John Harrison's Marine Timekeepers, H1 to H4

Latitude can be read from the sun. Longitude cannot: it is a measure of time difference, and to know how far east or west a ship is, the navigator must know what time it is at a reference meridian at the moment of a local observation. On land that is easy. On a wooden ship, in temperature swings, salt air and constant motion, no clock of 1714 could hold a rate well enough to be worth carrying.

The Longitude Act of 1714 put public money behind the problem in three tiers: 10,000 pounds for a method accurate to one degree of longitude, 15,000 pounds for forty minutes of arc, and 20,000 pounds for half a degree. Half a degree at the equator is about thirty nautical miles, and to hold it across a voyage to the West Indies a timekeeper has to keep within roughly three seconds a day for weeks.

The Act created a body of commissioners drawn from politics, the Navy, astronomy and mathematics. It is a fact worth holding on to that the Board of Longitude did not convene at all until about 1737, twenty-three years after it was created, and that what brought it together was John Harrison.

Harrison built H1 at Barrow upon Humber between 1730 and 1735, probably with the help of his brother James, and brought it to London in 1735. The Royal Museums Greenwich catalogue calls it the first relatively successful marine timekeeper of any kind.

It is 620 millimetres high, 680 wide and 450 deep as displayed, in brass, bronze and steel with oak, mahogany and lignum vitae. Two interconnected swinging balances, coupled so that the motion of the ship acts on both equally and cancels, replace the pendulum that no seagoing clock could use. It compensates for temperature. And it runs without lubrication, through an extensive set of anti-friction devices, because Harrison had understood that eighteenth-century oils thickened, dried and changed viscosity with heat, and that a clock which depends on oil is a clock whose rate depends on the weather.

That decision is the intellectual centre of the whole project. Lignum vitae is self-lubricating, which is why a tropical hardwood is a structural material in a precision instrument.

H1 went to sea in May 1736, out to Lisbon aboard HMS Centurion and home aboard the Orford. It is the only one of the four large machines ever tested at sea.

The episode that made Harrison's name happened on the return leg. Harrison, reading his own timekeeper, identified a landfall as the Lizard where the ship's officers had reckoned the Start. He was right and they were about sixty miles out. A single instance of a navigator's dead reckoning being corrected is not a proof of anything statistically, but it was the demonstration that mattered politically, and on 30 June 1737 the Commissioners of Longitude approved a payment of 500 pounds, with 250 pounds paid up front, against Harrison's promise to build a better machine within two years.

That meeting is the first the Board ever held on a submission. The organisation existed because of this problem and it was activated by this object.

H2 was finished in 1739 to the promised timetable. It is a larger and more solidly built version of H1, 686 millimetres overall and weighing 40 kilogrammes, with a remontoire added so that the drive to the two balances is as uniform as possible. It never went to sea. Harrison had found a fundamental fault in the behaviour of its balances and withdrew it himself.

He began H3 in 1740 and worked on it for nineteen years. It is catalogued at 1759 and stands 622 millimetres high. It never achieved the accuracy he needed, because the problem was in the physics of the balance and the spring rather than in the workmanship, and he did not have the theory.

Two things came out of those nineteen years and both are still in use. The bimetallic strip, made by bonding two metals with different expansion rates so that the compound bends with temperature, is in thermostats worldwide. The caged roller bearing, which keeps rolling elements from touching each other, is in nearly all modern mechanical engineering. A record that describes H3 only as a failure has missed the two most consequential objects Harrison ever produced, neither of which is a clock.

Around 1751 to 1753 Harrison commissioned a pocket watch from the London watchmaker John Jefferys, incorporating his own ideas about the balance. It performed better than it had any right to, and it told him that the direction he had been travelling in for twenty years was wrong: the answer was not a bigger machine with slower balances but a small one with a fast, heavy, high-amplitude balance.

He began H4 in 1755. It is catalogued at 1759. The dial is 102 millimetres across, the whole object 165 by 124 by 28 millimetres, and it weighs 1.45 kilogrammes: a watch the size of a dessert plate. It beats five times a second, far faster than an ordinary watch of the period and with larger oscillations, which is what makes it insensitive to the motion of a ship. Its materials include diamond and ruby, used as bearing and pallet jewels, which is why the Royal Museums Greenwich materials list for a timekeeper reads like the description of a piece of jewellery.

The Jefferys watch matters for attribution. H4 is Harrison's, but the object that showed him the way was made by another man to Harrison's specification, and the record should say so.

H4 went to Jamaica in 1761 and 1762 in the care of Harrison's son William. Over eighty-one days it lost three minutes thirty-six and a half seconds in total; after the rate was applied the residual error was five seconds slow, which corresponds to an error in longitude of about one and a quarter minutes of arc, roughly one nautical mile. That is inside the top tier of the 1714 Act by a wide margin.

The Board did not pay. It ruled that the result could have been luck and ordered a second trial. H4 went to Barbados in 1764, and kept time to within thirty-nine seconds, giving an error in the longitude of Bridgetown of less than ten miles. Nevil Maskelyne, who was appointed to conduct the astronomical observations against which the watch was judged, was at that date the leading practitioner and advocate of the lunar-distance method, which is to say the champion of the competing solution was placed in charge of assessing the timekeeper. He became Astronomer Royal the following year.

The Board accepted that the watch had met the terms of the Act. It then declined to pay the prize.

What Parliament passed in 1765 was not the 1714 Act being honoured. It was a new bargain imposed on a man who had already met the old one.

Harrison would receive 10,000 pounds, less what he had already been paid, when he demonstrated the principles of H4 to nominated experts. He would receive a second 10,000 pounds only when other makers had shown that watches of the same performance could be made by somebody other than him. To collect the first half he had to dismantle H4 in front of a committee and explain every part of it, and he had to surrender all of his timekeepers.

The second condition is worth reading carefully, because it is a condition no inventor can satisfy by his own effort: it made Harrison's payment contingent on the future behaviour of his competitors. The Board's stated purpose was to ensure the nation got a reproducible method rather than a single irreplaceable object, which is a defensible public policy and was also, in practice, the instrument by which an eighty-year-old man was kept out of his money.

The disclosure produced a book. In 1767 the Board of Longitude published The Principles of Mr Harrison's Time-Keeper, with plates. A private trade secret was converted by statute into a public specification, and that publication is why the H4 escapement is documented at all.

The surrender was real. The Royal Museums Greenwich catalogue entry for H2 states it plainly: Harrison kept H2 at his own house until 1766, when the Astronomer Royal claimed it under the conditions of the longitude prize. The same applies to the rest.

Maskelyne received H4 on 5 May 1766 and tested it at the Royal Observatory for ten months. It did not perform well, and he published the results. Harrison disputed them, and the objections he raised were not frivolous: a watch that has been dismantled by a committee, transported, and run in a fixed location by a man with a declared interest in the competing method is not being tested under the conditions the Act contemplated.

No source consulted here resolves whether the poor result was the watch, the handling or the observer, and this record does not arbitrate it. What is not in doubt is the custody: from 1766 the objects were in the hands of the state, in the keeping of the principal advocate of the rival system, and their maker never had them back.

The 20,000 pound prize was never awarded to anybody, for any method, in the form the 1714 Act described. Harrison received 23,065 pounds in total across roughly forty years, in instalments, most of them discretionary encouragements rather than prize money.

The sequence: 500 pounds voted in 1737 with 250 paid immediately; a further 500 in 1741; further sums across the middle decades; 10,000 pounds under the 1765 Act on disclosure, less prior payments; and finally 8,750 pounds voted directly by Parliament in 1773, when Harrison was eighty, after he petitioned the King. He died three years later.

Two things follow for anyone reading a price into this. First, the 20,000 pounds is a figure that is quoted constantly and was never paid as a prize; the modern equivalents attached to it in popular accounts, of the order of several million pounds, depend entirely on which conversion method is used and should not be repeated without naming the method. Second, none of these sums is the value of the objects. They are the price of a method, paid by a government, to a man it did not much want to pay.

The Board's second condition, that others must be able to make such a watch, was tested on an object. Larcum Kendall was commissioned to make a copy of H4. K1 went with James Cook on his second and third voyages and performed well enough that Cook's published praise settled the question in the Navy's mind. A later Kendall, K2, was aboard the Bounty at the mutiny.

K1 is the reason marine chronometry became a trade rather than a single man's achievement, and it is also the object most often photographed and captioned as Harrison's watch. It is not. It is Kendall's copy, made after H4 had been dismantled and published, and any record on H1 to H4 has to say so, because the two objects look alike by design.

The four timekeepers went into official custody in 1765 and 1766 and then, in large part, into neglect. What rescued them was one man working unpaid.

Rupert Gould, a retired Royal Navy officer, spent years after the First World War restoring all of them at his own cost and on his own time. He gave them the designations H1 to H5 that the whole world now uses, which means that the very names in this record's title are his, not Harrison's. He published The Marine Chronometer: Its History and Development in 1923, still the standard account.

The Board of Longitude itself was abolished in 1828, its work considered done, and replaced by a Resident Committee for Scientific Advice to the Admiralty whose members included Thomas Young, Michael Faraday and Edward Sabine.

H2, H3 and H4 are on display at the Royal Observatory Greenwich, catalogued to the National Maritime Museum as ZAA0035, ZAA0036 and ZAA0037. H1, ZAA0034, is not currently in the gallery: the museum states that it is undergoing conservation at the Prince Philip Maritime Collections Centre, and a replica stands in its place.

That detail is the kind of thing this archive exists to record. A visitor who goes to Greenwich to see H1, or a writer who describes the four machines running side by side, is describing a display that is not currently there, and the object in the case is a replica. The museum says so; almost nothing written about Harrison does.

H5, the second watch, is not part of this record and is not in Greenwich at all. It belongs to the Worshipful Company of Clockmakers and has been shown at the Science Museum in London since 2015.