Related photographs
No confidently matched photograph of this subject has been found yet. These photographs show a closely related record, and are not examples of the type described on this page.
The Earth turns through 360 degrees in roughly twenty-four hours, so each hour of difference between local time and a reference meridian corresponds to fifteen degrees of longitude. Local time could be found at sea by astronomical observation. The difficulty was knowing the reference time accurately after weeks at sea, when small daily errors accumulate into large positional errors.
The British Longitude Act of 1714 offered rewards for a practicable method. Two broad approaches competed: the astronomical lunar-distance method, which used the Moon's position among the stars together with published tables, and the mechanical approach of carrying a timekeeper. These were not simply rivals; for decades navigators used lunar observations and timekeepers to check one another.
John Harrison's sea clocks H1 to H3 and his large watch H4 are the best-known stage of the mechanical story. H4's performance on trial voyages in the 1760s showed that a portable timekeeper could hold time well enough to be useful. Its construction was complex and costly, however, and later practice moved toward designs that could be made and repaired more routinely.
French makers including Pierre Le Roy and Ferdinand Berthoud, and English makers such as John Arnold and Thomas Earnshaw, developed features that became characteristic of the later marine chronometer: a detent escapement, a compensation balance to reduce the effect of temperature on rate, and a fusee to even out mainspring force. The instrument was typically mounted in gimbals inside a box so that it stayed level as the ship moved.
A chronometer was not expected to be perfect. What mattered was a known, stable rate. Before a voyage it was rated against an observatory or time signal, and the navigator applied the recorded daily gain or loss. Ships often carried more than one chronometer so that an instrument that changed its behaviour could be identified by comparison.
For collectors and researchers, a surviving chronometer's box, maker's number, rating certificates and naval or observatory records can matter as much as the movement. Chronometers are precision instruments with fragile detent springs; they should not be transported unsecured or run after long storage without assessment by a specialist conservator or chronometer maker.
Trials, rewards and the first copy
Royal Museums Greenwich recounts that H1 was tested on a voyage to Lisbon in 1736 and that, in 1737, the Commissioners of Longitude agreed a payment of £500, half in advance, for Harrison to build an improved clock. H2 never went to trial because Harrison found a fundamental flaw, and H3 occupied him for 19 years from 1740. After the 1764 trial to Barbados, run alongside tests of lunar distances and Jupiter's satellites, the Board recommended in 1765 that Harrison receive £10,000 once he demonstrated H4's principles, with the remainder due only when other makers had shown they could produce similar timekeepers. H4 was then tested for ten months at the Royal Observatory from May 1766 under Nevil Maskelyne, where it did not perform well, and the published results reopened the dispute with Harrison.
The copy the Board required was made by Larcum Kendall, whom Harrison recommended; Kendall agreed to copy H4 part for part but would give no guarantee of its performance. Its record at Greenwich says that K1 went with James Cook on his second voyage of 1772–75 and performed so well that Cook called it his 'never failing guide', while Kendall's simplified K2 of 1771, which omitted the remontoir, was later aboard the Bounty. Formal rating continued at the Observatory: from 1822 the Admiralty ran thirteen annual Premium Trials, each lasting twelve months and buying the best chronometers, until they were discontinued in 1836.
Written from the guidance and records listed under Sources, credited to each publisher and put in our own words; each statement was checked against the cited page.