Pierre le Roy, eldest son and successor to Julien le Roy, the companion and friend of H. Sully, had the merit of accomplish ing the great desideratum of making the compensation on the balance itself. In the chronometer, which was presented to the king of France the 5th August, 1766, and obtained the prize of the Academy of Sci ences of Paris the 31st of the same month, that author executed a compensation in the balance, which he has fully explain ed in his description of that machine. (" Memoires stir la meilleure Maniere de mesurer le Temps en Mer, qui a rempor te le Prix double au judgement de 1'Aca demie Royale des Sciences. Contenant la Description de la Montre a Longitudes, presentee I sa Majeste le 5 Aout, 1766." Par M. le Roy, Horloger du Roi, pp. 41 to 44.) This memoir accompanies the ac count of Cassini's voyage in 1668, pub lished in 1770, This compensation is composed (fig. 4.) of two thermometers, t K t A K, of mercury and spirits of,wine, made each in the form ofa parallelgogram, except in the upper branch, which bears the ball containingthe spirits of wine, and is a little bent downwards ; the mercury is in the lower part, and the vertical branch of the tube, A K, is open at the upper end. These two thermometers are placed opposite one another, the axis of the balance being in the same plane with the centrical lines of the tubes ; and the thermometers and balance are solidly attached together, and form a well pois ed and steady regulator. .Ert the middle temperature of the atmosphere, the quicksilver stands at K A t K; but, when an increase of heat occurs, the alcohol, byits expansion, forces the mercury from the exterior branch, t K t, towards A K, and a portion of the mass of the regulator contracts by approaching the centre. On the contrary, if the variation consist of an additional degree of cold, the mercury moves towards the exterior branch, and the weight towards the circumference of the balance becomes greater. Thus, if the thermometers are well adjusted, the effects of all the changes of temperature in the balance will be compensated, and the regulator will act with the same uni formity as if its figure were not liable to such alterations. • Peter le Roy mentions his knowledge of Harrison's expansion curb, acquired soon after he made his thermometrical balance ; and he constructed a balance accordingly on the principle of Harrison's curb, the arms or arcs of which act by flexure, and are adjusted by moveable weights. He effect of such arms by experiment ; but gave the pre ference to his own mercurial compensa tion. Peter le Roy's second invention is the same as is now used : but though so publicly declared in the face of the French government and academy in 1766, Arnold took an English patent for it in 1782.
We have not yet taken any notice of improvements made in the escapement, because after all the plans proposed for this most essential part of chronometers, the principle of what is called the de tached escapement is the only one now used ; and, being established upon long experience, seems to merit the prefer ence given to it over all the constructions proposed till now. We shall content our selves with stating in a general manner the beginning and progress of that es. capement.
In all the escapements known till the middle of the last century, the escape wheel was in continual contact with the pallets belonging to the axis of the ba lance wheel ; and the friction proceeding from this circumstance may be consider ed as a principal source of irregularity in the going of the watches. Suppose that a regulator should be made so perfect as to be exactly isochronal, while vibrating in a free situation, that advantage would be diminished or lost as soon as it was pla ced in connection with a train of wheels; and the errors would be more or less,'ac cording to the nature and quantity of fric tion in the escapement. It would be, therefore, extremely useful to secure to the regulator a perfect liberty of vibra tion, except during the short intervals of time which may be necessary for the ac tion of the escape wheel, to give it a new impulse. This ingenious idea was also started by P. le Roy, and carried into exe cution by the same artist, in a model which he presented in 1748 to the Academy of Sciences of Paris, and is described in the collection of machines approved by that society for the same year, vol. vii. p. 385.
The effect or action of le Roy's escape ment in few words is this : An escape wheel is kept in repose bya lever detent.
The balance unlocks the detent, and re ceives an impulse or stroke on a pallet through a part of every second vibration ; and during great part of its course it is free and detached.
About the year 1755, according to Count de Bruhl, the late Mr. Thomas Mudge invented a detached escapement, and applied it to a watch which he made for the king of Spain, Ferdinand VT. This is the same escapement that was used by the late Josiah Emery, in his chro nometers, some of which have gone very well. It differs from the constructions which we have already explained, both in the detent and in the communication of the impulse, which in this mechanism takes place at every vibration : but the date will not suffer us to consider it as the first invention of the detached escape ment.
This justly celebrated artist afterwards made a chronometer, in which the vibra tions of balance were kept up by second ary springs attached to two pallets, each of which was wound up by the last wheel of the train during the time employed by the balance in its vibration, unconnected with that pallet. Though this inventioft is highly ingenious, and was rewarded by parliament, it is now generally considered as unsafe in the locking of the hooks, or detent parts, which terminate the pallets. Mr. Alexander Cumming executed a si milar escapement for clocks long before Mudge actually carried his idea into ef fect, though it has been contended that Mudge hack the mere notion much earlier. But Mr. Cumming, to whom our art and the other branches of mechanics are highly indebted for his labours, and his Treatise on Clock and Watch Making, made his detents separate from his pal lets, and by that means avoided the chief defect of the construction afterwards adopted by Mudge. Our limits will not, however, allow us to pursue these and other improvements and variations, adopt ed by our own and by foreign artists.
We must confine ourselves to the con struction used at present by the English watchmakers ; and shall begin with that of the late Mr. Anold, as described in his statement, presented by his son to the board of longitude.
The teeth of the escape wheel (fig. 5.) are of a cycloidal shape, in the face part, which is intended for action, the section of which, with those of the two other sides, form a sort of mixed triangle. B B d represents the detent, which is formed of a flexible piece or spring, bending be 'ween C and N ; and in the part N B d, which is stronger than the other, is fixed the locking pallet, a, opposite an adjust ing screw F. The pallet, projecting be low the spring detent, locks upon the in terior angle of the tooth; suspending the motion of the escape wheel, and leaving the balance to vibrate free, as pointed out in the preceding escapements. The ac tion of the spring detent (for the joint of the detent is itself a spring) presses the locking pallet against the screw, F, ex cept at the time of unlocking the wheel. A very delicate spring N e, called the discharging or unlocking spring (and al so the tender spring,) is attached by one cud, N, to the spring detent, C BNB a; and, passing under the adjusting screw, F, extends a little beyond the extremity, d, of the detent itself. H H is a circu lar piece attached to the axis of the ba lance, and, o, the discharging pallet. This pallet, when the balance is in motion from e to d, presses against the end of the dis charging spring, n e ; and carrying it, to gether with the locking spring, B B d, disengages the locking piece, a, out of the internal angle of the tooth, with which it was in contact ; and the escape wheel then communicates a new power to the balance, by its impulse on a pallet, m, which is fixed or set in the aperture of the circular piece. As soon as this is done, the springdetent, or locking spring, falls back to its position against the ad justing screw, F ; and the pallet, by re ceiving or intercepting the next tooth, stops the motion of the escape wheel. When the balance returns from d to e, the unlocking pallet acts again on the extre mity of the discharging spring ; but this being very delicate, gives way without disturbing the detent or locking spring ; and the balance, after suffering a trifling degree of resistance by that contact, con tinues its free vibrations. At the next vi bration the unlocking takes place ; and the action of the escapement proceeds successively, as explained before.