Bells, Gears and Pendulums: How the Mechanical Clock Changed Daily Life

Inside Salisbury Cathedral stands an iron-framed machine with no face and no hands. It cannot show you the time. It was built to count the hours and strike a bell, using two sets of gears, one for counting and one for striking, each driven by a falling weight that has to be wound up every day. The clock is thought to date from about 1386, and it is often called the oldest working mechanical clock in the world, a title other clocks contest.

That faceless machine is a good place to start the second part of our series on the history of time. In part one, people read the hour from the sky and from flowing water. The Salisbury clock does something new. It makes the hours by itself, in iron, at any time of day or night, and announces them to everyone within earshot. This part follows that idea from monastery bells to the pendulum clock of 1656.

Why the first clocks were bells, not faces#

Ask where the word clock comes from and the answer is a sound. English clock traces back to medieval Latin clocca, meaning bell, the same root as French cloche and Dutch klok. The Museum of the History of Science at Oxford, in its account of the mechanical clock's origin, explains the link: the earliest clocks announced time not by moving a pointer round a dial but by sounding a bell.

The bell mattered because of the monastic day. Monks prayed at set times by day and night, the canonical hours, and a community had to be gathered at the right moment. Oxford's museum points out that keeping that cycle in step needed a timekeeper that worked at night and in all weathers, and that freezing northern winters limited how useful water clocks could be. Britannica adds that the first all-mechanical clocks may have been used in monasteries to toll a bell calling monks to prayer, though their origin is unknown.

The earliest clocks that sources describe clearly were large machines driven by falling weights and fitted into towers. Britannica says they struck only the hours and had no hands or dial, exactly as at Salisbury. For centuries a clock was something you heard before it was something you looked at.

The escapement: a brake that ticks#

A falling weight makes a poor clock on its own. Hang it from a rope wound round a drum and it will spin the gears in a rush until it hits the floor. To keep time you need something that lets the gears move forward in small, even steps. That part is the escapement: it repeatedly catches the gear train and lets it go, one tooth at a time. The tick of a clock is the sound of the catching.

The classic early escapement was the verge and foliot. Britannica describes it this way: a weight turns a crown wheel (a wheel whose teeth stand up like the points of a crown), a pair of small flags called pallets on an upright rod, the verge, lets one tooth slip past at a time, and a horizontal bar with weights on it, the foliot, swings back and forth to control the pace.

The foliot has a weakness. It has no natural rhythm of its own, so its speed depends on how hard the weight pulls and how much the parts rub. A physics analysis of the mechanism calls it a rather bad timekeeper and puts typical error at about 15 minutes a day. Estimates for the early tower clocks of the 1300s vary widely, and Britannica is the source usually cited for an error of up to about half an hour.

Who invented it, and when?#

Nobody knows. One clue is a remark in a commentary written in 1271 by the English astronomer Robertus Anglicus, who wrote that clockmakers were trying to make a wheel that turns once a day but could not quite perfect their work. Some writers read this as a sign that the escapement had not yet been solved; others point out that the passage never names an escapement, so historians read it in different ways. The stronger evidence is indirect: a sudden rise in the number and cost of clocks in the late 1200s. Sources still differ. Britannica says the verge was probably invented in 13th-century Europe, while Oxford's museum puts it in the early 14th century.

The oldest surviving clocks usually cited come from the 1380s and 1390s. Guinness World Records lists Salisbury's as the oldest surviving working clock, dated to 1386 "or possibly earlier". Britannica reports a movement from 1389 at Rouen. The Wells Cathedral mechanism, dated about 1390 by the Science Museum in London and 1392 in its collection database, may have come from the same makers as Salisbury's. It is now in the Science Museum in London.

Towns buy the hour#

Before mechanical clocks, an hour was not a fixed length. Day and night were each divided into twelve hours, a system How Ancient People Told Time: Shadows, Sundials and Water Clocks traces back to Egypt, so a daytime hour in summer was much longer than one in winter. At Salisbury's latitude, about 51 degrees north, our own calculation gives a midsummer daytime hour of about 83 minutes and a midwinter one of about 40. A clock ticks at one steady pace, so towns that bought clocks began to count hours of equal length all year. Equal hours did not take over overnight: historians describe the two systems running side by side for a long time, with the old hours lingering longest in church life.

Italy shows the idea at its boldest. The Milan chronicler Galvano Fiamma described a clock whose bell struck according to the twenty-four hours of the day and night. Accounts date his description to 1335 or 1336, with 1336 the date we found most often, so call it the mid-1330s. Counting straight through to twenty-four is an old habit, and our guide to the 24-hour clock shows how it survives. In 1344, according to secondary accounts, a public clock went up in a tower in Padua under the supervision of a physician and astronomer named Jacopo Dondi. Accounts of what its dial showed vary, so we leave that out.

Who wanted these machines? Not merchants alone, if the historian Gerhard Dohrn-van Rossum is right. A Bryn Mawr Classical Review notice of his German study of the hour, Die Geschichte der Stunde, reports his view that merchants were not at the forefront of introducing clocks and that churchmen adopted modern time reckoning early. Whatever the motive, a bell can be heard by everyone in a town at once, so the hour stopped being something each person estimated and became something a whole community shared.

Ambition in brass: Giovanni Dondi's astrarium#

If a striking clock was a marvel, Jacopo's son Giovanni set out to build a wonder. Oxford's Museum of the History of Science dates his work on an astrarium, a clock that showed the heavens, to 1348 to 1364, about sixteen years. The dating is not settled: an Eton College catalogue entry puts the building between 1365 and 1381, and Google Arts & Culture notes a view that moves the whole project later. It had seven faces, one each for the Sun, the Moon and the five planets known at the time, and 107 moving parts, all driven by a single weight. It also showed religious feast days.

The machine itself is lost. One suggestion is that it was destroyed in the sack of Mantua in 1630. What survives is Giovanni's detailed written description, the Tractatus astrarii, and that text has been enough to let modern craftsmen rebuild it. Oxford's museum records a reconstruction made by Luigi Pippa in 1961 to 1963, one of several. The early clock was never just a bell machine: some makers wanted it to be a model of the sky.

Springs, pockets and the myth of the Nuremberg egg#

A weight-driven clock has to stand still, with room for the weight to fall. Replace the weight with a coiled spring and the clock can shrink and move. Britannica says clockmakers, probably in southern Germany or northern Italy, were making small spring-driven clocks by about 1450. The oldest surviving one is often said to be the Burgundy Clock, now in Nuremberg's Germanisches Nationalmuseum, linked to Philip the Good of Burgundy and dated to about 1430 since an 1877 treatise by Maximilien de Leber. We could not check that date against the museum's own catalogue, so treat it as traditional. Oxford's museum names a table clock of about 1435 made for the same duke as the first known clock with a fusee, a cone-shaped pulley that evens out the pull of a spring as it unwinds. The sources checked here do not say whether that is the same clock. Early spring clocks were less accurate than weight-driven ones, but they could be carried.

This is where Peter Henlein of Nuremberg usually enters, as the man who invented the pocket watch around 1505 to 1510. The truth is thinner.

Galileo and the pendulum: what is documented#

The story everyone knows runs like this. A young Galileo sits in the cathedral at Pisa, watches a lamp swing on its chain, times the swings against his own pulse, and notices that wide swings and narrow ones take the same time. Here is what can be traced.

The source is Vincenzo Viviani, Galileo's pupil and first biographer. Viviani wrote his account in 1654, as the Museo Galileo in Florence cites it, and it was printed only in 1717. That is roughly seven decades after the supposed event, which the museum places around 1583, when Galileo was a student.

What is documented is less romantic and more solid. The earliest surviving document in which Galileo discusses the regularity of the pendulum is a letter of 29 November 1602 to Guidobaldo del Monte, a patron and mathematician. At about the same time, the physician Santorio Santorio, who worked in Padua and Venice, was developing a pulse-timing device, the pulsilogium. Sources differ on the dates: a first description in print in 1603 (some say 1602), and a fuller account with illustrations in 1625 or 1626, in which he explained that a pendulum regulated it. So one of the earliest recorded practical uses of the pendulum was to time the pulse, not to watch a cathedral lamp.

One point is easy to miss. The idea that every swing takes the same time is only approximately true. For small swings it is very close, but by the standard correction a swing reaching 10 degrees each side takes about 0.2 percent longer than a tiny one, which over a day would cost a clock about 2.7 minutes. That matters later in this story.

Late in life, when he was blind, Galileo conceived of regulating a clock with a pendulum and described the plan to Viviani and to his son Vincenzo. Sources date the design anywhere from about 1637 to 1641, and a drawing survives in Florence. Galileo died in 1642 without seeing it built. The practical pendulum clock came from Christiaan Huygens, who was still a boy when Galileo died.

Huygens and the pendulum clock, 1656 to 1657#

Huygens, a Dutch mathematician and astronomer who lived from 1629 to 1695, needed accurate time for his astronomy. The Metropolitan Museum of Art's essay on 17th-century clocks says the principle was discovered by Galileo, but for the practical purposes of European clockmaking, the pendulum's development began with Huygens. His first design dates from toward the end of 1656.

Huygens did not build the clocks himself. Salomon Coster, a clockmaker in The Hague, did. Within a year Coster held exclusive rights in the Netherlands, and sources differ on whether Huygens held a patent too. Huygens published a short book, Horologium, in 1658, and the full theory in 1673 in Horologium Oscillatorium. The idea spread fast: John Fromanteel of London went to work in Coster's shop, and by November 1658 the Fromanteel family was advertising pendulum clocks in London. The first ones kept the old verge escapement, now controlling a short pendulum in place of a foliot.

How accurate was it?#

Here the sources do not agree, so here is the spread.

A drift of 15 minutes a day is an error of about 1 percent, and it would add up to about 91 hours over a year if nobody corrected it. A clock that is off by 15 seconds a day is wrong by about 0.017 percent, or about an hour and a half over a year. Against the 15-minute baseline, a minute a day is a fifteen-fold gain, 15 seconds is sixty-fold and 10 seconds is ninety-fold. The sources agree on the direction and on the leap, if not on the exact size.

• 15 minutesdaily error of verge-and-foliot clocks, quoted for the better clocks of the mid-1600s
• 15 secondsthe daily error most often quoted for early pendulum clocks, though some sources say a minute or more
• 107moving parts in Giovanni Dondi's astrarium, which Oxford's museum dates to 1348 to 1364
• 0.994 meterslength of a pendulum that takes two seconds for one full swing (our calculation)

The anchor escapement, the long case and the minute hand#

The next improvement was the anchor escapement. Its earliest dated examples are tower clocks of 1670 to 1671, though some accounts put the idea in the late 1650s. It lets a pendulum swing through a much smaller arc, only a few degrees according to a Seiko museum page, which is where Galileo's near-equal swings become nearly exactly equal. The Met says the anchor, paired with a pendulum of slightly more than 39 inches, became the standard. That is the seconds pendulum: it takes two seconds for a full swing, one second each way, and by our calculation it is 0.994 meters or 39.1 inches long. Such a pendulum needs a tall box, and the longcase clock, often called a grandfather clock, grew up around it.

Who invented the anchor is disputed. The British Museum notes that a turret clock supplied to Wadham College, Oxford, in early 1670 is probably by Joseph Knibb, and that William Clement made one for King's College, Cambridge, in 1671. The museum gives Knibb the strongest claim but says neither clock proves who invented it. Robert Hooke is often named, but the museum says he was not a clockmaker and finds no evidence that he worked on such an idea.

Better accuracy changed what clocks showed. Until the 17th century most clocks had only an hour hand, because they were not accurate enough to justify more. Minute hands came into regular use around 1690, according to several accounts, though a few earlier examples are mentioned. The Met describes a pendulum clock of about 1670 in Museum Boerhaave in Leiden, which by tradition belonged to Huygens, with a dial showing hours, minutes and seconds.

1. About 1300The first verge-and-foliot clocks appear, somewhere between the late 1200s and the early 1300s. Sources differ, and no inventor or exact date is known.
2. About 1335 to 1336Galvano Fiamma describes a Milan clock that strikes twenty-four times a day.
3. 1348 to 1364Giovanni Dondi builds his astrarium, by Oxford's dating; an Eton College catalogue says 1365 to 1381.
4. About 1386 to 1392The clocks at Salisbury, Rouen (1389) and Wells (about 1390) date from these years.
5. About 1430The Burgundy Clock, often said to be the oldest surviving spring-driven clock, is linked to Philip the Good. The date is traditional.
6. Late 1656Christiaan Huygens designs a pendulum clock.
7. 1670 to 1671The earliest dated anchor escapements, in tower clocks at Wadham College and King's College.
8. About 1690Minute hands come into regular use.

Try it: time a pendulum with a stopwatch#

You can test Galileo's and Huygens's physics at your kitchen table. Tie a small weight, such as a few washers or a key, to a string and hang it from a fixed point. Measure from the pivot to the middle of the weight. Pull the weight a few inches to one side, keeping it under about 10 degrees each way, and let go gently.

Start the stopwatch as you release the weight and stop it after twenty full swings, counting out and back as one. Then divide by twenty. Timing many swings helps because your reaction time, which the stopwatch's own FAQ says is usually a tenth of a second or more, is spread across all of them. These are our calculated values for standard gravity:

Quadrupling the length doubles the time. A length of 99.4 cm gives exactly 2.00 seconds, the seconds pendulum behind the tall clocks above. Gravity varies slightly from place to place, so expect small differences.

How the clock changed work and daily life#

It is tempting to say the clock created modern time discipline in one stroke. Historians are more careful.

The classic argument comes from Jacques Le Goff, whose 1960 essay "Merchant's Time and Church's Time in the Middle Ages" is collected in Time, Work, and Culture in the Middle Ages. In outline, church time was tied to prayer and natural rhythms, while merchants needed predictable schedules, and town clocks symbolized the second kind. Later scholars have argued the two outlooks were more entwined.

Dohrn-van Rossum's study complicates it from another side. As the Bryn Mawr review of the German original summarizes, he argues that monastic timekeeping was about the order of events, prayer following prayer, not about measuring equal units.

A study of the cloth towns of Flanders and Artois by the historian Peter Stabel adds a twist. It found that the working day was set by town authorities and announced by bells, with regulation growing more detailed in the late 1200s and early 1300s. Bells already told cloth workers when to start and stop, which suggests the mechanical clock joined an existing system rather than inventing one.

What can be said safely? Clocks spread through European towns across the 1300s, counted equal hours and made the hour audible to everyone at once. After 1656 they became accurate enough that minutes mattered.

The pendulum clock solved the problem of steady time on land. Huygens had pendulum clocks made for finding longitude at sea. They were tested at sea in the early 1660s and again on a Dutch East India Company voyage to the Cape that left in 1686. The results were contested: the glowing reports of the English captain Robert Holmes from 1663 and 1664 were later questioned, and doubts about the 1686 longitudes led the company to order a second trial in 1689. A pendulum depends on gravity pulling steadily on something that stays put, and a ship is not that. What replaced it is the story of the next part, the longitude problem. Later in the series, railways and telegraphs forced whole countries to agree on one time.

Frequently asked questions

Sources and further reading

1. Clock (Britannica, introduction and history) — Encyclopaedia Britannica
2. The mechanical clock: origin and purpose — Museum of the History of Science, University of Oxford
3. Key innovation: the escapement — Museum of the History of Science, University of Oxford
4. Toward miniaturisation: the Fusee — Museum of the History of Science, University of Oxford
5. Two historical clocks re-displayed (Wells Cathedral clock) — Science Museum Group
6. Wells Cathedral clock, 1392 (collection record) — Science Museum Group
7. Dover Castle turret clock (collection record) — Science Museum Group
8. Oldest clock — Guinness World Records
9. Striking clock — Wikipedia
10. Giovanni Dondi's Astrarium, 1364 — Museum of the History of Science, University of Oxford
11. Cathedral of Santa Maria (Pisa itinerary) — Museo Galileo, Florence
12. Pisa: following the path of Galileo — Museo Galileo, Florence
13. Galileo and the pendulum clock — Royal Holloway, University of London
14. European Clocks in the Seventeenth and Eighteenth Centuries — The Metropolitan Museum of Art
15. First operational pendulum clock — Guinness World Records
16. Dynamics of the verge and foliot clock escapement — arXiv
17. Verge escapement — Wikipedia
18. The invention of the pendulum clock — Seiko Museum
19. Christian Huygens (Scientist of the Day) — Linda Hall Library
20. British Museum collection record on the anchor escapement — The British Museum
21. Spherical Table Watch (Melanchthon's Watch) — The Walters Art Museum
22. The so-called Henlein pocket watch — Google Arts & Culture
23. Giovanni Dondi's astrarium (Eton College catalogue entry) — Eton College Library
24. Nuremberg eggs — Wikipedia
25. The Nuremberg Egg: Myth and Reality — Antique Pocket Watch
26. Time, Work, and Culture in the Middle Ages — University of Chicago Press
27. Peter Stabel on working hours in the Flemish and Artois cloth industry — Tijdschrift voor Sociale en Economische Geschiedenis
28. Review of Gerhard Dohrn-van Rossum, Die Geschichte der Stunde (German original of History of the Hour) — Bryn Mawr Classical Review