being found, it is not desirable to alter it in bringing the machine to time. To effect an adjustment, the two screws c c have been introduced, the drawing out of which from the centre causes the machine to lose, and the screwing them in to gain. the bell; and as soon as the pin leaves the tail of the hammer, the force of the spring u acting on the lower part of the hammer produces Many important matters relating to chronometers apply in a smaller degree to spring clocks and to pocket-watches, and will receive notice as we go on. But it is necessary here to say a few words concerning the competitive trials of chronometers, which have partly resulted from, and have partly suggested the beautiful inventions perfected within the last few years by Eiffe, Frodsham, Loseby, Dent, and others. Between the years 1822 and 1835, the Admiralty gave rewards as prizes for the good performance of individual chronometers: two or three sums being given annually to the makers of those which showed the smallest amount of error. Nothing was given for any new principle; but the general construction was improved by this competition. The relative order of merit was arrived at by taking the number of seconds in the greatest difference between one week's indications and the next, and adding to it twice the greatest difference between one week's rate and the next; the sum was the trial number; and that chronometer gained the prize which exhibited the smallest trial number. The trials were conducted at Greenwich Observatory, under the Astronomer Royal. After the year 1835 the prizes were withdrawn; but the makers were equally ready to compete, on the score of reputation and commercial advantage. The test was rendered more severe; for every chronometer was exposed in the open air to the fiercest cold of winter; and at other times to the air of a chamber heated to 100° Fahr. In successive years different makers obtained the post of honour; for instance Poole in 1845, Hutton in 1846, Frodsham in 1847, Hewett in 1848, Eiffe in 1849. In the five years here indicated, 219 chronometers were tested at Greenwich; of which 79 were bought for public use, at a maximum price of 621. During the course of these refined experiments, a fact was observed which amounts almost to a new discovery. However perfect may be the compensation of a chronometer for certain temperatures, it was found that it did not remain constant for all temperatures. If a chronometer be planned to resist great cold and great heat, it gains at medium temperatures; but if specially adapted for the latter, it errs slightly in very cold and very hot weather. The late Mr. Dent explained this fact by supposing that the inertia of a spring balance, as usually constructed, cannot be made to vary uniformly according to temperature, but will vary more rapidly in cold weather than in hot. Chronometer-makers at once set about devising a remedy; and this remedy, however produced, is called the secondary compensation. Loseby, Dent, Eiffe, Molyneux, and other makers, have invented highly-curious arrangements of spring balance, all of which have been severely tested at Greenwich, and some of them rewarded by government grants. Loseby's arrangement, which received high commendation from the Astronomer Royal, may be thus briefly described. Attached to the balance is a curved tube containing mercury. The mercury, on expanding with an increase of temperature, arrives at certain parts of the tube inclined in different degrees to the radii of the balance; and therefore its successive expansions produce successive effects of different magnitudes on the momentum of the balance. By giving different forms to the tube, the law of the successive alteration of the momentum may be made to adapt itself to the law of alteration of the elasticity of the spring, whatever that law may be. Spring Clocks.-In describing spring-clocks, we shall at the same time treat of many parts of the mechanism which are applicable likewise to chronometers, but not described in the foregoing paragraphs. a blow on the bell. The number of strokes on the bell is regulated by mechanism placed on the outside of the front plate of the clock, but is removed from the figure just described, to prevent confusion. (See fig. 5.) On the centre wheel-pinion k (fig. 4), whose arbor comes Fig. 5. In fig. 4, A B C D may be taken as representing the front plate of an eight-day spring clock (which is supposed to be transparent), and is attached to another plate of similar form by five strong pillars between which the wheels here shown are placed. EE are two barrels containing springs; the one on the right gives motion to the train of wheels called the going or watch train; the other to the striking train of wheels e, f, g, h, and fly i. In producing the series of movements, a is the main wheel of 96 teeth, acting in the centre wheel-pinion k of eight leaves, to which is attached the centre wheel b; this revolves in an hour, and acts in the third wheel-pinion 1, on which is fixed the third wheel c, acting in the swing or scape wheel-pinion m (not seen in the cut); to this pinion is fixed the swing-wheel d, whose teeth act alternately on the two pallets n o, and thereby give motion to the pendulum by means of a piece attached to the arbor of the pallets, one end of which enters a slit made in the pendulum for its reception. F F are the two fusees, the use of which has already been described in connection with chronometers. The method in which the fusees are attached to their respective wheels a and e is shown in fig. 6, where is the main wheel of 96 teeth hollowed out to receive the click b and its spring c which are attached to the wheel, the ratchet d being attached to the under side of the fusee by two screws. In fig. 4, e is the striking main wheel, haying 84 teeth; this drives the pinion p of eight leaves, on which is the pin-wheel ƒ of 64 teeth, into the rim of which are put eight pins to lift the hammer s, by acting upon its tail t. The pin-wheel ƒ drives the pallet-pinion q of eight leaves, on which is fixed the pallet-wheel g of 56 teeth; this pallet-wheel acts in the warning wheel-pinion r of seven leaves, on which is the warning-wheel h of 48 or 50 teeth, acting in the fly-pinion i. When in action a pin in the pin-wheel catches the tail of the hammer t, and raising it, the hammer-head & recedes from 8 through the front plate about one and a half inches, is placed the minute wheel a (fig. 5), which revolves with the centre wheel in an hour, and carries the minute-hand of the clock. This wheel has a pipe nearly as long as the centre-wheel arbor, the upper end of which is squared to receive the minute hand; and by means of a small spring beneath the wheel, which rests upon a shoulder just above the upper surface of the front plate, and acts against the upper surface of the wheel, the wheel, together with the hand, is forced against a pin over Fusee-wheel. nism of the clock, or rather that part of it which is called the going or watch train; for the minute-wheel a gives motion to another minutewheel b, which, as it must revolve also in an hour, has the same number of teeth as a. A pinion in the centre of the wheel b has six leaves, and acts in the hour-wheel c of 72 teeth, which is placed over the minute-wheel a, and consequently revolves once in 12 hours, and has screwed to its socket, at the upper end, the hour-hand. To the socket of the hour-wheel, about one-eighth of an inch above the wheel, is fixed a piece in which are 12 steps, each of which includes an angle of 30°, or a twelfth part of a circle; this piece is called the snail, and is represented by d. e is a rack whose centre of motion is a stud or pin f, on which it acts by means of a pipe about half an inch long, and on to the upper end of which is riveted the rack-tail g; in this rack-tail is a short pin h, pointing perpendicularly downwards to the front plate of the clock. The rack lies about the tenth of an inch above the front plate; but the pipe which acts on the stud is long enough to carry the rack-tail just clear of the snail when the rack is forced back by the spring i; whilst the pin h is long enough to strike against the steps in the snail, and yet so short as to be perfectly free of the hour-wheel c. Near this is k, the rack-hook moving freely on a stud; m the lifting piece, also moving freely on a stud p; n the tail of the lifting-piece firmly pinned on to the upper part, and moving with it; o the gathering pallet, which has a square hole through it, and is fixed upon the square end of the arbor q of the wheel g (fig. 4) which revolves once for every blow given by the hammer. A pin in the warning-wheel h always stands in the same position when the striking part is at rest, which is the position represented in fig. 4. On the end of the lifting-piece is a small piece q (fig. 5), which passes through a slit in the front plate, and resting on the bottom of the slit, keeps the lifting piece in its proper position. The gathering pallet o rests on a pin r in the rack, and thereby prevents any motion in the internal wheel-work of the striking train. Such being the mechanism, the mode of action will be understood from a study of the figures, without a detailed description. A small piece x is called the pull-piece, by pulling a string at the end of which the lifting-piece is raised, and the clock is made to repeat the hour last struck at any required time. y is a spring to force the pull-piece against the pin z fixed in the plate of the clock; s is another pin to limit the motion given to the pull-piece when the string t is pulled. Pocket-watches.-Without attempting to notice the numerous improvements in spring clocks, we shall proceed briefly to describe those features in a pocket watch which are of a distinctive w Fig. 7. W Endless Cord. A, and also in slight spring s. B character. When clocks and watches had acquired a certain degree of accuracy in their performance, the time lost in winding up (especially when it had to be done every twenty-four hours) became a matter of importance; and there have been several inventions to remedy this evil, by producing what is called a maintaining power. By Huyghens the clock was kept going while winding by means of an endless cord, as in fig. 7. B is the clock-barrel; c, that portion of the line which comes from the barrel to the weight; P, a pulley for the line to run over; Q, a pulley for the line to run under, and to which is attached a small weight w. It will be seen by inspection that the hand applied to that part of the line marked a will be able to raise the weight w without depriving the barrel B of any portion of the power by which it is urged forward, and which power in this arrangement is equal to onehalf of the weight w. Another kind of maintaining power is the forcing spring, shown in fig. 8. A lever A, whose centre of motion is o, has a notch cut in its end, into which is jointed a small lever c, whose centre of motion is ; this small lever is kept in its proper position against the bottom of the notch, as shown in (which is only another position of the lever), by a D is a strong spring which acts constantly on the that of a; and in its progress in passing a tooth of the wheel the small lever c assumes the position represented in fig. 9, which it is allowed to do by the very slender spring s. As soon as the tooth is passed, the pressure of s obliges the lever c to return to its original place; and by the pressure of its opposite end on the bottom of the notch in which it is inserted, the lever A is prevented from regaining its former position by the pressure of the piece c on the tooth of the wheeluntil the wheel shall have advanced so far as to have allowed its escape, when the lever regains its position B, where it remains till another winding becomes necessary. It will be evident that so long as c remains on a tooth, the wheel will be urged forward by the action of the spring D. ee are two pins which are fixed in the plate of the clock, and serve to determine the quantity of motion given to the lever A. Harrison's contrivance for the same purpose, however, is the one now in general use, both in clocks and watches, and is admirably adapted to the purpose. When this principle is applied to a fusee, it is termed a going fusee; but maintaining power, as a more comprehensive term, is now generally applied. Into the hollow of the fuseewheel is placed a circular spring a b c, (Fig. 10), which is secured to 1, Harrison's Maintaining Power-Going Fusee. Fusee auxiliary ratchet, and fusee-wheel attached; 2, fusee-wheel and auxiliary spring, separate; 3, 4, upper and under sides of the fusee, separate; 5, auxiliary ratchet q, with clicks a a, and springs e e, attached. the wheel by a pin at about one-fourth of its circumference from the end a, namely, at b. The wheel has a short notch cut through it, near the other end of the spring. The spring passes over this notch; and by means of a pin c, fixed firmly in the spring and projecting through the notch in the wheel, a motion is allowed to the spring, which in extent is equal to the difference between the length of the notch in the wheel and the thickness of the pin which passes through it. It is the reaction of this spring through the short distance already mentioned which maintains the motion in the watch during the time of winding up. Instead of any click and spring being attached to this fusee-wheel, as in an ordinary eight-days' clock, there is a circular disc of steel, rather larger than the bottom of the fusee, and smaller than the fusee-wheel, having very fine ratchet-teeth cut in its edge, and two clicks a a and springs e e on its upper surface, in which the ratchet fixed on the under side of the fusee, and called the fusee-ratchet, acts. The steel-ratchet is called the auxiliary-ratchet, and its teeth stand in a direction opposed to those of the fusee-ratchet. We will now suppose the auxiliary ratchet to be laid on to the fuseewheel over the spring a b c; a hole in its centre passing over a short pipe in the centre of the fusee-wheel retaining it in its situation; the pin c, which we have described as projecting through the notch in the fusee-wheel, also projecting upwards just equal to the thickness of the auxiliary-ratchet, through which it likewise passes; and the pin exactly fitting the hole in the ratchet. In this situation the wheel and ratchet We must next describe the escapement, so important in horology. This term is applied to a combination of parts in a clock or watch, which has for its object the conversion of the circular motion of the wheels into a vibratory motion, as exhibited in the pendulum. In the description we are about to give, the term will be made to include the scape-wheel, the pallets with their arbor or axis, and a bent lever attached thereto, called the crutch, which last piece maintains the motion of the pendulum. In a watch this combination consists of the scape-wheel, together with all those parts lying between it and the balance, and which are concerned in converting the circular motion of the wheels into the alternating one of the balance. In Graham's deadbeat escapement the distance between the centre of motion p of the pallets and the centre of the scape-wheel is equal to one diameter of the scape-wheel. In fig. 11, the tooth i has just given impulse to the pallet P and escaped from it; the tooth o has in consequence fallen upon that part of the pallet q called its arc of rest, which, in both pallets, is formed by a circle struck from the centre of motion p of the pallets. The impulse given by i causes the pendulum, and with it the pallets, to vibrate some distance after i has left P and o has fallen on Q; but the arc of rest being concentric with the centre of motion of the pallets, the wheel ceases to rotate, or remains dead, until the pendulum by its the tooth o escapes from the pallet ; when another tooth h, on the opposite side of the wheel, falls on the arc of rest of the pallet P, which arc is in this pallet on the outside, and on which the tooth rests until by the return of the pendulum the pallet P is lifted so high as to allow h to get on the inclined plane or face of the pallet r, upon which it acts. The tooth raises the pallet and with it the pendulum, till it escapes and gains the position i, when the same process is repeated: the wheel alternately giving impulse to one pallet and resting on the circular part of the other, which we have denominated the arc of rest. When the pendulum is in a state of rest, some one tooth is always resting on one of the circular arcs; the pendulum being put in motion brings a pallet into a position to receive an impulse from the wheeltooth, when the process already described commences. Most of the great improvements in pocket watches have been made in the horizontal or flat forms; but the common vertical watch must be first described. The annexed cut (fig. 12) represents such a watch as it would appear if the dial (which is here omitted) were turned downwards. A is the barrel; B, the fusee; b, the chain by which motion is communicated from the barrel to the fusee, on which is the great or fusee wheel acting on the centre-wheel pinion D, on which is riveted the centrewheel: the arbor of the pinion D being prolonged through the plate of the watch as far as l. The centre-wheel and its pinion revolve in an hour. Upon that part of the arbor D which is on the outside of the plate or frame is placed the cannon-pinion, which has a hole quite through it for the reception of the centre-wheel arbor, on which it turns spring-tight. The cannon-pinion is secured in its place by a small pin through the end of the centre-wheel arbor 1, the end of the pinion being squared to receive the minute-hand h. The cannon-pinion has 12 leaves acting in the minute-wheel, of 48 teeth, causing the latter to revolve once in four hours. Concentric with the minute-wheel, and attached to it, is a pinion, having a hole through their common centre, through which passes a stud fixed on the plate. This pinion, having 14 leaves, drives the hour-wheel m, of 42 teeth, once round in 12 hours; this wheel is placed over the cannon-pinion by a socket, which has a hole through it for the cannon-pinion to pass through; on this socket is fixed the hour-hand k. It will be perceived that by this arrangement the cannon-pinion, minute-wheel, pinion, and hour-wheel, together with the hands, can all be turned backward or forward without affecting the interior mechanism of the watch, simply by the application of a key to the squared end of the cannon-pinion. The assemblage of wheels, &c., thus put in motion is called the motion work of the watch; that between the plates, the movement, which we shall next describe. The centre-wheel gives motion to the third ARTS AND SCI. DIV. VOL. IV. h I L wheel-pinion, to which is attached the third wheel G, acting upon the contrate-wheel pinion. On this is placed the contrate-wheel acting in the pinion of the balance-wheel L, which is also called the scape-wheel. We have already explained the mode by which the balancewheel teeth act upon the pallets, so as to cause an alternating motion in the balance M. One end of the balance-wheel arbor works in a piece called the dovetail, which is inserted in a piece p, called the potence, firmly attached by a screw to one of the plates of the watch; the other end works in a piece called the follower, which is inserted in another piece riveted into the plate called the counter-potence, (left out of the figure to prevent confusion). Another part of the potence, called the foot, receives one end of the balance arbor or spindle, called the verge (on which are the pallets); the other end works in a hole in the pin o, which passes through the centre of the cock q q, which is secured to the upper plate of the watch; the pendulum-spring (also called the regulating-spring and hair-spring) has one end, immediately below the balance, secured to a stud fixed in the plate, and the other pinned fast to a small collet, which goes spring-tight unto the axis of the verge, and is seen just under the balance. Figs. 13, 14, and 15, represent some of the parts separately. Fig. 13 shows the main-spring in a relaxed state as it would appear out of the barrel; to which, when in, one end of it is attached, the other being held by a hook in the arbor of the barrel, which comes through the plate, as shown in fig. 12, and is kept from turning by a ratchet and click. The spring is wound up by the chain acting on the barrel and pulling it round, which operation is performed by turning a key placed on the squared end of the fuseearbor. The effort of the spring to unbend itself after being wound up 3 A 1, 2, 3, being the teeth of repose, and a, b, c the teeth of impulse, which are triangular, and stand perpendicular to the plane or surface the wheel. CD, the impulse pallet, fixed upon the arbor of the balance, and standing just above the surface of the wheel A A, receives itsmotion from the teeth a, b, c, &c. After the tooth a has passed the pallet c D, the tooth b comes in contact with a small roller made of ruby, and placed on the lower part of the axis of the balance, where it remains till the balance is brought back by the balance-spring to such a position that the notch, shown by the dotted line in the ruby roller, will allow the tooth 1 to enter it, and thereby pass the balancearbor, or escape, which it does by the wheel A A being constantly urged in the direction from 3 to 1. As soon as tooth 1 escapes from the notch, tooth b gives a fresh impulse to the pallet CD, and the act of escapement is thus repeated; the wheel moving forward one whole tooth, and the balance making two vibrations for each impulse given by the upright teeth. Another effective variety, the lever escapement, is shown in fig. 17. The lever is placed on the pallets in a position at right angles to that in which it is usually placed in a watch, by which means we think the principle will be more apparent to the general reader. A A is the scapewheel moving in the direction of the arrow; bd the pallets, whose centre of motion is c. To the pallets is pinned the lever l, in which is the guard-pin e, pointing upwards from the lever ; the roller f is fixed on the axis of the balance, and stands just above the lever 1, having a piece cut off from its circumference to allow the guard-pin e to pass and repass the roller, which it does when the escape takes place. o is a ruby pin fixed in the roller, and pointing downwards through the notch in the end of Lever Escapement. guard-pin e coming in contact with the circular edge of the roller. When an impulse is given by a tooth to the other pallet d, the lever impels the ruby pin o to the left hand, where precisely the same effects take place with regard to the guard-pin e, &c., as have been already described. If the pallets b and d were of the form shown by the dotted lines (which are supposed to be circular ares concentric to the centre of motion c of the pallets), it is evident it would be a perfect dead beat, like the clock escapement previously described; but in order, after the escape has taken place, that the guard-pin e may be retained at a small distance from the roller, that part of each pallet on which the tooth rests when it falls on the pallet is taken off, as shown in the figure; and as the faces of the wheel-teeth are considerably undercut, the wheel advances a small distance, after having fallen on that part of either of the pallets which is within the dotted line. This further advance of the wheel draws the pallet down towards the centre of the wheel, and thereby keeps the guard-pin e at a slight distance from the edge of the roller f. As soon as the balance has performed so much of the returning vibration as to bring the ruby pin o into the notch in the lever, the momentum of the balance, acting through the medium of the ruby pin o upon the lever, moves it a short distance, and thereby lifts the pallet outwards from the centre of the wheel and unlocks it. During this unlocking the wheel retrogrades (before it can get upon the face of the pallet to give a fresh impulse) just as much as it had previously advanced after falling on the pallet. By this retrograde motion the tooth gains the inclined plane or face of the pallet, and gives a new impulse; and the same process is repeated by another tooth on the opposite pallet. PP are two pins, called banking-pins, against which the lever presses when locked, and which prevent the guard-pin e from being drawn too far away from the edge of the roller f, when the locking takes place. In fig. 18 is shown a horizontal escapement. ABC represents the part of the cylinder which has the least portion of its circumference taken away, when a tooth is in the cylinder, the point rubs against the internal surface until the balance by its vibration gets into such a situation that the inclined plane can act upon its edge. It then impels the cylinder in the direction from D to A; until the highest part of the plane escapes from the inside of the cylinder, and the next tooth falls upon the outside. This tooth continues to rub until the balance completes its vibration and has returned so far as to permit the point of the tooth, which has been rubbing on the outside of the cylinder, to get upon its edge, where it gives impulse to the cylinder; and when its heel escapes, the point falls on the inside of the cylinder, and the former process is repeated. 1, 2, 3, &c., are teeth of the horizontal or scape-wheel, one of which is seen inside the cylinder; the dotted lines represent the face or inclined plane of the tooth, which is just coming in contact with the edge of the cylinder. The direction of the motion of the wheel is from 1 to 3; the proportion of the cylinder to the wheel is such, that a tooth of the wheel, when in the cylinder, may just have sensible shake; and the outside diameter must be sensibly less than the distance between two teeth. The detached escapement, such as is used in a modern chronometer, is shown in fig. 19. AAA is the scape-wheel, made either of brass or Fig. 19. Detached Escapement. steel, the teeth 1, 2, 3, 4, &c., of which are considerably undercut on the face. The steel-roller or main-pallet BBB, which is fixed on the arbor of the balance, has an opening in it, the face of which is also much undercut as shown near B, and has set in it a piece of hard stone, such as a ruby, for the points of the teeth to act upon. s is a stud firmly fixed to one of the plates to which the detent-spring E E is secured by a screw c. This spring is made extremely slender and weak in the part E near the stud; and it is only by the yielding of this thin part of the detent-spring that any motion can be given to the detent for the purpose of unlocking the wheel; so that some part of this spring may be considered as the centre of motion of the detent. D is a stud also fixed to the plate of the watch, into which is inserted a screw d, against the head of which the detent rests. o is a ruby pin inserted in the detent, pointing downwards from the detent; so that one of the teeth of the wheel which is supposed to pass under the detent may rest on the pin; and in this state the wheel is said to be locked. To the inner side of the detent is attached a very delicate spring, called the lifting-spring, which rests upon and extends a little beyond the end of the detent. Concentric with the main pallet, and just above it, is a small lifting-pallet 9, which should be flat on its face or lifting-side, and rounded off on the other side. In the position given in the figure, the lifting-pallet q is just coming with its face in contact with the lifting-spring p; which in the course of vibration it lifts, and with it the detent (on whose point the lifting-spring presses), so as to raise the pin o clear of the wheel-tooth 5. By the time the wheel is free from the ruby-pin, the main-pallet has advanced so far as to be ready to receive an impulse from the tooth 1; and before the tooth escapes the lifting pallet q, parts with the spring p, and the detent resumes its place on the head of the screw d. In this position the ruby-pin receives the point of tooth 6, as soon as tooth 1 has escaped from the ruby-face of the mainpallet BBB. The balance, having performed this vibration by the impulse given to the main-pallet, returns by the force of the balance The rounded side of the latter spring, and with it the lifting-pallet q. The name repeating-watch, or repeater, is applied to a watch which, in addition to showing the time on a dial, is supplied with mechanism by putting which in action the wearer is enabled to ascertain the time within certain limits. We have shown, in describing an eightday clock, how the number of blows given by the hammer to the bell is made to correspond with the hour denoted by the hands of the dial; and also that, by pulling a string, the clock will at any time repeat the hour last struck. But this will not be the case where the minute hand has approached within about ten minutes of twelve o'clock, for from that time till the hand comes to twelve the clock is on the warning, and is in such a position that it cannot strike at all. This defect is remedied in repeaters. Most repeaters are watches which are capable of striking on a bell or spring the hours and quarters; but there are others which also strike the minutes, and these by way of distinction are called minute repeaters. In a repeater, besides the goingtrain and the motion-work, there is an additional train of wheels between the frame-plates, called the runners or little wheel-work, or sometimes the repeating-train. This train serves the purpose of regulating the rapidity with which the successive blows shall be given to the bell, and consists generally of five wheels and five pinions. The last pinion in the train, performing the office of a fly-wheel, is generally called the fly-pinion; and, when the striking is regulated to its ordinary rate, makes about two hundred revolutions to every blow of the hammer. The chief use of these intricate pieces of mechanism is to furnish the means of knowing the hour of the night in the dark. All the more delicate pivots of chronometers, and of the better kind of watches, work in jewelled holes, which will be found described under JEWELLING. Pendulum Clocks.-We now arrive at the consideration of those horological machines which receive their regulating adjustment by means of the pendulum. The sensible equality of the oscillations of a weight suspended by a string or wire was first applied as a regulator to a clock by Huyghens about 1657. The successive improvements in the escapement, which sustains the motion of the pendulum and records its vibrations, and those in the pendulum itself, which secure a perfect equality in the duration of each oscillation, have finally produced the astronomical clock, the most accurate machine which man has hitherto constructed, and one of the most essential instruments in a modern observatory. We shall suppose that the dead-beat, or Graham's escapement, is that adopted. The pallets PQ, fig. 11, have motion on an arbor which passes through p, and has its pivots resting in holes in the clock-frame. A slender bar or wire, called the crutch, is attached to this arbor, and a notched piece projecting outwards and backwards from the crutch clasps the rod of the pendulum. The pendulum is hung from a cock at the back of the frame, and moves with the crutch. In a well-made clock, the error arising from expansion from temperature is the most considerable, and is that which must be guarded against. Before explaining more accurate and costly contrivances, it will be well to point out one recommended by Mr. Francis Baily. (Mem. Astron. Soc.,' vol. i., p. 381.) Take a cylinder of lead about 14 inches long, and pierced through its axis, as a bead, with a hole large enough to admit freely the rod of a wooden pendulum. This hollow cylinder rests on a nut, which works on a screw in the continuation of the rod below. The rod itself, from the centre of motion to the nut, will be about 46.0 inches. As it is easier to cut the cylinder shorter than to lengthen it, and as the expansion of the spring is not allowed for, and that of the wood is somewhat uncertain, it will be better to make the leaden cylinder an inch longer for a first trial; but even if the pendulum should turn out to be under compensated, an additional ring of lead may be added, above or below, of the thickness required. To the best clocks it is usual to apply either the gridiron pendulum of Harrison (which was once chiefly used in England, and is still in repute abroad), or the mercurial pendulum of Graham. The annexed figure (fig. 20) is not exactly the pendulum as arranged by Harrison, but accords with his principle. The steel rods 1 and 5 are pinned into two brass cross-pieces, Aa, Bb. The zinc rods 2 and 4 are pinned below into Bb, and carry a cross-piece above, into which the steel rod 3 is pinned. Rod 3 passes freely through a round hole in Bb (this is shown by dotted lines), and is tapped into a screw below; the bob rests upon the nut, which works on the screw. The steel rods 1 and 5 expand downwards, the zinc rods 2 and 4 expand upwards, and the steel rod 3 downwards; and it is possible so to adjust their lengths (the expansion of zinc being more than double that of steel) that the effects of the expansion downwards and upwards shall have no effect on the length of the pendulum or time of oscillation. Harrison used brass instead of zinc for the upward expansion; and in order to produce a perfect compensation, was forced to use four more rods, a second pair of brass |