature of the water discharged by the air-pump is about 100°, which, taken from 1172°, leaves 1072°, which must be taken up by such a quantity of cold water that its temperature will not rise above 100°. If the temperature of the injection water be 50°, then the difference between that and 100°, viz. 50°, is available for the absorption of the heat; and 1072 divided by 50 = 21:44, which is the number of times the quantity of injection water must exceed the quantity of water in the steam. To condense a cubic inch of water therefore in the shape of steam, 21:44 cubic inches of injection water are necessary; but inasmuch as the water may not always be as cold as 50°, Mr. Watt's practice was to allow a wine pint, or 28-9 cubic inches of injection water for every cubic inch of water converted into steam. The capacity of the cold-water pump is usually made from one thirty-sixth to one forty-eighth of the capacity of the cylinder. The injection orifice should have an area of about one fifteenth of a square inch per horse power. The capacity of the hot-water pump should be about one 240th of that of the cylinder, supposing that the engine is double-acting, and the pump singleacting. The air-pump is usually made half the diameter of the cylinder, and half the stroke, or one eighth of the capacity. The power requisite to work the air-pump is from one thirtieth to one fortieth of the power of the engine. The openings through the foot and delivery valves are made of about one fourth of the area of the pump. The internal diameter of the steam-pipe may be found by dividing the horse power by 8, and extracting the square root of the quotient. We shall reserve what we have to say on the subject of bolts until we come to speak of the holding-down bolts of marine engines.

There are many other kinds of parallel motion besides those which we have mentioned, but there are none of them of sufficient importance to justify a lengthened description. Fig. 290. represents a species of parallel

Fig. 290.

motion invented by Mr. James White, and published in his "New Century of Inventions," in 1801. It depends on the principle that an encycloidal curve, formed by one circle rolling within another, becomes a straight line when the diameter of the outer circle is just twice that of the inner one. A large wheel, with teeth on its inner circumference, is fixed on a frame concentric with the axis and circle of the crank. A wheel with external teeth is fixed freely on the crank-pin and the point of attachment of the piston-rod. By this arrangement the small wheel is compelled, by the pressure of the piston-rod upwards, to roll round the great circle, ascending on the one side, and descending on the other, so that the distance of the end of the piston-rod from the point of contact of the circles is always equal to the distance of the circle from the diameter. The fault of this species of parallel motion is, that the socket in the centre of the revolving wheel is exposed to a strain equal to twice that on the piston, and which it cannot conveniently be made long enough to resist, so that it is liable to break or speedily shake loose. In the plate of directaction engines, various modifications of the parallel motion will be observed. In the Gorgon engine, by Messrs. Seaward, the parallel motion is formed by the application of a radius-bar to the air-pump lever, whereby one radius is made to counteract the other the centre of the lever resting upon a jointed pillar, in order to enable the cylinder end of the lever to move up and down in a vertical line. This species of parallel motion is sometimes made with a horizontal slide for the centre to move in, instead of a vibrating pillar

or link; but a slide works slack sideways, and is not satisfactory in practice. The combination might be improved by causing the sliding ends of the rods, which in some cases are used instead of a lever, to work into stuffing-box tubes hung on a centre, so as to enable them to swivel. The rods, so soon as any wear took place, could be tightened afresh by screwing up the packing.

MARINE ENGINES.

Cylinder.-In the marine engine the cylinder-bottom is more frequently cast in, than in land engines, and a plug of metal is fitted into a hole in the centre of the bottom, which is left to allow the boring-bar to pass through. It is necessary that the cylinder should be bolted very firmly to the sole plate, as in engines which exhaust at the under side of the valve-casing, an air-tight joint has to be made between the sole plate and the part of the cylinder-bottom next to the valve. A cylinder of about 6 feet diameter is usually made about 1 inch thick, and the metal should be hard as well as solid. Messrs. Maudslay's practice in side-lever engines is to cast the cylinder-bottom in, up to 60 inches diameter, and above that size they prefer casting the cylinder open at the bottom, and making the bottom out of the sole-plate. A projection is cast on the sole-plate, to go a certain distance into the cylinder, with a space hollowed out for the cylinder-port. The bottom joint should not be of rust, but metal to metal. the bottom flange of the cylinder and the place on which it stands on the sole-plate being both faced in the boring-mill. The cylinder-cover should fit so nicely as to be tight, by interposing a piece of lead or a ring of wire gauze, smeared with white or red lead. In oscillating engines the cylinder-bottom is generally cast in, whatever be the size of the cylinder.

The valve-casing should be attached to the cylinder by means of a metallic joint, or, in other words, by fitting the surfaces so accurately that a little red lead interposed will make them tight. The valve-casing can thus be easily removed at any time to repair the valve faces; whereas, if the joint of the casing be of rust, the removal of the casing is an operation of much difficulty. The attachments of the cylinder to the diagonal stay are very generally made too small; that is, the surface is too small, and the flange too thick. A very thick flange cast on any particular part of a cylinder endangers the soundness of the cylinder by inducing an unequal contraction. It is much the best way to make the flange for the framing thin, and the surface large. The bolts, too, should be turned bolts, and nicely fitted. Some persons make them with a nut at both ends, the body of the bolt being made with a little taper; and the nut which answers to the head is screwed up after the conical part of the bolt has been drawn into the hole by the nut at the point. The object of this plan is to facilitate the fitting; but if the fitting be well done, it is unimportant whether it is done in this way or any other.

Cylinders are not now usually made with steam-casings, yet experiment has satisfactorily proved that there is a loss of power consequent on their relinquishment. It is not very easy to discern the cause of this loss, as there is more radiating surface in the casing than in the cylinder; yet the existence of the loss is very certain. Mr. Watt, in some of his early trials, discontinued the steam-jacket; and he found the consumption of fuel to be materially increased. He therefore again resumed it, but it has been again discarded in most of the modern engines, except those of the Cornish construction. Escape-valves, for letting out any water that may enter with the steam, are now usually employed in marine engines: they may in most cases be applied conveniently to the ports of the cylinder, as shown in the details of engines of the West India packets, and may be kept shut by a spring, in the same manner as the safety-valve of a locomotive. Escape-valves should be placed on that side of the cylinder which is nearest the side of the ship; so that the attendants may not be scalded by the issuing water if the engine primes. The escape-valve is shown beneath the "Plan of Cylinder" in the West India mail engine details; and to those details the remarks which follow are to be understood to refer, except where specified to the contrary. In boring cylinders of 74 inches diameter, the boringbar must make one revolution in about 4 minutes, so that the cutters will move at the rate of about 5 feet per minute. In boring brass the speed must be slower the common rate at which the tool moves in boring brass air-pumps is about 3 feet per minute. If this speed be exceeded the tool will be spoiled, and the pump made taper. The speed proper for boring a cylinder will answer for boring the brass air-pump of the same engine. A brass air-pump of 36 inches diameter requires the bar to make one turn in about 3 minutes, which is also the speed proper for a cylinder 60 inches in diameter. To bore a brass air-pump 36 inches in diameter requires a week, an iron one requires 48 hours, and a copper one 24 hours. În turning a malleable iron shaft, 12 inches in diameter, the shaft should make about five turns per minute, which is equivalent to a speed in the tool of about 16 feet per minute. A boring-mill, of which the speed may be varied from one turn in six minutes to twenty-five turns in one minute, will be suitable for all ordinary wants that can occur in practice.

Piston. We have nothing further to say on the subject of pistons, after the various specimens we have already given. The proportion of taper given to the piston-rod where it fits into the piston, in the West India mail engines is a good one; if the taper be too small, the rod is drawn through the hole, and the piston is split. Small grooves are turned out of this piston

rod above and below the cutter-hole, and hemp is introduced, in order to make the piston-eye tight. Most piston-rods are fixed to the piston by means of a gib and cutter, as shown in the plates of details, but in some cases the upper portion of the rod within the eye is screwed, and it is fixed into the piston by means of an indented nut. This nut is in some cases hexagonal, and in other cases the exterior forms a portion of a cone, which

Fig. 291.

PISTON-ROD.

completely fills a corresponding recess in the piston. But nuts made in this way become rusted into their seat after some time, and cannot be started without much difficulty. Messrs. Miller, Ravenhill, and Co. fix in their piston-rods by means of an indented hexagonal nut, which may be started by means of an open box key. The thread of the screw is made flat upon the one side, and much slanted on the other, whereby a greater strength is secured, without any disposition to split the nut. When pistons are made of a single ring, or of a succession of single rings, the strength of each ring is tested previously to its introduction into the piston, by means of a lever loaded by a TOP OF DON JUAN'S heavy weight. The old practice was to depend chiefly upon grinding, as the means of making the rings tight upon the piston or upon one another; but scraping is now chiefly relied on. A slight grinding, however, with powdered Turkey stone appears to be expedient, which may be most conveniently accomplished by setting the piston on a revolving table, and holding the ring stationary by a cross piece of wood while the table turns round. Pieces of wood may be interposed between the ring and the body of the piston, to keep the ring nearly in its right position; but these pieces of wood should be fitted so loosely as to give some side-play, else the ring will wear itself into a groove on the piston. Messrs. Penn grind their cylinders after they are bored, by laying them on their side, and rubbing a heavy piece of lead, made to the curve of the cylinder, and smeared with emery and oil, backwards and forwards by hand, the cylinder being gradually turned round, so as to subject every part successively to the operation. The pistons are also ground into the cylinders with great care, so that they are perfectly tight from the commencement. Messrs. Penn's piston for oscillating engines has a single packing ring, with a tongue-piece, as in Messrs. Maudslay's and Messrs. Miller's arrangements, figured in pages 197 and 193. The ring is packed behind with hemp-packing, and the piece which covers the joint is made of sheet copper, and is indented into the iron of the ring, so as to offer no obstruction to the application of the hemp. The ring is ground to the piston only on the under edge: the top edge is rounded from the inside to a point, and the junk-ring does not bear upon it, but the junk-ring squeezes down the hemp-packing between the packing-ring and the body of the piston. The metallic packing of the piston consists of a double tier of rings, cut into numerous segments. We approve of the plan of adding a nut to the top of the piston-rod, in addition to the cutter, for securing the piston-rod to the cross-head, as shown in fig. 291., where the piston-rod is 7 inches in diameter, and the screw 5 inches: the part of the rod which fits into the cross-head eye is 1 ft. 5 inches long, and tapers from 6 to 6 inches diameter. The proportion of taper is a good one: if the taper be less, or if a portion of the piston-rod within the cross-head eye be left untapered, as is sometimes the case, it is very difficult to detach the parts from one another, and we have known great inconvenience to be thus occasioned.

Cylinder Cover. The cylinder cover in plate of details is cast close, and a few holes are left for taking out the core by, which holes are afterwards plugged up. An annular recess is left in the under side of the cover, for the accommodation of the heads of the piston-bolts. The gland of the stuffing

stuffing-box of the Don Juan steamer, cylinder 68 inches diameter. This appears to us better than that of the West India packets: there is a great advantage in a deep stuffing-box, especially in the case of vessels intended to perform long voyages. Fig. 293. represents the cylinder cover of a Cornish

Fig. 293.

CYLINDER COVER OF CORNISH ENGINE WITH LANTERN BRASS.

Fig. 294.

engine. The stuffing-box is provided with a lantern brass, into which steam is admitted by a small pipe. There is packing both above and below the lantern brass, the purpose of which is to prevent any leakage of air by the stuffing-box; for even if the packing be defective, it will be steam that leaks in, which is condensable; and such a leak, though it will increase the consumption of fuel, will not diminish the power of the engine. It is the usual practice to interpose between the cylinder cover and the cylinder flange a gasket-ring as a joint; but a joint of this kind leaks air imperceptibly, and it is better to make the surfaces very true, and to interpose either a ring of sheet-lead or a little red-lead putty. It appears expedient to us that all marine engines should be furnished with steam-jackets, and should also be furnished with spaces in the cylinder-cover and cylinder-bottom for the admission of steam. Large engines, too, we conceive should be fitted with the lantern brass stuffing-boxes. Fig. 294. represents the stuffingbox of the Trident, the engines of which are of the oscillating kind by Messrs. Boulton and Watt. The extra depth of this stuffing-box is necessary, to counteract the tendency to wear oval. This tendency, although its existence is undoubted, has not been found to occasion much inconvenience in oscillating engines, although great fears were entertained on that score by the adherents of more antiquated engineering mechanisms. It should be borne in mind also, that even in engines with the ordinary parallel motion the stuffing-box has a tendency to wear oval, which may be perceived if attention be paid to the setting of the parallel motion. If the piston be moved through a stroke, the gland will be found to move easily upon the piston-rod at some points, and be jammed up tight in the stuffing-box at other points of the stroke; inequalities which clearly show the existence of a very sensible deviation from a perfectly vertical motion, The brass dome attached to the gland and embracing the rod, is an excellent addition; it keeps the grease employed to lubricate the rod, from being spilt, and prevents grit and dust from getting into the gland, whereby in common engines the rod is frequently much scratched and injured. Metallic packing in the stuffing-box has been used in some engines, consisting in most instances of one or more rings, cut, sprung out, and slipped upon the pistonrod, before the cross-head is put on, and packed with hemp behind. This species of packing answers very well when the parallel motion is true, and the piston-rod free from scratches, and it accomplishes a material saving of tallow. In some cases a piece of sheet-brass, packed behind with hemp, has been introduced with good effect, a flange, notched to permit the bending, being turned over on the under edge of the brass, to prevent it from slipping up or down with the motion of the rod.

STUFFING-BOX OF OSCILLATING ENGINE.

Scale inch=1 foot.

Slide-valve. The slide-valve represented in the plate of details is that known as the long D. The valve-rod is attached to a cross-bridge in the plane of the under face, and the spring upon the rod is sufficient to allow the valve to be tightened up as the face wears. In short D valves, where the valve-rod is very short, the eye which attaches the valve to the rod has to be made oblong, fig. 295., or else the holes of the casing-cover have to be made oval, so as to enable the valve-cover to be advanced nearer to the cylinder, as the valve face wears. The valve-packings are introduced by doors at the back of the valve-casing, and are pressed by blocks, of which one is shown in the same plate as the valve. This block, it will be remarked, is in three pieces, which are tongued to one another; it is pressed

cast therein. The purpose of this bar is to prevent the strain requisite for tightening the packing from being thrown upon the packing-door, which would spring it out, and might cause the joint of the door to leak. In some cases the screws by which the packing is pressed, pass through the door, and are made tight by a jam-nut, with a recess, into which a turned part of the nut enters, as shown in fig. 297.—a hemp washer being interposed at the point of contact, and this we think is the preferable practice. In the plate, however, a different plan is shown, which is again represented in fig. 298. The packing-screws do not pass through the doors, but are kept short; and opposite to each screw a plug is situated in the packing-door, which has to be withdrawn when the packing is to be tightened. In each of these plugs a small recess is turned out, for the reception of a ring of hemp. This recess is better half-round than square.

Some packing-blocks are tightened sideways, by screws which are inserted in the sides of the packing-ports; but in the plate the block is tightened by its own wedge-shape point, which presses against another wedgeformed piece, cast on the valve-casing, as will be understood by a reference to fig. 299. In some cases the end of the packing abuts upon the cylinder face, but generally it overlaps two or three inches in large engines, and a piece, a, fig. 300., is cast on each side of the cylinderport, in continuation of the circle of the valve, to furnish a surface upon which the packing may press. By this expedient the chance of leakage at the corner of the valve is diminished, and the length of the packing need not be adjusted with such critical exactitude as is necessary by the other arrangement. In some engines the packing of the valve is put in like that of a piston, and is pressed down by means of a junk-ring, but that plan is now little resorted to. Metallic packing has been tried in D valves, but only with very moderate success. The kind that has answered best is a piece of sheet brass, thinned at the ends, bent to the shape of the valve, and packed behind with hemp. We believe the D valve will now give place to the equilibrated valve employed by Messrs. Penn, and of which we have already given a description at page 199.

We have experienced a good deal of trouble with every modification of valve faces; but cast iron working upon cast iron is perhaps the best combination yet introduced. The usual practice is to pin brass faces on the cylinder, allowing the valve to retain its cast-iron face. Some makers employ brass valves, and others pin brass on the valves leaving the cylinder. with a cast-iron face. Speculum metal and steel have been tried for the cylinder faces, but only with moderate success. In some cases the brass gets into ruts; but the most prevalent affection is a degradation of the iron, owing to the action of the steam, and the face assuming a granular appearance, something like loaf sugar. This action shows itself only at particular spots, and chiefly about the angles of the port, or valve face. At first the action is slow; but, once the steam has worked a passage for itself, the cutting away becomes very rapid, and in a short time it will be impossible to prevent the engine from heating when stopped, owing to the leakage of steam through the valve into the condenser. However truly the D valve may be formed at first, the face will become slightly hollow by the application of heat, as the circular will expand more than the straight part, and the packing resists the enlargement of the circle. The cross-section will therefore assume something of the form shown in fig. 301, where the dotted line represents the original position of the face; and on examining a valve newly put in action, it will generally be found that it presses hardest on the tails. The face therefore should be made slightly rounded in the manufacture; and if the engine is a large one, the cylinder must not be faced when lying on its back, unless it has been wedged up to the form it assumes when standing on end, else the partial collapse of the cylinder will cause the face to become untrue. Copper steam-pipes seem to have some galvanic action on valve faces, and malleable iron pipes have sometimes been substituted; but they are speedily worn out by oxidation, and the scales of rust which are carried on by the steam, scratch the valves and cylinders, so that the use of copper pipes is the least evil. The valve-rod in that part opposite to the steam-port is often much wasted by the steam; it therefore appears expedient to surround it by a copper pipe where an injurious action is to be apprehended. The valve-casing shown in the plates is made close at the bottom, the exhaustion being accomplished by the upper eductionpipe. In cases in which exhaustion is performed from below, it is expedient to cast two projections on the sole-plate, to prevent the valve from falling down inconveniently far when the valve links are taken off. There is no expansion joint introduced in the valve-casing of these engines, which is a serious defect, as the steam gains admission to the valve-casing before it can enter the cylinder, and the joints are damaged, and in some cases the cylinder is cracked, by the inequality of expansion of the cylinder and valve-casing.

In facing a valve recourse is had to the use of a face plate, to ensure the accuracy of the work. To ascertain whether the face plate bears equally, smear it over with a little red ochre and oil, and move the face plate slightly, which will fix the colour upon the prominent points. This operation is to be repeated frequently, and as the work advances, the quantity of colouring matter is to be diminished, until finally it is spread over the face plate in a thin film, which only dims the brightness of the plate. The surfaces at this stage must be rubbed firmly together to make the points of contact visible, and the higher points will become slightly clouded, while the other parts are left more or less in shade. If too small a quantity of colouring matter be used at first, it will be difficult to form a just conception of the general state of the surface, as the prominent points will alone be indicated, whereas the use of a large quantity of colouring matter in the latter stages would destroy the delicacy of the test the face plate affords. The scraping tool should be of the best steel, and should be carefully sharpened at short intervals on a Turkey stone, so as to maintain a fine edge. A flat file bent, and sharpened at the end, makes an eligible scraper for the first stages; and a three-cornered file, sharpened at all the corners, is the best instrument for finishing the operation. The number of bearing points desirable on the surface of the work depends on the use to which it is to be applied, but in any case the bearing points should be distributed equally over the surface. Great care must be taken in fitting valve-faces that the valve be not made conical; unless the back be exactly parallel with the face, it will be impossible to keep the packing from being rapidly cut away. When the valve is laid upon the face plate, the back must be made quite fair along the whole length, by draw-filing, according to the indications of a straight edge; and the distance from the face to the extreme height of the back must be made identical at each extremity. Should a hole occur either in the valve, in the cylinder, or any other part where the surface requires to be smooth, it may be plugged up with a piece of cast-iron as nearly as possible of the same texture. Bore out the faulty part, and afterwards widen the hole with an eccentric drill, so that it will be of the least diameter at the mouth. The hole may go more than half through the iron: fit then a plug of cast-iron roughly by filing, and hammer it into the hole, whereby the plug will become rivetted in, and its surface may then be filed smooth. Square pieces may be let in after the same fashion, the hole being made dovetailed, and the pieces thus fitted will never come out.

Brass faces are put upon valves or cylinders by means of small brass screws, tapped into the iron with conical necks for the retention of the brass: they are screwed in by means of a square head, which, when the screw is in its place, is cut off and filed smooth. In some cases the face is made of extra thickness, and a rim not so thick runs round it, forming a M M

step or recess for the reception of brass rivets, the heads of which are clear of the face.

Air-pump.-The air-pump is attached to the sole-plate by a rust faucit joint, which is preferable to a rust flange joint, as the salt-water eats away the heads of the bolts, unless they are copper; and if they are copper, they waste the iron. The oil and grease which fall from the crank-pin upon the sole-plate, deoxidise the rust of a flange joint, whereas with a faucit joint, suitably made, they cannot remain in the same intimate contact. Short steel keys should be driven into the faucit in several places before the joint is made, but they should not rise to the top of the faucit so as to divide the joint into segments.

The air-pump bucket is made with a junk-ring, whereby the packing of the bucket may be easily screwed down. The valve is of the spindle or pot-lid kind. The foot and delivery are of the flap or hanging kind. These valves all make a considerable noise in working, and are objectionable in many ways. Valves of the same construction as those known as Harvey and West's, which are similar to those shown in fig. 242. have been employed with advantage by Messrs. Rennie; and valves on Belidor's construction, which is in effect that of a throttle-valve hung off the centre, were some years ago proposed by us for the delivery and foot-valves. Some delivery-valve seats are bolted into the mouth of the air-pump, apparently in the conviction that the pump-bucket never requires to be looked at. If delivery-valves be put in the mouth of the air-pump at all, the best mode of fixing them appears to be that adopted by Messrs. Maudslay. The top of the pump-barrel is made quite fair across, and upon this flat surface a plate containing the delivery-valve is set, there being a small ledge all round to keep it steady. Between the bottom of the stuffingbox of the pump cover and the eye of the valve-seat, a short pipe extends, encircling the pump-rod, its lower end checked into the eye of the valveseat, and its upper end widening out to form the bottom of the stuffing-box of the pump-cover. Upon the top of this pipe some screws press, which are accessible from the top of the stuffing-box gland, and the packing also aids in keeping down the pipe, the function of which is to retain the valveseat in its place. When the pump-bucket has to be examined the valveseat may be slung with the cover so as to come up with the same purchase. For the bucket-valyes Messrs. Maudslay employ two or more concentric ring-valves, with a small lift. These valves have given a good deal of trouble, in consequence of the frequent fracture of the bolts which guide and confine the rings; but their principle appears to us superior to that of any of the other air-pump valves at present in common use, with the exception of the equilibrated-valve, known as Harvey and West's, in which it is preferable that the face should fall in a groove filled with end-wood. It would not be difficult to make this groove so that the water would have to be forced out of it during the descent of the valve, whereby the shock would be still further diminished. It would be preferable, however, if all these valves could be discarded in favour of a slide-valve, which it appears to us might be applied to the air-pump with much advantage.

The air-pump bucket and valves are all of brass, and the chamber of the pump is lined with copper. It is now a common practice to make the chamber of the air-pump wholly of brass, whereby a single boring suffices. When a copper lining is used, the pump is first bored out, and a bent sheet of copper is introduced, which is made accurately to fill the place, by hammering the copper on the inside. Muntz's metal is sometimes used instead of copper, and Muntz's metal air-pump rods are now as generally used as copper rods or iron rods covered with brass. Iron rods covered with brass are not to be commended; they generally are wasted away where the bottom cone fits into the bucket-eye, and if the casing be at all porous, the water will sometimes insinuate itself between the casing and the rod, and eat away the iron. If iron rods covered with brass be used, the brasscasing should come some distance into the bucket-eye; the cutter should be of brass, and a brass washer should cover the under side of the eye, so as to defend the end of the rod from the salt water. Rods of Muntz's metal are, probably, on the whole to be preferred; and it is a good practice to put a nut on the top of the rod to secure it more firmly in the cross-head eye. The part which fits into the cross-head eye should have more taper when made of copper or brass than when made of iron; as if the taper be small, the rod may get staved into the eye, whereby it will be so firmly fixed as to make its detachment a difficult operation. Metallic packing has in some instances been employed in air-pump buckets, but its success has not been such as to lead to its further adoption.

Sole-plate and Condenser. - Every marine engine, of the side-lever kind, should be constructed with a sole-plate; and we think it the best way that the condenser be cast upon the sole-plate. Engines unfurnished with soleplates, and with joints between the valve-casing and condenser below the level of the keelsons, are extremely objectionable. Those joints, either from the working of the ship-the movement of the cylinder or condenser -the deoxidising effect of the oil spilt about the machinery-or the combination of all these causes-will be sure, sooner or later, to become leaky; and it is almost impossible to remake or effectually staunch them. Messrs. Maudslay and Co., in their West India mail packet engines, bolted the condenser to the under-side of the sole-plate; by which expedient the rust-joints are preserved, in a great measure, from the action of the grease, and from any strain or vibration consequent upon the yielding of the condenser or cylinder. Messrs. Miller and Co. adopted the same arrangement of condenser, but cast the condenser upon the sole-plate.

There are very few of the engines made in Scotland in which the con denser is not cast upon the sole-plate; and in almost all of them the greater part of the condenser is situated above the sole-plate, and the main-centre passes through it. The height of the condenser, in this arrangement, has the advantage of enabling the air-pump to drain it of water very effectually; but the same object is accomplishable by the use of a very large eductionpipe immediately behind the valve-casing into which the injection-water is admitted, and which thus becomes, in effect, a tall condenser. This latter is the arrangement of Messrs. Maudslay and Messrs. Miller. It has the advantage of leaving the space usually occupied above the sole-plate by the condenser, free and unperplexed by any species of machinery except the main-centre, which is supported by pillow-blocks resting or cast on the sole-plate.

The practice of passing the main-centre through the condenser, either with or without a pipe, is objectionable. A pipe is calculated to make the sides of the condenser crack by unequal contraction, and the absence of a pipe endangers a leakage of air round the main-centre joint. The keys employed to fix the main-centre will sometimes occasion trouble, from becoming loose; and, in some instances, we have known a main-centre boss to be split, from the keys being driven too hard. It all cases the thickness of metal requisite in the condenser sides for resisting the strain of the main-centre, will make the sides more liable to crack, in consequence of being suddenly cooled. Upon the whole, the practice of securing the main-centres by plummer-blocks appears greatly preferable: when the main-centre is made to pass through the condenser, the hole should be bored out, and the main-centre ground in with a little taper. It is the usual practice in engines which have the main-centre passing through the condenser, to set the hot well on the top of the condenser, and this is the arrangement in the engines of which we have given the details. A part of the hot well is divided off, to serve as an eduction-passage for the conveyance of the steam from the superior part of the valve-casing. By this arrangement there is no danger of water running from the condenser back into the cylinder. Projections are cast in the foot-valve passage for the reception of the foot-valve seat, by means of which it is keyed into its place; and similar projections are cast in the mouth of the air-pump for the reception of the delivery-valve.

There does not appear to be any manhole-door to the condenser in this engine, which is in our eyes a defect. It would have been easy to make a manhole in the curved nozzle leading from the air-pump to the hot well: a door in that situation would have been easy of access from the hot well manhole, and it would have always been covered with water when the engine was at work, so that a leakage of air could not have taken place. The injection rose-pipe runs across the condenser, near the mouth of the eduction-pipe. A cock, by which water may be injected from the bilge should the vessel spring a leak, is universally employed in marine engines, and is shown in the sectional drawing of the engines of which we give the details, dotted in. This cock should never be furnished with a rose within the condenser, and should never be joined on to the injection-pipe proceeding from the sea. We have known various cases in which a vessel has been nearly lost from the internal roses of the bilge injection becoming choked up with refuse drawn out of the bilge, and which, but for those roses, would have passed into the condenser and been delivered by the air-pump without creating any obstruction.

Framing.-Cast-iron framing is now given up in marine engines, and malleable iron framing alone is employed. Of malleable iron framing as applicable to the side-lever engine, we have given a specimen in the engines of the City of London steamer; and of cast iron framing it is needless, under the circumstances, to say much. It is a bad plan, in our judgment, to attach the diagonal stay to the hot well, as is sometimes done, as the working of the stay breaks the hot-well joint. It is a bad plan, too, to attach the framing to the sides of the ship, as the working of the ship in a sea will strain and may break it. In iron steamers a plan now prevails of running the deep beams before and abaft the crank-hatch (which are also made of iron) through the ship's side, joining the extremities of those beams by curved cross-beams, on which the shaft plummer-blocks are made to rest The paddle-wheel, by this plan, is overhung, and the whole of the arms radiate from a triple centre. A very substantial framing may be made by adopting this arrangement, and it is one which is applicable in the case of direct-action engines as well as in those on the side-lever plan. The brasses of the paddle-shaft plummer-blocks should not be made with fitting strips on the backs, but the whole of their exterior should be planed, and the interior of the plummer-blocks should also be planed for their reception. Brasses fitted with fitting strips soon wear slack sideways. Octagonal bottom brasses are not so good as those which are square, as they cannot be lined up so conveniently if the shaft gets out of truth. Square-bottom brasses, with flanges, as shown in fig. 302., we have found preferable to any other variety.

Fig. 302.

SQUARE BUSH.

Side-lever. The drawing we have given of the side-lever will sufficiently explain its general form and dimensions. In some of the more recent side-lever engines, the side-levers are made of malleable iron, each lever being composed of two plates, set on edge, the length of the bearings apart from each other. The studs in the side-lever should be steeled and should be of larger dimensions than is necessary for strength; as if they