The celebrated George Graham im proved this escapement very much, by taking off part of the slope furthest from the points of the pallets ; instead of which part, he formed a circular or cylindrical face, having its axis in the centre of mo tion. Pallets of this kind are seen on the opposite side of the wheel at E and G. having H for their centre or axis. A tooth of the wheel is seen resting upon the circular inner surface of the pallet G, which is not therefore affected by the wheel, excepting so far as its motion, arising, from any other cause, may be af fected by the friction of the tooth. If the vibration of the pendulum be supposed to carry G outwards, the slope surface will be brought to the point of the tooth, which will slide along it, and urge the pallet outwards during this sliding action. When the tooth has fallen from the point of this pallet, an opposite tooth will be received on the circular surface of E, and will not affect the vibration, except ing when the slope surface of E is carried out so as to suffer the tooth to slide along it. In the two former escapements, there is always a certain portion of vibration takes place after the drop which drives the pallets back, and causes the index also to recede through a small arc this has been distinguished by the name of a recoil. Other considerable objections, besides that of the continued action of the maintaining power, have been made against escapements with a recoil ; but it I would lead us too far into the minute de , partments of this subject to discuss them. The escapement of Graham, and all such as have no recoil, have been called dead beat escapements, because the index for seconds falls directly through its arc, and remains motionless on the line of division till the next vibration. It may be observ ed, that the maintaining power in Gra ham's escapement may be applied during a small portion only of the vibration ; and that an increase of the maintaining power tends to enlarge the arc of vibra tion, but scarcely interferes with its velo city.
The effect of the escapement which has been termed horizontal, because the last wheel in watches of this construction has its plane parallel to the rest of the sys tem, is similar to that of the dead beat escapement of Graham. In fig. 3, the ho rizontal wheel is seen with twelve teeth, upon each of which is fixed a small wedge supported above the plane of the wheel, as may be seen at the letters A and B. On the verge of the balance there is fixed part of a hollow cylinder of steel, or other hard material, the imaginary axis of which passes through the pivots of the verge. C. represents this cylindrical piece, into which the wedge D. may be supposed to have fallen. While tile vibration causes the cylindrical piece to revolve in the di rection which carries its anterior edge to wards the axis of the wheel, the point of the wedge will merely rub the internal surface, and no otherwise affect the vibra tion of the balance than by retarding its motion. But when the return of the vi bration clears the cylinder of the point of the wedge 3), the wheel will advance, and the slope surface of the wedge, act ing against the edge of the cylinder, will assist the vibration of the balance. When the edge of the cylinder- arrives at the outer point of the wedge D, its posterior edge must arrive at the position denoted by the dotted lines of continuation ; im mediately after which the wedge or tooth E will arrive at the position e, and rest on the outer surface oft he cylinder, where it will produce no other effect than that of retardation from friction, as was re marked with regard to the wedge D, un til the course of the vibration shall bring the posterior edge of the cylinder clear of the point of the wedge. In this last si
tuation the wedge will act on the edge of the cylinder, and assist the vibration, as in the former case, until that edge shall ar rive at the outer or posterior point of the wedge ; immediately after which the leading point will tall on the inner sur face of the cylinder in the first position, as was shewn in the wedge D.
Time-pieces, with a pendulum regula tor, are certainly the most perfect, when they are kept in a fixed situation ; and, for that reason, these are the only sort used in astronomical observatories. But external motion is so contrary to the re gularity of their performance, that no sea chronometer has been since attempted to be constructed upon that principle. The balance regulator remained, as affording the only method by which the desired uniformity might be obtained in portable machines ; and the great improvement made in that regulator, by the addition of a spiral spring, may be' considered as one principal cause of the perfection which has been since attained in them. The first invention of attaching a spring, to give to the balance, by its elasticity, a power which renders the action of this sort of regulator similar to that of gravity in the pendulum, is undoubtedly due to Dr. Hooke, though it is not so clear whether he ever applied it in the shape of a spiral, as has been so long practised since. F. Berthoud, in his " Histoire de la Mesure du Temps," (vol. i. pp. 134 to 141,) gives a body of extracts from several works re lative to this subject ; and concludes, that Dr. Hooke only applied a straight spring to the balance, and that M. Huygens im proved upon that idea, and contrived the spiral spring, which is more favourable to the vibrations of the balance. M. Huy gens, indeed, applied in France a balance spring, the account of which has been published in the Philosophical Transac tions for 1675, No. 112 ; but Dr. Hooke, in the Postscript to his Description of He lioscopes, asserts, that the hint was taken from the experiments he had made in 1664, in Gresham College, where he ex plained above twenty several ways, by which springs might be applied to do the same thing. • In relating the progress of time-piece making, we must not omit mentioning the use of precious stones particularly rubies, to form the holes in which the pivots of the wheels turn, and the pallets upon which the action of the teeth is exercis ed. These jewels, by the high polish given to them, reduce the quantity of friction ; and, not being liable to the wear which takes place in metal rubbing upon metal, the machine, with that addition, not only becomes more durable, but ac quires a degree of uniformity in the mo tion of the pieces, which is very favoura ble to the regularity of its going.