Of all the instruments used by a watchmaker in the prosecution of his business, there is probably none more important than his regulator. Its purpose is to divide time into seconds, and it is the standard by which the practical results of his labors are tested; the guide which all the other timekeepers in his possession are made to follow and the arbitrator which settles all disputes regarding the performance of his watches.

No regulator has yet been constructed that contains within itself every element for producing absolutely accurate timekeeping. At intervals they must all be corrected from some external source, such as comparison with another timekeeper, the error of which is known, or by the motion of the heavenly bodies, when instruments for that purpose are available. Before beginning to make a regulator, the prudent watchmaker will first reflect on the various plans of constructing all the various details of an accurate timekeeper, and select the plan which, in his opinion, or in the opinion of those whom he may consult on the subject, will best accomplish the object he has in view.

In former years a regulator case was made with the sole object of accommodating the requirements of the regulator, and every detail in the construction of the case was made subservient to the necessities of the clock. The plain, well made cases of former years are now almost discarded for those of more pretentious design. If the general change in the public taste demands so much display, there can be no objection. It is perfectly harmless to the clock, if the designers and makers of the cases would only remember that narrow waists or narrow necks on a case, although part of an elegant design, do not afford the necessary room for the weight and freedom of the pendulum; that the doors and other openings in the case must be constructed with a view to exclude dust; and that the back should be made of thick, well-seasoned hardwood, such as oak or maple, so as to afford the means of obtaining as firm a support for the pendulum as possible.

When a regulator case is known to have been made by an inexperienced person, which sometimes happens, or when we already have a case, it is always the safest course for those who make the clock to examine the case personally and see the exact accommodation there is for the clock. Sometimes, when we know beforehand, we can, without violating any principle, vary the construction a little, so as to make the weight clear the woodwork of the inside of the case, and in other respects complete the regulator in a more workmanlike manner by making the necessary alterations in the clock at the beginning of its construction, instead of after it has been once finished agreeably to some stereotyped arrangement.

The arrangement of the mechanism of an ordinary regulator is a simple operation compared with some other horological instruments of a more complex character. We are not limited in room to the same extent as in a watch, and the parts being few in number a regulator is more easily planned than timekeepers having striking or automatic mechanism for other purposes combined with them; yet it often happens that the inexperienced make serious blunders in planning a regulator, and, as the clock approaches completion, many errors make themselves visible, which might have been avoided by the exercise of a little more forethought. It may be that, when the dial is being engraved, the circles do not come in the right position, or the weight comes too close to the pendulum, or the case, or the cord comes against a pillar, or other faults of greater or less importance appear, all of which might have been obviated by taking a more comprehensive view of the subject before beginning to make the clock. The best way to do this is to draw a plan and side and front elevations to a scale.

The position which the barrel and great wheel should occupy is worthy of serious consideration. In most of the cheap regulators, as well as in a few of a more expensive order, the barrel is placed in a direct line below the center wheel, as is shown in Fig. 152. This arrangement admits of a very compact movement, and it also allows the weight to hang exactly in the center of the case, which some think looks better than when it hangs at the side, especially when there is a glass door in the body of the case. But while a weight hanging in the center of a case may be more pleasing to the eye than when it hangs at the side, this is an instance where looks can, with great propriety, be sacrificed for utility, because when the weight hangs in the center it comes too close to the pendulum, and is very liable to disturb its motion. In proof of this statement, let any reader who has a regulator with a light pendulum and a comparatively large weight hanging in front of it, closely watch the length of the arc the pendulum vibrates when the weight is newly wound up and when it is down opposite the pendulum ball, and he will observe that the length of vibration of the pendulum varies from five to fifteen minutes of arc, according to the position in which the weight is placed; that the pendulum will vibrate larger arcs when the weight is above or below the ball than when it is opposite it; and if the clock has a tendency to stop from any cause, that it will generally do so more readily when the weight is opposite the pendulum ball than when it is in any other position. For this reason I would dispense with the symmetrical looks of the weight hanging in the center of the case, which, after all, is only a matter of taste, and construct the movement so that the weight will hang at the side, and as, far away from the pendulum as possible.

Fig. 153 is intended to represent the effect which placing the barrel at either side has on throwing the weight away from the pendulum. A is the center wheel; B and C are the great wheels and barrels with weights hanging from them; D is the pendulum. It will be noticed by the diagram that the weight at the left of the pendulum is exactly the diameter of the barrel farther away from the pendulum than the weight on the right. On close inspection it will also be observed that on the barrel C the force of the weight is applied between the axis of the barrel and the teeth of the wheel, while on the barrel B the axis of the barrel lies between the point where the force is applied and the point where the teeth act on the pinion; consequently a little more of the effective force of the weight is consumed by the extra amount of pressure and friction on the pivots of the barrel B than there is in C.

Notwithstanding this disadvantage, I would for a regulator recommend the barrel to be placed at the left side of the center wheel, because the weight may thereby be led a sufficient distance from the pendulum in a simple manner. If we place the barrel at the right, and thereby secure the greatest effective force of the weight, and then lead the weight to the side by a pulley, we will lose a great deal more by the friction of the pulley than we gain by the proper application of the weight.

In a regulator with a Graham escapement but little force is required to keep it going, and there is usually accommodation for an abundance of power; therefore we cannot use a little of this superabundant available force to better advantage than by placing the barrel at the left side of the clock, and thereby throw the weight a sufficient distance from the pendulum in the simplest manner.

The escapement we assume to be the old dead beat, as for timekeeping it is equal to a gravity escapement while possessing advantages undesirable to sacrifice for a doubtful improvement. The advantages it possesses over any form of gravity escapement are: it has fewer pieces and not so many wheels; it takes very much less power to drive; is not liable to fail in action while winding, if the maintaining power should be rather weak; while for counting, seconds and estimating fractions, its clear, definite, and equable beat has great superiority over the complication of noises made by a gravity escapement.

Full directions for making this and other escapements have already been given, but in a regulator there are some considerations which will not be encountered in connection with the escapements of ordinary clocks, where fine timekeeping is not expected. We have previously stated that the center of suspension of the pendulum should be exactly in line with the axis of the escapement and we will now endeavor to state plainly how important this is in a fine clock and the reasons for it. Mr. Charles Frodsham, the noted English chronometer maker, has conducted a series of careful experiments and the results were communicated in a report to the British Horological Society, as follows:

When we talk of detached escapements, or any escapement applied to a pendulum, it is necessary to bear in mind that there is always one-third at the least of the pendulum’s vibration during which the arc of escapement is intimately mixed up with the vibration, either in locking, unlocking, or in giving impulse; therefore, whatever inherent faults any escapement may possess are constantly mixed up in the result; the words “detached escapement” can hardly be applied when the entire arc of vibration is only two degrees; or, in other words, what part of the vibration is left without the influence of the escapement?—at most one degree. In chronometers the arc of vibration is from ten to fifteen times greater than the arc of escapement.

The dead-beat escapement has been accused of interfering with the natural isochronism of the pendulum by its extreme friction on the circular rests, crutch, and difficulty of unlocking, etc., all of which we shall show is only so when improperly made.

When the dead-beat escapement has been mathematically constructed, and is strictly correct in all its bearings, its vibrations are found to be isochronous for arcs of different extent from 0.75 of a degree to 2.50 degrees; injurious friction does not then exist; the run up on the locking has no influence, nor is there any friction at the crutch; oil is not absolutely necessary, except at the pivots; and there is no unlocking resistance nor any inclination to repel or attract the wheel at its lockings.

The general mode of making this escapement is very defective and indefinite, and entirely destroys the naturally isochronous vibration of the pendulum.

The following is the usual rate of the same pendulum’s performance in the different arcs of vibration with an escapement as generally constructed after empirical rules:

Thus for a change of vibration of 1°, we have a daily error of 3.5. No change of suspending spring will alter inherent mechanical errors destructive of the laws of motion. With clocks made in the usual manner, whether you apply a long or short spring, strong or weak, broad or narrow, you will not remove one fraction of the error; so the sooner the fallacy of relying upon the suspending spring to cure mechanical errors is exploded the better.

That the suspending spring plays a most important part must be admitted, since, when suspended by a spring, a pendulum is kept in motion by a few grains only, whereas, if supported on ordinary pivots, 200 lbs. weight would not drive it 2' beyond its arc of escapement, so great would be the friction at the point of suspension.

The conditions on which alone the vibrations of the pendulum will be isochronous are the following:

To accomplish the foregoing conditions, there is but one fixed point or line of distance between the axis of the escape wheel and that of the pallet, and that depends upon the number of teeth embraced by the pallets and only one point in which the pallet axis can be placed from which the several lines of the escapement can be correctly traced and properly constructed with equal angles, and equal rectangular lockings on both sides, so that each part travels with the same degree of angular velocity, which are the three essential points of the escapement.

Much difference of opinion has been expressed upon the construction of the pallets, as to whether the lockings or circular rests should be at equal distances from the pallet axis, with arms and impulse planes of unequal length, or at unequal distances from the pallet axis, with arms and impulse planes of equal length. In the latter case the locking on one side is three degrees above, and on the other three degrees below the rectangle, whereas in the former the tooth on both sides reposes at right angles to the line of pressure; but the length of the impulse planes is unequal. When an escapement is correctly made upon either plan, the results are very similar.

It is possible to obtain equal angles by a false center of motion or pallet axis; but then the arcs of repose will not be equal. This, however, is not of so much consequence as that of having destroyed the conditions Nos. 2, 3, 4; for even at correct centers, if the angles are not drawn off correctly by the protractor, and precisely equal to each other, the isochronous vibrations of the pendulum will be destroyed, and unequal arcs will no longer be performed in equal times; the quiescent point is not the center of the vibration, except when the driving forces are equal on both sides of the natural quiescent point of the pendulum at rest.

Now this is the very pith of the subject, and which few would be inclined to look for with any hope of finding in it the solution of this important question, the isochronism of the pendulum.

One would naturally suppose that unequal arcs on the two sides of the vertical lines would not seriously affect the rate of the clock, but would be equal and contrary, and consequently a balance of errors, and so they probably are for the same fixed vibration, but not for any other; because different angles are driven with different velocities, the short angle has a quicker rate of motion than the long. Five pounds motive weight will multiply three times the pendulum’s vibration over an arc of escapement of 0.75°; but the same pendulum, with an arc of escapement of 1°, would require 11.20 lbs. to treble its vibration; the times of the vibration vary in the same ratio as the sum of the squares of the differences of the angles of each pallet, compared with the spaces passed over.

From this it will be seen that the exact bending point of the pendulum spring should be opposite the axis of the fork arbor when regulating the clock and this may have to be determined by trial, raising or lowering the plates by screws in the arms of the suspending brackets until the proper position is found, when the movement may be clamped firmly in position by the binding screws, see Fig. 158.

On common clocks the crutch is simply riveted on its collet and bent as required to set the clock in beat, but for a first-class clock a more refined arrangement is usually adopted. There are other plans, but perhaps none so thoroughly sound and convenient as the following. The crutch itself is made of a piece of flat steel cut away so as to leave a round boss at the bottom for the fork, and a round boss at the top to fit on a collet on the pallet arbor, a part projecting above to be embraced between a pair of opposing screws. On the collet is fixed a thin brass plate with two lugs projecting backwards from the frame, these lugs being drilled and tapped to receive the opposing screws in a line. The boss of the crutch lies flat against this plate, and is held up to it by a removable collet. The collet may be pinned across or fitted keyhole fashion, in either case so as to hold the crutch firmly, allowing it to move with a little stiffness under the influence of the screws. With this arrangement the adjustment to beat may be made with the utmost delicacy by slacking one screw and advancing the other, taking care that in the end they are well set home so as to make the crutch practically all one piece with the arbor. Milled heads are most convenient for these screws, and being placed at the top they are easily got at. The crutch should always be fitted with a fork to embrace the pendulum rod, as this ensures the impulse being given directly through the center, and with the same object the acting sides of the fork should be truly square to the frame. A slot in the pendulum rod with a pin acting in it is never so sure of being correct, as, although the surfaces may be rounded, it is very unlikely that the points of contact will be truly in the plane of the axis of the rod. The slightest error in this respect will tend to cause wobbling of the bob, although, to avoid this, great attention must also be given to the suspension spring, the pin on which it hangs, and the pin and the hole at the top of the pendulum rod. All these points must be in a true line, and the spring symmetrical on both sides of the line in order that the impulse may be given exactly opposite the center of the mass, otherwise wobbling must occur, although perhaps of an amount so small as to be difficult of detection, and this is not a matter of small importance, as it has an effect on the rate which could be mathematically demonstrated.

The frames of many regulators are made too large and heavy. In some cases there may be good reasons for making them large and heavy, but in most instances, and especially when the pendulum is not suspended from the movement, it would be much better to make the frames lighter than we frequently find them. Very large frames present a massive appearance, and convey an idea of strength altogether out of proportion to the work a regulator is required to perform. They are more difficult and more expensive to make than lighter ones, and after they are made they are more troublesome to handle, and the pivots of the pinions are in greater danger of being broken when the clock is being put together than when they are moderately light.

In a clock such as we have under consideration, where the frame is not to be used as a support for the pendulum, but simply to contain the various parts which constitute the movement, the thickness of the frames may with propriety be determined on the basis of the diameter of the majority of the pivots which work into the holes of the frames. The length of the bearing surface of a pivot will, according to circumstances, vary from one to two and a half times the diameter of the pivot. The majority of the pivots of our regulator will not be more than .05 or .06 of an inch in diameter; consequently a frame 0.15 of an inch thick will allow a sufficient length of bearing for the greater portion of the pivots, and will also allow for countersinks to be made for the purpose of holding the oil. If thin plates are used one or two of the larger pivots should be run in bushes placed in the frame, as described in Fig. 155.

The length and breadth of the frame, and also its shape, should be determined solely on the basis of utility. There can be no better shape for the purpose of a regulator than a plain oblong, without any attempt whatever at ornament. For our regulator a frame nine inches long and seven inches broad will allow ample accommodation for everything, as may be seen on referring to Fig. 157.

The plates are made of various alloys: cast-brass, nickel-silver, and hard-rolled sheet brass. It is difficult to make plates of cast-brass which would be even, free from specks, etc., but cast plates may very well be made of ornamental patterns and bushings of brass rod inserted, or they may be jeweled as shown in Figs. 154, 155, 156. Nickel, or German silver, makes a fine plate, but it is difficult to drill the small holes through plates of four-tenths of an inch in thickness, on account of the peculiar toughness of the metal, so that bushings are necessary. The best material where the holes are to be in the plates is fine, hard-rolled sheet brass; it should have about 4 oz. of lead to the 100 lbs., which will make it “chip free,” as clockmakers term it, rendering it easy to drill; the metal is so fine and condensed to that extent by rolling, that the holes can be made with the greatest degree of perfection. The many improvements in tools and machinery have effected great changes and improvements in clock-making. It once was quite a difficult task to drill the small holes in the plates with the ordinary drills and lathes; now we lay the plates after they are soldered together at the edges (which is preferable to pinning), on the table of an upright drill, and with one of the modern twist-drills the task is rendered a very easy one. After the pivot holes are drilled, we run through from each side a round broach, finished lengthwise and hardened, which acts as a fine reamer, straightening and polishing the holes exquisitely. A little oil should be used on the reamer to prevent sticking. The method of fitting up the pivot holes invented by LeRoy, a French clockmaker of some note, is shown in Fig. 154. It is a sectional view of the plate at the pivot hole. It will be observed that, instead of countersinking for the oil, the reverse is the case. A is a hardened steel plate counterbored into the clock plate B, and held in its place by the screws. There should be a small space between the steel plate and the crown of the arch for the oil. After the clock has been put together it is laid down on its face or side, a drop of oil is put to the pivot end, and the steel plate immediately put on; and the oil will at once assume the shape of the shaded spot in the drawing, being held in the position at the center of the pivot by capillary attraction, until it is exhausted by the pivots; the steel plates also govern the end play of the pinions. The pivot ends being allowed to touch the plates occasionally, the shoulders of the pinions are turned away into a curve, and, of course, do not bear against the plate, as in most clocks.

Glass plates may be used instead of steel, or rose cut thin garnets, or sapphires, with the flat sides smoothly polished, may be bought of material dealers and set in bezels like a cap jewel. They are very hard and smooth for the pivot ends, and the state of the oil at the pivots can be seen at any time. Clocks fitted up in this manner have been running many years without oiling.

When fitted up in this way the plates may be thicker. We have made the clock plates about four-tenths of an inch in thickness, which allows of counterboring, and admits of long bearings for the barrel arbor, which are so liable to be worn down in the holes by the weights; and the pivots of the pinions, by being a little longer, do not materially increase the friction. In first-class clocks, when all the materials are as hard as possible, the wheels and pinions high numbered, the teeth, pinions, pivots, and holes smooth, true, and well polished, the amount of wear is very slight, especially if the driving weight has no useless excess. Yet there are advantages in having some parts jeweled, such as the pallets and the four escapement holes. The cost of such jeweling is not an objection, while the diminished friction of the smooth, hard surfaces is worth the extra outlay. The holes can be set in the bushes described in Fig. 156, the end stones being cheap semi-precious stones, either rose cut or round.

For jeweling the pallets, dovetailed slots may be made so that the stones will be of a wedge shape; there is no need for cutting the slots right through as in lever watch pallets. The stones will be held more firmly if shaped as wedges lying on a bed of the steel and exposing only the circular resting curve and the driving face. The slots can be filed out and the stones ground on a copper lap to fit, fixed with shellac and pressed firmly home while warm. The grinding and polishing of the acting surfaces are done exactly as described for hard steel, only using diamond powder instead of emery. The best stones are pale milky sapphires, such as are useless as gems, this kind of stone being the hardest.

The holes may be much shorter when jeweled, as the amount of bearing surface required with stones is less than with brass; this results in less adhesion through the oil, and less variation of force through its changes of consistency. The ’scape wheel may also be thinner with similar results, and less weight to be moved besides. So the advantages of jeweling are worth consideration.

It is important to finish the wheels and pinions before drilling any holes in the plates and then to definitely locate the holes after trial in the depthing tool.

For the clockmaker’s use the next in value to the wheel cutting engine is a strong and rigid depthing tool, for it is by means of this instrument that the proper center distances of wheels and pinions can be ascertained, and all errors in sizes of wheels and pinions, and shapes of teeth, are at once detected before the holes are drilled in the plates. In fact, this tool becomes for the moment the clock itself; and if the workman will consider that as the wheels and pinions perform in the tool for the little time he is testing them, so they will continue to run during the life of the clock, he will not be too hasty in allowing wheels to go as correct when a hundredth of an inch larger or smaller, and another test, would, perhaps, make the pitching perfect.

There are various kinds of depthing tools in use, but many of them are objectionable for the reason that the centers are so long that the marking points on their outer ends, are too far from the point where the pitching or depthing is being tested, and the slightest error in the parallelism of these centers is, of course, multiplied by the distance, so that it may be a serious difference. Having experienced some trouble from this cause, we made an instrument with very short centers, on the principle that the marking points, or centers, should be as near the testing place as possible. We succeeded in making one with a difference of only three-fourths of an inch, which was so exact that we had no further trouble. It was made on the Sector plan, but upright, so that the work under inspection, whether wheels and pinions, or escapements, could be observed closely, and with a glass, if necessary.

It is very important that the posts or pillars and side-plates of clocks should be made and put together in the most thorough manner; the posts should be turned exact to length and have large shoulders, turned true, so that the plates, when put together without screws should fit accurately, for if they do not, when the screws are driven, some of the pivots will be cramped. We prefer iron for the posts, it being stiffer, and better retaining the screw threads in the ends, which in brass are liable to strip unless long and deep holes are tapped. Steel pillars should be blued after being finely finished, thus presenting a pleasing contrast. The plate screws should also be of steel, with large flat heads, turned up true, and having a washer next to the plate. Brass pillars are favored by many and are easier turned in a small lathe, but they should be much larger than the steel ones.

When the pillars are made of brass round rod of proper diameter is the best stock. If this cannot be procured, a pattern is turned from wood, and a little larger in every respect than the pillar is desired to be. If there is to be any ornament put on the pillar, it is never made on the pattern, because it makes it more difficult to cast, and besides, the ornamentation would all be spoiled in the hammering. The pattern must be turned smooth, and the finer it is the better will be the casting. After the casting is received the first thing to be done is to hammer the brass, and then center the holes, because it will be seen from Fig. 159 that there are holes for screws at each end of the pillar. Holes of about .20 of an inch are then bored in the ends of the pillars, and should be deep, because deep holes do no harm and greatly facilitate the tapping for the screws. After the holes are tapped, run in a bottoming tap and then countersink them a little, to prevent the pillar from going out of truth in the turning. It will depend a great deal on the conveniences which belong to the lathe the pillars are turned in as to how they will be held in the lathe and turned. If the holes in the ends of the pillars have been bored and tapped true, and if the lathe has no kind of a chuck or face plate with dogs, suitable for holding rods, the best way is to catch a piece of stout steel wire in the chuck and turn it true, cut a true screw on it, and on this screw one end of the pillar, and run the other end in a male center. However, if the screws are not all perfectly true, and the centers of the lathe not perfectly in line, this plan will not work well, and it will be necessary to catch a carrier on to the pillar and turn it between two male centers.

The dial feet are precisely the same as the pillars, only smaller. These dial feet are intended to be fastened in the frame by a screw, the same as the pillars; but it will be observed that the screw which is intended to hold the dial on the pillar is smaller. The dial feet will be turned in precisely the same manner as the pillars. For finishing the plain surfaces of the pillars and dial feet, an old 6 or 7-inch smooth file makes a good tool. The end of the file is ground flat, square or slightly rounded, and perfectly smooth. The smoother the cutting surface the smoother the work done by it will be. It is difficult to convey the idea to the inexperienced how to use this tool successfully. In the first place, a good lathe is necessary, or at least one that allows the work to run free without any shake. In the second place, the tool must be ground perfectly square, that is, it is not to be ground at an angle like an ordinary cutting tool. Then the rest of the lathe must be smooth on the top, and the operator must have confidence in himself, because if he thinks that he cannot turn perfectly smooth, it will be a long time before he is able to do it. A tool for turning the rounded part of the pillar, if a pattern of this style is decided on, is made by boring a hole, the size of the desired curve, in an old file, or in a piece of flat steel, and smoothing the hole with a broach and then filing away the steel. The shoulders should be smooth and flat, or a very little undercut, and the ends of the pillars should be rounded as is shown in Fig. 159, because rounded points assist greatly in making the frames go on to the pillars sure and easy, and greatly lessen the danger of breaking a pivot when the clock is being put together.

When a washer is used the points of the pillars project half the thickness of the washer through the frames, the hole in the washer being large enough to go on to the points of the pillars.

Figure 160 is an outline of the cock required for the pallet arbor, and the only cock that will be required for the regulator. It is customary, in some instances, to use a cock for the scape wheel and also for the hour wheel arbors, but for the scape wheel arbor I consider that a cock should never be used when it can be avoided. The idea of using a cock for the scape wheel arbor is to bring the shoulder of the pivot near to the dial and thereby make the small pivot that carries the seconds hand so much shorter; and so far this is good, but then the distance between the shoulders of the arbor being greater, when a cock is used the arbor is more liable to spring and cause the scape wheel to impart an irregular force to the pendulum through the pallets. This is the reason why I prefer not to use a cock except when the design of the case is such that long dial feet are necessary, and renders the use of a cock indispensable. In the present instance, however, the dial feet are no longer than is just necessary to allow for a winding square on the barrel arbor, and therefore a cock for the scape wheel is superfluous. It is better to use a long light socket for the seconds hand than put a cock on the scape wheel arbor in ordinary cases. Except for the purpose of uniformity a cock on the hour wheel is always superfluous, although its presence is comparatively harmless. The front pivot of the hour wheel axis can always be left thick and strong enough should the design of the case require the dial feet to be extra long.

For the pallet arbor, however, a cock is always necessary, and it should always be made high enough to allow the back fork to be brought as near to the pendulum as possible, so as to prevent any possibility of its twisting when the power is being communicated from the pallets to the pendulum. This cock should be made about the same thickness as the frames, and about half an inch broad. Make the pattern out of a piece of hard wood, either in one solid piece or by fastening a number of pieces together. The pattern should be made a little heavier than the cock is required to be when finished, and it should also be made slightly bevelled to allow it to be easily drawn from the sand when preparing the mould for casting. After it is cast the brass should be hammered carefully, and then filed square, flat, and smooth.

Screws are better and cheaper when purchased, but they may be made of steel or brass rod by any workman who is provided with a set of fine taps and dies. If purchased they should be hardened, polished and blued before using them in the regulator. The threads of screws vary in proportion to the size of the screw and the material from which it is made. A screw with from 32 to 40 turns to the inch, and a thread of the same shape as the fine dies for sale in the tool shops make, is well adapted for the large screws in a regulator. However, it is not threads of the screws I desire to call attention to so much, although it must be admitted that the threads are of primary importance. It is the shape of the heads and the points which is too often neglected.

A thread, or a thread and a half, cut down on the point of a screw, will allow it to enter easier than when the point is flat, round, or shaped like a center. This is not a new idea for making the points of screws, but the plan is either not known to many, or it is not practiced to the extent it ought to be.

The shape of the head of a screw should also always be based on utility, and the shape that will admit of a slit into it that will wear well should be selected. A round head ought never to be used, because a head of this shape does not present the same amount of surface to the screwdriver that a square head does. It is the extreme end of the slit that is most effective, and in round-headed screws this part is cut away and the value of the head for wearing by the use of the screwdriver is the same as if the head of the screw was so much smaller. A chamfered head may suit the tastes of some people better than a perfectly flat head, but in a head of this shape the slit must be cut deeper than in a square head, because the chamfered part of the head is of little or no use for the screwdriver to act against. The slits should always be cut carefully in the center of the head and the sides of the slit filed perfectly flat with a thin file and the slight burr filed off the edge to prevent the top of the head getting bruised by the action of the screwdriver. The shape of the slit which is best adapted for wearing is one slightly tapered, with a round bottom. The round bottom gives greater strength to the head, and prevents the heads of small screws from splitting.

I have dwelt at some length on these little details because a proper attention to them goes a long way in the making of a clock in a workmanlike manner, and it is desirable that the practical details should be as minute as possible.

The construction of the barrel is a subject which requires a greater amount of consideration than is sometimes bestowed upon it. We often meet with regulator barrels which have considerable more brass put into them than is necessary. The value of this extra metal is of little or no consequence. It is the unnecessary pressure the weight of it causes on the barrel pivots, and the consequent increase of friction, which is objectionable. For this reason the weight of the barrel, as well as the weight of every other part of the clock that moves on pivots, should be made no heavier than is absolutely necessary to secure the required amount of strength. In every instance, except when the diameter is required to be very small, the barrel should be made of a piece of thin brass tubing with two ends of cast-brass fastened into it.

Figure 161 is a sectional view of the ends of a barrel; the diagram on the right is the end where the great wheels rest against, and the one on the left is the other end. The insides of both these ends are precisely the same, but the outsides differ a little. It will be observed that there is a little projection near the hole on the outside of the front end. This projection is left with the view of making the hole in the center longer, and thereby causing this end to take a firmer hold on the barrel arbor. The back end, or the end that the great wheels rest against, and where the ratchet teeth are cut, is shaped precisely like the diagram on the right of Fig. 161. If you cannot get brass plate of sufficient thickness for the ends of the barrel they must be cast.

The patterns for these barrel ends should be made without any hole in the center, and in every way heavier and thicker than they are to be when finished, because it is difficult to obtain good and solid castings when the patterns are made thin, although it is by no means impossible to make them so. Like all brass castings used for the clockmaker’s purpose, they should be carefully hammered, and, although these pieces are of an irregular shape, they can be easily hammered regularly with the aid of narrow-faced hammers or punches, and with the exercise of a little patience. After hammering, the castings should be placed on a face plate in the lathe, and the tube which is to form the top part of the barrel fitted easy and without shake on to the flanges and the other parts of the castings turned down to the required thickness, and a hole a little less than 0.3 of an inch diameter bored in the center of each before it is removed from the face plate. The tube which is to form the top of the barrel should be no heavier than is just necessary to cut a groove for the cord, and for this regulator it should be 1.5 inch diameter outside measurement, 1.5 inch long, and turned perfectly true on the ends.

The hole in the front end of the barrel, which is the end nearest to the dial, should be broached a little from the inside, and the other end broached a little larger from the outside. The reason for broaching the holes in this manner is to cause the thickest part of the barrel arbor to be at the place where the great wheels work, because, in making a barrel for a regulator, it will generally be found that the arbor requires to be thickest in this particular place. The arbor should be made from a piece of fine cast steel a little more than 0.3 of an inch thick, and not less than four inches long. It is always well to have the steel long enough. This steel should be carefully centered and turned true, and of the same size and taper as the holes in the barrel ends. It is not necessary that the barrel arbor should be hardened and tempered, except on special occasions. In most cases it will last as long as any other part of the clock if it is left soft, and it is much easier to make when soft. Before fitting the arbor to the barrel ends it is well to place the ends into the tube that is to form the top of the barrel, because a better fit can be made in this way than when each is fitted separately. When the arbor has been fitted, a good and convenient way of fastening it together is, to use soft solder. It can be easily heated to the required degree of heat with the blow-pipe. A very little solder is sufficient for the purpose, and if the joints have been well fitted the solder will not show when the work is finished. Care should be taken to notice that the solder adheres to the arbors properly. Perhaps it would be well to mention here that, should the clockmaker not have access to a cutting engine with conveniences attached to it for cutting the barrel ratchet after the barrel has been put together, the ratchet should be cut first.

When the different pieces which constitute a barrel have been fastened together the brass work has next to be turned true, and the grooves cut for the cord to run in. It is best not to turn anything off the arbor till the grooves are cut, because they are usually cut smoother when the arbor is strong. The most important points to notice when turning a barrel is to be sure that the top is of equal diameter from the one end to the other, and that the bearing where the great wheels rest against are perfectly true, because, if the top of a barrel is of unequal thickness, the weight will pull with unequal force as it runs down, and if the bearing on the end be out of truth the great wheels will also be very liable to get out of truth, as their position on the barrel is altered by winding the clock up.

The shape of the outside of the barrel ends, as is represented in Fig. 161, will be found to be good and serviceable. AA is the bearing for the great wheels to rest against; BB is where the ratchet teeth are to be cut. There must be a little turned off the face of BB, as is shown in the diagram, so as to prevent the great wheel from rubbing on the teeth. The space between AA and the barrel arbor is turned smooth.

Although it is by no means an absolute necessity to have a groove cut in the top of the barrel, yet it is extremely desirable that there should be one, so that the cord may always be guided with certainty as the clock is wound up. It has long been a disputed question whether the cord should be fastened at the front end of the barrel and wind towards the back, or whether it should be fastened at the back and wind towards the front. I am not aware that there is any violation of principle, so far as the regularity of the power is concerned, whether the cord runs one way or the other. I understand it to be solely a question of keeping the weight clear of the case and the pendulum ball. In ordinary constructed regulator cases this object will be best attained by cutting the screw so that the cord can be fastened at the front of the barrel and wind towards the back; because in making it in this way, the weight is the length of the barrel farther away from the front of the case when it is wound up, and about the same distance farther away from the pendulum ball when it is nearly run down, than if the cord was fastened at the back end of the barrel and wound towards the front. The cutting of the groove is usually done in an ordinary screw cutting lathe.

In making the pivots on a barrel it is the usual custom to make the back pivot smaller than the front one but, with all due respect for this time-honored custom, I would direct a little attention to the philosophy of continuing to make the barrel pivots of a regulator in this manner. Friction varies with pressure; a large pivot has a greater amount of friction than a smaller one, because the pressure on the sliding surface of the revolving body is farther away from the center of motion in one case than in the other. In regulators where the barrel pivots are of a different size, the effective force of the weight will vary slightly according as the weight is fully wound up or nearly run down. In one instance the pressure of the weight is more directly on the large pivot than it is on the smaller one; and in the other instance the pressure is more directly on the small pivot than it is on the larger one, and when the weight is half wound up, or half run down, the pressure is equal on both pivots.

In the center pinion and in some of the other arbors of a clock, it is sometimes necessary to make one pivot considerably larger than the other; but in these cases the difference in the size of the pivots does not affect the regularity of the transmission of the power, because the pressure that turns the wheel is always at the same point. In a regulator barrel, however, the pressure of the cord and weight shifts gradually from one end of the barrel to the other, as the clock runs down, and when the pivots are of unequal thickness the power is transmitted nearly as irregular as if the top of the barrel was slightly conical and both pivots of the same size. For the above reason, I think, that it will be plain to all that in a fine clock both of the barrel pivots should be made of an equal diameter. The front pivot should be made no larger than is absolutely necessary for a winding square, and when we take the fact into consideration that a fine clock with a Graham escapement requires considerable less power to keep it in motion than an eight-day marine chronometer does, we may safely conclude that the winding squares of many regulators of the Graham class might be made smaller. A pivot about 0.2 of an inch will secure a sufficient amount of strength. For the reasons mentioned above, the back pivot should be exactly the same diameter, and although the effects of friction will be slightly greater when both pivots are of an equal size, still the force of the weight will be transmitted more regularly, which is the object aimed at. Where the plates are bushed a length of two to three diameters is long enough for the pivot holes.

The stop works, maintaining powers and general arrangement of the great wheel, ratchets and clicks, have been so fully described and illustrated on pages 282 to 290, Figs. 83 to 87, that it would be useless duplication to repeat them here, and the reader is therefore referred to those pages, for full particulars. This is also the case with the purely mechanical operations of cutting the wheels and pinions, hardening, polishing, staking, etc.; all have been fully treated; but there are some further considerations which may be mentioned here. The practical value of making pinions with very high numbers is very much over-rated. I know of two clocks situated in the same building that are compared every other day by transit observation. They both have Graham escapements and mercurial pendulums, and are equally well fitted up, and as far as the eye can detect, they are about equally well made in all the essential points, with only this difference: one clock has pinions of eight, and the other pinions of sixteen leaves, yet for two years one clock ran about equally as well as the other. In fact, if there was any difference, it was in favor of the clock with the eight-leaved pinions. In giving this example, I must not be understood to be placing little value on high numbered pinions. I know that in some instances they can be used to advantage. The idea that I want to illustrate at present is, that it is not in this direction that we are to search for the means of improving the rates of regulators.

A pinion as low as eleven leaves can be made so that the action of the tooth will begin at or beyond the line of centers; but as eleven is an inconvenient number to use in clockwork, we may with great propriety decide upon twelve as being a sufficient number of leaves for all the pinions used in a regulator having a Graham escapement.

In arranging the size of the wheels in a regulator, the diameters of the center and third wheels are determined by the distance between the center of the minute and the center of the seconds hand circle on the dial. As the dials of regulators are usually engraved after the dial plates have been fitted, and as the position of the holes in the dial for the center and scape wheel pivots to come through determines the size of the seconds circle, it may be well to mention here that, for a twelve-inch dial, two and a half inches is a good distance for the center of the minute circle to be from the center of the seconds circle. Consequently the center and third wheels must be made of such a diameter as will raise the scape wheel arbor two and a half inches from the center arbor, and the other wheels must be made proportionally larger, according to the number of teeth they contain.

We all know what a difficult matter it is to make a cutter that will cut a tooth of the proper shape; but when the cutter is once made and carefully used, we also know that it will cut or finish a great number of wheels without injury. For this reason, those who are contemplating making only one, or at most but a few regulators, will find the work will be greatly simplified by making the wheels of a diameter proportionate to the number of teeth they contain, and for all practical purposes the cutter that cuts or finishes the teeth of one wheel will be sufficiently accurate for the others. If we make all the pinions with the same number of leaves they will also all be nearly of the same diameter, and may be cut, or rather the cutting operation may without any great impropriety be finished with one cutter.

An opinion prevails among a certain class of workmen that the teeth of the great wheel and leaves of the center pinion should be made larger and stronger than the other wheels and pinions, because there is a greater strain upon them than on the other. However reasonable this idea may seem, a little consideration will show that in the case of a regulator, with a Graham escapement, where so little motive power is required to keep it in motion, an arrangement of this nature is altogether unnecessary. The smallest teeth ever used in any class of regulators are strong enough for the great wheel; and if there be a greater amount of strain on the teeth of the great wheel in comparison with the teeth of the third wheel, for example, then make the great wheel itself proportionately thicker, as is usually done, according to the extra amount of strain that it is to bear. The teeth of wheels and the leaves of pinions wear more from imperfect construction than from any want of a sufficient amount of metal in them.

If we assume the distance between the center of the minute and the center of the seconds circle to be 2½ inches, and also assume that the clock will have a seconds pendulum, and all the pinions have 12 leaves, and the barrel make one turn in 12 hours, then the following is the diameter the wheels will require to be, so that the teeth may all be cut with one cutter, and also the number of teeth for each wheel:

Great wheel 144 teeth.
Diameter 3.40 inches for the pitch circumference.

Hour wheel 144 teeth.
Diameter 3.40 inches for the pitch circumference.

Center wheel 96 teeth.
Diameter 2.26 inches for the pitch circumference.

Third wheel 90 teeth.
Diameter 2.11 inches for the pitch circumference.

Scape wheel 30 teeth.
Diameter 1.75 inches for the pitch circumference.