How Ancient People Told Time: Shadows, Sundials and Water Clocks

A sundial from the workers' huts#

In 2013 a Swiss team from the University of Basel announced a find from Egypt's Valley of the Kings: a flat piece of limestone. On it someone had painted a black half circle, its straight edge about 16 centimeters (6 inches) long. The half circle was divided into twelve sections of about 15 degrees each, and twelve times 15 is 180, exactly half a circle (Bickel and Gautschy, University of Basel).

A small dent in the middle of the straight edge would have held a peg of wood or metal. As the sun crossed the sky, its shadow would have swept over the sections one by one. According to the published study, the stone lay among the ruins of the huts of workmen who built the royal tombs, between tombs KV 29 and KV 61, although early press reports said it came from a tomb entrance. The team dated it to the 19th Dynasty, in the 13th century BCE (Live Science).

Nobody knows what it was for. The excavation leader, Susanne Bickel, wondered whether it helped fix the workmen's hours or breaks, and the published study agrees that its precision would have suited that job. The team also raised a religious reading: that the half circle stood for the sun god's journey through the underworld.

Either way, it points at a problem that people in Egypt, Mesopotamia, Greece, China and the Islamic world worked on for thousands of years. A day has no marks on it. Sunrise and sunset need no instrument, but agreeing on when things happen does: when the temple rite begins, when the shift ends, when a speech in court must stop. One review of Chinese water clocks lists their early jobs as astronomy, religious sacrifices, military affairs, timing court cases and sharing out water rights (Hwang, Yan and Lin, 2021).

This is the first of five stories in the History of Time series. It follows the oldest answers to that problem, up to the edge of the mechanical clock, where Bells, Gears and Pendulums: How the Mechanical Clock Changed Daily Life picks up the thread.

Shadows first: sticks, shadow clocks and sundials#

The simplest clock is a stick pushed into the ground. Its shadow is long in the morning, shortest at midday and long again in the evening. The Greek word for the stick, gnomon, still names the shadow-casting part of a sundial, and a sundial has three parts: the gnomon, a base plate, and hour lines marked on the plate (Britannica).

Nobody can say who first used a stick this way, because sticks leave no trace. Among the earliest instruments that survive are Egyptian shadow clocks, usually dated to around 1500 BCE. A shadow clock in Berlin is tied to the reign of Thutmose III, which Berlin's museum records date to about 1479 to 1425 BCE, and the History of Science Museum at Oxford calls the oldest known sundial one from his reign (Archaeoastronomy and Ancient Technologies, Oxford). That makes it the oldest one that survives, not proof that nothing older existed. The Science Museum in London keeps a plaster copy of a Berlin shadow clock whose base carries a scale of hours; its record says the crosspiece that cast the shadow is missing from the original (Science Museum Group).

Shadows change with the seasons, so a useful instrument had to allow for them. A portable shadow clock from the Ptolemaic period at the Metropolitan Museum has six parallel lines on its face, and each line served two months of the year (Metropolitan Museum of Art).

Babylonian astronomers kept time too. The astronomical compendium called MUL.APIN, whose oldest surviving copies date from the 7th century BCE (the text itself is thought to be older), includes a section on a water clock (Brack-Bernsen, Centaurus). The Greek historian Herodotus says the Greeks learned the sundial, the gnomon and the division of the day into twelve parts from the Babylonians (Herodotus 2.109). Modern scholars read that sentence with care. The classicist Sofie Remijsen calls the passage problematic and offers a new reading of it. She argues that an hour-like unit was worked out by Greek astronomers in the late fifth century BCE under Babylonian influence, and that it moved into everyday civic life only in the second half of the fourth century (Remijsen, Klio).

A shadow also depends on where you stand. For a sundial to show clock time, its gnomon must be tilted to match the latitude (Britannica). The Roman writer Pliny tells a story about that. Around 263 BCE a sundial taken as war booty from Catania in Sicily was set up on a column in the Roman Forum (Smith's Dictionary, Perseus). It had been made for a more southerly latitude, so it never showed the hours correctly in Rome, yet Pliny says the Romans went by it for 99 years until a better one was built (Stuart, Antiquities of Athens (1825), p. 42, quoting Pliny, Heidelberg University Library). Treat 263 BCE as approximate: most sources give that year, and at least one gives 264 BCE.

Egypt's twelve and twelve, and the hour that stretched#

The Egyptian word for "hour" was wnwt. According to a McMaster University analysis, it appears as early as the Pyramid Texts, spelled with a star sign. The earliest known set of those texts is in the pyramid of Unas, the last king of the 5th Dynasty (Britannica), and the oldest texts are usually dated to roughly 2400 to 2300 BCE. The authors take that as a hint that the hour began as a unit of the night (McMaster University).

Around 2000 BCE, wooden coffin lids from Asyut carry star tables. Each column covers a ten-day stretch of the year and lists 12 stars, which the analysis calls the earliest systematic division of the night into 12 parts. The word wnwt is not yet attached to them. That link is made explicit at the Osireion in Abydos, a temple of the 13th century BCE, where the 12 rows of a star table are labeled wnwt. The McMaster authors date that table to about 1210 BCE.

By the New Kingdom, Egypt counted 12 night hours and 12 day hours. Daylight hours were read from shadows and night hours mainly from stars, and the two sets were counted separately. The McMaster authors argue that the day hours and night hours were always given separately, never as one 24-hour day, and that the twilight periods around sunrise and sunset fell outside both counts. That is their reading of the evidence.

Stars needed their own instruments. A merkhet in the Science Museum's collection, dated to about 600 BCE, belonged to a man named Bes, an astronomer-priest of Horus at Edfu. Its catalog entry says it was used to find the hour by the sun in the day and by watching chosen stars cross the north-south line at night (Science Museum Group).

The key point is what "twelve" meant. If daylight is divided into 12 equal parts, a longer day gives longer hours. Hours defined this way are called seasonal or unequal hours, and they were in use in Egypt from at least the second millennium BCE (Remijsen, Klio). Noon always fell exactly between the sixth and seventh hour, but everything around it moved.

• 12 + 12hours of daylight and of night in the Egyptian system by the New Kingdom, counted separately
• 53 to 69 minuteslength of one daylight hour at Karnak's latitude, from the winter to the summer solstice (calculated for this article)
• 46 to 76 minutesthe same swing in Rome
• 39 to 83 minutesand in London, from December to June

Seen that way, a short winter hour is not a faulty clock. It is an even share of a short day.

Where 60 and 24 came from#

The 60 in our hour has a different home: Mesopotamia. Babylonian astronomers counted in base 60, a habit they inherited from the Sumerians. One study of Sumerian tablets finds scribes calculating in base 60 by the 26th century BCE (arXiv), and the place-value system built on it is dated to around 2000 BCE or the 19th century BCE, depending on the source. Nobody knows why 60 was chosen, but it is a handy number. It is the smallest number that divides evenly by 1, 2, 3, 4, 5 and 6, and 12 whole numbers divide it exactly, against only four for 10 (NIST-hosted Scientific American answer).

The Babylonians also had their own time unit, the beru, about two of our hours, used at least from the seventh century BCE to say how much time had passed since sunrise or sunset (Remijsen, Klio).

The base-60 habit reached us through Greek astronomers. Ptolemy, writing in the second century CE, divided each degree of a circle into 60 parts and each part into 60 smaller ones. The Latin names, which came later, were pars minuta prima, the "first small part", and secunda pars minuta, the "second small part", and from these come "minute" and "second" (Saturday Evening Post, Online Etymology Dictionary). At first they divided angles, not time, and minutes and seconds did not enter everyday timekeeping until many centuries later.

The 24 has a messier history. Equal time units were older than that, as the beru shows. Hipparchus, who worked roughly between 147 and 127 BCE, is credited with proposing a division of the day into 24 equal "equinoctial" hours, based on the equal daylight and darkness of the equinox. Ordinary people kept using seasonal hours for centuries (NIST-hosted Scientific American answer). The credit is not settled: Wikipedia's article on equinoctial hours says equal hours were already used for the full day in Babylonian astronomy about 2,400 years ago (Wikipedia). Remijsen argues that Greco-Roman clock time took its final form only after it absorbed Egyptian ideas of the hour in the Ptolemaic kingdom, in the early third century BCE.

Water clocks: telling time in the dark and in court#

A sundial fails at night, indoors and under cloud. Water does not. The simplest water clock, called a clepsydra (Greek for "water thief"), is a vessel with a small hole near the bottom. As the water leaks out, its level falls past marks on the inside (Britannica).

The oldest surviving water clock is usually said to be a vessel found broken at Karnak and now in Cairo. It dates to the reign of Amenhotep III, which museum records place around 1400 BCE, roughly 3,400 years ago. Their exact years differ by a few decades (McMaster Ancient Egyptian Astronomy database). The vessel is made of alabaster and was probably for the hours of the night. Because those hours changed length, its inside carries twelve scales, each marked with the name of a month (Science Museum Group). Whether priests used it in daily ritual, or whether it was mainly a valuable gift to the temple, is an open question (Gautschy).

A tomb text points to an older origin. The autobiography of a court official named Amenemhet, who served under Amenhotep I around 1500 BCE, has been read as crediting him with inventing the water clock (Cotterell, Dickson and Kamminga, 1986; abstract on an author page). The tomb was found in 1885 and is now lost, and a 2016 study by Alexander von Lieven argues that the text describes a more elaborate mechanical device with moving figures (von Lieven, 2016; abstract). The 1986 study suggests Egyptian water clocks had the potential to be accurate to within about 15 minutes a day, a best case that other writers doubt, and a NIST-hosted magazine article repeats the figure for ancient water clocks (NIST).

In Athens the water clock became a courtroom tool. Each side in a trial received a fixed speaking time measured with a water clock, and Aristotle's Constitution of the Athenians gives the allotments in a unit of volume called the chous (Aristotle, The Stoa). A clay vessel from the Athenian marketplace, the Agora, has a spout near the bottom and a hole that let it be filled to the same level each time. It matches what the texts describe for the courts from the late fifth century BCE (Young, Hesperia). One Agora water clock, whose inscription appears to name the tribe Antiochis, held about two choes, roughly 6.4 liters, and ran for about six minutes (Cornell College). One scholarly note adds that pots marked this way may have timed speeches in the council rather than in the law courts, so read it as an example of the vessel type rather than as proof of courtroom practice.

The Greeks also improved the design. A simple outflow clock slows as the vessel empties. Vitruvius says Ctesibius of Alexandria, whom Britannica dates to about 270 BCE (Britannica), was the first to investigate water clocks, and the History of Science Museum at Oxford credits him, on Vitruvius's authority, with the inflow clock, in which an overflowing upper vessel keeps the level, and so the flow, steady (Oxford, Vitruvius). A large public water clock stood inside the Tower of the Winds in Athens, an octagonal marble tower with sundials on its walls. The tower takes its name from the eight wind figures carved on its frieze, and the astronomer Andronicus of Cyrrhus is credited with it. Its date is disputed between the second century and the early first century BCE (World History Encyclopedia).

Romans used water too. Caesar writes that his inquiries turned up nothing about a reported island where winter nights lasted thirty days, but that accurate measurements with water showed nights in Britain were shorter than on the continent (Gallic War 5.13). A satire traditionally ascribed to Seneca jokes that it is easier to get philosophers to agree than clocks (Apocolocyntosis). The joke hints at a problem that would last for centuries: clocks that disagree.

Towers and elephants: China and the Islamic world#

Chinese tradition places the invention of the water clock in the time of the legendary Yellow Emperor (China Science and Technology Museum). A state science newspaper dates the earliest written mention to the 6th century BCE and says that from around 200 BCE the outflow type was widely replaced by an inflow clock with a float and an indicator rod (Science and Technology Daily). The 2021 review places the float and the feedback idea in the 3rd century BCE. In the 8th century, it says, China began building water-powered mechanical clocks driven by waterwheels (Hwang, Yan and Lin). The monk Yi Xing and the official Liang Lingzan built a water-powered celestial globe in that era (Mechanical Sciences).

The best-documented example comes from the Song dynasty, because Su Song's illustrated book survives. In the capital, Bianjing (today's Kaifeng), the scholar-official Su Song directed the building of a clock tower. Britannica describes an armillary clock, an astronomical clock built around a model of the heavens made of metal rings, in a tower about 35 feet (11 meters) tall, powered by a waterwheel and a chain drive (Britannica). The tower had three levels: an armillary sphere on top, a celestial globe in the middle, and a pagoda-like display below where figures rang a bell and struck a gong to mark the hours (National Museum of Natural Science, Taiwan).

The heart of it was a big wheel with scoops. Water ran into one scoop at a time; when a scoop was full, the wheel was released to advance one step, and the next scoop moved under the spout (Revolution Watch). That step-by-step release is the job of an escapement, the part of a mechanical clock that lets its drive move in regular steps. The idea was not new with Su Song: accounts trace the escapement tradition back to the monk Yi Xing and the official Liang Lingzan, whose earlier device is dated to 725 CE in some sources.

Sources give different dates. One places the start of the project in 1088 and its completion in 1092 (Taiwan Panorama). Another gives a small wooden model in 1088, bronze parts by 1090 and the finished tower in 1094 (Instituto Português de Relojoaria), and Wikipedia notes that sources put the construction anywhere from 1086 to 1092 (Wikipedia). Su Song's illustrated book on the tower is usually said to have been written in 1092 and officially printed in 1094, which is why 1092 is sometimes mistaken for the date of the tower itself. The tower was dismantled in 1127, when the Jin army took the city; one account says the parts were carried north but could not be put back together, though the details differ between accounts (INCOIS).

The book outlived the machine. In the 1990s builders in Taichung, Taiwan, used it to make a working replica, which the National Museum of Natural Science describes as a true-to-scale working model of the tower.

Further west, the engineer al-Jazari completed a book in 1206 describing about fifty mechanical devices, though counts vary by source (Muslim Heritage). One of his best-known devices is a water clock built as an elephant. A page from a copy made in 1315, now at the Metropolitan Museum, shows what happens every half hour: a bird on the dome whistles, a figure drops a ball into a dragon's mouth, and the driver strikes the elephant (Metropolitan Museum of Art).

The caliph Harun al-Rashid is recorded as sending Charlemagne a water clock. The Frankish annals describe such a clock under the year 807, in which a falling ball struck a cymbal and twelve horsemen appeared at twelve windows (Muslim Heritage). Later tales about the clock's fate are best treated as legend (Ithra).

Candles, incense and sand#

Not every timer needed water. A biography of King Alfred of Wessex, attributed to his contemporary Asser, says Alfred had six candles made, each 12 inches tall and marked in inches. Each inch took about 20 minutes to burn, so one candle lasted four hours and the six together lasted a full day and night. A lantern of wood and thin horn kept drafts from making them burn unevenly (Thijs Porck).

In China, incense kept time. A trail of powdered incense laid in a winding groove or a stenciled pattern burned at a steady pace, and the position of the glow told the time. Sources trace incense clocks to around the 6th century CE and say they flourished in the Song era, but differ on where the idea began (JSTOR Daily, Wikipedia).

The hourglass has a shorter paper trail. The earliest known European picture is a 1338 fresco by Ambrogio Lorenzetti in Siena, where a figure of Temperance holds one. An early written reference to a sandglass used at sea is a receipt of about 1345 for the English king's ship La George, which records glass timers bought in Flanders. Wikipedia calls it the earliest reference that can be said with certainty to a marine sandglass, and Encyclopedia.com also mentions the receipt (Wikipedia, Encyclopedia.com). The old story that a monk named Luitprand invented the hourglass in the 8th century has no evidence behind it (Guinness World Records). Britannica's clock summary lists sandglasses among the devices of ancient Egypt, Greece and Rome, so sources disagree on how early they existed (Britannica).

All of these are timers rather than clocks. They measure how much time has passed since you lit the candle or turned the glass. Remijsen draws the same line for the first Greek water clocks, which measured the length of a period instead of letting people pin a moment onto the day.

From shadows to gears#

Before mechanical clocks spread, there were many ways to measure time, and each had a catch. Sundials needed sun and a gnomon matched to the latitude. Water clocks needed a steady flow. Candles needed shelter from drafts, and an hourglass needed someone to turn it. In daily life the hour still stretched and shrank with the seasons.

Then a different kind of clock arrived, driven by weights and an escapement. Britannica says the first mechanical clocks to which clear references exist were large weight-driven machines fitted into towers, and its summary dates the first European public clock that struck the hours to Milan in 1335 (Britannica clock article, Britannica clock summary). Other accounts of the Milan clock, on the tower of San Gottardo, put its completion in 1336 (Lapham's Quarterly), so treat the year as 1335 or 1336. Hours of fixed length became commonplace only after such clocks arrived (NIST-hosted Scientific American answer). Part 2 follows them, including why historians trace the first escapement clocks to the late 1200s: the bells, the gears and the pendulum that made them accurate. Later parts take the story to the ship's chronometer (The Longitude Problem: How a Clock Taught Ships Where They Were), the railway timetable (How Time Zones Were Born: Railways, Telegraphs and the Day of Two Noons) and the atomic clock (How Your Phone Knows the Time: Atomic Clocks, Leap Seconds and the Network Behind Them).

1. About 2000 BCECoffin lids from Asyut carry star tables that divide the night into 12 parts.
2. About 1500 BCEEgyptian shadow clocks are in use; a tomb text of this period has been read as crediting Amenemhet with the water clock.
3. About 1400 BCEThe Karnak water clock is made in the reign of Amenhotep III (dates vary by source).
4. 13th century BCEThe Valley of the Kings sundial is made (reported in 2013).
5. About 1210 BCEThe McMaster authors date a star table at the Osireion in Abydos, whose rows are labeled with the word for "hour", to about this time.
6. About 700 BCEMUL.APIN copies include a water clock section.
7. Late 5th century BCEWater clocks are in common use in the Athenian law courts.
8. Around the 3rd century BCECtesibius of Alexandria improves the water clock (dating varies).
9. About 147 to 127 BCEHipparchus is active and is credited with proposing 24 equal hours.
10. 8th century CEYi Xing and Liang Lingzan build a water-driven celestial globe in Tang China.
11. About 1088 to 1094Su Song's water-driven clock tower is built in Kaifeng (dates vary).
12. 1206Al-Jazari completes his book of mechanical devices, including the elephant clock.
13. 1338Lorenzetti paints the earliest known European picture of an hourglass.

Frequently asked questions

Sources and further reading

1. Analysis: Ancient Egyptians measured the first hour, and changed how we related to time — McMaster University
2. Eine ramessidische Sonnenuhr im Tal der Könige — University of Basel, Zeitschrift für Ägyptische Sprache und Altertumskunde
3. Egyptian Water Clock — Science Museum Group
4. WCO 1 Amenhotep III Karnak — McMaster University, Ancient Egyptian Astronomy database
5. Living by the Clock: The Introduction of Clock Time in the Greek World — Sofie Remijsen, Klio, University of Amsterdam repository
6. Why is a minute divided into 60 seconds, an hour into 60 minutes, yet there are only 24 hours in a day? — NIST, hosting a Scientific American answer
7. An Athenian Clepsydra — American School of Classical Studies at Athens
8. Inflow water clock — History of Science Museum, University of Oxford
9. Historical development of water-powered mechanical clocks — Mechanical Sciences
10. Su Song — Encyclopaedia Britannica
11. Ctesibius Of Alexandria — Encyclopaedia Britannica
12. Sexagesimal Calculations in Ancient Sumer — arXiv
13. Water-powered Armillary Sphere and Celestial Globe Tower — National Museum of Natural Science, Taichung
14. The Elephant Clock, folio from al-Jazari's Book of the Knowledge of Ingenious Mechanical Devices — The Metropolitan Museum of Art