788 



HOKOLOGY 



escape-wheel, being released, overtakes the impulse- 

 pallet and drives it on till their paths diverge and 

 they separate. The wheel is again brought to a 



Fig. 12. 



stand by the locking-pallet of the detent, which, 

 on being released by the discharge-pallet, has sprung 

 back to its original position. The roller, having 

 made its vibration, is brought back by the spring. 

 In the return the discharge-pallet forces itself past 

 the end of the gold-spring, the impulse-pallet clears 

 the teeth of the escape- wheel, and the balance goes 

 on till the momentum is exhausted, when the 

 spring induces another vibration, the wheel is again 

 unlocked, and the impulse-pallet gets another 

 blow. By receiving impulse in one direction and 

 unlocking at every alternate vibration only, the 

 chronometer-balance is more thoroughly detached 

 than any other. It is very delicate, however, and, 

 though the most perfect known, it cannot stand 

 rough usage, and is not so suitable for ordinary 

 pocket-watches as a good lever. At sea the 

 chronometer is hung in Gimbals (q.v.), so as to 

 be always horizontal whatever the motion of 

 the vessel. 



In watches, even more than in clocks, varia- 

 tions of temperature, unless provided for, produce 

 variations in the rate of going. A rise in the 

 temperature makes the balance expand, and there- 

 fore augments its moment of inertia. It diminishes 

 the elasticity of the spring ; and the time of vibra- 

 tion of the balance, which depends upon the 

 moment of inertia directly, and upon the elastic 

 force of the spring inversely, is increased the 

 watch, that is, goes more .slowly. A fall in the 

 temperature is attended by opposite results, the 

 watch going more rapidly than before. Compensa- 

 tion can obviously be made in either of two ways 

 by an expedient for shortening the effective length 

 of the balance-spring as the temperature rises, so 

 as to increase the elastic force of the spring ; or by 

 an expedient for diminishing the moment of inertia 

 of the balance as the temperature rises, so as to 

 correspond to the diminution of the force of the 

 spring. The first method was that made use of 

 by John Harrison (q.v.) in his chronometer, and 



Fig. 13. 



it depended on a laminated bar of brass and steel 

 fixed at one end, called a compensation curb ; the 

 free end carries two curb pins, which embrace the 

 balance-spring, and, as the bar shrinks and expands, 

 regulate the length of the spring. It is never used 



now. An adaptation of the other method, invented 

 in 1782 by John Arnold, and improved by Thomas 

 Earnshaw, is that which is always employed now. 

 Fig. 13 shows the 

 form employed for 

 marine chronometers, 

 and fig. 14 that for 

 pocket chronometers 

 and watches : t, a, t' 

 (fig. 13) is the main 

 bar of the balance ; 

 and t, b, t', b' are two 

 compound bars, of 

 which the outer part 

 is of brass and the 

 inner part of steel, 

 carrying weights, c, 

 c', whose position may 

 be shifted to or from 

 the fixed end, according as the compensation is 

 found on trial to be less or more than is desired. 

 Brass expands more with heat and contracts more 

 with cold than steel ; consequently, as the tempera- 

 ture rises the bars with their weights, being fixed at 

 one end to the main bar, bend inwards at the free 

 end, and so the moment of inertia of the balance is 

 diminished ; as it falls they bend outwards, and the 

 moment of inertia is increased ; and of course the 

 diminution or the increase must be made exactly 

 to correspond to the diminution or increase in the 

 force of the spring. The screws, d, d, fitted to the 

 fixed end of each of the compound bars are used 

 for bringing the chronometer to time ; sometimes 

 the smaller ones are dispensed with. In fig. 14 the 

 principle is the same : a, a, a, a are the time screws 

 (equally distributed in the watch-balance); the 

 others are for compensation, and their positions may 

 be shifted or larger ones substituted if necessary. 



The modern marine chronometer is just a large 

 watch fitted with all the contrivances which ex- 

 perience has shown to be conducive to accurate 

 time-keeping e.g. the cylindrical balance-spring, 

 the detached spring-detent escapement, and the 

 compensation-balance. Harrison's chronometer, 

 mentioned above, was the first, and was completed 

 after many years of study in 1736. For a de- 

 scription, see British Horological Journal, vol. xx. 

 page 120. After many trials and improvements, 

 and two test voyages to America, undertaken for 

 the satisfaction of the commissioners, the last of 

 which was completed on the 18th September 1764, 

 the reward of 20,000, which had been offered by 

 government for the best time-keeper for ascertain- 

 ing the longitude at sea, was finally awarded to 

 him. Harrison made many other inventions and 

 improvements in clocks and watches, including 

 his maintaining spring to the fusee, to keep the 

 works going while being wound ; a form of remon- 

 toire escapement, &c. 



Somewhat later than this several excellent chron- 

 ometers were produced in France by Berthoud and 

 Le Roy, to the latter of whom was awarded the 

 prize by the Academic Royale des Sciences. Pro- 

 gress was still made in England by Mudge, Arnold, 

 and Earnshaw, to whom prizes were awarded by 

 the Board of Longitude. The subsequent progress 

 of watch -making has been chiefly directed to the 

 construction of pocket-watches on the principle of 

 marine chronometers, and such accuracy has been 

 obtained that the average error is reduced to one 

 second a day. 



The compensation of an ordinary balance chron- 

 ometer cannot be made perfectly accurate for all 

 degrees of temperature, but only for two points. 

 The explanation of this lies in the fact that, while 

 the variations of elastic force in the spring go on 

 uniformly in proportion to the rise or fall of the 

 temperature, the inertia of the balance varies, not 



