Britain's Momentous Contribution To Horology

An accurate way to tell the time has been a cornerstone of business, warfare, government, transport and social activities for the past two hundred years. The techniques were developed over the previous five hundred years. Nearly all the important components were invented and developed in Britain, the rest being invented in France. It allowed Britain to dominate world timepiece markets for 300 years. It employed tens of thousands of skilled workmen. It even played a small part in the growth of the British Empire, through pioneering work on marine chronometers. Here we review Britain's momentous contribution.

MB used to be a jeweller. We loved working on mechanical clocks and watches. Their history is linked with some of the greatest scientists and inventors that have ever lived. The most famous of them is John Harrison, who (sort of) won the Longitude Prize by devising an accurate marine chronomter. His contribution - as we explain in our blog about him here - was not a great as most people think. A dozen or so others made more significant contributions, none of whom particularly stand out. We have therefore decided to write a blog about them collectively.

How mechanical clocks work

Mechanical clocks do not work as most people think. We tend to get the basics right: the gears and hands are connected to a spikey thing known as an 'escape wheel', which usually tick-tocks with some pincers known as an 'escapement'. In most cases, the escapement is connected to a pendulum or spring balance wheel. The natural assumption is that this pendulum or balance wheel drives the movement. Actually, the movement drives the pendulum or balance wheel. This is shown on the Wiki 'anchor' escapement animation below. The two rounded pallet faces that are in contact with the escape wheel are known as 'lifts', because the escape wheel lifts each side alternately to keep the pendulum swinging.

So, a mechanical clock works roughly like this. An escape wheel is attached to some sort of energy source - typically a wound spring, a weight or a battery - that would make it spin freely, were it not held in place by an escapement. The escapement is linked to something that oscillates, typically a pendulum, spring balance wheel or verge post. As the oscillator passes a particular point in its cycle, the escapement releases the escape wheel, which progresses exactly one tooth before getting caught by the escapement on the other side. Hence the terms: the 'escapement' temporarily allows the 'escape wheel' to 'escape'. Each time the escape wheel turns, it drives a gear chain to move the hands and it pushes the oscillator to keep the clock running.

In early clocks the escape wheel pushed constantly on the escapement, first in one direction then stopping the swing before pushing in the other direction. In later clocks, the oscillator swings freely, just receiving a tiny kick from the escape wheel on each advance to replace the tiny amount of energy that it loses through friction.

The first mechanical clocks

Mechanical clocks date back to the late 13th century. We can be pretty sure that they did not exist before 1271, because that was when medieval astronomer Robert the Englishman (sometimes known as Robertus Anglorum) said that horologists were trying to make mechanical clocks but had yet to succeed.

The oldest surviving mechanical clock in the world belongs to Salisbury Cathedral. They claim that it was installed in 1387. The Wells Cathedral clock is only a few years younger. Both are on display in their cathedrals. The Salisbury clock movement is still working (picture above). The Wells clock movement is on display at the Science Museum.

C F C Beeson and John D North, horologists, reckon that mechanical clocks had already been around for 100 years before the surviving Salisbury clock was installed. Beeson lists what he thinks are the the seven oldest mechanical clocks. They were at Dunstable Priory (1283), St Paul's (1286), Merton College (1288), Norwich Cathedral Priory (1290), Ely Abbey (1291), Canterbury Cathedral (1292) and Salisbury Cathedral (1306). Another one is recorded at Lincoln Cathedral in 1324, with a note saying that clocks were now commonplace in English cathedrals, abbeys and churches.

There is no proof that any pre-1327 clock was mechanical. Beeson's list is based on diary notes and sales receipts which just talk about the installation of clocks. They might have been clepsydrae (water clocks), which had been around for 200 years. But there is some circumstantial evidence that they were mechancial.

The Annals of Dunstable Priory say that their clock was installed above the Rood Screen. The plumbing for a water clock on a rood screen would be complicated, noisy and undevout. North notes that the seven oldest clocks listed above were too expensive and took too long to build to have been clepsydrae: "Taken singly the records are easy to view with scepticism, but taking them together, and noting especially that relatively large sums of money are involved in payment for the materials used, they persuade us that the mechanical clock had arrived on the scene."

The first definitive account of a mechanical clock is in Richard of Wallingford's 1327 manuscript Tractatus Horologii Astronomici. John D North, who wrote Wallingford's biography entitled 'God's Clockmaker', explains his history. He was orphaned at age 10 and raised by monks. He spent 15 years, on and off, studying mathematics, astronomy and theology at Oxford. John Leland's 16th century diary says that he was a fellow at Merton College. From humble beginnings, he became Britain's most brilliant medieval astronomer and mathematician. In between his stints at Oxford he became a monk. Soon after graduating he became abbot of St Albans. During a visit to confirm his appointment with the Pope, he contracted leprosy and died six years later. He spent his last years designing and commissioning the clock described in Tractatus Horologii Astronomici. He named it 'Albion'.

Albion's primary purpose was to mark times of prayer and service with a bell, but it could also forecast astronomical and astrological events, including lunar eclipses. It was an incredibly sophisticated device for its day. In our opinion, there is no way even someone as brilliant as Wallingford could have designed it completely from scratch. We guess that the movement was based on a clock that he already knew intimately, presumably the one at Merton College. We guess his contribution was not much more than designing the astrological and astronomical gearing, presumably driven by a link to one of the main clock pinions.

Wallingford commissioned Roger and Lawrence of Stoke to build his clock. Stoke is a long way from St Albans by medieval standards. The most obvious reason for choosing them is that they had already built successful mechanical clocks elsewhere. If we are right that Wallingford learned of mechanical clocks at Merton, we guess that it was made by the Stokes and that this is how Wallingford knew of them. The earlier Norwich Cathedral clock was commissioned to a Roger of Stoke; presumably the same person.

If the clocks at Norwich, Merton and St Albans were made by the Stokes, perhaps they made the others too. All of Beeson's clocks had similarly high costs and similarly long installation periods. Note that there was a regular and falling gap between the construction of the clocks up to Canterbury. It looks like they were built by the same clockmakers, presumably the Stokes family, who gradually improved their construction speed. If St Albans was mechanical, exactly as Besson and North argue, all the others were probably mechanical too.

As far as anyone knows, there were no mechanical clocks outside England until 1335. It is often assumed that England had a greater need than other major countries, because our dismal weather made sundials pretty useless for most of the year. But early clocks were not used to tell the time. They did not have hands. They were designed to ring bells that marked times for prayer and service. Clepsydrae had been doing this in Britain and the rest of Europe for centuries. We think that early mechanical clocks were exclusively English because Roger of Stoke invented the technology and, like WD40, kept the secret of how to make them within his family.

The history of escapements

Escapements are the core of all mechanical clocks. It is the escapement that converts energy stored in the power source - whether it is a weight, spring or battery - into isochronous repetitions of the movement. More than 300 different designs have been invented over the years. Mostly they were for niche purposes or for fun. We will run through the most important of the rest in historical order. 

According to North, Albion had what is now known as a 'strob' escapement. There are no surviving examples of an original strob escapement clock. North built one from Wallingford's description. He reckons the motion was similar to one sketched 170 years later by Leonardo da Vinci in Codex Atlantica (above). Similar but not the same. They both have two parallel escape wheels with offset perpendicular pegs, but Albion had a semi-circular double pallet whereas da Vinci had a parallel bar pallet, which he helpfully explodes for us (outlined in red above). We will return to this below.

At roughly the same time that Albion was being built, mechanical clocks started to appear in continental Europe: in Milan (1335), Padua (1344), Genoa (1353), Bologna (1356), Chartre (1359) and Ferrara (1362), according to Dr Karen Smyth. None of these clocks survives either. We can however be fairly sure that they had crown wheel verge escapements, like the animation above, because Giovanni Dondi gives a detailed description of the verge in his Astrarium, an astronomical clock he built between 1348 and 1363.

Dondi's father, Jacopo, built the 1344 clock in Padua. Giovanni learned the technology from his father, so the Padua clock probably had a verge escapement too. If the second and the last of Smyth's clocks had verge escapements, it is reasonable to assume that those in the middle did too. From Genoa, there is a regular three year gap between the clocks, the same as in England. It suggests that one team built all the post-Padua clocks. Probably not Jacapo or Giovanni because they had full time professional jobs; they made clocks as a hobby. We guess they were made by extended family members or by someone they had trained.

Every known clock and watch in Britain and continental Europe for the next two hundred and fifty years had a crown wheel verge escapement. This raises the interesting question of why the strob escapement died out in England and why it never took off in Europe. It is not as if the verge is notably superior. They were similar sizes, similarly inaccurate and similarly difficult to make. But the strob movement could be wood or iron, whereas the verge would quickly lose accuracy if made of wood. Wooden movements should have been cheaper to manufacture, which should have given them a reasonable market share. They would be made by carpenters and wood turners, which should have created competition with those made by metalworkers. It never happened. Why?

Early clocks had no faces. They were specifically designed to ring bells that alerted the faithful that it was time for services or prayer. Other devices, including clepsydrae, had been used for this purpose for hundreds of years. Some of them triggered bell ringing machines that hit the bell at regular intervals. Some of those bell ringing machines used escapements to implement the regularity of the mechanism. 

There is no material difference between the mechanism needed to strike a bell at, say, two second intervals and the mechanism needed to move a cog at two second intervals. It makes sense then that the first clocks were adapted from bell ringing machines and that the first clockmakers previously made bell ringing machines. There are surviving examples of verge escapement bell ringing machines in Europe. John North speculates that the strob mechanism was used in some English bell ringing machines.

If we are right about some of our conclusions above, strob escapements appeared in early English mechanical clocks because the Stokes family made strob escapement bell ringing machines, while verge escapements appeared in early Europe machanical clocks because Jacopo Dondi made verge escapement bell ringing machines or reverse engineered one.

As for why the strob died out, we suspect the Black Death was culpable. It arrived in England in June 1348. The Stokes would have been still working on Albion at the time. This is the last clock known to have used a strob escapement. It is feasible that it is a one off, perhaps designed by Wallingford specifically for Albion. We think it more likely that Black Death wiped out the Stokes or forced their retirement, and that they took the technology to their graves.

So, how come Leonardo was drawing a wooden strob escapement in the early 16th century? He liked to draw clock escapements. Half a dozen of them appear in his various works. He didn't invent them all. There is no evidence that he invented any of them. Perhaps he just liked taking clocks apart and drawing their escapements. So, perhaps he took a strob escapement clock apart. We think not. Leonardo's pallet is different from that described by Wallingford. There is no record of a wooden clock movement before Leonardo's. We guess that he re-invented the strob escapement.

The weighted bar across the top of the earliest verge escapements is known as a 'foliot'. In the mid-14th century the foliot bar was replaced by a foliot balance wheel, which was smaller and more reliable. Salisbury Cathedral's clock was probably made with a foliot balance wheel. The one there now is not original. Its escapement was replaced by an anchor escapement and pendulum in the 17th century, to improve accuracy. Curators restored what they believe to be the original design of verge and foliot wheel escapement, but they cannot really be sure. Much the same seems to have happened to the 1392 Wells Cathedral clock and, most probably, to every verge and foliot clock that could be adapted to use a pendulum. As far as we know, the oldest timepieces with their original verge and foliots are those that could not have a pendulum, like chamber clocks.

Early clocks had to be large and stationary because they were powered by weights. The advent of spring power in the 15th century allowed clocks to be smaller and portable. Perhaps the earliest known spring powered timepiece is 'Burgunderuhr' (below right), a chamber clock now in the Germanisches Nationalmuseum. They claim it was made in 1435, although there is a lot of scepticism. The oldest in Britain, albeit incomplete and incredibly drab in comparison, is a French table clock dating from 1450 (above), that belongs to the V&A (which we last saw on loan to the British Museum).

Verge escapements are in constant frictional contact with the escape wheel, which makes their timing sensitive to the amount of power applied to them. Early clocks were powered by weights, which applied a constant force, so the clocks ran at a constant speed. Clockmakers refer to this as 'isochronism'. Springs lose torque as they unwind. A spring powered verge escapement clock will therefore tend to run increasingly slow as its mainspring unwinds. Some form of compensation is required.

The most common mainspring torque compensation system is known as a 'fusée'. There was a less common alternative system known as a 'stackfreed'. A fusée uses cat gut or chain that is wound around a cone shaped post to connect the mainspring to the escape wheel - as seen on the Tompion watch above. The technology originated with crossbow windlasses, to increase leverage as the tension grows in the bowstring.

Burgunderuhr has a fusée. If it was really made in 1435, it would be the oldest clock fusée in the world. The oldest clock fusée in the world with a secure date is the 1525 Jacob Zech astronomical clock (above left), on display in the Watch and Clock Gallery at the British Museum.

Watches first appeared at the start of the 16th century. The first watchmaker was German, Peter Henlein. The oldest watch with a definitive date is 'Melanchthon’s Watch' (above left), from 1530 and attributed to Henlein. It is on display at the Walters Art Museum in Baltimore.

These early watches were worn around the neck, as a pendant. They were known, rather unkindly, as Nuremburg eggs, because they were so big. They were also horribly inaccurate, typically losing an hour or more a day. The oldest English made watch of this style is by Michael Nouwen dating from the turn of the 17th century (above centre). It is in the Met Museum in New York. The oldest made in England and still here (above right) is also by Michael Nouwen, dating from 1609 and in the British Museum.

Clock design changed little until 1657, when the great Dutch mathematician, physicist and inventor Christiaan Huygens patented the pendulum clock. Pendulums improved the accuracy of clocks some 60-fold. He was not the first to come up with the idea. Galileo Galilei, the great Italian astronomer and inventor, had the same concept 20 years earlier. He had also invented an improved escapement to work with it. Sadly, he had gone blind by this time and he died soon after. He passed the design and construction details to his son, but he too died before finishing it. As far as anyone knows, Huygens never saw their design. He therefore must have independently devised the pendulum clock, albeit based on scientific work on pendulums published by Galileo, Wren and Hooke.

Huygens' pendulum clock had a verge escapement. The pendulum had to swing through 100° to make it work. If the pendulum was long enough for the clock to be reasonably accurate, it would swing too much to fit in a case. Moreover, the pendulum had the capability to be accurate to perhaps 15 seconds a day but verge escapements could not maintain accuracy to more than a few minutes a day. There was a need for a better escapement. Galileo realised this and had designed a new escapement for his pendulum clock. That design was lost following the death of his son.

Robert Hooke, yet another great name in the history of science, came to the rescue with his 'anchor' escapement - Wiki animation above - towards the end of 1657. He demonstrated it to the Royal Society a few years later. The earliest known anchor escapement clock is the one commissioned by Christopher Wren, made by Joseph Knibb, that was installed at Wadham College in 1671. The clock face is still there in the outer quad but apparently the movement has been transferred to the Museum of Science History, over the road. It was not on display when we were last there in 2019. Most standing clocks and most tower clocks used the anchor escapement through to the mid-19th century.

The Wadham clock has caused some confusion about the invention of the anchor escapement. Some people assume that Knibb invented the anchor escapement, because most of his long case clocks have anchor escapements, and he was making long case clocks before Hooke's presentation. They would be wrong. Photos show that his early cased clocks have mini-verge escapements with no pendulum.

Other experts ascribe the anchor escapement to William Clement, not least because he publicly claimed to have invented it. He might even have applied for a patent in the early 1670s. The best that can be said is that he re-invented it. We suspect he did nothing of the sort. We guess that he was spoofing - pretending that he had invented the anchor escapement and that he had a patent pending - in order to disuade competitors from building rival anchor escapement clocks.

It makes perfect sense to us that Wren would commision Knibb to make the Wadham clock with Hooke's anchor escapement. He had been close friends with Hooke since they were at university. They were both alumni of Wadham College. Wren would have been at Hooke's Royal Society presentation where the anchor escapement was revealed. He would certainly have known its advantages over the verge. Indeed, it might originally have been his idea, since the impetus many of Hooke's best inventions came from Wren innovations.

Anchor escapement clocks were a lot more accurate than verge escapement clocks but not accurate enough for many requirements. They also had two practical drawbacks, both of which can be seen on the animation above. One is that the escapement is in almost constant frictional contact with the escape wheel. This makes it sensitive to the power being applied from the escape wheel - i.e. it is far from isochronous - which necessitates the use of a fusée and which causes excessive wear on the pallets and bearings. Another is that the escape wheel is forced backwards during the cycle which upsets the consistency of the pendulum's natural beat and which causes more wear on the pallets, bearings and cogs.

Pendulums work through gravity. Watches cannot use them because of the angles at which they are used. Small clocks cannot use them because a consistently rhythmic swing needs a long pendulum.

Before 1660, watches and small clocks were still using verge escapements. They were horribly inaccurate. In 1660, Hooke and Huygens independently invented the spring balance wheel. It performed the same oscillating function as a pendulum but did not rely on gravity. The oscillation is shown on the Gliphy animation above (with a modern lever escapement).

Spring balance wheels and anchor escapements improved pocket watch accuracy from an hour a day to a few minutes a day. Hooke and Huygens disputed who had come up with the idea first. Recently discovered Royal Society papers confirm that Hooke had priority, although Huygens was first to file a patent and the first to put a spring balance wheel into a working timepiece.

in 1675, Richard Towneley, a great scientist in his own right, addressed the two major issues with the anchor in his 'deadbeat' escapement. In effect, the deadbeat is an enhancement of the anchor, with a more acute angle on the escape wheel teeth and with angular rather than rounded pallets. Its low-friction non-recoil operation can be seen in the Mfrasca animation above. The deadbeat is more accurate and more durable than the anchor, but was initially both difficult and expensive to manufacture.

Thomas Tompion was the first to use the deadbeat in a practical timepiece when he put them into year-going precision regulators, including two sent to Astronomer Royal John Flamstead at the Greenwich Observatory in 1676. A replica has recently been installed in the location originally reserved for it at the Royal Observatory. One of the original clocks is at the British Museum, the other is at the National Maritime Museum, Greenwich.

Deadbeat escapements were popularised in normal high-end clocks from 1715 by Tompion's protégé George Graham. It was only used in niche high-end products until the mid-19th century, when better machine tools reduced manufacturing costs enough for them to be used in even the cheapest of pendulum clocks. Subsequently, almost all pendulum clocks have used the deadbeat escapement.

The deadbeat also proved popular in tower clocks, where it could be accurate to within a few seconds a day, as long as the weather was favourable and birds absent. In the 1750s, the French clockmaker Jean-André Lepaute modified the deadbeat to use pins instead of teeth, which allowed the escapement to be easily retrofitted to older anchor escapement tower clocks. His pin version of the deadbeat is common in French tower clocks. The Graham version is more common elsewhere.

Meanwhile, in 1704 the English (Hugenot descent) watchmaker Pierre Debaufre came up with an enhancement of the deadbeat escapement to make it work in pocket watches. This became known as the 'Debaufre' escapement, which is easily identifiable by having two escape wheels. A Debaufre watch with a Debaufre escapement was given to Sir Isaac Newton, who was impressed by its performance. He showed it to Henry Sully, who used it as the basis for the world's first marine chronometer (we do not count the one supposedly made by Jeremy Thacker, which is now thought to have been a hoax). Sully actually modified the escapement, but most experts thought original was better. Sully's chronometers were accurate to a few seconds a day, a huge improvement over what had gone before.

The requirement to accurately measure longitude became pressing after the appalling loss of life when Admiral Shovell's fleet grounded off the Isles of Scilly in 1714. The government set up the Board of Longitude to encourage solutions. John Harrison made the best known submissions. His first three Longitude submissions used his own 'grasshopper' escapement. It can be seen in the H1, H2 and H3 clocks at the Royal Observatory. The grasshopper was accurate enough to win the Longitude prize in principle, but it proved to be too delicate for sea voyages. Harrison's winning H4 clock used a hybrid verge-anchor escapement of his own invention. Despite its accuracy, it was not low friction and it had many of the practical drawbacks of standard verge escapements. It was replaced in subsequent chronometers.

Harrison returned to the grasshopper escapement late in life, when he incorporated it into a clock that he claimed would be the most accurate in the world. Unfortunately, he died before it could be made. Nearly 300 years later, British watchmaker Martin Burgess made a clock using Harrison's design that did indeed prove to be the most accurate mechanical atrmospheric pressure clock ever created. It lost less than a second over three months.

In 1748 French clockmaker Pierre Le Roy invented the 'detent' escapement specifically for chronometers. A minor design flaw detracted from their accuracy. English watchmaker Thomas Earnshaw fixed the flaw by replacing Le Roy's pivot detent with a spring detent. His unscrupulous employer leaked the innovation to John Arnold who patented it in 1775. Earnshaw had to set up on his own and wait for Arnold's patent to expire before he could announce further improvements to the escapement. His eventual design, developed with sponsorship from the Board of Longitude, was incorporated into virtually all marine chronometers until the 1970s.

Pocket watches were still two centimeters thick in the mid-18th century; cumbersome for men and unsuitable for women, who had no pockets in which to put them. In 1764 French watchmaker Jean-Antoine Lépine devised the 'Lépine calibre', which was a way of assembling the components that allowed a watch to be far thinner; thin enough to be worn on the wrist, which was aristocratic women's preference, as a form of jewellery. Moreover, the Lépine calibre was adaptable to mass production. When affordable consumer pocket watches and wristwatches arrived in the 19th century, they were mass produced using the Lépine calibre layout. Indeed, most watches still use the Lépine calibre, even though they no longer have a mainspring or winding mechanism.

None of the existing escapements were ideal for use in a thin Lépine calibre watch. Conversely, Lépine calibre watches were not as thin as they could be because of their bulky escapements. In 1795, Thomas Tompion invented the low profile 'cylinder' escapement specifically for thin Lépine calibre watches. Like the deadbeat, it was perfected by his protégé George Graham. Cylinder escapement watches could be just a few millimetres thick, like a modern watch, but they suffered from excessive wear. Graham tried to extend their durability by making the cylinders from rubies, but this proved to be prohibitively expensive. In the mid-19th century, French and Swiss watchmakers solved the problem by making the cylinders from cheap carbon steel. They sold in large numbers for the next 50 years. 

Affordable accurate watches were elite status symbols during the 18th century. By affordable, we do not mean for anyone that had to work for a living; we mean something that cost less than a house. The mass markets were virtually untapped. Accurate pocket watches cost more than a theatre; the only other pocket watches cost more than a house and were not terribly accurate. Robert Hooke solved part of the problem in 1700 with his 'duplex' escapement. It was accurate enough for personal use, reasonably robust and no more expensive than the cylinder escapement. His design was improved by French watchmakers and then by English watchmaker Thomas Tyrer, who patented his design in 1782. Tyrer's duplex escapement was popular in high quality pocket watches from 1790 until the mid-1800s.

There were still no consumer watches until 1750, when George Graham's protégé Thomas Mudge invented the 'lever' escapement (see Volker Vyskocil animation above), the single most important horology invention since Wallingford. The lever from which it gets its name is a separate component with pallets that are shaped like those of the deadbeat. It gets flicked side to side by a lug on the balance wheel. Each flip releases the escape wheel to advance one tooth. The escape wheel gives the lever pallet a tiny nudge at the end of each movement. The other end of the lever transmits the nudge back to the balance wheel to keep it oscillating. The escapement is simple, reliable, cheap to manufacture and almost frictionless which makes it durable and accurate. The lever escapement gradually replaced all the alternatives, except in chronometers, tower clocks, grandfather clocks and prestige watches. Today, the lever escapement still has more than 90% of the timepiece market.

Tower clock escapements proved to be particuarly troublesome, thanks to wind, rain, ice, birds and heavy hands. Thomas Mudge invented the 'gravity' escapement in the mid-18th century to solve these issues. Versions of it have been adopted by almost every tower clock built since and it has been retro-fitted to many others. The clock that drives Big Ben has a version of the gravity escapement that is snappily known as the 'double three-legged gravity escapement'. It was invented specifically for the purpose by Edmund Beckett Denison. French tower clocks are an exception. During the 18th and 19th centuries they generally used a different escapement that had been developed in France.

Other timepiece innovations

Although pendulum or balance spring wheel clocks and watches with post-verge escapements were fairly accurate in ideal conditions, they deteriorated over time and were less accurate in less ideal conditions. Some inaccuracies could be compensated with a regulator. French Thuret watches had the first useful regulators. Their design was adopted and improved upon by Thomas Tompion in 1680. His design was further improved by English watchmaker Joseph Bosley in 1755. Bosley regulators are still used in modern watches.

John 'Longitide' Harrison solved three other sources of inaccuracy in H4: 1) That a spring's modulus of elasticity changes with temperature, which affected its accuracy; 2) That springs tend to lose elasticity as they work harden; and 3) That clocks stopped when they were being wound. All three issues were deal-breakers for John Harrison because the Longitude Prize needed a marine chronometer that remained accurate on long voyages that passed through tropical seas.

Harrison's solution for temperature compensation was to invent the bimetallic strip, which he used to adjust the length of the balance spring in H4. It worked but proved too fiddly and too expensive for subsequent chronometers. Pierre Le Roy improved the design by moving the bimetallic strip to the rim of the balance wheel. In 1805 Thomas Earnshaw, sponsored by the Board of Longitude, improved Le Roy's design. Earnshaw's design was used in all marine chronometers until the 20th century when some switched to 'elinvar', an alloy unaffected by temperature that was discovered by Nobel prize winning Swiss physicist Charles Guillaume.

His solution for work hardening in H4 was to use purer iron that was made into flexible steel by repeated quenching and tempering. This technique is probably his most important horological legacy. It is his only invention that was widely used in timepieces after his death. Indeed, his technique for making watch springs is still used in the vast majority of watches. A few top end manufactuers have switched to semiconductor materials.

His solution for the winding issue was a new type of 'maintaining power'. It not only worked perfectly but became standard in all chronometers and many top quality watches through to the electronic era.

Harrison's maintaining power did not work terribly well in tower clocks. They preferred an earlier system invented by Swiss clockmaker Jost Burgi around the turn of the 16th century. It is now known as the 'gravity train remontoire'. It was developed into its modern form, like so much else, by Huygens in the late 17th century.

Another important advance was the use of jewel bearings to reduce wear and friction, especially on the pallets. The technique was developed in 1702 by London-based Swiss inventor Nicolas Fatio de Duillier and his English watchmaking colleagues Pierre and Jacob Debaufre. Jewel bearings enormously increased the longevity and later life accuracy of watches. They are still used in top end watches today. The development of jewel bearings in London gave British watchmakers a huge competitive advantage for the next fifty years.

Finally, there was isochronism. We mention above that nearly all early watches used a fusée or stackfreed to compensate for the mainspring's loss of torque as it unwinds. They both worked, but could not be precision made for high accuracy. They also added an extra source of complexity and an extra component to go wrong. In 1795, the great Swiss watchmaker Abraham-Louis Breguet permanently solved the problem with his amazingly simple and elegant 'Breguet overcoil'. It has been used in virtually every spring balance wheel timepiece for the last 200 years.