The Carpenter Who Tamed the Sea: How John Harrison’s Clocks Solved Longitude

On the night of October 22, 1707, a British fleet coming home from the Mediterranean ran straight into the Isles of Scilly, the scatter of granite islands and reefs off the southwestern tip of England. Admiral Sir Cloudesley Shovell’s flagship, HMS Association, struck the Gilstone Ledge around 8 p.m. and went down in minutes. HMS Eagle, HMS Romney and the fireship Firebrand followed. Somewhere between 1,400 and 2,000 sailors drowned within sight of home. Shovell’s body washed up the next morning in a cove on St Mary’s, seven miles from the wreck.

The fleet’s navigators had believed they were safely south of the islands, heading up the English Channel. They were wrong by a few dozen miles, and that was all it took. The wreck became one of the worst disasters in British naval history, and it fed a national anxiety that had been building for two centuries: once a ship lost sight of land, nobody aboard could say exactly where it was.

The answer, when it finally came, was not a new star chart or a better compass. It was a clock, built by a self-taught carpenter from rural Lincolnshire named John Harrison, who spent more than 40 years fighting the scientific establishment to prove that a machine could carry the time across an ocean.

Latitude was easy. Longitude was a clock problem

Sailors had long known how to find their latitude, their position north or south. Measure how high the Sun stands at noon, or the height of the Pole Star at night, look up a table, and you have it. The equator gives you a natural starting line.

Longitude, your position east or west, has no such line. What it does have is the Earth’s steady spin. The planet turns 360 degrees in 24 hours, which works out to 15 degrees every hour, or one degree every four minutes. If you know the time at home, say Greenwich, and you can work out your own local time from the Sun, the difference between the two clocks is your longitude. Local noon arrives three hours after Greenwich noon? You are 45 degrees west. The Dutch mathematician Gemma Frisius proposed exactly this method in 1530. The physics was never the obstacle. The hardware was.

Here is how unforgiving the arithmetic is. At the equator, one degree of longitude spans about 60 nautical miles. A clock that drifts by just one minute puts you 15 nautical miles off. To find a ship’s position within half a degree after a six-week Atlantic crossing, a timekeeper had to stay within about two minutes over the whole voyage, or roughly three seconds a day. The best clocks of the early 1700s were pendulum clocks, and a pendulum is useless on a pitching deck. Heat and cold changed the length of metal parts and the stiffness of springs. Salt air rusted them, and the oils of the day turned to gum.

In 1714, two mathematicians, William Whiston and Humphry Ditton, summed up the expert view: watches were “so influenc’d by heat and cold, moisture and drought” that “all wise Men give up their Hopes from them in this Matter.” Isaac Newton agreed that a watch might help for a few days, but “when longitude at sea is lost, it cannot be found again by any watch.” The same principle, carried to absurd precision, is what lets GPS satellites find you with atomic clocks today. In 1714 it looked hopeless.

It is worth being honest about Scilly. Naval historians who have studied the surviving logbooks argue that the fleet’s worst error was probably in latitude, made worse by charts that put the islands in the wrong place and by compasses in poor condition. The famous tale of a sailor hanged for warning Shovell he was off course has no contemporary evidence behind it. But the disaster still dramatized how primitive navigation was, and the pressure it added helped push Parliament to act.

A £20,000 question

In July 1714, Parliament passed the Longitude Act. It offered £10,000 for any method that found longitude at sea to within one degree, £15,000 for within 40 minutes of arc, and £20,000 for within half a degree. The top prize was a fortune, worth millions of pounds in today’s money. The test was a voyage from Britain to the West Indies, about six weeks of sailing. A panel later known as the Board of Longitude judged the entries and could hand out smaller grants to promising ideas.

Spain had offered a longitude prize as early as 1567, and the Dutch soon followed. Neither had been won. Galileo had pitched the moons of Jupiter as a celestial clock; their eclipses worked beautifully for mapmakers on land but were impossible to observe through a telescope on a rolling deck. The favorite of the astronomers was the “lunar distance” method: the Moon moves across the background of stars about 13 degrees a day, roughly its own width every hour, so with good enough tables a navigator could measure its angle from a star and read off Greenwich time. The tables did not yet exist, and even when they did, each sighting could take hours of arithmetic.

Very few people bet on a clock. John Harrison did.

A carpenter who built clocks out of wood

Harrison was born in 1693 in Foulby, Yorkshire, and grew up in Barrow upon Humber in Lincolnshire, where he followed his father into carpentry. He was also a choirmaster, and he built clocks in his spare time. His first longcase clock, finished in 1713 when he was 20, had a movement made almost entirely of wood. Three of those early wooden clocks still survive.

What he lacked in formal training he made up for with an obsessive feel for friction and temperature. He built his movements from oak and lignum vitae, a dense tropical hardwood that is naturally oily, and designed them to run without any added oil. He invented the grasshopper escapement, a nearly frictionless way of releasing the clock’s power one tick at a time. He also invented the gridiron pendulum, which used alternating rods of brass and iron. Brass expands more than iron when it warms, and Harrison arranged the rods so that their expansions cancelled out, keeping the pendulum the same effective length in summer and winter. By the late 1720s, his wooden precision clocks kept time to about a second a month, far better than anything else in Britain.

In 1730 Harrison travelled to London with plans for a sea clock. The Astronomer Royal, Edmond Halley, sent him to George Graham, the country’s leading clockmaker, who was impressed enough to lend him money.

H1, H2, H3: thirty years of brass

Harrison’s first marine timekeeper, now called H1, took about five years to build. Instead of a pendulum, it used two big dumbbell-shaped balances linked so that the ship’s rolling pushed them in opposite directions and the effects cancelled out. It was huge, brass and gloriously strange, and in 1736 it was sent on a trial voyage to Lisbon aboard HMS Centurion. On the way home aboard HMS Orford, Harrison told the officers that the headland ahead was not Start Point, as they believed, but the Lizard. He was right, which meant the ship was about 60 miles off its reckoning.

Lisbon was not the West Indies, so the prize was not on the table, but the Board granted £500 for development. Harrison moved to London and built H2, a sturdier version, then abandoned it when he realized a flaw: his bar balances could be thrown off when the ship swung around its vertical axis while changing course. He started H3 in 1740 and kept working on it for nearly two decades. It never quite met his standard, but along the way he invented two things engineers still use: the bimetallic strip, two bonded metals that bend with temperature, and the caged roller bearing.

The watch that changed everything

The breakthrough came from an unexpected direction. Around 1752, Harrison had the watchmaker John Jefferys build him a pocket watch to his own design, with temperature compensation and a mechanism that kept it running while being wound. It performed so well that Harrison realized the answer might not be a giant sea clock after all, but a watch.

He began H4 in 1755 and finished it in 1759. It looks like an oversized pocket watch, about 13 centimetres across and weighing 1.45 kilograms, but inside it is unlike any watch of its day. Its large balance swings fast, five ticks a second, which makes it far less sensitive to the ship’s motion. Its escapement pallets are made of diamond. A bimetallic “thermometer kirb” adjusts for temperature, and a tiny remontoire evens out the driving force.

In November 1761, H4 sailed for Jamaica aboard HMS Deptford in the care of Harrison’s son William; John, then 68, stayed home. On the way out, William used the watch to predict an earlier landfall at Madeira than the crew expected, and he was right. When the ship reached Kingston after 81 days at sea, the watch, corrected for its known daily rate, was just five seconds off. That is an error in longitude of about one nautical mile.

The Board moves the goalposts

Harrison expected the £20,000. He did not get it. The Board argued the result might have been luck, and that a machine that took six years to make was hardly a practical solution. A second trial was arranged, this time to Barbados in 1764, and this time the lunar distance method would be tested side by side. The astronomer in charge was a young clergyman named Nevil Maskelyne, the lunar method’s champion.

H4 did it again. It kept time to within 39 seconds, fixing the longitude of Bridgetown to within about 10 miles. Maskelyne’s lunar observations came in at around 30 miles, also within the Act’s limits but far more work. Then, in February 1765, Maskelyne was appointed Astronomer Royal, which automatically made him a member of the Board judging the contest.

The Board agreed H4 had met the strictest terms of the 1714 Act, but it changed the deal. Harrison would get £10,000 once he revealed how H4 worked, and the rest only after other makers proved they could copy it. In 1765 Harrison had to take the watch apart in front of a panel of experts. In 1766 Maskelyne received H4, sealed in a box, and tested it at Greenwich for ten months. He reported that it performed poorly. Harrison was furious, and the feud became a public pamphlet war.

It is easy to cast Maskelyne as the villain, and Harrison’s supporters did. The fuller picture is more interesting. The Board had a fair worry that one hand-built masterpiece was not a fleet-wide solution, and Maskelyne’s own project, the Nautical Almanac, which published lunar distance tables from 1767, gave ordinary navigators a cheaper way to find longitude for decades. It also made Greenwich the reference meridian on British charts, which is a big part of why the world adopted Greenwich as the prime meridian in 1884.

The King steps in

By 1772 Harrison was 79, worn out, and convinced the Board was holding his watch hostage. He had finished a second sea watch, H5, and appealed to King George III. The King tested H5 himself over ten weeks of daily checks and found it accurate to within a third of a second a day. Reportedly furious with the Board, he encouraged Harrison to petition Parliament directly.

In 1773, at 80, Harrison received £8,750 from Parliament. Counting the Board’s earlier payments, he collected £23,065 over his lifetime, more than the prize itself, but he was never formally declared its winner. Nobody ever was. He died on March 24, 1776, just shy of his 83rd birthday.

“Our trusty friend the Watch”

The final vindication came from the other side of the world. The watchmaker Larcum Kendall, who had witnessed H4’s disassembly, built a faithful copy, now called K1. It cost £450. Captain James Cook took it on his second voyage around the world, from 1772 to 1775, aboard HMS Resolution. After initial skepticism, Cook was won over. In his log he called it “our trusty friend the Watch” and “our never-failing guide,” and he reported that “Kendall’s watch has exceeded the expectations of its most zealous advocate.” Three timekeepers by John Arnold sent on the same voyage did not hold up as well. K1 proved H4 was no fluke.

A later copy, K2, went with Lieutenant William Bligh on HMS Bounty and was carried off by the mutineers. It was not recovered from Pitcairn Island until 1808.

Harrison’s watches were too complex to mass-produce. At first a chronometer cost something like 30 percent of the price of a ship. It took the next generation, makers such as John Arnold and Thomas Earnshaw, to simplify the design with the spring detent escapement and bring prices down. By the early 1800s, sailing without a chronometer was considered reckless, and insurers agreed. Ships in the Thames set their clocks by the red time ball on the Royal Observatory, which John Pond installed in 1833 and which still drops at 1 p.m. The U.S. Navy kept mechanical Hamilton chronometers in service as backups until 1988, when GPS finally took over.

Harrison’s last laugh: Clock B

Near the end of his life, Harrison made a claim that got him mocked in print. He said that a pendulum clock built on his principles “shall perform to a second in 100 days.” In an age when the best observatory clocks drifted far more than that, it sounded absurd.

In the 1970s, the artist and clockmaker Martin Burgess began building two clocks from Harrison’s late writings, using some modern materials such as duraluminium and invar. The second, Clock B, sat unfinished in his workshop for decades until a collector bought it in 2009 and the London firm Charles Frodsham & Co. completed it. In April 2014 it was sealed in a clear case at the Royal Observatory Greenwich. On January 6, 2015, a formal 100-day trial began. When it ended in April, Clock B had lost just five-eighths of a second. Guinness World Records named it the most accurate mechanical clock with a pendulum swinging in free air. Harrison, 239 years after his death, had been telling the truth.

Why a clock was the answer

It is tempting to see longitude as a story about navigation. Really it is a story about time. The Earth turns at a steady rate, the same rotation that Foucault’s pendulum later made visible in Paris, so a reliable clock is a map of the sky’s motion that you can carry in a box. Harrison’s machines belong to a long line of gear-driven calculators that stretches back to the Antikythera mechanism, and forward to the satellites overhead. The lesson about tiny errors adding up never went away, either; ask the engineers behind the Mars Climate Orbiter.

Today H2, H3 and H4 are in the collection of Royal Museums Greenwich, and H1 is undergoing conservation. H4 is usually kept stopped, because unlike the earlier sea clocks it needs oil and would wear as it ran. Since 2006, a memorial stone in the nave of Westminster Abbey, near the graves of the clockmakers George Graham and Thomas Tompion, has carried his name. Running through it is an inlaid strip of brass and stainless steel, a nod to his bimetallic strip, inscribed with the stone’s own longitude: 0° 7′ 35″ W. It took a carpenter, a watch and nearly 60 years of argument, but the man who was told that no watch could find a lost ship’s position ended up remembered with a line of longitude of his own.

Further reading